Pyrrolopyrazines as selective spleen tyrosine kinase inhibitors

Article abstract of DOI:10.1021/jm301720p

We describe the discovery of several pyrrolopyrazines as potent and selective Syk inhibitors and the efforts that eventually led to the desired improvements in physicochemical properties and human whole blood potencies. Ultimately, our mouse model revealed unexpected toxicity that precluded us from further advancing this series.

Full text of DOI:10.1021/jm301720p

Article
pubs.acs.org/jmc
Pyrrolopyrazines as Selective Spleen Tyrosine Kinase Inhibitors
Fernando Padilla,*,† Niala Bhagirath,† Shaoqing Chen,† Eric Chiao,† David M. Goldstein,†,∥
Johannes C. Hermann,† Jonathan Hsu,† Joshua J. Kennedy-Smith,† Andreas Kuglstatter,† Cheng Liao,†
Wenjian Liu,‡ Lee E. Lowrie, Jr.,† Kin Chun Luk,† Stephen M. Lynch,† John Menke,† Linghao Niu,†
Timothy D. Owens,† Counde O-Yang,‡ Aruna Railkar,† Ryan C. Schoenfeld,† Michelle Slade,†
Sandra Steiner,† Yun-Chou Tan,† Armando G. Villasenor,† Ce Wang,‡ Jutta Wanner,† Wenwei Xie,‡
̃
Daigen Xu,† Xiaohu Zhang,‡,⊥ Mingyan Zhou,† and Matthew C. Lucas*,†,§
†Hoffmann-La Roche Inc., pRED, Pharma Research & Early Development, 340 Kingsland Street, Nutley, New Jersey 07110,
United States
‡BioDuro Beijing Co. Ltd., Building E, No. 29, Life Science Park Road, Changping District, Beijing 102206, P.R. China
S
* Supporting Information
ABSTRACT: We describe the discovery of several pyrrolopyrazines as potent and
selective Syk inhibitors and the efforts that eventually led to the desired improve-
ments in physicochemical properties and human whole blood potencies. Ultimately,
our mouse model revealed unexpected toxicity that precluded us from further
advancing this series.
(in macrophages).6 In addition, because Syk is positioned
upstream in the cell signaling pathway, therapies targeting
INTRODUCTION
■
Protein kinases constitute one of the largest families of human
enzymes and regulate many different signaling processes by
appending phosphate groups to proteins. Tyrosine kinases, in
Syk might be more advantageous than drugs that inhibit a
single downstream event.7
Syk’s potential as a therapeutic target for an array of
inflammatory diseases has led to many pharmaceutical
companies, as well as academic institutions, becoming involved
in the development of inhibitors of Syk. Fostamatinib
(1, Figure 1), a Syk inhibitor, has demonstrated in phase 2
studies that it can reduce disease activity in patients with
rheumatoid arthritis (RA). However, it showed adverse
events including diarrhea, hypertension, and neutropenia.8
A more selective inhibitor 2 is currently in phase 2, but
safety data has not yet been reported.9 We, and others,
have commented that some of the observed side effects of
1 are likely due to off-target activities and that inhibitors
with an enhanced selectivity profile compared to that shown
by 1 might improve the safety profile.9,10 The availability
of a compound with such selectivity will facilitate further
evaluation of Syk as a target for the treatment of diseases
such as rheumatoid arthritis, asthma, systemic lupus erythe-
matosus (SLE), idiopathic thrombocytopenic purpura (ITP),
particular, phosphorylate proteins on the alcohol moiety of
tyrosine residues. The tyrosine kinase family includes mem-
bers that control cell growth, migration, and differentiation.1
Abnormal kinase activity has been implicated in a variety of
human diseases including cancers and autoimmune and
inflammatory diseases.2 Because protein kinases are among
the key regulators of cell signaling, they are attractive drug
targets, and selective and efficacious kinase inhibitors provide a
means to modulate cellular function. These inhibitors are not
only relevant for the treatment of kinase-mediated disease pro-
cesses but are also highly useful for investigation of cell signal-
ing processes and the identification of other cellular targets of
therapeutic interest.
Spleen tyrosine kinase (Syk) inhibitors possess a high poten-
tial for the treatment of inflammatory and autoimmune dis-
orders.3 This cytoplasmic protein kinase, discovered in 1991,4
associates with receptors on the surface of immune cells including
B cells, mast cells, macrophages, and neutrophils, and nonimmune
cells such as osteoclasts.5 The engagement of these receptors with
their ligands activates Syk, which in turn orchestrates different
cellular processes, including cytokine production (in T cells and
monocytes), bone resorption (in osteoclasts), and phagocytosis
Received: November 21, 2012
Published: January 28, 2013
© 2013 American Chemical Society
1677
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Figure 1. Fostamatinib 1, PRT-062607 2, and selective Syk inhibitors 3−5.
Crohn’s disease, allergic rhinitis, multiple sclerosis, leukemia, and
had been previously identified by us as a Syk/Jak kinase inhibitor
platform.11 The pyrrolopyrazines were revisited as a starting
point in our effort to find Syk-specific inhibitors because of their
good ligand efficiency (ligand efficiency data not shown) and
high permeability, two qualities that had proved elusive in some
of our early scaffolds (3−5). On the basis of our newly developed
insight into how to obtain good Syk selectivity profiles,9 we felt
that there were new opportunities to rationally introduce selec-
tivity into the scaffold. In short, we had identified four amino acid
residues that rarely appeared in other kinase active sites: Met450,
Leu453, Pro455, and Asn457 (each having an incidence in less
than 5% of all kinases). Of these, we had demonstrated that
optimizing interactions of ATP-competitive Syk inhibitors with
Pro455, present in only nine aligned kinases of a total of 433, was
an attractive way of introducing high levels of Syk specificity.
carcinomas.3
DISCUSSION
■
Our earliest efforts to develop small molecule Syk inhibitors (3−5,
Figure 1) were successful with respect to finding potent scaffolds with
excellent selectivity profiles, yet their poor physicochemical properties
prevented us from assessing them in in vivo models of RA.9 As a
consequence, we sought new chemical scaffolds as starting points with
similar potency and selectivity but with improved potential to achieve
the biopharmaceutical properties required to deliver orally available
drug molecules. The pyrrolopyrazine scaffold (e.g., 6, Figure 2)
RESULTS
■
The pyrrolopyrazine lead molecule 6 showed good permeability,
very attractive ligand efficiency, and excellent cell-based potency
(Tables 2, 3). It also had a potency in our human whole blood
(HWB) assay that was superior to that of 1 (Tables 1, 2). The
high human whole blood potency was very appealing to us
because we consider HWB potency as an excellent predictor for
efficacy in the clinic, and obtaining good HWB activity had been a
significant challenge during our optimization of compounds such
as 3−5. Early on, we identified some challenges with the pyrrolo-
pyrazine scaffold that needed to be addressed, including a poor
selectivity profile (Figure 2b and Table 2) and the formation of
GSH adducts during in vitro incubation with human liver micro-
somes (e.g., 7, Table 3). At the outset, though, we were optimistic
Figure 2. (a) Pyrrolopyrazine kinase inhibitor 6; (b) DiscoverX kinase
screening dendrogram for 6 at 10 μM concentration. The KINOMEscan
screening platform employs an active site-directed competition binding assay
to quantitatively measure interactions between test compounds and 386 kinase
assays. Images generated using TREEspot Software Tool. Reprinted with
permission from KINOMEscan, a division of DiscoveRx Corporation, 2010.
Table 1. Inhibition of Syk Activity, Selectivity, and Human Whole Blood Potency of Compounds 1−3
a
b
c
d
compd
Syk IC50 (nM)
JAK3 selectivity
JAK2 selectivity
JAK1 selectivity
kinase selectivity
B cell IC50 (μM)
HWB IC50 (μM)
1
2
3
17
1
2
0.1
>700
63
0.4
>300
561
145/290
116/385
14/386
0.267
0.223
0.156
12.0
0.540
11.9
>1000
>139
7
a
b
See Experimental Section for details. DiscoverX kinase screening dendrogram (number of kinases inhibited >90% at 10 μM). The KINOMEscan
screening platform employs an active site-directed competition binding assay to quantitatively measure interactions between test compounds and
c
d
multiple kinase assays. Ramos B cell IC50. Human whole blood IC50.
1678
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
that these challenges could be overcome. First, on the basis of our
analysis of an X-ray structure on the initial lead 6 and our ex-
perience with previous scaffolds,11 we expected that replacement
or modification of the phenyl ring to optimize interaction with
Pro455 would improve the selectivity of the compounds (Figure 3).
Second, the GSH adduct formation observed for 7 was not
Table 2. Inhibition of Syk Activity, Selectivity, B cell Potency, Human Whole Blood Potency, and clogP Values for
Pyrrolopyrazines 6−18
1679
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Table 2. continued
a
b
See Experimental Section for details. DiscoverX kinase screening dendrogram (number of kinases inhibited >90% at 10 μM). The KINOMEscan
screening platform employs an active site-directed competition binding assay to quantitatively measure interactions between test compounds and
c
d
e
multiple kinase assays. Ramos B cell IC50. Human whole blood IC50. Calculated log P.
a
b
c
d
e,f
Table 3. Solubility , Permeability , Human Plasma Protein Binding , CYP450 Inhibition , and Liver Microsomal Stability ,
g
h
hERG Channel Inhibition , and GSH-Adduct Formation
a
b
c
d
e
f
g
h
compd
LYSA (μg/mL)
CACO2 AB (ER)
PPB (% free)
CYP450 IC50 (μM)
HLM
RLM
hERG IC50 (μM)
GSH
1
<1
<1
<1
16
0.4
0
1.5
1.4
0.94
<1
>50
>25
18
3
47
10
>1
yes
no
2
9.1
3
>50
6
15
2
16 (2C9)
>50
14
1
42
32
7
2
4
yes
no
no
8
0.2
0.2
0.2
1.8
9
9
<1
<1
<1
<1
<1
<1
<1
<1
<1
5
0.3
>50
10
11
12
13
14
15
16
17
18
0.14
0.8
14
16
14
17
18
17
10
7
48
47
44
43
51
41
30
51
no
no
0.3
>7
>50
0.1
>19
8
2
>8 (2D6)
>19
yes
no
12 (2D6)
5 (2D6)
2
47
>2
a
b
c
Lyophilized solubility assay. Apical to basolateral CACO2 permeability, followed by measured efflux ratio. Human plasma protein binding % free
d
e
fraction. CYP450 inhibition at isoforms 3A4, 2C9, and 2D6. Isoform with the highest percent inhibition listed. Stability in human liver microsomes
(mL/min/kg). Stability in rat liver microsomes (mL/min/kg). hERG channel binding. GSH adduct formation flag.
f
g
h
unexpected due to the presence of the dimethoxyphenyl front
group. Compound 7 also showed low plasma protein binding, an
attribute that we felt might be critical in our quest to obtain good
potency in the human whole blood assay. Indeed, compound 7
had HWB potency values within 2-fold of competitor molecule 2.
Molecular modeling was used to seek ways of optimizing
interactions with Pro455. Among the replacements for the
phenyl front group that were modeled and docked, we felt that
bicyclic heteroaromatic systems showed particularly strong
potential to improve the van der Waals interactions with
Pro455, and therefore exhibit enhanced selectivity profiles. From
rigorous mining of our internal database of JAK inhibitor
compounds, we knew that the pyrrolopyrazine scaffold also had
the propensity to inhibit Syk. Further analysis helped us to
identify a handful of compounds that possessed some of the
heterocyclic replacements with the best fit based on modeling.
Indazoles, such as compound 8 in Figure 3, were particularly
appealing. As previously reported by our group,12 the lowest
energy conformation of the indazole core (Figure 4) is predicted
to favor the desirable van der Waals interaction with Pro455.
5-Chloroindazole, analogue 8, showed in on our modeling
studies a good steric fit for optimal van der Waals interactions
with Pro455. The good fit prediction was confirmed by the
observed high potency as a Syk inhibitor in vitro (Table 2). On
further profiling, we were excited to find that 8 had a superior in
vitro profile overall when compared to compound 6. It not only
showed vastly improved kinase selectivity (Table 2), but in
addition it showed no glutathione adduct formation in vitro
(Table 3). We obtained an X-ray crystal structure of 8 that
confirmed the predicted binding mode and suggested that
optimal interactions with the Pro455 were driving the observed
selectivity because all other interactions appeared unchanged
compared to 6 (Figure 3). That the crystal structure supported
our selectivity hypothesis encouraged further optimization of this
scaffold. Our renewed enthusiasm for the scaffold was sustained
when we found that 8 had a potency of ∼16 μM in the HWB
assay. This value was close to that measured for 1 and certainly
appeared within reach of 2.
Our profiling also clearly highlighted areas in which further
optimization efforts were needed. Solubility was a major concern,
and despite having significantly improved selectivity over the
kinome, reducing the potent JAK family activity was important
(e.g., compound 8 was 0.2-, 1-, and 2-fold selective over JAK3,
JAK2, and JAK1, respectively). Although there is evidence to
1680
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Figure 3. Crystal structure of compound 6 in complex with Syk (PDBaccession number 4I0R, 2.1 Å resolution). Carbon atoms of the protein are colored in
pink and compound carbon atoms colored in orange (6). Overlayed is compound 8 in cyan (PDB 4I0S, 1.98 Å resolution). Nitrogen atoms are colored in
blue and oxygen atoms in red. Proline 455 is highlighted with carbon colored in green. All molecular visualizations are generated with MOE.13.
rather than the indazole moieties, had previously been assessed
for kinase selectivity in a small panel of 40 kinases which included
both Syk and JAK3. Data mining and analysis of this selectivity
data for a number of matched pairs suggested that an increase in
the steric bulk at the amide portion should manifest itself through
a favorable increase in selectivity over JAK3 and, presumably, the
other JAK family members. Thus, starting with the isopropyl
amide as our benchmark, we made the next logical step and
synthesized the t-butyl amide compound 9 and then related
Figure 4. Preferred conformation of the indazole pyrazine core.
analogue 10. Gratifyingly, potency for Syk was improved, out-
standing selectivity over the kinome was obtained (Figure 5), and
most importantly, reduced inhibition of the JAK family of kinases
was observed (e.g., 9 showed 61-, 250-, and >1373-fold selectivity
for Syk over JAK3, JAK2, and JAK1, respectively). However,
further profiling of 9 showed that it had no potency in either the
cell-based assay or the HWB assay (Table 2). We attributed the
lack of activity in these assays to poor properties due to its
elevated lipophilicity (clogP = 3.9). Thus we pursued amide
substituents that would lower the clogP into a more desirable
range. We investigated the introduction of a hydroxyl group first
because the early SAR suggested this could positively impact
properties and improve human whole blood potency (for
example, compare 6 and 7, Table 2). Compound 11 was
suggest that a combination of Syk and JAK inhibition might lead
to greater efficacy in vivo (for example, Rigel’s Fostamatinib
inhibits JAK as well as Syk, and Pfizer’s Tofacitinib is a pan-JAK
kinase inhibitor that shows strong efficacy in phase III clinical
trials for RA),14 our goal was to increase the selectivity for Syk
over the JAK family, as well as other kinases, to improve our
chances of delivering a clinically differentiated medicine.
Because the indazole moiety had provided such an excellent
boost to selectivity over the kinome, we first maintained this
group relatively constant and directed our initial optimization
efforts toward the amide moiety. Many of the earliest examples in
this pyrrolopyrazine scaffold, which contained phenyl substituents
1681
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
(2) The introduction of polar or basic groups that might
project out of the binding pocket into bulk solvent at the
indazole nitrogen (26−38, Table 4).
(3) The reduction in sp2 carbon count (39−40, Table 4).
Introduction of basicity at the amide substituent proved to be a
somewhat successful strategy for improving solubility. For many
compounds, improved solubility was achieved while maintaining
reasonable HWB potency (for example, compounds 20, 21, and
24, Tables 4 and 5). The amines were selected based on model-
ing studies that suggested that the amine functionality should be
tolerated at the specific position. We were pleased that these
changes were usually also accompanied by improved stability in
HLM and RLM (Table 5).
Compound 24 showed excellent selectivity on further
profiling, and no human ether-a-go-go-related gene (hERG)
channel inhibition was detected. Several of these analogues were
assessed in SDPK experiments, and, unfortunately, all showed
very low oral bioavailability (e.g., 24, Table 8). This might be due
to the fact that the permeability of these compounds, with the
additional hydrogen bond donors and the positive charge at
physiological pH, was generally very poor.
The results of our alternative efforts of introducing a solvent-
exposed amine via a flexible linker or as a cyclic amine, and of
introducing polar groups via flexible linkers to the indazole N-1
position, are also shown in Tables 4 and 5. Although we did not
see the solubility improvement that we had anticipated, in some
cases we observed significantly enhanced HWB potencies while
maintaining excellent to good selectivity profiles (e.g., 26 and 31,
Table 4). In contrast to the basic amines at the amide position,
which had been free from a cardiovascular toxicity flag in the
hERG assay, these compounds showed remarkably high, and
consequently unacceptable, inhibition of the hERG channel.
Inhibition of the hERG channel has been associated with QT
prolongation and fatal arrhythmias.15
Because the overall profile of these compounds looked promis-
ing, we embarked on an exploration of ways to reduce the hERG
binding affinity. In the case of the amine-containing compounds,
we investigated reducing the basicity of the amine. This is a
strategy that has improved countless series for many discovery
groups including our own,16 and examples of some of the com-
pounds with attenuated basicity are shown (28−29, Table 4).
We were disappointed to find that lowering the pKa did not
translate into improved hERG safety. For example, 3-
cyanoazetidine compound 28 and morpholino analogue 29
showed only minimal reduction of the hERG signal while
maintaining good HWB potencies.
Figure 5. (a) Pyrrolopyrazine kinase inhibitor 9; (b) DiscoverX kinase
screening dendrogram for 9 at 10 μM concentration. The KINOMEscan
screening platform employs an active site-directed competition binding
assay to quantitatively measure interactions between test compounds
and 386 kinase assays. Images generated using TREEspot Software Tool.
Reprinted with permission from KINOMEscan, a division of DiscoveRx
Corporation, 2010.
prepared and did show significantly improved HWB potency.
Permeability improved slightly, and plasma protein binding was
acceptable, suggesting that the improved physicochemical
properties were important contributors to achieving better
human whole blood potency. Next, we evaluated the 2-hydroxy-
1,1-dimethyl-ethyl amide analogue 12. This modification also
resulted in a large improvement in HWB potency, and we were
pleased that the selectivity over the JAK family remained
respectable (12-, 35-, and 34-fold over JAK3, JAK2, and JAK1,
respectively). We were further encouraged by analogue 13, which
showed very good cell and HWB potency. Thus, we determined
that bulky amide substituents were required to ensure that we
obtained Syk selectivity over the JAK family, and that the
addition of polar functionality could improve the HWB potency,
presumably through lowering clogP and improving overall
physicochemical properties.
SAR with respect to indazole substituents was also assessed,
and selected results are highlighted in Tables 2 and 3. In sum-
mary, substituents of various sizes and electronics in the 5-, 6-,
and 7-position were well tolerated. Similar to what we had
observed with the amide modifications, reduction of the clogP
generally improved whole blood potencies. For example, com-
pound 15 showed submicromolar activities in the HWB assay.
Demethylation of the indazoles to expose the polar indazole
hydrogen offered an alternative way of reducing the clogP (16−18,
Table 2) and led to the identification of 17, a very potent com-
pound in the HWB assay (IC50 = 412 nM).
Unfortunately, despite the improved clogP for these com-
pounds (Table 2), all representatives that were tested showed
poor oral bioavailability as assessed in rodent single-dose PK
studies (e.g., 15, Table 8). Because most of these examples had
moderate or good permeability but very low solubility, we also
sought to improve the solubility by disrupting the aromatic
π-stacking ability of these planar molecules by substituting at the
benzylic position. Doing so generated 18, which had excellent cell
potency and HWB potency. However, essentially no improvement
in solubility was achieved, and the PK data showed, not surprisingly,
only very minor improvements (Table 8).
We next investigated whether the amphiphilicity of these
compounds was contributing to the high hERG channel blocking
effect. When we replaced the hydrophobic tert-butyl amide with
the more hydrophilic N-(3-hydroxy-2,2-dimethyl-propyl) amide
to obtain 30, we were pleased to find that a substantial reduction
of the hERG signal was achieved. However, although the
cardiovascular toxicity risk had been reduced, it remained higher
than we were comfortable with, and we continued our search for
compounds with an improved therapeutic window.
Results from mutagenesis and molecular modeling studies
have shown that aromatic substitutions can strongly influence
interactions with the hERG receptor.17 Thus, while we had been
very encouraged by the higher potency in HWB of the OCHF2
group, we investigated some of the alternative substituents on the
indazole that we had previously identified in order to evaluate
whether the indazole-OCHF2 moiety could be acting as some
Seeking a greater impact on solubility in order to obtain com-
pounds with balanced PK/potency profiles, we implemented the
following strategies:
(1) The introduction of basic functionalities at the amide
substituent (19−25, Table 4).
1682
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
kind of recognition motif for the hERG channel. A comparison of
compounds 30 with 37 and 31 with 33 (Table 5), for example,
suggested that the 5-difluoromethoxy group was indeed
responsible for much of the hERG activity. Thus, we moved
away from this substituent. We also prepared molecules where
the amine was part of a five- or six-membered ring directly
attached to the N-1 position of the indazole (e.g., 38, Table 4).
We anticipated that in addition to potentially impacting
Table 4. Inhibition of Syk Activity, Selectivity, Human Whole Blood Potencies, and clogP Values for Pyrrolopyrazines 19−40
1683
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Table 4. continued
1684
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Table 4. continued
a
b
See Experimental Section for details. DiscoverX kinase screening dendrogram (number of kinases inhibited >90% at 10 μM). The KINOMEscan
screening platform employs an active site-directed competition binding assay to quantitatively measure interactions between test compounds and
c
d
e
f
multiple kinase assays. Ramos B cell IC50. Human whole blood IC50. Calculated log P. Tested at 1 μM concentration.
a
b
c
d
e,f
Table 5. Solubility , Permeability , Human Plasma Protein Binding , CYP450 Inhibition , Liver Microsomal Stability , hERG
g
h
i
Channel Inhibition , GSH-Adduct Formation , and cpKa
a
b
c
d
e
f
g
h
i
compd
LYSA (μg/mL)
CACO2 AB (ER)
PPB (% free)
4
CYP450 IC50 (μM)
HLM
RLM
hERG IC50 (μM)
GSH
cpKa
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
<1
429
101
3
>50
3 (2D6)
>50
12
5
25
15
19
23
26
13
0
0.1
5
0.2
2
4
10
3
0.1
0.1
8
4 (2C9)
>50
2
>30
no
2
>1
<1
<1
<1
<1
586
<1
<1
<1
<1
<1
<1
58
<1
<1
<1
>50
9
46
17
58
58
41
58
49
43
38
23
40
23
6
<0.3
0.293
0.414
1.44
6.32
0.429
4.15
>10
no
no
9.36
0.2
28
1.1
19 (2D6)
5 (2C9)
4 (2C9)
48 (2D6)
11 (2D6)
8 (2C9)
>50
5
20
20
12
5
6.41
7.33
1.3
no
no
1.1
3
0.7
0.7
8
8
>50
5
1.2
6.5
>20
4
>3
>2
0.1
0.1
>50
0
no
no
no
12
>50
9
>30
>3
0.95
>50
3
19
14
50
36
19
4
>50
4.1
yes
a
b
c
Lyophilized solubility assay. Apical to basolateral CACO2 permeability, followed by measured efflux ratio. Human plasma protein binding % free
d
e
fraction. CYP450 inhibition at isoforms 3A4, 2C9, and 2D6. Isoform with the highest percent inhibition listed. Stability in human liver microsomes
(mL/min/kg). Stability in rat liver microsomes (mL/min/kg). hERG channel binding. GSH adduct formation flag. Calculated pKa, Moka Linux
f
g
h
i
version 20090616 T.
solubility, this modification might also provide an opportunity to
address the microsomal stability of the molecules compared to
the dimethylpropylamine containing compounds. The dime-
thylpropylamine and other linear solubilizing pieces all tended to
exhibit moderate to low microsomal stability, most likely due to
N-demethylation at the terminal amine or N-dealkylation at the
N1-position of the indazole. We hypothesized that an increase in
the steric bulk around the N1 position might restrict the access of
metabolizing enzymes to this site, and these new cycles would not
have the N-demethylation problem. While we did see the desired
improvements to microsomal stability, these changes tended to
adversely affect human whole blood potency (Tables 4, 5).
The third approach to improve pharmacokinetic properties
involved the reduction of the number of sp2 centers (Table 4,
compounds 39−40). We, and others, have successfully used this
strategy in the past to improve solubility of scaffolds.6,18 Because
the pyrrolopyrazine core acted as our kinase hinge binder, and
the amide group was critical for potency, we felt our best option
was to saturate one of the aromatic rings of the heterocyclic
“selectivity” piece. We hoped that a partially saturated bicycle,
1685
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Figure 6. Crystal structure of compound 40 in complex with Syk (PDB accession number 4I0T, 1.7 Å resolution). Carbon atoms of the protein are
colored in gray, and compound carbon atoms and molecular surface are colored in gold. Hydrogen bonds between protein (hinge atoms) and
compound are highlighted in magenta.
Table 6. Inhibition of Syk Activity, Selectivity, B-Cell Potency, and Human Whole Blood Potency for Pyrrolopyrazines 41−43
a
b
See Experimental Section for details. DiscoverX kinase screening dendrogram (number of kinases inhibited >90% at 10 μM). The KINOMEscan
screening platform employs an active site-directed competition binding assay to quantitatively measure interactions between test compounds and
c
d
e
f
multiple kinase assays. Ramos B cell IC50. Human whole blood IC50. Calculated log P. Tested at 1 μM concentration.
appropriately substituted, would continue to form favorable van
der Waals interactions with Pro455. Unfortunately, none of
these analogues showed improved solubility in our LYSA assay.
Our prototype, pyrazole 39, was highly permeable, and retained
cell-based potency. Further improvements were realized when
we switched from the pyrazole to the weakly basic imidazole
(e.g., 40). Despite the low solubility values in the LYSA assay, the
pharmacokinetic properties were significantly enhanced, as is
evident from the single dose PK data: 40 showed good oral
bioavailability (Table 8). We also obtained a crystal structure
which showed that the saturated ring of compound 40 was
interacting favorably with the Pro455 (Figure 6). However, we
1686
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
believe that these interactions are suboptimal because profiling
determined that these compounds showed affinity for many
other kinases. This, coupled with the discovery that several of
these compounds were hERG inhibitors and showed GSH
adduct formation, led us to halt further investigation.
A simultaneous effort involved the evaluation of regioisomeric
scaffolds (Tables 6, 7; compounds 41−43). Applying the same
Molecular modeling was used to explore ways to achieve the
requisite interactions with Pro455, and ultimately we found
that incorporation of an indazole moiety provided excellent
selectivity over the kinome. Furthermore, an increase in the steric
bulk at the amide portion produced favorable increases in potency
and selectivity, particularly over the JAK family of kinases. We
demonstrated that a reduction of lipophilicity improved HWB
potency, a predictor for efficacy in the clinic. We also successfully
improved solubility, but despite this we were unable to improve
oral bioavailability. Intraperitoneal dosing to assess the viability of
the scaffold led to the discovery of unexpected toxicity, which
ultimately led to the deprioritization of this series in favor of more
soluble starting points with improved PK properties.
a
b
Table 7. Solubility , Permeability , Human Plasma Protein
c
d
Binding , CYP450 Inhibition , and Liver Microsomal
Stability , hERG Channel Inhibition , and GSH-Adduct
e
,f
g
h
Formation
CACO2 CYP450
hERG
g
a
LYSA
AB
IC
IC50
b
50d
e
f
h
compd (μg/mL) (ER)
(μM)
HLM RLM
(μM) GSH
EXPERIMENTAL SECTION
■
41
42
43
<1
<1
19
>8 (2C9)
>19
11
14
7
37
35
42
Chemistry. Compounds 6−43 were prepad following one of the
no
general synthetic schemes (Schemes 1, 2, 3, 4, and 5) below:
0.3
12 (2D6)
>10
no
a
b
Lyophilized solubility assay. Apical to basolateral CACO2
Scheme 1. General Synthetic Scheme 1
c
permeability, followed by measured efflux ratio. Human plasma
protein binding % free fraction. CYP450 inhibition at isoforms 3A4,
2C9 and 2D6. Lowest inhibition value indicated. Stability in human
liver microsomes (mL/min/kg). Stability in rat liver microsomes
d
e
f
g
h
(mL/min/kg). hERG channel binding. GSH adduct formation flag.
SAR learnings from the main series, we were able to generate
compounds with exceptionally good potency. The discovery that
these N-linked indazoles could also offer highly potent Syk
inhibitors did represent an important finding that led to the
identification of a new subseries. The details of these ongoing
efforts to identify safe, selective Syk inhibitors will be the subject
of future communications.
In evaluating all the compounds made in this series with
respect to overall best profiles, we selected compound 24 for IP
dosing in rodents in order to reach a decision as to whether to
continue to pursue this scaffold. We envisioned that if we could
show efficacy with this compound, the basic amine would permit
us to use a pro-drug strategy to improve oral PK behavior.
Compound 24 did show good IP single dose PK (5 mg/kg), and
we evaluated this compound further in our in vivo efficacy model.
Unfortunately, very soon after repeat dosing at 100, 30, and 10
mg/kg in our mouse model (collagen induced arthritis), the
compound showed toxicity that precluded us from continuing
the study. The etiology of this toxicity was not determined but
was of sufficient concern to us that we decided that this scaffold
should be deprioritized.
General Methods. All reagents and solvents were purchased from
commercial sources and used without further purification. 1H NMR
spectra were measured using a Bruker NMR Avance 400 MHz or Bruker
NMR Avance 300 MHz spectrometer, and chemical shifts were expressed
in δ (ppm) units using tetramethylsilane as an internal standard (in 1H
NMR description: s = singlet, d = doublet, t = triplet, q = quartet, m =
multiplet, and br = broad peak). Mass spectra were recorded on and
Agilent AG1 Single Quad G6140A spectrometer.
Purity values for all tested compounds were found to be above 95%
from the high-performance liquid chromatography (HPLC) analyses. In
general, the nomenclature used in this section is based on
AUTONOMTM v.4.0, as noted above, or, alternatively, based on
ChemDraw. Protein crystal structures were determined as described
previously.19 Commonly used abbreviations: acetyl (Ac), atmospheres
(Atm), tert-butoxycarbonyl (Boc), benzyl (Bn), butyl (Bu), benzyloxy-
carbonyl (CBZ or Z), dibenzylideneacetone (dba), dichloromethane
(DCM), di-iso-butylaluminumhydride (DIBAL or DIBAL-H), di-iso-
propylethylamine (DIPEA), 4-N,N-dimethylaminopyridine (DMAP),
CONCLUSIONS
■
We demonstrated that seeking to optimize the interaction
between ATP-competitive Syk inhibitors and a specific Pro455
residues constitutes a reliable approach toward the discovery of
drug candidates with high levels of Syk specificity and potency.
a
Table 8. Single-Dose Pharmacokinetic Parameters in Mouse of Selected Compounds
compd
species
rat
CLiv (mL/min/kg)
T1/2 h
Vdss (L/kg)
AUC/dose (ng/mL·h/(mg/kg))
F (%)
15
18
24
24
40
25.8
44.2
55.6
NA
0.71
4.93
5.33
NA
3.7
8.4
3.07
5.69
0
0.5
2.2
0
rat
rat (po)
mouse (ip)
rat
24.4
NA
1.73
286
3.76
NA
100
48.8
0.52
a
n = 2−3 per time point. IV dose = 0.5 or 1 mg/kg. PO dose = 10 or 20 mg/kg. IP dose = 5 mg/kg.
1687
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Scheme 2. General Synthetic Scheme 2
Scheme 3. General Synthetic Scheme 3
N,N-dimethylformamide (DMF), dimethyl sulfoxide (DMSO), 1,1′-bis-
(diphenylphosphino)ferrocene (dppf), 1-(3-dimethylaminopropyl)-3-
ethylcarbodiimide hydrochloride (EDCI), ethylenediaminetetraacetic
acid (EDTA), ethyl (Et), ethyl acetate (EtOAc), ethanol (EtOH),
diethyl ether (Et2O), O-(7-azabenzotriazole-1-yl)-N,N,N′N′-tetrame-
thyluronium hexafluorophosphate acetic acid (HATU), 1-N-hydrox-
ybenzotriazole (HOBt), high pressure liquid chromatography (HPLC),
liquid chromatography−mass spectrometry (LCMS), methanol
(MeOH), melting point (mp or MP), methyl (Me), mass spectrum
(ms or MS), N-methylmorpholine (NMM), N-methylpyrrolidone
(NMP), phenyl (Ph), propyl (Pr), iso-propyl (i-Pr), pyridine (pyr),
room temperature (rt or RT), triethylamine (TEA or Et3N),
trifluoroacetic acid (TFA), thin layer chromatography (TLC), and
tetrahydrofuran (THF). Conventional nomenclature including the
prefixes normal (n), iso (i-), secondary (sec-), tertiary (tert-), and neo- have
their customary meaning when used with an alkyl moiety.20
Step 1. (2-Bromo-7-hydroxymethyl-pyrrolo[2,3-b]pyrazin-5-yl)-
methanol 50.
To a suspension of 2-bromo-5H-pyrrolo[2,3-b]pyrazine (5.0 g,
25.2 mmol) in 1,4-dioxane (100 mL) was added 2.0 M aqueous
NaOH (25 mL, 50.0 mmol) and 37% aqueous formaldehyde (19 mL,
252 mmol). The dark homogeneous reaction mixture was stirred at
room temperature overnight. The organics were evaporated under
reduced pressure. The aqueous layer was neutralized with 1.0 M HCl
and extracted with EtOAc (2×). The combined organics were
concentrated to afford 2.6 g of an orange solid. Upon standing, a
thick brown precipitate formed in the aqueous layer. The precipitate
was collected by filtration and dried. The brown solid was extracted
with hot 10% MeOH/EtOAc (3 × 200 mL). The extracts were
combined and evaporated to provide an additional 3.05 g of orange
solid to yield 50 (5.65 g, 87%). LCMS (M + H)+ = 258, 260. 1H NMR
(DMSO-d6, 400 MHz) δ: 8.43 (s, 1H), 7.96 (s, 1H), 6.71 (t, J = 7.3 Hz,
1H), 5.59 (d, J = 7.6 Hz, 2H), 5.10 (t, J = 5.3 Hz, 1H), 4.66 (d, J = 5.6
Hz, 2H).
The preparation of 9 represents a typical procedure used for the
synthesis of the pyrrolopyrazines.
N-tert-Butyl-2-(6-chloro-1-methyl-1H-indazol-3-yl)-5H-pyrrolo-
[2,3-b]pyrazine-7-carboxamide 9.
1688
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Scheme 4. General Synthetic Scheme 4
Scheme 5. General Synthetic Scheme 5
1689
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
Step 2. (2-Bromo-5H-pyrrolo[2,3-b]pyrazin-7-yl)-methanol 51.
In a round-bottomed flask, 53 (167 mg, 0.57 mmol) was dissolved in
THF (2 mL) and the solution was cooled to 0 °C. Potassium tert-
butoxide (90 mg, 0.80 mmol) was added, and the reaction mixture was
stirred at 0 °C for 1.25 h. Methyl iodide (0.045 mL, 0.72 mmol) was
added dropwise and then the ice bath was removed and the reaction
mixture was stirred at room temperature for 2 h. The reaction mixture
was quenched with water and extracted with EtOAc (2×). The com-
bined organic layers were washed with water and brine and then dried
over sodium sulfate, filtered, and concentrated. The residue was purified
by flash chromatography eluting with 0−5% EtOAc/hexanes to afford
54 (110 mg, 63%) as a yellow solid. 1H NMR (DMSO-d6) δ: 7.91 (dd,
J = 1.8, 0.5 Hz, 1H), 7.44 (dd, J = 8.7, 0.6 Hz, 1H), 7.22 (dd, J = 8.5, 1.8
Hz, 1H), 4.05 (s, 3H).
To a suspension of 50 (5.65 g, 21.9 mmol) in THF (150 mL) was added
a solution of 2.0 M aqueous NaOH (33 mL, 66 mmol). The homo-
geneous reaction mixture was stirred overnight, and then the organics
were removed under reduced pressure. The aqueous residue was
brought to pH 4 with 1.0 M aqueous HCl. The resulting precipitate was
collected by filtration and rinsed with H2O to afford 3.68 g of a yellow
solid. The filtrate was extracted with EtOAc (2×), and the organics were
concentrated under reduced pressure to provide an additional 0.92 g of
yellow solid to yield 51 (4.60 g, 92%). 1H NMR (DMSO-d6, 300 MHz)
δ: 12.19 (br s, 1H), 8.33 (s, 1H), 7.85 (s, 1H), 4.96 (t, J = 5.3 Hz, 1H),
4.62 (d, J = 4.9 Hz, 2H).
Step 6. 6-Chloro-1-methyl-3-tributylstannanyl-1H-indazole 55.
Step 3. 2-Bromo-5-(2-trimethylsilanyl-ethoxymethyl)-5H-pyrrolo-
Compound 54 (108 mg, 0.35 mmol) was dissolved in THF (2 mL),
and the solution was cooled to −16 °C using a NaCl/ice bath.
Isopropylmagnesium chloride (2.0 M in THF, 0.20 mL, 0.40 mmol) was
added dropwise, and the reaction mixture was stirred at −16 °C for
15 min. Tributylchlorostannane (0.11 mL, 0.40 mmol) was slowly added,
and the reaction mixture was allowed to warm to room temperature over
1.5 h. The reaction mixture was quenched with saturated NH4Cl solution
and extracted with EtOAc (2×). The combined organic layers were
washed with water and brine and then dried over sodium sulfate, filtered,
and concentrated to afford 55 as a brown oil, which was used
immediately into the next step without further purification.
[2,3-b]pyrazine-7-carbaldehyde 52.
A stock solution of Jones reagent (2.67 M) was prepared by carefully
adding concentrated H2SO4 (2.3 mL) to CrO3 (2.67 g) amd then
diluting to 10 mL with H2O. To a suspension of 51 (4.6 g, 20.1 mmol) in
acetone (300 mL) was slowly added Jones reagent (9 mL, 24.0 mmol).
During the addition, the starting material gradually dissolved and a thick
green solid was formed. The reaction mixture was stirred for 15 min and
then quenched with i-PrOH (2 mL) and filtered over Celite, rinsing with
acetone. The filtrate was concentrated to provide 4.76 g of 2-bromo-5H-
pyrrolo[2,3-b]pyrazine-7-carbaldehyde as a yellow−orange solid that
was used without further purification. To a solution of this solid in DMF
(50 mL) at 0 °C was added NaH (60% in mineral oil, 1.2 g, 30.1 mmol).
The reaction mixture was stirred at room temperature for 30 min and
then cooled back to 0 °C, and 2-(trimethylsilyl)ethoxymethyl chloride
(4.3 mL, 24.1 mmol) was slowly added. The reaction mixture was
warmed to room temperature and stirred for 1 h and then quenched
with H2O and extracted with EtOAc (3×). The combined organics were
washed with H2O (3×) and brine and then dried over MgSO4 and
concentrated. The residue was purified by flash chromatography eluting
with 20−30% EtOAc/hexanes to give 52 (3.82 g, 53%) as a yellow solid.
1H NMR (CDCl3, 300 MHz) δ: 10.37 (s, 1H), 8.50 (s, 1H), 8.33 (s, 1H),
5.73 (s, 2H), 3.53−3.70 (m, 2H), 0.90−1.05 (m, 2H), 0.00 (s, 9H).
Step 4. 6-Chloro-3-iodo-1H-indazole 53.
Step 7. 2-(6-Chloro-1-methyl-1H-indazol-3-yl)-5-(2-trimethylsi-
lanyl-ethoxymethyl)-5H-pyrrolo[2,3-b]pyrazine-7-carbaldehyde 56.
In a round-bottomed flask, 52 (400 mg, 1.12 mmol) and 55 (1.11 g, 1.7
mmol) were dissolved in DMF (10 mL). The flask was evacuated and
backfilled with argon, and then tetrakis(triphenylphosphine)palladium
(0) (65 mg, 0.056 mmol) and copper(I) iodide (43 mg, 0.23 mmol)
were added. The reaction mixture was stirred at 80 °C in an oil bath
overnight and then cooled to room temperature, quenched with water,
and extracted with diethyl ether (2×). The combined organic layers
were washed twice with water and once with brine and then dried over
sodium sulfate, filtered, and concentrated. The residue was adsorbed on
silica gel and purified by flash chromatography eluting with 0−30%
EtOAc/hexanes to give 56 (416 mg, 84%) as a yellow solid. 1H NMR
(chloroform-d) δ: 10.52 (s, 1H), 9.33 (s, 1H), 8.78 (d, J = 8.7 Hz, 1H),
8.33 (s, 1H), 7.49 (s, 1H), 7.34 (dd, J = 8.7, 1.5 Hz, 1H), 5.80 (s, 2H),
4.19 (s, 3H), 3.58−3.77 (m, 2H), 0.91−1.05 (m, 2H), 0.00 (s, 9H).
Step 8. 2-(6-Chloro-1-methyl-1H-indazol-3-yl)-5-(2-trimethylsi-
lanyl-ethoxymethyl)-5H-pyrrolo[2,3-b]pyrazine-7-carboxylic Acid 57.
In a round-bottomed flask, 6-chloro-1H-indazole (90 mg, 0.59 mmol,
available commercially from Ark Pharm, Inc.) was dissolved in DMF
(1.4 mL). Iodine (300 mg, 1.18 mmol) was added, followed by
potassium hydroxide (128 mg, 2.28 mmol). The dark reaction mixture
was stirred at room temperature for 3 h and then was quenched with
10% aqueous NaHSO3 and extracted with diethyl ether (2×). The
combined organic layers were washed with water and brine and then
dried over sodium sulfate, filtered, and concentrated to give 53 (169 mg,
98%) as an orange solid. 1H NMR (chloroform-d) δ: 7.52 (d, J = 1.5 Hz,
1H), 7.46 (s, 1H), 7.44 (s, 1H), 7.23 (d, J = 1.5 Hz, 1H), 7.20 (d, J =
1.5 Hz, 1H)
Step 5. 6-Chloro-3-iodo-1-methyl-1H-indazole 54.
In a round-bottomed flask, 56 (412 mg, 0.93 mmol) was suspended in
1,4-dioxane (15 mL) and water (3 mL). The suspension was cooled to
0 °C, and sulfamic acid (543 mg, 5.59 mmol) was added. Then, a
solution of sodium chlorite (80%, 137 mg, 1.21 mmol) and potassium
dihydrogen phosphate (1.52 g, 11.2 mmol) in water (9 mL) was added
1690
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
via dropping funnel over 20 min. After the addition was complete, the ice
bath was removed and the reaction mixture was stirred at room tempera-
ture for 3 h. THF (15 mL) was added, and the reaction mixture was
stirred at room temperature for an additional 3 h. The reaction mixture
was diluted with water and extracted with EtOAc (2×). The combined
organic layers were washed with water and brine and then dried over
sodium sulfate, filtered, and concentrated. The residue was triturated
with ethyl acetate/hexanes to afford 57 (358 mg, 84%) as a light-yellow
powder. 1H NMR (chloroform-d) δ: 9.37 (s, 1H), 8.42 (t, J = 4.3 Hz,
2H), 7.52 (d, J = 1.5 Hz, 1H), 7.34 (dd, J = 8.5, 1.7 Hz, 1H), 5.79 (s, 2H),
4.20 (s, 3H), 3.52−3.75 (m, 2H), 0.93−1.06 (m, 2H), 0.00 (s, 9H).
Step 9. N-tert-Butyl-2-(6-chloro-1-methyl-1H-indazol-3-yl)-5H-
pyrrolo[2,3-b]pyrazine-7-carboxamide 9. To a stirred suspension of
ACKNOWLEDGMENTS
We thank Karen Lackey, Satwant Narula and Julie DeMartino for
their guidance and strategic discussions.
■
ABBREVIATIONS USED
■
ATP, adenosine triphosphate; Btk, Bruton’s tyrosine kinase;
CD69, cluster of differentiation 69; GSH, glutathione; hERG,
human ether-a-go-go-related gene; HLM, human liver micro-
somes; HWB, human whole blood; IP, intraperitoneally; ITP,
idiopathic thrombocytopenic purpura; Jak, Janus kinase; LYSA,
lyophilized solubility assay; PK, pharmacokinetics; PPB, plasma
protein binding; RA, rheumatoid arthritis; RLM, rat liver
microsomes; SDPK, single dose pharmacokinetics; SLE,
systemic lupus erythematous; Syk, spleen tyrosine kinase
REFERENCES
■
(1) Lawrence, D. S.; Niu, J. Protein Kinase Inhibitors: The Tyrosine-
Specific Protein Kinases. Pharmacol. Ther. 1998, 77, 81−114.
́
(2) Mocsai, A.; Ruland, J.; Tybulewicz, V. L. J. The SYK tyrosine
kinase: a crucial player in diverse biological functions. Nature Rev.
Immunol. 2010, 10, 387−402.
(3) Pamuk, O. N.; Tsokos, G. Spleen tyrosine kinase inhibition in the
treatment of autoimmune, allergic and autoinflammatory diseases.
Arthritis Res. Ther. 2010, 12, 222.
57 (111 mg, 242 μmol) in DMF (4 mL) was added EDC (107 mg,
557 μmol), HOBT (107 mg, 557 μmol), and 2-methylpropan-2-amine
(78.6 mg, 0.113 mL, 1.08 mmol) at 20 °C. After 3 h, the reaction mixture
was diluted with ethyl acetate, washed with 10% citric acid solution,
saturated sodium bicarbonate, and brine and then dried (sodium
sulfate), filtered, and concentrated in vacuo. Crude N-tert-butyl-2-
(6-chloro-1-methyl-1H-indazol-3-yl)-5-((2-(trimethylsilyl)ethoxy)-
methyl)-5H-pyrrolo[2,3-b]pyrazine-7-carboxamide (140 mg, 273
μmol) was dissolved in dichloromethane (5 mL), and to this solution
was added trifluoroacetic acid (3.11 g, 2.1 mL, 27.3 mmol) at 20 °C.
After 6 h, the mixture was concentrated in vacuo. The residue was
suspended in dichloromethane (5 mL) and ethylenediamine (1.23 g,
1.38 mL, 20.5 mmol) added and the mixture stirred at room temperature
for 15 h and then concentrated in vacuo. The solid residue was triturated
with water, separated by filtration, dried, and then purified by flash
chromatography eluting with 0−5% methanol in dichloromethane to
give 9 (63 mg, 60%) as a pale-yellow powder. LCMS (M + Na)+ = 405.
1H NMR (CDCl3) δ: 11.84 (br s, 1H), 9.05 (s, 1H), 8.38 (d, J = 8.7 Hz,
1H), 8.14 (d, J = 3.4 Hz, 1H), 7.96 (s, 1H), 7.39 (d, J = 1.5 Hz, 1H), 7.08
(dd, J = 8.5, 1.7 Hz, 1H), 4.04 (s, 3H), 1.48 (s, 9H).
(4) Taniguchi, T.; Kobayashi, T.; Kondo, J.; Takahashi, K.; Nakamura,
H.; Suzuki, J.; Nagai, K.; Yamada, T.; Nakamura, S.; Yamamura, H.
Molecular cloning of a porcine gene Syk that encodes a 72-kDa protein−
tyrosine kinase showing high susceptibility to proteolysis. J. Biol. Chem.
1991, 266, 15790−15796.
(5) Riccaboni, M.; Bianchi, I.; Petrillo, P. Spleen tyrosine kinases:
biology, therapeutic targets and drugs. Drug Discovery Today 2010, 15,
517−530.
(6) Kyttaris, V. C.; Tsokos, G. C. Syk kinase as a treatment target for
therapy in autoimmune diseases. Clin. Immunol. 2007, 124, 235−237.
(7) Wong, B. R.; Grossbard, E. B.; Payan, D. G.; Masuda, E. S.
Targeting Syk as a treatment for allergic and autoimmune disorders.
Expert Opin. Invest. Drugs 2004, 13, 743−762.
(8) (a) Tam, F. W. K.; McAdoo, S. P. Fostamatinib disodium. Drugs
Future 2011, 36, 273−280. (b) Weinblatt, M. E.; Kavanaugh, A.;
Genovese, M. C.; Musser, T. K.; Grossbard, E. B.; Magilavy, D. B. An
oral spleen tyrosine kinase (Syk) inhibitor for rheumatoid arthritis. N.
Engl. J. Med. 2010, 363, 1303−1312.
(9) Lucas, M.; Goldstein, D.; Hermann, J.; Kuglstatter, A.; Liao, C.; Liu,
W.; Padilla, F.; Slade, M.; Wanner, J.; Xie, W.; Zhang, X.; Luk, K. C.
Rational design of highly selective spleen tyrosine kinase inhibitors. J.
Med. Chem. 2012, 55, 10414−10423.
(10) Coffey, G.; DeGuzman, F.; Inagaki, M.; Pak, Y.; Delaney, S. M.;
Ives, D.; Betz, A.; Jia, Z. J.; Pandey, A.; Baker, D.; Hollenbach, S. J.;
Phillips, D. R.; Sinha, U. Specific Inhibition of Spleen Tyrosine Kinase
Suppresses Leukocyte Immune Function and Inflammation in Animal
Models of Rheumatoid Arthritis. J. Pharmacol. Exp. Ther. 2012, 34, 350−
359.
(11) (a) Hendricks, R. T.; Hermann, J. C.; Jaime-Figueroa, S.; Kondru,
R. K.; Lou, Y.; Lynch, S. M.; Owens, T. D.; Soth, M.; Yee, C. W.
Pyrrolo[2,3-b]pyrazine-7-carboxamide derivatives as JAK and SYK
inhibitors and their preparation and use for the treatment of
autoimmune and inflammatory diseases. Patent WO 2011144585,
2011; (b) Hendricks, R. T.; Hermann, J.; Kondru, R.; Lou, Y.; Lynch, S.
M.; Owens, T. D.; Soth, M. Preparation of pyrrolopyrazine as kinase
inhibitors useful for the treatment of autoimmune and inflammatory
diseases. U.S. Patent US 20110230462, 2011; (c) Bamberg, J. T.;
Bartlett, M.; Dubois, D. J.; Elworthy, T. R.; Hendricks, R. T.; Hermann,
J. C.; Kondru, R. K.; Lemoine, R.; Lou, Y.; Owens, T. D.; Park, J.; Smith,
D. B.; Soth, M.; Yang, H.; and Yee, C. W. Preparation of
pyrrolopyrazines as JAK and SYK inhibitors. U.S. Patent US
20090215750, 2009; (d) Dubois, D. J.; Hendricks, R. T.; Hermann, J.
C.; Kondru, R. K.; Lou, Y.; Owens, T. D.; Yee, C. W. Preparation of
ASSOCIATED CONTENT
* Supporting Information
Full experimental detail and characterization of all new synthetic
intermediates and final products along with experimental
procedures for the Syk biochemical and cell-based assays. This
material is available free of charge via the Internet at http://pubs.
acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*For F.P.: phone, (973) 235-7319; E-mail, fernando.padilla@
roche.com. For M.C.L.: phone, (408) 585-8051; E-mail,
mattlucas70@hotmail.com.
■
Present Addresses
§Cubist Pharmaceuticals, 65 Hayden Avenue, Lexington,
Massachusetts 02421, United States.
∥Principia BioPharma, Inc., 400 E. Jamie Court #302, South San
Francisco, California 94080, United States.
⊥School of Pharmaceutical Sciences, Soochow University, 199
Ren Ai Rd. Suzhou, Jiangsu, P.R.China.
Notes
The authors declare no competing financial interest.
1691
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692

Journal of Medicinal Chemistry
Article
pyrrolopyrazines as JAK and SYK inhibitors. U.S. Patent US
20090215724, 2009.
(12) Lynch, S. M.; DeVicente, J.; Hermann, J. C.; Jaime-Figueroa, S.;
Jin, S.; Kuglstatter, A.; Li, H.; Lovey, A.; Menke, J.; Niu, L.; Patel, V.;
Roy, D.; Soth, M.; Steiner, S.; Tivitmahaisoon, P.; Vu, M. D. and Yee, C.
Identification of Potent, Selective JAK3 Inhibitors through Structure
Based Design and Strategic use of Conformational Bias. Bioorg. Med.
Chem. Lett. submitted for publication.
(13) Molecular Operating Environment (MOE), 2011.10; Chemical
Computing Group Inc.: 1010 Sherbooke St. West, Suite #910, Montreal,
QC H3A 2R7, Canada, 2011.
(14) Garber, K. Pfizer’s JAK inhibitor sails through phase 3 in
rheumatoid arthritis. Nature Biotechnol. 2011, 29, 467.
(15) Cubeddu, L. X. Iatrogenic QT Abnormalities and Fatal
Arrhythmias: Mechanisms and Clinical Significance. Curr. Cardiol.
Rev. 2009, 5, 166−176.
(16) (a) Cumming, J. G.; Tucker, H.; Oldfield, J.; Fielding, C.;
Highton, A.; Faull, A.; Wild, M.; Brown, D.; Wells, S.; Shaw, J. Balancing
hERG affinity and absorption in the discovery of AZD5672, an orally
active CCR5 antagonist for the treatment of rheumatoid arthritis. Bioorg.
Med. Chem. Lett. 2012, 22, 1655−1659. (b) Shultz, M. D.; Cao, X.;
Chen, C. H.; Cho, Y. S.; Davis, N. R.; Eckman, J.; Fan, J.; Fekete, A.;
Firestone, B.; Flynn, J.; Green, J.; Growney, J. D.; Holmqvist, M.; Hsu,
M.; Jansson, D.; Jiang, L.; Kwon, P.; Liu, G.; Lombardo, F.; Lu, Q.;
Majumdar, D.; Meta, C.; Perez, L.; Pu, M.; Ramsey, T.; Remiszewski, S.;
Skolnik, S.; Traebert, M.; Urban, L.; Uttamsingh, V.; Wang, P.;
Whitebread, S.; Whitehead, L.; Yan-Neale, Y.; Yao, Y.-M.; Zhou, L.;
Atadja, P. Optimization of the in vitro cardiac safety of hydroxamate-
based histone deacetylase inhibitors. J. Med. Chem. 2011, 54, 4752−
4772. (c) Mendez-Andino, J. L.; Wos, J. A. MCH-R1 antagonists: what is
keeping most research programs away from the clinic? Drug Discovery
Today 2007, 12, 972. (d) Rokosz, L. L.; Hobbs, D. W. Biological
Examination of Melanin Concentrating Hormone Receptor 1: Multi-
tasking from the Hypothalamus. Drug News Perspect. 2006, 19, 273.
(17) Perry, M.; Stansfeld, P. J.; Leaney, J.; Wood, C.; de Groot, M. J.;
Leishman, D.; Sutcliffe, M. J.; Mitcheson, J. S. Drug binding interactions
in the inner cavity of HERG channels: molecular insights from
structure−activity relationships of clofilium and ibutilide analogs. Mol.
Pharmacol. 2006, 69, 509−519.
(18) Ritchie, T. J.; Macdonald, S. J. F.; Young, R. J.; Pickett, S. D. The
impact of aromatic ring count on compound developability: further
insights by examining carbo- and hetero-aromatic and aliphatic ring
types. Drug Discovery Today 2011, 16, 164−171.
(19) Villasenor, A. G.; Kondru, R.; Ho, H.; Wang, S.; Papp, E.; Shaw,
̃
D.; Barnett, J. W.; Browner, M. F.; Kuglstatter, A. Structural Insights for
Design of Potent Spleen Tyrosine Kinase Inhibitors from Crystallo-
graphic Analysis of Three Inhibitor Complexes. Chem. Biol. Drug Des.
2009, 73, 466−470.
(20) Rigaudy, J.; Klesney, D. P. Nomenclature in Organic Chemistry:
IUPAC; Pergamon Press: Oxford, 1979.
1692
dx.doi.org/10.1021/jm301720p | J. Med. Chem. 2013, 56, 1677−1692