Fragment-based drug discovery of potent and selective MKK3/6 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2015.11.054

The MAPK signaling cascade, comprised of several linear and intersecting pathways, propagates signaling into the nucleus resulting in cytokine and chemokine release. The Map Kinase Kinase isoforms 3 and 6 (MKK3 and MKK6) are responsible for the phosphorylation and activation of p38, and are hypothesized to play a key role in regulating this pathway without the redundancy seen in downstream effectors. Using FBDD, we have discovered efficient and selective inhibitors of MKK3 and MKK6 that can serve as tool molecules to help further understand the role of these kinases in MAPK signaling, and the potential impact of inhibiting kinases upstream of p38.

Full text of DOI:10.1016/j.bmcl.2015.11.054

Accepted Manuscript
Fragment-Based Drug Discovery of Potent and Selective MKK3/6 Inhibitors
Mark Adams, Toshitake Kobayashi, J. David Lawson, Morihisa Saitoh,
Kenichiro Shimokawa, Simone Bigi, Mark S. Hixon, Christopher R Smith,
Takayuki Tatamiya, Masayuki Goto, Joseph Russo, Charles E. Grimshaw,
Steven Swann
PII:
S0960-894X(15)30262-6
DOI:
http://dx.doi.org/10.1016/j.bmcl.2015.11.054
BMCL 23310
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
5 October 2015
10 November 2015
16 November 2015
Please cite this article as: Adams, M., Kobayashi, T., David Lawson, J., Saitoh, M., Shimokawa, K., Bigi, S., Hixon,
M.S., Smith, C.R., Tatamiya, T., Goto, M., Russo, J., Grimshaw, C.E., Swann, S., Fragment-Based Drug Discovery
of Potent and Selective MKK3/6 Inhibitors, Bioorganic & Medicinal Chemistry Letters (2015), doi: http://
dx.doi.org/10.1016/j.bmcl.2015.11.054
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Fragment-Based Drug Discovery of Potent and Selective
MKK3/6 Inhibitors.
Mark Adams*, Toshitake Kobayashi#, J. David Lawson‡, Morihisa Saitoh#, Kenichiro Shimokawa#, Simone Bigi*, Mark S.
Hixon§, Christopher R Smith*, Takayuki Tatamiya#, Masayuki Goto#, Joseph Russoϑ, Charles E. Grimshaw‡ and Steven
Swann*
*Medicinal Chemistry, ‡Structure Based Discovery Research, §Global DMPK§, CNS & Novel Target IDϑ; Takeda California Inc.,
10410 Science Center Drive San Diego California 92121, United States.
#Pharmaceutical Research Division, Takeda Pharmaceutical Company Ltd., 26-1, Muraoka-higashi 2-chome, Fujisawa, Kana-
gawa 251-8555, Japan.
ABSTRACT: The MAPK signaling cascade, comprised of several linear and intersecting pathways, propagates signal-
ing into the nucleus resulting in cytokine and chemokine release. The Map Kinase Kinase isoforms 3 and 6 (MKK3 and
MKK6) are responsible for the phosphorylation and activation of p38, and are hypothesized to play a key role in regulat-
ing this pathway without the redundancy seen in downstream effectors. Using FBDD, we have discovered efficient and
selective inhibitors of MKK3 and MKK6 that can serve as tool molecules to help further understand the role of these
kinases in MAPK signaling, and the potential impact of inhibiting kinases upstream of p38.
Treatment of chronic inflammatory diseases such as
rheumatoid arthritis (RA), psoriasis, and Crohn’s disease
remains a significant unmet medical need. The mitogen-
activated protein kinase (MAPK) family is implicated in
perpetuating inflammation and tissue injury in RA, psori-
asis, and Crohn’s disease and therefore is an attractive
target for small molecule drug intervention. 1,2,3 The
MAPK family is comprised of several linear and inter-
secting pathways, connecting Jun N-terminal kinase
(JNK), extracellular signal-regulated kinase (ERK) and
p38 MAPK, which propagate signaling in the nucleus
and are essential components in a number of critical
cellular processes.4
describes the discovery of a chemical probe using frag-
ment-based drug design (FBDD). We envisaged the
probe should inhibit MKK3/6 with an IC50 < 10nM and
selective (> 50-fold) over other members of the MAPK
signaling cascade.
An internal library of 11,012 fragments (heavy atom
count between 11-19 atoms) was screened at 100µM
against phosphorylated MKK3 using a Caliper-based
biochemical probe.13 From the primary screen, single-
point inhibition values were used to calculate approxi-
mate pIC50 (ApIC50)14 values for each fragment. These
values were then used to estimate ligand efficiency (LE)
and lipophilic ligand efficiency (LLE) values for each
compound that exhibited >30% inhibition. Using this
criteria, 93 fragments were identified with LE > 0.35 and
an LLE >4. A subset of these fragments was then eval-
uated by full dose response curves against both MKK3
and MKK6, and direct binding was measured using sur-
face plasmon resonance.15
Over the past two decades, a large number of p38
MAPK inhibitors have demonstrated efficacy in multiple
rodent models of arthritis and several of these molecules
(VX-7025, VX-745, pamapimod6, iosmapimod,
dilmapimod, doramapimod, BMS-582949, ARRY-797,
PH797804, SCIO-469) have been investigated in Phase
II for RA or Crohn’s disease.7,8 Despite the compelling
rodent efficacy, the clinical results revealed modest to
nominal efficacy. Explanations include p38a activation
itself having anti-inflammatory effects, and/or the pres-
ence of significant redundancy in the intersecting MAPK
pathways.
One fragment of interest, compound 1, was a highly effi-
cient inhibitor of both MKK3 (37µM; LE = 0.44, LLE =
4.21) and MKK6 (16µM; LE = 0.47,LLE = 4.8) and
demonstrated direct binding to MKK3 with a Kd of 2µM
(Shown in Figure 1).
An emerging proposal in the literature, supported by ro-
dent MKK3-/- and MKK6-/- knockout data 9,10,11, is to explore
protein targets upstream of p38 kinase. MKK3 and
MKK6 are phosphorylated by upstream MKKKs such as
TAK1 and ASK1, which in turn propagate the signal
through p38 to the nucleus resulting in cytokine and
chemokine release.12
Figure 1. Structure and SPR isotherm of 1 binding to
MKK3.
Further study of the role of MKK3 and MKK6 in p38
MAPK signaling is hampered by the lack of potent, se-
lective, and well-characterized inhibitors. This article

pose of 2 in MKK6 suggested it bound similarly to 1 as
described above (Figure 3). Using this binding mode,
further optimization of 2 was undertaken.
NH2
O
NH2
N
N
1
MKK3 IC50 = 37uM
MKK3 Kd = 2uM
MKK6 IC50 = 16uM
A putative binding mode for 1 was determined by molec-
ular modeling using the MOE software package16 . The
,
compound was hand-docked into the protein and solvent
of the MKK6/AMPPNP co-complex (PDB code: 3VN9)17
using the in-house co-crystal structure of 1 in the kinase
domain of BTK kinase (PDB code: 4ZLY)18 as a
guide. The binding pose suggested the primary amide
and the free amine on the bicyclic ring formed a triden-
tate interaction with the hinge of the protein (Figure 2).
Figure 3. Model of 2 (orange) overlaid with 1 (blue) bound
to MKK6.
Making use of the similarity between MKK3 and MKK6,
we drove our optimization process using MKK3 IC50 data.
Once a compound was identified with a MKK3 IC50 < 10
nM, a selectivity profile was established by screening
against a wider panel of MAPK enzymes.
Highlighted in Scheme 2, further optimization of 2 was
achieved by substituting the phenyl ring of 2 with a
pyridin-4-yl (3), which had a MKK3 IC50 of 30 nM (LE =
0.64). Based on the docked pose of 2, this boost in po-
tency is likely a result of the pyridine nitrogen lone pair
interacting favorably with the binding pocket Lysine-82.
Adding a methyl group to 3 led to 4, which gave an ap-
proximate 10-fold increase in potency and met the initial
goal of discovering a MKK3 inhibitor with an IC50 < 10
nM.
Figure 2. Model of 1 bound to MKK6.
The process of ‘hit expansion’ was initiated by testing a
series of compounds structurally related to 1, from either
internal or commercial sources. The goal of this exercise
was to establish structural activity relationships (SAR)
and refine our docking studies to validate the putative
binding mode.
Scheme 2. MKK3 in vitro enzyme inhibition data for 2-4
with reference to 1.
Compound
MKK3 IC50
LE
1
2
3
4
Scheme 1. Identification of 2 from fragment hit 1.
30nM
0.64
6.9
4nM
0.67
8.2
27µM
0.44
3.9
4.2µM
0.45
4.3
LLE
Modifications to the core structure of 4 were subsequent-
ly undertaken to potentially identify another series of
MKK3 inhibitors that were structurally differentiated from
the 3-aminopicolinamide chemotype. One design strat-
egy involved cyclization of the picolinamide ring into oth-
er fused heterocycles (Scheme 3). Through this process
a potent and efficient inhibitor of MKK3 (IC50 = 3 nM)
From this exercise 2, containing a 3-amino picolinamide
core, was identified as an efficient inhibitor (IC50 = 27µM;
LE = 0.44) of MKK3 (Shown in Scheme 1). A docked
2

was identified that contained an imidazolopyridine (5).
Further optimization resulted in 6, which contained a
unique isothiazole ring and resulted in a 30-fold boost in
potency.
cleaved during the workup. Finally, oxidation of the nitrile
group and sequential Suzuki-coupling afforded 4.
The syntheses of 5 and 6 are shown in Scheme 4B.
Bromination
of
ethyl
imidazo[1,5-A]pyridine-3-
carboxylate (E) with NBS provided intermediate F.
Subesquent hydrolysis followed by amidation with 1-
hydroxy-1H-benzotriazole ammonium salt gave com-
pound G. Suzuki-coupling with 4-pyridineboronic acid
provided 5. 5-Bromo-isothiazole (H) was treated with n-
BuLi and reacted with tri-n-butyltin chloride to give
stannane I, which upon Stille coupling with G afforded
the final product (6).
Scheme 3. MKK3 in vitro enzyme inhibition data for 4-6.
O
O
NH
O
NH2
NH2
N
N
NH2
N
N
S
N
N
N
N
4
5
6
Scheme 4. (A) Synthesis of 3 & 4. (B) Synthesis of 5 &
6.
Compound
MKK3 IC50
LE
4
5
6
NH2
N
O
(A)
4nM
0.67
8.2
150nM
0.51
6.0
4nM
0.67
7.6
NH2
N
NH2 O
NH2
CN
a
b
NH2
N
LLE
Cl
Cl
B
A
N
3
A docked pose of 6 suggests a similar binding mode to 4
(Figure 4). In this model, the isothiazole is within an
acceptable distance for a favorable interaction with Ly-
sine-82. Additionally, the sulfur likely maintains the pla-
narity of the system, as well as the preferred confor-
mation of the ring. The boost in potency could also result
from a potential sulfur-sulfur interaction between Methio-
nine-129 and the sulfur atom of the isothiazole ring.
c, d
NH
NH
O
NH
O
NH2
CN
e
f
NH2
N
N
N
Cl
Cl
D
C
N
4
(B)
O
O
O
NH2
N
R
OEt
g
i
N
N
N
N
N
Br
E
F
: R = OEt
G: R = NH2
N
h
5
O
NH2
N
Br
Sn(n-Bu)3
k
j
N
S
S
N
N
S
H
I
N
6
Reagents and conditions: (a) H2O2 aq., K2CO3 aq.,
DMSO, 77%; (b) 4-pyridinylboronic acid, Pd(dppf)Cl2-
CH2Cl2, Na2CO3, DME, 90°C, 85% (c) TFAA, Py, DMAP,
THF 93%; (d) MeI, K2CO3, DMF, 67%; (e) H2O2 aq.,
K2CO3 aq., DMSO, 67%; (f) 4-pyridinylboronic acid,
Pd(dppf)Cl2-CH2Cl2, Na2CO3, DME, 90°C, 65% (g) NBS,
CH3CN, 0°C, 92%; (b) (i) NaOH, EtOH, rt; (ii) EDCI,
HOBtNH3, DMSO, 70 °C, 71%; (h) 4-pyridinylboronic
acid, Pd(dppf)Cl2-CH2Cl2, Na2CO3, DME, 90°C, 65%; (i)
n-BuLi, tributylchlorostannane, THF, -78°C, 97%; (k) 5-
Figure 4. Model of 6 bound to MKK6.
The syntheses of 3 and 4 are shown in Scheme 4A.
Picolyl nitrile A, a key starting material for this series,
was converted into carboxamide B under oxidative con-
ditions. Introduction of the pyridine ring was accom-
plished by employing a Suzuki-Miyaura coupling reaction
providing 3. With regard to 4, N-acylation of A with TFAA
followed by N-methylation gave C whose TFA group was
3

crystal structure in a related kinase, this information
helped guide the optimization process and shed light on
key SAR. Both compounds show inhibition of P-p38a in
cells and are highly selective against other kinases in the
MAK pathway.
(tributylstannyl)isothiazole Pd(PPh3)4, PhMe, 120°C,
34%
Having identified two potent MKK3 inhibitors, their inhibi-
tory activity was tested against p38a phosphorylation in
cells (Figure 5). Human leukemic monocyte U937 cells
were treated with compounds 4 and 6 one hour prior to
TNFα stimulation; after which phospho-p38a (P-p38a)
was quantified using ELISA. Both compound 4 and
compound 6 showed dose dependent inhibitory activity
against p38a phosphorylation with IC50 values 2.7 uM
and 0.73 uM, respectively. The large shift between in
vitro inhibition and cell potency is likely a result of
MKK3’s Km for ATP (400-600nM)19.
The use of these compounds to help understand the role
of MKK in MAPK signaling, as well as the therapeutic
hypothesis behind MKK inhibition, will be reported in due
course.
1 Kumar, S.; Boehm, J.; Lee, J.C.; Nat Rev Drug Discov.,
2003, 9, 717-26.
2 Lee, J.C.; Kumar, S.; Griswold, D.E.; Underwood, D.C.;
Votta, B.J.; Adams, J.L.; Immunopharmacology, 2000, 47(2-
3),185-201.
3 Wagner, G.; Laufer, S.; Med Res Rev., 2006, 26 (1):1-62.
4 Zarubin, T.; Han, J.; Research, 2005, 15, 11–18.
5 Mjanov, N.; Kauffman, R.S.; Spencer-Green, G.T.; Arthritis
Rheum., 2009 60(5):1232-41.
6 Hen, S.B.; Cheng, T.T.; Chindalore, V.; Damjanov, N.;
Burgos-Vargas, R.; Delora, P.; Zimany, K.; Travers, H.;
Caulfield, J.P.; Arthritis Rheum., 2009, 60(2), 335-44.
Figure 5. Dose-dependent inhibitory activity of com-
pound 4 and compound 6 against p38a phosphorylation
in cells.
7
Simon, J.; Arthur, C.; Ley, S.; Nature Reviews Immunolo-
gy, 2013, 13,679–692.
8
Goldstein, D.M.; Gabriel, T.; Curr Top Med Chem., 2005;
5(10), 1017-29.
Characterization of the role of MKK3 and MKK6 in MAPK
signaling requires an in-depth understanding of the se-
lectivity profile across other kinases in the pathway.
Thus, compounds 4 and 6 were tested against a com-
prehensive panel of kinases found within the p38 signal-
ing pathway, including MKK6. As shown in Table 1,
both compounds were approximately 10x less potent
against the MKK6 isoform. In addition, compound 4 ex-
hibited >100-fold selectivity over all of the kinases in the
panel, whereas compound 6 showed moderate potency
against TAK1 kinase.
9
Hammaker, D.; Firestein, G.S.; Ann. Rheum. Dis., 2010,
69, i77-i82.
10
Hammaker, D.; Boyle, D.L.; Topoloewski, K.; Firestein,
G.S.; J Inflamm (Lond), 2014, 11: 14.
11
Yoshizawa, T.; Hammaker, D.; Boyle, D.L.; Corr, M.;
Flavell, R.; Davis, R.; Schett, G.; Firestein, G.S.; J
Immunol., 2009, 183(2),1360-7.
12 Raingeaud, J.; Whitmarsh, A.J.; Barrett, T.; Derijard, B.;
Davis, R.J.; Mol. Cell. Biol.,1996 ,16(3), 1247-1255.
13 Caliper Life Sciences, A Perkin Elmer Company. Hopkin-
ton, MA http://www.perkinelmer.com/
14
ApIC50 = -log([inhibitor]/(Vo/V-1)) where Vo = enzyme
activity in the absence of inhibitor and V = enzyme activity
in the presence of inhibitor
MKK6 p38a
p38g JNK1 JNK2 ERK1
15 SPR work conducted under contract at BIOSesnsor Tools
LLC. See Supporting Information.
4
6
<9
12
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
16 Molecular Operating Environment (MOE) software, MOE,
Chemical Computing Group Inc.
17 Han, J.; Lee, J-D.; Jiang, Y.; Li, Z.; Feng, L.; Ulevitch,
R.J.; J Biol. Chem., 2012, 151 (5):541-549.
18 Smith, C.R.; Dougan, D.R.; Komandala, M.; Kanouni, T.;
Knight, B.; Lawson, J.D.; Sabat, M.; Taylor, E.R.; Vu, P.;
Wyrick, C.; J.Med.Chem., 2015, 58, 5437-5444.
19 Internal Data no shown
ERK2 MEK1 MEK2 TAK1 MLK3 ASK1
4
6
>100
>100
>100
86
>100
>100
>100
23
>100
>100
>100
>100
Table 1. Fold Selectivity of 4 and 6 against a panel of
MAPK signaling kinases (Invitrogen)
FBBD has long been established as a reliably efficient
means to identifying novel chemotypes that can be rap-
idly optimized using structural information. Herein we
described the rapid optimization of two highly potent and
efficient MKK3/6 inhibitors from compound 1. Although
the binding pose of the fragment was based on a co-
4