Discovery and optimization of a novel series of dyrk1b kinase inhibitors to explore a MEK resistance hypothesis

Article abstract of DOI:10.1021/acs.jmedchem.5b00098

Potent and selective inhibitors of Dyrk1B kinase were developed to explore the hypothesis, based on siRNA studies, that Dyrk1B may be a resistance mechanism in cells undergoing a stress response.

Full text of DOI:10.1021/acs.jmedchem.5b00098

Article
pubs.acs.org/jmc
Discovery and Optimization of a Novel Series of Dyrk1B Kinase
Inhibitors To Explore a MEK Resistance Hypothesis
Jason G. Kettle,* Peter Ballard, Catherine Bardelle, Mark Cockerill,† Nicola Colclough,
Susan E. Critchlow, Judit Debreczeni, Gary Fairley, Shaun Fillery, Mark A. Graham, Louise Goodwin,
Sylvie Guichard, Kevin Hudson, Richard A. Ward, and David Whittaker
Oncology iMed, AstraZeneca, Alderley Park, Macclesfield, SK10 4TG, United Kingdom
S
* Supporting Information
ABSTRACT: Potent and selective inhibitors of Dyrk1B kinase were developed to explore the hypothesis, based on siRNA
studies, that Dyrk1B may be a resistance mechanism in cells undergoing a stress response.
pathway under serum-free conditions by treatment with a MEK
inhibitor caused upregulation of Dyrk1B by 18−20-fold.1a In-
house studies confirmed upregulation of Dyrk1B mRNA and
protein levels when cells are treated with MEK inhibitors,
consistent with the hypothesis that Dyrk1B may be a resistance
mechanism in cells undergoing a stress response. After
knockdown of Dyrk1B, A375 cells were treated with MEK
inhibitors AZD62449 or AZD833010 for 48 h. Total levels of
Dyrk1B protein were analyzed by western blot (Figure 1),
showing that MEK inhibition causes significant upregulation of
Dyrk1B and that this is reduced by prior treatment with siRNA
to Dyrk1B. Approximately 60% knockdown of Dyrk1B protein
was observed in these experiments. The resulting viability of
cells was also evaluated, and it can be seen that treatment of
A375 cells with either Dyrk1B siRNA or MEK inhibitor
independently results in no appreciable effects on viability;
however, the combination of the two results in a significant
enhancement of cell death (Figure 2).
Despite this interesting biology, there remains a lack of
highly potent and selective tool molecules with which to
elucidate the role of Dyrk family kinases in a disease setting,
particularly with regard to recapitulation of siRNA knockdown
studies, and also ultimately as in vivo probes.11 The beta-
carboline alkaloid harmine 1 (Figure 3) has been most studied
in the context of Dyrk1A inhibition in a neuritogenesis setting,
although this tool compound may be more accurately described
as a mixed Dyrk1A and 1B inhibitor, with reported enzyme
IC50’s of 33 and 166 nM, respectively. It has much lower
potency against Dyrk2 (2 μM) and Dyrk4 (74 μM).12 Despite
INTRODUCTION
■
Dyrk1B (also known as Mirk or Minibrain-related kinase) is a
dual-specificity serine−threonine kinase that has been shown to
mediate survival and differentiation in skeletal muscle, in which
it is expressed at higher levels relative to that in other normal
tissues.1 It is a member of the Minibrain/Dyrk family of protein
kinases that includes Dyrk1A, Dyrk1C, Dyrk2, Dyrk3, Dyrk4A,
and Dyrk4B, and, unusually, this family retains the ability to
autophosphorylate at a conserved tyrosine residue on the
activation loop.2 Dyrk1B is expressed in several different tumor
types and appears to play a key role in maintaining cells in
quiescence.3 Amplification of Dyrk1B expression has been
observed in a large subset of ovarian cancers when compared to
normal tissue, and depletion of Dyrk1B levels sensitizes these
lines to cisplatin treatment, possibly through elevated reactive
oxygen species content.4 Dyrk1B was also found to be highly
expressed in osteosarcoma cell lines, and shRNA-mediated
target knockdown led to a decrease in proliferation and
induction of apoptosis across a range of tumor lines.
Osteosarcoma tumors display high metastatic potential, and
the relative level of Dyrk1B expression has been correlated with
poor outcome in such patients.5 Additionally, a recent study has
implicated a mutation in Dyrk1B kinase with increased
adipogenesis and glucose homeostasis in an inherited form of
metabolic syndrome.6 Inhibition of Dyrk1A, by contrast, has
been investigated as a strategy for therapeutic intervention in
Down’s syndrome,7 where overexpression of the Dyrk1A gene
on chromosome 21 has been linked to the impaired neuronal
development associated with this genetic disorder through
studies in rodents.8
Our interest in this kinase was linked to the observation that
Dyrk1B is regulated by the kinase Erk and that inhibition of this
Received: January 19, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. MEK inhibitor treatment of A375 melanoma cells results in significant upregulation of Dyrk1b, which is reduced by siRNA targeting
Dyrk1b. Western blot is representative of at least two separate experiments.
subsequent step, which, in the case of 2,4-dichloropyrimidine,
occurred selectively at C-4 to give pyrimidine 6. N-Methyl
azaindole analogues were prepared from 6 in a sequence
involving deprotection to give 7 and methylation to give
chloropyrimidine 8. Introduction of the 2-anilino moiety was
realized under mild acid catalysis to give 19. Styrene analogue
31 was obtained directly from 6 via Suzuki cross-coupling and
deprotection of the azaindole in situ. A modified route was
adopted for synthesis of 2-methylpyrimidine analogues 32 and
33 (Scheme 2). Here, azaindole 2 was first acetylated to give 3.
Protection of the azaindole nitrogen to give 10 was followed by
a short sequence to build up the pyrimidine ring directly.
Reaction of 10 with DMF acetal to give 11 was followed by
condensation with acetimidamide and resultant loss of the
protecting group to yield 2-methylpyrimidine 32. Alkylation of
the azaindole nitrogen of 32 was achieved under Mitsunobu
conditions to give benzyl analogue 33.
Figure 2. siRNA-mediated knockdown of DYRK1b increases cell
death in AZD6244-treated A375 cells. The percentage of dead cells
plotted is the mean and SEM from four separate experiments.
Statistically significant differences, compared with control (scrambled)
siRNA, are indicated.
6-AZAINDOLE ANILINOPYRIMIDINES AS
INHIBITORS OF DYRK1B
■
Kinases Dyrk1A, Dyrk2, and Dyrk3 were available to us for
screening through our collaboration with the MRC Protein
Phosphorylation Unit at the University of Dundee together
with a range of other serine−threonine kinases. This panel of
enzymes was utilized by discovery projects to assess broader
kinase selectivity as part of the lead optimization process.
Consequently, there existed a significant amount of data against
these isoforms for current and past project compounds.
Compound 18, initially synthesized as part of an internal
program to target IGF, had been profiled in this panel and
showed significant inhibition at 1 μM (90, 92, and 83%
inhibition of Dyrk1A, 2, and 3, respectively), albeit with only a
moderately selective profile. Of the 71 kinases tested in this
panel, 18 showed >75% inhibition against 11 kinases (and
>50% inhibition against 27/71). Internal testing confirmed 18
to be a highly potent inhibitor of Dyrk1B with an IC50 of 4 nM.
This initial lead compound was followed up with screening of
further analogues, with the aim of improving the Dyrk1B
potency and overall kinase selectivity and delivering tool
molecules suitable for exploring the role of Dyrk1B inhibition
in vivo.
Figure 3. Literature reported inhibitor of Dyrk1A and Dyrk1B,
harmine.
its generally good selectivity across other kinases,13 the utility of
harmine as a probe compound, particularly in vivo, may be
limited by other off-target pharmacologies, which include
cytotoxicity,14 PPARgamma induction,15 and potent inhibition
of monoamine oxidase A.16 Herein, we report our studies
aimed at generating potent and selective small molecule
inhibitors of the Dyrk family enzymes, in particular Dyrk1B,
for use as in vivo probes to further delineate the role of these
targets in disease.
CHEMISTRY
■
Azaindole 2 was brominated at C-3 to give compound 3, which
was subsequently protected to give benzenesulfonyl azaindole 4
(Scheme 1). Boronate ester 5 served as the key intermediate for
all anilinopyrimidine inhibitors and was derived from 4 via
coupling with bis(pinacolato)diboron. Capping of 5 with a
sulfonyl group was required to ensure efficient coupling in the
Compound 12 (Table 1), based on a 4,6-pyrimidine scaffold
and an undecorated indole at C-4, has modest enzyme potency
but weak activity in cells, with reasonable (28-fold) selectivity
over IGF. Selectivity over cell cycle kinase CDK2 was also
B
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1. Synthesis of Key 6-Azaindole Intermediates and Inhibitors 19 and 31
a
Reagents and conditions: (i) Br2, NaHCO3, MeOH, 0 °C, 3 h; (ii) PhSO2Cl, aq. NaOH, TBAHS, DCM, 0 °C, 1 h; (iii) PdCl2(dppf),
Bis(pinacolato)diboron, KOAc, DME, 80 °C, 18 h; (iv) Pd(PPh3)4, 2,4-dichloropyrimidine, aq. Na2CO3, dioxane, 80 °C, 1 h; (v) aq. Na2CO3,
dioxane, 120 °C, 0.5 h; (vi) NaH, MeI, 0 °C, 2 h; (vii) pTSA, 2-methoxy-4-(4-methylpiperazin-1-yl)aniline, NMP, 200 °C, 2 h; (viii) PdCl2(PPh3)2,
trans-stilbeneboronic acid, K2PO4·H20, 130 °C, 0.5 h, then aq. Na2CO3, 140 °C, 40 min.
selectivity over CDK2. The impact of an ortho-substituent in
diminishing CDK2 potency has been documented,18 and
because such a strategy had previously been employed by us
to gain CDK2 selectivity in the IGF targeting series, it is
unsurprising to see measurable, albeit weak, IGF inhibition for
this change. In vivo clearance for this compound again remains
above liver blood flow, however. The much increased enzyme
potency for 13 and 14 results in submicromolar cellular activity
for these agents. Conversion of the basic side chain in 14 to a
neutral, acetoxy-capped piperazine, as in 15, was investigated
because the transition from a basic to a neutral chemotype in
optimization of the IGF series was commensurate with
significantly improved DMPK. In the case of 15, unfortunately,
in vivo clearance remained stubbornly high and was coupled
with a slight attenuation of Dyrk1B potency. Methylation of 15
to give 16 was undertaken in an effort to understand the role, if
any, of the azaindole NH on clearance. Gratifyingly, in vivo
clearance was observed to drop significantly to a more
acceptable 16 mL/min/kg for this compound, with a slight
drop in potency relative to that of 15 (2-fold). However, when
dosed orally, 16 exhibited a low bioavailability of 5%, consistent
with poor absorption. Transformation of 4,6-pyrimidine 15 to
the matched 2,4-pyrimidine 17 highlights the role the hinge
binding group plays in both potency and selectivity. Dyrk1B
inhibition is similar for both scaffolds, but this compound, as
expected, shows much greater activity versus IGF and CDK2.
The in vivo clearance of 17 appears lower compared to that of
15, although it is still in excess of liver blood flow. Finally,
reversion of the side chain back to a basic piperazine, as in 18,
gives a highly potent Dyrk1B inhibitor with potent inhibition in
cells, with an IC50 of 85 nM. IGF and CDK2 appear to be
largely indifferent to this change, resulting in an improved
selectivity margin for 18. Compound 18 is the pyrimidine
matched pair to 14; this switch to a 2,4-pyrimidine again
confirms the largely similar Dyrk1B profile but with attenuated
selectivity versus IGF and CDK2.
Scheme 2. Synthesis of Diverse Dyrk1B Inhibitors 32 and
33
a
a
Reagents and conditions: (i) AcCl, AlCl3, 25 °C, 18 h; (ii) PhSO2Cl,
aq. NaOH, TBAHS, DCM, 0 °C, 40 min; (iii) 1,1-di-tert-butoxy-N,N-
dimethylmethanamine, DMF, 80 °C, 1 h; (iv) acetimidamide·HCl,
KO-t-Bu, DMA, 150 °C, 1 h; (v) PhCH2OH, DIAD, PPh3, THF, 25
°C, 18 h.
routinely assessed since compounds from this chemotype had
provided the original leads for exploitation in IGF,17 and 12
showed good selectivity over CDK2. Introduction of a 6-aza
substituent to the indole as in 13, however, has a dramatic effect
on Dyrk1B potency, increasing 40-fold to 2 nM, and is
suggestive of a specific interaction within the kinase active site.
This change also appears to be highly unfavorable in IGF: the
compound is inactive up to 100 μM, although selectivity against
CDK2 is slightly reduced. Comparison of the in vivo clearances
of these two compounds indicates exceptionally high clearance
for the aza-analogue, which is not explained from in vitro
clearance values, suggestive of an extrahepatic component to
clearance. Introduction of an ortho-methoxy group at the hinge
aniline, as in 14, maintains Dyrk1B potency and improves
This initial exploration highlighted the azaindole moiety as
the likely cause of the high observed in vivo clearance,
C
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Profile of Anilinopyrimidine Inhibitors of Dyrk1B
a
b
All IC50 data are reported as micromolar and are the mean of at least n = 2 independent measurements. Each has a SEM 0.2 log units. Inhibition
c
of phosphorylation of Dyrk pS421 of transiently overexpressed cMyc-Dyrk in Cos-1 cells following a 5 h treatment with compound. Clint (μL/min/
d
e
1 × 106 cells) in Wistar Han rat hepatocytes. Plasma Cl (mL/min/kg) in Wistar Han rats dosed at 2 μmol/kg. Bioavailability of an oral solution in
f
Wistar Han rats dosed at 3 μmol/kg. Blood Cl.
particularly in light of the lowered clearance with indole 12 and
methylated azaindole 16. As it appeared that both an azaindole
and basic side chain are required for high potency, we further
examined the role of substitution on the heterocycle to
modulate metabolism (Table 2). N-Methyl azaindole 19, the
basic analogue of 16, is a potent inhibitor of Dyrk1B kinase
with good selectivity over IGF and CDK2. As with 16, this
compound showed low in vitro and in vivo clearance in rat and,
gratifyingly, an improved, yet modest, bioavailability of 22%. It
is tempting to speculate that the reduced bioavailability seen
with neutral 16 is as a consequence of its much reduced
aqueous solubility (pH 7.4) relative to that of 19 (53 versus
>755 μM, respectively). Methylation adjacent to the aza-
nitrogen at C5 to give 20, or C7 to give 21, gave compounds
with very low turnover in rat hepatocytes, although in both of
these compounds, potent IGF inhibition was observed.
Consistent with earlier compounds, and despite low clearance
in vivo for 21, no oral exposure was seen for this compound
where the azaindole nitrogen was not methylated. Combining
the features of 19 and 21 to give dimethylated azaindole 22
gave a low clearance compound with the selectivity profile
somewhat restored, although bioavailability was lower at 14%,
again possibly a consequence of lower solubility (4 μM).
D
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 2. Profile of Methylated Azaindole Inhibitors of Dyrk1B
IC50 (μM)
ab
Rat DMPK
a
a
a
c
d
e
,
ID
R1
R2
R3
Dyrk1B
Cell
IGF
CDK2
Clint
Cl
F%
19
20
21
22
Me
H
H
H
0.003
0.004
0.005
0.009
0.042
0.779
0.072
0.051
3.174
4.473
10.476
17.368
13.713
18
6
27
22
H
Me
H
H
Me
Me
0.083
0.168
<5
6
38
24
0
Me
H
14
a
b
All IC50 data are reported as micromolar and are the mean of at least n = 2 independent measurements. Each has a SEM 0.2 log units. Inhibition
c
of phosphorylation of Dyrk pS421 of transiently overexpressed cMyc-Dyrk in Cos-1 cells following a 5 h treatment with compound. Clint (μL/min/
d
e
1 × 106 cells) in Wistar Han rat hepatocytes. Plasma Cl (mL/min/kg) in Wistar Han rats dosed at 2 μmol/kg. Bioavailability of an oral solution in
Wistar Han rats dosed at 3 μmol/kg.
STRUCTURAL BASIS FOR POTENCY AND
SELECTIVITY
Insights into the binding mode of the series benefited from the
solved kinase crystal structures of closely related compound 23
in both TTK and IGF1R (Figures 4 and 5), which had been
■
Figure 5. Compounds used to infer the binding mode of azaindoles in
Dyrk1B.
IGF1R. The basic piperidine side chain on the anilino group is
directed into the solvent channel, as is commonly observed in
kinase inhibitor bound crystal structures. Interactions in this
region can impact inhibitor potency and selectivity but most
notably allow for modulation of physical properties. Interest-
ingly, there appear to be differences in the binding mode
conformation of the 6-azaindole group between the two
structures. The IGF1R structure shows the 6-azaindole group
directed back toward the solvent-exposed region of the pocket;
Figure 4. Crystal structures of compound 23 bound in IGF1R (pink)
the azaindole itself does not appear to make any specific
interactions to explain the role of the nitrogen on the azaindole
ring. Conversely, in the TTK structure, this group is directed
toward the selectivity pocket region and the aza-nitrogen
appears to make a specific interaction with the conserved lysine
residue (Lys-553) that is present in most kinases. It is perhaps
nonintuitive that making a specific interaction with this lysine
residue would lead to improved selectivity; it is, however, likely
that the environment around the lysine is variable across the
kinome, and the ability of the azaindole group to bind may vary
between different kinases. The increased potency of compound
23 in TTK (and selectivity versus IGF1R) may initially suggest
and TTK (cyan) showing the change in orientation of the C-4
azaindole headgroup.
solved in-house previously (PDB codes 4D2S and 4D2R,
respectively). This compound exhibited an IC50 of 0.02 μM
against TTK and 3.3 μM against IGF1R in biochemical assays,
allowing these systems to act as potential structural surrogates.
The Dyrk1B potency of this compound was 0.14 μM. The
template binds in a traditional anilino-pyrimidine binding mode
with a hydrogen-bond acceptor−donor pair interaction with the
hinge region of each kinase, Gly-605 in TTK and Met-1082 in
E
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
this binding mode is also the preferred one in Dyrk1B.
However, we additionally observed that introduction of a 5-
chloro group to the pyrimidine template, as in compound 24,
maintains potency in Dyrk1B (IC50 of 0.025 μM) but shows
much reduced selectivity (with IC50’s of 0.02 μM in CDK2 and
0.03 μM in IGF1R). The addition of this chlorine would be
expected to cause a steric clash with the 6-azaindole and
enforce a binding mode similar to the IGF1R structure,
explaining the lowered selectivity. The presence of the chlorine
itself may also increase the potency of compound 24 against
other kinases directly, especially kinases with lipophilic
gatekeeper residues such as methionine, which is the gatekeeper
of IGF1R, the original target of this series. That Dyrk1B
potency is maintained with both a 5-H and 5-Cl pyrimidine
may support a hypothesis that the preferred Dyrk1B binding
mode to gain selectivity is as observed in TTK, but that the
alternative, IGF1R-like, binding mode is also tolerated in
Dyrk1B if additional substituents prohibit access to the more
selective TTK orientation.
agents lacking a donor in the established 1,3-donor−acceptor
hinge interaction.
Wider testing of kinase templates from our screening
collection that lacked the hydrogen-bond donor in this position
identified a published MK2 inhibitor, 26,19 which, in our hands,
had an IC50 0.043 μM in Dyrk1B. A crystal structure (PDB
code 2P3G) of a compound from this series bound in MK2
confirms interaction with the hinge region via the pyridine, with
the C-2 phenyl ring passing close to the hinge region and
directed toward solvent. It is also interesting to note that this
molecule appears to interact with the conserved lysine (Lys-93)
of MK2 (and therefore most likely Dyrk1B also) in a similar
fashion as that of compound 23 in TTK through the amide of
the C-4 group. Subsequent to this work, crystal structures of
Dyrk1A have been released into the public domain, and one
example is bound with harmine 1 (PDB code 3ANR; Figure 7).
To further develop our understanding of the binding modes
of these compounds, a number of kinases were identified that
had homologous kinase domain sequences. At the time this
work was undertaken, there were no published Dyrk family
crystal structures available; therefore, of most interest were
kinases that were likely to share structural similarity to Dyrk1B
in the ATP pocket. The published CLK1 structure bound with
10z-hymenialdisine 25 (PDB code 1Z57; Figure 6) appeared to
Figure 7. Crystal structure of harmine 1 bound to Dyrk1A.
This structure shows the inhibitor interacting with the salt
bridge (Lys-188) via the pyridine group, again reminiscent of
the interaction of the 6-azaindole with this group in the TTK
structure, and provides further evidence to support the
proposed binding mode of this series in Dyrk1B. Harmine
appears to interact with the hinge region via a hydrogen-bond
acceptor from the methoxy group to the backbone of Leu-241.
Gratifyingly, this structure also demonstrates the twist of the
carbonyl in the hinge region that we had predicted and suggests
that such a feature is likely to be observed in Dyrk1B also.
Reviewing the psi angle from the carbonyl across the kinase
structures described, IGF1R showed ψ = −92.6°, TTK, ψ =
−17.2°, CLK1, ψ = 0.2°, and Dyrk1A, ψ = 6.7°. This highlights
the most significant twist was observed in the CLK1 and
Dyrk1A structures, although the TTK structure also showed
the carbonyl to be out-of-plane relative to IGF1R (Figure 8).
Such subtle differences in hinge region conformation appear
to allow selectivity to be achieved across certain kinases. A
more significant conformational change in the hinge region has
been reported in P38α, where a flip of the hinge region residues
Leu-108 and Met-109, which changes the interaction pattern of
the hinge, led to the identification of inhibitors with increase
selectivity.20
Figure 6. Structure of 10z-hymenialdisine 25 bound to CLK1.
provide a suitable structural surrogate with ∼92% sequence
homology to Dyrk1B in the ATP pocket, although the overall
sequence homology was only ∼31%. Interestingly, the
aminoimidazolone group of 25 directly interacts with the
conserved lysine (Lys-191), as is similarly observed for the 6-
azaindole group of compound 23 in the TTK structure.
Additionally, the carbonyl group of Leu-244 on the hinge is
observed with a pronounced twist, resulting in this group being
less accessible to an inhibitor relative to that which is
commonly observed across kinases. The equivalent carbonyl
group was shown to form a direct hydrogen bond with the C-2
anilino group of 23 in both the IGF1R and TTK structures, but
this interaction appeared unlikely to be optimal if this more
signficant twist was observed in Dyrk1B. In addition, the
sequence overlay of CLK1 with Dyrk1B highlights a leucine
residue in Dyrk1B in the same position as Leu-244 of CLK1.
We therefore speculated that Dyrk1B may accommodate a
greater variety of functionality in this position, more specifically
A HYBRID 2-PHENYLPYRIMIDINE SERIES
■
Hybridization of the favored Dyrk1B azaindole pharmacophore
with reported MK2 inhibitor 26, in a manner that places a
direct phenyl group linked to the C-2 position of the
pyrimidine hinge binder, gave 27. As hypothesized, deletion
F
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
similar overall profile, albeit with reduced potency and
compromised bioavailability, which is consistent with the in
vivo profile of this headgroup in compound 21, the equivalent
compound in the aniline series.
Encouraged by the unusual observation that the donor can
be dispensed with in the usual donor−acceptor hinge
interaction of this anilinopyrimidine series, we sought to
further explore the tolerance for functionality in this region.
Extending the aromatic group at C-2 out by spacing with a
trans-alkene gave compound 31 (Table 4), which was a potent
inhibitor of Dyrk1B in enzyme and cell assays, with good
selectivity versus IGF and CDK2. Compound 32 indicates that
substantial activity is retained in this template even when the C-
2 group is pared to as small a group as methyl. Inhibitor 32,
with a molecular weight of just 210 Da, represents a highly
ligand efficient (LE), almost fragment like, yet potent inhibitor
of Dyrk1B enzyme with an IC50 of 0.017 μM (LE defined as
pIC50/heavy atom count = 0.49). Alkylation of 32 with a benzyl
group gave 33, which recovered a significant amount of the
cellular potency lost in compound 32 and illustrates the scope
for additional optimization in this region of the binding site.
Finally, in exploration of alternative hinge binding motifs,
compound 34, a structurally interesting fusion of 6- and 7-
azaindole moieties, also represents a highly potent scaffold for
Dyrk1B inhibition. In this case, restoration of the 1,3-donor−
acceptor motif leads to increased CDK2 inhibition, however.
Figure 8. Structure of 10z-hymenialdisine 25 bound to CLK1
(orange) overlaid with compound 23 bound in IGF1R (pink). The
atoms used in the measurement of the psi torsion are indicated in
bold.
of the hinge hydrogen-bond donor was tolerated and led to
potent inhibition of Dyrk1B (Table 3). Selectivity against IGF
and CDK2 was also improved, and it was anticipated that wider
kinase selectivity would also be enhanced. Further optimization
of 27 was undertaken, with an emphasis on reducing
lipophilicity. Replacement of the terminal amide of 27 with a
methanesulfonyl moiety gave 28, which showed modest cellular
activity but with excellent selectivity. Despite this, in vitro
clearance was high in rat hepatocytes. Building upon the SAR
for methyl substitution on the azaindole delineated earlier for
the anilinopyrimidine series, N-Methyl analogue 29 and C7-
Methyl analogue 30 were also synthesized. Compound 29 with
submicromolar cellular potency and increased selectivity versus
IGF and CDK2 benefited from low in vitro and in vivo turnover
and a reasonable bioavailability of 37%. Unlike that in the
aniline series, the low solubility of 29 (2 μM) did not preclude
good absorption, presumably due to the excellent permeability
as measured in Caco2 cells. The C7-Methyl analogue 30 had a
SELECTIVITY PROFILES
■
To assess the broader kinase selectivity of the Dyrk inhibitors,
representative compounds were screened in a diverse panel of
124 kinases at a single concentration of 1 μM. Inhibitors 19, 31,
and 33 were chosen as being a potent, structurally diverse set,
representing examples from the anilinopyrimidine and hybrid
series, and the kinome profile of these compounds is shown in
summary form in Table 5. Complete panel data is available as
Supporting Information. Generally, all compounds showed a
reasonable selectivity profile. Anilinopyrimidine 19 showed a
moderately selective profile, although it was the least selective
of the three, with 10% of the panel inhibited >75%. Hybrid 33
appeared to be the most selective of these, with no kinases
Table 3. 2-Phenylpyrimidine Inhibitors of Dyrk1B
IC50 (μM)
ab
Rat DMPK
a
a
a
c
d
e
,
ID
R1
R2
R3
Dyrk1B
Cell
IGF
CDK2
Clint
32
Cl
F%
0
27
28
29
30
H
H
C(O)NH-c-Pentyl
SO2Me
0.098
0.021
0.012
0.014
67.727
>100
>10
45.885
16.884
>100
H
H
0.898
0.810
2.118
129
8
Me
H
H
SO2Me
23
24
37
4
Me
SO2Me
>29
>100
9
a
b
All IC50 data are reported as micromolar and are the mean of at least n = 2 independent measurements. Each has a SEM 0.2 log units. Inhibition
c
of phosphorylation of Dyrk pS421 of transiently overexpressed cMyc-Dyrk in Cos-1 cells following a 5 h treatment with compound. Clint (μL/min/
d
e
1 × 106 cells) in Wistar Han rat hepatocytes. Plasma Cl (mL/min/kg) in Wistar Han rats dosed at 2 μmol/kg. Bioavailability of an oral solution in
Wistar Han rats dosed at 3 μmol/kg.
G
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 4. 2-Alkylpyrimidine Inhibitors of Dyrk1B
a
b
All IC50 data are reported as micromolar and are the mean of at least n = 2 independent measurements. Each has a SEM 0.2 log units. Inhibition
of phosphorylation of Dyrk pS421 of transiently overexpressed cMyc-Dyrk in Cos-1 cells following a 5 h treatment with compound.
CELLULAR PROFILING
Table 5. Kinase Selectivity Profiles of Representative
Dyrk1B Inhibitors
■
A primary hypothesis was developed that a Dyrk1B inhibitor in
combination with a MEK inhibitor would result in increased
cell death in solid tumors, with melanoma being an example of
a tumor type where a strong response might be expected. It was
observed that tumor cells upregulate Dyrk1B in response to
MEK inhibitors and that siRNA to Dyrk1B sensitized multiple
tumor cell lines to MEK inhibitors, including induction of cell
death in the most sensitive models. A selection of potent,
structurally diverse Dyrk1B inhibitors assembled in the course
of this study was profiled in A375 melanoma cells in
combination with MEK inhibitor AZD6244 (Figure 9). The
combination of MEK inhibitor and siRNA leads to an increase
in cell death over that with either modality alone. In contrast to
this, we observed no effect of structurally diverse and potent
Dyrk1B inhibitors 19, 31, and 33 in combination with MEK
inhibitor in this assay. It is noteworthy that 33, shown to be a
percent kinases
inhibited >75%
percent kinases
inhibited >50%
kinases most potently
inhibited
ID
19
31
33
10
6
16
15
0
CaMKIIβ, Flt3, DRAK1
Flt1, Flt3, MLK1
0
Mnk2, PI3KC2γ, Flt3
inhibited above the 50% level. By comparison, hybrid 31 was
slightly less selective, and it is tempting to speculate that an
alternative binding modality through the azaindole nitrogen
acceptor might be accessible for some of the kinases inhibited
by this compound. It was determined that compound 19 was
also able to potently inhibit other Dyrk family members, albeit
with selectivity in favor of Dyrk1B (IC50’s of 3, 30, and 190 nM
against Dyrk1B, Dyrk1A, and Dyrk2, respectively).
Figure 9. Impact on survival of the combination of MEK inhibitor AZD6244 and Dyrk1B inhibition on A375 melanoma cell line. Dyrk1B inhibition
is affected with inhibitors 19, 31, and 33. A375 cells were treated with AZD6244 in the presence or absence of DYRK1b inhibitors 19, 31, or 33 at a
concentration of 1.33 μM for 72 h. The viability of cells was evaluated using CellTraceTM calcein green AM and propidium iodide (PI) stains. The
percentage of dead cells plotted is the mean and SEM from at least two separate experiments.
H
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
obtained by diligent process development; preparations were repeated
if more material was required.
highly selective inhibitor of Dyrk1B, with cellular activity,
showed no effect in sensitization. Taken together, we conclude
that Dyrk1B kinase inhibition with a small molecule ATP
competitive inhibitor does not recapitulate the effects seen in
combination with target knockdown through siRNA. That gene
silencing through target knockdown leads to a different
phenotype from that seen with small molecule inhibition is
perhaps not unexpected, since the former may disrupt critical
protein complexes or inhibit other functional activities
independent of kinase activity.21 Compound 19 has recently
been profiled in depth for its role in inhibiting Dyrk1B in a
cellular setting and was shown to potently inhibit Dyrk1B
induced phosphorylation of cyclin D1 at T286.22 Dyrk1B is also
known to autophosphorylate at Y273, and, in contrast, this
activity is not abolished by either compound 19 or, indeed,
harmine 1. This is one example of a functionally different
consequence from that observed with gene knockdown
approaches, but it is not possible to attribute with certainty
the lack of observed sensitization to MEK inhibition to this
specific observation.
3-Bromo-1H-pyrrolo[2,3-c]pyridine (3). A solution of bromine
(8.7 mL, 169 mmol) in methanol (300 mL) was added dropwise to a
stirred mixture of 1H-pyrrolo[2,3-c]pyridine 2 (20 g, 169 mmol) and
sodium hydrogen carbonate (42.7 g, 508 mmol) in methanol (600
mL) at −5 °C at such a rate that the internal temperature remained
below 0 °C. The resulting mixture was allowed to warm to room
temperature and stirred for 3 h. The reaction mixture was evaporated
to dryness, redissolved in ethyl acetate (1 L), and washed sequentially
with water (2 × 1 L) and saturated brine (1 L). The organic layer was
dried, filtered, and evaporated to afford crude product. The crude
product was purified by recrystallization from toluene (1 L) to afford 3
(17.7 g, 53%) as a beige solid. 1H NMR δ 7.41 (1H, dd), 7.81 (1H, s),
8.21 (1H, d), 8.78 (1H, d), 11.98 (1H, brs); m/z (ES+) (M + H)+ =
199; HPLC tR = 1.57 min.
3-Bromo-1-(phenylsulfonyl)-1H-pyrrolo[2,3-c]pyridine (4).
Benzenesulfonyl chloride (13.7 mL, 108 mmol) was added dropwise
to compound 3 (17.7 g, 90 mmol), tetrabutylammonium hydro-
gensulfate (0.9 g, 2.7 mmol) in dichloromethane (200 mL), and
sodium hydroxide (10.78 g, 269 mmol) in water (20 mL) at 0 °C over
a period of 5 min. The resulting mixture was stirred at 0 °C for 1 h.
The reaction mixture was diluted with dichloromethane (100 mL) and
washed with water (2 × 100 mL). The organic layer was dried, filtered,
and evaporated to afford crude product, which was triturated with
methanol to give a solid that was collected by filtration and dried
CONCLUSIONS
■
Treatment of a variety of cell lines either with a MEK inhibitor
or under conditions of serum starvation leads to significant
upregulation of Dyrk1B, and combination of this with silencing
of Dyrk1B leads to enhanced cell death. A diverse array of
potent, selective, and cellular active Dyrk1B inhibitors has been
developed through a combination of targeted screening and
structure-based design, but these do not recapitulate the
sensitization seen with siRNA. Nevertheless, these compounds
may serve as useful tools to further delineate the role of Dyrk1B
in cells and may provide useful tools in further exploring the
role of the Dyrk family enzymes in an in vivo disease model
setting.
1
under vacuum to give 4 (23.8 g, 78%) as a beige powder. H NMR δ
7.53 (1H, dd), 7.61−7.69 (2H, m), 7.72−7.79 (1H, m), 8.12−8.18
(2H, m), 8.44 (1H, s), 8.51 (1H, d), 9.28 (1H, d); m/z (ES+) (M +
H)+ = 339; HPLC tR = 2.38 min.
1-(Phenylsulfonyl)-3-(4,4,5,5-tetramethyl-1,3,2-dioxaboro-
lan-2-yl)-1H-pyrrolo[2,3-c]pyridine (5). PdCl2(dppf) (1.3 g, 1.8
mmol) was added in one portion to 4,4,4′,4′,5,5,5′,5′-octamethyl-2,2′-
bi(1,3,2-dioxaborolane) (5.4 g, 21 mmol), compound 4 (6 g, 17.8
mmol), and potassium acetate (5.9 g, 60.5 mmol) in dimethoxyethane
(170 mL) at 25 °C under nitrogen. The resulting mixture was stirred
at 80 °C for 18 h. The reaction mixture was evaporated to dryness and
dissolved in ethyl acetate (600 mL), water was added (500 mL), and
the mixture was shaken and filtered through dicalite. The ethyl acetate
layer was separated and washed with saturated brine (500 mL). The
organic layer was dried and filtered. This was poured onto a plug of
silica and eluted with ethyl acetate (1.5 L) to give a pink solid, which
was triturated with diethyl ether to afford 5 (5.2 g, 76%) as a white
EXPERIMENTAL SECTION
■
Chemistry. All reactions were performed under inert conditions
(nitrogen) unless otherwise stated. Temperatures are given in degrees
celsius (°C); operations were carried out at room or ambient
temperature, that is, at a temperature in the range of 18−25 °C. All
solvents and reagents were purchased from commercial sources and
used without further purification. For coupling reactions, all solvents
were dried and degassed prior to reaction. Reactions performed under
microwave irradiation utilized either a Biotage Initiator or CEM
Discover microwave. Upon work up, organic solvents were typically
dried prior to concentration with anhydrous MgSO4 or Na2SO4. Flash
silica chromatography was typically performed on an Isco Companion,
using Silicycle silica gel, 230−400 mesh 40−63 μm cartridges, Grace
Resolv silica cartridges or Isolute Flash Si or Si II cartridges. Reverse-
phase chromatography was performed using a Waters XBridge Prep
C18 OBD column, 5 μ silica, 19 mm diameter, 100 mm length, using
decreasingly polar mixtures of either water (containing 1% NH3) and
acetonitrile or water (containing 0.1% formic acid) and acetonitrile as
eluents. Analytical LC-MS was performed on a Waters 2790 LC with a
996 PDA and 2000 amu ZQ single quadrupole mass spectrometer
using a Phenomenex Gemini 50 × 2.1 mm 5 μm C18 column; UPLC
was performed on a Waters Acquity binary solvent manager with
Acquity PDA and an SQD mass spectrometer using a 50 × 2.1 mm 1.7
μm BEH column from Waters, and purities were measured by UV
absorption at 254 nm or TIC and are ≥95% unless otherwise stated.
NMR spectra were recorded on a Bruker Av400 or Bruker DRX400
spectrometer at 400 MHz in DMSO-d6 at 303 K unless otherwise
indicated. 1H NMR spectra are reported as chemical shifts in parts per
million (ppm) relative to an internal solvent reference. Yields are given
for illustration only and are not necessarily those which can be
1
solid. H NMR δ 1.33 (12H, s), 7.65 (2H, t), 7.73−7.80 (2H, m),
8.18−8.25 (3H, m), 8.43 (1H, d), 9.24 (1H, d); m/z (ESI+) (M + H)+
= 303; HPLC tR = 0.83 min.
3-(2-Chloropyrimidin-4-yl)-1-(phenylsulfonyl)-1H-pyrrolo-
[2,3-c]pyridine (6). Tetrakis(triphenylphosphine)palladium(0) (0.78
g, 0.7 mmol) was added in one portion to 2,4-dichloropyrimidine (3 g,
20 mmol), compound 5 (5.2 g, 13.5 mmol), and 2 M sodium
carbonate (27 mL, 54 mmol) in dioxane (80 mL) at 25 °C under
nitrogen. The resulting mixture was stirred at 80 °C for 1 h. The
reaction mixture was allowed to cool and then quenched with water
(750 mL), and the precipitate was collected by filtration, washed with
water, and dried under vacuum to afford crude product as a green
solid. This was preadsorbed onto silica and purified by flash silica
chromatography (elution gradient 0−100% ethyl acetate in isohexane).
Pure fractions were evaporated to dryness to afford 6 (1.33 g, 27%) as
1
a yellow solid. H NMR δ 7.64−7.71 (2H, m), 7.75−7.81 (1H, m),
8.19−8.26 (2H, m), 8.29 (1H, d), 8.39 (1H, dd), 8.57 (1H, d), 8.82
(1H, d), 9.23 (1H, s), 9.32 (1H, d); m/z (ES+) (M + H)+ = 371;
HPLC tR = 2.35 min.
3-(2-Chloropyrimidin-4-yl)-1H-pyrrolo[2,3-c]pyridine (7).
Compound 6 (1.26 g, 3.4 mmol) and 2 M aqueous sodium carbonate
(8.5 mL, 17 mmol) were suspended in dioxane (80 mL) and sealed
into a microwave tube. The reaction was heated at 120 °C for 30 min
in the microwave reactor and cooled to room temperature. The
mixture was diluted with water (80 mL), and the organic solvent was
removed by rotary evaporation. This gave a solid that was collected by
filtration and dried under vacuum to give 7 (0.8 g, 100%) as a yellow
I
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
solid. 1H NMR δ 7.97 (1H, d), 8.29 (1H, dd), 8.33 (1H, d), 8.61 (1H,
d), 8.70 (1H, s), 12.53 (1H, brs); m/z (ES+) (M + H)+ = 231; HPLC
tR = 2.16 min.
solid. 1H NMR δ 2.49 (3H, s), 8.05 (1H, dd), 8.28 (1H, d), 8.49 (1H,
s), 8.83 (1H, d), 12.32 (1H, s); m/z (ES+) (M + H)+ = 160.79, HPLC
tR = 1.39 min.
3-(2-Chloropyrimidin-4-yl)-1-methyl-1H-pyrrolo[2,3-c]-
pyridine (8). Sodium hydride (151 mg, 3.8 mmol) was added in one
portion to compound 7 (790 mg, 3.4 mmol) in DMF (20 mL) at 0 °C
under nitrogen. The resulting mixture was stirred at 0 °C for 30 min,
and then iodomethane (0.22 mL, 3.6 mmol) was added dropwise. The
mixture was stirred and allowed to warm to ambient temperature over
2 h. The reaction mixture was diluted with ethyl acetate (100 mL) and
washed sequentially with water (3 × 100 mL) and saturated brine (100
mL). The organic layer was dried, filtered, and evaporated to afford
crude product, which was purified by flash silica chromatography
(elution gradient 0−5% 7 M ammonia/methanol in dichloromethane).
Pure fractions were evaporated to dryness to afford 8 (146 mg, 17%)
as a yellow solid. 1H NMR δ 4.02 (3H, s), 7.88 (1H, d), 8.27 (1H, dd),
8.38 (1H, d), 8.62 (1H, d), 8.68 (1H, s), 8.98 (1H, d); m/z (ES+) (M
+ H)+ = 245; HPLC tR = 2.29 min.
1-(1-(Phenylsulfonyl)-1H-pyrrolo[2,3-c]pyridin-3-yl)-
ethanone (10). Benzenesulfonyl chloride (5.1 mL, 40 mmol) was
added dropwise to compound 9 (5.3 g, 33 mmol), tetrabutylammo-
nium hydrogensulfate (0.34 g, 1 mmol), sodium hydroxide (4 g, 100
mmol) in dichloromethane (100 mL), and water (20 mL) cooled to 0
°C over a period of 5 min. The resulting suspension was stirred at 0 °C
for 40 min. The layers were separated, and the aqueous layer was
extratced with dichloromethane (100 mL). The combined organic
layers were then washed with water (100 mL) and concentrated to
dryness. The crude residue was triturated with methanol to give a
solid, which was collected by filtration and dried under vacuum to give
1
10 (5.8 g, 58%) as a white solid. H NMR δ 2.61 (3H, s), 7.68 (2H,
dt), 7.75−7.82 (1H, m), 8.10 (1H, dd), 8.25 (2H, dt), 8.49 (1H, d),
9.00 (1H, s), 9.25 (1H, d); m/z (ES+) (M + H)+ = 301.09; HPLC tR =
2.49 min.
N-(2-Methoxy-4-(4-methylpiperazin-1-yl)phenyl)-4-(1-meth-
yl-1H-pyrrolo[2,3-c]pyridin-3-yl)pyrimidin-2-amine (19). 2-Me-
thoxy-4-(4-methylpiperazin-1-yl)aniline (132 mg, 0.6 mmol), com-
pound 8 (73 mg, 0.3 mmol) and 4-methylbenzenesulfonic acid hydrate
(170 mg, 0.9 mmol) were dissolved in NMP (3 mL) and sealed into a
microwave tube. The reaction was heated at 200 °C for 2 h and
cooled. The reaction mixture was diluted with water, and saturated
sodium bicarbonate was added. This was extracted with ethyl acetate,
and the combined organic extracts were washed with water and
saturated brine. The organic layer was dried, filtered, and evaporated to
afford crude product, which was purified by flash silica chromatog-
raphy (elution gradient 0−6% 7 N ammonia/methanol in dichloro-
methane). Pure fractions were evaporated to dryness to afford 19 (30
mg, 23%) as a solid. 1H NMR δ 2.26 (3H, s), 2.48−2.54 (4H, m), 3.17
(4H, t), 3.81 (3H, s), 3.98 (3H, s), 6.56 (1H, dd), 6.68 (1H, d), 7.12
(1H, d), 7.71 (1H, d), 7.95 (1H, s), 8.18−8.25 (2H, m), 8.28 (1H, d),
8.44 (1H, s), 8.90 (1H, s); HRMS: ESI+ m/z calcd, 430.2277; found,
430.23438 (M + H)+; HPLC tR = 2.86 min.
(E)-3-(Dimethylamino)-1-(1-(phenylsulfonyl)-1H-pyrrolo[2,3-
c]pyridin-3-yl)prop-2-en-1-one (11). 1,1-Di-tert-butoxy-N,N-dime-
thylmethanamine (0.4 mL, 1.7 mmol) was added in one portion to
compound 10 (100 mg, 0.33 mmol) in DMF (2 mL). The resulting
solution was stirred at 80 °C for 2.5 h. The mixture was then
concentrated to dryness and triturated with ether to give a dark green
solid, which was purified by flash silica chromatography )elution
gradient 0−10% methanol in dichloromethane). Pure fractions were
evaporated to dryness to afford 11 (76 mg, 64%) as a yellow solid. 1H
NMR δ 2.98 (3H, s), 3.15 (3H, s), 5.96 (1H, d), 7.61−7.72 (3H, m),
7.77 (1H, t), 8.12−8.19 (2H, m), 8.22 (1H, dd), 8.43 (1H, d), 8.85
(1H, s), 9.23 (1H, d); m/z (ES+) (M + H)+ = 356.37; HPLC tR = 1.85
min.
3-(2-Methylpyrimidin-4-yl)-1H-pyrrolo[2,3-c]pyridine (32).
Compound 11 (0.5 g, 1.4 mmol), acetimidamide hydrochloride
(0.67 g, 7 mmol), and potassium 2-methylpropan-2-olate (0.95 g, 8.4
mmol) were suspended in DMA (20 mL) and sealed into a microwave
tube. The reaction was heated at 150 °C for 1 h in the microwave
reactor and cooled. The reaction mixture was filtered through Celite
and concentrated to dryness. The crude product was purified by flash
silica chromatography (elution gradient 0−10% 7 M ammonia/
methanol in DCM). Pure fractions were evaporated to dryness to
afford 32 (0.22 g, 73%) as a cream solid. 1H NMR δ 3.29 (3H, s), 7.72
(1H, d), 8.27 (1H, d), 8.40 (1H, dd), 8.54−8.58 (2H, m), 8.83 (1H,
d), 12.26 (1H, br s); HRMS: ESI+ m/z calcd, 211.0905; found,
211.09802 (M + H)+; HPLC tR = 1.86 min.
1-Benzyl-3-(2-methylpyrimidin-4-yl)-1H-pyrrolo[2,3-c]-
pyridine (33). Tri-n-butylphosphine (0.24 mL, 0.95 mmol) was
added in one portion to compound 32 (100 mg, 0.5 mmol), benzyl
alcohol (0.1 mL, 0.95 mmol), and (E)-diisopropyl diazene-1,2-
dicarboxylate (0.19 mL, 0.95 mmol) in degassed THF (4 mL)
under nitrogen. The resulting solution was stirred for 18 h and then
diluted with methanol, and the crude product was was purified by
preparative HPLC using decreasingly polar mixtures of water
(containing 1% ammonia) and acetonitrile as eluents. Fractions
containing the desired compound were evaporated to dryness to afford
33 (55 mg, 39%) as a white solid. 1H NMR δ 2.66 (3H, s), 5.63 (2H,
s), 7.24−7.40 (5H, m), 7.69 (1H, d), 8.30 (1H, d), 8.38 (1H, dd), 8.59
(1H, d), 8.75 (1H, s), 8.95 (1H, d); HRMS: ESI+ m/z calcd, 301.1374;
found, 301.14432 (M + H)+; HPLC tR = 1.06 min.
(E)-3-(2-Styrylpyrimidin-4-yl)-1H-pyrrolo[2,3-c]pyridine (31).
Tetrahydrofuran (2.15 mL) was added in one portion to trans-beta-
styreneboronic acid (0.048 g, 0.32 mmol), compound 6 (0.1 g, 0.27
mmol), PdCl2(PPh3)2 (9.5 mg, 0.01 mmol), and tri-potassium
phosphate trihydrate (0.11 g, 0.54 mmol) at 25 °C and sealed into
a microwave tube. The reaction was heated at 130 °C for 30 min in the
microwave reactor and cooled. 2 N sodium carbonate (600 μL) was
added to the reaction and sealed into a microwave tube. The reaction
was heated at 140 °C for 1 h and cooled. The reaction was evaporated
to dryness, and DMF (3 mL) was added, the mixture was filtered, and
the filtrate was purified by preparative HPLC using decreasingly polar
mixtures of water (containing 1% ammonia) and acetonitrile as
eluents. Fractions containing the desired compound were evaporated
to dryness, and the material was further purified by ion exchange
chromatography, using a 1g SCX-3 column. The desired product was
eluted from the column using 7 M ammonia in methanol, and pure
fractions were evaporated to dryness to afford 31 (0.03 g, 39%) as a
1
solid. H NMR δ 7.41 (1H, d), 7.42−7.55 (3H, m), 7.81−7.88 (3H,
m), 8.09 (1H, d), 8.39 (1H, d), 8.58 (1H, dd), 8.70 (1H, s), 8.74 (1H,
d), 8.92 (1H, d), 12.38 (1H, s); HRMS: ESI+ m/z calcd, 298.1218;
found, 299.12888 (M + H)+; HPLC tR = 2.13 min.
1-(1H-Pyrrolo[2,3-c]pyridin-3-yl)ethanone (9). Aluminum
chloride (16.9 g, 127 mmol) was added in one portion to 1H-
pyrrolo[2,3-c]pyridine 2 (3 g, 25.39 mmol) in DCM (596 mL) at 25
°C under nitrogen. The resulting mixture was stirred at 25 °C for 1 h.
To this was added acetyl chloride (9 mL, 127 mmol) dropwise over 15
min. The reaction mixture was stirred at 25 °C for 18 h under
nitrogen. The reaction mixture was quenched with methanol (80 mL)
dropwise (cautious) with cooling in a water bath. The reaction mixture
was evaporated to dryness, dissolved in water, and purified by ion
exchange chromatography, using an SCX-2 column, washing the
column well with water and then MeOH. The desired product was
eluted from the column using 7 M ammonia in methanol, and pure
fractions were evaporated to dryness to afford 9 (3 g, 74%) as a yellow
ASSOCIATED CONTENT
■
S
* Supporting Information
Protocols for the enzyme and cell assays, synthetic methods for
the remaining examples, crystallographic information, and
kinase panel selectivity data for compounds in Table 5. This
material is available free of charge via the Internet at http://
pubs.acs.org.
J
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
as inhibitors of DYRK1A/DYRK1B as potential treatment for Down’s
syndrome or Alzheimer’s disease. ACS Med. Chem. Lett. 2013, 4, 502−
503. (d) Schmitt, C.; Kail, D.; Mariano, M.; Empting, M.; Weber, N.;
Paul, T.; Hartmann, R. W.; Engel, M. Design and synthesis of a library
of lead-like 2,4-bisheterocyclic substituted thiophenes as selective
Dyrk/Clk inhibitors. PLoS One 2014, 9, e87851.
AUTHOR INFORMATION
Corresponding Author
*Tel.: +441625 517920. E-mail: jason.kettle@astrazeneca.com.
■
Present Address
†(M.C.) Cancer Research UK Manchester Institute, University
of Manchester, Wilmslow Road, Manchester M20 4BX, United
Kingdom.
(12) Gockler, N.; Jofre, G.; Papadopoulos, C.; Soppa, U.; Tejedor, F.
̈
J.; Becker, W. Harmine specifically inhibits protein kinase Dyrk1A and
interferes with neurite formation. FEBS J. 2009, 276, 6324−6337.
(13) Bain, J.; Palter, L.; Elliott, M.; Shpiro, N.; Hastie, C. J.;
McLauchlan, H.; Klevernic, I.; Arthur, J. S.; Alessi, D. R.; Cohen, P.
The selectivity of protein kinase inhibitors: a further update. Biochem.
J. 2007, 408, 297−315.
Notes
The authors declare no competing financial interest.
REFERENCES
■
(14) Song, Y.; Kesuma, D.; Wang, J.; Deng, Y.; Duan, J.; Wang, J. H.;
Qi, R. Z. Specific inhibition of cyclin-dependent kinases and cell
proliferation by harmine. Biochem. Biophys. Res. Commun. 2004, 317,
128−132.
(15) Waki, H.; Park, K. W.; Mitro, N.; Pei, L.; Damoiseaux, R.;
Wilpitz, D. C.; Reue, K.; Saez, E.; Tontonoz, P. The small molecule
harmine is an antidiabetic cell-type-specific regulator of PPARgamma
expression. Cell Metab. 2007, 5, 357−370.
(16) Kim, H.; Sablin, S. O.; Ramsay, R. R. Inhibition of monoamine
oxidase A by beta-carboline derivatives. Arch. Biochem. Biophys. 1997,
337, 137−142.
(17) Ducray, R.; Simpson, I.; Jung, F. H.; Nissink, J. W. M.; Kenny, P.
W.; Fitzek, M.; Walker, G. E.; Ward, L. T.; Hudson, K. Discovery of
novel imidazo[1,2-a]pyridines as inhibitors of the insulin-like growth
factor-1 receptor tyrosine kinase. Bioorg. Med. Chem. Lett. 2011, 21,
4698−4701.
(18) Kothe, M.; Kohls, D.; Low, S.; Coli, R.; Rennie, G. R.; Feru, F.;
Kuhn, C.; Ding, Y.-H. Selectivity-determining residues in Plk1. Chem.
Biol. Drug Des. 2007, 70, 540−546.
(19) Anderson, D. R.; Meyers, M. J.; Vernier, W. F.; Mahoney, M.
W.; Kurumbail, R. G.; Caspers, N.; Poda, G. I.; Schindler, J. F.; Reitz,
D. B.; Mourey, R. J. Pyrrolopyridine inhibitors of mitogen-activated
protein kinase-activated protein kinase 2 (MK-2). J. Med. Chem. 2007,
50, 2647−2654.
(20) Wrobleski, S. T.; Lin, S.; Dhar, T. G.; Dyckman, A. J.; Li, T.;
Pitt, S.; Zhang, R.; Fan, Y.; Doweyko, A. M.; Tokarski, J. S.; Kish, K. F.;
Kiefer, S. E.; Sack, J. S.; Newitt, J. A.; Witmer, M. R.; McKinnon, M.;
Barrish, J. C.; Dodd, J. H.; Schieven, G. L.; Leftheris, K. The
identification of novel p38α isoform selective kinase inhibitors having
an unprecedented p38α binding mode. Bioorg. Med. Chem. Lett. 2013,
23, 4120−4126.
(21) Medina, J. R. Selective 3-phosphoinositide-dependent kinase 1
(PDK1) inhibitors: dissecting the function and pharmacology of
PDK1. J. Med. Chem. 2013, 56, 2726−2737.
(1) (a) Lee, K.-M.; Deng, X.; Friedman, E. Mirk protein kinase is a
mitogen-activated protein kinase substrate that mediates survival of
colon cancer cells. Cancer Res. 2000, 60, 3631−3637. (b) Mercer, S. E.;
Ewton, D. Z.; Deng, X.; Lim, S.; Mazur, T. R.; Friedman, E. Myrk/
Dyrk1B mediates survival during the differentiation of C2C12
myoblasts. J. Biol. Chem. 2005, 280, 25788−25801.
(2) Becker, W.; Joost, H. G. Structural and functional characteristics
of Dyrk, a novel subfamily of protein kinases with dual specificity. Prog.
Nucleic Acid Res. Mol. Biol. 1999, 62, 1−17.
(3) Friedman, E. Mirk/Dyrk1B in cancer. J. Cell. Biochem. 2007, 102,
274−279.
(4) Hu, J.; Deng, H.; Friedman, E. A. Ovarian cancer cells, not
normal cells, are damaged by Mirk/Dyrk1B kinase inhibition. Int. J.
Cancer 2013, 132, 2258−2269.
(5) Yang, C.; Ji, D.; Weinstein, E. J.; Choy, E.; Hornicek, F. J.; Wood,
K. B.; Liu, X.; Mankin, H.; Duan, Z. The kinase Mirk is a potential
therapeutic target in osteosarcoma. Carcinogenesis 2010, 31, 552−558.
(6) Keramati, A. R.; Fathzadeh, M.; Go, G.-W.; Singh, R.; Choi, M.;
Faramarzi, S.; Mane, S.; Kasaei, M.; Sarajzadeh-Fard, K.; Hwa, J.; Kidd,
K. K.; Babaee Bigi, M. A.; Malekzadeh, R.; Hosseinian, A.; Babaei, M.;
Lifton, R. P.; Mani, A. A form of the metabolic syndrome associated
with mutations in DYRK1B. N. Engl. J. Med. 2014, 370, 1909−1919.
(7) (a) Loidreau, Y.; Marchand, P.; Dubouilh-Benard, C.;
Nourrisson, M.-R.; Duflos, M.; Loaec, N.; Meijer, L.; Besson, T.
̈
Synthesis and biological evaluation of N-aryl-7-methoxybenzo[b]furo-
[3,2-d]pyrimidin-4-amines and their N-arylbenzo[b]thieno[3,2-d]-
pyrimidin-4-amine analogues as dual inhibitors of CLK1 and
DYRK1A kinases. Eur. J. Med. Chem. 2013, 59, 283−295. (b) Wang,
D.; Wang, F.; Tan, Y.; Dong, L.; Chen, L.; Zhu, W.; Wang, H.
Discovery of potent small molecule inhibitors of DYRK1A by
structure-based virtual screening and bioassay. Bioorg. Med. Chem.
Lett. 2012, 22, 168−171.
(8) Altafaj, X.; Dierssen, M.; Baamonde, C.; Marti, E.; Visa, J.;
Guimera, J.; Oset, M.; Gonzalex, J.; Florez, J.; Fillat, C.; Estivill, X.
Neurodevelopmental delay, motor abnormalities and cognitive defects
in transgenic mice overexpressing Dyrk1A (minibrain), a murine
model of Down’s syndrome. Hum. Mol. Genet. 2001, 10, 1915−1923.
(9) Metro, G.; Chiari, R.; Baldi, A.; De Angelis, V.; Minotti, V.;
Crino, L. Selumetinib: a promising pharmacologic approach for KRAS-
mutant advanced non-small-cell lung cancer. Future Oncol. 2013, 9,
167−177.
(22) Ashford, A. L.; Oxley, D.; Kettle, J.; Hudson, K.; Guichard, S.;
Cook, S. J.; Lochhead, P. A. A novel DYRK1B inhibitor AZ191
demonstrates that DYRK1B acts independently of GSK3β to
phosphorylate cyclin D1 at Thr286, not Thr288. Biochem. J. 2014,
457, 43−56.
(10) Cohen, R. B.; Aamdal, S.; Nyakas, M.; Cavallin, M.; Green, D.;
Learoyd, M.; Smith, I.; Kurzrock, R. A phase I dose-finding, safety and
tolerability study of AZD8330 in patients with advanced malignancies.
Eur. J. Cancer 2013, 49, 1521−1529.
(11) (a) Anderson, K.; Chen, Y.; Chen, Z.; Dominique, R.; Glenn,
K.; He, Y.; Janson, C.; Luk, K.-C.; Lukacs, C.; Polonskaia, A.; Qiao, Q.;
Railkar, A.; Rossman, P.; Sun, H.; Xiang, Q.; Vilenchik, M.; Wovkulich,
P.; Zhang, X. Pyrido[2,3-d]pyrimidines: discovery and preliminary
SAR of a novel series of DYRK1B and DYRK1A inhibitors. Bioorg.
Med. Chem. Lett. 2013, 23, 6610−6615. (b) Coombs, T. C.; Tanega,
C.; Shen, M.; Wang, J. L.; Auld, D. S.; Gerritz, S. W.; Schoenen, F. J.;
Thomas, C. J.; Aube, J. Small-molecule pyrimidine inhibitors of the
cdc2-like (Clk) and dual specificity tyrosine phosphorylation-regulated
(Dyrk) kinases: development of chemical probe ML315. Bioorg. Med.
Chem. Lett. 2013, 23, 3654−3661. (c) Rosse, G. Tricyclic pyrimidines
K
DOI: 10.1021/acs.jmedchem.5b00098
J. Med. Chem. XXXX, XXX, XXX−XXX