Maximizing diversity from a kinase screen: Identification of novel and selective pan-Trk inhibitors for chronic pain

Article

pubs.acs.org/jmc

Maximizing Diversity from a Kinase Screen: Identiﬁcation of Novel

and Selective pan-Trk Inhibitors for Chronic Pain

Shawn J. Stachel,* John M. Sanders, Darrell A. Henze, Mike T. Rudd, Hua-Poo Su, Yiwei Li,

Kausik K. Nanda, Melissa S. Egbertson, Peter J. Manley, Kristen L. G. Jones, Edward J. Brnardic,

Ahren Green, Jay A. Grobler, Barbara Hanney, Michael Leitl, Ming-Tain Lai, Vandna Munshi,

Dennis Murphy, Keith Rickert, Daniel Riley, Alicja Krasowska-Zoladek, Christopher Daley, Paul Zuck,

Stephanie A. Kane, and Mark T. Bilodeau

Departments of Medicinal Chemistry, Biological Chemistry, Pain & Migraine, Molecular Systems, and Structural Biology, Merck

Research Laboratories, P.O. Box 4, West Point, Pennsylvania 19486, United States

S

* Supporting Information

ABSTRACT: We have identiﬁed several series of small

molecule inhibitors of TrkA with unique binding modes.

The starting leads were chosen to maximize the structural and

binding mode diversity derived from a high throughput screen

of our internal compound collection. These leads were

optimized for potency and selectivity employing a structure

based drug design approach adhering to the principles of

ligand eﬃciency to maximize binding aﬃnity without overly

relying on lipophilic interactions. This endeavor resulted in the

identiﬁcation of several small molecule pan-Trk inhibitor series

that exhibit high selectivity for TrkA/B/C versus a diverse panel of kinases. We have also demonstrated eﬃcacy in both

inﬂammatory and neuropathic pain models upon oral dosing. Herein we describe the identiﬁcation process, hit-to-lead

progression, and binding proﬁles of these selective pan-Trk kinase inhibitors.

members of the Trk family: TrkA, TrkB, and TrkC. All three of

the Trks’ polypeptides are composed of three distinct

topological domains: an extracellular domain, a transmembrane

domain, and a cytoplasmic kinase domain. Activation of the

receptor occurs when a complementary neurotrophic partner

binds with the extracellular domain of these kinases. Binding

results in dimerization and subsequent autophosphorylation,

thereby initiating the intracellular signaling pathway. Each of

the Trks binds to a speciﬁc neurotrophic factor, and the

complementary neurotrophin for each of the Trks is NGF for

TrkA, BDNF for TrkB, and NT-3 for TrkC. We were

particularly interested in the NGF/TrkA pathway because it

plays a central role in the biology of chronic pain.^3,4Activation

of the receptor tyrosine kinase TrkA by NGF triggers

intracellular signaling cascades and protein expression that

increases the sensitivity of nociceptors leading to chronic

sensitization and pain. Inhibition of the NGF/TrkA pathway

has been validated clinically using NGF-neutralizing mono-

clonal antibodies that are currently in phase II−III trials for

osteoarthritis pain.⁵We were interested in identifying small-

molecule inhibitors of TrkA as an alternative intervention

modality for the treatment of chronic pain.

INTRODUCTION

■

Kinases are a class of enzymes whose primary function is to

transfer the terminal phosphate group of ATP to an acceptor.

For a major subclass of kinases the acceptor is another protein.

This phospho-transfer event, termed phosphorylation, induces

a conformational change in the receiving protein, resulting in a

change in functional activity. This change in function is

manifested in a diverse array of signal transduction processes

within a cell and inﬂuences a wide array of cellular processes.

Kinases are grouped into two main classes by cellular location,

receptor and intracellular kinases. Receptor kinases propagate

transmembrane signaling events, whereas intracellular kinases

initiate signal transduction to the nucleus. In fact, kinases are so

integral to biological processes that greater than 500 kinases

have been identiﬁed, representing close to 2% of the human

genome.¹In totality, it has been stated that approximately 30%

of all human proteins may be modiﬁed by kinase activity. Given

that kinase signaling plays a ubiquitous role in physiological

processes and the fact that ATP is the conserved ligand for all

kinases, selectivity becomes paramount for the development of

a safe and eﬀective pharmaceutical agent. As such, achieving

high selectivity against other kinases is a prerequisite for the

development of clinical kinase inhibitor candidates.

Tropomysin receptor kinases (Trks) are cell surface receptor

kinases that are highly expressed in neurons and play an

important role in cell signaling.²There are three known

Received: May 1, 2014

A

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

than 20% activity at 1 μM and <10% activity against the

remaining kinases tested.⁶Initial eﬀorts focused on obtaining

increased potency toward TrkA to better gauge its relative

selectivity toward the greater kinome. In parallel with our

screening eﬀorts, we also pursued methods to obtain X-ray

crystallographic data to support follow-up SAR development

using a rational structural design approach. Toward this end, we

succeeded in obtaining protein crystals of the TrkA kinase

domain that were suitable for soaking with inhibitors to obtain

structural data in a relatively high-throughput manner.⁷The

crystallographic structure of 1 was obtained and was found to

bind in the ATP-binding site as seen in Figure 2.⁸Figure 3

HTS EFFORT

■

As part of our eﬀort to identify selective inhibitors of TrkA, we

conducted an HTS campaign with the intent of identifying

nontraditional chemotypes as a starting point for discovery

eﬀorts to maximize the potential for increased selectivity versus

the known kinome. Approximately 30K compounds were

identiﬁed in the initial HTS campaign that displayed at least

30% activity toward TrkA, the kinase of interest (median Z′ is

0.81; hit rate at 3σ is 0.5%). To narrow the number of active

hits derived from the high throughput screen, we employed a

number of ﬁlters. We removed hits that displayed promiscuous

activity, deﬁned as being active in more than ﬁve HTS

campaigns or more than two kinase HTS campaigns. In

addition, compounds were also counterscreened against the

receptor tyrosine kinase insulin-like growth factor receptor 1

(IGFR1) serving as a general kinase for initial selectivity

assessments. Lastly, activity against TrkA was measured using

both low (0.1 mM) and high (2 mM) ATP concentrations to

inform on potential ATP competitive binders. We succeeded in

identifying a number of potentially interesting candidates to

investigate further. Four particularly noteworthy compounds

that showed modest activity against TrkA and at least a 50-fold

selectivity window versus IGFR1, emerged from this eﬀort, are

exempliﬁed in Figure 1. Compound 1 originated from a general

Figure 2. Crystal structure of 1 with TrkA. The compound binds in a

DFG-in conformation.

Figure 1. HTS leads.

kinase library but was chosen for further investigation because,

while it contained a known chemotype used in orthosteric

kinase inhibitors, it exhibited high selectivity versus IGFR1

(IC₅₀> 50 μM) and as such was followed-up on as a starting

reference point for our crystallization eﬀorts. In contrast,

compounds 2−4 were all derived from non-kinase programs

and, upon initial ocular inspection, did not resemble chemo-

types typical of ATP-competitive binders. We chose to

investigate these four compound classes in more detail as hit

classes of interest with the goal of elucidating their binding

proﬁles and further optimizing both their intrinsic potency

toward TrkA and their physical properties.

Class I: Pyridopyrimidine Series. The pyridopyrimidine

series, exempliﬁed by the high-throughput screening hit 1,

displayed modest potency toward TrkA, IC₅₀≈ 1 μM, but as

mentioned, 1 displayed favorable selectivity against IGFR1.

When assayed against a broader 92-member kinase counter-

sceening panel, compound 1 exhibited only two hits of greater

Figure 3. Hydrogen bonding interactions of 1 within the active site of

TrkA.

depicts a magniﬁed view of the ATP-site highlighting the

binding interactions of 1 in the active site. The pyridopyr-

imidine was found to bind primarily within the hinge region of

the enzyme and would be described as a type I, or DFG-in,

inhibitor (vide infra).⁹The main hydrogen bonding inter-

actions between the enzyme and inhibitor 1 occurs through a

direct hydrogen bond between the N1 nitrogen of the

pyridopyrimidine and the backbone NH of Met 592. The

B

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

benzylic pyridine portion of the molecule is tucked in the front

pocket and forms a face-to-face π-stacking interaction with the

guanidine group of the activation loop residue, Arg 673. The

morpholine extends from the binding pocket into the solvent,

which is consistent with the observation that morpholine and

piperazine groups are commonly preferred substituents in this

region of the binding pocket.¹⁰

Having elucidated the binding orientation of 1 to TrkA in the

active site, we were able to pursue a rational approach for ligand

optimization (Scheme 1). To track the progress of our

(compound 5 LLE = 1.38 vs compound 1 LLE = 2.19) revealed

that much of the improved potency was derived from lipophilic

interactions.

Upon further optimization, the benzylamine was replaced

with a diﬂuoropyrrolidine as part of a scaﬀold hopping exercise

incorporating structural motifs found in other known kinase

inhibitors.¹²Next, screening a variety of aniline replacements

identiﬁed a benzylic imidazothiazole as a preferred substitution

resulting in compound 6.^13,14The synthesis of compound 6 is

described in Scheme 2. Compound 6 displayed impressive

a

Scheme 1. Evolution of Pyridopyrimidine HTS Lead 1

Scheme 2. Synthetic Route for Compound 6

a

(a) (S)-2-(2,5-diﬂuorophenyl)pyrrolidine, EtOH, 69%; (b) POCl₃,

70%; (c) 1-(1H-benzimidazol-2-yl)methanamine, EtOH, 58%.

single digit nanomolar activity toward TrkA, IC₅₀= 1.5 nM, as

well as greatly improved potency in our cell-based TrkA assay,

IC₅₀= 13 nM. The increased potency realized with 6 was also

commensurate with improvements in ligand eﬃciency with

respect to both molecular mass (LE = 0.36) and lipophilicity

(LLE = 3.16).

Compound 6 was a very interesting lead that exhibited a

typical type I kinase inhibitor binding mode. However, in

parallel with optimization of 1, eﬀorts also continued on some

of the other hits identiﬁed in the screen that originated from

non-kinase programs with the hopes of elucidating additional

inhibitors with more novel binding characteristics.

Class II: Phenyltriazole Series. A second series of interest

focused on the HTS lead exempliﬁed by compound 2. This

compound was originally synthesized as part of a legacy

neuroscience program and did not constitute a lead originating

from a prior kinase program. While 2 displayed modest potency

toward TrkA, IC₅₀= 662 nM, it displayed a very favorable

selectivity proﬁle registering only one hit greater than 20% at 1

μM in the panel of 92 diverse kinases with the majority showing

<10% activity against the kinases tested. Since the structure

bore little resemblance to traditional ATP-site binding motifs

for kinase inhibitors, we conducted a more detailed

investigation of this structural series. Gratifyingly, we were

able to obtain crystallographic data of 2 binding with TrkA.¹⁵In

contrast to the type I, or DFG-in, binding mode described

above for 1, compound 2 was found to bind to a very diﬀerent

conformational form of the enzyme, commonly referred to as a

type II, or DFG-out, conformation.¹⁶The diﬀerences between

optimization eﬀorts in each series and to ensure that we were

balancing any improvements in potency with the physicochem-

ical properties of the molecule, we relied on optimizing ligand

eﬃciency metrics based on both molecular mass (LE) and

lipophilicity (LLE). Our optimization eﬀorts in the pyridopyr-

imidine series began with replacing the morpholine ligand in

the solvent exposed region with an N-methylpiperazine group,

which as previously mentioned is another generic preferred

substituent in the hinge/solvent front region of kinase

inhibitors. In addition, since the pyridyl nitrogen in the western

portion of the molecule did not appear to be making a

productive hydrogen-bonding interaction with the protein, we

investigated alternative substituents in the region by screening

of a variety of benzylic amines and identiﬁed meta-

chlorobenzylamine as a superior replacement. These two key

modiﬁcations resulted in compound 5 that displayed a 90-fold

enhancement in intrinsic inhibitory activity toward TrkA

(Scheme 1). In addition, while 1 displayed activity of greater

than 10 μM in the TrkA cell-based assay,¹¹the potency

improvements realized in 5 now aﬀorded activity in our cellular

assay. More importantly, 5 retained a high degree of selectivity

displaying only three hits of greater than 20% activity at 1 μM

against the 92-member kinase panel with a majority still

showing <10% activity against the kinases tested. However,

while overall ligand binding eﬃciency increased with these

modiﬁcations (compound 5 LE = 0.33 vs compound 1 LE =

0.27), comparison of the lipophilic ligand eﬃciencies, LLEs

C

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

type I and type II inhibitors go beyond the interactions of the

inhibitor with the enzyme and involve substantial diﬀerences in

the tertiary structure of the kinase itself. The primary function

of a kinase is to transfer a phosphate group from ATP to

another protein substrate in order to initiate a signaling cascade.

For a number of kinases, this is regulated by a conformational

change in a conserved triad of amino acids, composed of an

Asp-Phe-Gly, or DFG sequence, at the start of the activation

loop. To propagate the signaling cascade, the inactive kinase is

ﬁrst phosphorylated on its activation loop by an upstream

kinase. In the case of the Trks this is accomplished through an

autophosphorylation event brought about by dimerization of

the enzyme upon binding to its corresponding neurotrophic

factor. Once the activation loop is phosphorylated, it undergoes

a conformational change that results in an opening of the active

site. This conformation of the enzyme is termed the DFG-in

conformation. This opening of the active site allows ATP to

bind and subsequently transfer its terminal phosphate group.

Conversely, in the inactive or nonphosphorylated state, the

activation loop occupies the active site and prevents ATP from

binding.¹⁷This closed conformation is termed the DFG-out

conformation where the DFG segment is rotated out and

toward the ATP binding site. The conserved DFG triad plays

an integral part in inducing the requisite tertiary structure

involved in kinase function. In the active form of the kinase the

Asp (DFG) residue plays a critical role by coordinating with a

magnesium ion that is also bound to the γ-phosphate of the

ATP, thus serving to facilitate the phospho-transfer event. From

a structural perspective, in the active form of the enzyme, the

Phe of the DFG triad normally occupies a hydrophobic pocket

located behind the gatekeeper residue allowing for proper

positioning of the catalytic aspartate for the phospho-transfer

event. However, in the inactive form the phenylalanine of the

DFG triad is rotated out of its binding site behind the

gatekeeper residue, thereby revealing a large hydrophobic

pocket commonly referred to as the selectivity pocket. The fact

that TrkA is able to adopt a DFG-out conformation is

particularly unusual, since a small gatekeeper, typically valine or

threonine, is often required to allow for this conformational

transition, whereas in TrkA the gatekeeper residue is a large

phenylalanine. Scientists at GNF have also reported a DFG-out

binding conformation for their TrkA inhibitor series in a recent

publication.¹⁸In fact, much of the recent eﬀorts in the

development of kinase inhibitors have focused on exploiting

DFG-out binders, since they have been thought to hold

potential advantages over active site orthosteric inhibitors.

Since all kinases bind ATP as their orthosteric ligand, targeting

the selectivity pocket has been viewed as a strategy to improve

selectivity against the broader kinome. Several important new

therapeutic agents have already been approved that selectively

bind to the DFG-out kinase conformation, with the tyrosine

kinase inhibitor imatinib being one of the most notable

examples.¹⁹However, while the majority of human kinases

contain the DFG-sequence as the start of the activation loop,

only a relative few have been observed in both the DFG-in and

DFG-out conformations.

Figure 4. Hinge-binding interactions of 2.

Figure 5. DFG-out binding of 2.

inhibitor in the active site. The amide carbonyl forms a

hydrogen bond with the NH of the catalytic Asp (668) residue

and the amide NH hydrogen bonds to the carboxylate of Glu

560. The N-phenylpyrazole occupies the hydrophobic

selectivity pocket normally occupied by Phe 669 in the active,

DFG-in, kinase conformation, with the pyrazole nitrogen

making an interesting water-mediated hydrogen bond to the

carboxylate of the catalytic Asp 668.

SAR Optimization of 2. Having secured a crystal structure

of 2 in the TrkA active site, we were able to leverage the

structural information gleaned from our analysis of the binding

mode to rapidly progress the series (Scheme 3). First, from the

structure it was evident that the methoxy group on the western

aryl ring of 2 was oriented toward the solvent front and was not

participating in any meaningful interaction with the enzyme. As

such, removal of the methoxy group (10) resulted in an

improvement in both enzymatic and cellular potency, perhaps

due to the conformational eﬀect of the methoxy group that

twists the imidazole/phenyl dihedral angle from coplanarity.

Removal of methoxy allows the imidazole to attain coplanarity

with the aryl ring, the proposed bioactive conformation,

aﬀording a 10-fold increase in potency. This modiﬁcation also

resulted in increased metrics for ligand eﬃciency with the LE

increasing from 0.23 to 0.28 and LLE increasing from 2.08 to

3.16, signaling that the increase in potency was not being driven

As mentioned, compound 2 was found to bind to the DFG-

out conformation of TrkA (Figure 5). Figure 4 depicts some of

the key binding elements observed in the structure. First, the

imidazole was observed to make one hydrogen bond to the

hinge backbone NH of Met 592. The central triazole ring

makes hydrophobic interactions with the gatekeeper phenyl-

alanine residue, thus serving as a hydrophobic clamp for the

D

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Scheme 3. Evolution of Triazole HTS Lead 2

Scheme 4. Synthetic Route for Compound 11

a

(a) Imidazole, L-proline, CuI, K₂CO₃, DMSO, 61%; (b) TMS−

acetylene, PdCl₂(PPh₃)₂, CuI, Et₃N, DMF, 78%; (c) KF, MeOH,

100%; (d) 2-chloroacetyl chloride, Et₃N, CH₂Cl₂, 72%; (e) NaN₃,

EtOH, 92%; (f) Cu powder, CH₂Cl₂−t-BuOH, H₂O, aq CuSO₄, 72%.

solely by lipophilicity changes. As mentioned in the above

discussion, it was evident that 2 formed only one direct

hydrogen bond with the hinge backbone. For our next

modiﬁcation we sought to increase the number of hinge

binding interactions through the introduction of a nitrogen into

the aryl ring in the form of an aminopyridine. However, instead

of a direct interaction with the backbone, the X-ray cocrystallo-

graphic structure revealed that the nitrogen lone pair interacted

indirectly with the carbonyl of Glu 590 through an intermediary

water molecule. Modiﬁcations were also made to the portion of

the molecule occupying the hydrophobic selectivity pocket. As

such, the cyclopropyl group was changed to a bulkier tert-butyl

group in order to more optimally ﬁll the hydrophobic void

previously occupied by the DFG-phenylalanine in the DFG-out

conformation. In addition to this change, the introduction of a

nitrogen to the pendent N-phenyl group served to decrease the

overall cLogD by 1 log unit while not imparting additional

hydrogen bonding interactions with the enzyme. The impact of

this modiﬁcation is apparent in the LLE value, as it increased to

5.78 from 3.16 for compound 11 when compared to 10,

indicating more eﬃcient binding with lower overall lip-

ophilicity. Taken together, these modiﬁcations that aﬀorded

compound 11 (synthesis described in Scheme 4) resulted in an

additional improvement in both enzymatic and functional

activity providing an inhibitor with single digit nanomolar

activity in both the enzymatic and cellular assays. The changes

to the starting structure not only led to increased potency but

were accomplished by maintaining and improving ligand

eﬃciency in a per atom binding basis and also with respect

to the overall lipophilic character. Compound 11 retained the

exquisite selectivity against the kinase panel showing high

selectivity with the improved potency registering only nine hits

with >20% inhibition at 1 μM with a majority showing <10%

activity against the kinases tested. Compound 11 possessed the

requisite potency and physical properties to be assessed in vivo

and was found to be eﬃcacious in our pain models (vide infra).

It should be noted that the superior cellular potency compared

with the enzymatic activity of this, and other series, is most

likely a result of the eﬀect of conducting the enzymatic activity

assay at elevated ATP levels to facilitate an increased dynamic

range in the enzymatic assay but that are not necessarily

reﬂective of the native levels of ATP present within the cells.

Class III: Urea Series. A third series of interest focused on

the HTS lead exempliﬁed by compound 3.²⁰This compound

displayed remarkable selectivity against all other kinases in the

general counterscreen with 0 hits of >20% activity at 1 μM.

Compound 3 was originally synthesized as part of our general

screening library. While compound 3 did not overtly resemble a

typical ATP-competitive chemotype, we now recognized that

TrkA was capable of binding both DFG-in and DFG-out type

inhibitors. Since amide and urea-based kinase inhibitors have

been well documented in the literature and are typically DFG-

out binders, at ﬁrst glance it seems plausible that this inhibitor

could also be interacting through a similar binding

paradigm.²¹⁻²⁴However, all of the reported inhibitors of this

class are diphenyl amides or ureas and are involved in hydrogen

bond donor and acceptor roles with the unsubstituted nitrogens

which are both involved in key hydrogen bonding interactions

with the enzyme. The potential for an atypical binding

paradigm for 3 was reinforced when we performed docking

experiments to suggest potential binding conformations.

Unfortunately we were unable to achieve such convincing

poses with the substituted ureas analogues. In fact, the fully

unsubstituted urea (19) was found to be completely inactive in

the enzymatic assay with an IC₅₀of >100 μM. By analyzing

structurally similar compounds contained within the collection,

we quickly found that an isopropyl group oﬀered a 30-fold

potency enhancement over the cyclopropyl derivative (17)

(Scheme 5). Compound 17 also retained excellent selectivity

with only one hit of greater than 20% activity when assayed

against the 92-member kinase counterscreen. Concurrent with

the aforementioned similarity assessment, we succeeded in

obtaining a crystal structure of compound 3 with the TrkA

kinase domain (Figure 6).²⁵The structural information turned

out to be quite revealing: the inhibitor was indeed found to

bind to the DFG-out conformation of the kinase, as other

amide and urea kinase inhibitors were already known to do, but

E

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Scheme 5. SAR Progression of HTS Lead 3

it did so without some of the more commonly observed

characteristics. The thiazole occupied the ATP-binding site but

was only found to form a weak hydrogen bond with Met 592.

The benzylic aryl ring was involved in hydrophobic interactions

with the gatekeeper residue, Phe 617. The para-triﬂuorome-

thoxyphenyl group was situated in the hydrophobic selectivity

pocket normally occupied by Phe in the DFG-in conformation,

as is typical in DFG-out inhibitors, and the amide carbonyl was

shown to form a hydrogen bond with the catalytic aspartate

(D668) as is typical with the literature precedence. The

cyclopropyl group on the adjacent nitrogen was shown to

occupy a hydrophobic cleft and was tucked neatly against Val

524. This is interesting, since inhibitors of most kinases are

typically unsubstituted and involved in a hydrogen bond with

the homologous residue of α-C helix Glu 560. Our

interpretation of the SAR around this group is that a critical

hydrophobic mass is required to ﬁll the small pocket near Val

524, Ala 542, Lys 544, and Phe 589, perhaps to displace a water

molecule that may otherwise unfavorably occupy this region.

Having solved the crystal structure and elucidated that the

thiazole, while occupying the ATP binding site, was not making

optimal interactions with the protein, we sought to replace it

with an alternative heterocycle with a greater potential to make

favorable hydrogen bonding interactions with the hinge. This

was accomplished by replacement of the thiazole with an

azaindole group and resulted in compound 18 (synthesis

described in Scheme 6). Here the azaindole functions as a

hydrogen-bond acceptor interacting with the NH of Met 592.

Conversely, the indole NH serves as a hydrogen bond donor

interacting with the carbonyl of Glu 590 in the backbone.

Hence, optimization of the interaction with the hinge region

and increasing the hydrophobic size of the heterocycle in the

adenine site resulted in 10-fold improvement in both enzymatic

and functional activity. It should also be noted that, similar to

the two series previously described, optimization of the binding

aﬃnity for this series was also accomplished while maximizing

the ligand binding and the lipophilic binding eﬃciencies.

Figure 6. (a) Crystal structure of 3 bound with TrkA. (b) Hinge-

binding interactions of 3. (c) View of 3 bound in the selectivity pocket.

Class IV: Azaindole Series. The fourth and ﬁnal series

described here focused on the HTS lead exempliﬁed by

compound 4. Pursuant to our strategy of investigating novel

chemotypes, compound 4 was chosen for further investigation,

since it displayed no activity against IGFR1 as part of the HTS

counterscreening protocol. In addition, 4 showed no activity

against any of the broader 92-membered kinase panel which

was supported by our inability to dock this structural class into

the TrkA active site, in either a DFG-in or DFG-out

conformation, in a reasonable way. Intriguingly, while

compound 4 was identiﬁed as a hit possessing moderate

potency toward TrkA in the high throughput screen, i.e., IC₅₀

=

F

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

based assay we employed a higher ATP concentration (i.e., 10

mM) compared to the 2 mM ATP concentration used in the

HTS screen allowing more eﬀective binding competition and

an upward shift in IC₅₀values. The shift to higher ATP

concentrations for use in the follow-up assay was done to allow

for more dynamic range in the assay for our optimization

eﬀorts. However, it should also be noted that it is diﬃcult to

determine if the original 17 μM IC₅₀observed when assayed

using phosphorylated substrate was derived from residual levels

of nonphosphorylated enzyme in the supplied protein and may

not reﬂect an accurate IC₅₀against pure phosphorylated

material. As such, the IC₅₀values for all subsequent compounds

for this particular series were determined using nonphosphory-

lated enzyme to have measurable and consistent activity by

which to conduct SAR analysis. Prior to determining the

importance of the phosphorylation state for the enzymatic assay

the series was progressed using the cell-based potencies, since

the cellular system exists in a dynamic equilibrium between the

active and inactive forms of the protein and provides a more

physiological assessment of a compound’s activity in the

presence of endogenous ATP levels.

Scheme 6. Synthetic Route for Compound 18

a

(a) Isopropylamine, NaCNBH₃, MeOH; (b) 1-isocyanato-4-

(triﬂuoromethoxy)benzene, THF, 98% over two steps.

Binding of Azaindole Series. To elucidate the binding

mode and obtain the structural insights necessary to design

more potent inhibitors, we obtained the structure of compound

4 complexed with the TrkA kinase domain (Figure 7).²⁸The

structure clearly showed that 4 binds to the DFG-out

conformation of TrkA. The structure revealed several unusual

binding features for compound 4. First, the inhibitor was found

not to have a direct interaction with hinge backbone. Instead,

the N5 nitrogen of the azaindole was shown to participate in a

water-mediated hydrogen bond to two hinge residues, Asp 590

and Met 592. The naphthalene moiety was buried in the

hydrophobic cleft at the front of the hinge ﬂanked by the

activation loop residue Met 670. This conformation was further

stabilized through hydrophobic interactions with the DFG

phenylalanine just behind the naphthyl ring inside the

adenosine site. The isoxazole ring was situated under the

gatekeeper phenylalanine residue, stabilized by hydrophobic

van der Waals interactions and orienting the terminal phenyl

ring into the selectivity pocket. One of the more unusual

binding interactions was that of the indole carboxylic acid. Here

the carboxylate was interacting with two backbone NHs in the

activation loop, donated by Gly 671 and Met 670. A third

hydrogen bonding interaction was also evident between the

carboxylate and Lys 544 on the roof of the adenosine binding

pocket. As stated previously, several of the inhibitors described

previously were also found to bind to a DFG-out conformation

but they did not show the phosphorylation-dependent binding

that compound 4 exhibited. We believe that it is this unusual

binding conformation between the indole carboxylate and the

activation loop that leads to the phosphorylation-state depend-

ent binding. This can be visualized in Figure 8, which depicts

the interaction between Tyr 676 and Asp 595 in the

crystallographic structure of TrkA bound to compound 4. In

this structure tyrosine 676 is involved in a hydrogen-bonding

interaction with Asp 596. Since Tyr 676 is a known

phosphorylation site on TrkA, it is apparent that phosphor-

ylation of Tyr 676 would introduce a repulsive electrostatic

interaction and disrupt this interaction with Asp 596, thereby

forcing the activation loop away from the active site and thus

stabilizing the active conformation. As such, the phosphorylated

form of the enzyme is not able to adopt the tertiary

conformation needed to bind compound 4 because of this

163 nM, when compound 4 was tested in our conﬁrmation

assay, it was found to display a severe disconnect between the

screening and follow-up assays, exhibiting approximately a 100-

fold diﬀerence in activity. The team was initially perplexed at

the disparity between the two assays and sought to identify the

cause for the diﬀerence in activity. In turned out that two

slightly diﬀerent assays were used for the HTS screen and

subsequent conﬁrmation assay, with the HTS screen utilizing a

homogeneous time-dependent ﬂuorescence (HTRF) assay,

whereas the conﬁrmation assay employed a caliper assay. More

importantly, upon reinvestigation of the assay protocols, it was

discovered that the TrkA protein used for the high throughput

screen was nonphosphorylated protein whereas the protein that

was used in the conﬁrmation assay was provided as a

phosphorylated TrkA protein. To investigate this phenomenon

further, Thermoﬂuor analysis was conducted using both forms

of the protein, phosphorylated and nonphosphorylated.

Compound 4 displayed a ΔT_mof 11 °C when assayed using

the unphosphorylated protein. In contrast, no change in

melting temperature was observed when the phosphorylated

protein was utilized. These results provided support that

compound 4 was exhibiting phosphorylation-state dependent

binding. Phosphorylation state-dependent binding to kinases is

an emerging avenue for inhibitor design. Since the DFG-out

conformation represents the inactive, unphosphorylated

protein, many compounds that exhibit this type of binding

proﬁle have shown diﬀerential binding activities between the

active and inactive forms, but most DFG-out inhibitors are

capable of binding both forms of the protein. However, there

are several recent reports in the literature of inhibitor series that

are active only versus the unphosphorylated form of the kinase

of interest and have unmeasurable, or weak, inhibition of the

phosphorylated form.^26,27Having discerned the cause of

discrepancy between the activity in the two assay formats, we

subsequently assayed compound 4 in our caliper-based assay

using nonphosphorylated TrkA protein. Under these con-

ditions 4 displayed IC₅₀= 2300 nM. While the potency was

improved under these conditions versus the 17 μM IC₅₀

obtained using the phosphorylated enzyme, the IC₅₀was still

∼10-fold higher than the original HTS HTRF IC₅₀of 0.16 μM.

This shift can partially be explained by the diﬀerence in ATP

concentrations between the assays. For the standard caliper-

G

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 8. View of stabilizing interaction of inactive conformation.

Scheme 7. SAR Progression of HTS Lead 4

Figure 7. Binding pose of 4 in the TrkA active site.

repulsive interaction. A second, similar interaction also exists

between Asp 650 and Tyr 680 (not depicted), another known

TrkA phosphorylation site.

leaving an unfavorable water molecule at the bottom of the

pocket. As such, a para-chloro group was added, providing

compound 23, which eﬀectively displaced the water molecule

and more eﬃciently occupied the hydrophobic cavity resulting

in a further improvement in potency. This interaction was

further optimized through the use of a para-triﬂuoromethoxy

group (compound 24) to more optimally ﬁll the unoccupied

space, as inspired by crystallographic overlays with compound

3, resulting in an additional increase in potency (synthesis

described in Scheme 9). While advances were realized toward

increasing potency with concomitant increases in ligand

eﬃciencies, it was apparent that the majority of the binding

aﬃnity was being derived from lipophilic interactions as

represented in the low LLE values. In order to obtain more

acceptable druglike physiochemical properties, we endeavored

SAR Optimization of 4. Having solved the structure and

elucidated the key interactions of 4 with TrkA, we were now

positioned to follow a rational design approach toward ligand

optimization (Scheme 7).²⁹First, the lipophilic naphthyl group

that was found to bind in the hydrophobic cleft at the front of

the hinge was replaced with a 3-ethoxyphenyl group. It was also

discovered that the isoxazole could be replaced by a β-ethoxy

linker (compound 22), which maintained the linear orientation

and trajectory for positioning of the arene ring into the

selectivity pocket. These modiﬁcations resulted in approx-

imately 10-fold improvements in both functional and intrinsic

activities. It was also evident that the unsubstituted aryl ring did

not optimally ﬁll the selectivity pocket, presumably instead

H

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

to improve on this parameter by lowering the clogP of the

series.

Scheme 9. Synthetic Route for Compound 24

Introduction of Hinge Binding Interaction. In order to

address the high lipophilicity of the series, we envisaged that

further optimizing hydrogen bonding interactions with the

hinge could provide increased aﬃnity without adding to the

lipophilicity of the molecule. We began by removing the very

hydrophobic ethoxy arene that was involved in hydrophobic

van der Waals interactions with the hydrophobic cleft to allow

room for alternative substitution of the indole ring. While

removal of this hydrophobic interaction resulted in a

precipitous 35-fold loss in potency (compound 25, Scheme

8), it should be noted that ligand binding eﬃciency stayed on

Scheme 8. Optimization of Hinge Binding Interaction with

Azaindole Series

a

(a) 3-Ethoxyphenylboronic acid, Pd(OAc)₂, XPhos, KF, 1,4 dioxane

73%; (b) NaH, 2-bromophenyl ether, DMF, 1 N NaOH; (c) 2 N

NaOH, 75% over two steps.

was assayed against the full kinase panel, there was no activity

noted against any of the 92 kinases. This was not entirely

surprising, since all of the kinases used in the panel were

presumed to be the active form of the protein; i.e., they were

phosphorylated kinases. Since we already determined that 4 did

not bind to the active, phosphorylated form of TrkA, it was

conjectured that it would not bind to the activated forms of

other kinases either. These results were also conﬁrmed with

compound 24. In order to more fully assess the functional

selectivity of this series, we submitted 4 and 24 to a panel of

cell-based assays where the kinase should be in a dynamic

equilibrium between the activated and unactivated states, but

here again we were unable to detect any signiﬁcant activity.

Binding Kinetics. Surface plasmon resonance (SPR)

studies using Biacore were performed with several inhibitors

to investigate diﬀerences in their binding kinetics. Rates for

inhibitor association (k_on) and disassociation (k_off) with the

enzyme were measured to reveal the underlying drivers of

diﬀerences in the disassociation constant K_D(where K_D= k_on/

k_off, Table 1). Compound 5, which bound in a DFG-in

par with compound 24 while the lipophilic contributions to

aﬃnity were improved, as is apparent in the increase in LLE to

0.61 for compound 25 compared with −0.48 for compound 24.

Since the indole N5 nitrogen lone pair was directed toward the

backbone residue Met 592 with an intervening water molecule,

we decided to introduce a nitrile group in an attempt to form a

direct hydrogen bonding interaction with the backbone

(compound 26). Gratifyingly, this modiﬁcation led to a 100-

fold increase in both enzymatic and cellular potency. Finally,

replacement of the nitrile with a pyrazole, a common hinge

binding motif, led to compound 27 which has an IC₅₀value of

51 nM in the enzymatic TrkA assay and an IC₅₀of 29 nM in

the TrkA cell-based assay. Thus, by removing a large

contributor to lipophilic binding and focusing on optimizing

the binding interactions with the hinge, we were able to

signiﬁcantly improve the LLE value while maintaining good

overall per atom binding eﬃciency.

Table 1. Binding Aﬃnity of Inhibitor Series

compd

k_on(M⁻¹s⁻¹

)

k_off(s⁻¹

)

K_D(nM)

5

63

2

62 000

1.7 × 10⁻³

1.7 × 10⁻⁴

1.5 × 10⁻²

1.9 × 10⁻³

1.7 × 10⁻⁴

9.0 × 10⁻⁵

27

11

3

180 000

0.94

∼400

63

17

4

30 000

22 000

26 000

7.9

23

3.5

conformation, displayed very fast on and oﬀ rates (yielding a

“square wave” in the Biacore sensogram); ﬁtting the plateau

value gave a K_Din line with its intrinsic activity (K_D= 25 nM vs

IC₅₀= 10 nM). In contrast, all of the inhibitor series that bound

in the DFG-out conformation displayed signiﬁcantly slower on

and oﬀ rates with the enzyme. The triazole HTS lead 2 was

shown to have a k_off= 1.7 × 10⁻³s⁻¹(t_1/2= 6.8 min); further

optimization to 11 led to a dramatic reduction in the oﬀ rate,

k_off= 1.7 × 10⁻⁴s⁻¹(t_1/2= 68 min). Similarly, in the urea

Selectivity against other kinases was somewhat diﬃcult to

assess for this series because of its unique phosphorylation-state

dependent binding paradigm. As mentioned, 4 was inactive

against IGFR1 in our original HTS screen. In addition when 4

I

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

inhibitor series, the initial hit compound 3 displayed a relatively

fast oﬀ rate of 1.5 × 10⁻²s⁻¹(t_1/2= 0.8 min) and K_D≈ 400

nM. Potency of the series was improved with compound 17,

and this tracked with a 10-fold improvement in disassociation

constant, providing K_D= 63 nM that was driven by a 10-fold

improvement in oﬀ rate, k_off= 1.9 × 10⁻³s⁻¹(t_1/2= 6 min).

Finally, the azaindole series had the longest oﬀ rates of the four

series investigated where the initial hit, 4, displayed k_off= 1.7 ×

10⁻⁴s⁻¹(t_1/2= 68 min). Interestingly, the dissociation

constants for this series were much lower than the intrinsic

potency compared to the other series. Here the disassociation

constant for 4 had K_D= 7.9 nM versus IC₅₀= 2300 nM.

Further optimization of this series resulted in compound 23

which displayed a 2-fold improvement in both K_Dand

Figure 9. CFA-induced mechanical hypersensitivity assay (inﬂamma-

tory pain) with compound 11.

residence time, K_D= 3.5 nM; k_off= 9.0 × 10⁻⁵s⁻¹; t_1/2

=

Table 2. Plasma Concentrations in CFA Study at 2 h

128 min. Since the on rates remained relatively constant

between the series, these results suggest that the increases in

potency realized during lead optimization were largely driven

by decreased oﬀ rates. This also suggests that DGF-out

inhibitors should have signiﬁcantly longer residence times on

the receptor compared with DFG-in inhibitors. These data

align with other type II binders that have also been reported to

display lower disassociation rates and extended residence

times.^30,31This further highlights the advantages of targeting

the DGF-out conformation in inhibitor design, since they not

only have the potential to exhibit increased selectivity toward

other kinases but may have the added advantage of slower oﬀ

rates prolonging binding to the receptor.

dose (mg/kg)

[plasma] (μM)

C_u,plasma(nM)

fold IC₅₀

10

1−2

5−10

1−2 ×

30

4−9

16−45

20−45

80−226

4−9 ×

16−45 ×

100

(Figure 10), similar to the eﬀect observed with the positive

control, pregabalin. Here, similar plasma levels were obtained at

In Vivo Activity. While eﬃcacy for the TrkA pathway has

already been established both clinically and preclinically, those

studies were conducted using an anti-NGF antibody approach.

We sought to validate the use of a small molecule TrkA

inhibitor in preclinical in vivo pain models.^32,33To support

these studies, compound 11 was chosen because it exhibited

excellent cellular potency and an extended residence time on

the enzyme. As such, compound 11 was proﬁled in rats and

displayed pharmacokinetic properties suitable for in vivo

studies with a clearance of 16 mL min⁻¹kg⁻¹, a half-life of

2.6 h, an oral bioavailability of 30%, and a plasma protein

binding of 99.5%. To fully explore the in vivo eﬃcacy of a small

molecule TrkA inhibitor, we investigated the eﬀects of

compound 11 in both an inﬂammatory and a neuropathic

pain model. For the inﬂammatory pain studies, we employed

the complete Freund’s adjuvant-induced hind-paw (CFA)

model. Compound 11 was dosed orally b.i.d. for 7 days after

administration of the Freund’s adjuvant, similar to previous

studies using an anti-NGF antibody where reversal of allodynia

was reported on day 3 through day 7 after treatment.³²Figure 9

displays the percentage reversal of mechanical hypersensitivity

at both the predose and 2 h postdose time points on day 7 of

the study. As exempliﬁed in Figure 9, compound 11 inhibited

mechanical allodynia with naproxen-like eﬃcacy at both the 30

and 100 mg/kg dose levels. Table 2 lists the plasma levels of

compound 11 at the 2 h postdose time point. When corrected

for plasma protein binding, signiﬁcant eﬀects are achieved at

unbound drug concentration at or above the IC₅₀values

obtained in the cell-based TrkA assay, with residual eﬀects on

mechanical allodynia remaining at the trough (predose, plasma

levels were not obtained) prior to compound readministration.

To investigate in vivo eﬃcacy in a rat model for neuropathic

pain, we employed the Chung spinal nerve ligation model

where compound 11 demonstrated approximately a 50%

reversal of mechanical hypersensitivity at the 100 mg/kg dose

Figure 10. Spinal nerve ligation assay (neuropathic pain) after oral

dosing of compound 11.

the 2 h time point compared to those measured at 2 h in the

CFA study. Again, eﬃcacy was observed at plasma concen-

tration in moderate excess of the cellular IC₅₀when corrected

for plasma protein binding (Table 3).

Table 3. Plasma Concentrations in SNL Study

dose (mg/kg)

[plasma] (μM)

C_u,plasma(nM)

fold IC₅₀

10

1

5

1×

30

6.7

23

34

115

6×

23×

100

SUMMARY

■

The intention of this manuscript is to serve as an overview of

our eﬀorts to identify novel chemotypes for the development of

selective kinase inhibitors. We have identiﬁed several series of

small molecule inhibitors of TrkA with unique binding modes

in the ﬁeld of kinase inhibitors. The crystallographic structures

have been solved for all four inhibitor series with the TrkA

kinase domain, and their unique binding modes have been

elucidated. These leads were optimized for potency and

selectivity. We employed a rational structure based drug design

approach to progress these series. During the optimization

process we strived to abide by the principles of ligand eﬃciency

J

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

7.30 (m, 1H), 6.98−6.94 (m, 3H), 5.33 (m, 1H), 4.75 (d, J = 5.6 Hz,

2H), 3.88 (m, 4H), 3.15 (m, 4H). HRMS (ESI): m/z calculated for

C₂₃H₂₃N₇O = 413.1964. Found [M + H]⁺: 414.2048.

to maximize binding aﬃnity without relying excessively on

lipophilic interactions. While all of the described inhibitor series

were found to display very high selectivity against a diverse

panel of kinases, because of the high homology in the Trk

active site, they exhibited no selectivity versus the other Trk

homologues, i.e., TrkB and TrkC, and as such are considered

pan-Trk inhibitors. We have also demonstrated eﬃcacy in both

inﬂammatory and neuropathic pain models upon oral dosing.

More detailed discussions with respect to the SAR of the

individual series and optimization of their physical properties

will be the subject of future communications.

N-(3-Cyclopropyl-1-phenyl-1H-pyrazol-5-yl)-2-(4-(3-me-

thoxy-4-(4-methyl-1H-imidazol-1-yl)phenyl-1H-1,2,3-triazol-1-

yl)acetamide (2). Step 1. To a stirred solution of 3-cyclopropyl-1-

phenyl-1H-pyrazol-5-amine (2.00 g, 10.0 mmol) and Et₃N (1.32 g,

13.1 mmol) in THF (50 mL) at room temperature was added 2-

chloroacetyl chloride (1.36 g, 12.1 mmol) dropwise. After being stirred

overnight, the reaction mixture was concentrated. The crude product

was puriﬁed by silica gel chromatography (80 g SiO₂, 3−50% EtOAc

in hexanes) to aﬀord 2-chloro-N-(3-cyclopropyl-1-phenyl-1H-pyrazol-

5-yl)acetamide (2.3 g, 82%). LRMS: [M + H]⁺276.0.

Step 2. A mixture of 2-chloro-N-(3-cyclopropyl-1-phenyl-1H-

pyrazol-5-yl)acetamide (2.77 g, 10.1 mmol) and NaN₃(0.98 g, 15.1

mmol) in EtOH (35 mL) was heated at 60 °C for 16 h. The reaction

mixture was concentrated, diluted with water, and extracted with

EtOAc (3 × 30 mL). The combined organic extracts were dried over

Na₂SO₄and concentrated. Puriﬁcation by silica gel chromatography

(80 g SiO₂, 3−60% EtOAc in hexanes) aﬀorded 2-azido-N-(3-

cyclopropyl-1-phenyl-1H-pyrazol-5-yl)acetamide (2.0 g, 71%). LRMS:

[M + H]⁺283.1.

EXPERIMENTAL SECTION

■

General Chemistry Information. Reagents and solvents,

including anhydrous THF, dichloromethane, and DMF, were

purchased from Aldrich, Acros, or other commercial sources and

were used without further puriﬁcation. Reactions that were moisture

sensitive or that required the use of anhydrous solvents were

performed under either nitrogen or argon atmosphere. Analytical

thin layer chromatography (TLC) was performed on precoated silica

gel plates obtained from Analtech. Visualization was accomplished

with UV light or by staining with basic KMnO₄solution or ethanolic

H₂SO₄solution. Compounds were puriﬁed by ﬂash chromatography

using an automated puriﬁcation system (ISCO, Biotage, Analogix)

using disposable silica gel prepacked cartridges. Alternatively,

compounds were puriﬁed by preparative reverse-phase HPLC using

a Gilson 215 liquid handler and a Phenomenex Luna C18 column

(150 mm × 20 mm i.d.) with a linear gradient over 15 min (95:5 to

0:100 H₂O containing 0.1% triﬂuoroacetic acid/acetonitrile or 0.1%

formic acid/acetonitrile). NMR spectra were recorded on 400 or 500

MHz on a Bruker or Varian spectrometer with CDCl₃or DMSO-d₆as

solvent. The chemical shifts are given in ppm, referenced to the

deuterated solvent signal. Purity of target compounds was determined

using LCMS and HPLC. LCMS analyses were performed using an

Applied Biosystems API-150 mass spectrometer and Shimadzu SCL-

10A LC system: column, Phenomenex Gemini C18, 5 μm, 50 mm ×

4.6 mm i.d.; gradient, from 90% water, 10% CH₃CN, 0.05% TFA, 5

min, to 5% water, 95% CH₃CN, 0.05% TFA in 5 min; UV detection,

254 nm. Compound purity was determined by integrating peak areas

of the liquid chromatogram, monitored at 254 nm. Purity of targets

compounds was ≥95%.

Step 3. To a solution of 2-azido-N-(3-cyclopropyl-1-phenyl-1H-

pyrazol-5-yl)acetamide (0.596 g, 2.11 mmol) and 1-(4-ethynyl-2-

methoxyphenyl-4-methyl-1H-imidazole (0.471 g, 2.22 mmol) in 1:1

tert-butanol/water (12 mL) was added a 1 M solution of copper(II)

sulfate (0.422 mL, 0.422 mmol) and copper dust (0.107 g, 1.69

mmol), and the mixture was heated in a sealed vial at 125 °C for 1 h.

The mixture was cooled and diluted with water. Concentrated

ammonium hydroxide was added to convert all of the remaining

copper to copper−hexaamine complex and extracted with EtOAc (3 ×

15 mL). The combined organics were dried over sodium sulfate,

ﬁltered, and concentrated. Puriﬁcation by silica gel chromatography

(isocratic 10% MeOH/EtOAc) aﬀorded N-(3-cyclopropyl-1-phenyl-

1H-pyrazol-5-yl)-2-(4-(3-methoxy-4-(4-methyl-1H-imidazol-1-yl)-

1

phenyl-1H-1,2,3-triazol-1-yl)acetamide (0.6 g, 58%). H NMR (400

MHz, DMSO-d₆) δ 10.59 (s, 1 H), 9.38 (s, 1H), 8.76 (s, 1H), 7.79 (m,

2H), 7.67 (m, 2H), 7.52 (m, 3H), 7.40 (m, 1H), 6.20 (s, 1H), 5.42 (s,

2H), 3.96 (s, 3H), 2.36 (s, 3H), 1.94 (m, 1H), 0.91 (m, 2H), 0.69 (m,

2H). HRMS calculated for C₂₇H₂₆N₈O₂, 494.2179; found (ES) m/z

[M + H]⁺, 495.2251.

1-Cyclopropyl-1-(3-(thiazol-2-yl)benzyl)-3-(4-

(triﬂuoromethoxy)phenyl)urea (3). Step 1. A 20 L ﬂask ﬁtted with

an overhead mechanical stirrer was charged with 4.1 L of toluene,

followed by aqueous K₂CO₃(4.5 L, 1M). The reaction mixture was

stirred and degassed with nitrogen over 15−30 min, after which 2-

bromothiazole (325 g, 1.8 mol) was quickly charged to the ﬂask. The

solution was degassed with stirring for a further 30 min. Tetrakis-

(triphenylphosphine)palladium(0) (41.6 g, 0.036 mol) was added to

the mixture under an inert atmosphere and the mixture degassed over

30 min. 3-(Formylphenyl)boronic acid (270 g, 1.98 mol) was

dissolved in ethanol (2 L) in a 5 L conical ﬂask and the suspension

degassed with nitrogen (30 min). The ethanol mixture was rapidly

charged to the 20 L ﬂask and the reaction mixture degassed for a

further 30 min. With vigorous stirring the mixture was heated at 80 °C

under an atmosphere of nitrogen overnight (12 h). Approximately 2−

4 L of solvent was initially distilled from the reaction to aid the workup

procedure. Upon cooling, the reaction solution was transferred to a

sinter using a siphon and ﬁltered through Celite. The Celite was

washed with toluene (2 × 500 mL) and the combined organic layer

separated. The aqueous layer was then washed with successive

portions of toluene (2 × 2 L), and the organic layers were combined

and washed with 1 M aqueous K₂CO₃(2 × 2.5 L). The organic phase

was separated, dried over MgSO₄and ﬁltered through Celite. The

solvent was removed in vacuo to provide 3-(thiazol-2-yl)benzaldehyde

(356.9 g, 59%).

N4-(4-Morpholinophenyl)-N6-(pyridin-3-ylmethyl)pyrido-

[3,2-d]pyrimidine-4,6-diamine (1). Step 1. Commercially available

4,6-dichloropyrido[3,2-d]pyrimidine (0.20 g, 1.0 mmol), 4-morpholi-

noaniline (0.19 g, 1.1 mmol), and DIPEA (0.18 mL, 1.1 mmol) were

dissolved in DMF (3 mL), and the mixture was stirred at 80 °C for 12

h. The reaction mixture was cooled to room temperature, and water

(30 mL) was added. The aqueous solution was extracted using ethyl

acetate (3 × 50 mL). The combined organic layers were dried over

MgSO₄, ﬁltered, and concentrated. The crude product was puriﬁed

using silica gel chromatography (24 Gm Si gel, 0−20% MeOH in

DCM) to aﬀord 6-chloro-N-(4-morpholinophenyl)pyrido[3,2-d]-

1

pyrimidin-4-amine (0.34 g, 99% yield). H NMR (400 Hz, CDCl₃)

δ 8.76 (s, 1H), 8.72 (s, 1 H), 8.10 (d, J = 8.8 Hz, 1 H), 7.76 (d, J = 8.8

Hz, 2 H), 7.67 (d, J = 8.8 Hz, 1H), 6.99 (d, J = 9.0 Hz, 2 H), 3.88 (m,

4H), 3.18 (m, 4 H). LRMS, M + H: 342.8.

Step 2. A solution of 6-chloro-N-(4-morpholinophenyl)pyrido[3,2-

d]pyrimidin-4-amine (0.15 g, 0.44 mmol), pyridin-3-ylmethane amine

(0.10 g, 0.88 mmol), and DIPEA (0.15 mL, 0.88 mmol) in NMP (3

mL) was heated to 90 °C for 48 h. The reaction mixture was then

cooled, diluted with water (30 mL), and extracted with EtOAc (3 × 50

mL). The combined organic layers were dried over MgSO₄, ﬁltered,

and concentrated. The crude product was puriﬁed using silica gel

chromatography (24 Gm Si gel, 0−20% MeOH in DCM) to aﬀord

N4-(4-morpholinophenyl)-N6-(pyridin-3-ylmethyl)pyrido[3,2-d]-

Step 2. 3-(Thiazol-2-yl)benzaldehyde (13.1 g, 0.069 mol) was

charged to a 250 mL ﬂask, followed by 65 mL of THF and TMOF

(7.51 mL, 0.069 mol). Cyclopropylamine (5.90 g, 0.102 mol) was

subsequently added to the ﬂask and the reaction solution shaken

1

pyrimidine-4,6-diamine (0.09 g, 51%). H NMR (400 Hz, CDCl₃) δ

8.74 (s, 1H), 8.57 (d, J = 4.9 Hz, 1 H), 8.54 (s, 1H), 8.44 (s, 1H), 7.87

(d, J = 8.8 Hz, 1H), 7.75 (d, J = 7.8 Hz, 1H), 7.70 (d, J = 9.0 Hz, 2H),

K

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

overnight at room temperature. The solvent was removed in vacuo,

and the imine was used directly in the reduction step without further

puriﬁcation. The imine (69 mmol) was dissolved in MeOH (100 mL)

and charged to a 250 mL ﬂask. The solution was cooled in an ice bath

and NaBH₄(5.20 g, 0.138 mol) was carefully added, maintaining the

reaction temperature below 30 °C. The mixture was shaken overnight

at room temperature. Reduction of the imine was monitored by

HPLC−MS, and upon reaction completion the solution was

concentrated to dryness in vacuo, redissolved in ethyl acetate (100

mL), and washed with 1 M aqueous K₂CO₃(2 × 100 mL). Following

separation, the aqueous layer was washed with two further portions of

ethyl acetate (2 × 50 mL) and the organic phases were combined,

dried over MgSO₄, and concentrated in vacuo. PS-CHO (43g, 1.3

mmol/g, 0.6 equiv) and PS-AMPS (29.3 g, 1.9 mmol/g, 0.6 equiv)

were transferred to the vessel, and the mixture was shaken at room

temperature for 1−2 days. The mixture was ﬁltered. The resin was

washed with acetonitrile (2 × 50 mL) and the ﬁltrate concentrated

under vacuum. The crude amine was puriﬁed by silica gel

chromatography (heptane/ethyl acetate; DCM/ethyl acetate) to

aﬀord N-(3-(thiazo-2-yl)benzyl)cyclopropanamine (9.30 g, 61%).

Step 3. The secondary amine was prepared as a stock solution in

DCE to give 0.4 mmol (∼100 mg) per 0.5 mL of solvent and

dispensed into a 10 mL plate. A stock solution of 1-isocyanato-4-

(triﬂuoromethoxy)benzene in DCE (0.6 mmol) was added to the

amine, and the reaction plate was agitated for 2 days at room

temperature. By use of the Radley’s Titan resin loader, PS isocyanate

(350 mg, 0.4 mmol, 1 equiv, 1.2 mmol/g), and PS-AMPS (210 mg, 0.4

mmol, 1 equiv, 1.9 mmol/g), scavenger resins were added to the

reaction plates. A further 2−3 mL of DCE was added to the well to

induce resin mobility and the plate was shaken for 2 days at room

temperature. The reaction well was individually agitated after 1 day of

scavenging. After this scavenging period the DCE solution was

removed using a Lissy liquid handler to a 5 mL plate using the

following protocol for Lissy transfer from 10 mL plate to 5 mL plate:

(4-methylpiperazin-1-yl)aniline (0.201 g, 1.05 mmol), and DIPEA

(0.183 mL, 1.05 mmol) were dissolved in DMF (3 mL), and the

mixture was stirred at 80 °C for 12 h. The reaction mixture was then

cooled to room temperature, and water (30 mL) was added. The

aqueous solution was extracted with EtOAc (3 × 50 mL). The

combined organic layers were dried over MgSO₄, ﬁltered, and

concentrated. The crude product was puriﬁed using silica gel

chromatography (24 Gm Si gel, 0−20% MeOH in DCM) to aﬀord

6-chloro-N-(4-(4-methylpiperazin-1-yl)phenyl)pyrido[3,2-d]-

1

pyrimidin-4-amine (0.21 g, 59% yield). H NMR (400 Hz, CDCl₃) δ

8.73 (s, 1 H), 8.72 (s, 1H), 8.09 (d, J = 8.8 Hz, 1H), 7.74 (d, J = 8.8

Hz, 2H), 7.66 (d, J = 8.8 Hz, 1H), 7.00 (d, J = 8.8 Hz, 2H), 3.25 (m,

4H), 2.62 (m, 4H), 2.39 (s, 3H). LRMS, M + H: 355.1.

Step 2. A solution of 6-chloro-N-(4-(4-methylpiperazin-1-yl)-

phenyl)pyrido[3,2-d]pyrimidin-4-amine (0.10 g, 0.28 mmol), (3-

chlorophenyl)methanamine (0.08 g, 0.56 mmol), and DIPEA (0.098

mL, 0.56 mmol) was dissolved in NMP (3 mL) and heated to 90 °C

for 48 h. The reaction mixture was cooled, diluted with water (30 mL),

and extracted with EtOAc (3 × 50 mL). The combined organic layers

were dried over MgSO₄, ﬁltered, and concentrated. The crude product

was puriﬁed using silica gel chromatography (24 Gm Si gel, 0−20%

MeOH in DCM) to aﬀord N6-(3-chlorobenzyl)-N4-(4-(4-methyl-

piperazin-1-yl)phenyl)pyrido[3,2-d]pyrimidine-4,6-diamine (0.038 g,

1

29%). H NMR (400 Hz, CDCl₃) δ 8.53 (s, 1H), 8.44 (s, 1H), 7.84

(d, J = 9.0 Hz, 1H), 7.67 (d, J = 8.8 Hz, 2H), 7.43 (s, 1 H), 7.31−7.26

(m, 3H), 6.98 (d, J = 8.8 Hz, 2H), 6.93 (d, J = 9.0 Hz, 1H), 5.30 (s,

1H), 4.68 (d, J = 5.4 Hz, 2H), 3.21 (m, 4H), 2.61 (m, 4H), 2.37 (s,

3H). LRMS, M + H: 460.2. HRMS (ESI): m/z calculated for

+

C₂₅H₂₆ClN₇459.1928. Found [M + H] : 460.2037.

N-(1H-Benzimidazol-2-ylmethyl)-6-[(2R)-2-(2,5-

diﬂuorophenyl)pyrrolidin-1-yl]pyrido[3,2-d]pyrimidin-4-amine

(6). Step 3. 4-Chloro-6-[(2R)-2-(2,5-diﬂuorophenyl)pyrrolidin-1-yl]-

pyrido[3,2-d]pyrimidine (compound 9) (10 mg, 0.029 mmol), 1-(1H-

benzimidazol-2-yl)methanamine (21 mg, 0.087 mmol), and DIPEA

(30 μL, 0.17 mmol) were dissolved in 2 mL of EtOH. The vial was

irradiated in a microwave reactor at 110 °C for 15 min. The mixture

was concentrated and the crude product was puriﬁed by reverse phase

HPLC (0−95% 0.01% TFA−acetonitrile gradient over 20 min) to

aﬀord N-(1H-benzimidazol-2-ylmethyl)-6-[(2R)-2-(2,5-

diﬂuorophenyl)pyrrolidin-1-yl]pyrido[3,2-d]pyrimidin-4-amine (7.9

(1) Add 1 mL of MeOH to each well, and manually agitate the well.

(2) Transfer 3.5 mL of solvent from the 10 mL master plate to a 5

mL plate using level tracking.

(3) Remove solvent from 5 mL plate in a Genevac.

(4) Add 2 mL of methanol to the well of the 10 mL master plate

and mix.

(5) Transfer remaining solvent from the 10 mL master plate to the

5 mL plate using level tracking.

(6) Remove solvent from 5 mL plate in a Genevac.

1

mg, 58%) . H NMR (400 MHz, CDCl₃): δ 10.51 (NH, s, 1 H),

8.65 (s, 1 H), 7.82 (m, 1 H), 7.73 (m, 2 H), 7.46 (dd, J = 6.4, 3.2 Hz, 2

H), 7.07 (m, 1 H), 6.93 (m, 1 H); 6.68 (m, 2 H), 5.52 (d, J = 6.7 Hz, 2

H), 5.17 (m, 1 H), 4.12 (m, 1 H), 3.95 (m, 1 H), 2.68−2.46 (m, 1 H),

2.10 (m, 4 H). HRMS (ES) m/z calculated for C₂₄H₂₁F₂N₇S 477.1547.

Found [M + H⁺]⁺: 478.1610.

Puriﬁcation was performed on a reverse-phase Gilson Prep LC module

Xterra Prep MS C18, 50 mm × 19 mm i.d., 5 μm column, with 10 mM

ammonium bicarbonate, buﬀered to pH 10 (water/ACN) over 7 min

at [20 mL/min]. Product fractions were concentrated under reduced

pressure to aﬀord 1-cyclopropyl-1-(3-(thiazol-2-yl)benzyl)-3-(4-

4-Chloro-6-[(2R)-2-(2,5-diﬂuorophenyl)pyrrolidin-1-yl]-

pyrido[3,2-d]pyrimidine (8). Commerically available 6-

chloropyrido[3,2-d]pyrimidin-4(3H)-one (compound 7) (250 mg,

1.38 mmol) and (2R)-2-(2,5-diﬂuorophenyl)pyrrolidine (504 mg, 2.75

mmol) were dissolved in EtOH (2 mL). The vial was irradiated in a

microwave reactor at 150 °C for 2 h. The mixture was concentrated

and the crude product was puriﬁed by reverse phase HPLC (0−95%

0.01% TFA−acetonitrile gradient over 20 min) to give 4-chloro-6-

[(2R)-2-(2,5-diﬂuorophenyl)pyrrolidin-1-yl]pyrido[3,2-d]pyrimidine

(317 mg, 69%). LRMS (ES) m/z M + H calcd, 329; found, 329.

4-Chloro-6-[(2R)-2-(2,5-diﬂuorophenyl)pyrrolidin-1-yl]-

pyrido[3,2-d]pyrimidine (9). 6-[(2R)-2-(2,5-Diﬂuorophenyl)-

pyrrolidin-1-yl]pyrido[3,2-d]pyrimidin-4(3H)-one (compound 8)

(465 mg, 1.42 mmol) was suspended in POCl₃(132 μL, 1.42

mmol). The mixture was heated at 100 °C for 2 h. The reaction

mixture was added to a mixture of ice, ethyl acetate (20 mL), and

aqueous sodium hydrogen carbonate (saturated, 20 mL). The mixture

was shaken well, and the aqueous solution was adjusted to be around

pH 8. The aqueous layer was extracted with EtOAc (2 × 20 mL). The

combined organic fractions were washed with aqueous sodium

hydrogen carbonate (saturated, 20 mL), followed by brine and then

dried, ﬁltered and the solvent was evaporated under reduced pressure

to provide 4-chloro-6-[(2R)-2-(2,5-diﬂuorophenyl)pyrrolidin-1-yl]-

pyrido[3,2-d]pyrimidine (343 mg, 70%). ¹H NMR (400 MHz,

1

(triﬂuoromethoxy)phenyl)urea (158 mg, 46%). H NMR δ (DMSO-

d₆): 8.59 (s, 1H), 7.92 (d, J = 3.2 Hz, 1H), 7.78−7.86 (m, 3H), 7.67−

7.69 (m, 2H), 7.48 (t, J = 7.67 Hz, 1H), 7.37 (d, J = 7.6 Hz, 1H), 7.27

(d, J = 8.77 Hz, 2H), 4.62 (s, 2H), 2.62−2.65 (m, 1H), 0.90 (t, J =

6.38 Hz, 2H), 0.77−0.79 (m, 2H). HRMS (ESI), m/z calculated for

C₂₁H₁₈F₃N₃O₂S: 433.1100. Found [M + H]⁺: 434.1151.

4-(Naphthylen-10-yl)-1-((5-phenyl-1,2,4-oxadiazol-3-yl)-

methyl)-1H-pyrrolo[3,2-c]pyridine-2-carboxylic Acid (4). Com-

pound 4, 4-(naphthylen-10-yl)-1-((5-phenyl-1,2,4-oxadiazol-3-yl)-

methyl)-1H-pyrrolo[3,2-c]pyridine-2-carboxylic acid, was part of the

HTS screen collection and was not independently synthesized again

after validation, as such full experimental procedures are not available,

1

but the ﬁnal compound has been characterized. H NMR (CDCl₃): δ

8.78 (d, J = 6.6 Hz, 1H), 8.15−7.95 (m, 3H), 7.86 (d, J = 8.1 Hz, 1H),

7.82 (d, J = 6.8 Hz, 1H), 7.73 (d, J = 7.1 Hz, 1H), 7.59−7.51 (m, 3H),

7.48−7.44 (m, 3H), 7.36 (t, J = 7.6 Hz, 1H), 7.13 (s, 1H), 6.06 (s,

2H). HRMS (ESI): m/z calculated for C₂₇H₁₈N₄O₃446.1379. Found:

[M + H⁺]⁺: 447.1474.

N6-(3-Chlorobenzyl)-N4-(4-(4-methylpiperazin-1-yl)phenyl)-

pyrido[3,2-d]pyrimidine-4,6-diamine (5). Step 1. Commercially

available 4,6-dichloropyrido[3,2-d]pyrimidine (0.20 g, 1.0 mmol), 4-

L

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

CDCl₃): δ 2.05−2.20 (m, 3 H), 2.48−2.61 (m, 1 H), 3.75−4.25 (m, 2

H), 5.20−5.75 (bs, 1 H), 6.74−6.80 (m, 1 H), 6.86−6.96 (m, 1 H),

7.10−7.50 (m, 1 H), 7.92−8.02 (m, 1 H), 8.76 (s, 1 H). LRMS (ES)

m/z M + H calcd, 347; found, 347.

g, 8.44 mmol), imidazole (0.575 g, 8.44 mmol), ^L-proline (0.194 g,

1.69 mmol), CuI (0.080 g, 0.42 mmol), and K₂CO₃(2.33 g, 16.89

mmol) in DMSO (8 mL) was heated at 90 °C, under nitrogen, for 15

h. The reaction mixture was then cooled to room temperature and

partitioned between ether (200 mL) and water (100 mL). The

aqueous layer was separated and extracted with ether (2 × 100 mL).

Combined organic extracts were washed with saturated aqueous

NaHCO₃, dried over Na₂SO₄, and concentrated to aﬀord 5-bromo-2-

(1H-imidazol-1-yl)pyridine (1.16 g, 61%) which was used as is in the

next step. LRMS: [M + H]⁺225.9.

5-Ethynyl-2-(1H-imidazol-1-yl)pyridine (14). To a mixture of 5-

bromo-2-(1H-imidazol-1-yl)pyridine (200 mg, 0.893 mmol),

PdCl₂(PPh₃)₂(50 mg, 0.071 mmol), and CuI (27 mg, 0.143 mmol),

under nitrogen, were added DMF (4 mL) and Et₃N (1 mL, 7.17

mmol) followed by trimethylsilylacetylene (0.5 mL, 3.57 mmol). The

reaction mixture was heated at 90 °C, under nitrogen, for 3 h and then

cooled to room temperature, diluted with ether (30 mL), and ﬁltered

through a pad of Celite. The ﬁltrate was washed with water (1 × 20

mL) and brine (1 × 20 mL), dried over Na₂SO₄, and concentrated.

Puriﬁcation by silica gel chromatographic (12 g SiO₂, 70−100%

EtOAc in hexanes) aﬀorded 2-(1H-imidazol-1-yl)-5-((trimethylsilyl)-

ethynyl)pyridine (167 mg, 78%). LRMS: [M + H]⁺242.1.

Step 2. Potassium ﬂuoride (48 mg, 0.83 mmol) was added to a

solution of 2-(1H-imidazol-1-yl)-5-((trimethylsilyl)ethynyl)pyridine

(166 mg, 0.688 mmol) in MeOH (1 mL) and stirred at room

temperature for 1 h. The reaction mixture was diluted with water and

extracted with DCM (3 × 10 mL). Sodium chloride was then added to

the aqueous solution until saturation and extracted with EtOAc (2 ×

10 mL). The combined organic extracts were dried over Na₂SO₄and

concentrated to provide 5-ethynyl-2-(1H-imidazol-1-yl)pyridine (116

mg, 100%). LRMS: [M + H]⁺170.1.

2-Azido-N-(3-(tert-butyl)-1-(pyridin-3-yl)-1H-pyrazol-5-yl)-

acetamide (15). To a stirring solution of commercially available 3-

(tert-butyl)-1-(pyridin-3-yl)-1H-pyrazol-5-amine (compound 16) (700

mg, 3.24 mmol) and Et₃N (426 mg, 4.21 mmol) in DCM (12 mL) at

0 °C was added 2-chloroacetyl chloride (439 mg, 3.88 mmol)

dropwise. After 20 min, the reaction mixture was concentrated and

puriﬁed by silica gel chromatography (40 g SiO₂, 10−100% EtOAc in

hexanes) to aﬀord N-(3-(tert-butyl)-1-(pyridin-3-yl)-1H-pyrazol-5-yl)-

2-chloroacetamide (682 mg, 72%). LRMS: [M + H]⁺293.0.

A mixture of N-(3-(tert-butyl)-1-(pyridin-3-yl)-1H-pyrazol-5-yl)-2-

chloroacetamide (682 mg, 2.33 mmol) and NaN₃(227 mg, 3.49

mmol) in EtOH (25 mL) was heated at 60 °C for 16 h. The reaction

mixture was concentrated, diluted with water, and extracted with

EtOAc (3 × 100 mL). Combined organic extracts were dried over

Na₂SO₄and concentrated. Puriﬁcation by silica gel chromatographic

(24 g SiO₂, 30−100% EtOAc in hexanes) aﬀorded 2-azido-N-(3-(tert-

butyl)-1-(pyridin-3-yl)-1H-pyrazol-5-yl)acetamide (640 mg, 92%).

LRMS: [M + H]⁺300.1.

1 - I s o p r o p y l - 1 - ( 3 - ( t h i a z o l - 2 - y l ) b e n z y l ) - 3 - ( 4 -

(triﬂuoromethoxy)phenyl)urea (17). Step 1. To a solution of N-(3-

(thiazol-2-yl)phenyl)methanamine (1.23 g, 6.46 mmol) in dichloro-

ethane (20 mL) was added acetone (0.32 mL, 4.30 mmol) followed by

acetic acid (0.26 mL, 4.30 mmol). The reaction mixture was allowed to

stir for 10 min after which a white solid precipitated. To the mixture

was then added sodium triacetoxy borohydride (1.37 g, 6.46 mmol),

and the mixture was allowed to stir at ambient temperature for 16 h.

The reaction mixture was concentrated under reduced pressure and

partitioned between 1 N NaOH (50 mL) and ethyl acetate (3 × 50

mL). The combined organic layers were washed with water (1 × 50

mL) followed by brine (1 × 50 mL). The combined organic layers

were dried over sodium sulfate, ﬁltered, and concentrated under

reduced pressure. Puriﬁcation by silica gel chromatography (80 g SiO₂,

5−10% MeOH/DCM over 45 min) aﬀorded N-(3-(thiazol-2-yl)-

benzyl)propan-2-amine (782 mg, 78%). ¹H NMR δ (CDCl₃): 7.94 (s,

1 H), 7.81−7.86 (m, 2 H), 7.39−7.40 (m, 2 H), 7.33 (d, J = 3.3 Hz, 1

H), 3.85 (s, 2 H), 2.86−2.92 (m, 1 H), 1.12 (d, J = 6.2 Hz, 6 H).

LCMS: m/z [M + H⁺] = 233.1.

N-(3-Cyclopropyl-1-phenyl-1H-pyrazol-5-yl)-2-(4-(4-(4-

methyl-1H-imidazole-1-yl)phenyl-1H-1,2,3-triazol-1-yl)-

acetamide (10). Step 1. To a solution of 4-ﬂuorobenzaldehyde (5.00

g, 40.3 mmol) and 4-methyl-1H-imidazole (6.68 g, 81 mmol) in DMF

(20 mL), under nitrogen, was added K₂CO₃(8.35 g, 60.4 mmol). The

reaction mixture was heated at 100 °C for 19 h. The reaction mixture

was cooled to room temperature and diluted with EtOAc (200 mL).

The organic solution was washed with saturated aqueous NaHCO₃(2

× 50 mL), dried over Na₂SO₄, and concentrated. Puriﬁcation by ﬂash

chromatography (120 g SiO₂, 0−10% MeOH in DCM) aﬀorded 4-(4-

methyl-1H-imidazole-1-yl)benzaldehyde (822 mg, 11%). LRMS: [M +

H]⁺187.1.

Step 2. To a solution of 4-(4-methyl-1H-imidazole-1-yl)-

benzaldehyde (500 mg, 2.69 mmol) and dimethyl (1-diazo-2-

oxopropyl)phosphonate (619 mg, 3.22 mmol) in anhydrous MeOH

(10 mL) was added K₂CO₃(742 mg, 5.4 mmol). The reaction mixture

was stirred at room temperature for 24 h, diluted with DCM (60 mL),

and washed with water (1 × 50 mL). The aqueous solution was

extracted with DCM (1 × 50 mL). Combined organic extracts were

dried over Na₂SO₄and concentrated. Puriﬁcation by silica gel

chromatography (24 g SiO₂, 0−10% MeOH in DCM) aﬀorded 1-

(4-ethynylphenyl)-4-methyl-1H-imidazole (466 mg, 95%). LRMS: [M

+ H]⁺183.1.

Step 3. To a solution of 1-(4-ethynylphenyl)-4-methyl-1H-imidazole

(43 mg, 0.237 mmol) and 2-azido-N-(3-cyclopropyl-1-phenyl-1H-

pyrazol-5-yl)acetamide (67 mg, 0.237 mmol) in t-BuOH−H₂O (1:1, 3

mL) were added copper powder (15 mg, 0.237 mmol) and aqueous

CuSO₄(1 M, 0.237 mL). The reaction mixture was heated at 125 °C,

under microwave irradiation, for 10 min. After cooling to room

temperature, concentrated NH₄OH (10 mL) was added. The mixture

was stirred for 5 min, and then the aqueous portion was extracted with

EtOAc (3 × 20 mL). Combined organic extracts were dried over

Na₂SO₄and concentrated. Puriﬁcation by reverse-phase HPLC (5−

95% MeCN in water, 0.05% TFA modiﬁer used for the mobile phase)

aﬀorded N-(3-cyclopropyl-1-phenyl-1H-pyrazol-5-yl)-2-(4-(4-(4-

methyl-1H-imidazole-1-yl)phenyl-1H-1,2,3-triazol-1-yl)acetamide (37

1

mg, 34%). H NMR (400 MHz, DMSO-d₆) δ 10.52 (s, 1H), 8.59

(s, 1H), 8.26 (s, 1H), 7.98 (d, J = 7.2 Hz, 2H), 7.71 (d, J = 7.2 Hz,

2H), 7.53−7.48 (m, 5H), 7.41−7.37 (m, 1H), 6.20 (s, 1H), 5.37 (s,

2H), 2.18 (s, 3H), 1.93−1.87 (m, 1H), 0.92−0.87 (m, 2H), 0.71−0.67

(m, 2H). HRMS calculated for C₂₆H₄₆N₈O, 464.2073; found (ES) m/

z [M + H]⁺, 465.2145.

2-(4-(6-(1H-Imidazol-1-yl)pyridin-3-yl-1H-1,2,3-triazol-1-yl)-

N-(3-tert-butyl)-1-(pyridin-3-yl)-1H-pyrazol-5-yl)acetamide

(11). To a solution of 5-ethynyl-2-(1H-imidazol-1-yl)pyridine

compound (14) (2.30 g, 13.4 mmol) and 2-azido-N-(3-cyclopropyl-

1-phenyl-1H-pyrazol-5-yl)acetamide (compound 15) (4.07 g, 13.6

mmol) in DCM−t-BuOH (1:12, 65 mL) was added copper powder

(0.691 g, 10.9 mmol) followed by water (20 mL) and aqueous CuSO₄

(1 M, 2.72 mL). The reaction mixture was heated at 90 °C for 2 h,

cooled to room temperature, and diluted with EtOAc (300 mL). To

the resulting solution were added concentrated NH₄OH (60 mL) and

water (20 mL). The biphasic mixture was stirred vigorously for 30

min. Layers were separated. The aqueous solution was saturated with

NaCl and extracted with EtOAc (4 × 200 mL). Combined organic

extracts were dried over Na₂SO₄and concentrated. The resulting solid

was recrystallized from hot EtOAc to aﬀord 2-(4-(6-(1H-imidazol-1-

yl)pyridin-3-yl-1H-1,2,3-triazol-1-yl)-N-(3-tert-butyl)-1-(pyridin-3-yl)-

1H-pyrazol-5-yl)acetamide (4.60 g, 72%). ¹H NMR (400 MHz,

DMSO-d₆) δ 10.67 (br, 1H), 8.99 (d, J = 0.9 Hz, 1H), 8.81 (d, J = 1.4

Hz, 1H), 8.71 (s, 1H), 8.59 (s, 2H), 8.46−8.43 (m, 1H), 8.01−7.93

(m, 3H), 7.56−7.53 (m, 1H), 7.15 (s, 1H), 6.46 (s, 1H), 5.42 (s, 2H),

1.28 (s, 9H). HRMS calculated for C₂₅H₂₄N₁₀O, 468.2138; found (ES)

m/z [M + H]⁺, 469.2207.

5-Bromo-2-(1H-imidazol-1-yl)pyridine (13). Step 1. A mixture

of commercially available 2,5-dibromopyridine (compound 12) (2.00

Step 2. To a solution of N-(3-(thiazol-2-yl)benzyl)propan-2-amine

(150 mg, 0.65 mmol) in dichloroethane (2 mL) was added 4-

M

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(triﬂuoromethoxy)phenyl isocyanate (131 mg, 0.65 mmol). The

mixture was stirred at ambient temperature for 2 h, concentrated

under reduced pressure. Puriﬁcation by silica gel chromatography (40

g SiO₂, 0−50% ethyl acetate/hexanes, over 45 min) aﬀorded 1-

isopropyl-1-(3-(thiazol-2-yl)benzyl)-3-(4-(triﬂuoromethoxy)phenyl)-

{[(benzyloxy)carbonyl]amino}-3-(2-chloro-4-iodopyridin-3-yl)prop-

2-enoate (11.46 g, 86%). ¹H NMR (400 MHz, CDCl₃): δ 7.89 (d, J =

5.1 Hz, 1 H); 7.70 (d, J = 5.1 Hz, 1 H); 7.33 (m, 3 H); 7.24 (m, 2 H);

7.06 (s, 1 H); 6.96 (s, 1 H); 4.97 (s, 2 H); 3.91 (s, 3 H).

Step 2. Methyl (2Z)-2-{[(benzyloxy)carbonyl]amino}-3-(2-chloro-

4-iodopyridin-3-yl)prop-2-enoate (11.46 g, 24.25 mmol), K₂C0₃

(10.05 g, 72.7 mmol), CuI (0.462 g, 2.43 mmol), and L-proline

(0.558 g, 4.85 mmol) were combined in 1,4-dioxane (120 mL). The

mixture was degassed (3 × pump/N₂) and heated to reﬂux. After

being stirred at reﬂux overnight, the mixture was cooled to room

temperature and diluted with saturated aqueous 1 N HCl and

extracted with dichloromethane (3 × 50 mL). The combined organic

layers were dried over MgSO₄, ﬁltered through a pad of Celite, and

concentrated. The resulting solid was triturated with Et₂O/hexanes to

give methyl-4-chloro-1H-pyrrolo[3,2-c]pyridine-2-carboxylate (4.6 g,

1

urea (263 mg, 94%). H NMR δ (CDCl₃): 8.00 (1 H, s), 7.88−7.91

(m, 2 H), 7.37−7.51 (m, 3 H), 7.20 (d, J = 8.7 Hz, 2 H), 7.06 (d, J =

8.6 Hz, 2 H), 6.23 (s, 1 H), 4.70−4.77 (m, 1 H), 4.53 (s, 2 H), 1.26 (d,

J = 6.8 Hz, 6 H). HRMS (ESI), m/z calculated for C₂₁H₂₀F₃N₃O₂S:

436.1301. Found: [M + H]⁺436.1301.

1-((1H-Pyrrolo[2,3-b]pyridin-3-yl)methyl)-1-isopropyl-3-(4-

(triﬂuoromethoxy)phenyl)urea (18). Step 1. To a solution of

isopropylamine (0.360 g, 6.09 mmol) in methanol (10 mL) was added

commercially available 1-(tert-butoxycarbonyl)-3-formyl-7-azaindole

(compound 20) (1.50 g, 6.09 mmol). The mixture was stirred at

ambient temperature for 2 h to allow for imine formation. Sodium

borohydride was added (0.576 g, 15.2 mmol) and stirred for 16 h at

ambient temperature. The solution was concentrated under reduced

pressure and partitioned between ether (50 mL) and 10% potassium

carbonate (50 mL). Extraction was with ether (2 × 50 mL). The

combined organic layers were washed with water (10 mL) and brine

(10 mL), dried over sodium sulfate, ﬁltered, and concentrated under

reduced pressure to provide N-((1H-pyrrolo[2,3-b]pyridin-3-yl)-

methyl)propan-2-amine (compound 21) that was used for step 2

without additional puriﬁcation. LCMS: m/z [M + H⁺] = 290.2.

Step 2. To a solution of the crude N-((1H-pyrrolo[2,3-b]pyridin-3-

yl)methylpropan-2- amine (100 mg, 0.528 mmol) in dichloromethane

(20 mL) was added 1-isocyanato-4-(triﬂuoromethoxy)benzene (0.08

mL, 0.528 mmol). The mixture was allowed to stir at ambient

temperature for 30 min. The mixture was concentrated under reduced

pressure, dissolved in 2 mL of DMSO, and injected on a Gilson

preparative reversed-phase HPLC instrument (Waters Sun Fire Prep

C18 5 μm, 30 mm × 150 mm column, 5−95% CH₃CN/water + 0.1%

TFA over 15 min at 40 mL/min with a 2 min start collection time).

The desired product fractions were combined, and saturated sodium

carbonate was added until pH 12 (5 mL) was obtained. The aqueous

portion was extracted with EtOAc (3 × 50 mL). The combined

organic layers were washed with water (25 mL) and brine (50 mL),

dried over sodium sulfate, ﬁltered, and concentrated to aﬀord 1-((1H-

pyrrolo[2,3-b]pyridin-3-yl)methyl)-1-isopropyl-3-(4-

1

90%). H NMR (400 MHz, CDCl₃): δ 9.25 (s, 1 H), 8.18 (d, J = 5.8

Hz, 1 H), 7.35 (d, J = 2.1 Hz, 1 H), 7.29 (d, J = 5.9 Hz, 1 H), 3.99 (s, 3

H).

Step 3. Methyl 4-chloro-1H-pyrrolo[3,2-c]pyridine-2-carboxylate

(1.00 g, 4.75 mmol), 3-ethoxyphenylboronic acid (1.03 g, 6.21

mmol), Pd(OAc)₂(0.064 g, 0.285 mmol), XPhos (0.226 g, 0.475

mmol), and potassium ﬂuoride (0.828 g, 14.2 mmol) were combined

in 1,4-dioxane (20 mL). The mixture was degassed (3 × pump/N₂)

then heated to 100 °C. After 4 h the mixture was cooled to room

temperature, diluted with EtOAc, ﬁltered through a pad of Celite

washing with EtOAc, and concentrated. Puriﬁcation by silica gel

chromotography (Biotage-SNAP, 50 g SiO₂, 50% EtOAc/hexanes)

provided methyl 4-(3-ethoxyphenyl)-1H-pyrrolo[3,2-c]pyridine-2-car-

1

boxylate (1.02 g, 72.5%). H NMR (400 MHz, CDCl₃): δ 9.20 (bs, 1

H), 8.51 (d, J = 5.8 Hz, 1 H), 7.54−7.49 (m, 3 H), 7.47−7.38 (m, 1

H), 7.29 (d, J = 5.9 Hz, 1 H), 7.04−7.00 (m, 1 H), 4.14 (q, J = 6.83

Hz, 2 H), 3.97 (s, 3 H), 1.45 (t, J = 7.12 Hz, 3 H).

Step 4. To a solution of methyl 4-(3-ethoxyphenyl)-1H-pyrrolo[3,2-

c]pyridine-2-carboxylate (30 mg, 0.10 mmol) in DMF (0.5 mL) was

added NaH (60% dispersion in mineral oil, 5.5 mg, 0.148 mmol) at

room temperature. After gas evolution had ceased, 2-bromoethyl

phenyl ether (27 mg, 0.13 mmol) was added all at once as a solid.

After the mixture was stirred overnight 2 M NaOH (200 mL, 0.40

mmol) was added, and stirring continued at room temperature. After 4

h the mixture was concentrated. The crude material was taken up in

DMSO and acidiﬁed with TFA. The resulting solution was puriﬁed by

preparative reversed-phase HPLC (30 mm × 100 mm Phenomenex

AXIA-Gemini-NX, 15−40% CH₃CN/water containing 0.1% TFA over

18 min at 50 mL/min) to aﬀord 4-(3-ethoxyphenyl)-1-(2-phenox-

yethyl)-1H-pyrrolo[3,2-c]pyridine-2-carboxylic acid as the TFA salt

(39 mg, 75%). ¹H NMR (500 MHz, DMSO-d₆): δ 8.59 (d, J = 6.6 Hz,

1 H), 8.17 (s, 1 H), 7.60−7.56 (m, 2 H), 7.52 (m, 1 H), 7.48 (m, 1 H),

7.27−7.19 (m, 3H), 6.92−6.87 (m, 1 H), 6.81 (d, J = 8.1 Hz, 2 H),

5.16 (m, 2 H), 4.37 (m, 2 H), 4.16 (q, J = 6.9 Hz, 2 H), 1.38 (t, J = 6.6

Hz, 3 H). HRMS (ESI) calcd (M + H)⁺= 403.1652; found 403.1656.

1-(2-(4-Chlorophenoxy)ethyl)-4-(3-ethoxyphenyl)-1H-

pyrrolo[3,2-c]pyridine-2-carboxylic Acid (23). Step 1. To a

solution of ( )-benzyloxycarbonyl-α-phosphonoglycine trimethyl

ester (12.1 g, 36.5 mmol) in 80 mL of DCM was added DBU (6.34

mL, 42.1 mmol) slowly at room temperature. After 30 min a solution

of 2-chloro-3-formyl-4-iodopyridine (7.50 g, 28.0 mmol) in 20 mL of

DCM was added slowly at room temperature. During the addition the

reaction became slightly exothermic and the mixture was cooled via a

cold water bath. After 4 h the mixture was diluted with DCM and

washed with 1 N HCl (2 × 50 mL) and H₂O (1 × 50 mL). The

organic layer was dried over MgSO₄, ﬁltered, and concentrated.

Puriﬁcation by silica gel chromatography (330 g SiO₂, 20% EtOAc/

hexanes) aﬀorded (Z)-methyl 2-(((benzyloxy)carbonyl)amino)-3-(2-

chloro-4-iodopyridin-3-yl)acrylate (11.4 g, 86%). ¹H NMR (400 MHz,

CDCl₃): δ 7.89 (d, J = 5.2 Hz, 1 H), 7.70 (d, J = 5.2 Hz, 1 H), 7.33 (d,

J = 6.8 Hz, 3 H), 7.25−7.26 (m, 2 H), 7.06 (s, 1 H), 6.96 (bs, 1 H),

4.97 (s, 2 H), 3.91 (s, 3 H).

1

(triﬂuoromethoxy)phenyl)urea (97 mg, 98%). H NMR δ (CDCl₃):

9.35 (s, 1 H), 8.38 (d, J = 4.6 Hz, 1 H), 7.99 (d, J = 7.9 Hz, 1 H), 7.32

(s, 1 H), 7.13−7.16 (m, 3 H), 7.03 (d, J = 8.6 Hz, 2 H), 6.52 (s, 1 H),

4.69−4.76 (m, 1 H), 4.59 (s, 2 H), 1.23−1.28 (d, J = 6.9 Hz, 6 H).

HRMS (ESI): m/z calculated for C₁₉H₁₉F₃N₄O₂: 392.1460. Found:

[M + H]⁺393.1533.

1-(3-(Thiazol-2-yl)benzyl)-3-(4-(triﬂuoromethoxy)phenyl)-

urea (19). To a solution of (3-1,3-thiazol-2-yl)phenyl)methylamine

(170 mg, 0.893 mmol) in hexane (10 mL) and EtOAc (3 mL) was

added 4-(triﬂuoromethoxy)phenyl isocyanate (200 mg, 0.985 mmol).

A white solid precipitated, and the mixture was stirred for 16 h at

ambient temperature. The white precipitate was ﬁltered, washed with

hexanes, and dried to provide 1-(3-(thiazol-2-yl)benzyl)-3-(4-

1

(triﬂuoromethoxy)phenyl)urea. H NMR δ (DMSO-d₆): 8.86 (s, 1

H), 7.92−7.93 (m, 2 H), 7.77−7.82 (m, 2 H), 7.41−7.53 (m, 4 H),

7.23 (d, J = 8.6 Hz, 2 H), 6.81 (t, J = 6.0 Hz, 1 H), 4.39 (d, J = 5.9 Hz,

2 H). LCMS [M + H]⁺= 394.1.

4-(3-Ethoxyphenyl)-1-(2-phenoxyethyl)-1H-pyrrolo[3,2-c]-

pyridine-2-carboxylic Acid (22). Step 1. To a solution of (racemic)-

benzyloxycarbonyl-α-phosphonoglycine trimethyl ester (12.1 g, 36.5

mmol) in 80 mL of dichloromethane was added DBU (6.34 mL, 42.1

mmol) slowly at room temperature. After 30 min a solution of 2-

chloro-3-formyl-4-iodopyridine (7.5 g, 28.0 mmol) in CH₂Cl₂(20

mL) was added slowly at room temperature. During the addition the

reaction became slightly exothermic and the mixture was cooled via a

cold water bath. After 4 h the mixture was diluted with dichloro-

methane and washed with 1 N aqueous HCl (2 × 20 mL) and water

(1 × 30 mL). The organic layer was dried over MgSO₄, ﬁltered, and

concentrated. Puriﬁcation by silica gel chromatography (ISCO Redi-

Sep, 330 g SiO₂, 20% EtOAc/hexanes) aﬀorded methyl (2Z)-2-

Step 2. (Z)-Methyl 2-(((benzyloxy)carbonyl)amino)-3-(2-chloro-4-

iodopyridin-3-yl)acrylate (11.46 g, 24.25 mmol), K₂CO₃(10.05 g, 72.7

mmol), copper(I) iodide (0.462 g, 2.425 mmol), and ^L-proline (0.558

N

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

g, 4.85 mmol) were combined in 1,4-dioxane (120 mL). The mixture

was degassed (3 × pump/N₂) and then heated to reﬂux. After

reﬂuxing overnight, the mixture was cooled to room temperature,

diluted with saturated aqueous NH₄Cl, and extracted with dichloro-

methane (3 × 100 mL). The combined organic layers were dried over

MgSO₄, ﬁltered through a pad of Celite, and concentrated. The

resulting solid was triturated with Et₂O/hexanes to provide methyl 4-

chloro-1H-pyrrolo[3,2-c]pyridine-2-carboxylate (4.6 g, 90%). ¹H NMR

(400 MHz, CDCl₃): δ 9.25 (bs, 1 H), 8.18 (d, J = 5.8 Hz, 1 H), 7.35

(d, J = 2.1 Hz, 1 H), 7.29 (d, J = 5.9 Hz, 1 H), 3.99 (s, 3 H).

Step 3. Methyl 4-chloro-1H-pyrrolo[3,2-c]pyridine-2-carboxylate

(1.00 g, 4.75 mmol), 3-ethoxyphenylboronic acid (1.03 g, 6.21

mmol), Pd(OAc)₂(0.064 g, 0.285 mmol), XPhos (0.226 g, 0.475

mmol), and KF (0.828 g, 14.24 mmol) were combined in 1,4-dioxane

(20 mL). The mixture was degassed (3 × pump/N₂) and then heated

to 100 °C. After 4 h the mixture was cooled to room temperature,

diluted with EtOAc, ﬁltered through a pad of Celite washing with

EtOAc, and concentrated. Puriﬁcation by silica gel chromatography

(50 g SiO₂, 50% EtOAc/hexanes) aﬀorded methyl 4-(3-ethoxyphen-

Phenomenex AXIA-Gemini-NX, 20−45% CH₃CN/water containing

0.1% THF over 18 min at 20 mL/min) to aﬀord 4-(3-ethoxyphenyl)1-

{2-[4-(triﬂuoromethoxy)phenoxy]ethyl}-1H-pyrrolo[3,2-c]pyridine-2-

carboxylic acid as the TFA salt (5 mg, 15%). ¹H NMR (500

MHz,DMSO-d₆): δ 8.58 (d, J = 6.6 Hz, 1 H), 8.16 (bs, 1 H), 7.58 (m,

2 H), 7.52 (d, J = 7.7 Hz, 1 H), 7.47 (m, 1 H), 7.23 (m, 3H), 6.91 (d, J

= 8.8 Hz, 2 H), 5.16 (m, 2 H), 4.40 (m, 2 H), 4.15 (q, J = 6.7 Hz, 2

H), 1.38 (t, J = 7.0 Hz, 3H). HRMS (ESI) calcd (M + H) = 487.1475,

found 487.1482.

1-(2-[4-(Triﬂuoromethoxy)phenoxy)ethyl)-1H-pyrrolo[3,2-c]-

pyridine-2-carboxylic Acid (25). Methyl 4-chloro-1-{2-[4-

(triﬂuoromethoxy)phenoxy]ethyl}-1H-pyrrolo[3,2-c]pyridine-2-car-

boxylate (25 mg, 0.060 mmol) was hydrogenated (balloon) with 10%

Pd/C (10 mg, 9.40 mmol) in MeOH (3 mL) at room temperature.

After 4 h the mixture was ﬁltered using a 0.45 μm PTFE syringe ﬁlter

and then concentrated. The residue was taken up in 0.5 mL of MeOH.

To this was added 2 M NaOH (0.15 mL, 0.300 mmol), and then the

mixture was heated to 60 °C. After being stirred overnight the mixture

was cooled to room temperature and concentrated. The crude material

was taken up in DMSO and acidiﬁed with TFA. The resulting solution

was ﬁltered using a 0.45 μm PTFE syringe ﬁlter and then puriﬁed by

preparative reversed-phase HPLC (21 mm × 100 mm Phenomenex

AXIA-Gemini-NX, 20−45% CH₃CN/water containing 0.1% TFA over

18 min at 20 mL/min) to aﬀord 1-(2-[4-(triﬂuoromethoxy)phenoxy)-

ethyl)-1H-pyrrolo[3,2-c]pyridine-2-carboxylic acid as the TFA salt (22

mg, 75%). ¹H NMR (500 MHz, DMSO-d₆): δ 9.36 (s, 1 H), 8.58 (d, J

= 6.7 Hz, 1 H), 8.26 (d, J = 6.8 Hz, 1 H), 7.70 (s, 1 H), 7.22 (d, J = 8.5

Hz, 2H), 6.87 (d, J = 8.6 Hz, 2 H), 5.14 (m, 2 H), 4.38 (m, 2 H).

HRMS (ESI) calcd (M + H)⁺= 367.0900, found 367.0907.

1

yl)-1H-pyrrolo[3,2-c]pyridine-2-carboxylate (1.02 g, 72%). H NMR

(400 MHz, CDCl₃): δ 9.20 (s, 1 H), 8.51 (d, J = 5.8 Hz, 1 H), 7.50−

7.53 (m, 3 H), 7.43 (t, J = 7.9 Hz, 1 H), 7.29 (d, J = 5.9 Hz, 1 H), 7.02

(dd, J = 8.2, 2.5 Hz, 1 H), 4.14 (q, J = 7.0 Hz, 2 H), 3.97 (s, 3 H), 1.45

(t, J = 7.0 Hz, 3 H).

Step 4. To a solution of methyl 4-(3-ethoxyphenyl)-1H-pyrrolo[3,2-

c]pyridine-2-carboxylate (30 mg, 0.10 mmol) in 0.5 mL of THF was

added NaH (60% in mineral oil, 6.0 mg, 0.15 mmol) at room

temperature. After gas evolution had ceased 1-(2-bromoethoxy)-4-

chlorobenzene (31 mg, 0.13 mmol) was added and the mixture heated

to 60 °C. After stirring overnight 0.16 mL of 2 M NaOH was added

and heating continued. After 4 h the mixture was cooled to room

temperature and concentrated. The crude material was taken up in

DMSO (1.0 mL), ﬁltered using a 0.45 μm PTFE syringe ﬁlter, then

puriﬁed by preparative reversed-phase HPLC (CH₃CN/H₂O with

0.1% TFA modiﬁer). Fractions containing pure product were

concentrated to aﬀord 1-(2-(4-chlorophenoxy)ethyl)-4-(3-ethoxy-

phenyl)-1H-pyrrolo[3,2-c]pyridine-2-carboxylic acid (9.7 mg, 17%)

as a TFA salt. ¹H NMR (500 MHz, DMSO-d₆): δ 8.51 (d, J = 6.2 Hz,

1 H), 7.88 (bs, 1 H), 7.51 (m, 3 H), 7.46 (s, 1 H), 7.27 (d, J = 8.7 Hz,

2 H), 7.14 (bs, 1 H), 6.84 (d, J = 8.7 Hz, 2 H), 5.07 (m, 2 H), 4.35 (m,

2 H), 4.13 (q, J = 7.0 Hz, 2 H), 1.38 (t, J = 6.9 Hz, 3 H). HRMS

C₂₄H₂₁ClN₂O₄(M + H)⁺calculated, 437.1263; found, 437.1262.

4-(3-Ethoxyphenyl)1-{2-[4-(triﬂuoromethoxy)phenoxy]-

ethyl}-1H-pyrrolo[3,2-c]pyridine-2-carboxylic Acid (24). Step 1.

To a solution of methyl-4-(3-ethoxyphenyl)-1H-pyrrolo[3,2-c]-

pyridine-2-carboxylate (230 mg, 0.776 mmol) in DMF (7 mL) was

added NaH (60% dispersion in mineral oil, 40 mg, 1.00 mmol) at

room temperature. After gas evolution had ceased 1,2-dibromoethane

(1 mL, 11.60 mmol) was added rapidly. After being stirred overnight,

the mixture was diluted with water and extracted with EtOAc (3 × 30

mL). The combined organic layers were washed with water and brine

and then dried over MgSO₄, ﬁltered, and concentrated. Puriﬁcation by

silica gel chromatography (Biotage-SNAP, 25 g SiO₂, 0−50% EtOAc/

hexanes) aﬀorded methyl-1-2(bromoethyl)-4-(3-ethoxy-1H-pyrrolo-

5-Cyano-1-{2-[4-(triﬂuoromethoxy)phenoxylethyl}-1H-in-

dole-2-carboxylic Acid (26). To a solution ethyl 5-cyano-1H-indole-

2-carboxylate (37 mg, 0.17 mmol), 2-[4-(triﬂuoromethoxy)phenoxy]-

ethanol (42 mg, 0.19 mmol), and triphenylphosphine (227 mg, 0.518

mmol) in toluene (1 mL) was add DIAD (0.044 mL, 0.225 mmol)

dropwise at room temperature, and the mixture was stirred overnight.

The mixture was diluted with EtOAc and quenched with water. The

two layers were separated, and the aqueous phase was further extracted

twice with EtOAc. The combined organic layers were washed with

brine, dried over Na₂SO₄, ﬁltered, and concentrated in vacuo.

Puriﬁcation by silica gel chromatography (0−30% EtOAc in hexanes)

provided ethyl-5-cyano-l-{2-[4-(triﬂuoromethoxy)phenoxy]ethyl}-1H-

indole-2-carboxylate (15 mg, 0.04 mmol) that was subsequently

dissolved in THF (1 mL), MeOH (300 μL), and 1 M NaOH (200

μL). The reaction mixture was stirred at room temperature for 5 h,

after which water was added, the pH was adjusted to 3 with 1 N HCl,

and the solvent was removed in vacuo. EtOAc was added to the

resulting aqueous solution. The two layers were separated, and the

aqueous phase was further extracted with EtOAc (2 × 10 mL). The

combined organic layers were washed with brine, dried over Na₂SO₄,

ﬁltered, and concentrated in vacuo to aﬀord 5-cyano-1-{2-[4-

(triﬂuoromethoxy)phenoxylethyl}-1H-indole-2-carboxylic acid (13.4

1

mg, 20%). H NMR (400 MHz, CDCl₃): δ 8.03 (s, 1H), 7.66 (d, J

= 8.8 HZ, 1H), 7.57 (d, J = 8.8 Hz, 1H), 7.39 (s, 1H), 7.07 (d, J = 8.8

Hz, 2H), 6.73 (d, J = 9.2 Hz, 2H), 4.97 (t, J = 5.1 Hz, 2H), 4.46−4.34

(m, 4H), 1.43 (t, J = 7.1 Hz, 3H). HRMS: calcd for C_l9H₁₃F₃N₂O₄

413.0720. Found: 413.0718.

1

[3,2-c]pyridine-2-carboxylate (135 mg, 43%). H NMR (400 MHz,

CDCl₃): δ 8.51 (d, J = 6.0 Hz, 1 H), 7.60 (d, J = 1.0 Hz, 1 H), 7.49−

7.45 (m, 2H), 7.40 (m, J = 7.9 Hz, 1 H), 7.31 (dd, J = 6.0, 1.0 Hz, 1

H), 6.99 (ddd, J = 8.2, 2.6, 1.1 Hz, 1 H), 4.92 (t, J = 6.8 Hz, 2 H), 4.12

(q, J = 7.0 Hz, 2 H), 3.91 (s, 3 H), 3.71 (t, J = 7.1 Hz, 2 H), 1.43 (t, J =

7.12 Hz, 2 H).

Step 2. 4-(Triﬂuoromethoxy)phenol (15.4 mg, 0.086 mmol) and

K₂CO₃(24 mg, 0.17 mmol) were combined in a screw cap vial. To this

was added a solution of methyl-1-(2-bromoethyl)-4-(3-ethoxyphenyl)-

1H-pyrrolo[3,2-c]pyridine-2-carboxylate (22 mg, 0.056 mmol) in

DMF (0.5 mL). The vial was capped and then heated to 60 °C.

After being stirred overnight, the mixture was cooled to room

temperature. Then 2 N NaOH (100 mL) was added. After 4 h at room

temperature the mixture was concentrated. The crude material was

taken up in DMSO and acidiﬁed with TFA. The resulting solution was

puriﬁed by preparative reversed-phase HPLC (21 mm × 100 mm

5-(1H-Pyrazol-3-yl)-1-(2-(4-(triﬂuoromethoxy)phenoxy)-

ethyl)-1H-indole-2-carboxylic Acid (27). Step 1. To a solution of

ethyl 5-bromo-1H-indole-2-carboxylate (1.7 g, 6.3 mmol), 2-[4-

(triﬂuoromethoxy)phenoxy]ethanol (1.7 g, 7.6 mmol), and triphenyl-

phosphine (3.3 g, 12.7 mmol) in toluene (20 mL) was added DIAD

(1.85 mL, 9.5 mmol) dropwise at 0 °C. The mixture was left to warm

to room temperature and stirred overnight. The mixture was then

concentrated in vacuo and puriﬁed by silica gel chromatography (0−

20% EtOAc in hexanes) to aﬀord ethyl 5-bromo-1-{2-[4-

(triﬂuoromethoxy)phenoxy]ethyl}-1H-indole-2-carboxylate (1.6 g,

1

52%). H NMR (400 MHz, CDCl₃): δ 7.79 (s, 1H), 7.45 (s, 1H),

7.25 (m, 2H), 7.01 (d, J = 9.2 Hz, 2H), 6.75 (m, 2H), 4.92 (t, J = 5.6

Hz, 2H), 4.37 (q, J = 6.8 Hz, 2H), 4.32 (t, J = 5.6 Hz, 2H), 1.41 (t, J =

O

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

6.8 Hz, 3H). LRMS (M + H): calcd for C₂₀H₁₈BrF₃NO₄472.1/474.1.

Found: 472.1/474.1.

eﬃciency; SPR, surface plasmon resonance; CFA, complete

Freund’s adjuvant

Step 2. Ethyl 5-bromo-1-{2-[4-(triﬂuoromethoxy)phenoxy]ethyl}-

1H-indole-2-carboxylate (812 mg, 1.7 mmol), pinacol borane (873 mg,

3.4 mmol), XPhos (98 mg, 0.21 mmol), and KOAc (540 mg, 5.5

mmol) were combined in dioxane (8 mL). The mixture was degassed

under vacuum and back-ﬁlled with N₂. Palladium acetate (23 mg, 0.1

mmol) was added, and the mixture was degassed and back-ﬁlled again.

The mixture was left to stir overnight at 85 °C. EtOAc was added, and

the mixture was ﬁltered through a pad of Celite, concentrated in vacuo,

and puriﬁed by silica chromatography (0−30% EtOAc in hexanes) to

provide ethyl 5-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1-{2-[4-

(triﬂuoromethoxy)phenoxy]ethyl}-1H-indole-2-carboxylate (675 mg,

76%). ¹H NMR (400 MHz, CDCl₃): δ 8.19 (s, 1H), 7.78 (dd, J = 8.4

Hz, 1.2 Hz, 1H), 7.75 (d, J = 8.8 Hz, 1H), 7.33 (s, 1H), 7.05 (d, J = 8.8

Hz, 2H), 6.77 (m, 2H), 4.95 (t, J = 5.6 Hz, 2H), 4.38 (q, J = 7.2 Hz,

2H), 4.32 (t, J = 5.6 Hz, 2H), 1.42 (t, J = 7.2 Hz, 3H). LRMS (M +

H): calcd for C₂₆H₃₀BF₃NO₆520.3. Found: 520.3.

REFERENCES

■

(1) Manning, G.; Whyte, D. B.; Martinez, R.; Hunter, T.;

Sudarsanam, S. The Protein Kinase Compliment of the Human

Genome. Science 2002, 298, 1912−1934.

(2) Reichardt, L. F. Neurotrophin-Regulated Signalling Pathways.

Philos. Trans. R. Soc. B 2006, 361, 1545−1564.

(3) Watson, J. L.; Allen, S. J.; Dawbarn, D. Targeting Nerve Growth

Factor in Pain. BioDrugs 2008, 22, 349−359.

(4) Wang, T.; Yu, D.; Lamb, M. L. Trk Kinase Inhibitors as New

Treatments for Cancer Pain. Expert. Opin. Ther. Pat. 2009, 19, 305−

319.

(5) Lane, N. E.; Schnitzer, T. J.; Birbara, C. A.; Mokhtarani, M.;

Shelton, D. L.; Smith, M. D.; Brown, M. T. Tanezumab for the

Treatment of Pain from Osteoarthritis of the Knee. N. Engl. J. Med.

2010, 363, 1521−1531.

(6) Invitrogen 92 member kinase panel used as the general

counterscreen at 1 μM to assess kinase selectivity.

(7) Protein crystallography conditions and procedured are described

in the Supporting Information.

(8) Structure coordinates have been deposited in the PDB: code

4PMT.

Step 3. Ethyl 5-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1-{2-

[4-(triﬂuoromethoxy)phenoxy]ethyl}-1H-indole-2-carboxylate (75

mg, 0.14 mmol), 3-bromo-1-(phenylsulfonyl)-1H-pyrazole (83 mg,

0.29 mmol), K₂CO₃(60 mg, 0.43 mmol), and SPhos (6 mg, 0.014

mmol) were combined in dioxane (2 mL). The reaction was degassed

and ﬁlled with nitrogen 2×. Water (0.4 mL) was added and purged

again with nitrogen. Palladium acetate (1.6 mg, 0.007 mmol) was

added under nitrogen, and the mixture was heated to 80 °C overnight.

EtOAc was added and the crude mixture was ﬁltered through a pad of

Celite, concentrated in vacuo, and puriﬁed by silica gel chromatog-

raphy (0−40% EtOAc in hexanes) to provide ethyl 5-(1-(phenyl-

sulfonyl)-1H-pyrazol-3-yl)-1-(2-(4-(triﬂuoromethoxy)phenoxy)ethyl)-

1H-indole-2-carboxylate (50 mg, 58%). This material was then

dissolved in THF (1 mL) and MeOH (1 mL), and NaOH (1 M,

0.5 mL) was added. The reaction mixture was then heated to 80 °C

overnight. The solvent was removed in vacuo and the mixture was

dissolved in DMSO, puriﬁed by reverse phase HPLC (10−95% ACN

in H₂O (with 0.1% TFA)) to aﬀord 5-(1H-pyrazol-3-yl)-1-(2-(4-

(triﬂuoromethoxy)phenoxy)ethyl)-1H-indole-2-carboxylic acid (22

(9) Additional hinge binding TrkA inhibitors have also been

previously reported in the literature; see the following: (a) Wang,

T.; Lamb, M. L.; Scott, D. A.; Wang, H.; Block, M. H.; Lyne, P. D.;

Lee, J. W.; Davies, A. M.; Zhang, H.-J.; Zhu, Y.; Gu, F.; Han, Y.; Wang,

B.; Mohr, P. J.; Kaus, R. J.; Josey, J. A.; Hoffmann, E.; Thress, K.;

MacIntyre, T.; Wang, H.; Omer, C. A.; Yu, D. Identification of 4-

Aminopyrazolylpyrimidines as Potent Inhibitors of Trk Kinases. J.

Med. Chem. 2008, 51, 4672−4684. (b) Hong, S.; Kim, J.; Seo, J. H.;

Jung, K. H.; Hong, S.-S.; Hong, S. Design, Synthesis and Evaluation of

3,5-Disubstituted 7-Azaindoles as Trk Inhibitors with Anticancer and

Antiangiogenic Activities. J. Med. Chem. 2012, 55, 5337−5349.

(10) Van Linden, O. P. J.; Kooistra, A. J.; Leurs, R.; de Esch, I. J. P.;

de Graff, C. KLIFS: A Knowledge-Based Structural Database To

Navigate Kinase−Ligand Interaction Space. J. Med. Chem. 2014, 57,

249−277.

1

mg, 28%). H NMR (500 MHz, DMSO-d₆): δ 8.07 (s, 1H), 7.77

(d, J = 9.0 Hz, 1H), 7.70 (m, 2H), 7.27 (s, 1H), 7.22 (d, J = 9.0 Hz,

2H), 6.92 (d, J = 9.0 Hz, 2H), 6.69 (d, J = 2 Hz, 1H), 4.98 (t, J = 5.5

Hz, 2H), 4.32 (t, J = 5.5 Hz, 2H). HRMS: calcd for C₂₁H₁₇F₃N₃O₄

432.1166. Found: 432.1183.

(11) The cell-based assay is performed using the Discover X-

Pathhunter assay platform. Experimental details are described in the

Supporting Information.

ASSOCIATED CONTENT

■

(12) Bouhana, K. S.; Impastato, R.; Pheneger, J.; Jiang, Y.; Wallace, R.

D.; Do, M. G.; Zaute, N. A.; Andrews, S. W. Analgesic Eﬀects of a

Potent and Selective Kinase Inhibitor of Neurotrophin Receptors

TRKA, TRKB and TRKC. WO2010033941, 2010; WO2010048312,

2010; WO2011006074, 2011; www.arraybiopharma.com.

(13) Green, A.; Li, Y.; Stachel, S. J. Preparation of Pyridotriazole and

Benzotriazole Compounds as TrkA Kinase Inhibitors. PCT Int. Appl.

WO 2012125667 A1, 2012.

(14) Green, A.; Li, Y.; Stachel, S. J. TrkA kinase Inhibitors

Compositions and Methods of Use PCT Int. Appl. WO2012125668

A1, 2012.

(15) Structure coordinates have been deposited with the PDB: code

4PMM.

(16) For an excellent review on type I and type II kinase inhibitors

see the following: Zuccotto, F.; Ardini, E.; Casale, E.; Angiolini, M.

Through the “Gatekeeper Door”: Exploiting the Active Kinase

Conformation. J. Med. Chem. 2010, 53, 2681−2694.

(17) Huse, M.; Kuriyan, J. The Conformational Plasticity of Protein

Kinases. Cell 2002, 109, 275−282.

(18) Albaugh, P.; Fan, Y.; Mi, Y.; Sun, F.; Adrian, F.; Li, N.; Jia, Y.;

Sarkisova, Y.; Kreusch, A.; Hood, T.; Lu, M.; Liu, G.; Huang, S.; Liu,

Z.; Loren, J.; Tuntland, T.; Karanewsky, D. S.; Seidel, H. M.; Molteni,

V. Discovery of GNF-5837, a Selective TRK Inhibitor with Efficacy in

Rodent Cancer Tumor Models. ACS Med. Chem. Lett. 2012, 3, 140−

145.

S

* Supporting Information

Experimental procedures for protein crystallography, SPR, and

details of in vitro and in vivo assays. This material is available

free of charge via the Internet at http://pubs.acs.org.

Accession Codes

Structure coordinates have been deposited in the PDB: codes

4PMT, 4PMP, 4PMS, 4PMM.

AUTHOR INFORMATION

■

Corresponding Author

*Phone: 215-652-2273. Fax: 215-652-7310. E-mail: shawn_

stachel@merck.com.

Notes

The authors declare no competing ﬁnancial interest.

ABBREVIATIONS USED

■

Trk, tropomyosin-related kinase; ATP, adenosine 5′-triphos-

phate; NGF, nerve growth factor; BDNF, brain-derived

neurotrophic factor; NT-3, neurotrophin 3; HTS, high-

throughput screening; IGFR1, insulin-like growth factor

receptor 1; LE, ligand eﬃciency; LLE, lipophilic ligand

P

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(19) Schindler, T.; Bornmann, W.; Pellicena, P.; Miller, W. T.;

Clarkson, B.; Kuriyan, J. Structural Mechanism for STI-571 Inhibition

of Abelson Tyrosine Kinase. Science 2000, 289, 1938−1942.

(20) Stachel, S. J.; Egbertson, M.; Brnardic, E.; Jones, K.; Sanders, J.

M.; Henze, D. A. TrkA Kinase Inhibitors, Composition and Methods

Thereof. PCT Int. Appl. WO 2013/176970, Nov 2013.

Established Tactile Allodynia in Rodent Models of Neuropathic Pain

without Tolerance. J. Pharmacol. Exp. Ther. 2007, 322, 282−287.

(21) Ghose, A. K.; Herbertz, T.; Pippin, D. A.; Salvino, J. M.;

Mallamo, J. P. Knowledge Based Predictions of Ligand Binding Modes

and Rational Inhibitor Design for Kinase Drug Discovery. J. Med.

Chem. 2008, 51, 5149−5171.

(22) Dar, A. C.; Lopez, M. S.; Shokat, K. M. Small Molecule

Recognition of c-Src via the Imantinib-Binding Conformation. Chem.

Biol. 2008, 15, 1015−1022.

(23) Bhat, R.; Xue, Y.; Berg, S.; Hellberg, S.; Ormo, M.; Nilsson, Y.;

Radesater, A.-C.; Jerning, E.; Markgren, P.-O.; Borgegard, T.; Nylof,

M.; Gimenez-Cassina, A.; Hernandez, F.; Lucas, J. J.; Diaz-Nido, J.;

Avila, J. Structural Insights and Biological Effects of Glycogen Synthase

Kinase 3-Specific Inhibitor AR-A014418. J. Biol. Chem. 2003, 278,

45937−45945.

(24) Boije af Gennas, G.; Mologni, L.; Ahmed, S.; Rajaratnam, M.;

Marin, O.; Lindholm, N.; Viltadi, M.; Gambacorti-Passerini, C.;

Scappozza, L.; Yli-Kauhaluoma, J. Design, Synthesis, and Biological

Activity of Urea Derivatives as Anaplastic Lymphoma Kinase

Inhibitors. ChemMedChem 2011, 6, 1680−1692.

(25) Structure coordinates have been deposited in the PDB: code

4PMP.

(26) (a) Behshad, E.; Klabe, R. M.; Margulis, A.; Becker-Pasha, M.;

Rupar, M. J.; Collier, P.; Liu, P. C.; Hollis, G. F.; Burn, T. C.; Wynn, R.

Phosphorylation State-Dependent High Throughput Screening of the

c-Met Kinase. Curr. Chem. Genomics 2010, 4, 27−33. (b) Wang, W.;

Marimuthi, A.; Tsai, J.; Kumar, A.; Krupka, H. I.; Zhang, C.; Powell,

B.; Suzuki, Y.; Nguyen, H.; Tabrizizad, M.; Luu, C.; West, B. L.

Structural Characterization of Autoinhibited c-Met Kinase Produced

by Coexpression in Bacteria with Phosphatase. Proc. Natl. Acad. Sci.

U.S.A. 2006, 103, 3563−3568.

(27) Whelligan, D. K.; Solanki, S.; Taylor, D.; Thomson, D. W.;

Cheung, K.-M. J.; Boxall, K.; Mas-Droux, C.; Barillari, C.; Burns, S.;

Grummitt, C. G.; Collins, I.; van Montfort, R. L. M.; Aherne, G. W.;

Bayliss, R.; Hoelder, S. Aminopyrazine Inhibitors Binding to Unusual

Inactive Conformation of the Mitotic Kinase Nek2: SAR and

Structural Characterization. J. Med. Chem. 2010, 53, 7682−7698.

(28) Structure coordinates have been deposited in the PDB: code

4PMS.

(29) Hanney, B.; Manley, P.; Rudd, M. T.; Sanders J. M.; Stachel, S.

J.; Henze, D. TrkA Kinase Inhibitors, Compositions and Methods

Thereof. PCT. Int. Appl. WO2013009582 A1, 2013.

(30) Regan, J.; Pargellis, C. A.; Cirillo, P. F.; Gilmore, T.; Hickey, E.

R.; Peet, G. W.; Proto, A.; Swinamer, A.; Moss, N. The Kinetics of

Binding to p38 MAP Kinase by Analogues of BIRB 796. Bioorg. Med.

Chem. Lett. 2003, 13, 3101−3104.

(31) Pargellis, C.; Tong, L.; Churchill, L.; Cirillo, P. F.; Gilmore, T.;

Graham, A. G.; Grob, P. M.; Hickey, E. R.; Moss, N.; Pav, S.; Regan, J.

Inhibition of p38 MAP Kinase by Utilizing a Novel Allosteric Binding

Site. Nat. Struct. Biol. 2002, 9, 268−272.

(32) Array BioPharma recently reported on the use of a small

molecule pan-Trk inhibitor in models of bone cancer and skeletal

fracture pain: (a) Ghilardi, J. R.; Freeman, K. T.; Jimenez-Andrade, J.

M.; Mantyh, W. G.; Bloom, A. P.; Kuskowski, M. A.; Mantyh, P. W.

Administration of a Tropomysin Receptor Kinase Inhibitor Attenuates

Sarcoma-Induced Nerve Sprouting, Neuroma Formation and Bone

Cancer Pain. Mol. Pain 2010, 6, 87−97. (b) Ghilardi, J. R.; Freeman,

K. T.; Jimenez-Andrade, J. M.; Mantyh, W. G.; Bloom, A. P.; Bouhana,

K. S.; Trollinger, D.; Winkler, J.; Lee, P.; Andrews, S. W.; Kuskowski,

M. A.; Mantyh, P. W. Sustained Blockage of Neurotrophin Receptors

TrkA, TrkB, and TrkC Reduces Non-Malignant Skeletal Pain but Not

the Maintenance of Sensory and Sympathetic Nerve Fibers. Bone

2011, 28, 389−398.

(33) Wild, K. D.; Bian, D.; Zhu, D.; Davis, J.; Bannon, A. W.; Zhang,

T. J.; Louis, J.-C. Antibodies to Nerve Growth Factor Reverse

Q

dx.doi.org/10.1021/jm5006429 | J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

Maximizing diversity from a kinase screen: Identification of novel and selective pan-Trk inhibitors for chronic pain

DOI: 10.1021/jm5006429

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm5006429

Full text of DOI:10.1021/jm5006429

Products guided by the article

R&D Labs maybe for 1614229-09-2

Relevant to this article

Hot Product