Exploration of structure-activity relationships at the two C-terminal residues of potent 11mer Glucagon-Like Peptide-1 receptor agonist peptides via parallel synthesis

Article abstract of DOI:10.1016/j.peptides.2010.04.013

We report the identification of potent agonists of the Glucagon-Like Peptide-1 receptor (GLP-1R) via evaluation of two positional scanning libraries and a two-dimensional focused library, synthesized in part on SynPhase Lanterns. These compounds are 11 amino acid peptides containing several unnatural amino acids, including (in particular) analogs of biphenylalanine (Bip) at the two C-terminal positions. Typical activities of the most potent peptides in this class are in the picomolar range in an in vitro functional assay using human GLP-1 receptor.

Full text of DOI:10.1016/j.peptides.2010.04.013

Peptides 31 (2010) 1353–1360
Contents lists available at ScienceDirect
Peptides
journal homepage: www.elsevier.com/locate/peptides
Exploration of structure–activity relationships at the two C-terminal residues of
potent 11mer Glucagon-Like Peptide-1 receptor agonist peptides via
parallel synthesis
Tasir S. Haquea,∗, Rogelio L. Martineza, Ving G. Leeb, Douglas G. Riexingerb, Ming Leib, Ming Fengc,
Barry Koplowitzd, Claudio Mapellib, Christopher B. Coopera, Ge Zhangb, Christine Huangd,
William R. Ewinga, John Krupinskic
a Department of Discovery Chemistry, Bristol-Myers Squibb Research & Development, PO Box 4000, Princeton, NJ 08543-4000, USA
b Department of Applied Biotechnologies, Bristol-Myers Squibb Research & Development, PO Box 4000, Princeton, NJ 08543-4000, USA
c Department of Discovery Biology, Bristol-Myers Squibb Research & Development, PO Box 4000, Princeton, NJ 08543-4000, USA
d Department of Metabolism and Pharmacokinetics, Bristol-Myers Squibb Research & Development, PO Box 4000, Princeton, NJ 08543-4000, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We report the identification of potent agonists of the Glucagon-Like Peptide-1 receptor (GLP-1R) via eval-
uation of two positional scanning libraries and a two-dimensional focused library, synthesized in part on
SynPhaseTM Lanterns. These compounds are 11 amino acid peptides containing several unnatural amino
acids, including (in particular) analogs of biphenylalanine (Bip) at the two C-terminal positions. Typical
activities of the most potent peptides in this class are in the picomolar range in an in vitro functional
assay using human GLP-1 receptor.
Received 22 January 2010
Received in revised form 19 April 2010
Accepted 19 April 2010
Available online 24 April 2010
Keywords:
GLP-1R
© 2010 Elsevier Inc. All rights reserved.
Agonist
11mer
Biphenylalanine
Combinatorial
Library
Parallel
1. Introduction
tic target [6,10,15]. However, the enzyme dipeptidyl peptidase IV
(DPP-IV) readily cleaves GLP-1 at the 2-position alanine residue,
Type 2 diabetes is the most prevalent form of diabetes, account-
ing for 90–95% of the estimated 20.8 million people with diabetes in
the United States alone [5]. Patients with type 2 diabetes initially
develop resistance to the effects of insulin, resulting in an eleva-
tion of blood glucose levels. The patient’s production of insulin by
pancreatic ␤-cells eventually decreases, leading to an even greater
elevation of glucose levels in the blood.
Glucagon-Like Peptide-1 (GLP-1) was first identified in 1983
[3,4], and has become an important subject of interest for the treat-
ment of type 2 diabetes [14]. GLP-1 is an incretin hormone that is
normally secreted in response to food intake and regulates insulin
release in a glucose-dependent manner. GLP-1 has been implicated
in promoting satiety after meals, beta-cell growth, delayed gastric
emptying, and decreased glucagon and gluconeogenesis [9]. The
broad, beneficial effects of GLP-1 make it an attractive therapeu-
inactivating the agonist activity of GLP-1. Consequently, the short
plasma half-life of GLP-1 in vivo (less than 5 min) precludes the
parent peptide from being a useful treatment for type 2 diabetes.
GLP-1 binds to and activates the GLP-1 receptor (GLP-1R).
GLP-1R is a member of the class B family of G-protein-coupled
receptors (GPCRs) [8]. Receptors in this family typically consist
of two domains; an N-terminal extracellular binding domain (N-
domain), and a C-terminal domain of 7 ␣-helices that span the
membrane (juxtamembrane domain, or J-domain). It is believed
that high affinity binding of ligands to the receptor occurs at the
N-domain, inducing a conformational change that allows the N-
terminal domain of the ligand to bind to the J-domain to activate
the receptor and intracellular G protein. With few exceptions, the
class B GPCRs have proven to be difficult targets for development
of small molecule agonist therapeutics; nearly all agonists reported
to date consist of peptides of moderate length (30–60 amino acids)
[8].
Most agonists reported for GLP-1R are closely related in both
sequence and size to the endogenous peptide GLP-1(7–37), the
∗
Corresponding author. Tel.: +1 609 252 4000.
E-mail address: tasir.haque@bms.com (T.S. Haque).
0196-9781/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.peptides.2010.04.013

1354
T.S. Haque et al. / Peptides 31 (2010) 1353–1360
Fig. 1. Sequences of GLP-1(7–37) agonist peptide (top), and Exendin-4 (bottom).
Fig. 2. Lead peptide 1, with sequence His1-Aib-Glu-Gly-Thr5-[␣-MePhe(2-F)]-Thr-Ser-Asp-Bip(2-Et,4-OMe)10-Bip(2-Me)11-NH2.
sequence of which is shown in Fig. 1. A number of organiza-
tions are working to develop drugs which act as metabolically
stabilized GLP-1R agonists [13]. One example of a potent and rel-
atively stable GLP-1 agonist, recently approved for use with type
2 diabetes by the US Food and Drug Administration, is exenatide
(marketed under the name ByettaTM). Exenatide is a synthetic ver-
sion of the naturally occurring peptide Exendin-4, which is found
in Gila Monster lizard saliva (Fig. 1). Another peptide reported
to be in clinical development by Novo Nordisk is Liraglutide
(NN2211), which also has a sequence closely related to that of GLP-1
(Liraglutide = ␥-L-gutamoyl(N-␣-hexadecanoyl)-Lys26, Arg34-GLP-
1(7–37)) [1]. These larger peptide agonists have high binding
affinities for the GLP-1 receptor (typically in the picomolar range).
The primary focus of our research has also focused on finding com-
pounds with a longer plasma half-life than GLP-1, while retaining
or improving upon GLP-1’s potent functional activity.
with a chemical linker to allow organic synthesis to occur. Each
lantern used in this project incorporated a radiofrequency tag. This
tag accommodated the application of deterministic “split and mix”
synthesis techniques with identification and isolation of individual
final products. Lanterns functionalized with the Rink amide linker,
a strong acid labile linker, were selected to allow for the depro-
tection of the tert-butyloxycarbonyl (Boc) protecting group under
mildly acidic conditions without cleaving the amino acid itself from
the support. (Several early attempts to accomplish the Suzuki reac-
tion with Fmoc-protected 4-iodophenylalanine on the Lantern solid
support failed to yield the desired product, and resulted in loss of
the Fmoc group from the amino acid.)
2. Methods
2.1. Abbreviations
We have reported the identification of 11 amino acid peptides
that are potent GLP-1 agonists when tested against human and
mouse GLP-1 receptors (1, Fig. 2) [11]. These peptides contain
four non-proteinogenic amino acids, display GLP-1-like functional
agonist activity in the low picomolar range, and have greatly
improved plasma half-lives over that of GLP-1 itself. The exact
binding mode of these peptides to GLP-1R is currently unknown.
The first nine amino acids (positions 1–9 in our series, as shown
in Fig. 2) are closely related to those found in the N-terminal
region of GLP-1. The two C-terminal biphenyl amino acids (posi-
tions 10 and 11 in our series) essentially replace the 22 Val16-Gly37
amino acids of GLP-1. We sought to identify novel position 10
and position 11 biphenylalanine (Bip) amino acids as part of our
efforts to clarify structure–activity relationships around lead pep-
tide 1.
We elected to examine a variety of substituted Bip amino acids at
either position 10 or 11, while holding the rest of the 11mer peptide
sequence constant. The “parent sequence” used for the constant
part of the peptides was taken from 1. SynPhaseTM Lanterns (Mimo-
topes, Melbourne, Australia) were selected for use as a solid support
to facilitate the generation of the libraries. The SynPhase-PSTM
Lantern consists of an inert polymer base onto which a polystyrene
surface has been grafted. The polystyrene surface is functionalized
Bip = Biphenylalanine; Boc = tert-butoxycarbonyl; cAMP = cyclic
adenosine monophosphate; DIC = diisopropylcarbodiimide; DIEA =
N,N-diisopropylethylamine; DMA=N,N-dimethylacetamide; DMF=
N,N-dimethylformamide;
Fmoc = fluorenylmethoxycarbonyl;
HOBt = 1-hydroxybenzotriazole; HPLC = high pressure liquid
chromatography; K3PO4 = potassium phosphate; LC–MS = liquid
chromatography/mass spectroscopy; MeOH = methanol; PyBop =
benzotriazol-1-yl-oxytripyrrolidinophosphonium
hexaflu-
[Pd(PPh3)4] = tetrakistriphenylphosphine
SAR = structure–activity relationship;
orophosphate;
palladium(0);
SPA = scintillation proximity assay; TFA = trifluoroacetic acid;
THF = tetrahydrofuran.
2.2. Solid phase synthesis of position 10 and 11 biphenyl amino
acid dipeptides
Synthesis of diverse position 11 Bip analogues was accom-
plished as shown in Scheme 1. Access to a diverse set of position 10
Bips was achieved in an analogous method after first coupling the
completed position 11 Bip amino acid (synthesized in bulk using
standard solution-phase methods) to the Lantern. Generation of the
“positional scanning” sets was accomplished as shown in Scheme 3,

T.S. Haque et al. / Peptides 31 (2010) 1353–1360
1355
Scheme 1. Reagents and conditions: (a) DMF/piperidine (8:2, excess), 1 h. (b)
Boc-L-Phe(4-I)-OH (2 equiv.), DIC (2.1 equiv.), HOBt (2.1 equiv.), DIEA (4 equiv.),
DMF/CH2Cl2 (1:1), 18 h. (c) Arylboronic acid (4 equiv.), aq. K3PO4 (2 M solution,
8 equiv.), Pd[P(Ph)3]4 (0.1 equiv.), DMA, 80 ◦C, 16 h. (d) 1N HCl in 1,4-dioxane
(10 equiv.), 1 h. (e) Fmoc-Bip(2ꢀ-Et,4ꢀ-OMe)-OH (1.3 equiv.), PyBOP (1.4 equiv.), HOBt
(1.4 equiv.), DIEA (4 equiv.), DMF/CH2Cl2 (2:1) 24–48 h. (f) DMF/piperidine (8:2), 1 h.
(g) 1:1 TFA/CH2Cl2 (1:1), 2 h.
Scheme 2. Reagents and conditions: (a) DMF/piperidine (8:2, excess), 1 h. (b)
Boc-L-Phe(4-I)-OH (2 equiv.), DIC (2.1 equiv.), HOBt (2.1 equiv.), DIEA (4 equiv.),
DMF/CH2Cl2 (1:1) 18 h. (c) Arylboronic acid (4 equiv.), aq. K3PO4 (2 M solution,
8 equiv.), Pd[P(Ph)3]4 (0.1 equiv.), DMA, 80 ◦C, 16 h. (d) 1N HCl in 1,4-dioxane
(10 equiv.), 1 h. (e) TFA/CH2Cl2 (1:1) 2 h, repeat.
using a convergent synthesis to take advantage of the fact that posi-
tions 1–9 were held constant throughout the peptides described.
The solid-phase synthetic route allowing us to access sub-
stituted Bips at positions 10 and 11 was developed based
on Suzuki-Miyaura coupling chemistry [12]. For example,
for position 11 Bips, Boc-L-4-iodophenylalanine was loaded
onto SynPhase-PSTM Lanterns with Rink amide linker via
activation with 1,3-diisopropylcarbodiimide in the presence
of 1-hydroxybenzotriazole and N,N-diisopropylethylamine. The
palladium-catalyzed Suzuki cross-coupling was achieved using
[Pd(PPh3)4] as catalyst, K3PO4 as base, at least 4 equiv. of aryl-
boronic acid, and heating in N,N-dimethylacetamide at 80 ◦C
overnight. We determined that the Boc-carbamate can be depro-
tected without significant cleavage of amino acid from the Rink
linker using 1N HCl in 1,4-dioxane for 60 min [16]. The Lanterns
were washed with DMF, MeOH, and CH2Cl2 between reactions and
after deprotections (3 washes per solvent, 3–5 min per wash), as
well as washing with dilute base (typically 10% Et3N in DMF) after
removal of the Boc protecting group.
The synthesis of the C-terminal dimers for the 2D combina-
torial library is described in Scheme 2. The variations for both
positions 10 and 11 were introduced on the SynPhaseTM Lanterns
via Pd-catalyzed coupling of arylboronic acids to support-bound
Boc-L-4-iodophenylalanine. After cleavage of the dimers from the
Lanterns, the peptides were completed via the solution-phase frag-
ment coupling shown in Scheme 3.
Scheme 3. Convergent synthesis of 11mer peptides. Reagents and conditions:
(a) standard solid-phase Fmoc-based peptide synthesis on 2-chlorotrityl chloride
resin. (b) dichloromethane/acetic acid/2,2,2-trifluoroethane (8:1:1), 1 h. (c) 5:5:0.25
TFA:CH2Cl2:triisopropylsilane, 2 h (excess, repeat). (d) macroporous carbonate resin
(3 equiv.) in THF, 2 h. (e) DIC (1.0 equiv.), HOBt (1.1 equiv.), 9:1 CHCl3:DMF, 16 h. (f)
95:2.5:2.5 TFA/H2O/triisopropylsilane (excess), 2 h.

1356
T.S. Haque et al. / Peptides 31 (2010) 1353–1360
Table 1
2.3. Solution-phase 9 + 2 fragment coupling to generate desired
11mer peptides
In vitro GLP-1 agonist functional activity for human receptor, for peptides 1–47 con-
taining reported 11mer sequence with Aib in position 2 and Bip(2ꢀ-Me) at position
11. All position 10 amino acids were of the L-configuration.
A convergent approach to the synthesis of the full peptides
was adopted, where the positions 1–9 segment of the peptide
(free C-terminal carboxylic acid, side chains and N-terminal amine
protected) was coupled to the positions 10 + 11 dimer in solu-
tion [2] (Scheme 3). Both the side chain-protected N-terminal
9mer and the C-terminal dimer were synthesized on solid sup-
ports (resin and Lanterns, respectively), then cleaved and purified
prior to the solution-phase fragment condensation reaction. Prod-
uct 11mer peptides then underwent acidic side chain deprotection
and preparative purification via HPLC.
Compound
Position 10 second aryl group
Human GLP-1 cAMP EC50
pMa
,
1
2
3
2-Ethyl-4-methoxyphenyl
2-Methoxypyridin-3-yl
2-Ethylphenyl
146
135
157
4
Quinolin-5-yl
184
5
6
7
2-Methyl-4-methoxyphenyl
3,5-Dimethylisoxazol-4-yl
o-Tolyl
281
303
336
8
9
2,3-Dimethylphenyl
2-Methyl-4-fluorophenyl
Isoquinolin-4-yl
338
351
388
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
3,5-Dimethoxyphenyl
2-Chlorophenyl
Quinolin-6-yl
2-Methyl-5-fluorophenyl
2,3-Dihydrobenzofuran-5-yl
2-Methyl-4-chlorophenyl
2-Fluoro-3-methoxyphenyl
2-Methoxy-5-isopropylphenyl
3-Fluoro-4-methoxyphenyl
Naphthalen-1-yl
388
392
410
475
553
560
589
624
631
676
689
729
801
876
1090
1190
1230
1270
1300
1330
1360
1420
1440
1500
1510
1880
1890
2100
2120
2530
3840
3970
4190
4610
11,700
15,300
30,400
2.4. Characterization of product peptides
The products were characterized by analytical LC–MS chro-
matography and observation of the M+H+ ion via electrospray mass
spectroscopy.
Conditions for analytical LC–MS chromatography: 10–20 ␮L of
a 1 mg/mL solution of product in 1:1 CH3CN/H2O was injected
onto a Shimadzu/Waters LC–MS system equipped with a Phe-
nomenex Luna C-18 4.6 mm × 50 mm column, and running a
10–90% aq. MeOH + 0.1% trifluoroacetic acid gradient over 4 min.
Product purity was determined via UV detection of peaks at 220 nm,
and identity was confirmed by observation of the (M+2H+)/2 ion in
the mass trace.
Conditions for electrospray mass spectroscopy of product pep-
tides: Product peptide was dissolved at 1 mg/mL in 1:1 CH3CN/H2O
and injected onto a Finnigan SSQ7000 mass spectrometer equipped
with an electrospray ionization source. Both ES+ and ES− spectra
were obtained to confirm identity of product peptides.
Pyridin-4-yl
m-Tolyl
2-Fluoro-6-methoxyphenyl
2,4-Dichlorophenyl
3-Isopropylphenyl
6-Methoxynaphthalen-2-yl
4-Methoxymethylphenyl
3-Methoxyphenyl
4-Methoxypyridin-3-yl
4-Phenoxyphenyl
Benzo[d][1,3]dioxol-5-yl
6-Ethoxynaphthalen-2-yl
3,4-Dimethylphenyl
Quinolin-3-yl
2.5. GLP-1 agonist functional assay
4-Chlorophenyl
3,5-Dimethylphenyl
3-Methyl-4-methoxyphenyl
3-Chlorophenyl
4-Methylnaphthalen-1-yl
2,3-Difluorophenyl
4-Chloro-4-fluorophenyl
4-Isopropylphenyl
3,5-Difluorophenyl
3-Methyl-4-chlorophenyl
3,5-Dichlorophenyl
2-Phenoxyphenyl
The compounds were assayed for GLP-1 agonist functional activ-
ity as follows: a Chinese Hamster Ovary (CHO) cell line expressing
human GLP-1R was utilized in primary cell based assays that mea-
sure the intracellular cAMP generated by GLP-1 analog stimulation.
This scintillation proximity assay (SPA) was performed in a 96-well
adherent cell format using a radioimmunoassay (RIA) based kit.
Upon treatment with peptidyl agonists, the non-radioactive cAMP
synthesized in the cell lysate competed with the 125I-labeled cAMP
tracer for the limited amount of anti-cAMP antibody, which binds
to SPA beads that are covalently coated with anti-rabbit antibody,
causing the light signal emitted by the SPA beads to decrease. The
decrease of signal is proportional to the amount of intracellular
cAMP generated by GLP-1 analog stimulation.
Quinolin-8-yl
a
Values are means of four experiments, with assay-to-assay variability typically
within 2-fold based on results with a control compound.
data were normalized and plotted as the percentage of the response
induced by 10 nM GLP-1. The EC50 value of compounds was defined
as the concentration of compound that stimulated 50% of maximal
cAMP synthesis by 10 nM GLP-1 in CHO cells as the positive control.
The values shown in the tables are the means of four experiments,
with assay-to-assay variability typically within 2-fold based on
results with a control compound.
CHO cells stably expressing human GLP-1R were plated at
2 × 104 cells/well in sterile 96-well white clear-bottom plates
and incubated overnight before assaying. On the assay day, after
aspirating the growth media, the cells were treated with 50 ␮L
of compound at varying concentrations, or buffer control in
phosphate-buffered saline (PBS) without MgCl2 and CaCl2, with
0.1 mM 3-isobutyl-1-methylxanthine and 0.05% bovine serum
albumin for 20 min at room temperature. The solution was then
aspirated, and 50 ␮L lysis buffer was added immediately, followed
by addition of 70 ␮L of assay buffer containing 125I-labeled cAMP
tracer, rabbit anti-cAMP antibody and SPA beads that were cova-
lently coated with anti-rabbit antibody (provided in the Amersham
cAMP SPA assay kit). The plates were incubated at room temper-
ature for 12 h before counting on a TriLux Microbeta reader. The
cAMP standard curve with 12 concentrations was established inde-
pendently using known amounts of non-radioactive cAMP. The
amount of cAMP from treated cells was converted to picomoles
(pmol) of cAMP by interpolating from the cAMP standard curve. The
2.6. Pharmacokinetics in dogs
The pharmacokinetic parameters of peptide 49 were deter-
mined in male beagle dogs. Following an overnight fast, each
animal received peptide 49 by subcutaneous injection given near
the shoulder blades (67 ␮g/kg). The dosing vehicle was propylene
glycol:phosphate buffer (50:50). Serial blood samples were col-
lected in EDTA-containing microcentrifuge tubes at predose, 0.25,
0.5, 0.75, 1, 2, 4, 6, 8, 24, and 30 h post-dose after subcutaneous
administration. Approximately 0.3 mL of blood was collected at

T.S. Haque et al. / Peptides 31 (2010) 1353–1360
1357
Table 2
In vitro GLP-1 agonist functional activity for human receptor, for peptides 48–62
containing reported 11mer sequence with Aib in position 2 and Bip(2ꢀ-Et,4ꢀ-OMe)
at position 10. All position 11 amino acids were of the L-configuration.
Compound
Position 11 second aryl group
Human GLP-1 cAMP EC50
pMa
,
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
3,5-Dimethylphenyl
2-Chlorophenyl
Quinolin-5-yl
3,4-Dimethoxyphenyl
3-Chlorophenyl
3-Methoxyphenyl
Pyridin-3-yl
3-Chloro-4-fluorophenyl
4-Methoxypryidin-3-yl
Quinolin-6-yl
78
88
89
100
103
108
118
123
144
160
220
228
241
244
249
Quinolin-8-yl
3-Isopropylphenyl
4-Isopropylphenyl
Quinolin-3-yl
Fig. 3. Plasma concentration vs. time profiles of peptide 49 in the dog.
2,4-Dichlorophenyl
a
each time point. Blood samples were immediately centrifuged at
4 ◦C. The obtained plasma was frozen with dry ice and stored at
−20 ◦C. Plasma drug levels were determined using the LC–MS/MS
assay.
Values are means of four experiments, with assay-to-assay variability typically
within 2-fold based on results with a control compound.
Table 3
Pharmacokinetic profile of peptide 49.
3. Results
Species
Dog
SC dose
Tmax (h)
Cmax (nM)
AUC (nM h)
168
SC t1/2 (h)
67 ␮g/kg
5
11
8
Previous research within our group suggested that Bip(2ꢀ-Et,4ꢀ-
OMe) (from lead peptide 1) was a preferred residue at position 10
[11]. We wished to exhaustively examine the SAR at position 10 and
explore this hypothesis. Therefore, a wide variety of arylboronic
acids were used in this first “positional scanning” library. The first
set of compounds explored a variety of Bip analogs at position 10,
while holding positions 1–9 constant (as shown in Fig. 2), and posi-
tion 11 constant as Bip(2ꢀ-Me). The results for these peptides in an
agonist assay using human GLP-1 receptor are shown in Table 1.
We next proceeded to explore position 11. This position was
moderately explored in previous research within our group [7,11],
so we selected a limited set of arylboronic acids to complete our SAR
analysis. Functional activity for the resulting 15 peptides is shown
in Table 2. Both these results and those obtained previously suggest
that position 11 is considerably more “permissive” than position 10
regarding substitution on the distal ring of the Bip.
The pharmacokinetic parameters of peptide 49 in male bea-
gle dogs (n = 2), following subcutaneous (SC) administration are
summarized in Table 3. The plasma concentration–time profiles
are presented in Fig. 3. The time to reach peak concentrations
(Tmax) after a subcutaneous dose of 67 ␮g/kg occurred at 5 h.
The maximum plasma concentration (Cmax) after subcutaneous
administration was 11 nM. The estimated apparent elimination
half-life was 8 h. These results encouraged us to pursue our two-
dimensional (2D) matrix library, which is described below.
Fig. 4. Aryl groups used at positions 10 and 11 in the 10 × 9 matrix library.

1358
T.S. Haque et al. / Peptides 31 (2010) 1353–1360
Fig. 5. Graph of matrix results, displaying – pEC50 activity on the z-axis (i.e. taller cones indicate more potent GLP-1 agonists), position 10 amino acid on the x-axis, and
position 11 amino acid on the y-axis. “Not obtained” values are at the base of the graph.
The aryl groups for the 90 compound 2D library were chosen
based either on the GLP-1R agonist potency displayed in the posi-
tional scan libraries, or on similarity to the corresponding side chain
from the lead compound 1. An effort was made to include side
chains with a range of properties, e.g. electron-withdrawing sub-
stituents, electron-donating substituents, heteroaromatic, etc. The
Bip side chains ultimately incorporated into the library are shown
in Fig. 4.
Upon completion of the synthesis and purification of the 2D
library, we were able to obtain 75 out of 90 anticipated peptides
in sufficient yield and purity for inclusion in our GLP-1 agonist in
vitro assay. Results for the peptides in the GLP-1 functional assay
against human receptor are shown in Table 4 , and are depicted
graphically in Fig. 5 in Section 4.
to observe that the peptide had a half-life of 8 h after SC adminis-
tration, much longer than the less than 5 min typically seen with
the native GLP-1 peptide. The long apparent half-life of the 11mer
suggests that potent agonists in this class could have potential as
practical therapeutic agents for the treatment of type 2 diabetes.
We ultimately wished to identify new, more potent peptides
with novel Bip analogs at positions 10 and 11. Much of our previ-
ous research, including the two initial positional scanning libraries
detailed in Tables 1 and 2, focused on modifying a single amino acid
while holding the rest of the peptide constant [7,11]. While this
approach is ideal for side chains that bind to the receptor indepen-
dently, it does not allow for the exploration of possible cooperative
effects. We elected to synthesize a 90 compound two-dimensional
(10 × 9) library focusing on positions 10 and 11 with two key pur-
poses in mind: (1) identification of any cooperative effects resulting
in highly potent agonists; (2) production of a number of potent ago-
nists that could be subsequently pursued in more advanced (e.g. in
vivo) studies.
Several trends in activity are apparent after examination of the
data in Table 4 and Fig. 5 (higher cones equal more potent ago-
nists). At position 10, a substituent at the ortho position of the
second aromatic ring of the biphenyl system promotes agonist
activity, as seen with the 2-ethylphenyl, 2-methyl-4-fluorophenyl,
2-methyl-4-methoxyphenyl, and 2-chlorophenyl Bips. The activ-
ity of the quinolin-5-yl substituent at position 10 observed with
4 in the first positional scan library was validated by the potent
activities of compounds 135–139, all of which also contain the
quinolin-5-yl substituent. The potency of the 3,4-dimethoxyphenyl
substituent at position 10 was likewise affirmed by the activity
of compounds 99, 100, and 102. Overall, side chains bearing close
4. Discussion
From the positional scanning library exploring position 10 it’s
perhaps not surprising that compounds 3 and 5, containing Bips
that are close analogs to Bip(2ꢀ-Et,4ꢀ-OMe) (from lead compound 1),
display good GLP-1 agonist activity. Several interesting and unex-
pected results were also observed; the good activities of peptides 4,
6, 10, and 13 (products of quinoline or isoquinoline boronic acids)
suggested the possibility that larger Bip analogs at position 10 could
promote binding to the GLP-1 receptor. The position 11 scanning
library was less definitive, with a number of side chains displaying
good agonist potency.
One of the more potent compounds from the second library, 49,
was examined via cassette dosing in dogs to evaluate the phar-
macokinetic properties of this series (Table 3). We were pleased

T.S. Haque et al. / Peptides 31 (2010) 1353–1360
1359
Table 4
In vitro GLP-1 agonist functional activity for human receptor, for peptides 63–152 obtained from 10 × 9 matrix library, containing variation at positions 10 and 11.
Compound
Position 10 second aryl group
Position 11 second aryl group
Human GLP-1 cAMP EC50, pMa
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
2-Fluoro-3-methoxyphenyl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
6-Methoxypyridin-3-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
3,5-Dimethylisoxazol-4-yl
Pyridin-4-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
148
159
851
214
574
994
331
1670
11,700
190
402
270
827
110
322
6020
2240
Not obtained
Not obtained
352
Not obtained
542
1870
991
557
562
2070
54
166
52
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
Pyridin-4-yl
Pyridin-4-yl
92
93
Pyridin-4-yl
3-Methoxyphenyl
216
94
Pyridin-4-yl
Quinolin-5-yl
Not obtained
95
96
97
98
Pyridin-4-yl
Pyridin-4-yl
Pyridin-4-yl
Pyridin-4-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
1350
Not obtained
2990
Not obtained
55
67
128
50
1000
259
1550
170
2360
86
62
187
Not obtained
208
106
198
427
1570
143
97
146
Not obtained
48
63
304
Not obtained
446
294
99
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
3,4-Dimethoxyphenyl
2-Chlorophenyl
2-Chlorophenyl
2-Chlorophenyl
2-Chlorophenyl
2-Chlorophenyl
2-Chlorophenyl
2-Chlorophenyl
2-Chlorophenyl
2-Chlorophenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-methoxyphenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
2-Methyl-4-fluorophenyl
Quinolin-5-yl
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
154
Not obtained
49
138
Not obtained
97
167
446
28
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
Quinolin-5-yl
Quinolin-5-yl
72
73

1360
T.S. Haque et al. / Peptides 31 (2010) 1353–1360
Table 4 (Continued )
Compound
Position 10 second aryl group
Position 11 second aryl group
Human GLP-1 cAMP EC50, pMa
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
Quinolin-5-yl
Quinolin-5-yl
Quinolin-5-yl
Quinolin-5-yl
Quinolin-5-yl
Quinolin-5-yl
2-Ethylphenyl
2-Ethylphenyl
2-Ethylphenyl
2-Ethylphenyl
2-Ethylphenyl
2-Ethylphenyl
2-Ethylphenyl
2-Ethylphenyl
2-Ethylphenyl
3-Methoxyphenyl
Quinolin-5-yl
27
50
157
260
785
Not obtained
52
45
Not obtained
61
38
Not obtained
240
63
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
3-Chloro-4-fluorophenyl
3,5-Dimethylphenyl
2-Chlorophenyl
3-Methoxyphenyl
Quinolin-5-yl
2-Methyl-4-methoxyphenyl
2-Methoxy-3,5-difluorophenyl
3,4-Dimethoxyphenyl
6-Methoxypyridin-3-yl
305
a
Values are means of four experiments, with assay-to-assay variability typically within 2-fold based on results with a control compound.
resemblance to the parent position 10 side chain in 1 are potent,
and the activity of several more novel position 10 side chains was
confirmed.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.peptides.2010.04.013.
A wider variety of side chains at position 11 displayed high
levels of activity, as expected from our previous results. Substitu-
tion is tolerated at nearly every position on the second aryl ring,
though less so at the para position: peptides with the 2-methyl-4-
methoxyphenyl and 3,4-dimethoxyphenyl side chains at position
11 tend towards lower activity versus peptides with substituted
side chains at position 11. Given the ortho-substituted position 11
side chain in 1, we were interested to observe that many of the more
potent side chains from the 2D library lacked an ortho-substituent
at this position (e.g. representative peptides 99, 102, 109, 135, 138,
and 145).
Several side chains included in the 2D library are generally
unfavorable in terms of conferring agonist activity. These include
3,5-dimethylisoxazol-4-yl, 5-methoxypyridin-3-yl, and 2-fluoro-
3-methoxyphenyl at position 10, and 3,4-dimethoxyphenyl and
6-methoxypyridin-3-yl at position 11.
References
[1] Agerso H, Jensen LB, Elbrond B, Rolan P, Zdravkovic M. The pharmacokinetics,
pharmacodynamics, safety and tolerability of NN2211, a new, long-acting GLP-
1 derivative, in healthy men. Diabetologia 2002;45(2):195–202.
[2] Barlos K, Gatos D. 9-Fluorenylmethoxycarbonyl/t-butyl-based convergent pro-
tein synthesis. Biopolymers 1999;51(4):266–78.
[3] Bell GI, Sanchez-Pescador R, Laybourn PJ, Najarian RC. Exon dupli-
cation and divergence in the human preproglucagon gene. Nature
1983;304(5924):368–71.
[4] Bell GI, Santerre RF, Mullenback GT. Hamster preproglucagon contains the
sequence of glucagon and two related peptides. Nature 1983;302(5910):716.
[5] Centers for Disease Control and Prevention. “National diabetes fact sheet:
general information and national estimates on diabetes in the United States,
2005.” Atlanta, GA: US Department of Health and Human Services, Centers
for Disease Control and Prevention; 2005. Available on the worldwide web at
http://www.cdc.gov/diabetes/pubs/factsheet05.htm.
[6] Drucker DJ, Nauck MA. The incretin system: Glucagon-Like Peptide-1 recep-
tor agonists and dipeptidyl peptidase-4 inhibitors in type 2 diabetes. Lancet
2006;368(9658):1696–705.
[7] Haque TS, Ruan Z, Lee VG, Riexinger DG, Lei M, Malmstrom S, et al. Iden-
tification of potent 11mer Glucagon-Like Peptide-1 agonist peptides with
novel C-terminal amino acids. Homohomophenylalanine analogues. Peptides
2010;31(5):950–5.
[8] Hoare SRJ. Mechanisms of peptide and nonpeptide ligand binding to Class B
G-protein-coupled receptors. Drug Discov Today 2005;10(6):417–27.
[9] Lenhard JM, Gottschalk WK. Preclinical developments in type 2 diabetes. Adv
Drug Deliv Rev 2002;54:1199–212.
[10] Lovshin JA, Drucker DJ. Incretin-based therapies for type 2 diabetes mellitus.
Nat Rev Endocrinol 2009;5(5):262–9.
[11] Mapelli C, Natarajan S, Meyer J-P, Bastos M, Bernatowicz M, Lee V, et al. Eleven
amino acid Glucagon-Like Peptide-1 receptor agonists with antidiabetic activ-
ity. J Med Chem 2009;52(23):7788–99.
5. Conclusions
In conclusion, a series of short (11 amino acid) GLP-1 receptor
agonist peptides were optimized for agonist potency against the
human GLP-1 receptor using iterative parallel synthesis. Through
the use of two positional scanning libraries, followed by a two-
dimensional library focused primarily on highly active side chains,
we were able to identify several new position 10 and position 11
Bip analogues which result in peptides with high agonist potency.
One analog displayed an apparent half-life of 8 h following sub-
cutaneous administration in a dog PK study. The final, 2D (10 × 9)
library that was focused on optimization of compound potency was
also highly successful, with 22 compounds (29% of 76 compounds
tested) having agonist activity of less than 100 pM. By extensively
exploring SAR at the two C-terminal positions via positional scan-
ning libraries, then using that data to design a final 2D library,
we were able to rapidly explore SAR, identify new biphenylala-
nine analogs at both positions 10 and 11, and generate a number of
highly active agonists for potential use in follow-up in vitro and in
vivo studies.
[12] Miyaura N, Yamada K, Suzuki A. A new stereospecific cross-coupling by the
palladium-catalyzed reaction of 1-alkenylboranes with 1-alkenyl or 1-alkynyl
halides. Tetrahedron Lett 1979;20(36):3437–40.
[13] Nielsen LL. Incretin mimetics and DPP-IV inhibitors for the treatment of type 2
diabetes. Drug Discov Today 2005;10(10):703–10.
[14] Perry T, Greig NH. The Glucagon-Like Peptides: a double-edged therapeutic
sword? Trends Pharmacol Sci 2003;24(7):377–83.
[15] Roges OA, Baron M, Philis-Tsimikas A. The incretin effect and its potentiation
by Glucagon-Like Peptide 1-based therapies: a revolution in diabetes manage-
ment. Expert Opin Investig Drugs 2005;14(6):705–27.
[16] Trivedi HS, Anson M, Steel PG, Worley J. A method for selective N-Boc depro-
tection on Wang resin. Synlett 2001:1932–4.