Discovery and optimization of (R)-prolinol-derived agonists of the Growth Hormone Secretagogue receptor (GHSR)

Article abstract of DOI:10.1016/j.bmcl.2008.07.120

The discovery and optimization of a novel series of prolinol-derived GHSR agonists is described. This series emerged from a 11,520-member solid-phase library targeting the GPCR protein superfamily, and the rapid optimization of low micromolar hits into single-digit nanomolar leads can be attributed to the solid-phase synthesis of matrix libraries, which revealed multiple non-additive structure-activity relationships. In addition, the separation of potent diastereomers highlighted the influence of the α-methyl stereochemistry of the phenoxyacetamide sidechain on GHSR activity.

Full text of DOI:10.1016/j.bmcl.2008.07.120

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5083–5086
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and optimization of (R)-prolinol-derived agonists of the Growth
Hormone Secretagogue receptor (GHSR)
Weixu Zhai a, Neil Flynn c, Daniel A. Longhi c, Joseph A. Tino d, Brian J. Murphy c, Dorothy Slusarchyk c,
David A. Gordon c, Anna Pendri a, Shuhao Shi a, Robert Stoffel e, Baoqing Ma b, Michael J. Sofia a,
Samuel W. Gerritz a,
*
a Bristol Myers Squibb Co., Early Discovery Chemistry, Wallingford, CT 06492, USA
b Bristol Myers Squibb Co., Analytical R&D, Wallingford, CT 06492, USA
c Bristol Myers Squibb Co., Metabolic Disease Research, Hopewell, NJ 08543, USA
d Bristol Myers Squibb Co., Discovery Chemistry, Hopewell, NJ 08543, USA
e Bristol Myers Squibb Co., Lead Discovery, Hopewell, NJ 08543, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery and optimization of a novel series of prolinol-derived GHSR agonists is described. This ser-
ies emerged from a 11,520-member solid-phase library targeting the GPCR protein superfamily, and the
rapid optimization of low micromolar hits into single-digit nanomolar leads can be attributed to the
solid-phase synthesis of matrix libraries, which revealed multiple non-additive structure–activity rela-
Received 16 June 2008
Revised 30 July 2008
Accepted 31 July 2008
Available online 3 August 2008
tionships. In addition, the separation of potent diastereomers highlighted the influence of the
stereochemistry of the phenoxyacetamide sidechain on GHSR activity.
a-methyl
Keywords:
Solid-phase library
Hit to lead
Ó 2008 Published by Elsevier Ltd.
Combinatorial chemistry
Non-additive SAR
Non-linear SAR
Growth hormone secretagogue receptor
GHSR
G-protein coupled receptor
GPCR
Solid-phase synthesis
Parallel synthesis
GPCR agonist
Matrix library
The Growth Hormone Secretagogue receptor (GHSR) has been
implicated in a number of important biological functions, including
the regulation of growth, metabolism, and food intake.1 Small mol-
ecule agonists of GHSR have been studied in the clinic2 for the
treatment of growth hormone disorders, frailty, and cachexia.
GHSR is a member of the G-protein-coupled receptor (GPCR) pro-
tein superfamily3, and as part of an in-house effort to identify novel
GHSR agonists4 we screened a series of GPCR-targeted libraries,
including one based on a secondary amine chemotype.5 This
library afforded a number of low micromolar GHSR agonist hits,
and the identification and optimization of these hits is described
herein.
of cyclic and acyclic secondary amino alcohols; R2, consisting of
primary amines and anilines; and R3, consisting of carboxylic acids.
The overall dimensions of the library were 11 R1 Â 31 R2 Â 42 R3,
which in theory would afford 14,322 discrete products. However,
each library sample underwent a rigorous quality control (QC) pro-
cess: each sample was analyzed by HPLC (with UV and ELS detec-
tion) and mass spectrometry. Samples which did not afford P70%
purity by HPLC and a mass ion corresponding to the desired prod-
uct were discarded. Upon completion of the QC process, 11,520
samples were submitted for screening against a variety of GPCR
targets and resulted in the identification of a series of prolinol-de-
rived GHSR agonists as described in Table 1.
Given that 11,520 library members were screened in the GHSR
assay, the high degree of structural homology for the active sam-
ples described in Table 1 was compelling. Indeed, a modest amount
of SAR could be elucidated from the HTS data, with the caveat that
the original library samples were submitted as crude (but P70%
The secondary amine GPCR-targeted library5 was synthesized
on solid support and comprised three reagent sets: R1, consisting
* Corresponding author. Tel.: +1 203 677 6884; fax: +1 203 677 6984.
E-mail address: samuel.gerritz@bms.com (S.W. Gerritz).
0960-894X/$ - see front matter Ó 2008 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2008.07.120

5084
W. Zhai et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5083–5086
Table 1
a-substituted phenoxyacetamide moieties described in Table 1
Initial hits from secondary amine GPCR library
were critical for activity. As a result, we initiated the synthesis of
follow-up libraries utilizing the solid-phase chemistry (outlined
in Scheme 1) which provided the original library5 Starting with
commercially available Wang7 linker-equipped polystyrene resin
1, the Wang hydroxyl group was converted to the corresponding
p-nitrophenyl carbonate 2 under standard conditions8 followed
by p-nitrophenol displacement with (R)-prolinol to afford solid-
supported prolinol carbamate 3. The primary hydroxyl group of 3
was oxidized to aldehyde 4 using pyridine–sulfur trioxide9 and
the aldehyde subsequently underwent reductive amination with
primary anilines (n = 0) and benzylamines (n = 1) under standard
conditions10 to afford secondary amine 5. It should be noted that
this seemingly innocuous reductive amination step does not pro-
ceed smoothly in solution, particularly with anilines. In a deviation
from the amide coupling conditions used in the original library
(carboxylic acid + coupling reagent), we elected to utilize acid chlo-
rides owing to their enhanced electrophilicity. Thus, treatment of 5
with the appropriate acid chloride and Hunig’s base afforded 6,
which was liberated from the solid support via treatment with tri-
fluoroacetic acid to provide the final product 7. All follow-up sam-
ples were purified by reverse-phase HPLC and characterized via LC/
MS (P95% pure by UV) and 1H NMR.
O
H
O
N
N
R2
n
R1
a,b
Entry
R1
n
GHSR EC50
(lM)
R2 = H
R2 = Me
R2 = Et
1
2
3
4
5
6
7
H
0
0
0
0
0
0
1
>10
2
N.a.
>10
N.a.
5
3
0.3
9
4
2
N.a.
4
14
>10
>10
2
p-AcNH
p-NCCH2
m-MeO
m-MeS
o-MeO-m-Ph
H
N.a.
N.a.
>10
0.9
a
See Ref. 6 for experimental details.
N.a., no sample (failed QC).
B
pure by HPLC) samples. Moreover, the 31 primary amines (30 ani-
lines and 1 benzylamine) used in the library spanned a wide range
of chemical property space and thus the R1 substituents listed in
Entries 1–6 of Table 1 do not constitute a focused library by any
stretch of the imagination. The p-AcNH substituent (Entry 2) was
the most active R1 substituent in the aniline series (n = 0), and it
appeared that a methyl group was the optimal R2 substituent,
although the R2 SAR was confounded by a significant number of
missing samples—these samples were not screened because they
failed our QC process and were discarded. In addition to the aniline
series exemplified by Entries 1–6, a benzylamine analog (n = 1)
afforded sub-micromolar GHSR agonist activity (Entry 7, R2 = Et).
Because only a single benzylamine reagent was present in the li-
brary, the absence of the R2 = Me sample (which was anticipated
to be more potent than the R2 = Et sample) served to heighten
our interest in the benzylamine series.
Through the synthesis of a number of small libraries (containing
<20 samples per iteration), we were able to rapidly elucidate the
key structure–activity relationships described in Figure 1, includ-
ing the observations that substitution at the prolinol nitrogen
abrogated GHSR activity, R stereochemistry was preferred at the
2-position of the pyrrolidine, and an a-methyl group was optimal
on the phenoxyacetamide sidechain. As a result, our SAR efforts fo-
cused on the two aromatic groups: the aniline (or benzylamine)
sidechain and the phenoxy group of the
mide sidechain.
a-methyl 2-phenoxyaceta-
At this point, we elected to synthesize a ‘matrix’ library, in
which the aniline/benzylamine and phenoxyacetamide substitu-
ents were varied simultaneously. This library was straightforward
to synthesize on solid support, as the use of IRORI MicroKansTM and
radiofrequency tags facilitated a ‘split-mix’ synthesis protocol11
We synthesized a library comprising approximately 40 different
Analysis of the GHSR screening data for inactive library mem-
bers indicated that the prolinol-derived secondary amine and
OH
NO2
NO2
O
O
OH
O
OH
HN
Cl
O
O
N
O
O
OH
Wang
Wang
Wang
DCM/NMM
20 h, RT
Et3N/DCM
20 h, RT
=
O
3
2
1
O
O
Cl
O
N
Py.SO3
Et3N
CHO
NH2
Me
R1
O
n
NH
n
R2
O
O
N
DMSO/DCM
4 h, RT
Wang
R1
NaBH(OAc)3, AcOH
DMF/TMOF, 2d, RT
DIEA, DCM
2d, RT
Wang
4
5
O
O
R2
O
N
H
N
O
N
Me
O
50% TFA/DCM
1h , RT
O
Wang
N
n
Me
R2
n
R1
7
6
R1
Scheme 1. Solid-phase synthesis of the secondary amine library.

W. Zhai et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5083–5086
5085
2 (R1 = o-Me) and 3 (R1 = m-Me); in this case, similar potencies
were observed when R2 = H or R2 = o-Cl, but when R2 = m-Cl the
SAR diverged dramatically, R1 = o-Me (Entry 2) afforded a 3-nM full
agonist, whereas R1 = m-Me (Entry 3) provided a 120-nM weak
partial agonist.
(R) stereochemistry
is preferred.
O
Alkylation/acylation
is not tolerated.
H
N
O
N
As shown in Table 3, non-additive SAR was also observed in the
aniline series. For example, the most potent GHSR agonist when
R2 = o-Cl or R2 = m-Cl was Entry 3 (R1 = p-AcNH), affording full ago-
nists with EC50s of 50 and 30 nM, respectively. However, when
R2 = H, the most potent GHSR agonist was Entry 5 (R1 = m-Cl). It
is also noteworthy that even though a significant loss of activity
was observed when R2 = o-Cl—three analogs (Entries 1, 4, and 6)
were inactive and two others (Entries 2 and 5) were partial ago-
nists—Entry 3 (R1 = p-AcNH) was one of the most potent com-
pounds in the entire matrix.
Me
AcNH
Me is optimal.
Figure 1. SAR of the secondary amine and phenoxyacetamide sidechains.
Table 2
SAR matrix for the benzylamine series
Although the ramifications of non-additive SAR are context-
dependent, the knowledge that perturbations in one region of a
chemical series are not synergistic with perturbations in another
region of the same chemical series is clearly pertinent to any
medicinal chemistry effort. In our experience, non-additive SAR
is often observed with GPCR targets, possibly due to the inherent
O
H
N
O
N
Me
R2
R1
mobility of the membrane-spanning
a-helices which compose
the small molecule binding sites. The detection of non-additive
SAR is greatly facilitated by the synthesis of matrix libraries, an
observation which has been described in detail elsewhere.12
All of the aforementioned compounds were synthesized as dia-
Entry
R1
GHSR EC50 (% Control)a (nM)
R2 = H
R2 = o-Cl
R2 = m-Cl
1
2
3
4
5
6
H
90 (110)
40 (110)
80 (90)
150 (100)
50 (80)
20 (60)
140 (90)
90 (70)
260 (60)
10 (90)
60 (80)
7 (80)
o-Me
m-Me
o-Cl
m-Cl
p-Cl
3 (100)
120 (20)
40 (80)
20 (70)
30 (90)
stereomeric mixtures because the
a-methyl phenoxyacetyl chlo-
rides used in the library synthesis were commercially available
only as racemates or were derived from racemic carboxylic acid
precursors. We were interested in determining the influence of
220 (80)
the
a-methyl stereocenter on the potency and efficacy of these
a
See Ref. 6 for experimental details.
GHSR agonists, and so we performed a chromatographic separation
of compound 8 to afford optically pure diastereomers 8a and 8b, as
shown in Table 4.
anilines and benzylamines, and 11 different phenoxyacetamides
and screened it for GHSR activity. The most interesting SAR trends
are summarized in Table 2 (benzylamine series) and Table 3 (ani-
line series).
In the case of the benzylamine series (Table 2), it was apparent
that the R1 and R2 SAR were non-additive; that is, the most potent
R1 substituent when R2 = H does not continue to be the most po-
tent R1 substituent when R2 is varied, and vice-versa. This trend
can be clearly seen by comparing Entries 2 and 5: when R2 = H or
R2 = m-Cl, the most potent GHSR agonists were found when
R1 = o-Me (Entry 2, EC50 = 40 and 3 nM, respectively); when
R2 = o-Cl, the most potent GHSR agonist was found where R1 = m-
Cl (Entry 5, EC50 = 10 nM). It is also interesting to compare Entries
The stereochemistry of the a-methyl group of the phenyoxyac-
etamide sidechain had a dramatic effect on GHSR potency, as illus-
trated by the 60-fold difference in EC50 and P40-fold difference in
Ki observed between the R (8a) and S (8b) isomers. The R stereo-
chemistry for 8a was assigned via X-ray crystallography, as shown
in Figure 2.13
Table 5 describes similar stereochemical effects in the benzyl-
amine series. In the case of the diastereomers of compound 9, we
observed that virtually all GHSR activity resided in 9b, as 9a was
P300-fold less active in both functional and binding assays. With
10, the single diastereomer 10b was 90-fold more potent than 10a
in the functional assay, and 70-fold more potent in the binding as-
say. Combined with the data from Table 4, it is clear that the ste-
reochemistry of the
a-methyl group of the acetamide sidechain
makes a significant contribution to the potency of this chemotype.
Table 3
SAR matrix for the aniline series
O
H
O
N
Table 4
N
GHSR activity of diastereomers of compound 8
Me
R1
R2
O
H
O
Cl
N
*
N
Entry
R1
GHSR EC50 (% Control)a,b (nM)
Me
R2 = H
R2 = o-Cl
R2 = m-Cl
8
1
2
3
4
5
6
H
440 (80)
I.a.
120 (60)
140 (60)
30 (90)
780 (70)
40 (110)
120 (90)
m-AcNH
p-AcNH
o-Cl
m-Cl
p-Cl
1000 (80)
330 (80)
1260 (80)
90 (110)
600 (80)
200 (70)
50 (90)
I.a.
250 (60)
I.a.
AcNH
Compound
Stereochem (*)
EC50 (nM)
% Control
Ki (nM)
260
8
8a
8b
R and S
R
S
30
1200
20
90
100
100
P10,000
a
See Ref. 6 for experimental details.
I.a., inactive (EC50 P 10 lM).
250
b

5086
W. Zhai et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5083–5086
Acknowledgment
The authors acknowledge Ed Kozlowski for performing the
chromatographic separation of diastereomers 8–10.
O
O
N
HN
Cl
Me
Supplementary data
=
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2008.07.120.
O
HN
References and notes
1. Smith, R. G. Endocrine Rev. 2005, 26, 346.
Figure 2. X-ray structure of 8a.
2. (a) Rocha-Sousa, A.; Henriques-Coelho, T.; Leite-Moreira, A. F. Expert Opin. Ther.
Patents 2007, 17, 909; (b) Carpino, P. A. Expert Opin. Ther. Patents 2002, 12, 1599.
3. Thomsen, W. J.; Behan, D. P.. In Comprehensive Medicinal Chemistry II; Taylor, J.
B., Triggle, D. J., Eds.; Elsevier: Oxford, 2006; Vol. 2, p 771.
Table 5
4. (a) Li, J. J.; Wang, H.; Qu, F.; Musial, C.; Tino, J. A.; Robl, J. A.; Slusarchyk, D.;
Golla, R.; Seethala, R.; Dickinson, K.; Giupponi, L.; Grover, G.; Sleph, P.; Flynn,
N.; Murphy, B. J.; Gordon, D.; Kung, M.; Stoffel, R. Bioorg. Med. Chem. Lett. 2005,
15, 1799; (b) Hernandez, A. S.; Cheng, P. T. W.; Musial, C. M.; Swartz, S. G.;
George, R. J.; Grover, G.; Slusarchyk, D.; Seethala, R. K.; Smith, M.; Dickinson, K.;
Giupponi, L.; Longhi, D. A.; Flynn, N.; Murphy, B. J.; Gordon, D. A.; Biller, S. A.;
Robl, J. A.; Tino, J. A. Bioorg. Med. Chem. Lett. 2007, 17, 5928; (c) Li, J.; Chen, S. Y.;
Li, J. J.; Wang, H.; Hernandez, A. S.; Tao, S.; Musial, C. M.; Qu, F.; Swartz, S.;
Chao, S. T.; Flynn, N.; Murphy, B. J.; Slusarchyk, D. A.; Seethala, R.; Yan, M.;
Sleph, P.; Grover, G.; Smith, M. A.; Beehler, B.; Giupponi, L.; Dickinson, K. E.;
Zhang, H.; Humphreys, W. G.; Patel, B. P.; Schwinden, M.; Stouch, T.; Cheng, P.
T. W.; Biller, S. A.; Ewing, W. R.; Gordon, D.; Robl, J. A.; Tino, J. A. J. Med. Chem.
2007, 50, 5890.
5. Andres, C. J.; Antal Zimanyi, I.; Deshpande, M.; Iben, L. G.; Grant-Young, K.;
Mattson, G. K.; Zhai, W. Bioorg. Med. Chem. Lett. 2003, 13, 2883.
6. All EC50 and Ki values represent the mean of three independent experiments.
Percent control (% Control) values describe agonist activity relative to a known
full agonist; 100% Control = full agonist. The functional and binding assays
were conducted as described in Ref. 4b.
GHSR activity of diastereomers of compounds 9 and 10
O
H
O
Cl
N
*
N
Me
9 (R=H)
R
10 (R=Me)
Compound
R
Stereochem (*)a
EC50 (nM)
% Control
Ki (nM)
300
P10,000
30
9
9a
9b
10
10a
10b
H
H
H
Me
Me
Me
R and S
R
S
R and S
R
S
7
1500
4
3
180
2
90
100
90
100
90
90
2200
30
100
7. Wang, S. S. J. Am. Chem. Soc. 1973, 95, 1328.
a
8. Marsh, I. R.; Smith, H.; Bradley, M. Chem. Commun. 1996, 941.
9. (a) Chen, C.; Ahlberg Randall, L. A.; Miller, R. B.; Jones, A. D.; Kurth, M. J. J. Am.
Chem. Soc. 1994, 116, 2661; (b) Rotella, D. P. J. Am. Chem. Soc. 1996, 118, 12246.
10. Gordon, D. W.; Steele, J. Bioorg. Med. Chem. Lett. 1995, 5, 47.
11. Nicolaou, K. C.; Pfefferkorn, J. A.; Mitchell, H. J.; Roecker, A. J.; Barluenga, S.;
Cao, G. -Q.; Affleck, R. L.; Lillig, J. E. J. Am. Chem. Soc. 2000, 122, 9954.
12. (a) McClure, K.; Hack, M.; Huang, L.; Sehon, C.; Morton, M.; Li, L.; Barrett, T. D.;
Shankley, N.; Breitenbucher, J. G. Bioorg. Med. Chem. Lett. 2006, 16, 72; (b)
Sehon, C.; McClure, K.; Hack, M.; Morton, M.; Gomez, L.; Li, L.; Barrett, T. D.;
Shankley, N.; Breitenbucher, J. G. Bioorg. Med. Chem. Lett. 2006, 16, 77.
13. X-ray coordinates for 8a have been deposited at the Cambridge
Crystallographic Data Centre and allocated the deposition number CCDC
691452.
Stereochemistry was assigned by analogy to 8a and 8b.
In summary, this report describes the discovery and optimiza-
tion of a potent series of GHSR agonists which were identified from
a large GPCR-targeted screening library. Owing to the library pedi-
gree of these screening hits, we were able to rapidly optimize
micromolar partial agonists to full GHSR agonists with single-digit
nanomolar potencies. In addition, the a-methyl stereochemistry of
the phenoxyacetamide sidechain was found to have a dramatic ef-
fect on potency.