Structure-efficiency relationship of [1,2,4]triazol-3-ylamines as novel nicotinamide isosteres that inhibit tankyrases

Article abstract of DOI:10.1021/jm400826j

Tankyrases 1 and 2 are members of the poly(ADP-ribose) polymerase (PARP) family of enzymes that modulate Wnt pathway signaling. While amide- and lactam-based nicotinamide mimetics that inhibit tankyrase activity, such as XAV939, are well-known, herein we report the discovery and evaluation of a novel nicotinamide isostere that demonstrates selectivity over other PARP family members. We demonstrate the utilization of lipophilic efficiency-based structure-efficiency relationships (SER) to rapidly drive the evaluation of this series. These efforts led to a series of selective, cell-active compounds with solubility, physicochemical, and in vitro properties suitable for further optimization.

Full text of DOI:10.1021/jm400826j

Article
pubs.acs.org/jmc
Structure−Efficiency Relationship of [1,2,4]Triazol-3-ylamines as
Novel Nicotinamide Isosteres that Inhibit Tankyrases
Michael D. Shultz,* Dyuti Majumdar, Donovan N. Chin, Pascal D. Fortin, Yun Feng, Ty Gould,
Christina A. Kirby, Travis Stams, Nigel J. Waters,† and Wenlin Shao
Novartis Institutes for Biomedical Research Incorporated, 250 Massachusetts Avenue, Cambridge, Massachusetts 02139, United
States
S
* Supporting Information
ABSTRACT: Tankyrases 1 and 2 are members of the poly(ADP-ribose) polymerase
(PARP) family of enzymes that modulate Wnt pathway signaling. While amide- and
lactam-based nicotinamide mimetics that inhibit tankyrase activity, such as XAV939,
are well-known, herein we report the discovery and evaluation of a novel
nicotinamide isostere that demonstrates selectivity over other PARP family members.
We demonstrate the utilization of lipophilic efficiency-based structure−efficiency
relationships (SER) to rapidly drive the evaluation of this series. These efforts led to a
series of selective, cell-active compounds with solubility, physicochemical, and in vitro
properties suitable for further optimization.
whether the biology of the Wnt pathway and the potential
advantages or limitations of targeting this pathway could be
understood in the absence of an in vivo active compound.
Therefore, we sought to rapidly develop an in vivo active
compound with low off-target activity in order to evaluate
which cancers may be dependent on TNKS activity. With these
requirements in hand, we were motivated to explore hit
prioritization and optimization approaches that would allow us
to generate the desired profile in the most efficient manner
possible.
The use of efficiency indices, also known as composite
parameters, has been promoted as a way to more rapidly
identify and advance quality chemical matter.22,23 At the time
we began this effort, high ligand efficiency (LE)22 and lipophilic
efficiency (LipE)24 scores were used to rationalize high-quality
hits and highly optimized compounds, respectively. To our
knowledge, however, there was no study demonstrating the
real-time use of a single composite parameter to enhance
decision-making throughout the optimization process. The
conventional wisdom is that LE should be used during early hit
identification to maximize potency versus molecular size, and
LipE should be used during later optimization stages to control
for lipophilicity.25 Since we had a variety of hits from our
phenotypic cellular and high-throughput biochemical screens,
in addition to structural and computational tools, we sought to
focus on a single composite parameter to help simplify and
accelerate the decision-making process from start to finish. We
wondered if LipE could be utilized continuously from hit
INTRODUCTION
■
Tankyrase (TNKS) 1 and 2, two members of the poly(ADP-
ribose) polymerase (PARP) family, have been identified as
potential targets for biochemical inhibition of the Wnt signaling
pathway.1 As such, recent efforts to develop low molecular
weight inhibitors of these proteins have generated several new
classes of Wnt pathway modulators.2−6 The canonical
pharmacophore for PARP inhibition is well-known and
typically is achieved via nicotinamide isosteres such as lactams,
pyridone, and pyrimidone motifs that bind in the nicotinamide
binding pocket.7 TNKS proteins appear to differ from other
PARP family members by their ability to be inhibited by
molecules that bind exclusively to the adenosine binding
pocket.5,8−11 The association of TNKS activity with a range of
diseases such as cancer,5,12−17 brain injury,18 viral infections,19
and pulmonary fibrosis20 suggests TNKS as an attractive target
for developing small-molecule antagonists. To this end, we
engaged in efforts to identify a range of chemical scaffolds to
validate TNKS inhibition as a modality for therapeutic
intervention.
At the outset of hit finding efforts, we considered the
requirements for compounds that would be required to validate
TNKS as a cancer drug target.21 Numerous attempts at target
validation efforts have been confounded by the utilization of
compounds with low selectivity for their intended target, where
off-target activities led to erroneous conclusions regarding the
disease dependency on the target in question. Since inhibition
of other PARP family members are known to have anticancer
activity, a pan-PARP inhibitor would not serve our initial
purpose of target validation even though this could be a
desirable clinical profile. It was also unclear at the outset
Received: June 3, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. Early cellular phenotypic and biochemical high-throughput screening hits and partially optimized analogues used to generate
pharmacophore model.
identification, through lead optimization, to candidate selection,
and herein we report how this hypothesis was tested at the
outset of identification and optimization of a novel series of
tankyrase inhibitors.
RESULTS AND DISCUSSION
■
After a phenotypic cellular screen that identified XAV9391 (1)
as an antagonist of the Wnt pathway and tankyrase as the
enzymatic target, we initiated a high-throughput screen (HTS)
against tankyrase that identified multiple series, including novel
adenosine pocket binders (compounds 2−4, Figure 1).26 Our
phenotypic screen identified several distinct series of TNKS1/2
inhibitors, where structural biology efforts allowed us to
generate multiple high-resolution crystal structures prior to
completion of the HTS.4,27,28 Multiple structures (1, 2, and 5)
were aligned by the receptor structure by use of ICM.29−31 We
generated atomic property field (APF) values for the aligned
ligands that formed the basis for the pharmacophore.32 APFs
Figure 2. (A) Alignment of 4 with pharmacophore APF high density
regions (red, hydrogen-bond acceptor; blue, hydrogen-bond donor;
white, aromatic; yellow, hydrophobic). (B) Overlay of 4 with 1 based
on pharmacophore model.
are made up of three-dimensional property densities for several
properties (hydrogen-bond donors/acceptors, hydrophobicity,
and aromaticity). Each molecule has a set of APFs, and the
overlay will result in higher density around regions of space
with common field strength that can be visualized by a
graphical view of the high-density contours (Figure S1 in
Supporting Information), which constitutes the pharmacophore
density that spans a wide range of binding modes. Upon
completion of the HTS, interesting hits were docked into the
pharmacophore model, and to our delight, several novel
binding modes were identified, including compound 4 (Figure
2). This series has some of the highest LipE values identified
from the HTS. We were very intrigued by this novel inhibitor
of a PARP family member where the nicotinamide amide motif
was replaced by an aminotriazole, which to our knowledge is a
novel isostere of the nicotinamide moiety. The crystal structure
of the catalytic domain of human TNKS1 complexed with
compound 4 was determined to 2.3 Å resolution (PDB code
4KRS), and it revealed interactions that closely relate to those
of the TNKS1/1 complex.28 These include a hydrogen bond
between the triazole portion of 4 and the hydroxyl of Ser1221
as well as a hydrogen bond to the backbone amide of Gly1185
(Figure 3). The amide linker of 4 also forms a hydrogen bond
to the carbonyl oxygen of Gly1185 and engages in a water-
mediated hydrogen bond to the backbone amide of Tyr1213.
Van der Waals interactions are maintained between 4 and both
Tyr1213 and Tyr1224. The D-loop in the TNKS1/2 structure
is ordered, adopting a similar conformation as reported in the
TNKS1/1 complex structure,28 with Tyr1203 making an edge−
face stacking interaction with the phenyl ring of 4. Additional
van der Waals interactions exist between the t-butyl group of 4
and Pro1187, Phe1188, and Ile1228 of TNKS1. The solved
structure was very similar to the structure predicted by our
pharmacophore model (Figure S2 in Supporting Information)
Compound 4 was the most potent compound in this class
identified by our screen, and traditional structure−activity
relationship (SAR) studies revealed factors that modulated
potency (Tables 1 and 3). We were curious if we could
understand key interactions better and progress this series more
rapidly by examining LipE-driven structure−efficiency relation-
ships (SER) rather than potency-driven SAR. We sought to
understand what maximized the efficiency of binding and
utilized LipE in real time for decision-making. We include LE in
this paper to highlight some interesting findings with this
metric; however, LE was not utilized as a decision-making tool.
One structural feature that substantially affects binding is the
size of the thiazine ring (Table 1). As with compounds 4 and 6,
all molecular matched pairs (MMP) within this series
demonstrated that five-membered rings were approximately
10-fold more potent and had improved LE/LipE relative to the
corresponding six-membered rings. We observed that, with
B
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
reduced. The next two compounds highlight interesting aspects
about SER-driven decision-making. Replacing the cyclohexyl
moiety with a phenyl ring (18) or a tetrahydrofuran ring (19)
reduces lipophilicity (ΔclogP of −1.2 and −2.4, respectively)
such that an approximate 10-fold reduction in potency with 18
and 19 results in increased LipE of 0.1 and 1.4, respectively, but
with a decrease in LE of 0.08 for both compounds. From a
potency and LE perspective these two modifications are
significantly less favorable, but from a LipE perspective 17
and 18 are similar while 19 is significantly more favorable than
17.
Increasing the tether length by an additional methylene (20−
23) provided the first single-digit nanomolar analogues in this
series having similar biochemical activity as 1. Analogues with
longer alkyl chains (24 and 25) were 10-fold less active,
suggesting an optimal tether length had already been achieved.
A similar trend was observed where smaller cycloalkyl rings
have more favorable potency, LipE, and LE (20 and 21 vs 22)
while the phenyl ring (23), by virtue of its lower lipophilicity,
was favored from a LipE perspective despite lower potency and
LE than 21. Modifications such as urea (26) and piperidine
(27) demonstrated further dynamic SER. The lower LipE of 26
suggested that either conformational effects of the urea lowered
affinity or the TNKS protein could not offer a polar interaction
with the urea NH to compensate for the desolvation penalty
introduced by this polar group. Compound 27 maintained
constant LipE relative to 21 despite a 300-fold decrease in
potency.
Despite a lower LE relative to 21, compound 23 has higher
LipE and therefore became our starting point for further
analogue expansion (Table 4). While 19 had a more attractive
LipE than 23, we chose to sacrifice efficiency for the ability to
generate analogues rapidly, with the assumption that the phenyl
ring could be made more efficient by incorporating more
efficient substituents and/or replacing with a heteroaryl motif.
Our aim was to learn more about which types of interactions
with TNKS were favored and improve the cellular activity of
these compounds to the low nanomolar range. The low
lipophilicity of 23 (clogP = 1.9) gave us the advantage of
focusing on analogues with increased lipophilicity (28−33), a
tactic that was well aligned with our goal for increasing cell
potency. Halogen substituents at the ortho position (28 and
29) increased potency and LipE and, in the case of 29, resulted
in activity in the cellular assay that was slightly more potent
than 1, but they were less effective at other positions (data not
shown). With a clogP of 2.6 and encouraging cell activity,
compound 29 has improved physical chemical properties versus
both 1 and 4 and a highly favorable LipE of 5.7. We undertook
a brief examination of disubstituted aryl modifications to
generate a tool compound that could be used to help us decide
if this series had the potential for a successful lead optimization
campaign.
Figure 3. X-ray crystal structure of 4 in TNKS1 (PDB code 4KRS).
Hydrogen bonds are shown as dashed lines. H-bond distances
(clockwise from Ser1221, in angstroms) are 3.01, 2.87, 2.89, and 2.58
(4 to H2O) and 2.87 (H2O to Tyr1213).
each MMP, the approximate 10-fold difference in potency from
the removal of a methylene group (Δheavy atom count = +1,
ΔclogP33 = −0.2) led consistently to a ΔLipE of approximately
+1.1, while for the lowest MW matched pair (9 vs 10) the LE
improvement (0.13) was significantly greater than that of the
highest MW matched pair (0.09 for 4 vs 6). The LE
dependence on MW has been noted elsewhere and illustrates
how its use can be misleading.34−36 It appears obvious that the
five-membered ring is superior to the six-membered ring on the
basis of potency. However, the SER tells us that although the 4-
tert-butylphenyl amide is the most potent analogue within this
series, it is also the least efficient, with compound 9 having an
improved LipE (ΔLipE = +1.4) and LE (ΔLE = +0.16) relative
to compound 4. Further truncation of the alkyl amide (11 and
12) obliterated activity, and therefore further optimization was
based on 9 due to the improved LipE and lower clogP despite
being severalfold less potent than 4. Analogues at this stage
were spot-checked for activity against PARP1 and PARP2 and
pathway inhibition. The PK profile of 9 in C57 BL/6 mice was
determined (Table 2) and was characterized by low to
moderate clearance and good oral bioavailability, and the
Cmax exceeded the cellular IC50 by 2-fold at a 5 mg/kg oral dose.
The excellent in vitro safety profile (target hit rate of 0%, n =
122)37,38 and high solubility coupled with the negative results
versus hERG, Ames, and CYP450 panel suggests this novel
series has excellent potential for further optimization.
We began investigating cycloalkyl amides (13−16, Table 3)
by modification of 5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-
amine (Scheme 1) and found the cyclobutyl amide (14, LipE =
5.8) to have the most efficient ring size, while cyclohexyl (16)
was the most potent of this subset. While the amplitude of SAR
appeared substantial (ΔpIC50 = +1.5, approximately 30-fold
range of potency) the SER amplitude was minimal (ΔLipE =
+0.5, approximately 3-fold range of efficiency). Extending the
ring further by a methylene (17) resulted in the first analogue
in this series with submicromolar activity in the HEK293-STF
reporter gene cellular assay. In this case potency was clearly
improved and LE was slightly improved, but LipE was slightly
The 2,6-dichloro analogue 30 resulted in lower LipE and
slightly higher potency, while the lower activity of 31 clearly
demonstrated the importance of at least one ortho chloro
substituent. After intravenous administration in rats, 30 showed
high CL (45.6 mL·min−1·kg−1) and a moderate volume of
distribution (1.1 L/kg), in good agreement with the high in
vitro microsomal clearance values. We examined additional
dichloro analogues and found that the 2,3-dichlorophenyl
moiety (32) was among the most efficient and cell-potent
analogues in this series. This series demonstrates the dynamic
SAR required for successful lead optimization; however,
C
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Table 1. Relationship between Thiazine Ring Size and Amide Substituents and Lipophilic and Ligand Efficiency
a
b
c
d
LipE = pIC50 (TNKS2) − clogP. LE = 1.4pIC50/(no.of heavy atoms). HEK293 SuperTopFlash reporter gene assay. LipE = pIC50 (TNKS2) −
e
log D. Measured log D. nd = not determined.
a
Table 2. Pharmacokinetic Profile of 9 in C57 BL/6 Mice
We next explored a heterocyclic subseries (34−38) based on
compound 23 (Table 5). The most significant consequence of
utilizing SER for decision-making is that one has to accept that
modifications that reduce potency can be more favorable and
should be retained in further rounds of analogue design. This
point is illustrated with compounds 34−38. In each case the
heterocycle is less lipophilic than 23 so that it would be
expected that if LipE were to remain constant, potency would
necessarily decrease. For example, compounds 34 and 35 could
be 30-fold less potent (ΔpIC50 = −1.5) than 23 and maintain
constant LipE since they are 30-fold less lipophilic (ΔclogP =
−1.5). These compounds are only 3.5-fold less active than 25,
and thus their greater lipophilic efficiency (ΔLipE = 1)
indicates the modifications exceeded expectations from a LipE
perspective. While both the LipE and LE values are in highly
attractive territory, it was believed that cellular activity needed
to be improved. Combined with our earlier observation that
polar functionality at this region improved LipE without
enhancing potency,39 we surmised that on this scaffold we were
not orienting our heterocyclic substituents in a manner that
1 mg/kg iv dose
5 mg/kg po dose
AUC (nM·h)
AUC extrap (nM·h)
CL (mL·min−1·kg−1
Vss (L/kg)
T1/2 (h)
4730
4747
17
8595
8617
)
0.1
0.1
T1/2 limits (h)
Tlast (h)
1−1
1
MRT (0 − t) (h)
Cmax (nM)
Tmax (h)
0.1
19 074
0.3
13 679
0.3
F (%)
36
a
Following a single 1 mg/kg iv or 5 mg/kg po dose.
lipophilicity had increased significantly relative to compound 9
and further modifications to decrease lipophilicity were
explored.
D
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 3. Variation of Ring Size and Tether Length and the
Relationship with Lipophilic and Ligand Efficiency
Scheme 1. General Synthetic Scheme for Preparation of
a
a
Aminotriazole Analogues
a
(a) RCOCl, DMAP, collidine. (b) RCOOH, EDCI, DMAP, Et3N,
THF.
Table 4. Structure−Efficiency Relationships of Mono- and
a
Disubstituted Phenyl Rings
a
LipE = pIC50 (TNKS2) − clogP. bLE = 1.4pIC50/(no. of heavy
c
atoms). HEK293 SuperTopFlash reporter gene assay.
detected, and in the case of 30, submicromolar potency was
achieved (Table 6). In contrast to other series,4,39 the
correlation between biochemical and cellular potency had
some inconsistencies in this current series during the early
stages of evaluation and optimization, but these diminished as
on-target potency and efficiency improved. The most potent
compounds had low solubility, which we speculate could
contribute to the inconsistent cellular activity and highlights the
importance of properties during hit series evaluation.
Compound 30 lacks appreciable activity against PARP1/2 or
other off targets; when tested against an internal panel of 120
safety targets, <50% inhibition at maximal concentration
occurred for 118 panel members (target hit rate = 1.7%). We
also evaluated the in vitro ADME properties, and in contrast to
compound 9, the more potent and lipophilic analogues in this
series have significantly increased in vitro clearance in mouse
liver microsomes. While rat and mouse liver microsomal
stability data correlated well (data not shown), we observed
significantly lower extraction ratios with human liver micro-
somes, suggesting the potential of species-dependent metabo-
a
LipE = pIC50 (TNKS2) − clogP. bLE = 1.4pIC50/(no. of heavy
c
atoms). HEK293 SuperTopFlash reporter gene assay.
allowed productive or specific interactions with the TNKS
binding pocket.
To further evaluate the ability of these compounds to
modulate tankyrase function in a cellular context, the direct
stabilization of Axin2 was measured in an ELISA format
developed in-house. As the direct substrate for TNKS, Axin
provided a proximal readout for cellular inhibition of TNKS1
and 2.1 For multiple analogues, clear stabilization of Axin2 was
E
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
pharmacophore model, favorable properties, and LipE-driven
optimization led to this series progressing into lead
optimization only 4 months after 4 was confirmed as a hit.
The use of LipE-based SER-driven decision-making facilitated
this rapid evaluation and potency improvement without the
identification of any new off-target liabilities.
Table 5. Structure−Efficiency Relationships of Heterocyclic
a
Analogues
MATERIALS AND METHODS
■
Tankyrase AutoPARsylation Assay. PARP catalytic activity was
monitored by quantitative liquid chromatography/mass spectrometry
(LC-MS) detection of nicotinamide, as previously described.40 The
autoPARsylation reactions were performed at room temperature in
384-well Greiner flat-bottom plates. The final reaction mixture
contained 2.5% dimethyl sulfoxide (DMSO) and inhibitors with
concentrations ranging from 0.0001 to 18.75 μM. GST-TNKS2P,
GST-TNKS1P, PARP1, and PARP2 enzymes were used at final
concentrations of 5, 5, 5, and 2 nM, respectively. The nicotinamide
concentration in the resulting supernatants was measured by LC-MS.
The percent inhibition was calculated as (control − sample)/(control
− background) × 100. “Control” is the average value of eight wells
without compound, and “background” is the average of eight wells
mixed with 5× quenching solution and measured prior to initiation of
the reaction.
SuperTopFlash Reporter Gene Assay. Compound activity in
inhibiting Wnt ligand-induced signaling was measured by a Wnt-
responsive SuperTopFlash (STF) luciferase reporter gene assay in
HEK293 cells. The percent inhibition was calculated as (maximum
Wnt-induced signaling − sample)/ (maximum Wnt-induced signaling
− background) × 100. Maximum Wnt-induced signaling is the STF
signal level induced by 20% Wnt3A CM without compound, and
“background” is the STF signal level without the addition of Wnt3A
CM or compound. To demonstrate the specific activity of inhibitors
regulating Wnt signaling, a counterscreen was performed in HEK293T
cells expressing a cAMP-response element (CRE) luciferase reporter
gene. Compound activity on the CRE reporter was measured in the
presence of 10 μM forskolin, which is an activator of cAMP signaling.
Axin2 Protein ELISA. Compound activity in stabilizing the Axin2
protein was measured by sandwich enzyme-linked immunosorbent
(ELISA) assay in the colorectal cell line SW480. Cell lysates were
prepared from cells treated with compounds in six-point dilution
starting at 10 μM for 24 h. For the ELISA assay, anti-Axin2 capture
antibody was diluted to a concentration of 1 μg/mL (1:1000) in
carbonate coating buffer, pH 9.2 (Sigma, C3041-50CAP). Then 100
μL of the diluted anti-Axin2 capture antibody per well was used to coat
the 96-well plate overnight at 4 °C. Plates were then washed three
times with wash solution, PBST20 (phosphate-buffered saline, PBS, +
0.05% Tween), and blocked with 1% bovine serum albumin (BSA) in
PBS for 1.5 h at room temperature while shaking gently. After
blocking, plates were then washed three times with wash solution.
Then 100 μL of prepared SW480 cell lysate was added to each well
and incubated at room temperature for 2 h while being shaked gently.
After washing, 100 μL of biotinylated anti-Axin2 antibody was added
a
LipE = pIC50 (TNKS2) − clogP. bLE = 1.4pIC50/(no. of heavy
c
atoms). HEK293 SuperTopFlash reporter gene assay.
lism (Table 6). Despite decreased microsomal stability and
solubility activity, there was no significant erosion of the other
favorable properties of compound 9 during our hit to lead
optimization efforts, suggesting that SER may be effective for
guiding decision-making early in the drug discovery process.
Taken together, the data suggest potential for further
optimization of this series, and the results of these efforts will
be disclosed in due course.
Summary. Following our cell-based phenotypic screen for
Wnt pathway antagonists, we initiated a biochemical HTS to
enlarge the pool of available hits. A pharmacophore model was
built prior to the confirmation of the HTS hits that allowed us
to rapidly identify novel binding modes and interesting
scaffolds to pursue. These efforts identified 4, a highly efficient
tankyrase inhibitor that represents a novel isostere of
nicotinamide-based PARP inhibitors with substantial selectivity
versus the other PARP family members tested. By utilizing
structure−efficiency relationships to drive the optimization
based on LipE rather than absolute potency, a novel class of
highly potent and selective tankyrase inhibitors with highly
favorable properties was generated. The combination of
Table 6. Cellular Axin2 Stabilization and in Vitro Metabolic Stability Data of Select Analogues
IC50 (μM)
a
b
compd
TNKS1
TNKS2
PARP1
PARP2
axin2
7.67
mER
hER
clogP
solubility pH 6.8 (μM)
c
9
0.395
0.299
0.028
0.032
0.014
0.012
0.007
0.024
0.103
0.038
0.009
0.008
0.005
0.003
0.001
0.006
55.3
nd
nd
0.29
0.84
0.91
0.95
0.90
0.91
0.92
0.91
0.17
0.36
0.68
0.20
0.66
0.61
nd
1.3
3.1
2.4
3.0
2.6
3.3
3.2
3.3
>1000
>100
630
520
32
4
>19
nd
>10
3.08
>10
20
21
29
30
32
33
nd
nd
nd
nd
nd
1.40
>19
>19
>19
>19
3.8
4.7
0.377
16
>10
>10
nd
nd
nd
a
b
c
Mouse liver microsome extraction ratio. Human liver microsome extraction ratio. nd = not determined.
F
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
to each well and the plates were incubated at room temperature for 2
h. Signal was detected by chemiluminescence (Pierce SuperSignal
ELISA Femto, no. 3704) with streptavidin−horseradish peroxidase
(HRP) (R&D Systems, DY998) and measured on Perkin-Elmer
Wallac 1420 plate reader.
Rat Pharmacokinetic Studies. Intravenous PK studies were
performed with male Sprague-Dawley rats weighing 220−270 g that
are approximately 6−10 weeks old, obtained from Harlan Research
Laboratories (South Easton, MA), each bearing dual implanted jugular
vein cannulas. The intravenous dose for compound 20 was prepared at
0.2 mg/mL in a solution containing 10% PG and 90% PBS. Each
animal received 1 mL of the solution/kg of body weight via
intravenous injection into the jugular vein cannula.
Mouse Pharmacokinetic Studies. Mouse PK studies were
performed with male C57BL/6 mice weighing 25−30 g that are
approximately 6−8 weeks old, obtained from Harlan Research
Laboratories (South Easton, MA). The intravenous dose for 9 was
prepared at 0.2 mg/mL in a solution containing 10% PG and 90%
PBS. Each animal received 1 mL of the solution/kg of body weight via
intravenous injection into the lateral tail vein. The oral dose was
prepared at 0.5 mg/mL in a solution in the same formulation as the iv
dose. Each animal received 10 mL of the dosing solution/kg of body
weight by oral gavage.
For both rat and mouse PK, approximately 50 μL of whole blood
was collected from the tail of each animal by use of a microvette
EDTA (ethylenediaminetetraacetic acid) tube, at 5 min (iv dose only)
and (oral 0.25 only) 0.5, 1, 2, 4, and 7 h postdose and transferred to a
EDTA tube. The blood was centrifuged at 3000 rpm and plasma was
transferred to a PCR-96-AB-C 96-well plate, capped with PCR strip
cap, and stored frozen (−20 °C) for parent compound analysis.
Measurement of Parent Compound. An LC-MS/MS method was
used to detect 9 or 20 standards and in plasma. Nontreated C57BL/6
or Sprague-Dawley rat plasma was obtained from Bioreclamation,
LLC. Compounds 9 and 20 as powder was dissolved in dimethyl
sulfoxide to achieve the target free-base stock concentration of 1 mg/
mL. Ten microliters of 1 mg/mL stock solution was spiked into 990
μL of nontreated C57BL/6 or Sprague-Dawley plasma. An automated
Protein Purification and Crystallization. Human TNKS1
protein was expressed, purified, and crystallized as previously
described.27 To obtain crystals containing 4, TNKS1/PJ34 complex
crystals were transferred into a soaking solution containing 100 mM
Bis-tris, pH 5.8, 18% poly(theylene glycol) (PEG) 3350, 320 mM
ammonium sulfate, and 200 μM 4. Transfer of the crystals into fresh
solution occurred once per hour over a 4-h period. Crystals were
cryoprotected in the same soaking solution with the addition of 20%
glycerol, followed by flash freezing directly into liquid nitrogen.
Data Collection and Structure Determination. Diffraction data
for the TNKS1/4 complex were collected on a Dectris Pilatus 6 M
Detector at the Swiss Light Source beamline X06SA at a wavelength of
1 Å. The data were measured from a single crystal, and the reflections
were indexed, integrated, and scaled by use of XDS and XSCALE.34
The space group of the complex was C2 with two molecules in the
asymmetric unit. The structure was determined by Fourier methods,
with the TNKS1/PJ34 structure (PDB ID 3UH2) with active-site
waters and inhibitors removed as the starting model.
Structure determination was achieved through iterative rounds of
positional and simulated annealing refinement using PHENIX35 with
model building using COOT.36 Individual B-factors were refined and a
bulk solvent correction was applied. The solvent, sulfate ions, and
inhibitor were built into the density in later rounds of the refinement.
Data collection and refinement statistics are shown in Table 7.
Pharmacokinetic Analysis. All animal-related procedures were
conducted under a Novartis IACUC-approved protocol in compliance
with Animal Welfare Act regulations and the Guide for the Care and
Use of Laboratory Animals.
serial dilution was performed with Tecan EVO 150 (Mannedorf,
̈
Switzerland) to yield final plasma standard concentrations of 1, 5, 10,
50, 100, 500, 1000, and 5000 ng/mL. Aliquots (25 μL) of plasma
treated with 9 or 20 and from the prepared standards and quality
controls (QCs) were deproteinated by adding acetonitrile (150 μL)
containing 50 ng/mL of the internal standard (IS; Glyburide). They
were then vortex-mixed for 10 min and centrifuged for 5 min at 4 °C
at 4000 rpm (2800g). An aliquot (125 μL) of the supernatant was
transferred to a new plate, to which 50 μL of water was added. Then
10 μL of sample extract was injected for HPLC-MS/MS analyses.
Chromatographic separation from interfering endogenous and
exogenous contaminants was achieved on a MAC-MOD ACE C18
HPLC column (particle size 3 μm, column dimensions 2.0 × 30 mm)
with 0.1% formic acid (FA) in water (A) and 0.1% FA in acetonitrile
(B) as mobile-phase solvents and a linear gradient. The column oven
temperature was set to 40 °C. The flow from the HPLC system
(Shimazdu 20AD LC pump) was held at 700 μL·min−1 from 0 to 3.01
min. The flow was directly introduced into the ion source of a Sciex
API4000 triple quadrupole (MS/MS) mass spectrometer (AB Sciex,
Framingham, MA) and subjected to electrospray ionization (ESI,
positive ion mode).
Table 7. Crystallographic Data and Refinement Information
parameters
TNKS1/2 complex
space group
a (Å)
C2
124.6
b (Å)
44.2
c (Å)
87.8
α (deg)
β (deg)
γ (deg)
90
90.1
90
resolution range (Å)
total observations
unique reflections
43.92−2.29
68 241
21 041
97.9 (99.2)
12.5 (3.3)
0.064 (0.360)
0.207/0.269
3330
a
completeness (%)
a
I/σ
a,b
Rsym
cryst/Rfree
c
R
All pharmacokinetic (PK) parameters were derived from concen-
tration−time data by noncompartmental analyses. All pharmacokinetic
parameters were calculated with the computer program WinNonlin
(Enterprise, Version 5.2) purchased from Pharsight Corp. (St. Louis,
MO).
protein atoms
heterogen atoms
solvent molecules
avg B-factor (Å2)
198
129
The peak concentrations (Cmax) and times they occurred (Tmax
)
protein atoms
39.4
35.8
were recorded. For the intravenous dose, the concentration of
unchanged compound at time 0 in pharmacokinetics and excretion
group was calculated from a log−linear regression of the first two data
points to back-extrapolate C(0). The area under the concentration−
time curve (AUClast) was calculated by use of the linear trapezoidal
rule.
The total body clearance from the plasma (CL) and the apparent
volume of distribution at steady state (Vss) were calculated from the
intravenous data, where AUCinf and AUMCinf are the area under the
concentration−time curve and area under the first moment
concentration−time curve from time 0 to infinity, respectively.
inhibitor atoms
rms deviations from ideal values
bond lengths (Å)
0.008
1.025
bond angle (deg)
a
b
Numbers in parentheses are for the highest resolution shell. Rsym
=
∑|Ih − ⟨Ih⟩|/∑Ih over all h, where Ih is the intensity of reflection h.
c
Rcryst and Rfree = ∑∥Fo| − |Fc∥/∑|Fo|, where Fo and Fc are observed
and calculated amplitudes, respectively. Rfree was calculated with 5% of
data excluded from the refinement.
G
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
doseiv
AUCinf
doseivAUMCinf
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (150 mg,
0.672 mmol) in N,N-dimethylformamide (3 mL) were added 1-ethyl-
3-(3-dimethylaminopropyl)carbodiimide hydrochloride (193 mg, 1.01
mmol), N,N-dimethylaminopyridine (123 mg, 1.01 mmol), and
triethylamine (0.24 mL, 177 mg, 1.75 mmol). The mixture was stirred
at ambient temperature for 16 h and concentrated to an oil in vacuo.
This was purified by silica gel column chromatography to give the title
compound (135 mg, 0.542 mmol, 81% yield, 90% purity) as a white
solid. 1H NMR (400 MHz, CD2Cl2) δ 4.39 (t, J = 7.3 Hz, 2H), 3.96 (t,
J = 7.3 Hz, 2H), 3.43 (quintet, J = 8.5 Hz, 1H), 2.23−2.41 (m, 4H),
2.00−2.13 (m, 1H), 1.83−1.99 (m, 1H); MS m/z 225.2 (M + H);
retention time 0.79 min; HRMS m/z (M + H) calcd for C9H12N4OS
225.0810, found 225.0802.
CL =
V =
ss
2
(AUCinf
)
AUCinf was calculated from the following equation, where AUClast was
calculated by use of the linear trapezoidal rule:
Clast
λz
AUCinf = AUClast
+
The terminal elimination rate constant (λz) for the unchanged
compound was the slope of the log linear line from at least the last
three data points, and the half-life (T1/2) was calculated. The
bioavailability was estimated as follows (AUClast was used if
extrapolation was >20%):
N-(5, 6-Dihydrothiazolo[2, 3-c][1, 2, 4]triazol-3-yl)-
cyclopentanecarboxamide (15). By general method A, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (150
mg, 0.672 mmol) and cyclopentanecarbonyl chloride (98 mg, 0.740
mmol) were converted to the title compound (110 mg, 0.462 mmol,
AUCinf ,po
doseiv
AUCinf ,iv dosepo
%F =
Results are expressed as mean. No further statistical analysis was
performed.
1
69% yield) as a white solid. H NMR (400 MHz, CD2Cl2) δ 4.29−
4.43 (m, 2H), 3.95 (t, J = 7.3 Hz, 2H), 3.00 (quintet, J = 7.9 Hz, 1H),
1.94−2.07 (m, 2H), 1.61−1.89 (m, 6H); MS m/z 239.2 (M + H);
retention time 0.92 min; HRMS m/z (M + H) calcd for C10H14N4OS
239.0967, found 239.0963.
Chemistry. All solvents employed were commercially available
“anhydrous” grade, and reagents were used as received unless
otherwise noted. A Biotage Initiator Sixty system was used for
microwave heating. Flash column chromatography was performed on
Analogix Intelliflash 280 using Si 50 columns (32−63 μm, 230−400
mesh, 60 Å). NMR spectra were recorded on a Bruker AV400 (Avance
400 MHz) or AV600 (Avance 600 MHz) instruments. Analytical LC-
MS was conducted on an Agilent 1100 series with UV detection at 214
and 254 nm, and an electrospray mode (ESI) coupled with a Waters
ZQ single quad mass detector. One of two methods was used: In
method A, 5−95% acetonitrile (ACN)/H2O with 5 mM ammonium
formate was used with a 2 min run and 3 μL injection through an
Inertsil C8 column, 3 cm × 5 mm × 3 μm. In method B, 20−95%
ACN/H2O with 10 mM ammonium formate was used with a 2 min
run and 3 μL injection through an Inertsil C8 column, 3 cm × 5 mm ×
3 μm. The purity of all compounds was determined by high-resolution
(HR) HPLC/MS characterization and found to be >95% pure except
as noted. HRMS data were recorded on a Waters LCT Premier mass
spectrometer with dual electrospray ionization source and Agilent
1100 liquid chromatograph. The resolution of the MS system was
approximately 12 000 [full width at half-maximum (fwhm) definition].
HPLC separation was performed at 1.0 mL/min flow rate with the
gradient from 10% to 95% in 2.5 min. Ammonium formate (10 mM)
was used as the modifier additive in the aqueous phase.
Sulfadimethoxine (Sigma; protonated molecule m/z 311.0814) was
used as a reference and acquired through the LockSpray channel every
third scan. Examples 4, 6−12, and 18 were purchased from ChemDiv
(http://us.chemdiv.com).
N-(5, 6-Dihydrothiazolo[2, 3-c][1, 2, 4]triazol-3-yl)-
cyclohexanecarboxamide (16). By general method A, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (150
mg, 0.672 mmol) and cyclohexanecarbonyl chloride (99 mg, 0.672
mmol) were converted to the title compound (60 mg, 0.238 mmol,
35% yield) as a white solid. 1H NMR (400 MHz, CD2Cl2) δ 4.34 (t, J
= 7.0 Hz, 2H), 3.94 (t, J = 7.3 Hz, 2H), 2.45−2.58 (m, 1H), 1.95 (d, J
= 12.6 Hz, 2H), 1.82 (td, J = 3.1, 12.4 Hz, 2H), 1.69−1.75 (m, 1H),
1.27−1.55 (m, 5H); MS m/z 253.2 (M + H); retention time 1.03 min;
HRMS m/z (M + H) calcd for C11H16N4OS 253.1123, found
253.1124.
2-Cyclohexyl-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-
acetamide (17). By general method A, 5,6-dihydrothiazolo[2,3-
c][1,2,4]triazol-3-ylamine hydrobromide (150 mg, 0.672 mmol) and
cyclohexylacetyl chloride (119 mg, 0.740 mmol) were converted to the
1
title compound (51 mg, 0.191 mmol, 28% yield) as a white solid. H
NMR (400 MHz, CD2Cl2) δ 4.37 (t, J = 7.3 Hz, 2H), 3.95 (t, J = 7.3
Hz, 2H), 2.39 (d, J = 7.0 Hz, 2H), 1.82−1.90 (m, 1H), 1.60−1.82 (m,
5H), 1.14−1.36 (m, 3H), 0.98−1.13 (m, 2H); MS m/z 267.4 (M +
H); retention time 1.12 min; HRMS m/z (M + H) calcd for
C12H18N4OS 267.1280, found 267.1283.
N-(5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-2-(tetrahydro-
pyran-4-yl)acetamide (19). By general method A, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (150
mg, 0.672 mmol) and (tetrahydropyran-4-yl)acetyl chloride (120
mg, 0.740 mmol) were converted to the title compound (9 mg, 0.03
Syntheses. 5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine.
The title compound was prepared in accordance with a literature
procedure.36
1
mmol, 5% yield) as a white solid. H NMR (400 MHz, CD3OD) δ
4.23 (t, J = 7.3 Hz, 2H), 4.05 (t, J = 7.0 Hz, 2H), 3.93 (dd, J = 3.8,
11.29 Hz, 2H), 3.37−3.50 (m, 2H), 2.36 (d, J = 7.5 Hz, 2H), 2.03−
2.16 (m, 1H), 1.69 (d, J = 13.1 Hz, 2H), 1.27−1.45 (m, 2H); MS m/z
269.3 (M + H); retention time 0.99 min; HRMS m/z (M + H) calcd
for C11H16N4O2S 269.1072, found 269.1080.
3-Cyclopentyl-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-
propionamide (20). By general method A, 5,6-dihydrothiazolo[2,3-
c][1,2,4]triazol-3-ylamine hydrobromide (150 mg, 0.672 mmol) and 3-
cyclopentylpropionyl chloride (119 mg, 0.740 mmol) were converted
to the title compound (96 mg, 0.36 mmol, 54% yield) as a white solid.
1H NMR (400 MHz, CD2Cl2) δ 4.38 (t, J = 7.0 Hz, 2H), 3.95 (t, J =
7.3 Hz, 2H), 2.54 (t, J = 7.5 Hz, 2H), 1.82 (m, 3H), 1.72 (m, 2H),
1.65 (m, 2H), 1.55 (m, 2H); MS m/z 267.4 (M + H); retention time
1.17 min; HRMS m/z (M + H) calcd for C12H18N4O6 267.1280,
found 267.1284.
3-Cyclohexyl-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-
propionamide (21). By general method A, 5,6-dihydrothiazolo[2,3-
c][1,2,4]triazol-3-yl-amine hydrobromide (100 mg, 0.448 mmol) and
3-cyclohexylpropionyl chloride (86 mg, 0.493 mmol) were converted
to the title compound (87 mg, 0.310 mmol, 69% yield) as a white
N-(5, 6-Dihydrothiazolo[2, 3-c][1, 2, 4]triazol-3-yl)-
cyclopropanecarboxamide (13): General Method A. To an ice-cold
suspension of 5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hy-
drobromide (150 mg, 0.672 mmol) in tetrahydrofuran (2 mL) were
added 4-dimethylaminopyridine (41 mg, 0.34 mmol) and collidine
(0.13 mL, 122 mg, 1.01 mmol). To this was added cyclo-
propanecarbonyl chloride (77 mg, 0.740 mmol) in tetrahydrofuran
(1 mL), and the mixture was warmed up to ambient temperature and
stirred for 16 h. The reaction was quenched with saturated aqueous
ammonium chloride solution and extracted with dichloromethane.
The extracts were dried (magnesium sulfate) and concentrated to an
oil that was purified by silica gel column chromatography to give the
1
title compound (44 mg, 0.209 mmol, 31% yield) as a white solid. H
NMR (400 MHz, CD2Cl2) δ 4.37 (t, J = 7.3 Hz, 2H), 3.94 (t, J = 7.0
Hz, 2H), 1.90−2.08 (m, 1H), 1.01−1.07 (m, 2H), 0.89−0.99 (m, 2H);
MS m/z 211.1 (M + H); retention time 0.64 min; HRMS m/z (M +
H) calcd for C8H10N4OS 211.0654, found 211.0648.
N-(5, 6-Dihydrothiazolo[2, 3-c][1, 2, 4]triazol-3-yl)-
cyclobutanecarboxamide (14): General Method B. To a suspension
of 3-(2-chlorophenyl)propanoic acid (74 mg, 0.740 mmol) and 5,6-
H
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
solid. 1H NMR (400 MHz, CD2Cl2) δ 4.37 (t, J = 7.2 Hz, 2H), 3.95 (t,
J = 7.2 Hz, 2H), 2.55 (t, J = 7.8 Hz, 2H), 1.53−1.82 (m, 7H), 1.10−
1.39 (m, 4H), 0.84−1.02 (m, 2H); MS m/z 281.4 (M + H); retention
time 1.26 min; HRMS m/z (M + H) calcd for C13H20N4OS 281.1436,
found 281.1440.
3-Cycloheptyl-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-
propionamide (22). By general method B, 5,6-dihydrothiazolo[2,3-
c][1,2,4]triazol-3-ylamine hydrobromide (150 mg, 0.672 mmol) and 3-
cycloheptylpropionic acid (114 mg, 0.672 mmol) were converted to
the title compound (135 mg, 0.459 mmol, 68% yield) as a white solid.
1H NMR (400 MHz, CD3OD) δ 4.22 (t, J = 7.0 Hz, 2H), 4.05 (t, J =
7.3 Hz, 2H), 2.42 (t, J = 7.5 Hz, 2H), 1.73−1.83 (m, 3H), 1.48−1.72
(m, 6H), 1.30−1.43 (m, 4H), 1.00−1.15 (m, 2H); MS m/z 295.3 (M
+ H); retention time 1.46 min; HRMS m/z (M + H) calcd for
C14H22N4OS 295.1593, found 295.1602.
N-(5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-3-phenylpropio-
namide (23). By general method A, 5,6-dihydrothiazolo[2,3-c][1,2,4]-
triazol-3-ylamine hydrobromide (150 mg, 0.672 mmol) and 3-
phenylpropionyl chloride (119 mg, 0.706 mmol) were converted to
the title compound (110 mg, 0.401 mmol, 60% yield) as a white solid.
1H NMR (400 MHz, CD2Cl2) δ 7.11−7.36 (m, 5H), 4.21−4.35 (m,
yield) as a white solid. 1H NMR (400 MHz, CD2Cl2) δ 4.42−4.53 (m,
2H), 4.06 (br s, 2H), 3.98 (t, J = 7.3 Hz, 2H), 2.54−2.77 (m, 4H),
1.61−1.77 (m, 4H), 1.41−1.50 (m, 10H), 1.11 (dq, J = 4.3, 12.3 Hz,
2H); MS m/z 382.2 (M + H); retention time 1.35 min.
N-(5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-3-piperidin-4-yl-
propionamide Hydrochloride (27). To a solution of 4-[2-(5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylcarbamoyl)ethyl]piperidine-1-
carboxylic acid tert-butyl ester (69 mg, 0.181 mL) in 1,4-dioxane/
v
ethanol mixture (2 mL, 1:1, /v) was added hydrogen chloride in 1,4-
dioxane (0.27 mL, 4M, 1.09 mmol), and the mixture was stirred at
ambient temperature for 24 h. The mixture was then concentrated to
give the monohydrochloride salt of the title compound (60 mg, 0.179
1
mmol, 99% yield) as a white solid. H NMR (400 MHz, CD3OD) δ
4.46 (t, J = 7.3 Hz, 2H), 4.14 (t, J = 7.3 Hz, 2H), 3.40 (d, J = 13.1 Hz,
2H), 2.98 (t, J = 11.8 Hz, 2H), 2.66 (t, J = 7.3 Hz, 2H), 1.99 (d, J =
14.1 Hz, 2H), 1.67−1.78 (m, 3H), 1.35−1.51 (m, 2H); MS m/z 282.2
(M + H); retention time 0.97 min.
N-(5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-3-(2-
fluorophenyl)propionamide (28). By general method D, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (150
mg, 0.672 mmol) and 3-(2-fluorophenyl)propionic acid (103 mg,
0.611 mmol) in tetrahydrofuran/N,N-dimethylformamide (3 mL, 2:1
v/v) were converted to the title compound (91 mg, 0.311 mmol, 51%
yield) as a white solid. 1H NMR (400 MHz, CD2Cl2) δ 7.32 (t, J = 7.5
Hz, 1H), 7.17−7.27 (m, 1H), 6.98−7.14 (m, 2H), 4.33 (t, J = 7.3 Hz,
2H), 3.94 (t, J = 7.3 Hz, 2H), 3.08 (t, J = 7.5 Hz, 2H), 2.86 (t, J = 7.5
Hz, 2H); MS m/z 293.1 (M + H); retention time 1.23 min; HRMS m/
z (M + H) calcd for C13H13FN4OS 293.0872, found 293.0886.
3-(2-Chlorophenyl)-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-
yl)propionamide (29). By general method B, 5,6-dihydrothiazolo[2,3-
c][1,2,4]triazol-3-ylamine hydrobromide (150 mg, 0.672 mmol) and 3-
(2-chlorophenyl)propanoic acid (124 mg, 0.672 mmol) were
converted to the title compound (180 mg, 0.583 mmol, 87% yield)
as a white solid. 1H NMR (400 MHz, CD2Cl2) δ 7.31−7.45 (m, 2H),
7.22 (dq, J = 5.8, 6.94 Hz, 2H), 4.22−4.38 (m, 2H), 3.94 (t, J = 7.3 Hz,
2H), 3.06−3.20 (m, 2H), 2.86 (t, J = 7.5 Hz, 2H); MS m/z 309.1 (M
+ H); retention time 1.19 min; HRMS m/z (M + H) calcd for
C13H13ClN4OS 309.0577, found 309.0591.
3-(2,6-Dichlorophenyl)-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]-
triazol-3-yl)propionamide (30). By general method B, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (150
mg, 0.672 mmol) and 3-(2,6-dichlorophenyl)propionic acid (147
mg, 0.672 mmol) were converted to the title compound (198 mg,
0.577 mmol, 86% yield) as a white solid. 1H NMR (400 MHz, DMSO-
d6) δ 11.06 (br s, 1H), 7.48 (d, J = 8.0 Hz, 2H), 7.30 (t, J = 8.0 Hz,
1H), 4.13 (t, J = 7.0 Hz, 2H), 4.01 (t, J = 7.3 Hz, 2H), 3.11−3.22 (m,
2H), 2.54−2.65 (m, 2H); MS m/z 343.1 (M + H); retention time 1.24
min; HRMS m/z (M +H) calcd for C13H12Cl2N4OS 343.0187, found
343.0201.
2H), 3.93 (t, J = 7.0 Hz, 2H), 2.98−3.11 (m, 2H), 2.87 (t, J = 7.5 Hz,
2H); MS m/z 275.1 (M + H); retention time 1.11 min; HRMS m/z
(M + H) calcd for C13H14N4OS 275.0967, found 275.0966.
4-Cyclohexyl-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-
butanamide (24). By general method B, 5,6-dihydrothiazolo[2,3-
c][1,2,4]triazol-3-ylamine hydrobromide (150 mg, 0.672 mmol) and 4-
cyclohexylbutyric acid (114 mg, 0.672 mmol) were converted to the
1
title compound (98 mg, 0.333 mmol, 49% yield) as a white solid. H
NMR (400 MHz, DMSO-d6) δ 10.84 (br s, 1H), 4.04−4.10 (m, 2H),
3.96−4.03 (m, 2H), 2.30 (t, J = 7.5 Hz, 2H), 1.50−1.72 (m, 7H),
1.10−1.27 (m, 6H), 0.78−0.93 (m, 2H); MS m/z 295.5 (M + H);
retention time 1.40 min; HRMS m/z (M + H) calcd for C14H22N4OS
295.1593, found 295.1594.
N-(5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-4-phenylbutana-
mide (25). By general method A, 5,6-dihydrothiazolo[2,3-c][1,2,4]-
triazol-3-ylamine hydrobromide (150 mg, 0.672 mmol) and 4-
phenylbutanoyl chloride (123 mg, 0.672 mmol) were converted to
the title compound (29 mg, 0.098 mmol, 15% yield) as a white solid.
1H NMR (400 MHz, CD2Cl2) δ 7.12−7.36 (m, 5H), 4.38 (t, J = 7.3
Hz, 2H), 3.95 (t, J = 7.3 Hz, 2H), 2.66−2.76 (m, 2H), 2.59 (t, J = 7.5
Hz, 2H), 2.03 (quintet, J = 7.7 Hz, 2H); MS m/z 289.3 (M + H);
retention time 1.29 min; HRMS m/z (M + H) calcd for C14H16N4OS
289.1123, found 289.1134.
1-Benzyl-3-(5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)urea
(26): General Method C. To a solution of 5,6-dihydrothiazolo[2,3-
c][1,2,4]triazol-3-ylamine hydrobromide (100 mg, 0. 450 mmol) in
dimethylacetamide (1 mL) was added isocyanatomethylbenzene (95
mg, 0.717 mmol), and the mixture was subjected to microwave
irradiation (120 °C for 15 min). The mixture was concentrated to an
oil that was purified by silica gel column chromatography to give the
N-(5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-yl)-3-(2,6-
dimethylphenyl)propionamide (31). By general method B, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (100 mg,
0.448 mmol) and 3-(2,6-dimethylphenyl)propionic acid (80 mg, 0.448
mmol) were converted to the title compound (87 mg, 0.288 mmol,
1
title compound (37 mg, 0.134 mmol, 30% yield) as a white solid. H
NMR (400 MHz, DMSO-d6) δ 9.85 (s, 1H), 8.16 (br s, 1H), 7.21−
7.38 (m, 5H), 4.37 (d, J = 6.0 Hz, 2H), 4.06−4.15 (m, 2H), 3.96−4.05
(m, 2H); MS m/z 276.1 (M + H); retention time 1.06 min; HRMS
m/z (M + H) calcd for C12H13N5OS 276.0919, found 276.0912.
4-[2-(5,6-Dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylcarbamoyl)-
ethyl]piperidine-1-carboxylic Acid tert-Butyl Ester: General Method
D. To a suspension of 4-(2-carboxyethyl)piperidine-1-carboxylic acid
tert-butyl ester (199 mg, 0.77 mmol) in tetrahydrofuran (3 mL) were
added EDC (222 mg, 1.16 mmol), HOBt (178 mg, 1.16 mmol), and
triethylamine (125 mg, 1.24 mmol). This misture was sonicated for a
few minutes and 5,6-dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine
hydrobromide (121 mg, 0.85 mmol) was added. The mixture was
stirred at ambient temperature for 16 h and then partitioned between
ethyl acetate and saturated aqueous ammounium chloride solution.
The aqueous layer was washed with more ethyl acetate and the
combined organic extracts were dried over magnesium sulfate, filtered,
and concentrated to an oil that was purified by silica gel column
chromatography to give the title compound (78 mg, 0.20 mmol, 26%
1
64% yield) as a white solid. H NMR (400 MHz, CD3OD) δ 6.98 (s,
3H), 4.21 (t, J = 7.3 Hz, 2H), 4.05 (t, J = 7.3 Hz, 2H), 2.98−3.09 (m,
2H), 2.52−2.63 (m, 2H), 2.35 (s, 6H); MS m/z 303.2 (M + H);
retention time 1.27 min; HRMS m/z (M + H) calcd for C15H18N4OS
303.1280, found 303.1286.
3-(2,3-Dichlorophenyl)-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]-
triazol-3-yl)propionamide (32). By general method B, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (100
mg, 0.448 mmol) and 3-(2,3-dichlorophenyl)propionic acid (98 mg,
0.448 mmol) were converted to the title compound (70 mg, 0.204
1
mmol, 45% yield) as a white solid. H NMR (400 MHz, CDCl3) δ
12.03 (br s, 1H), 7.31 (dd, J = 1.51, 8.0 Hz, 1H), 7.23 (dd, J = 1.5, 7.5
Hz, 1H), 7.10 (t, J = 7.8 Hz, 1H), 4.34 (t, J = 7.0 Hz, 2H), 3.91 (t, J =
7.3 Hz, 2H), 3.19 (t, J = 7.5 Hz, 2H), 2.88 (t, J = 7.5 Hz, 2H); MS m/z
343.0 (M + H); retention time 1.42 min; HRMS m/z (M + H) calcd
for C13H12Cl2N4OS 343.0187, found 343.0198.
I
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
3-(2,4-Dichlorophenyl)-N-(5,6-dihydrothiazolo[2,3-c][1,2,4]-
triazol-3-yl)propionamide (33). By general method B, 5,6-
dihydrothiazolo[2,3-c][1,2,4]triazol-3-ylamine hydrobromide (150
mg, 0.672 mmol) and 3-(2,4-dichlorophenyl)propionic acid (147
mg, 0.672 mmol) were converted to the title compound (127 mg, 0.37
mmol, 55% yield) as a white solid. 1H NMR (400 MHz, DMSO-d6) δ
10.98 (br s, 1H), 7.61 (s, 1H), 7.33−7.43 (m, 2H), 3.94−4.10 (m,
4H), 2.91−3.04 (m, 2H), 2.68 (t, J = 7.3 Hz, 2H); MS m/z 343.4 (M
+ H); retention time 1.28 min; HRMS m/z (M + H) calcd for
C13nH12Cl2N4OS 343.0187, found 343.0202.
relationship studies of small-molecule inhibitors of Wnt response.
Bioorg. Med. Chem. Lett. 2009, 19, 3825−3827.
(4) Shultz, M. D.; Kirby, C. A.; Stams, T.; Chin, D. N.; Blank, J.;
Charlat, O.; Cheng, H.; Cheung, A.; Cong, F.; Feng, Y.; Fortin, P. D.;
Hood, T.; Tyagi, V.; Xu, M.; Zhang, B.; Shao, W. [1,2,4]Triazol-3-
ylsulfanylmethyl)-3-phenyl-[1,2,4]oxadiazoles: Antagonists of the Wnt
pathway that inhibit tankyrases 1 and 2 via novel adenosine pocket
binding. J. Med. Chem. 2012, 55, 1127−1136.
(5) Waaler, J.; Machon, O.; Tumova, L.; Dinh, H.; Korinek, V.;
Wilson, S. R.; Paulsen, J. E.; Pedersen, N. M.; Eide, T. J.; Machonova,
O.; Gradl, D.; Voronkov, A.; von Kries, J. P.; Krauss, S. A novel
tankyrase inhibitor decreases canonical Wnt signaling in colon
carcinoma cells and reduces tumor growth in conditional APC mutant
mice. Cancer Res. 2012, 72, 2822−2832.
(6) Waaler, J.; Machon, O.; von Kries, J. P.; Wilson, S. R.; Lundenes,
E.; Wedlich, D.; Gradl, D.; Paulsen, J. E.; Machonova, O.; Dembinski,
J. L.; Dinh, H.; Krauss, S. Novel synthetic antagonists of canonical Wnt
signaling inhibit colorectal cancer cell growth. Cancer Res. 2011, 71,
197−205.
ASSOCIATED CONTENT
* Supporting Information
■
S
Two figures, showing aligned ligands used for generation of
pharmacophore and comparison of predicted and experimen-
tally determined binding mode of 4, and one table, listing safety
profiling results of 4 and 30. This material is available free of
charge via the Internet at http://pubs.acs.org.
(7) Woon, E. C.; Threadgill, M. D. Poly(ADP-ribose)polymerase
inhibition: Where now? Curr. Med. Chem. 2005, 12, 2373−2392.
(8) Gunaydin, H.; Gu, Y.; Huang, X. Novel binding mode of a potent
and selective tankyrase inhibitor. PLoS One 2012, 7, No. e33740.
(9) James, R. G.; Davidson, K. C.; Bosch, K. A.; Biechele, T. L.;
Robin, N. C.; Taylor, R. J.; Major, M. B.; Camp, N. D.; Fowler, K.;
Martins, T. J.; Moon, R. T. WIKI4, a novel inhibitor of tankyrase and
Wnt/ss-catenin signaling. PLoS One 2012, 7, No. e50457.
(10) Narwal, M.; Venkannagari, H.; Lehtio, L. Structural basis of
selective inhibition of human tankyrases. J. Med. Chem. 2012, 55,
1360−1367.
(11) Wahlberg, E.; Karlberg, T.; Kouznetsova, E.; Markova, N.;
Macchiarulo, A.; Thorsell, A. G.; Pol, E.; Frostell, A.; Ekblad, T.; Oncu,
D.; Kull, B.; Robertson, G. M.; Pellicciari, R.; Schuler, H.; Weigelt, J.
Family-wide chemical profiling and structural analysis of PARP and
tankyrase inhibitors. Nat. Biotechnol. 2012, 30, 283−288.
(12) Casas-Selves, M.; Kim, J.; Zhang, Z.; Helfrich, B. A.; Gao, D.;
Porter, C. C.; Scarborough, H. A.; Bunn, P. A., Jr.; Chan, D. C.; Tan,
A. C.; DeGregori, J. Tankyrase and the canonical Wnt pathway protect
lung cancer cells from EGFR inhibition. Cancer Res. 2012, 72, 4154−
4164.
(13) Gao, J.; Zhang, J.; Long, Y.; Tian, Y.; Lu, X. Expression of
tankyrase 1 in gastric cancer and its correlation with telomerase
activity. Pathol. Oncol. Res. 2011, 17, 685−690.
(14) Gelmini, S.; Poggesi, M.; Pinzani, P.; Mannurita, S. C.; Cianchi,
F.; Valanzano, R.; Orlando, C. Distribution of t.ankyrase-1 mRNA
expression in colon cancer and its prospective correlation with
progression stage. Oncol. Rep. 2006, 16, 1261−1266.
(15) Gelmini, S.; Poggesi, M.; Distante, V.; Bianchi, S.; Simi, L.;
Luconi, M.; Raggi, C. C.; Cataliotti, L.; Pazzagli, M.; Orlando, C.
Tankyrase, a positive regulator of telomere elongation, is over
expressed in human breast cancer. Cancer Lett. 2004, 216, 81−87.
(16) Seimiya, H.; Muramatsu, Y.; Ohishi, T.; Tsuruo, T. Tankyrase 1
as a target for telomere-directed molecular cancer therapeutics. Cancer
Cell 2005, 7, 25−37.
AUTHOR INFORMATION
Corresponding Author
*E-mail michael.shultz@novartis.com.
■
Present Address
†N.J.W.: Epizyme, 400 Technology Square, Cambridge,
Massachusetts 02139, United States.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge the work of James Groarke and Guoping Xiao
for protein purification of TNKS1. The TNKS data collection
was performed on the X06SA beamline at the Swiss Light
Source, Paul Scherrer Institut, Villigen, Switzerland.
ABBREVIATIONS USED
■
ADME, absorption distribution metabolism and excretion;
APF, atomic property field; CYP450, cytochrome P450; CL,
clearance; ELISA, enzyme-linked immunosorbent assay; HTS,
high-throughput screen; LE, ligand efficiency; LipE, lipophilic
efficiency; MMP, matched molecular pair; MW, molecular
weight; PARP, poly(ADP-ribose) polymerase; PK, pharmaco-
kinetic; SAR, structure−activity relationship; SER, structure−
efficiency relationship; STF, SuperTopFlash; TNKS, tankyrase
REFERENCES
(17) Seimiya, H. The telomeric PARP, tankyrases, as targets for
cancer therapy. Br. J. Cancer 2006, 94, 341−345.
(18) Fancy, S. P. J.; Harrington, E. P.; Yuen, T. J.; Silbereis, J. C.;
Zhao, C.; Baranzini, S. E.; Bruce, C. C.; Otero, J. J.; Huang, E. J.;
Nusse, R.; Franklin, R. J. M.; Rowitch, D. H. Axin2 as regulatory and
therapeutic target in newborn brain injury and remyelination. Nat
Neurosci. 2011, 14, 1009−1016.
■
(1) Huang, S. M.; Mishina, Y. M.; Liu, S.; Cheung, A.; Stegmeier, F.;
Michaud, G. A.; Charlat, O.; Wiellette, E.; Zhang, Y.; Wiessner, S.;
Hild, M.; Shi, X.; Wilson, C. J.; Mickanin, C.; Myer, V.; Fazal, A.;
Tomlinson, R.; Serluca, F.; Shao, W.; Cheng, H.; Shultz, M.; Rau, C.;
Schirle, M.; Schlegl, J.; Ghidelli, S.; Fawell, S.; Lu, C.; Curtis, D.;
Kirschner, M. W.; Lengauer, C.; Finan, P. M.; Tallarico, J. A.;
Bouwmeester, T.; Porter, J. A.; Bauer, A.; Cong, F. Tankyrase
inhibition stabilizes axin and antagonizes Wnt signalling. Nature 2009,
461, 614−620.
(2) Riffell, J. L.; Lord, C. J.; Ashworth, A. Tankyrase-targeted
therapeutics: expanding opportunities in the PARP family. Nat. Rev.
Drug Discovery 2012, 11, 923−936.
(3) Lu, J.; Ma, Z.; Hsieh, J. C.; Fan, C. W.; Chen, B.; Longgood, J. C.;
Williams, N. S.; Amatruda, J. F.; Lum, L.; Chen, C. Structure-activity
(19) Li, Z.; Yamauchi, Y.; Kamakura, M.; Murayama, T.; Goshima, F.;
Kimura, H.; Nishiyama, Y. Herpes simplex virus requires poly(ADP-
ribose) polymerase activity for efficient replication and induces
extracellular signal-related kinase-dependent phosphorylation and
ICP0-dependent nuclear localization of tankyrase 1. J. Virol. 2012,
86, 492−503.
(20) Ulsamer, A.; Wei, Y.; Kim, K. K.; Tan, K.; Wheeler, S.; Xi, Y.;
Thies, R. S.; Chapman, H. A. Axin pathway activity regulates in vivo
J
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
pY654-beta-catenin accumulation and pulmonary fibrosis. J. Biol.
Chem. 2012, 287, 5164−5172.
for use as tankyrase inhibitors. World Patent WO 2013010092 A1,
2013.
(21) Benson, J. D.; Chen, Y. N.; Cornell-Kennon, S. A.; Dorsch, M.;
Kim, S.; Leszczyniecka, M.; Sellers, W. R.; Lengauer, C. Validating
cancer drug targets. Nature 2006, 441, 451−456.
(22) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a
useful metric for lead selection. Drug Discovery Today 2004, 9, 430−
431.
(23) Abad-Zapatero, C.; Metz, J. T. Ligand efficiency indices as
guideposts for drug discovery. Drug Discovery Today 2005, 10, 464−
469.
(24) Leeson, P. D.; Springthorpe, B. The influence of drug-like
concepts on decision-making in medicinal chemistry. Nat. Rev. Drug
Discov. 2007, 6, 881−890.
(25) http://practicalfragments.blogspot.com/
(26) Shultz, M. D.; Kirby, C. A.; Stams, T.; Chin, D. N.; Blank, J.;
Charlat, O.; Cheng, H.; Cheung, A.; Cong, F.; Feng, Y.; Fortin, P. D.;
Hood, T.; Tyagi, V.; Xu, M.; Zhang, B.; Shao, W. [1,2,4]Triazol-3-
ylsulfanylmethyl)-3-phenyl-[1,2,4]oxadiazoles: antagonists of the Wnt
pathway that inhibit tankyrases 1 and 2 via novel adenosine pocket
binding. J. Med. Chem. 2012, 55, 1127−1136.
(27) Kirby, C. A.; Cheung, A.; Fazal, A.; Shultz, M. D.; Stams, T.
Structure of human tankyrase 1 in complex with small-molecule
inhibitors PJ34 and XAV939. Acta Crystallogr., Sect. F: Struct. Biol.
Cryst. Commun. 2012, 68, 115−118.
(28) Manuscript submitted for publication.
(29) Abagyan, R.; Totrov, M.; Kuznetsov, D. Icm, a new method for
protein modeling and design: Applications to docking and structure
prediction from the distorted native conformation. J. Comput. Chem.
1994, 15, 488−506.
(30) Abagyan, R.; Totrov, M. Biased probability Monte-Carlo
conformational searches and electrostatic calculations for peptides
and proteins. J. Mol. Biol. 1994, 235, 983−1002.
(31) Abagyan, R.; Orry, A.; Raush, E.; Budagyan, L.; Totrov, M. ICM
Manual 3.5, 2011.
(32) Totrov, M. Atomic property fields: Generalized 3D pharmaco-
phoric potential for automated ligand superposition, pharmacophore
elucidation and 3D QSAR. Chem. Biol. Drug Des. 2008, 71, 15−27.
(33) Ertl, P.; Muhlbacher, J.; Rohde, B.; Selzer, P. Web-based
cheminformatics and molecular property prediction tools supporting
drug design and development at novartis. SAR QSAR Environ. Res.
2003, 14, 321−328.
(34) Nissink, J. W. Simple size-independent measure of ligand
efficiency. J. Chem. Inf. Model. 2009, 49, 1617−1622.
(35) Reynolds, C. H.; Tounge, B. A.; Bembenek, S. D. Ligand
binding efficiency: Trends, physical basis, and implications. J. Med.
Chem. 2008, 51, 2432−2438.
(36) Reynolds, C. H.; Bembenek, S. D.; Tounge, B. A. The role of
molecular size in ligand efficiency. Bioorg. Med. Chem. Lett. 2007, 17,
4258−4261.
(37) Lounkine, E.; Keiser, M. J.; Whitebread, S.; Mikhailov, D.;
Hamon, J.; Jenkins, J. L.; Lavan, P.; Weber, E.; Doak, A. K.; Cote, S.;
Shoichet, B. K.; Urban, L. Large-scale prediction and testing of drug
activity on side-effect targets. Nature 2012, 486, 361−367.
(38) Bowes, J.; Brown, A. J.; Hamon, J.; Jarolimek, W.; Sridhar, A.;
Waldron, G.; Whitebread, S. Reducing safety-related drug attrition: the
use of in vitro pharmacological profiling. Nat. Rev. Drug Discovery
2012, 11, 909−922.
(39) Shultz, M. D.; Cheung, A.; Kirby, C. A.; Firestone, B.; Fan, J.;
Chen, Z.; Chin, D. N.; DiPietro, L.; Fazal, A.; Feng, Y.; Fortin, P. D.;
Gould, T.; Lagu, B.; Lei, H.; Lenoir, F.; Majumdar, D.; Ochala, E.;
Palermo, M. G.; Pham, L.; Pu, M.; Smith, T.; Stams, T.; Tomlinson,
R.; Toure, B. B.; Visser, M.; Wang, R.; Waters, N. J.; Shao, W.
Identification of NVP-TNKS656: The use of structure efficiency
relationships to generate a highly potent, selective and orally active
tankyrase inhibitor. J. Med. Chem.2013, DOI: 10.1021/jm400807n.
(40) Cheung, A.; Chin, D. N.; Fan, J.; Shultz, M. D.; Tomlinson, R.
4-Oxo-3,5,7,8-Tetrahydro-4H-pyrano{4,3-d}pyriminidinyl compounds
K
dx.doi.org/10.1021/jm400826j | J. Med. Chem. XXXX, XXX, XXX−XXX