2,4-Diaminopyrimidine MK2 inhibitors. Part II: Structure-based inhibitor optimization

Article abstract of DOI:10.1016/j.bmcl.2009.10.103

We describe structure-based optimization of a series of novel 2,4-diaminopyrimidine MK2 inhibitors. Co-crystal structures (see accompanying Letter) demonstrated a unique inhibitor binding mode. Resulting inhibitors had IC₅₀ values as low as 19 nM and moderate selectivity against a kinase panel. Compounds 15, 31a, and 31b inhibit TNFα production in peripheral human monocytes.

Full text of DOI:10.1016/j.bmcl.2009.10.103

Bioorganic & Medicinal Chemistry Letters 20 (2010) 334–337
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
2,4-Diaminopyrimidine MK2 inhibitors. Part II: Structure-based
inhibitor optimization
Christopher M. Harris, Anna M. Ericsson, Maria A. Argiriadi, Claude Barberis , David W. Borhani à,
Andrew Burchat, David J. Calderwood, George A. Cunha, Richard W. Dixon §, Kristine E. Frank,
Eric F. Johnson –, Joanne Kamens ––, Silvia Kwak, Biqin Li, Kelly D. Mullen, Denise C. Perron, Lu Wang,
||
àà
*
Neil Wishart, Xiaoyun Wu , Xiaolei Zhang , Tami R. Zmetra , Robert V. Talanian
Abbott Laboratories, 100 Research Drive, Worcester, MA 01605-5314, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We describe structure-based optimization of a series of novel 2,4-diaminopyrimidine MK2 inhibitors. Co-
crystal structures (see accompanying Letter) demonstrated a unique inhibitor binding mode. Resulting
inhibitors had IC50 values as low as 19 nM and moderate selectivity against a kinase panel. Compounds
Received 18 August 2009
Revised 21 October 2009
Accepted 26 October 2009
Available online 29 October 2009
15, 31a, and 31b inhibit TNF
a
production in peripheral human monocytes.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
MAPKAP-K2
MK2
TNFa
Diaminopyrimidine
Tumor necrosis factor-
a
(TNF
a
) is an important target for bio-
ring to interact with the extended hinge region; (2) substituents on
the pyrimidine-2-amine to access residues in the hydrophilic pock-
et and/or displace an ordered water behind gatekeeper residue
logical drugs that are efficacious in rheumatoid arthritis (RA) and
related autoimmune diseases.1 The Ser/Thr kinase MK2 is required
for TNF
a
production in an animal model of arthritis.2 Although
inhibitors of MK2 have been reported, none to our knowledge have
progressed to clinical studies.3–5
His106
Glu104
In Part I6, we described new in vitro tools to characterize MK2
inhibitors and enable structure-based drug design. We presented
crystal structures that show an unexpected binding mode for
2,4-diaminopyrimidine inhibitors in the MK2 ATP binding site.
The diaminopyrimidine ring was bound proximal to the DFG re-
gion, instead of adopting the commonly observed (for analogous
inhibitors of other kinases) donor-acceptor mode of binding to
the hinge region. This unexpected finding provided an opportunity
to improve potency and possibly selectivity by optimizing com-
pounds in several regions (Fig. 1): (1) substituents on the indazole
O
HN
N
H
Met138
NH
NH
O
2
O
H
H
O
S
N
H
N+
Lys93
H
NH
SH
H
H
Cys140
N
R1
N
O
N
N
H
Leu141
3
N
N
H
O
H
HN
O
O
* Corresponding author. Tel.: +1 508 849 2581.
4
R2
Asn191
E-mail address: bob.talanian@abbott.com (R.V. Talanian).
Present address: Sanofi-Aventis, Bridgewater, NJ 08807, USA.
Present address: DE Shaw Research, New York, NY 10036, USA.
Present address: Vertex Pharmaceuticals, Cambridge, MA 02139, USA.
Present address: Abbott Laboratories, Abbott Park, IL 60064, USA.
O
NH
H
2N
1
O
à
O
O
§
Glu190
HN
–
O
Asp142
––
Present address: RXi Pharmaceuticals, Worcester, MA 01605, USA.
N
H
O
||
Present address: Astra Zeneca, Waltham, MA 02451, USA.
Present address: Genzyme, Cambridge, MA 02142, USA.
Present address: Agilux Laboratories, Worcester, MA 01605, USA.
Figure 1. Structure-based optimization of compound 1, R1 = R2 = H.
àà
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.10.103

C. M. Harris et al. / Bioorg. Med. Chem. Lett. 20 (2010) 334–337
335
Table 1
H
N
H
N
SAR at 3- and 7-positions of the indazole ring
b
a
N
I
N
R1
NO2
NH2
N
NH2
I
2
3
HN
N
N
H
H
N
c
N
R2
N
N
N
H
NH
N
NH
N
Compound
R1
R2
MK2 IC50 (lM)
I
HN
N
N
1
4
5a
5b
5c
5d
10
11a
11b
11c
11d
H
I
H
H
H
H
H
H
I
24
14
5.4
0.72
1.6
1.3
9.1
3.4
H2N
H2N
4
5d
3-Thiophene
2-Naphthalene
2-Benzothiophene
2-Pyrrole
H
H
H
H
H
Scheme 1. Reagents and conditions: (a) HI (57% stabilized), 90 °C, 2 h, 82%; (b) 2-
amino-4-chloropyrimidine, EtOH, 80 °C, 2 h, 78%; (c) 1-(tert-butoxycarbonyl)pyr-
role-2-boronic acid, NaHCO3, DMF/H2O, Pd(PPh3)4, 150 °C/lW, 15 min, 29%.
3-Thiophene
2-Naphthalene
2-Benzothiophene
2-Benzofuran
0.31
0.16
0.16
O2N
O2N
H2N
O2N
a
b
N
N
H
NH2
NH2
I
I
6
7
8
N
H2N
N
c
N
N
H2N
H2N
H2N
N
N
N
N
H
N
H
N
H
e
d
I
I
S
11c
10
9
Scheme 2. Reagents and conditions: (a) IPy2BF4, CF3SO3H, DCM, 1 h, rt, 62%; (b)
NaNO2, AcOH/H2O, 4 h, 60%; (c) Fe powder, AcOH, 80 °C, 6 h, 64%; (d) 2-amino-4-
chloropyrimidine, EtOH, 80 °C, 2.5 h, 91%; (e) benzo[b]thiophene-2-boronic acid,
Na2CO3, DME/H2O, Pd(PPh3)4, 150 °C/lW, 10 min, 43%.
Screening of the Abbott compound collection produced 4-(2-
aminopyrimidin-4-ylamino) phenol, with an MK27 IC50 of 34
M.
l
Concerns about in vivo liabilities of the phenol moiety led us to test
bio-isosteres, resulting in compound 1, which substitutes a 5-
indazolyl group for the phenol (Table 1).
The binding mode (Fig. 1) suggested that 3- or 7-position sub-
stituents could make additional interactions with the MK2 ex-
tended hinge region (Tactic 1), requiring a 180° flip of the
indazole moiety which preserves hinge interactions with Leu141
(Fig. 2a).6 Representative routes to 3- and 7-substituted indazole
analogs are described in Schemes 1 and 2, respectively. 3- or 7-
iodo-substitutions had modest effects on inhibitor potency,
whereas mono- or bicyclic aromatic groups at these positions gave
significant gains (Table 1). 7-Position substitutions were slightly
preferred. The 2-benzothiophene moiety, as in compound 11c,
was the optimal indazole 7-substitution and was retained in subse-
H
N
N
Boc
I
Figure 2. (a) Model of 15 (b) model of 31b in MK2. Models were manually docked
based on optimal hinge and Lys93 interactions.
N
a, b
N
N
Boc
+
N
NH
N
N
B O
O
N
Met138; (3) replacement of the pyrimidine ring with constrained
bicyclic rings to reduce inhibitor conformational flexibility and to
interact with the glycine rich loop; and (4) substitutions of the
bridging nitrogen, or a similar position in a bicyclic system, to ac-
cess ribose pocket residues.
N
H2N
12
13
14
Scheme 3. Reagents and conditions: (a) Pd(PPh3)4, Cs2CO3, DME/H2O, 60 °C, 24 h;
(b) NaOH, EtOH, 100 °C, 2 h, 4% (two steps).

336
C. M. Harris et al. / Bioorg. Med. Chem. Lett. 20 (2010) 334–337
In many protein kinases, a ‘gatekeeper’ residue blocks access to
HN
HN
a
a hydrophobic pocket, occupancy of which has improved potency
and selectivity for some inhibitors.8 Our MK2 crystal structures6
revealed that the hydrophilic pocket beyond Met108 contains sol-
vent molecules (Fig. 1, Tactic 2). We attempted to occupy this
space by substituting the exocyclic amine with a series of aromatic
or aliphatic amides. All of these were inactive against MK2 (not
shown).
N
N
H2N
N
Cl
H2N
N
16
17
SEM
H
N
b
N
N
N
To reduce conformational flexibility between the pyrimidine
and indazole rings (Fig. 1, Tactic 3), the bridging nitrogen was con-
strained as a bicyclic system (Schemes 3 and 4). The pyrrolo[3,2-
d]pyrimidine template of 14 gave a significant improvement in po-
tency relative to the unconstrained pyrimidine 1 (Tables 1 and 2),
with a small increase in molecular weight. Appending 7-(2-benzo-
thiophene)-5-indazole onto the bicyclic template (15, Fig. 2a) re-
sulted in potency gains similar to those obtained in the
diaminopyrimidine scaffold. In contrast, the isomeric pyrrolo[2,3-
d]pyrimidine template in 21 decreased potency. Methylation of
the pyrrole N of 15 lead to the similarly active analog 30 indicating
this might be a suitable position for further exploration.
I
I
18
19
SEM
H
N
N
N
N
N
c
d
N
17 + 19
N
N
N
N
H2N
20
21
H2N
Scheme 4. Reagents and conditions: (a) 1 atm H2/Pd–C/MeOH, 16 h, 81%; (b) SEM-
Cl, TBABr, KOH, DCM, H2O, 0 °C, 30 min, quant; (c) CuI, -proline, K2CO3, DMSO,
L
Structural analyses6 suggest that ribose pocket residues Glu145
and Glu190 can be accessed from the pyrrole NH of the pyrrol-
o[3,2-d]pyrimidines (Figs. 1 and 2, Tactic 4). Such molecules were
synthesized according to Scheme 5. Only modest improvements in
enzymatic potency were achieved (ethylamino 31a, propylamino
31b, or propionic acid 31h substituents (Table 3)). In contrast to
the binding mode of 15, a model of 31b (Fig. 2a and b) shows pref-
erence for an inactive form of MK2 (glycine rich loop rotated up)
promoted by the pyrrole alkyl substitution. This may clash with
the glycine rich loop. The terminal amine can form a salt bridge
with Glu190. An analogous interaction was confirmed by an MK2
co-crystal structure with an alternate chemotype (not shown).
Our most potent MK2 inhibitor, 31a (Table 3), was tested in a
panel of 79 protein kinases. Compound 31a was 2–10-fold more
active against MK2 than eight other kinases (CHK2, PDK1, PKA,
90 °C, 16 h, 80%; (d) TBAF, ethylenediamine, THF, 50 °C, 16 h, 23%.
Table 2
Constrained A/B ring systems
R
X
NH2
N
N
Y
Compound
X
Y
R
MK2 IC50 (lM)
14
15
C
C
NH
NH
5-Indazole
7-(2-Benzothiophene)-5-
Indazole
1.4
0.056
21
30
N
C
CH
5-Indazole
49
1.9
NCH3 5-Indazole
PKCf, IKK-1, EMK, CDK2, and GSK3a), between 10- and 100-fold
selective towards 22 kinases, and >100-fold selective toward 49 ki-
nases (see Supplementary Table 1).
Several 5-substituted pyrrolo[3,2-d]pyrimidines inhibited lipo-
polysaccharide (LPS)-stimulated TNF
quent improvements of the series. Aliphatic groups at the benzo-
thiophene ring 5-position gave modest potency improvements at
the expense of a significant increase in molecular weight and so
were not pursued further (not shown).
a production in peripheral
human monocytes (PHM) (Table 3). Compound 31b had an IC50
I
I
H2N
H2N
H
N
a
b
c
d
S
N
S
Br
Br
SEM
Br
N
H
N
N
N
22
23
24
25
26
Br
Br
e
S
SEM
g
S
S
SEM
H
N
N
N
f
h, i
N
N
N
N
S
I
SEM
N
NH
N
NH
N
N
N
N
N
Boc
O
N
B
O
H2N
28
N
H2N
N
N
N
NH2
N
N
N
31b
29
12
27
Scheme 5. Reagents and conditions: (a) Py2IBF4, DCM, 1.5 h, 77%; (b) NaNO2, AcOH/H2O, 15 min, 95%; (c) benzo[b]thiophene-2-boronic acid, Na2CO3, Pd(PPh3)4, DME/H2O,
90 °C, 15 h, 70%; (d) t-BuOK, SEM-Cl, DMF, 100 °C, 18 h, 62%; (e) bis(pinacolato)-diboron, KOAc, Pd(dppf)Cl2, 80 °C, 2.5 h, 39%; (f) 12, Cs2CO3, Pd(PPh3)4, DME/H2O, 65 °C, 16 h;
NaOH/H2O, 65 °C, 1 h, 60%; (g) 10 equiv DMF-DMA, CH2Cl2/CHCl3, 45 °C, 16 h, 73%; (h) BocNH(CH2)3Br, NaH, DMF, 0 °C to rt, 16 h; (i) HCl, MeOH, 70 °C, 16 h, 10% (two steps).

C. M. Harris et al. / Bioorg. Med. Chem. Lett. 20 (2010) 334–337
337
Table 3
Ribose pocket
Cellular potencies for 15 and 31b were greater than predicted
by enzymatic data at 10 M ATP (Table 3), suggesting differences
l
N
HN
between recombinant MK2 and that in cells. Others have also ob-
served poor correlations between inhibition of recombinant MK2
enzyme and cell potency.2–4 We immunoprecipitated MK2 from
monocytes9 and showed that our compounds inhibited this mate-
rial like recombinant enzyme (Fig. 3). We conclude that recombi-
nant MK2 is not unlike MK2 in cells. We speculate that poor
enzyme to cell correlations are due to other factors such as poor
membrane penetration, inhibition of other targets that impact
the cellular readout, or altered properties of MK2 in cells when
in complex with other proteins.
NH2
N
N
N
S
R
Compound
R
MK2 IC50
PHM TNF
a
IC50 PHM MTT TC50
M)
(
lM)
(
l
M)
(l
15
H
0.056
0.019
0.035
0.26
0.055
0.28
0.086
>25
20
9.7
31a
31b
31c
31d
(CH2)2NH2
(CH2)3NH2
(CH2)4NH2
(CH2)2-4-
Piperidine
(CH2)3OH
(CH2)3OCH3
CH2CO2H
(CH2)2CO2H
(CH2)2CONH2
c-C5H6
In Sprague-Dawley rats, inhibitor 31b had no oral bioavailabil-
0.96
ity following a 10 mg/kg dose. A 3 mg/kg iv dose of 31b gave t1/2
=
5.5 h, Vb = 4 L/kg, AUC = 4.2
l
g h mLÀ1, and CL = 1.7 L hÀ1 kgÀ1. We
31e
31f
31g
31h
31i
0.037
0.82
1.9
0.02
0.041
>50
discontinued development of this series due to our difficulty in
further improving enzymatic potency, cellular potency, and in
achieving significant oral exposure.
>25
>25
>25
>25
31j
31k
Supplementary data
i-C3H7
0.49
1.0
>25
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.10.103.
8
7
6
5
References and notes
1. Gabay, C. Exp. Opin. Biol. Ther. 2002, 2, 135.
2. Gaestel, M.; Kotlyarov, A.; Kracht, M. Nat. Rev. Drug Disc. 2009, 8, 480.
3. Anderson, D. R.; Meyers, M. J.; Vernier, W. F.; Mahoney, M. W.; Kurumbail, R. B.;
Caspers, N.; Poda, G. I.; Schindler, J. F.; Reitz, D. B.; Mourey, R. J. J. Med. Chem.
2007, 50, 2647.
4. Wu, J.-P.; Wang, J.; Abeywardane, A.; Andersen, D.; Emmanuel, M.; Gautschi, E.;
Goldberg, D. R.; Kashern, M. A.; Lukas, S.; Mao, W.; Martin, L.; Morwick, T.; Moss,
N.; Pargellis, C.; Patel, U. R.; Patnaude, L.; Peet, G. W.; Skow, D.; Show, R. J.; Ward,
Y.; Werneburg, B.; White, A. Bioorg. Med. Chem. Lett. 2007, 17, 4664.
5. Schlapbach, A.; Feifel, R.; Hawtin, S.; Heng, R.; Koch, G.; Moebitz, H.; Revesz, L.;
Scheufler, C.; Velcicky, J.; Waelchli, R.; Huppertz, C. Bioorg. Med. Chem. Lett. 2008,
18, 6142.
4
4
5
6
7
8
6. Argiriadi, M. A.; Ericsson, A. M.; Harris, C. M.; Banach, D. L.; Borhani, D. W.;
Calderwood, D. J.; Demers, M. D.; DiMauro, J.; Dixon, R. W.; Hardman, J.; Kwak,
S.; Li, B.; Mankovich, J. A.; Marcotte, D.; Mullen, K. D.; Ni, B.; Pietras, M.;
Sadhukhan, R.; Sousa, S.; Tomlinson, M. J.; Wang, L.; Xiang, T.; Talanian, R. V.
Bioorg. Med. Chem. Lett. 2009, doi:10.1016/j.bmcl.2009.10.102.
7. Argiriadi, M. A.; Sousa, S.; Banach, D.; Marcotte, D.; Xiang, T.; Tomlinson, M. J.;
Demers, M.; Harris, C.; Kwak, S.; Hardman, J.; Pietras, M.; Quinn, L.; DiMauro, J.;
Ni, B.; Mankovich, J.; Borhani, D. W.; Talanian, R. V.; Sadhukhan, R. BMC Struct.
Biol. 2009, 9, 16.
Recombinant MK2
HTRF Assay, pIC50 (M)
Figure 3. MK2 immunoprecipitated from human macrophages and recombinant
MK2 exhibit similar inhibition behavior. Fit is a linear regression with slope = 0.96,
yint = À0.10, and r2 = 0.96.
8. Okram, B.; Nagle, A.; Adrian, F. J.; Lee, C.; Ren, P.; Wang, X.; Sim, T.; Xie, Y.; Wang,
X.; Xia, G.; Spraggon, G.; Warmuth, M.; Liu, Y.; Gray, N. S. Chem. Biol. 2006, 13,
779.
9. Monocytes were differentiated 5 days in RPMI with 50 ng/mL M-CSF then
stimulated 30 min with 50 ng/mL LPS. After PBS washing, cells were lysed in
50 mM Tris pH 7.5, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA,
20 mM NaF, 10 mM Na4P2O7, 1 mM Na3VO4, and Roche protease cocktail
11836153001. MK2 was immunoprecipitated with anti-MK2 agarose (Millipore)
and used in the HTRF MK2 assay.7
of 0.086
lation of the MK2 substrate Hsp27 in the same cell (2.4
than the enzyme IC50 (2.1 M) obtained at 1 mM ATP (presumably
closer to the cellular ATP concentration). Several inhibitors dis-
played toxicity determined using a 3-(4,5-dimethylthiazol-2-yl)-
2,5-diphenyltetrazolium bromide assay, preventing meaningful
PHM assay data interpretation (not shown).
l
M. This was substantially lower than that of phosphory-
l
M) and
l