Design, Synthesis, and Pharmacological Evaluation of Potent Positive Allosteric Modulators of the Glucagon-like Peptide-1 Receptor (GLP-1R)

Article abstract of DOI:10.1021/acs.jmedchem.9b01071

The therapeutic success of peptidic GLP-1 receptor agonists for treatment of type 2 diabetes mellitus (T2DM) motivated our search for orally bioavailable small molecules that can activate the GLP-1 receptor (GLP-1R) as a well-validated target for T2DM. Here, the discovery and characterization of a potent and selective positive allosteric modulator (PAM) for GLP-1R based on a 3,4,5,6-tetrahydro-1H-1,5-epiminoazocino[4,5-b]indole scaffold is reported. Optimization of this series from HTS was supported by a GLP-1R ligand binding model. Biological in vitro testing revealed favorable ADME and pharmacological profiles for the best compound 19. Characterization by in vivo pharmacokinetic and pharmacological studies demonstrated that 19 activates GLP-1R as positive allosteric modulator (PAM) in the presence of the much less active endogenous degradation product GLP1(9-36)NH₂ of the potent endogenous ligand GLP-1(7-36)NH₂. While these data suggest the potential of small molecule GLP-1R PAMs for T2DM treatment, further optimization is still required towards a clinical candidate.

Full text of DOI:10.1021/acs.jmedchem.9b01071

Subscriber access provided by Stockholm University Library
Article
Design, Synthesis and Pharmacological Evaluation of Potent Positive
Allosteric Modulators of the Glucagon-like Peptide-1 Receptor (GLP-1R)
Maria Mendez, Hans Matter, Elisabeth Defossa, Michael Kurz, Sylvain Lebreton, Ziyu Li, Matthias
Lohmann, Matthias Loehn, Hartmut Mors, Michael Alois Linus Podeschwa, Nils Rackelmann, Jens
Riedel, Pavel Safar, David Thorpe, Matthias Schaeffer, Dietmar Weitz, and Kristin Breitschopf
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01071 • Publication Date (Web): 09 Oct 2019
Downloaded from pubs.acs.org on October 12, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Design, Synthesis and Pharmacological Evaluation of
Potent Positive Allosteric Modulators of the Glucagon-like
Peptide-1 Receptor (GLP-1R)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
§
María Méndez*, Hans Matter*, Elisabeth Defossa, Michael Kurz, Sylvain Lebreton , Ziyu Li,
Matthias Lohmann, Matthias Löhn, Hartmut Mors, Michael Podeschwa, Nils Rackelmann,
§
†
Jens Riedel, Pavel Safar , David S. Thorpe , Matthias Schäfer, Dietmar Weitz, Kristin
Breitschopf
Sanofi-Aventis Deutschland GmbH, Industriepark Höchst, 65926 Frankfurt, Germany.
§Current address: Icagen, 2090 E Innovation Park Drive, Oro Valley, AZ 85755, USA.
†
Current address: University of Arizona, Clinical Sciences & Translational Medicine
Division of Cardiology, Dept of Medicine, Sarver Heart Center, 1501 N Campbell Ave.
Tucson, AZ 85724, USA
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
The therapeutic success of peptidic GLP-1 receptor agonists for treatment of type 2 diabetes
mellitus (T2DM) motivated our search for orally bioavailable small molecules that can
activate the GLP-1 receptor (GLP-1R) as a well-validated target for T2DM. Here, the
discovery and characterization of a potent and selective positive allosteric modulator (PAM)
for GLP-1R based on a 3,4,5,6-tetrahydro-1H-1,5-epiminoazocino[4,5-b]indole scaffold is
reported. Optimization of this series from HTS was supported by a GLP-1R ligand binding
model. Biological in-vitro testing revealed favorable ADME and pharmacological profiles for
the best compound 19. Characterization by in-vivo pharmacokinetic and pharmacological
studies demonstrated that 19 activates GLP-1R as positive allosteric modulator (PAM) in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
presence of the much less active endogenous degradation product GLP1(9-36)NH of the
2
potent endogenous ligand GLP-1(7-36)NH . While these data suggest the potential of small
2
molecule GLP-1R PAMs for T2DM treatment, further optimization is still required towards a
clinical candidate.
ACS Paragon Plus Environment
2

Page 3 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
The glucagon-like-peptide-1 receptor (GLP-1R), a class B G protein-coupled receptor
(GPCR), is a well-validated target for the treatment of type 2 diabetes mellitus (T2DM).1,2,3
Several peptidic GLP-1R agonists from the class of incretin mimetics, such as exenatide,
liraglutide, lixisenatide, albiglutide and dulaglutide, are approved today for treatment of
T2DM.4,5,6 The therapeutic success of these injectable GLP-1 analogs prompted us and other
groups to search for peptide variations or small molecules that act as GLP-1R agonist or
positive allosteric modulators (PAMs) leading to similar pharmacodynamic effects. PAMs
could have the potential for producing reduced side-effects such as nausea and vomiting
prevalent in injectable GLP-1R agonists, which in turn limits their maximal efficacy. Thus,
orally administered PAMs would likely result in higher patient convenience and compliance
with maximal therapeutic efficacy.7,8,9,10,1112
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The search for small molecular entities able to directly activate the GLP-1R is ongoing for
many years since mimicking the large interaction surface area involved in endogenous
peptide ligand binding to class B GPCRs is challenging. Two main approaches have been
pursued for activation of GLP-1R with small molecules, namely searching for direct agonists,
or for positive allosteric modulators (PAMs) able to enhance the activity of either the
endogenous ligand GLP-1(7-36)NH or its much less active degradation product GLP1(9-
2
3
6)NH , which lacks the first two amino acids His7 and Ala8.13,14 PAMs do not to interfere
2
with endogenous peptide levels and are expected not to induce tachyphylaxis, which could be
an advantage against full agonists.
Some of the earlier reported small molecule GLP-1R direct agonists suffer from non-drug-
like properties, such as high lipophilicity and molecular weight, which causes low plasma
exposure after oral administration.15,16,17 The most advanced compound is TTP273 from vTv
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Therapeutics, which is to our knowledge in clinical development for treatment of T2DM
1
8
(
Phase 2 completed). Although the chemical structure of TTP273 has not been disclosed,
dioxino-2,3-isoquinoline derivatives, such as 1 (Figure 1) have been disclosed by Transtec
now vTv-Therapeutics) in the patent literature.19,20 As expected for a small molecule with
these mentioned characteristics, considerable formulation efforts and high doses of TTP273
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
2
1
were necessary to conduct clinical trials. More recently, Pfizer and Lilly have also disclosed
2
2
23,24
small molecule GLP-1R agonists in clinical development (PF-06882961 and OWL833
in-licenced from Chugai), although structures have not yet been reported, they might be
related to those disclosed in the patent literature (compounds 7 and 8).25,26 Current data
suggest that some of the drawbacks from earlier molecules might have been overcome.
The best-characterized GLP-1R PAMs (BETP (2) and compound 2 (3), Figure 1) act by
covalently modifying a cysteine side chain in the GLP-1R, inducing thereby a conformational
change of the receptor which is able to trigger the GPCR signal cascade upon addition of less
active GLP1(9-36)NH2.27,28 However, the presence of reactive chemical functionalities in
both molecules 2 and 3, and the fact that these two compounds besides their PAM effect also
2
9
show direct agonistic activity on their own (they are allo-agonists), precludes them from
being good in-vivo tools for validating the potential of the solely positive modulation
approach in a T2DM-relevant disease context. Although several non-covalent compounds
have been reported as PAMs for the GLP-1R (compounds 4, 5 and 6 in Figure 1), only
limited information about their in-vivo pharmacology in T2DM disease-models is currently
available.30,31,32,33 Given that GLP1(9-36)NH accounts for ca. 80-90% of the postprandrial
2
3
4
GLP-1 total concentration in humans and that its half live in vivo is approximately five
times higher than the active GLP1(7-36)NH counterpart, we hypothesized that enhancing
2
the activity of the peptide ligand GLP1(9-36)NH could result in a therapeutic valuable
2
approach. In order to increase our success chances, we conducted a HTS campaign with two
ACS Paragon Plus Environment
4

Page 5 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
different experimental settings with the aim to identify both small molecule direct agonists
and PAMs for the GLP-1R. Unfortunately, this approach did not yield interesting starting
points as small molecule direct agonists which would justify a chemical optimization
program.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
However, the HTS outcome analysis led to the discovery and characterization of a potent and
selective positive allosteric GLP-1R modulator based on a 3,4,5,6-tetrahydro-1H-1,5-
epimino-azocino[4,5-b]indole scaffold (see Figure 2). This promising series was then
chemically optimized and we were able to show that the most potent compound and its
analogues indeed activate the GLP-1R as PAMs in combination with the endogenous ligand
GLP1(9-36)NH2.
N
O
OH
CF3
O
O
O
N
Cl
Cl
N
N
NH
S
N
H
O
N
N
S
O
O
O
Cl
Cl
O
3
- Compound 2
2- BETP
O
1
TTP273- close analogue
N
O
N
O
OH
N
CO2H
NH
N
HN
O
O
F
O
N
S
NH
N
O
N
N
4
5
6
O
N
N
N
NC
F
O
N
O
O
N
F
N
N
NH
O
N
N
O
O
N
N
N
OH
F
7
N
O
8
N
Figure 1. Small molecule GLP-1R agonists and allosteric modulators in literature.
Results and discussion
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Discovery of GLP-1R PAMs
We performed a HTS of the Sanofi compound collection to identify positive allosteric
modulators for the endogenous ligand GLP1(9-36)NH to interact with the GLP-1R. This was
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
done using an HTRF cAMP assay in a HEK293 cell line overexpressing the human GLP-1R.
In this assay, both GLP-1(7-36)NH and GLP1(9-36)NH peptides activate the GLP-1
2
2
receptor, albeit with significantly different potencies. The EC value of the GLP-1(7-36)NH
50
2
is approximately 10,000-fold better than the EC50 value for the GLP1(9-36)NH . When
2
performing the screening assay, all compounds were tested at the EC concentration of the
2
0
orthosteric ligand GLP1(9-36)-NH to better assess the potentiation of putative HTS-hits.
2
Subsequently, compounds were tested for their ability to shift the EC50 value of the
endogenous GLP1(9-36)NH at fixed compound concentrations of 10 and 3 µM, thereby
2
aiming to demonstrate improved biological action of the less active endogenous peptide.
Among the several active compounds from HTS, there were many with potentially reactive
chemical functionalities (not shown). Consequently, a glutathione (GSH) stability assay
served to focus on chemotypes able to activate GLP-1R in a non-covalent manner. Such a
stability assay is used as counterscreen to detect substances with electrophilic moieties
leading to covalent modification of nucleophiles.
One chemical series based on the 3,4,5,6-tetrahydro-1H-1,5-epiminoazocino[4,5-b]indole
scaffold showed a particularly high activity in our HTS assay (compound 8, PAM EC 600
50
nM/100% efficacy, see table 1). The best compound exhibited a very significant effect in
enhancing (shifting) the EC50 value of GLP1(9-36)NH (> 500 times @ 10 µM). In the
2
orthologue mouse receptor assay, which served as in-vitro surrogate for our in-vivo model,
the biological activity was similar or slightly improved (PAM EC 300 nM/100%). This
5
0
chemical series originated from a synthesized chemical library based on natural-product-like
ACS Paragon Plus Environment
6

Page 7 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
35
scaffold using a Pictet-Spengler intramolecular cyclization. This origin allowed for rapid
structure-activity-relationship (SAR) evaluation due to the internal availability of numerous
analogues plus a reasonable access to new derivatives by parallel-synthesis chemistry for
further exploring the SAR. Although the hit structure 8 is clearly outside of the property
range commonly assigned to lead-like structures from HTS (MW: 585.6),36 its robust
biological activity encouraged us to optimize this motif. Hit structures were always obtained
as single enantiomers and we could rapidly determine that the S,S stereochemistry at the
bridgehead carbon atoms is necessary for high GLP-1R activity in our primary assay. The
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
corresponding R,R isomers provided minor 2-4 EC50 shifts for GLP1(9-36)NH . Other
2
isomers were not tested.
F
F
H
O
N
H
O
O
O
N
N
NH
N
H
N
H
Boc
NH2
Boc
8
9
MW = 585.6
MW = 487.5
Figure 2. Structure of HTS hit (8) and advanced hit (9).
Chemistry
Synthesis of these compounds starts with the amide coupling of Z-protected N-Boc L-
tryptophan 9a (Scheme 1) and a diacetal-containing amine of the type 10 under standard
EDC-coupling conditions. The subsequent stereospecific Pictet-Spengler intramolecular
cyclisation in the presence of formic acid then delivered the polycyclic scaffold 12, mostly
with partial deprotection of the Boc-group at the indole-N1 moiety. A two-step protection /
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
deprotection protocol was established with crude mixtures to deliver the mono-Boc-protected
intermediate 12, which was then submitted to standard deprotection conditions of the Z-
group. Final amide coupling reaction with 2-(4,4-difluorocyclohexyl)-acetic acid rendered
compound 9 in good yields and as single (S,S) enantiomer. Subsequent functionalization of
the amide nitrogen of compound 9 was performed via a Michael addition with acrolein and
subsequent addition of MeMgBr, which delivered compound 14 as a 1:1 mixture of two
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
H
OEt
OEt
CO H
H
2
N
H
O
ZHN
H
a
ZHN
b,c,d
e
ZN
NH
OEt
OEt
H N
2
N
H
N
H
Boc
N
Boc
9a
10
11
12
H
O
R1
H
O
R1
H
O
N
g,h
HN
NH
f
O
O
OH
N
NH
N
N
H
N
H
N
H
Boc
Boc
Boc
1
3
9: R = CH difluoro-Cy;
1 2
14: R1 = CH2cyclohexyl; dr 1:1
diastereomers.
(a) EDC, Oxyma, DMAP, N-methyl morpholine, DMF, rt, 74%; Then DMAP,
Boc O, MeCN, 45%. (b) HCO H, rt, 80%. (c) Boc O, CH CN, DMAP, rt. (d)
2
2
2
3
HCO H, rt. (e) H , Pd/C, MeOH, rt, 90%. (f) R CO H, EDC, Oxyma, DMAP, N-
2
2
2
2
methyl morpholine, DMF, rt, 75%. (g) Acrolein, NaOH, THF (h) MeMgBr, THF,
7% over two steps, dr 1:1.
2
Scheme 1. Synthesis of compounds 9 and 14.
For the synthesis of compound 19, a slightly modified synthesis was employed as shown in
Scheme 2. Starting from the Fmoc-protected tryptophan 9b and the amine building block
ACS Paragon Plus Environment
8

Page 9 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
5,37 intermediate 16 was obtained via HATU coupling. Oxidation of the alcohol with Dess-
Martin periodinane and subsequent in-situ cyclisation of the crude reaction mixture led to
polycyclic intermediate 17, which was submitted to Fmoc-deprotection using MeNH to
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
render intermediate 18 as a mixture of diastereomeres, which could be separated at this stage
via HPLC. Final coupling with 2-(4,4-difluorocyclohexyl-)acetic acid and subsequent ester
hydrolysis delivered compound 19 as a single enantiomer. The relative stereochemistry of 19
could be unambiguously assigned by characterization and complete structure elucidation of
intermediate 18 by NMR spectroscopy (see supporting information).
(a) HATU, DIPEA, THF , rt, 37%; (b) Dess-Martin, CH Cl , rt, 40% ; c) H CO H, rt,
2
2
2
2
4
5%. (d) Me NH, THF, rt, 72%; (e) 2-(4,4-difluorocyclohexyl-)acetic acid, HATU,
2
DIPEA, THF , rt; (f) LiOH, MeOH:H O, rt (38% over two steps).
2
Scheme 2. Synthesis of compound 19.
Molecular pharmacology
Starting from the HTS hit 8, we first aimed to identify a simplified chemical analog retaining
biological activity in our assay, which led to identification of compound 8 (table 1). This new
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
molecule lacks the non-drug like basic side chain but was more active in the primary PAM
assay (EC / E 5 nM / 100%). Furthermore it shifted the EC value from GLP1(9-36)NH
2
5
0
max
50
by a factor of 3,791 at 10 µM and 1,183 at 3 µM, maintaining full agonistic efficacy (Emax
00%). For this and other compounds in this study, relevant data are summarized in table 2.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
Compound 8 is a specific ligand for GLP-1R without significant activity in the cAMP
accumulation assay with the parental, GIP-receptor (gastric inhibitory polypeptide), human
VIPR2 (vasoactive intestinal peptide receptor 2) or glucagon receptor overexpressing cell
lines, and it was also found to be highly stable in the GSH assay.
Figure 3. Dose-response curves from functional in-vitro assay data for cAMP
response of GLP-1(7-36)NH and GLP1(9-36)NH with or without test compounds in
2
2
PSC-HEK293 cells overexpressing GLP-1R using the shift assay format. Left: 10 µM
compound concentration. Right: 3 µM compound concentration.
Table 1: Summary of in vitro data for compounds 8-19. a) Compound EC50 at EC20 of
GLP1(9-36)NH2 in the cAMP response assay in PSC-HEK293 cells b) Shift factors of
GLP1(9-36)NH EC for compound concentrations of 10 and 3 µM in cAMP response assay
2
50
in the PSC-HEK293 c) Compound EC50 at EC20 of GLP1(9-36)NH2 in the cAMP response
ACS Paragon Plus Environment
10

Page 11 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2+
assay in 1.1B4 cells d) Compound effect on high glucose-mediated Ca influx in Min6 beta-
cells @ 10 µM vs. HG control (FLIPR assay) e) OK = No direct effect on cAMP in parental
HEK293 cells, HEK293 cells overexpressing GLP-1R, GCGR, VIPR or GIPR; no
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
modulation of cAMP response to GLP-1(7-36)NH or oxyntomodulin in GLP1R
2
overexpressing HEK293 cells, or to GLP-1(9-36)NH in GCGR overexpressing HEK293
2
cells.
EC Shift
5
0
1
.1B4 cells
PAM EC50
µM] / Emax
Fold change
Selectivity c)
[
PAM EC50
2+
ID
Ca Influx
1
0 µM 3 µM [µM] / Emax
[%]
b)
[%]
Cpd 8
Cpd 9
0.6 / 91%
1,681
nda)
nda)
526
nd
nd
0.005 / 99% 3,795
nd
0.9
1.2
nd
Cpd 14 0.0015 / 100% 3,143
0.04 / 100
OK
OK
Cpd 19 0.005 / 100% 8,188 2,215 0.130/102%
It is also a selective potentiator of GLP-1R in the presence of the GLP1(9-36)NH since it
2
neither induced a cAMP signal in the corresponding recombinant cell line when tested alone
nor in the presence of 10 nM GLP-1(7-36)NH or oxyntomodulin (data not shown).
2
Furthermore, in a radioactive ligand binding assay, compound 9 was not able to displace the
radioactive labelled peptide GLP-1(7-36)NH , thus supporting our presumed allosteric mode
2
of action. Compound 9 was also selective against other common channels and receptors.
3
8
When investigating in a standard receptor (CEREP-33 panel ) and ion channel antitarget
panel, only moderate agonistic effects on the KATP channel (EC 2.67 µM) and minor effects
50
3
9
on ion-channels such as N 1.2 and N 1.5 were observed.
av
av
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
However, when testing this compound for its ability to shift the EC value of GLP1(9-
5
0
3
6)NH in the human pancreatic 1.1B4 -cell line, which endogenously expresses the GLP-1
2
receptor, no EC value shift was observed. As shown in Figure 4, the ligand GLP1(9-36)NH
2
5
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
alone does not exhibit any activity on the endogenously expressed GLP-1R, whereas GLP-
(7-36)NH clearly activates its natural receptor.
1
2
Further optimization of the HTS hits led to analogues 14 and 19, bearing both a polar
substituent at the amide nitrogen. Our homology model (see below) suggested that a potential
polar interaction to Lys190 might be beneficial for potency. Interestingly, these compounds
could clearly potentiate GLP1(9-36)NH in our in-vitro assay (see table 1). In this primary
2
PAM assay, both compounds showed a similar potency (EC / E 1.5 nM / 100% for
max
5
0
compound 14 and 5 nM / 100% for compound 19). However, 19 showed superior activity in
the primary assay by shifting the EC value of GLP1(9-36)NH compared to 14 (8188 / 2215
50
2
fold at 10 and 3 µM for 19 versus 3143 / 526 fold at 10 and 3 µM for 14). To our knowledge,
this is the best potentiation ability described for a non-covalent PAM of the GLP-1R reported
4
0
so far. Moreover, in the endogenous human pancreatic -cell line 1.1B4, both compounds
4 and 19 showed robust PAM efficacy, although the difference between these two
compounds was not as pronounced as observed previously in the overexpressing cell line
Figure 4).
1
(
ACS Paragon Plus Environment
12

Page 13 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Dose-response curves from functional in-vitro assay for cAMP response of GLP-
1
(7-36)NH and GLP1(9-36)NH with or without test compounds in human pancreatic 1.1B4
2 2
cells using the shift assay format.
Both compounds 14 and 19 also proved to be selective potentiators of GLP1(9-36)NH . They
2
did neither alter the biological responses of GLP-1(7-36)NH nor of oxyntomodulin in the
2
recombinant cell line system (data not shown). Furthermore, they had no effect on cAMP
accumulation, when tested alone in HEK293 cells overexpressing the GLP-1R, in the parental
cell line or in HEK293 cells overexpressing the glucagon receptor either alone or in the
presence of GLP1(9-36)NH . Additional assays were performed using cells overexpressing
2
the human GIP-receptor and the human VIPR2. Both compounds were found inactive, when
tested tested for agonistic activity and for their effect as allosteric modulator using GLP1(9-
41
3
6)NH , GIP.
2
The good selectivity displayed by both compounds in our standard receptor and ion channel
panels and against other typical diabetes-relevant GPCRs, such as TGR5 or GPR119 (see
supporting information for full data), prompted us to investigate these compounds in relevant
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
functional assays like the glucose-stimulated insulin secretion (GSIS) assay in dispersed rat
and mouse pancreatic islets.
As shown in Figure 5, compound 14 significantly improved the GLP1(9-36)NH -mediated
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
glucose-stimulated insulin secretion in dispersed rat pancreatic islets. This GSIS potentiation
effect was much more pronounced at higher concentrations of the peptide ligand. However,
when higher concentrations (10 µM) of compound 14 were tested, its ability to potentiate the
effect of 10 nM GLP1(9-36)NH vanished and even turned into an inhibitory effect in the
2
absence of the peptide ligand GLP1(9-36)NH2 (Figure 5b).42 In contrast, compound 19
showed a distinct behavior in this assay. It robustly stimulated GSIS in dispersed rat
pancreatic islets even at high concentration (10 µM) in the presence of 10 nM GLP1(9-
3
6)NH to an extent, which is comparable to the positive control (10 nM GLP-1 (7-36)NH )
2 2
(Figure 4c). Moreover, when tested alone, it did not show any inhibitory effect. It is
important to note, that the insulinotropic response in this assay was absent at low glucose
concentrations (see supporting information for details).
ACS Paragon Plus Environment
14

Page 15 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Glucose-stimulated insulin secretion (GSIS) from dispersed rat islets for
PAM-compounds 14 (a, b) and 19 (c). Dispersed islets were incubated with high (15
mM) glucose in the presence of vehicle control (DMSO) +/- respective compound (1
µM) or GLP-1 ligand (1/10/100 nM) +/- respective compound (1 µM). Diluent for
GLP-1 ligand was DMSO. N ≥ 9 / condition. **p ≤ 0.001 vs. respective control;
****p ≤ 0.00001 vs. respective control. For b) and c), incubation was done +/-
respective compound (10 µM) or GLP-1 ligand (10 nM) +/- respective compound (10
µM).
To investigate whether the inhibitory effect on the glucose-stimulated insulin secretion
observed with compound 14 at high concentration is an effect, which is not mediated by
43
GLP-1R, we tested 14 in mouse islets derived from GLP-1R(-/-) mice (Figure 6). In this
model, neither GLP-1(7-36)NH nor GLP1(9-36)NH induced the insulinotropic response
2
2
either at low or high glucose. Compound 14 still displayed an inhibitory effect on the insulin
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
secretion, when tested in the presence or in the absence of GLP1(9-36)NH in a magnitude
2
comparable to the one observed in the rat islets. This finding suggests that the inhibitory
mechanism for compound 14 is likely linked to an off-target effect.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Effect on GSIS on islets of GLP-1(-/-) CD-1 mice. Dispersed mouse islets
were incubated at low (3 mM) and high (15 mM) glucose with: 100 nM GLP-1(7-
3
3
6)NH , 100 nM GLP1(9-36)NH ; 10 µM compound 14 and 100 nM GLP1(9-
2 2
6)NH + 10 µM compound 14.
2
As further selectivity testing of compound 14 did not reveal a single receptor or ion channel
responsible for this inhibitory effect, we hypothesized that it was possibly due to a
polypharmacology effect causing a counter-regulatory response in the insulin secretion
2
+
pathway. Indeed, when testing the influence of compound 14 alone in a Ca mobilization
2+
assay in the mouse pancreatic -cell line Min6-c4, it blocked glucose-mediated Ca influx in
a similar manner as the known KATP opener diazoxide, suggesting a detrimental influence on
metabolism-secretion coupling upstream of insulin granule exocytosis and independent of the
GLP-1 pathway. This phenotypic high-throughput assay allowed us then to efficiently
ACS Paragon Plus Environment
16

Page 17 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
exclude other compounds from further profiling that abrogated glucose-stimulated Ca2+
influx like compound 14.
Indeed, as shown in Figure 7, compound 19 tested under similar conditions did not inhibit
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2+
Ca mobilization even at the highest concentration, prompting us to evaluate its potential in
subsequent in-vivo studies.
2
+
Figure 7. Effect of compounds 19 (left) and 14 (right) on Ca mobilization. LG: Low
glucose (2.5 mM); HG: high glucose (15 mM), DIA: KATP opener diazoxide.
Ligand binding mode hypothesis in GLP-1R
The X-ray structure of the full-length GLP-1R with a peptidic agonist (peptide 5 with
44
45
modified GLP-1 residues 8-17, PDB 5NX2, resolution 3.7Å) in the active conformation
was used to derive binding hypothesis for the 3,4,5,6-tetrahydro-1H-1,5-
a
epiminoazocino[4,5-b]indole scaffold. However, due to the lack of a co-crystal structure and
dedicated GLP-1R mutants, this is only a hypothesis to illustrate our reasoning for ligand
optimization, while other potential binding modes for this series cannot be excluded.
The following arguments suggest that our series binds to a region overlapping with the
orthosteric pocket: 1) The ligands exhibit a positive allosteric effect with addition of the
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
truncated peptide GLP1(9-36)NH , but not in the presence of the full GLP-1(7-36)NH with
2
2
amino acids His7-Ala8 close to the orthosteric site, thereby blocking the entrance into this
area; 2) activation of GLP-1R is only possible by contacting both the ECD with the truncated
GLP-1 peptide derivative and the 7TM domain; 3) the binding site hypothesis allows to
sterically accommodate our ligand series.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
6
For compound 9, Induced Fit Docking (IFD) produced a hypothesis consistent with SAR
(Figure 8). The Boc tertiary-butyl group is possibly located next to Leu388 (helix 7). The
exocyclic ligand amide oxygen possibly interacts with Arg190 (helix 2) critical for GLP-1
binding affinity.44,47,48 The difluoro-cyclohexyl-group might be located in a hydrophobic
pocket at the bottom near Tyr241 (helix 3). The piperazinone amide might interact with
Gln234 (helix 3) and Met233 (helix 3), an important residue for peptide binding.Error! Bookmark
not defined.,Error! Bookmark not defined.
The indole is possibly located in the upper pocket area close to
Trp306 (helix 5) and Arg380 (helix7). The postulated binding modes for compounds 9, 14
and 19 are similar with the carboxylic acid in compound 19 possibly contacting the Lys197
(helix 2) side chain.
This model agrees to the GLP-1R active conformation with the indole oriented towards the
outward-shifted helix 6. It is also consistent with the Cryo-Electron Microscopy (Cryo-EM)
4
9
structure of activated GLP-1R with GLP-1 and G (5VAI, 4.1Å). Removing the first
s
residues of GLP-1 (His7-Ala8) in the Cryo-EM structure would allow accommodating the
ligands.
ACS Paragon Plus Environment
18

Page 19 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8. Postulated binding model for 9 (green carbon atoms) from Induced Fit
Docking into the GLP-1R X-ray structure from 5NX2 (resolution 3.7 Å). Potential
hydrogen bonds are indicated by dotted yellow lines,  interactions as dotted cyan
lines and cation- interactions as dotted green lines.
In-vitro ADME and in-vivo pharmacokinetic profiling of compound 19
Based on encouraging in-vitro pharmacological data of compound 19 in primary tissue, we
evaluated its ADME profile to assess its potential for in-vivo studies as tool compound
towards a proof-of-concept (PoC). Relevant in-vitro data for this compound are summarized
in Table 2. Compound 19 is metabolically stable in human and mouse liver microsomes and
exhibits high cell permeability in Caco-2 TC7 cell transwell experiments (Papp 32 * 10-7
-7
cm/sec with a threshold of 20 * 10 cm/sec for highly permeable compounds) under standard
test conditions conditions (pH 6.5 and 0.5% BSA apical and pH 7.4 and 5% BSA
basolateral). Furthermore, this compound did not inhibit CYP3A4, displays good aqueous
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
solubility and is chemically stable under different conditions including GSH and mouse
plasma. In mouse hepatocytes, compound 19 displays an intermediary clearance (0.145
6
mL/h/10 cells), which corresponds to 20% of hepatic blood flow.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. In-vitro and in-vivo profiling data for compound 19.
Physicochemical data
MW
Value
585.64
1.7
Unit
Da
logD (pH 7.4, 25°C)
clogP
4.5
Solubility (pH 7.4)
Chemical stability (pH 1.1)
Chemical stability (pH 7.4)
Chemical stability (GSH)
>853
98
µM
%
99
%
100
%
In-vitro ADME data
Value
Unit
Metabolic lability in microsomes (human)
Metabolic lability in microsomes (mouse)
20
5
%
%
CYP3A4 inhib. (IC , Substrate Midazolam)
>50
>50
32
µM
50
CYP3A4 inhib. (IC , Substrate Testosteron)
µM
5
0
Caco-2 Papp
hERG inhibition (IC , Patch clamp)
10-7 cm/s
>30
0.184
0.145
98
µM
5
0
Hepatic clearance (human)
Hepatic clearance (mouse)
Plasma stability (mouse)
mL/h/10e+06 cells
mL/h/10e+06 cells
%
ACS Paragon Plus Environment
20

Page 21 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In-vivo Pharmacokinetic data.
Treatment:
Mouse
Rat
Rat +
Rifampicin
1130
Unit
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
mg/kg i.v. / 10 mg/kg p.o.
Cmax (p.o.)
227
7.2
106
9.0
ng/mL
%
F (Bioavailability)
AUClast (p.o.)
Vdiss (i.v.)
100
109
110
2.29
7.06
7340
h*ng/mL
L/kg
0.934
6.52
0.815
1.15
Clearance (i.v.)
L/h/kg
However, after oral administration, the obtained mouse in-vivo pharmacokinetic (PK) data
revealed extensive clearance (6.52 L/h/kg) with associated low oral bioavailability (F = 7.2
%) and a moderate volume of distribution (V ). Relevant data for mouse and rat PK studies
diss
after oral and i.v. administration are also summarized in table 4. The half-life in mouse
plasma was low (0.29 h) and high concentrations of metabolically unchanged compound 19
were recovered in bile (rat fistula study, 74% of i.v. dose recovered in bile), suggesting that
biliary excretion plays an important role in the clearance process for this compound.
In order to understand the mechanism for biliary clearance of this compound, we performed
transporter studies employing HEK293 cells overexpressing the human organic anion
transporters hOATP1B1 (organic anion transporting polypeptide 1B1, SLCO1B1) and
hOATP1B3 (organic anion transporting polypeptide 1B3, SLCO1B3).50 In particular
OATP1B1 is one of the most highly expressed uptake transporters in human liver, and there
5
1
are numerous reports regarding OATP1B1-mediated clinical drug-drug interactions. The
structure of 19, containing an alkyl carboxylic acid moiety and relatively high molecular
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
weight, suggested that this compound could potentially be a substrate for these liver
transporters.52,53 Therefore we applied our internal in-silico transporter profile, a collection of
5
4
2D-QSAR models built from internal compounds and data, to this structure. As expected,
compound 19 was predicted to interact with OATP1B1 and OATP1B2 plus also classified as
substrate for P-gp.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Subsequently compound 19 was experimentally identified as a substrate of hOATP1B1 and
hOATP1B3 hepatic uptake transporter by a time-dependent increase of uptake in the
overexpressing cell lines compared to the control HEK293 wild type cells (background) and
with uptake ratios ranging from 6 to 22-fold, as obvious from inspection of Figure 9.
Figure 9. Compound 19 was identified as substrate of hOATP1B1 and hOATP1B3
transporter by a time-dependent incubation of 1 µM compound 19 with HEK cells
overexpressing hOATP1B1 (left) and hOATP1B3 (right) or HEK293 wild type cells
(control).
Additionally, we examined the efflux behavior of compound 19 in the presence and absence
of efflux transporter inhibitors. Studies with Caco-2 TC7 cells revealed that, under isocratic
ACS Paragon Plus Environment
22

Page 23 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
conditions (pH 7.4 and 0.5% BSA in apical and basolateral compartment), compound 19
shows a much lower permeability (0.18 * 10-7 cm/sec) and a high efflux ratio (ER = 247).
In presence of the efflux-transport-inhibitor cyclosporine A, the efflux could be inhibited (ER
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
=
2). This finding shows that compound 19 is indeed a substrate for efflux transporters, most
likely P-glycoprotein (P-gp) or MRP2 (multidrug resistance protein 1, MDR1, multidrug
resistance-associated protein 2, ABCC2),55,56 two important proteins in the cell membrane
acting as ATP-dependent efflux pump with broad substrate specificity. Additionally, this
result highly suggests an involvement of P-gp and MRP2 for active elimination of compound
1
9 in hepatocytes, as both transporters are expressed in the canalicular membrane.57
Although we could then identify the structural motif leading to a lowering of liver transporter
recognition, the resulting compounds significantly lost their GLP-1R PAM activity (data not
shown). Since a direct use of compound 19 for in-vivo proof-of-concept studies was
precluded due to its poor pharmacokinetic profile, we then envisioned a study design which
would still allow the evaluation of this compound in-vivo. Since the exposure of compound
1
9 was restricted by interactions with transporters in the small intestine, but especially in the
liver, we hypothesized that its co-administration with the known OATP/MRP2 inhibitor
Rifampicin would have a beneficial effect on plasma exposure of compound 19.58,59,60,61
Indeed, the mean exposure of compound 19 could be increased by a factor of 6 after i.v. and a
factor of 85 after p.o. administration, when dosed to rat which had been pretreated with
Rifampicin. This observation clearly supports our hypothesis that the above-mentioned
transporters are responsible for restricted absorption and high hepatic clearance of compound
1
9. Under these modified conditions, the plasma exposure of compound 19 was found
sufficient to allow for an in-vivo proof-of-concept study (Table 4).
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The rat PK study in presence and absence of Rifampicin was performed to confirm our in-
vitro data from overexpressing cell lines, showing that the clearance of compound 19 is
mainly driven by transporters especially of the OATP family in rodents. As this was clearly
proven in a rat model, we selected an equal dose of 60 mg/kg p.o. in the pharmacological
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
2
mice species. Moreover, the PK of Rifampicin in mice is described in the literature. Here
even at a lower dose of 10 mg/kg, Cmax is reached after 1.5 h with a total concentration of 13
µM. This value is well above the IC values described for OATPs. The higher dose was used
5
0
in our study to inhibit the efflux transporters in the intestine in parallel, which cannot easily
be predicted, as the intestinal concentration is not known.
In-vivo pharmacology for compound 19
As described above, compound 19 was applied in the presence of Rifampicin for a proof-of-
concept study in ob/ob mice, which is a well-accepted disease model for T2DM (Figure 10).63
Rifampicin was orally dosed 2.5 hours prior to the oral glucose-tolerance test (oGTT) to
establish appropriate blood plasma concentration (data not shown). Subsequently, one hour
prior to the oGTT, compound 19 was orally administered and, when indicated, GLP1(9-
3
6)NH was intraperitoneally co-administered 15 minutes prior to the oGTT. For the oGTT,
2
mice were fasted overnight and 2 g/kg glucose challenge was used.
ACS Paragon Plus Environment
24

Page 25 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 10. Proof-of-concept study for compound 19 in 11 weeks old ob/ob mice
(n=7), an animal model of T2DM. Compound 19 required the presence of Rifampicin
and GLP1(9-36)NH . GLP1(9-36)NH was given intraperitoneally. In the presence of
2
2
both GLP1(9-36)NH and compound 19, a significant reduction in blood glucose
2
concentration was observed.
In all models used, compound 19 alone did not show any effect on blood glucose
concentration in the absence of Rifampicin and/or GLP1(9-36)NH . Rifampicin at 60 mg/kg
2
alone or in combination with compound 19 at 30 mg/kg orally dosed to ob/ob mice (n=7) had
no effect on the blood glucose concentration in the oGTT. However, the intraperitoneal
administration of GLP1(9-36)NH to the combination of Rifampicin and compound 19
2
evoked a significant reduction in blood glucose concentration, as shown in Figure 11. In
contrast, the peptide GLP1(9-36)NH alone at a concentration of 4.5 mg/kg had no effect
2
with or without addition of Rifampicin on blood glucose concentration (see Supporting
Information). Furthermore, a dose-response relationship for the peptide GLP1(9-36)NH and
2
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compound 19 was established in normoglycemic C57Bl/6 mice (Figure 11). At all doses
tested for the peptide GLP1(9-36)NH , compound 19 evoked a significant dose-dependent
2
reduction of the blood glucose concentration during the oGTT.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 11. Dose-response relationship for compound 19 and the less active peptide
GLP1(9-36)NH in 11 weeks old normoglycemic C57Bl/6 mice (n=6). Compound 19
2
required the presence of Rifampicin and GLP1(9-36)NH for a biological effect.
2
GLP1(9-36)NH has been given intraperitoneally at 0.3, 1, 3, and 4.5 mg/kg.
2
Compound 19 evoked at all inactive GLP-1 doses a significant reduction in blood
glucose concentration.
To finally demonstrate that the in-vivo effects of compound 19 strongly depend on the GLP-1
-/-
receptor, an oGTT in the GLP-1R mouse was performed, as shown in figure 12. To explore
-/-
the effect of compound 19 in GLP-1R mice, the compound was tested in the presence of
Rifampicin at 60 mg/kg and in the absence or in combination of intraperitoneally applied
ACS Paragon Plus Environment
26

Page 27 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
GLP1(9-36)NH (4.5 mg/kg). Rifampicin and compound 19 were both applied orally in this
2
setting.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 12. Demonstration of the GLP-1R dependent effect of compound 19 in 10
-/-
weeks old GLP-1R mice (n=6). Compound 19 does not induce a significant blood
glucose reduction in GLP1R-/- mice in the presence of Rifampicin and GLP1(9-
3
6)NH at the highest dose tested.
2
-/-
Under all the conditions tested in GLP-1R mice, compound 19 did not show any effect on
blood glucose concentration, strongly supporting our hypothesis that the interesting
pharmacological effect of this positive allosteric modulator 19 is directly mediated by its
interaction with the GLP-1 receptor (Figure 12).
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Conclusion
The glucagon-like peptide 1 (GLP-1) hormone is secreted from intestinal L-cells after food
intake and plays a crucial role for blood glucose regulation and control of appetite and body
weight. Its activation of GLP-1R triggers glucose-dependent insulin secretion and leads to
suppression of glucagon secretion in the pancreas. For treatment of T2DM, several
approaches, like GLP-1 mimetics and DPP-IV inhibitors, are directed to enhance activation
of GLP-1R. In this contribution we report a new compound as positive allosteric modulator
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(PAM) for this important GPCR, namely compound 19 as potent, non-covalent and selective
PAM based on a 3,4,5,6-tetrahydro-1H-1,5-epiminoazocino[4,5-b]indole scaffold. This PAM
is able to turn the significantly less active metabolite GLP1(9-36)NH into a potent ligand
2
able activate GLP-1R in a similar manner as the active GLP-1 peptide itself does.
Our best compound 19 demonstrates high activity and selectivity in various functional assays.
Due to its favorable in-vitro ADME and ex-vivo profile, it was further evaluated in-vivo. We
could explain the observed high biliary clearance in animal PK studies and attributed this to
interaction of compound 19 with human organic anion transporters hOATP1B1 and
hOATP1B3. This finding required co-administration of the transporter inhibitor Rifampicin
for a successful in-vivo proof-of-concept study of compound 19 as tool for our proposed
mode of action.
Unfortunately, the need of exogenous addition of GLP1(9-36)NH to elicit a significant
2
response on glucose homeostasis experiments in-vivo clearly shows that the PAM ability of
our best compound 19 is not yet sufficient to induce a significant response under
physiological conditions without addition of this exogenous peptide. The requirement of
exogenous GLP1(9-36)NH for a successful in-vivo effect might not be so surprising, since
2
the EC -shift observed for GLP1(9-36)NH in the presence of 10 µM of our best PAM 19 in
5
0
2
ACS Paragon Plus Environment
28

Page 29 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pancreatic 1.1B4 cells was not sufficient to match the EC50 value of GLP1(7-36)NH2.
GLP1(7-36)NH is approximately 10,000 more potent than its less active counterpart
2
GLP1(9-36)NH Taking into account the higher concentration of the less active metabolite in
2
.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
vivo, a potentiation by a factor of ~2,500 is anticipated to be required at an in-vivo relevant
compound concentration to achieve full therapeutic utility.
Therefore even if compound 19 has proven to be a valuable tool to proof the feasibility of
GLP1(9-36)NH potentiation in-vivo, the validity of this approach under physiological
2
conditions remains an open question. The data presented herein suggest that a significant
amount of optimization would still be needed to arrive at a successful clinical candidate. In
our opinion, it is questionable, whether the challenging goal of identifying a non-covalent,
orally available GLP1(9-36)NH potentiator that works under physiological conditions can be
2
reached at all.
Experimental section
Biological Assays.
2
+
64
Min6 Ca mobilization assay. MIN6-c4 cells (Osaka University) were plated at a density
4
of 5x10 cells per well into black 96-well plates and cultured for 20-24 h at 37 °C and 5%
CO . For cell loading, culture supernatants were aspirated and 100 µL assay buffer (Krebs-
2
Ringer buffer, 10 mM HEPES, 0.1% BSA, 2.5 mM glucose) and an equal volume of Calcium
6
dye (FLIPR® Calcium 6 Assay Kit, Molecular Devices, R8191) dissolved in the same
buffer according to instructions of the manufacturer were added to each well. Cells were
incubated for 70 min at 37 °C / 5% CO and equilibrated for additional 10 min at room
2
temperature in the dark. To assess the effect of test compounds on glucose-mediated increase
2
+
in intracellular Ca , a volume of 50 L assay buffer containing 75 mM glucose (resulting in
ACS Paragon Plus Environment
29