Discovery of potent, reversible MetAP2 inhibitors via fragment based drug discovery and structure based drug design - Part 2

Article abstract of DOI:10.1016/j.bmcl.2016.04.072

Methionine aminopeptidase-2 (MetAP2) is an enzyme that cleaves an N-terminal methionine residue from a number of newly synthesized proteins. This step is required before they will fold or function correctly. Pre-clinical and clinical studies with a MetAP2 inhibitor suggest that they could be used as a novel treatment for obesity. Herein we describe the discovery of a series of pyrazolo[4,3-b]indoles as reversible MetAP2 inhibitors. A fragment-based drug discovery (FBDD) approach was used, beginning with the screening of fragment libraries to generate hits with high ligand-efficiency (LE). An indazole core was selected for further elaboration, guided by structural information. SAR from the indazole series led to the design of a pyrazolo[4,3-b]indole core and accelerated knowledge-based fragment growth resulted in potent and efficient MetAP2 inhibitors, which have shown robust and sustainable body weight loss in DIO mice when dosed orally.

Full text of DOI:10.1016/j.bmcl.2016.04.072

Bioorganic & Medicinal Chemistry Letters 26 (2016) 2779–2783
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of potent, reversible MetAP2 inhibitors via fragment based
drug discovery and structure based drug design—Part 2
a
a
a
Christopher McBride a, , Zacharia Cheruvallath , Mallareddy Komandla , Mingnam Tang ,
Pamela Farrell b, J. David Lawson c, Darin Vanderpool b, Yiqin Wu b, Douglas R. Dougan d,
Artur Plonowski b, Corine Holub b, Chris Larson b
⇑
a Medicinal Chemistry, Takeda California, United States
b Biological Sciences, Takeda California, United States
c Computational Sciences, Takeda California, United States
d Structural Biology, Takeda California, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Methionine aminopeptidase-2 (MetAP2) is an enzyme that cleaves an N-terminal methionine residue
from a number of newly synthesized proteins. This step is required before they will fold or function cor-
rectly. Pre-clinical and clinical studies with a MetAP2 inhibitor suggest that they could be used as a novel
treatment for obesity. Herein we describe the discovery of a series of pyrazolo[4,3-b]indoles as reversible
MetAP2 inhibitors. A fragment-based drug discovery (FBDD) approach was used, beginning with the
screening of fragment libraries to generate hits with high ligand-efficiency (LE). An indazole core was
selected for further elaboration, guided by structural information. SAR from the indazole series led to
the design of a pyrazolo[4,3-b]indole core and accelerated knowledge-based fragment growth resulted
in potent and efficient MetAP2 inhibitors, which have shown robust and sustainable body weight loss
in DIO mice when dosed orally.
Received 1 March 2016
Revised 22 April 2016
Accepted 23 April 2016
Available online 25 April 2016
Keywords:
Methionine aminopeptidase-2
MetAP2
Metalloprotease
Fragment-based drug discovery
FBDD
Ó 2016 Elsevier Ltd. All rights reserved.
Structure-based drug design
SBDD
Pyrazolo[4,3-b]indoles
In our previous communication1, we described the rationale for
our interest in MetAP2 as a target to treat obesity and disclosed an
indazole fragment hit with moderate affinity (1, MetAP2 pIC50 of
5.6), and a high LE and lipophilic ligand efficiency (LLE) due to its
small size and low lipophilicity. In that communication we identi-
fied interactions vital to potency, most importantly those at 6-posi-
tion of the indazole core, which can generate gains in affinity of
>100 fold over the unsubstituted core (Fig. 1, compound 2 vs 3).
In this Letter, we disclose the structure based design of a new
scaffold as part of our efforts in identifying multiple chemotypes
for our program. This led to the identification of several potent
pyrazolo[4,3-b]indoles, as selective, orally bioavailable, and reversi-
ble MetAP2 inhibitors.
methionine of the endogenous ligands for MetAP2. Filling this
pocket yields the aforementioned potency gains of up to 100-fold.
Installing an aryl group at the 4 position of the indazole core fills
a hydrophobic cleft between Tyr444 and His-339 and boosts
potency (5–10 fold).
In parallel to our work on the indazoles, we sought a novel
chemical scaffold and designed a pyrazolo[4,3-b]indole core which
we hypothesized could satisfy all of the key interactions seen in the
indazole series (Fig. 2, indazole core overlaid with model of pyra-
zolo[4,3-b]indole core). By utilizing the same pyrazole warhead
we expected to maintain the interactions with the metal ions in
the active site of MetAP2. The size of the tricyclic core would cause
the phenyl ring to protrude even deeper into the lipophilic pocket,
though still leaving room for small substitutions around the ring (7-
fluorinated example shown). Additionally, this new scaffold offered
the opportunity for substitution at N4, which would point in nearly
the same vector as the indazole 4-substitutions.
The co-crystal structure of an indazole1 bound to MetAP2
showed the N2-nitrogen of the indazole ring coordinates with
one of the active site metal ions while the N1-nitrogen makes a
water mediated interaction with the other active site metal. The
6-position substituent on the indazole ring is oriented to fill the
adjacent hydrophobic site that is typically filled by the terminal
Previous work by Pudlo et al. utilized a thermally decomposed
azide for nitrene insertion and ring formation of fused aromatic
indole heterocycles.2 We utilized a similar strategy to access our
target molecules. A representative synthesis for this series is
⇑
Corresponding author.
http://dx.doi.org/10.1016/j.bmcl.2016.04.072
0960-894X/Ó 2016 Elsevier Ltd. All rights reserved.

2780
C. McBride et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2779–2783
OH
N
Ph
N
N
Br
N
H
R
H
1
2
3
R = CF3
R = H
pIC50 5.6
LE 0.69
LLE 3.2
pIC50 7.7
LE 0.55
LLE 3.9
pIC50 5.3
LE
0.48
2.3
LLE
Figure 1. Indazole series elaboration.
described in Scheme 1. Starting with the diazotization of a
substituted 2-bromoaniline (A) followed by treatment with sodium
azide affords a substituted 1-azido-2-bromobenzene (B). This inter-
mediate was then cross-coupled with N1-THP protected pyrazole-
5-boronic ester and then thermally cyclized to the substituted
N1-THP protected pyrazolo[4,3-b]indole (D). The N4-aryl substitu-
tion was installed using copper thiophene carboxylate to catalyze
an Ullmann reaction, which was followed by removal of the THP
group with methanolic HCl to afford the desired product (F).3
Screening3 of the unsubstituted pyrazolo[4,3-b]indole (5)
showed that it had lower potency, though owing to the lower
MW and lipophilicity, still displayed LE and LLE which would make
this an attractive starting point (4, Fig. 3). As shown in Figure 2,
modeling4 also suggested that there was room for small substitu-
tions at the 6 and 7-positions of the tricyclic core. We were encour-
aged that when a fluorine atom was installed at the 7-position of
the tricyclic scaffold (6) we measured a greater than 10 fold
increase in potency and saw a corresponding boost in efficiency.
To explore the SAR at the 6 and 7 positions, we prepared a small
set of 4N-arylated cores with different halogen substitutions and
tested them in our enzymatic and cellular assays.3 The HUVEC
assay used is not a proliferation assay, but rather verifies target
engagement by measurement of the accumulation of N-Met 14-3-
From our previous work on the indazole series of MetAP2 inhi-
bitors, we had a wealth of data (SAR, in vitro safety, microsomal
stability, etc.) which has been discussed in the first part of this pub-
lication series (we found that aryl substitutions with ortho sub-
stituents were the most potent and stable in human liver
microsomes). As the vector from the 4 position of both cores points
in nearly the same direction, we were able to use knowledge-based
fragment growth to quickly elaborate this series to a number of
analogs (Table 2), many with suitable properties to study them
in vivo.
While the indazole SAR was useful in prediction of enzymatic
affinity, prediction of the cellular potency was not as straight for-
ward. Note that while compounds 10–12 are structurally quite sim-
ilar, and the pIC50’s are in the same range, the pEC50’s vary
somewhat depending on the placement of a methyl group. This
effect became more pronounced in cases such as compounds 13
and 14 which performed identically in the enzyme assay but
diverged in the cell assay by 100 fold.
It is of note that as the inhibitors grew modestly, the CYP1A2
inhibition exhibited by the smaller compounds from Table 1 was
alleviated. The smallest compound (15) displays potent inhibition
of CYP1A2, while a slight larger compound such as 10 showed mod-
erate inhibition while the rest showed none. This can be explained
by the relatively small binding pocket of CYP1A25 which cannot
accommodate larger ligands.
3c which is a client protein of MetAP2. While the fluoro and chloro
substitutions tested showed similar activity against MetAP2, the 7-
fluoro analog (7) showed superior cellular activity so this substitu-
tion was selected for use going forward (Table 1). Additionally,
these relatively small inhibitors do display potent inhibition of
CYP1A2.
A co-crystal structure6 of compound 10 confirmed our presump-
tion of its binding and interactions with the protein (Fig. 4). The
pyrazole warhead N2 interacts directly with one metal ion (M1)
while the N1–H makes a water mediated interaction with the other
Figure 2. Overlay of pyrazolo[4,3-b]indole model with indazole co-crystallized with MetAP2 (pdb code: 5JI6).

C. McBride et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2779–2783
2781
R
R
O
R
O
a, b
c
d
O
B
+
N
Br
Br
O
N
N
N
NH2
N3
N3
A
B
C
R
N
R
e, f
+
Ar-X
Ar
N
HN
O
NH
N
N
E
F
D
Scheme 1. Synthesis of 4-substituted pyrazolo[4,3-b]indoles. Reagents and conditions: (a) NaNO2, HCl; (b) NaN3 70–88% yield over 2 steps; (c) Pd(PPh3)4, Na2CO3, dioxane,
heat 60% yield; (d) dichlorobenzene, 170 °C, 2 h. 30% yield; (e) CuTC, Cs2CO3, DMF, 180 °C, 2 days; (f) MeOH, HCl, 10–70% yield over 2 steps.
H
N
H
N
N
N
N
H
N
F3C
N
H
N
H
F
4
5
6
pIC50 7.3
pIC50 6.0
pIC50 7.1
LE 0.74
LLE 5.5
LE
LLE
0.77
5.1
LE
LLE
0.68
4.6
Figure 3. Progression from indazole fragment to pyrazolo[4,3-b]indole core.
Table 1
SAR of 6 and 7 position halogenations
N
R2
R1
N
N
N
H
Compd
Structure (R1, R2)
clogD
MetAP2 pIC50
LE
pEC50
CYP1A2 pIC50
7
8
9
F, H
F, F
H, Cl
1.2
1.1
1.7
7.9
8.1
8.3
0.57
0.55
0.59
7.0
5.9
5.0
6.9
7.3
7.2
metal ion (M2). The phenyl ring of the core is well situated to make
an edge/face interaction with Tyr-444. The 7-fluoro substituent
protrudes into the lipophilic pocket occupied by the thioether of
the endogenous ligands, making the ligand very shape complemen-
tary with the target. The aza-benzimidazole substituent at the N-4
position points into a hydrophobic space between Tyr-444 and His-
339.
We elected to investigate the pharmacokinetic profile of com-
pound 10 due to its good enzyme and cellular potency, and good
stability in microsomes (rat Eh < 0.19, human Eh < 0.28) and hepato-
cytes (human Eh 0.27). When dosed orally at at 2.9 mg/kg in Spra-
gue Dawley rats, this compound’s pharmacokinetic profile showed
acceptable exposure (AUC = 2237 ng*h/ml), an oral half-life of 6 h,
clearance of 17.2 ml/min/kg, volume of distribution of 1597 ml/
kg, and an oral bioavailability of 70%.
With acceptable pharmacokinetics, compound 10 was further
evaluated in a DIO-mouse model for obesity. Dose-dependent body
weight loss (BWL) was observed when DIO mice were administered
10 orally for 7 days at 30 mg/kg qd (Fig. 5). At this dose, we
observed a 4% reduction in body weight relative to the vehicle.
Some BWL was seen in the vehicle in this study. This is a typical
response of DIO mice to the stress of oral gavage. For some com-
pounds we had the time acclimate the animals to the stress of oral
gavage for a few weeks, until the DIO weights were not changing
due to the stress of oral gavage before running the study. In these
studies we saw similar results in the BWL with respect to the vehi-
cle group. Importantly, one can see very little difference in the food
intake between dosed and vehicle cohorts, one indication that the
compounds are not toxic at this dose. Additionally, we monitored
the activity levels of the animals and there was no difference
between the vehicle and dosed groups.

2782
C. McBride et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2779–2783
Table 2
Selected N-aryl pyrazolo[4,3-b]indoles
R
N
F
N
N
H
*
Compd
Structure (R)
clogD
MetAP2 pIC50
8.3
LE
pEC50
7.9
CYP1A2 pIC50
5.5
Human Eh
<0.28
H
N
N
10
1.7
0.49
N
H
N
N
11
2.2
8.0
0.47
6.6
<4.3
<0.28
0.33
N
N
N
12
13
1.8
0.9
8.0
8.2
0.46
0.45
7.6
5.7
4.7
5.1
N
O
OH
OH
N
H
O
14
15
1.4
1.1
8.2
8.1
0.43
0.52
7.7
7.5
5.1
6.8
<0.28
0.76
N
H
N
NH2
Cl
*
Stability in human liver microsomes at 1 lM. Data reported as extraction ratios (Eh).
Figure 4. Crystal structure of 10 with MetAP2 (pdb code: 5JFR).
Compound 10 was evaluated in a Ricerca Comprehensive Phar-
good pharmacokinetic properties in pre-clinical species and have
shown efficacy in vivo.
macological Profile panel including 100 biological targets (GPCRs,
ion channels, transporters and enzymes) and a panel of proteases
and was found to have a no significant off target activity at
Acknowledgements
10
lM (<50% inhibition).
In summary, using a FBDD approach, a 6-substituted indazole
We acknowledge the ADME group for generating the microso-
mal and solubility data, Jennifer Matuszkiewicz for helping to gen-
erate the cellular data, and Lisa Rahbaek for generating the in vivo
PK data. The staff of the Berkeley Center for Structural Biology is
gratefully acknowledged for support of beam line 5.0.3 at the
Advanced Light Source. The Advanced Light Source is supported
core was identified as a potent inhibitor of MetAP2. From this frag-
ment hit, SBDD was used to design a pyrazolo[4,3-b]indole core
which was elaborated using an accelerated knowledge-based
approach to deliver small, potent, selective, and reversible MetAP2
inhibitors. Compounds from this series have been developed with

C. McBride et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2779–2783
2783
0
-2
4
3
2
1
0
-4
-6
-8
-10
-12
0
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
time (day)
time (day)
Vehicle (0.5%MC)
10 30mg/kg qd x7
Figure 5. Body weight loss resulting from oral dosing of 10.
Inhibitor Tablets. The lysate was centrifuged, and clarified supernatant was batch
bound with 10 ml of Probond Ni resin (Invitrogen). The protein was then eluted
with buffer containing 25 mM HEPES (pH 7.6), 150 mM NaCl, and 250 mM
imidazole. After tag cleavage, the protein was further purified by size-exclusion
chromatography utilizing a Superdex 200 column equilibrated in 25 mM HEPES
(pH 7.6), 150 mM NaCl, and 5% Glycerol. For compound co-crystallization
experiment, an aliquot of the frozen protein was incubated with the compound
(final concentration 1 mM) and MnCl2 (final 0.3 mM) in 4 °C overnight and then
further concentrated the protein to 18 mg/ml.
Crystals suitable for data collection are grown by vapor diffusion in sitting drops
at 4 °C with 50 nL of protein solution and 50 nL of reservoir solution containing
38–42% MPD and 0.1 M HEPES pH 6.8–7.2. Single crystals were immersed in
mother liquor solution containing 22% ethylene glycol for cryoprotection and
flash frozen. X-ray diffraction data to a resolution of 1.7 Å was collected on an
ADSC Q315R detector from single cryogenically protected crystals. Crystals of
Metap2 complex belong to the orthorhombic space group C2221 with unit cell
dimensions a = 89.64 Å, b = 100.82 Å, c = 99.44 Å, and contain one enzyme
molecule in the asymmetric unit. Data was reduced using the HKL2000
software package. The structure was determined by molecular replacement
with either MOLREP or PHASER of the CCP4 program suite and refined with the
program REFMAC. Several cycles of model building with XtalView or COOT and
refinement were performed for improving the quality of the model. The
coordinates and structure factors have been deposited in Protein Data Bank
with accession code 5JFR.
by the Director, Office of Science, Office of Basic Energy Sciences, of
the U.S. Department of Energy under Contract No. DE-AC02-
05CH11231.
References and notes
1. Cheruvallath et al., Bioorg. Med. Chem. Lett., preceding article.
2. Pudlo, Marc; Csanyi, Dorottya; Moreau, Fabien; Hajos, Gyorgy; Riedl, Zsuzsanna;
Sapi, Janos Tetrahedron 2007, 63, 10320.
3. Patent applications describing our MetAP2 inhibitors have been published. Full
experimental procedures for synthesis of all analogs in this Letter along with
enzymatic and cellular assay conditions are contained within the patent
application. The screen was run both in the presence of Co and Mn. The data
reported in this communication is from the Mn screen: WO2014039831(A1).
4. Molecular Operating Environment (MOE) software, MOE, Chemical Computing
Group Inc.
5. Sansen, Stefaan; Yano, Jason K.; Reynald, Rosamund L.; Schoch, Guillaume A.;
Griffin, Keith J.; David Stout, C.; Johnson, Eric F. J. Biol. Chem. 2007, 282, 14348.
6. The MetAp2 catalytic domain (residues 110-478) was cloned into the
pFastBacHTb vector, and recombinant baculovirus was generated and
expressed in fusion with an N-terminal 6Â poly-histidine tag and TEV cleavage
site. Large scale production of recombinant protein was carried out in Spodoptera
frugiperda Sf9 cells. The cell pellet was suspended into lysis buffer consisting of
25 mM HEPES (pH 8.0), 1 M NaCl, 20 mM imidazole, 3 Roche cOmplete Protease