Optimized protein kinase Cθ (PKCθ) inhibitors reveal only modest anti-inflammatory efficacy in a rodent model of arthritis

Article abstract of DOI:10.1021/jm5013006

We previously demonstrated that selective inhibition of protein kinase Cθ (PKCθ) with triazinone 1 resulted in dose-dependent reduction of paw swelling in a mouse model of arthritis.1,2 However, a high concentration was required for efficacy, thus providing only a minimal safety window. Herein we describe a strategy to deliver safer compounds based on the hypothesis that optimization of potency in concert with good oral pharmacokinetic (PK) properties would enable in vivo efficacy at reduced exposures, resulting in an improved safety window. Ultimately, transformation of 1 yielded analogues that demonstrated excellent potency and PK properties and fully inhibited IL-2 production in an acute model. In spite of good exposure, twice-a-day treatment with 17l in the glucose-6-phosphate isomerase chronic in vivo mouse model of arthritis yielded only moderate efficacy. On the basis of the exposure achieved, we conclude that PKCθ inhibition alone is insufficient for complete efficacy in this rodent arthritis model.

Full text of DOI:10.1021/jm5013006

Article
pubs.acs.org/jmc
Optimized Protein Kinase Cθ (PKCθ) Inhibitors Reveal Only Modest
Anti-inflammatory Efficacy in a Rodent Model of Arthritis
Dawn M. George,*,† Eric C. Breinlinger,† Maria A. Argiriadi,† Yang Zhang,∥ Jianfei Wang,∥
Pratima Bansal-Pakala,† David B. Duignan,† Prisca Honore,‡ QingYu Lang,§ Scott Mittelstadt,‡
Lian Rundell,† Annette Schwartz,† Jiakang Sun,∥ and Jeremy J. Edmunds†
†AbbVie Bioresearch Center, 381 Plantation Street, Worcester, Massachusetts 01605, United States
‡AbbVie, Inc., 1 North Waukegan Road, North Chicago, Illinois 60064, United States
§AbbVie China R&D Center, 5F, North Jin Chuang Building #1, 4560 Jinke Road, Pudong New District, Shanghai 201201, P. R.
China
∥WuXi AppTec (Shanghai) Co., Ltd., 288 Fute Zhong Road, Waigaoqiao Free Trade Zone, Shanghai 200131, P. R. China
S
* Supporting Information
ABSTRACT: We previously demonstrated that selective inhibition
of protein kinase Cθ (PKCθ) with triazinone 1 resulted in dose-
dependent reduction of paw swelling in a mouse model of
arthritis.1,2 However, a high concentration was required for efficacy,
thus providing only a minimal safety window. Herein we describe a
strategy to deliver safer compounds based on the hypothesis that
optimization of potency in concert with good oral pharmacokinetic
(PK) properties would enable in vivo efficacy at reduced exposures,
resulting in an improved safety window. Ultimately, transformation
of 1 yielded analogues that demonstrated excellent potency and PK
properties and fully inhibited IL-2 production in an acute model. In
spite of good exposure, twice-a-day treatment with 17l in the
glucose-6-phosphate isomerase chronic in vivo mouse model of arthritis yielded only moderate efficacy. On the basis of the
exposure achieved, we conclude that PKCθ inhibition alone is insufficient for complete efficacy in this rodent arthritis model.
remains to be determined. In our studies with 1, high
efficacious exposure and full coverage over the unbound
cellular EC80 were essential to drive efficacy in the chronic
study, and serious tolerability issues were encountered after
∼10 days of oral administration at doses only 3-fold higher than
that required for efficacy. Unable to identify a specific off-target
protein to which the toxicities could be attributed, we turned
our focus toward identifying compounds with reduced
efficacious exposures as an approach to potentially amplify
the therapeutic index. In this paper, we describe structure-based
optimization starting from compound 1 that led to the
discovery of compounds with improved cellular potency
coupled with desirable PK profiles. Progression of promising
compounds into a screening version of the acute in vivo Con A
model in mice enabled rapid selection of compounds for
advancement into the chronic in vivo GPI mouse model of
arthritis to test their efficacy and tolerability.
INTRODUCTION
■
Inhibition of protein kinase Cθ (PKCθ) blocks the activation of
T cells3−5 and thus is hypothesized to ameliorate T-cell-
mediated disease states such as rheumatoid arthritis (RA),
inflammatory bowel disease (IBD), and psoriasis. Multiple
companies have worked to identify suitable chemical matter to
inhibit this intriguing target that is predominately expressed in
T cells.6−17 In our previous paper, we detailed the discovery of
potent and selective PKCθ inhibitor triazinone 1.1 Compound
1 has moderate cellular potency (IC50 = 0.12 μM) and
permeability (Papp A to B = 2.1 × 10−6 cm/s), good oral
pharmacokinetics (PK) in mouse (CL = 1.5 L h−1 kg−1, t1/2
=
3.1 h; F = 108%), and efficacy in acute [concanavalin A (Con
A)-induced IL-2 production, ED80 = 29 mg/kg] and chronic
[glucose-6-phosphate isomerase (GPI)-induced arthritis,18,19
61% inhibition of paw swelling at 10 mg/kg and 100%
inhibition of paw swelling at 30 mg/kg when dosed in a full
prophylactic mode] in vivo mouse models. Although a
retrospective analysis of the utility of preclinical arthritis
models [rat adjuvant-induced arthritis (AIA), rat collagen-
induced arthritis (CIA), and mouse CIA] concluded that
efficacy with therapeutic treatment is more likely to predict
efficacy in human RA patients,20 the predictive value of the GPI
model, whether utilizing therapeutic or prophylactic dosing,
Special Issue: New Frontiers in Kinases
Received: August 26, 2014
Published: September 25, 2014
© 2014 American Chemical Society
333
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
a
Scheme 1
a
Reagents and conditions: (i) KHMDS, (S)-2-trifluoromethanesulfonyloxypropionic acid ethyl ester, THF, 0 °C to rt; (ii) Lawesson reagent,
toluene, 120 °C; (iii) NH2NH2·H2O, EtOH, rt.
a
Scheme 2
a
Reagents and conditions: (i) NH4Cl, Zn powder, MeOH, THF, 50 °C; (ii) 1-benzhydryl-3-methylazetidin-3-yl methanesulfonate, K2CO3, propan-
2-ol, 80 °C; (iii) nBu4NBr3, DCM, MeOH, rt; (iv) R1-phenylboronic acid or boronate ester, K2CO3, Pd(dppf)2Cl2, water, 1,4-dioxane, 80 °C; (v)
Pd(OH)2/C, HCl, MeOH, THF, H2 (55 psi), 50 °C; (vi) paraformaldehyde, MeOH, AcOH, reflux, then NaBH3CN, rt.
methylated by reductive amination to give analogues 11a and
11b.
CHEMISTRY
■
The synthesis of triazinone 1, prepared as a racemic mixture
followed by late-stage chiral separation, was described
previously.1,2 An improved synthesis was developed to
construct multifunctional triazinone intermediates 5a−c with
the R stereochemistry set at the 4-position (Scheme 1),
eliminating the need for a chiral separation step. Alkylation of
functionalized benzo[1,4]oxazinones 2a−c with (S)-2-trifluor-
omethanesulfonyloxypropionic acid ethyl ester [generated by
treatment of (S)-ethyl lactate with triflic anhydride] gave 3a−c.
Conversion to benzo[1,4]oxazinethiones 4a−c by treatment
with Lawesson reagent and subsequent cyclization with
hydrazine hydrate afforded versatile triazinone intermediates
5a−c. Bromide 5b could be further transformed to the
boronate ester (R1 = pinacolatoboron, R2 = H, 5d) by
treatment with bispinacolato(diboron). Elaboration of the
triazinone core to access 6-aminoazetidine-substituted ana-
logues proceeded as described in Scheme 2. Starting from 6-
nitro intermediate 5a, zinc-mediated reduction provided aniline
6, and reaction with a 3-methylazetidine methanesulfonate
yielded aminoazetidine 7. Conversion to bromide 8 followed by
Suzuki coupling with phenylboronic acids gave access to
compounds 9a and 9b. Palladium-catalyzed hydrogenation
yielded unmasked azetidines 10a and 10b, which were
A representative synthetic route to 6-piperidine-substituted
triazinones 17a−l is detailed in Scheme 3. From chiral 6-
bromide 5d, Suzuki cross-coupling with a tetrahydropyridin-4-
yl triflate and chiral separation gave 12. Subsequent hydro-
genation provided piperidine 13 stereoselectively. Treatment
with tetrabutylammonium bromide yielded bromide 14, which
underwent acid-mediated Boc deprotection to provide
piperidine 15. A penultimate methylation step yielded N-
methylpiperidine 16, and final Suzuki coupling afforded 7-aryl
substituted analogues 17a−l. Alternatively, the order of
functionalization steps may be reversed. Thus, Suzuki coupling
with bromide 14 followed by piperidine Boc deprotection and
methylation also allows access to 17.
7-Methyl-substituted analogue 18 (Figure 1) was prepared as
previously described.1,2 Methylated 6-aminoazetidine analogues
19−21 were prepared from 1 by sequential reductive amination
reactions (see the Supporting Information). Similarly, 7-Ph-
substituted analogue 22 was prepared by reductive amination of
azetidine 10a.
Preparation of 6-CH2-azetidine analogue 23 was accom-
plished by Suzuki coupling with 6-boronate ester 5d, reduction,
and further functionalization as detailed in the Supporting
334
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
a
Scheme 3
a
Reagents and conditions: (i) 3-methyl-4-trifluoromethansulfonyloxy-3,6-dihydro-2H-pyridine-1-carboxylic acid tert-butyl ester, K3PO4, Pd(Ph3P)4,
DMF, water, 80 °C, then chiral separation; (ii) Pd/C, H2 (1 atm), MeOH, rt; (iii) nBu4NBr3, DCM, MeOH, rt; (iv) HCl, EtOAc, rt; (v)
paraformaldehyde, NaBH3CN, MeOH, AcOH, rt; (vi) R1-boronic acid or boronate ester, K2CO3, PdCl2(dppf)−CH2Cl2, 1,4-dioxane, water, 90 °C.
Information. 6-Piperidine and 6-pyrrolidine analogues 24−26
were prepared from 7-bromide 5c via sequential Suzuki
coupling, bromination at the 6-position, a second Suzuki
coupling with the corresponding 5,6-dihydropyridine boronate
ester or 2,5-dihydropyrrole boronate ester, and reduction steps
(see the Supporting Information). Analogues 27−30 were
prepared using methods similar to those described above (see
the Supporting Information for details).
RESULTS AND DISCUSSION
■
Our previously reported PKCθ cocrystal structure with
compound 18 (2.6 Å, PDB accession code 4Q9Z)1 a structural
relative of analogue 1, confirmed the binding mode of the
triazinone core (Figure 1). The N−N−H of the triazinone ring
makes a hinge interaction with backbone residues Glu459 and
Leu461, the CO makes a water-mediated interaction with
Thr442 in the back pocket, and the azetidine NH tightly
contacts Asp522 and Asn509. Our initial design strategy was to
introduce small groups around the aminoazetidine fragment,
targeting occupation of lipophilic regions of the protein active
site to enhance potency while maintaining or improving on the
favorable PK profile observed with parent compound 1.
Methylation at either the anilino or azetidinyl nitrogen (19
or 20, respectively; Table 1) in the context of the 7-CF3 moiety
provided no improvement in potency. Notably, terminal
methylation to give tertiary amine 20 did significantly improve
the permeability (Papp A to B = 19 × 10−6 cm/s).
Polymethylation, as in 21, resulted in a significant deterioration
Figure 1. Key binding interactions of compound 18, generated using
in potency that could be explained by a potentially inefficient
ligand conformation that is unable to reach target residues
the ligand interaction tool in MOE.21
Asp522, Asn509, or Asp508. On the basis of our previous work,
we predicted that a 7-Ph moiety would improve potency but
the 7-Ph group, the introduction of a methyl group on the
could suffer from suboptimal PK. Accordingly, in the context of
azetidine nitrogen (22) was tolerated, similar to the activity
335
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
a
Table 1. Structure−Activity Relationships for Conformationally Constrained Variants
a
a
b
c
compd
R1
CF3
R2
R3
R4
PKCθ IC50 (μM)
IL-2 IC50 (μM)
permeability (10−6 cm/s)
hu/mu CLint,u (L h−1 kg−1
)
1
H
Me
Me
Me
Me
Me
Me
Me
H
0.023
0.059
0.057
0.70
0.12
0.49
0.37
3.0
2.1
−
19
<2.0/<5.7
6.4/46
−/−
6.0/39
5.0/17
15/68
19
20
21
22
11a
11b
CF3
CF3
CF3
Ph
Me
H
H
Me
Me
Me
Me
Me
Me
H
−
0.052
0.004
0.004
0.23
0.037
0.024
24
15
15
Ph
Me
Me
2-F-Ph
18/134
a
b
c
Means from n ≥ 2 experiments. MDCK-MDR1 with 40 μM CSA (pH 7.4) Papp A to B. In vivo predicted unbound CLint from human and mouse
microsome stability studies.
pattern observed with related 7-CF3-substituted triazinone 20.
Incorporation of an additional methyl group on the anilino-N
linker (11a) provided a significant improvement in cellular
activity, in stark contrast to 7-CF3-substituted analogue 21.
Ligand minimization calculations and torsion analysis with 11a
suggested that the combination of two neighboring methyl
groups on the aminoazetidine fragment and the presence of the
7-phenyl group result in an energetically favored minimum.
The resulting ligand conformation provides preferential access
to a different azetidine trajectory than observed with previous
analogues. On the basis of this hypothesis, the azetidine
nitrogen of 11a would be positioned with the basic amine
flipped “up” to target Asp508, whereas the azetidine nitrogen of
compound 22 is oriented “down” to interact with Asp522 and
Asn509 as seen with compound 18. A crystal structure of 11a in
PKCθ (2.61 Å, PDB accession code 4RA5; Figure 2a)
confirmed our hypothesis. The azetidine nitrogen has
interactions with both the backbone carbonyl and side chain
of Asp508. In comparison to our previously reported protein X-
ray structure of triazinone 18 bound to PKCθ,1 a protein
backbone shift (∼3.0 Å) is observed at the tip of the glycine-
rich loop (Figure 2b), most likely because of 7-phenyl
accommodation. Similar to observations from the previous
structure, an ordered C-terminal extension proximal to residues
Pro658−Leu669 is observed. A hydrophobic pocket formed as
a result of this protein ordering provides an optimal fit for the
7-phenyl substituent. The pocket is surrounded by aromatic
residues, including Phe391 (glycine-rich loop), Phe664 (C-
terminal extension), and Phe668 (C-terminal extension).
In spite of the excellent in vitro potency of 11a, in
comparison to parent molecule 1 this compound suffered
from higher microsomal clearance (mouse CLint,u = 68 L h−1
kg−1 compared with <5.7 L h−1 kg−1 for 1) and high in vivo
clearance (CL = 6.0 L h−1 kg−1 vs 1.5 L h−1 kg−1 for 1) in
mouse PK studies (Table 2). We suspected that demethylation
of the azetidine N-Me was occurring, but metabolite
identification (met ID) studies indicated that the major
metabolite was formed by aniline demethylation of 11a,
which was a potential cause of the increased CL. Although
the introduction of a 2-F group on the phenyl ring (11b) had a
minimal impact on potency (cell IC50 = 0.024 μM for 11b vs
0.037 μM for 11a) and resulted in a moderately higher
microsomal CLint,u (134 L h−1 kg−1 compared with 68 L h−1
kg−1 for 11a), it did afford improved in vivo clearance (CL =
Figure 2. X-ray crystal structure of 11a in PKCθ. Shown are (a) the X-
ray crystal structure of 11a (PDB accession code 4RA5) and (b) an
overlay of the X-ray crystal structures of 11a (blue) and 18 (gray) in
PKCθ.
3.1 L h−1 kg−1) and reasonable unbound oral exposure (AUC,u
= 102 ng h mL−1, 10 mg/kg dose). In contrast to 11a, minimal
aniline demethylation was observed in met ID studies in the
case of 11b, but significant demethylation of the azetidine N-
Me occurred. In mouse PK with 11b, exposure of the active
336
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
Table 2. Comparison of Mouse PK for Compounds with 6-N versus 6-C Linkers
a
b
c
Means from n ≥ 2 experiments. In vivo predicted unbound CLint from human and mouse microsome stability studies. Mouse PK reported (3 mg/
d
kg iv bolus in 50% PEG400; po dose of 10 mg/kg in 0.5% HPMC and 0.02% Tween80). Data are reported as averages of three animals. AUC of
desmethyl metabolite measured in the mouse PK study. The absolute stereochemistry at the C6 linker was not determined. The compound was not
tolerated with a 3 mg/kg iv bolus; the value shown is from a 1 mg/kg iv bolus. Not measured.
e
f
g
azetidine N-desmethyl metabolite (AUC,u = 83 ng h mL−1, 10
mg/kg dose) was ∼80% that of the parent (102 ng h mL−1).
Because of the reasonable clearance observed with 7-(2-
fluorophenyl)-substituted 11b, we chose to incorporate this
moiety in subsequent analogues while exploring alternative
linkers at the 6-position. In order to prevent potential
demethylation at the aniline position, the two isomeric
analogues with a methylene linker (exemplified by compound
23; Table 2) were prepared. Of the two isomers generated, one
was >250-fold more potent than the other (PKCθ IC50 = 0.003
and 0.77 μM; the absolute stereochemistry was not
determined). The linker change to CHMe from NMe was
tolerated in terms of potency, but no improvement in in vivo
clearance compared to 11b (CL = 3.5 L h−1 kg−1 for 23; CL =
3.1 L h−1 kg−1 for 11b) or bioavailability (F = 29 and 28%,
respectively) was observed. Considering the increasingly
complex synthetic route and purification required to access
analogues like 23 but encouraged by the promise of 6-C-linked
analogues, we turned our attention to conformationally
constrained pyrrolidine or piperidine moieties at the 6-position
(24−26; Table 2). This approach initially yielded a moderate
(∼10-fold) reduction in potency, with all three analogues
administered, 3 of 3 mice died rapidly. Curiously, in spite of
comparable exposure with iv dosing (data not shown), isomer
24 presented no obvious adverse events at 3 mg/kg. We
initially hypothesized that the S isomer 25 may inhibit
unidentified off-target kinases or receptors that could be
responsible for the observed severe adverse events. However,
off-target screening panel data failed to support this theory. For
compound 25, an internal panel of 19 receptors (see the
Supporting Information) identified no hits (EC50 > 10 μM for
all), and of 78 kinases tested (see the Supporting Information),
only three resulted in <100-fold selectivity compared with
PKCθ (Clk2, 6-fold; Cdk8, 6-fold; PKA, 19-fold).
In light of the unexplained toxicity observed with 25, we
turned our efforts toward the equipotent piperidine 26. Because
we had observed a significant amount of azetidine demethy-
lation in our studies with methyl azetidine 11b, we began
routinely monitoring for both parent and desmethyl metabolite
exposure in in vivo mouse PK studies. Typically the desmethyl
metabolite and parent molecules had similar in vitro potency, as
exemplified by 26 [PKCθ IC50 = 0.043 μM (parent) and 0.019
μM (metabolite); cellular IC50 = 0.20 μM (parent) and 0.19
μM (metabolite)], but the metabolite suffered from signifi-
cantly reduced permeability [MDCK-MDR1 + CSA (pH 7.4)
Papp A to B = 19 × 10−6 cm/s (parent) and 1.2 × 10−6 cm/s
(metabolite)]. In general, a minimal shift in activity (∼1- to 10-
fold) between enzyme and cell assays were observed as long as
the permeability remained greater than ∼1 × 10−6 cm/s. In
mouse PK, piperidine 26 exhibited low clearance (CL = 1.5 L
h−1 kg−1), resulting in a reasonable half-life (t1/2 = 5.1 h). Oral
bioavailability was good at 54%. Unbound oral AUC was 401
ng h mL−1 (10 mg/kg dose), and approximately 25% of the
desmethyl metabolite (AUC,u = 56 ng h mL−1) was observed.
displaying nearly identical profiles in PKCθ enzyme (IC50
=
0.041−0.043 μM) and cellular (IC50 = 0.20−0.21 μM) assays
and acceptable mouse microsomal clearance (CLint,u = 30−38 L
h−1 kg−1). Mouse PK studies revealed an interesting distinction
between the two pyrrolidine isomers, with the R isomer 24
exhibiting substantially improved oral exposure (AUC,u = 920
ng h mL−1) compared with the S isomer 25 (AUC,u = 120 ng h
mL−1). In the PK study with pyrrolidine 25, a steep tolerability
threshold was observed. A 1 mg/kg iv dose was tolerated with
no overt issues, but when a 3 mg/kg iv dose of 25 was
337
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
Consistent with good exposure, 26 inhibited IL-2 production in
the Con A model (52 4%, 10 mg/kg), on par with compound
1 (67 3%, 10 mg/kg). Thus, the piperidine was established as
a reasonable lead from which to improve potency in the context
of favorable PK. An overlay of the polymethylated azetidine
analogue (11a) with piperidine 26 highlighted the prospect of
introducing methyl groups around the piperidine ring in order
to occupy the van der Waals (VDW) space associated with the
glycine-rich loop or toward the pre-DFG region (Figure 3) as a
potency (PKCθ IC50 = 0.003 μM; IL-2 cell IC50 = 0.021 μM).
Compound 17a and desmethyl metabolite 27 had comparable
activity, clearance, and unbound oral exposure (albeit ∼2-fold
lower relative to parent 26) and adequate half-life (Table 3).
Approximately 10% of metabolite 27 was observed upon oral
dosing of 17a. In spite of similar in vitro and pharmacokinetic
profiles, 17a was more active (85 3% inhibition, 10 mg/kg
dose) than 27 (45
7% inhibition, 10 mg/kg dose) in our
screening Con A model, and thus, future analogues focused on
tertiary amines. The cis stereoisomer of 17a, compound 28,
also had compelling in vitro potency (IL-2 cell IC50 = 0.055
μM), notably higher oral exposure [AUC,u = 473 ng h mL−1
(parent) and 204 ng h mL−1 (metabolite)], and similar activity
in the Con A model (85
2% inhibition, 10 mg/kg dose).
Dimethylation at the piperidine 3-position (29) was tolerated
but provided no obvious advantage in terms of potency or PK.
Piperidine N-ethylated derivative 30, prepared in an attempt to
fully block terminal dealkylation, yielded only moderately
reduced in vivo clearance compared with parent 17a (CL = 1.9
vs 3.3 L h−1 kg−1) and no change in the amount of circulating
dealkylated metabolite (∼11%).
A crystal structure in surrogate kinase PKCα (PDB accession
code 4RA4) confirmed that the triazinone binding pose of 28 is
similar to that of compound 11a (Figure 4). As predicted, the
piperidine N is in close contact with Asp481 (Asp522 in
PKCθ). The cis-3-methyl substitution on the piperidine mimics
the 3-methyl moiety on the azetidine in 11a. In agreement with
a lipophilic pocket identified in the crystal structure of 28 and
consistent with our previous experience, a wide variety of
substituents around the 7-aryl substituent were tolerated for in
vitro potency (17a−l; Table 4). The C7 substituent therefore
provided a convenient handle to modulate the PK properties.
Figure 3. Overlay of the crystal structure of 11a (blue) with a model
of 26 (gray) in PKCθ, highlighting the lipophilic region for targeting
with substitution on the 6-piperidine.
potential approach to improve potency. Cis methylation at the
3-position of the piperidine as in 17a (the absolute structure
was confirmed by small molecule X-ray analysis; see the
Supporting Information for details) afforded desirable in vitro
Table 3. Structure−Activity Relationships for Piperidine Ring Substitution
a
b
c
Means from n ≥ 2 experiments. In vivo predicted unbound CLint from mouse and human microsome stability studies. Mouse PK reported (3 mg/
d
kg iv bolus in 50% PEG400; po dose of 10 mg/kg in 0.5% HPMC and 0.02% Tween80). Data are reported as averages of three animals. AUC of
desmethyl metabolite measured in the mouse PK study. Compound was dosed at 10 mg/kg in mice 2 h prior to Con A challenge. Readout is %
inhibition of IL-2 production. The absolute stereochemistry at the 4-position of the piperidine was not determined but is proposed to be as drawn;
e
f
the stereoisomeric compound had PKCθ IC50 = 0.19 μM.
338
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
with moderate Cmax,u values (17 and 91 ng/mL, respectively).
The 2-CH3 and 3-OCH3 analogues (17d and 17f), while
tolerated in terms of PKCθ activity, had markedly shorter half-
lives (3.3 and 1.6 h, respectively). Incorporation of a pyridyl
ring, as in the crossover 5-(2-methoxy)pyridyl analogue (not
shown), gave potency comparable to that of 4-methoxyphenyl
derivative 17g but with a higher Cmax,u (602 vs 91 ng/mL; data
not shown) that was outside of our target criterion. The 3,4-
dimethoxy analogue (17i) exhibited in a ∼10-fold loss in
potency (IL-2 IC50 = 0.24 μM), while the analogous
ethylenedioxy derivative 17j retained good in vitro activity
(IL-2 IC50 = 0.024 μM); apparently, the steric limitation of the
3,4-dimethoxy substitution could be resolved by conformational
constraint (as in 17j). However, this compound did not meet
our target t1/2 criterion with a moderate half-life (t1/2 = 3.2 h).
Combination of 2-F substitution with either the preferred 4-
CH3 or 4-OCH3 group (17k and 17l, respectively) provided
excellent cellular potency (IL-2 cell IC50 = 0.010−0.014 μM) as
well as a desirable half-life and low Cmax (t1/2 = 4.8 h and Cmax,u
= 78 ng/mL for 17k; t1/2 = 11.6 h and Cmax,u = 113 ng/mL for
17l). Both 17k and 17l inhibited IL-2 production in the Con A
model (85 3% and 86 6%, respectively, 10 mg/kg dose).
Incorporation of two fluorine groups, as in 2,6-difluoro
analogue 17h, also gave excellent in vitro potency and activity
in the acute Con A model (84 3% inhibition of IL-2, 10 mg/
kg dose). However, in a Con A dose−response study (at 1, 3,
10, and 30 mg/kg po), mortality occurred rapidly in several
mice in the 30 mg/kg group, while tolerability issues in the 10
mg/kg group were not apparent. Although similar adverse
events had been noted with pyrrolidine isomer 25 at a 3 mg/kg
iv dose, this was the first instance where a single oral dose
resulted in such a rapid and severe outcome. Because of the
speed and severity of the effect, a profound impact on the
cardiovascular system was suspected. In order to assess this
hypothesis, compound 17h was characterized in a conscious
Figure 4. X-ray crystal structure of 28 in PKCα (PDB accession code
4RA4).
Previously we demonstrated that a drop in mean arterial
pressure in rat and dog CV studies was observed with earlier
analogues.1 Our goal was to minimize the Cmax/Cmin ratio while
achieving extended coverage over the unbound cellular EC80 in
our chronic in vivo efficacy model by identifying compounds
with a long half-life (t1/2 > 5 h) and low unbound Cmax (Cmax,u
<
200 ng/mL). In PK studies, we observed that analogues of 17a
generally produced <10% of the corresponding desmethyl
metabolites. Thus, in order to more rapidly prioritize
compounds, a PK screening protocol (3 mg/kg iv bolus and
po dose of 10 mg/kg) was utilized that eliminated metabolite
measurement (Table 4). Fluorination of the aryl ring (17a−c)
was tolerated, with a slight potency preference for 2-F
substitution (17a), which gave a reasonable half-life (t1/2
=
3.8 h) and Cmax,u (55 ng/mL). Substitution at the 4-position
with either CH3 (17e) or OCH3 (17g) provided good in vitro
potency and protracted half-life (8.5 and 9.6 h, respectively)
Table 4. Structure−Activity Relationships of 7-Phenyl Substituents
b
mouse PK
a
a
c
compd
R1
PKCθ IC50 (μM)
IL-2 IC50 (μM)
t1/2 (h)
Cmax,u (ng/mL)
AUC,u (ng h mL−1
)
Con A (% inhibition)
17a
17b
17c
17d
17e
17f
17g
17h
17i
2-F
0.003
0.014
0.019
0.008
0.024
0.007
0.010
0.002
0.240
0.005
0.002
0.002
0.022
0.060
0.080
0.040
0.082
0.049
0.041
0.013
−
3.8
−
−
3.3
8.5
1.6
9.6
−
54.6
−
−
120
17.4
39.8
90.8
−
155
−
−
924
324
73.9
1290
−
85
−
−
74
49
69
82
3-F
4-F
2-CH3
4-CH3
3-OCH3
4-OCH3
d
2,6-di-F
84
3,4-di-OCH3
3,4-OCH2CH2O-
2-F-4-CH3
2-F-4-OCH3
−
−
−
−
17j
0.024
0.010
0.014
3.2
4.8
11.6
96.2
78.0
113
856
1020
1400
68
85
86
17k
17l
a
b
Means from n ≥ 2 experiments. Mouse PK reported (3 mg/kg iv bolus in 50% PEG400; po dose of 10 mg/kg in 0.5% HPMC and 0.02%
c
Tween80). Data are reported as averages of three animals. Compound was dosed at 10 mg/kg in mice 2 h prior to Con A challenge. Readout is %
inhibition of IL-2 production. During a Con A dose−response experiment, 3 of 15 mice died rapidly when given a 30 mg/kg po dose; the 10 mg/kg
po dose was tolerated without any overt adverse signs.
d
339
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
mouse CV model (see the Supporting Information for protocol
details). Oral administration of 17h to telemetrized mice at 30
mg/kg produced a significant decrease in mean arterial pressure
(MAP) to a maximum effect of −22 mmHg at the 2 h time
point (Figure 5), which could help to explain the mortality
observed in mice at 30 mg/kg in the Con A experiment
Figure 5. Conscious mouse CV data with 17h. Blue dots: compound
17h, po dose of 30 mg/kg in 0.5% HPMC and 0.02% Tween80 (n =
5). Black dots: vehicle (n = 5). Gray bar shows when compound 17h
was dosed.
Four compounds that did not display adverse effects (17a,
17k, 17l, and 28) were selected for profiling in our internal
kinome screening panel (Figure 6). Although the compounds
displayed good overall selectivity, inhibition of PKA, Clk2, and
Cdk8 was generally observed with these analogues bearing a 7-
aryl substituent. When profiled in an internal bioprofiling panel
that studies the functional effect of a compound on 19
receptors (see the Supporting Information for the receptors
tested), compound 28 did not interact with any of the receptors
at maximum concentrations of 10 μM. However, 17a and 17k
both inhibited the L-type calcium channel (IC50 = 5.87 and 6.8
μM, respectively) and were agonists of the AT1 receptor (EC50
= 5.1 and 5.5 μM, respectively). In addition, 17k and 17l were
evaluated in a battery of radioligand binding assays containing
representatives of most G-protein-coupled receptors and a set
of ligand and voltage-gated ion channel binding sites (CEREP;
see the Supporting Information for the receptors tested). In
addition to the receptors identified from internal testing,
moderate interaction (50−75% inhibition at 10 μM) was
observed with 17k at the alpha1 receptor, M5 receptor, and the
Na2+ current, and moderate interaction was observed with 17l
at the NK2 receptor and Na2+ current. Since all of the
interactions with additional receptors were observed at
concentrations greater than ∼250-fold the PKCθ enzyme
potency, the compounds were considered selective.
Figure 6. Comparison of kinome selectivity profiles of 1, 17a, 17k,
17l, and 28. The kinome selectivity data are colored by fold selectivity
over PKCθ IC50, with ≤10-fold selective in red and a color gradient to
≥100-fold selective in green. Gray = not tested. Kinases with <40-fold
selectivity are delineated on the right-hand side.
Figure 7. Acute in vivo activity (mouse Con A-induced IL-2
production) with 17l. Mice were predosed orally with compound
17l or 3 mg/kg cyclosporin A (CSA) 2 h prior to challenge. At 2 h
postchallenge, serum was analyzed for murine IL-2 levels. Statistical
significance was determined by one-way ANOVA using Dunnett’s
post-test.
With compelling in vitro activity, an acceptable selectivity
profile, and favorable PK profile, compound 17l was selected
for further profiling in the Con A model, giving dose responsive
inhibition of IL-2 production [ED50 = 1.9 mg/kg (95% CI,
1.2−3.0 mg/kg); ED80 = 9.4 mg/kg (95% CI, 4.7−18.8 mg/
kg)] (Figure 7). By comparison, a 10 mg/kg dose of 17a was
exposure of 17l was about 2.5-fold lower (66 ng/mL, 10 mg/kg
dose). At t = 2 h after Con A administration (3 h after dosing),
the exposure level of 17a dropped significantly (total = 135 ng/
mL; unbound = 20 ng/mL) while the exposure level of 17l (10
mg/kg dose) was maintained (total = 664 ng/mL; unbound =
similarly effective, inhibiting IL-2 production at 85
3%.
Whereas 17a unbound oral exposure was 153 ng/mL (1 h after
dosing, just prior to Con A administration), the unbound
340
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
Figure 8. Chronic in vivo efficacy data (mouse GPI-induced arthritis model) and exposure data with prophylactic dosing of 17a. GPI-immunized
mice were dosed starting on day 0 with 17a or with vehicle orally twice a day. Dexamethasone (Dex) was dosed orally starting on day 7. Statistical
significance was determined by two-way ANOVA using a Bonferroni post-test. Terminal plasma exposures from the mouse GPI study were
measured after the final dose on day 17 at 0, 0.25, 0.5, 1, 3, 7, and 12 h.
Figure 9. Chronic in vivo efficacy data (mouse GPI-induced arthritis model) and exposure data with prophylactic dosing of 17l. GPI-immunized
mice were dosed starting on day 0 with 17l or with vehicle orally twice a day. Dexamethasone (Dex) was dosed orally starting on day 7. Statistical
significance was determined by two-way ANOVA using a Bonferroni post-test. The 30 mg/kg group was terminated on day 4 because of severe side
effects. Terminal plasma exposures from the mouse GPI study were measured after the final dose on day 17 at 0, 0.25, 0.5, 1, 3, 7, and 12 h. The 12 h
time point for 17l is not reported because of an apparent sample mixup.
111 ng/mL). Because of the comparable activities in spite of
different exposure profiles and the short nature of the Con A
model, these results suggest that activity in the acute model is
Cmax-driven.
To compare the effects of Cmax versus time-over-target
coverage in our chronic model, we hypothesized that 12 h
coverage of the PKCθ unbound cellular IC80 would result in
maximal efficacy. PK modeling with 17l predicted that a 10
341
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
mg/kg BID dose would fully cover the unbound cellular IC80 in
the GPI model yet also maintain a minimal Cmax, whereas a 30
mg/kg dose of 17a was predicted to give ∼6 h of coverage over
the unbound cellular IC80. Treatment with 17a in the GPI
exposure with a low Cmax/Cmin ratio, as exemplified by
compound 17l. When advanced in vivo, in spite of full
coverage over the unbound cellular IC80, compound 17l
showed only marginal efficacy in the mouse GPI model, lending
credence to the argument that inhibition of PKCθ alone may be
insufficient for efficacy in this arthritis model. Multiple
compounds (including 17l) from the triazinone chemotype
caused severe side effects, including death, in in vivo studies.
Unable to definitively attribute these effects to a particular off-
target kinase or receptor, work on this chemical matter was
halted. Challenged with the inability to rationally design away
from the severe toxicity effects combined with a lack of
confidence in PKCθ as a target for arthritis, the PKCθ program
effort was halted.
model resulted in a full dose response, giving 77
2%
inhibition of paw swelling at the top dose (30 mg/kg).
Coverage over the unbound cellular IC80 lasted for approx-
imately 7 h. Similarly, administration of 17l in the GPI model at
10 mg/kg gave extended coverage over the unbound cellular
IC80, but only partial efficacy (43% inhibition of paw swelling)
was observed. With 17l, the 30 mg/kg dose group had to be
terminated on day 4; one animal was found dead and other
mice were exhibiting convulsive behavior and were otherwise
lethargic. However, no tolerability issue was seen with the 30
mg/kg cohort in the experiment with 17a. On the basis of the
exposure data with 17l, simple time-over-target coverage of
PKCθ appears insufficient for efficacy in this model of arthritis.
Interestingly, the Cmax with a 30 mg/kg dose of 17a (Figure 8)
was similar to that of the 10 mg/kg dose of 17l (Figure 9), but
at these doses 17a was significantly more efficacious than 17l.
PKCα is known to play a role in T cell proliferation and
differentiation, and thus, it could be expected to contribute to
efficacy. Since inhibition of PKCα has been associated with
increases in contractility as discussed in our previous paper,1 we
actively targeted minimal PKCα activity with our compounds in
order to avoid introducing additional cardiovascular risk. In
vitro, 17a and 17l inhibit PKCα to a comparable degree (IC50 =
0.31 μM for 17a and 0.45 μM for 17l), resulting in 90- and
200-fold selectivity margins, respectively. Thus, we cannot
attribute the superior efficacy of 17a to the contribution of
PKCα inhibition. Other analogues tested in the GPI model
demonstrated only partial efficacy, similar to 17l (data not
shown), in spite of compound levels sufficient for target
coverage. We concluded that PKCθ inhibition alone is
inadequate to give full efficacy in this chronic arthritis model,
but we are still unable to rationalize the two exceptions
(compounds 1 and 17a) that resulted in full efficacy in the
model.
In order to understand the genesis of the severe and rapid
adverse events that were observed with compounds 1, 17h, 17l,
and 25, the apparently well-tolerated compound 17a was
advanced to dose-escalating mouse PK (100 and 300 mg/kg
doses) in preparation for a mouse toxicology study. Following a
single dose of 100 or 300 mg/kg in the PK study, mice
exhibited seizurelike activity, and all of the mice were found
dead or were euthanized (within ∼30 min at 300 mg/kg and
1−3 h at 100 mg/kg). Upon demonstration of sufficient
exposure above the cellular IC80 with multiple compounds
(exemplified by 17l) that resulted in only partial efficacy in the
GPI model and for lack of better understanding of the
unpredictable nature of the adverse events observed with
multiple triazinones, work with this chemotype on the PKCθ
project targeting arthritic indications was halted.
MATERIALS AND METHODS
■
Chemistry. All reagents and anhydrous solvents were of the
highest grade available and were purchased from commercial sources
and used without further purification or distillation, unless otherwise
noted. All NMR spectra were recorded on Bruker Avance III 400 MHz
FT-NMR spectrometer with a 5 mm BBO(F) probe or a Varian 400
1
MHz spectrometer with an H/19F/31P/13C 5 mm PFG 4Nuc probe.
1H chemical shifts (δ) are reported in parts per million with MeOH-d3,
DMSO-d6, or CDCl3 as the reference standard. Data are reported as
follows: chemical shift (multiplicity, coupling constant in hertz,
integration). Multiplicities are denoted as follows: s = singlet, d =
doublet, t = triplet, q = quartet, br = broad, m = multiplet. Analytical
LC−MS analyses were conducted using Shimadzu LC-20AB pumps
and an SPD-M20 PDA detector set at 220 and 254 nm, and the MS
detection was performed with a MS-2010EV Micromass Platform LC
spectrometer in electrospray ionization mode. Unless alternative
column and eluting conditions are detailed in the experiment, the LC−
MS method used was as follows. Column: Ultimate XB-C18, 3 μm, 30
mm × 2.1 mm. Mobile phase: acetonitrile (0.02% TFA) in water
(0.04% TFA) from 10% to 80% within 4 min. Flow rate: 1.2 mL/min.
Wavelength: 220 nm. Analytical SFC-MS analyses were conducted
using Mettler pumps and an Agilent G1315B detector set at 190−370
nm, and the MS detection was performed with an Agilent 6110
detector following the detailed column and eluting conditions
specified below. Preparative reversed-phase (RP) HPLC was
performed using a Gilson 322 pump, a Gilson 156 UV detector set
at 220 and 254 nm, and a Gilson GX-281 liquid handler, following the
detailed column and eluting conditions specified below. Normal-phase
silica gel preparative purification was performed using an automated
Combiflash companion from ISCO with prepacked silica gel cartridges
supplied by Santai Technologies Inc. and Agela Inc. SFC separation
was performed using a Berger SFC Analytix system following the
detailed column and eluting conditions specified below. Microwave
reactions were performed using a Biotage INITIATOR60EXP
microwave synthesis system. Qualitative ( ) optical rotation data
were collected using a PDR-chiral advanced laser in-line polarimeter
(model ALP2002). Analytical thin-layer chromatography (TLC) was
performed using SGF254 TLC plates (0.2−0.25 mm) supplied by
Combinol Reagent (Yantai) Co., Ltd. Preparative TLC was performed
using SGF254 TLC plates (0.4−0.5 mm) supplied by Yucheng
Chemical (Shanghai) Co., Ltd. The purities of the final products were
≥95% as established by analytical liquid chromatography using a PDA
1
detector at 220 nm and confirmed by H NMR spectroscopy.
CONCLUSION
3-Methyl-4-trifluoromethanesulfonyloxy-3,6-dihydro-2H-
pyridine-1-carboxylic Acid tert-Butyl Ester. To a mixture of 3-
methyl-4-oxopiperidine-1-carboxylic acid tert-butyl ester (0.4 g, 1.876
mmol) in THF (5 mL) was added KHMDS (1 M in THF, 2.81 mL,
2.81 mmol) dropwise at −78 °C, and the reaction mixture was stirred
for 30 min, allowed to warm to ambient temperature, and stirred for 3
h. The reaction mixture was cooled at −78 °C, and a solution of N-(5-
chloropyridin-2-yl)-1,1,1-trifluoro-N-((trifluoromethyl)sulfonyl)-
methanesulfonamide (0.884 g, 2.251 mmol) in THF (5 mL) was
added dropwise. The reaction mixture was allowed to warm to
■
The cocrystal structures of compound 11a with PKCθ and the
surrogate structure of compound 28 in PKCα guided a design-
based effort to optimize the triazinone chemotype targeting
reduced efficacious exposure through a combination of
improved potency and optimal PK profile. A screening Con
A model enabled the rapid prioritization of multiple analogues,
ultimately leading to the identification of PKCθ inhibitors with
improved cellular activity (IC50 < 0.025 μM) and good oral
342
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
ambient temperature and stirred overnight. Saturated aqueous NH4Cl
(20 mL) was added, and the reaction mixture was extracted with
EtOAc (3 × 10 mL). The combined organic layers were washed with
brine (20 mL), dried over Na2SO4, and concentrated in vacuo. The
residue was purified by column chromatography on silica gel (eluting
with 1−15% EtOAc in petroleum ether) to give 3-methyl-4-
trifluoromethanesulfonyloxy-3,6-dihydro-2H-pyridine-1-carboxylic
acid tert-butyl ester (0.31 g, 47%) as a colorless oil. TLC: Rf = 0.70
yellow solid. The filtrate was concentrated to give an additional 70 g of
crude product, which was purified by chromatography on silica gel
(eluting with 10−25% EtOAc in petroleum ether) to give a second
batch of (R)-6-bromo-4-methyl-2,10-dihydro-9-oxa-1,2,4a-triazaphe-
nanthren-3-one (45.9 g, 71%) as a yellow solid. 1H NMR (400
MHz, CDCl3): δ 8.53 (br s, 1H), 7.10 (d, J = 8.4 Hz, 1H), 7.03 (s,
1H), 6.89 (d, J = 8.4 Hz, 1H), 4.69−4.52 (m, 3H), 1.52 (d, J = 6.8 Hz,
3H). LC−MS (column, Ultimate XB-C18, 3 μm, 30 mm × 2.1 mm;
mobile phase, acetonitrile (0.02% TFA) in water (0.04% TFA) from
5% to 95% within 1.5 min; flow rate, 1.2 mL/min; wavelength, 220
nm): tR = 0.745 min. MS: m/z 297 [M + H]+. SFC (column,
Chiralpak AS-H 250 mm × 4.6 mm i.d., 5 μm; mobile phase, EtOH
(0.05% DEA) in supercritical CO2 from 5% to 40%; flow rate, 2.35
mL/min; wavelength, 220 nm): tR = 6.075 min. [α]D20 −314.92 (c =
0.002 g/mL in MeOH).
1
(eluting with 10% EtOAc in petroleum ether). H NMR (400 MHz,
CDCl3): δ 5.73 (br s, 1H), 4.24−3.89 (m, 2H), 3.76−3.31 (m, 2H),
2.63 (br s, 1H), 1.48 (s, 9H), 1.16 (d, J = 7.0 Hz, 3H).
(R)-6-((3R,4R)-1,3-Dimethylpiperidin-4-yl)-7-(2-fluorophen-
yl)-4-methyl-2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-
one Hydrochloride (17a). Step A. To a mixture of 2-amino-4-
bromophenol (50 g, 266 mmol) and NaHCO3 (67.0 g, 798 mmol) in
water (200 mL) and DME (200 mL) was added 2-chloroacetyl
chloride (45.1 g, 399 mmol) at 0 °C. The reaction mixture was stirred
at 15 °C for 30 min and then heated at 80 °C overnight. The reaction
mixture was cooled to ambient temperature and quenched with water
(400 mL). The precipitate was collected by suction filtration, washed
with water (200 mL) and petroleum ether (300 mL), and dried to give
6-bromo-4H-benzo[1,4]oxazin-3-one (59 g, 97%) as a gray solid.
Step E. To a mixture of 3-methyl-4-trifluoromethanesulfonyloxy-
3,6-dihydro-2H-pyridine-1-carboxylic acid tert-butyl ester (5.07 g,
14.69 mmol), (R)-4-methyl-6-(4,4,5,5-tetramethyl[1,3,2]-
dioxaborolan-2-yl)-2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-
one2 (2.8 g, 8.16 mmol), and K3PO4 (3.46 g, 16.32 mmol) in DMF
(60 mL) and water (10 mL) was added Pd(Ph3P)4 (3.77 g, 3.26
mmol). The reaction mixture was stirred at 80 °C overnight and then
cooled to ambient temperature and concentrated in vacuo. The
residue was purified by column chromatography on silica gel (eluting
with 3−50% EtOAc in petroleum ether) to give a yellow liquid (3.5 g).
Further purification by preparative HPLC (column, Agella Venusil
ASB 150 mm × 25 mm, 5 μm; mobile phase, acetonitrile in water
(0.05% HCl) from 50% to 75% in 12 min; flow rate, 30 mL/min;
wavelength, 220 nm) gave 3-methyl-4-((R)-4-methyl-3-oxo-2,3,4,10-
tetrahydro-9-oxa-1,2,4a-triazaphenanthren-6-yl)-3,6-dihydro-2H-pyri-
dine-1-carboxylic acid tert-butyl ester (1.5 g, 44%) as a white solid. 1H
NMR (400 MHz, CDCl3): δ 8.03 (br s, 1H), 7.02−6.93 (m, 2H), 6.87
(s, 1H), 5.81 (br s, 1H), 4.77 (q, J = 6.8 Hz, 1H), 4.59 (d, J = 13.2 Hz,
2H), 4.47−4.18 (m, 1H), 3.95−3.77 (m, 2H), 3.32 (d, J = 13.2 Hz,
1H), 2.79 (br s, 1H), 1.53 (m, 3H), 1.51 (s, 9H), 1.02 (d, J = 6.8 Hz,
3H). LC−MS (column, Ultimate XB-C18, 3 μm, 30 mm × 2.1 mm;
mobile phase, MeCN (0.02% TFA) in water (0.04% TFA) from 10%
to 80% within 7 min; flow rate, 1.2 mL/min; wavelength, 220 nm): tR
= 3.339 min. MS: m/z 413 [M + H]+. The isomers were separated by
chiral SFC (column, Chiralpak AD 300 mm × 50 mm i.d., 10 μm;
mobile phase, supercritical CO2/EtOH (0.1% NH3·H2O) = 60/40;
flow rate, 220 mL/min; wavelength, 220 nm) to give (R)-3-methyl-4-
((R)-4-methyl-3-oxo-2,3,4,10-tetrahydro-9-oxa-1,2,4a-triazaphenanth-
ren-6-yl)-3,6-dihydro-2H-pyridine-1-carboxylic acid tert-butyl ester
(isomer 1) as a white solid (0.57 g, 38%; SFC (column, Chiralcel
OJ-H 250 mm × 4.6 mm i.d., 5 μm; mobile phase, MeOH (0.05%
DEA)/supercritical CO2 from 5% to 40%; flow rate, 2.35 mL/min;
wavelength, 220 nm): tR = 8.56 min). 1H NMR (400 MHz, CDCl3): δ
8.03 (br s, 1H), 6.99−6.94 (m, 2H), 6.86 (s, 1H), 5.81 (d, J = 15.6 Hz,
1H), 4.77 (q, J = 6.6 Hz, 1H), 4.67−4.53 (m, 2H), 3.85 (dd, J = 3.3,
13.1 Hz, 2H), 3.31 (dd, J = 3.9, 12.9 Hz, 1H), 2.79 (br s, 1H), 1.55 (d,
J = 6.8 Hz, 3H), 1.51 (s, 9H), 1.02 (d, J = 6.8 Hz, 3H). LC−MS
(column, Ultimate XB-C18, 3 μm, 30 mm × 2.1 mm; mobile phase,
acetonitrile (0.02% TFA) in water (0.04% TFA) from 5% to 95%
1
TLC: Rf = 0.24 (eluting with 30% EtOAc in petroleum ether). H
NMR (DMSO-d6, 400 MHz): δ 7.07 (d, J = 8.4 Hz, 1H), 7.02 (s, 1H),
6.90 (d, J = 8.4 Hz, 1H), 4.59 (s, 2H).
Step B. To a mixture of 6-bromo-4H-benzo[1,4]oxazin-3-one (59 g,
259 mmol) in THF (800 mL) was added KHMDS (1 M in THF, 259
mL, 259 mmol) dropwise at 0 °C, and the reaction mixture was stirred
for 30 min. (S)-Ethyl 2-(((trifluoromethyl)sulfonyl)oxy)propanoate1
(117 g, 466 mmol) was added. The reaction mixture was warmed to
ambient temperature and stirred for 2 h. The reaction mixture was
quenched by the addition of water (1000 mL). The organic phase was
separated, and the aqueous solution was extracted with EtOAc (4 ×
500 mL). The combined organic phases were washed with brine (1 L),
dried over Na2SO4, and concentrated in vacuo. The residue was
purified by column chromatography on silica gel (eluting with 5−10%
EtOAc in petroleum ether) to give (R)-2-(6-bromo-3-oxo-2,3-
dihydrobenzo[1,4]oxazin-4-yl)propionic acid ethyl ester (82.4 g,
97%) as a brown liquid. TLC: Rf = 0.37 (eluting with 30% EtOAc
1
in petroleum ether). H NMR (400 MHz, CDCl3): δ 7.12 (d, J = 8.8
Hz, 1H), 6.97 (s, 1H), 6.92 (d, J = 8.8 Hz, 1H), 5.23 (q, J = 6.8 Hz,
2H), 4.67−4.58 (d, J = 14.8 Hz, 2H), 4.26−4.20 (m, 2H), 1.60 (d, J =
6.8 Hz, 3H), 1.23 (t, J = 7.6 Hz, 3H). SFC (column, Chiralpak AD-H
250 mm × 4.6 mm i.d., 5 μm; mobile phase, MeOH (0.05% DEA) in
supercritical CO2 from 5% to 40%; flow rate, 2.35 mL/min;
wavelength, 220 nm): tR = 4.146 min. [α]D20 31.45 (c = 0.002 g/mL
in MeOH).
Step C. A solution of (R)-2-(6-bromo-3-oxo-2,3-dihydrobenzo-
[1,4]oxazin-4-yl)propionic acid ethyl ester (82.4 g, 252 mmol) and
Lawesson reagent (61.1 g, 151 mmol) in toluene (900 mL) was heated
at reflux for 3 h and then cooled to ambient temperature. The solvent
was removed in vacuo, and the residue was purified by
chromatography on silica gel (eluting with 5−10% EtOAc in
petroleum ether) to give (R)-2-(6-bromo-3-thioxo-2,3-dihydrobenzo-
[1,4]oxazin-4-yl)propionic acid ethyl ester (75.2 g, 87%) as a yellow
solid. TLC: Rf = 0.55 (eluting with 12% EtOAc in petroleum ether).
1H NMR (400 MHz, CDCl3): δ 7.19 (d, J = 8.8 Hz, 1H), 7.04 (s, 1H),
6.93 (d, J = 8.8 Hz, 1H), 6.57 (m, 2H), 4.96−4.87 (m, 2H), 4.29−4.16
(m, 2H), 1.71 (d, J = 6.8 Hz, 3H), 1.18 (t, J = 6.8 Hz, 3H). SFC
(column, Chiralpak AD-H 250 mm × 4.6 mm i.d., 5 μm; mobile phase,
MeOH (0.05% DEA) in supercritical CO2 from 5% to 40%; flow rate,
2.35 mL/min; wavelength, 220 nm): tR = 3.529 min. [α]D20 19.25 (c =
0.002 g/mL in MeOH).
within 1.5 min; flow rate, 1.2 mL/min; wavelength, 220 nm): tR
=
0.869 min. MS: m/z 413 [M + H]+. (S)-3-methyl-4-((R)-4-methyl-3-
oxo-2,3,4,10-tetrahydro-9-oxa-1,2,4a-triazaphenanthren-6-yl)-3,6-dihy-
dro-2H-pyridine-1-carboxylic acid tert-butyl ester (isomer 2) was also
obtained as a white solid (0.57 g, 38%; SFC (column, Chiralcel OJ-H
250 mm × 4.6 mm i.d., 5 μm; mobile phase, MeOH (0.05% DEA)/
supercritical CO2 from 5% to 40%; flow rate, 2.35 mL/min;
wavelength, 220 nm): tR = 6.86 min). 1H NMR (CDCl3, 400
MHz): δ 8.08 (br s, 1H), 7.01−6.93 (m, 2H), 6.87 (s, 1H), 5.82 (br s,
1H), 4.77 (q, J = 6.6 Hz, 1H), 4.60 (q, J = 13.1 Hz, 2H), 4.47−4.13
(m, 1H), 3.85 (dd, J = 3.0, 12.8 Hz, 2H), 3.32 (dd, J = 3.4, 12.9 Hz,
1H), 2.78 (br s, 1H), 1.55−1.51 (m, 3H), 1.51 (s, 9H), 1.02 (d, J = 6.8
Hz, 3H). LC−MS (column, Ultimate XB-C18, 3 μm, 30 mm × 2.1
mm; mobile phase, acetonitrile (0.02% TFA) in water (0.04% TFA)
from 5% to 95% within 1.5 min; flow rate, 1.2 mL/min; wavelength,
220 nm): tR = 0.869 min. MS: m/z 413 [M + H]+.
Step D. To a mixture of (R)-2-(6-bromo-3-thioxo-2,3-
dihydrobenzo[1,4]oxazin-4-yl)propionic acid ethyl ester (75.2 g, 219
mmol) in EtOH (800 mL) was added hydrazine hydrate (98%, 21.9 g,
438 mmol), and the reaction mixture was stirred at ambient
temperature overnight. The precipitate was collected by filtration
and washed with cold EtOH (3 × 80 mL) to give 6-bromo-4-methyl-
2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-one (6 g, 9%) as a
343
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
Step F. A mixture of (R)-3-methyl-4-((R)-4-methyl-3-oxo-2,3,4,10-
tetrahydro-9-oxa-1,2,4a-triazaphenanthren-6-yl)-3,6-dihydro-2H-pyri-
dine-1-carboxylic acid tert-butyl ester (isomer 1, 0.55 g, 1.333 mmol)
and Pd/C (10%, 0.142 g, 0.13 mmol) in MeOH (20 mL) was stirred
under an atmosphere of H2 (1 atm) at ambient temperature for 3 h.
The reaction mixture was filtered, and the filtrate was concentrated in
vacuo to give (3R,4R)-3-methyl-4-((R)-4-methyl-3-oxo-2,3,4,10-tetra-
hydro-9-oxa-1,2,4a-triazaphenanthren-6-yl)piperidine-1-carboxylic acid
tert-butyl ester (0.54 g, 98%) as a colorless solid. 1H NMR (400 MHz,
CDCl3) δ 8.35 (s, 1H), 6.97 (d, J = 8.4 Hz, 1H), 6.79 (d, J = 8.4 Hz,
1H), 6.71 (s, 1H), 4.77 (q, J = 6.8 Hz, 1H), 4.65−4.52 (d, J = 13.2 Hz,
2H), 4.48−3.98 (m, 2H), 3.18−2.73 (m, 3H), 2.13−1.92 (m, 2H),
1.61 (br s, 1H), 1.55−1.48 (m, 13H), 0.76−0.67 (d, J = 5.6 Hz, 3H).
LC−MS (column, Merck, 3 μm, 25 mm × 2 mm; mobile phase,
MeCN (0.02% TFA) in water (0.04% TFA) from 5% to 95% within
0.7 min, keep at 95% for 0.4 min, keep at 5% within 0.4 min; flow rate,
1.5 mL/min; wavelength, 220 nm): tR = 0.869 min. MS: m/z 415 [M
+ H]+. SFC (column, Chiralcel OD-3 150 mm × 4.6 mm i.d., 3 μm;
mobile phase, EtOH (0.05% DEA)/supercritical CO2 from 5% to
40%; flow rate, 2.5 mL/min; wavelength, 220 nm): tR = 4.910 min.
Step G. To a mixture of (3R,4R)-3-methyl-4-((R)-4-methyl-3-oxo-
2,3,4,10-tetrahydro-9-oxa-1,2,4a-triazaphenanthren-6-yl)piperidine-1-
carboxylic acid tert-butyl ester (0.2 g, 0.483 mmol) in DCM (20 mL)
and MeOH (10 mL) was added tetra-N-butylammonium tribromide
(0.233 g, 0.483 mmol), and the reaction mixture was stirred at ambient
temperature for 30 min. Saturated aqueous Na2S2O3 (6 mL) was
added, and the reaction mixture was extracted with DCM (4 × 10
mL). The combined organic phases were washed with brine (30 mL),
dried over Na2SO4, and concentrated in vacuo. The residue was
purified by preparative TLC (eluting with 50% EtOAc in petroleum
ether) to give (3R,4R)-4-((R)-7-bromo-4-methyl-3-oxo-2,3,4,10-tetra-
hydro-9-oxa-1,2,4a-triazaphenanthren-6-yl)-3-methylpiperidine-1-car-
boxylic acid tert-butyl ester (0.12 g, 50%) as a colorless solid. 1H NMR
(400 MHz, CD3OD): δ 7.23 (s, 1H), 6.84 (s, 1H), 4.87−4.81 (m,
1H), 4.64−4.54 (m, 2H), 4.32 (d, J = 12.8 Hz, 1H), 4.05 (d, J = 13.2
Hz, 1H), 3.20−2.75 (m, 3H), 2.37−2.17 (m, 3H), 1.48 (s, 9H), 1.44−
1.38 (m, 4H), 0.73 (d, J = 6.8 Hz, 3H). LC−MS (column, Merck, 3
μm, 25 mm × 2 mm; mobile phase, MeCN (0.02% TFA) in water
(0.04% TFA) from 5% to 95% within 0.7 min, keep at 95% for 0.4
min, keep at 5% within 0.4 min; flow rate, 1.5 mL/min; wavelength,
220 nm): tR = 0.934 min. MS: m/z 493 [M + H]+, 495 [M + H + 2]+.
SFC (column, Chiralpak AD-3, 150 mm × 4.6 mm i.d., 3 μm; mobile
phase, MeOH (0.05% DEA) in supercritical CO2 from 5% to 40%;
flow rate, 2.5 mL/min; wavelength, 220 nm): tR = 6.79 min.
ylpiperidine-1-carboxylic acid tert-butyl ester (0.076 g, 0.149 mmol) in
HCl (4 M in EtOAc, 15 mL) was stirred at ambient temperature for 1
h. The reaction mixture was concentrated in vacuo to give (R)-7-(2-
fluorophenyl)-4-methyl-6-((3R,4R)-3-methylpiperidin-4-yl)-2,10-dihy-
dro-9-oxa-1,2,4a-triazaphenanthren-3-one hydrochloride (27) (0.065
g, 98%) as a white solid. 1H NMR (400 MHz, CD3OD): δ 7.50−7.41
(m, 1H), 7.32−7.19 (m, 3H), 6.91 (br s, 1H), 6.86 (s, 1H), 4.80 (m,
1H), 4.70−4.59 (m, 2H), 3.48 (m, 1H), 3.15−2.98 (m, 2H), 2.82 (m,
1H), 2.43 (m, 1H), 1.88 (m, 2H), 1.50 (d, J = 6.4 Hz, 3H), 0.91 (br s,
3H). LC−MS: tR = 1.534 min. MS: m/z 409 [M + H]+. SFC (column,
Chiralpak AS-H, 150 mm × 4.6 mm i.d., 5 μm; mobile phase, MeOH
(0.05% DEA) in supercritical CO2 from 5% to 40%; flow rate, 3 mL/
min; wavelength, 220 nm): tR = 1.341 min.
Step J. A mixture of (R)-7-(2-fluorophenyl)-4-methyl-6-((3R,4R)-3-
methylpiperidin-4-yl)-2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-
one hydrochloride (3.05 g, 6.88 mmol) and paraformaldehyde (2.066
g, 68.8 mmol) in MeOH (60 mL) and AcOH (6 mL) was stirred at 80
°C for 16 h. Sodium cyanoborohydride (0.865 g, 13.76 mmol) was
added, and the reaction mixture was stirred at ambient temperature for
0.5 h. The reaction mixture was concentrated in vacuo, and the residue
was purified by column chromatography on basicified silica gel (eluting
with 10% MeOH in DCM) to give (R)-6-((3R,4R)-1,3-dimethylpiper-
idin-4-yl)-7-(2-fluorophenyl)-4-methyl-2,10-dihydro-9-oxa-1,2,4a-tria-
zaphenanthren-3-one (1.15 g, 40%) as a white solid. The solid was
stirred in HCl (4 M in EtOAc, 50 mL) for 30 min, and the reaction
mixture was concentrated in vacuo to give (R)-6-((3R,4R)-1,3-
dimethylpiperidin-4-yl)-7-(2-fluorophenyl)-4-methyl-2,10-dihydro-9-
oxa-1,2,4a-triazaphenanthren-3-one hydrochloride (17a) (1.215 g,
1
39%) as a white solid. H NMR (400 MHz, CD3OD): δ 7.44 (m,
1H), 7.32−7.16 (m, 3H), 6.90 (m, 1H), 6.84 (s, 1H), 4.62 (d, J = 13.2
Hz, 2H), 3.57 (m, 1H), 3.25 (m, 1H), 3.05 (m, 2H), 2.86 (m, 1H),
2.79 (s, 3H), 2.50 (m, 1H), 1.91 (m, 2H), 1.48 (d, J = 6.8 Hz, 3H),
0.91 (m, 3H). LC−MS (column, Ultimate XB-C18, 3 μm; 30 mm ×
2.1 mm; mobile phase, MeCN (0.02% TFA) in water (0.04% TFA)
from 0% to 60% within 4 min; flow rate, 1.2 mL/min; wavelength, 220
nm): tR = 2.275 min. MS: m/z 423 [M + H]+. SFC (column, Chiralcel
OJ-3, 50 mm × 4.6 mm i.d., 3 μm; mobile phase, MeOH (0.05%
DEA) in supercritical CO2 from 5% to 40%; flow rate, 4 mL/min;
wavelength, 220 nm): tR = 1.145 min.
(R)-6-((3S,4S)-1,3-Dimethylpiperidin-4-yl)-7-(2-fluorophen-
yl)-4-methyl-2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-
one Hydrochloride (28). Prepared according to the procedure for
compound 17a (steps F−J). Step F: using (S)-3-methyl-4-((R)-4-
methyl-3-oxo-2,3,4,10-tetrahydro-9-oxa-1,2,4a-triazaphenanthren-6-
yl)-3,6-dihydro-2H-pyridine-1-carboxylic acid tert-butyl ester (the
intermediate obtained in step E during the preparation of compound
17a). Step G: tetra-N-butylammonium tribromide (1 equiv). Step H:
(2-fluorophenyl)boronic acid (1.5 equiv), 90 °C overnight. Step I:
HCl (4 M in EtOAc) at ambient temperature. Step J: paraformalde-
hyde (2.5 equiv) then sodium cyanoborohydride (2 equiv) in 3%
Step H. To a mixture of (2-fluorophenyl)boronic acid (0.051 g,
0.365 mmol), (3R,4R)-4-((R)-7-bromo-4-methyl-3-oxo-2,3,4,10-tetra-
hydro-9-oxa-1,2,4a-triazaphenanthren-6-yl)-3-methylpiperidine-1-car-
boxylic acid tert-butyl ester (0.12 g, 0.243 mmol), and K2CO3 (0.067 g,
0.486 mmol) in 1,4-dioxane (12 mL) and water (2 mL) was added
PdCl2(dppf)−CH2Cl2 adduct (0.020 g, 0.024 mmol), and the reaction
mixture was stirred at 90 °C overnight. The reaction mixture was
cooled to ambient temperature and concentrated in vacuo, and the
residue was purified by column chromatography on silica gel (eluting
with 3−30% EtOAc in petroleum ether) to give (3R,4R)-4-[(R)-7-(2-
fluorophenyl)-4-methyl-3-oxo-2,3,4,10-tetrahydro-9-oxa-1,2,4a-triaza-
phenanthren-6-yl]-3-methylpiperidine-1-carboxylic acid tert-butyl ester
1
overall yield as a white solid. H NMR (400 MHz, MeOH-d4): δ
7.49−7.40 (m, 1H), 7.33−7.17 (m, 3H), 6.91 (s, 1H), 6.85 (s, 1H),
4.99−4.93 (m, 1H), 4.70−4.65 (m, 2H), 3.60−3.50 (m, 1H), 3.25−
3.00 (m, 3H), 2.90−2.85 (m, 1H), 2.80 (s, 3H), 2.55−2.45 (m, 1H),
2.01−1.84 (m, 2H), 1.49 (d, J = 6.4 Hz, 3H), 1.00−0.90 (m, 3H).
LC−MS (at wavelengths of 220 and 254 nm): tR = 1.654 min. MS: m/
z 423 [M + H]+. SFC (column, Chiralpak AS-H, 150 mm × 4.6 mm
i.d., 5 μm; mobile phase, EtOH (0.05% DEA) in supercritical CO2
1
(0.076 g, 61%) as a colorless solid. H NMR (CDCl3, 400 MHz): δ
8.34 (br s, 1H), 7.44−7.32 (m, 1H), 7.24−7.11 (m, 3H), 6.85 (s, 1H),
6.73 (br s, 1H), 4.80 (br s, 1H), 4.70−4.53 (m, 2H), 4.43−4.17 (m,
1H), 3.90−3.60 (m, 1H), 2.89 (br s, 1H), 2.60 (br s, 2H), 1.56 (d, J =
6.8 Hz, 3H), 1.46 (m, 10H), 0.72 (m, 3H). LC−MS (column, Merck,
3 μm, 25 mm × 2 mm; mobile phase, MeCN (0.02% TFA) in water
(0.04% TFA) from 5% to 95% within 0.7 min, keep at 95% for 0.4
min, keep at 5% within 0.4 min; flow rate, 1.5 mL/min; wavelength,
220 nm): tR = 0.948 min. MS: m/z 509 [M + H]+. SFC (column,
Chiralpak AD-3, 150 mm × 4.6 mm i.d., 3 μm; mobile phase, MeOH
(0.05% DEA) in supercritical CO2 from 5% to 40%; flow rate, 2.5 mL/
min; wavelength, 310 nm): tR = 6.21 min.
from 5% to 40%; flow rate, 3 mL/min; wavelength, 220 nm): tR
=
3.206 min.
(R)-6-((3R,4R)-1,3-Dimethylpiperidin-4-yl)-7-(2-fluoro-4-me-
thoxyphenyl)-4-methyl-2,10-dihydro-9-oxa-1,2,4a-triazaphe-
nanthren-3-one (17l). Step A. A solution of (3R,4R)-4-((R)-7-
bromo-4-methyl-3-oxo-2,3,4,10-tetrahydro-9-oxa-1,2,4a-triazaphe-
nanthren-6-yl)-3-methylpiperidine-1-carboxylic acid tert-butyl ester (3
g, 6.08 mmol) in HCl (4 M in EtOAc, 10 mL) was stirred at ambient
temperature for 1 h. The solvent was removed in vacuo to give crude
(R)-7-bromo-4-methyl-6-((3R,4R)-3-methylpiperidin-4-yl)-2,10-dihy-
dro-9-oxa-1,2,4a-triazaphenanthren-3-one hydrochloride (2.2 g, 82%)
as a white solid, which was used directly in the next step. LC−MS
Step I. A mixture of (3R,4R)-4-[(R)-7-(2-fluorophenyl)-4-methyl-3-
oxo-2,3,4,10-tetrahydro-9-oxa-1,2,4a-triazaphenanthren-6-yl]-3-meth-
344
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
(column, Merck, 3 μm, 25 mm × 2 mm; mobile phase, MeCN (0.02%
TFA) in water (0.04% TFA) from 5% to 95% within 0.7 min, keep at
95% for 0.4 min, keep at 5% within 0.4 min; flow rate, 1.5 mL/min;
wavelength, 220 nm): tR = 0.720 min. MS: m/z 393 [M + H]+, 395 [M
+ H + 2]+.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript. D.M.G., E.C.B., M.A.A., P.B.-P., D.B.D., P.H.,
Q.L., S.M., L.R., A.S., and J.J.E. are employees at AbbVie. The
design, study conduct, and financial support for the research
were provided by AbbVie. Y.Z., J.W., and J.S. are employees at
WuXi AppTec (Shanghai) Co., Ltd. AbbVie and WuXi
participated in the interpretation of data, review, and approval
of the publication.
Step B. To a solution of (R)-7-bromo-4-methyl-6-((3R,4R)-3-
methylpiperidin-4-yl)-2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-
one hydrochloride (2.2 g, 5.59 mmol) in AcOH (10 mL) and MeOH
(100 mL) was added paraformaldehyde (0.839 g, 28.0 mmol). The
reaction mixture was stirred at ambient temperature for 1 h.
NaBH3CN (0.557 g, 14.0 mmol) was added, and the reaction mixture
was stirred at ambient temperature overnight. Saturated aqueous
NaHCO3 (30 mL) was added, and the reaction mixture was extracted
with EtOAc (3 × 50 mL). The combined organic phases were dried
over sodium sulfate, filtered, and concentrated in vacuo. The residue
was purified by preparatory HPLC (column, Phenomenex Gemini
C18 200 mm × 25 mm, 10 μm; mobile phase, acetonitrile in water
(0.1% TFA) from 20% to 45% in 10 min; flow rate, 25 mL/min;
wavelength, 220 nm) to give (R)-7-bromo-6-((3R,4R)-1,3-dimethylpi-
peridin-4-yl)-4-methyl-2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-
one (1.8 g, 79%) as a light-yellow solid. 1H NMR (400 MHz,
CD3OD): δ 7.22 (s, 1H), 6.86 (s, 1H), 4.87−4.91 (m, 1H), 4.58 (d, J
= 13.2 Hz, 2H), 3.15 (m, 1H), 3.04 (m, 1H), 2.84 (m, 1H), 2.23−2.39
(m, 6H), 2.07−2.18 (m, 1H), 1.55 (m, 1H), 1.43 (d, J = 6.5 Hz, 3H),
0.83 (d, J = 7.0 Hz, 3H). LC−MS (column, Merck, 3 μm, 25 mm × 2
mm; mobile phase, MeCN (0.02% TFA) in water (0.04% TFA) from
5% to 95% within 0.7 min, keep at 95% for 0.4 min, keep at 5% within
0.4 min; flow rate, 1.5 mL/min; wavelength, 220 nm): tR = 0.681 min.
MS: m/z 407 [M + H]+, 409 [M + H + 2]+.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the following people from AbbVie: Li Li,
Yuanyuan Sun, and Zheng Zhou for in vitro testing; Kent
Stewart for helpful design and modeling discussions; Rodger
Henry for the small-molecule X-ray structure of compound
17a; Xiaoqin Liu for conscious telemetry mouse data with 17h;
Ron Pithawalla for expression and Anup Upadhyay for
purification of PKCα protein for cocrystallization; and Russell
Judge for growth of PKCα cocrystals with compound 28. The
X-ray crystal structure of PKCθ with compound 11a was
determined by Proteros Biostructures GmbH (Martinsried,
Germany). Data for compound 28 bound to PKCα were
collected at IMCA-CAT 17ID. Use of the IMCA-CAT
beamline 17-ID (or 17-BM) at the Advanced Photon Source
was supported by the companies of the Industrial Macro-
molecular Crystallography Association through a contract with
Hauptman-Woodward Medical Research Institute. Use of the
Advanced Photon Source was supported by the U.S. Depart-
ment of Energy, Office of Science, Office of Basic Energy
Sciences, under Contract DE-AC02-06CH11357.
Step C. To a mixture of (R)-7-bromo-6-((3R,4R)-1,3-dimethylpi-
peridin-4-yl)-4-methyl-2,10-dihydro-9-oxa-1,2,4a-triazaphenanthren-3-
one (0.050 g, 0.123 mmol), (2-fluoro-4-methoxyphenyl)boronic acid
(0.025 g, 0.147 mmol), and potassium carbonate (0.034 g, 0.246
mmol) in 1,4-dioxane (6 mL) and water (1 mL) was added
PdCl2(dppf)−CH2Cl2 adduct (0.020 g, 0.025 mmol), and the reaction
mixture was heated at 90 °C overnight. The reaction mixture was
cooled to ambient temperature and concentrated in vacuo. The
residue was purified by column chromatography on silica gel (eluting
with 10% MeOH in dichloromethane) to give (R)-6-((3R,4R)-1,3-
dimethylpiperidin-4-yl)-7-(2-fluoro-4-methoxyphenyl)-4-methyl-2,10-
dihydro-9-oxa-1,2,4a-triazaphenanthren-3-one (17l) (0.050 g, 90%) as
ABBREVIATIONS USED
■
CV, cardiovascular; Con A, concanavalin A; CSA, cyclosporin
A; GPI, glucose phosphate isomerase; IBD, inflammatory bowel
disease; IL-2, interleukin-2; MDCK-MDR, Madin Darby canine
kidney cells with the MDR gene; met ID, metabolite
identification; PK, pharmacokinetics; PKC, protein kinase C;
RA, rheumatoid arthritis; VDW, van der Waals
1
a brown solid. LC−MS: tR = 1.726 min. MS: m/z 453 [M + H]+. H
NMR (400 MHz, CD3OD): δ 7.18 (t, J = 8.5 Hz, 1H), 6.94−6.79 (m,
4H), 4.96−4.94 (m, 1H), 4.64 (d, J = 6.3 Hz, 2H), 3.86 (s, 3H), 3.51
(m, 1H), 3.24−2.78 (m, 4H), 2.74 (s, 3H), 2.48 (m, 1H), 2.00−1.82
(m, 2H), 1.50 (d, J = 6.5 Hz, 3H), 0.91 (d, J = 7.3 Hz, 3H). SFC
(column, Chiralpak AS-H, 150 mm × 4.6 mm i.d., 5 μm; mobile phase,
MeOH (0.05% DEA) in supercritical CO2 from 5% to 40%; flow rate,
3 mL/min; wavelength, 220 nm): tR = 3.175 min.
REFERENCES
■
Biology. Kinase and cellular assays and Con A and GPI in vivo
models were run following protocols described previously.1 To enable
higher throughput, a modified screening version of the Con A model
was utilized in which each compound was tested at a single 10 mg/kg
dose rather than in dose−response format. This allowed preliminary
analysis of up to five compounds per experiment wherein compounds
were dosed 1 h prior to challenge (t = −1 h), an orbital bleed for an
estimated Cmax was taken immediately prior to Con A administration
(t = 0 h), and 2 h later (t = 2 h) IL-2 levels and terminal compound
concentrations were measured.
(1) George, D. M.; Breinlinger, E. C.; Friedman, M.; Zhang, Y.;
Wang, J.; Argiriadi, M.; Bansal-Pakala, P.; Barth, M.; Duignan, D. B.;
Honore, P.; Lang, Q.; Mittelstadt, S.; Potin, D.; Rundell, L.; Edmunds,
J. J. Discovery of selective and orally bioavailable protein kinase Cθ
(PKCθ) inhibitors from a fragment hit. J. Med. Chem. 2014,
DOI: 10.1021/jm500669m.
(2) George, D. M.; Argiriadi, M. A.; Barth, M.; Bonafoux, D. F.;
Breinlinger, E. C.; Dai, L.; Edmunds, J. J.; Friedman, M.; Guillier, F.;
Hobson, A. D.; Potin, D.; Thomas, D.; Wang, J.; Wu, W.; Zhang, Y.
Triazinone compounds. Patent WO2014089904, 2014.
(3) Altman, A.; Isakov, N.; Baier, G. Protein kinase Cθ: A new
essential superstar on the T-cell stage. Immunol. Today 2000, 21, 567−
573.
(4) Isakov, N.; Altman, A. Protein kinase Cθ in T cell activation.
Annu. Rev. Immunol. 2002, 20, 761−794.
(5) Altman, A.; Villalba, M. Protein kinase Cθ (PKCθ): A key
enzyme in T cell life and death. J. Biochem. 2002, 132, 841−846.
(6) Cywin, C. L.; Dahmann, G.; Prokopowicz, A. S., III; Young, E. R.;
Magolda, R. L.; Cardozo, M. G.; Cogan, D. A.; DiSalvo, D.; Ginn, J.
D.; Kashem, M. A.; Wolak, J. P.; Homon, C. A.; Farrell, T. M.; Grbic,
H.; Hu, H.; Kaplita, P. V.; Liu, L. H.; Spero, D. M.; Jeanfavre, D. D.;
ASSOCIATED CONTENT
* Supporting Information
Additional experimental details. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*Telephone: 1-508-688-8001. E-mail: dawn.george@abbvie.
com.
■
345
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346

Journal of Medicinal Chemistry
Article
O’Shea, K. M.; White, D. M.; Woska, J. R., Jr.; Brown, M. L. Discovery
of potent and selective PKC-θ inhibitors. Bioorg. Med. Chem. Lett.
2007, 17, 225−230.
(7) Roughten, A.; Rong, Y.; Ho, K.-K.; Ohlmeyer, M.; Diller, D.
Purinones and 1H-imidazopyridinones as PKCθ inhibitors. Patent
WO2008143674 A1, 2008.
(20) Hegen, M.; Keith, J. C., Jr.; Collins, M.; Nickerson-Nutter, C. L.
Utility of animal models for identification of potential therapeutics for
rheumatoid arthritis. Ann. Rheum. Dis. 2008, 67, 1505−1515.
(21) Molecular Operating Environment (MOE), version 2013.08;
Chemical Computing Group: Montreal, QC, 2014.
(8) Roughton, A.; Ho, K.-K.; Ohlmeyer, M.; Diller, D.; Neagu, I.;
Kingsbury, C.; Chan, J.-H.; Lommerse, M.; Teerhuis, M.; Wijkmans,
M.; Plate, R. Pyrrolo[2,3-d]pyrimidin-2-yl-amine derivatives as PKCθ
inhibitors and their preparation. Patent CA2716202 A1, 2009.
(9) Cole, D. C.; Asselin, M.; Brennan, A.; Czerwinski, R.; Ellingboe, J.
W.; Fitz, L.; Greco, R.; Huang, X.; Joseph-McCarthy, D.; Kelly, M. F.;
Kirisits, M.; Lee, J.; Li, Y.; Morgan, P.; Stock, J. R.; Tsao, D. H.;
Wissner, A.; Yang, X.; Chaudhary, D. Identification, characterization
and initial hit-to-lead optimization of a series of 4-arylamino-3-
pyridinecarbonitrile as protein kinase C theta (PKCθ) inhibitors. J.
Med. Chem. 2008, 51, 5958−5963.
(10) Prashad, A. S.; Wang, D.; Subrath, J.; Wu, B.; Lin, M.; Zhang, M.
Y.; Kagan, N.; Lee, J.; Yang, X.; Brennan, A.; Chaudhary, D.; Xu, X.;
Leung, L.; Wang, J.; Boschelli, D. H. C-5 substituted heteroaryl-3-
pyridinecarbonitriles as PKCθ inhibitors: Part II. Bioorg. Med. Chem.
Lett. 2009, 19, 5799−5802.
(11) Subrath, J.; Wang, D.; Wu, B.; Niu, C.; Boschelli, D. H.; Lee, J.;
Yang, X.; Brennan, A.; Chaudhary, D. C-5 Substituted heteroaryl 3-
pyridinecarbonitriles as PKCθ inhibitors: Part I. Bioorg. Med. Chem.
Lett. 2009, 19, 5423−5425.
(12) Dushin, R. G.; Nittoli, T.; Ingalls, C.; Boschelli, D. H.; Cole, D.
C.; Wissner, A.; Lee, J.; Yang, X.; Morgan, P.; Brennan, A.; Chaudhary,
D. Synthesis and PKCθ inhibitory activity of a series of 4-
indolylamino-5-phenyl-3-pyridinecarbonitriles. Bioorg. Med. Chem.
Lett. 2009, 19, 2461−2463.
(13) Wu, B.; Boschelli, D. H.; Lee, J.; Yang, X.; Chaudhary, D.
Second generation 4-(4-methyl-1H-indol-5-ylamino)-2-phenylthieno-
[2,3-b]pyridine-5-carbonitrile PKCθ inhibitors. Bioorg. Med. Chem.
Lett. 2009, 19, 766−769.
(14) Wagner, J.; von Matt, P.; Sedrani, R.; Albert, R.; Cooke, N.;
Ehrhardt, C.; Geiser, M.; Rummel, G.; Stark, W.; Strauss, A.; Cowan-
Jacob, S. W.; Beerli, C.; Weckbecker, G.; Evenou, J. P.; Zenke, G.;
Cottens, S. Discovery of 3-(1H-indol-3-yl)-4-[2-(4-methylpiperazin-1-
yl)quinazolin-4-yl]pyrrole-2,5-dione (AEB071), a potent and selective
inhibitor of protein kinase C isotypes. J. Med. Chem. 2009, 52, 6193−
6196.
(15) Wagner, J.; von Matt, P.; Faller, B.; Cooke, N. G.; Albert, R.;
Sedrani, R.; Wiegand, H.; Jean, C.; Beerli, C.; Weckbecker, G.; Evenou,
J. P.; Zenke, G.; Cottens, S. Structure−activity relationship and
pharmacokinetic studies of sotrastaurin (AEB071), a promising novel
medicine for prevention of graft rejection and treatment of psoriasis. J.
Med. Chem. 2011, 54, 6028−6039.
(16) Jimenez, J. M.; Boyall, D.; Brenchley, G.; Collier, P. N.; Davis,
C. J.; Fraysse, D.; Keily, S. B.; Henderson, J.; Miller, A.; Pierard, F.;
Settimo, L.; Twin, H. C.; Bolton, C. M.; Curnock, A. P.; Chiu, P.;
Tanner, A. J.; Young, S. Design and optimization of selective protein
kinase C θ (PKCθ) inhibitors for the treatment of autoimmune
diseases. J. Med. Chem. 2013, 56, 1799−1810.
(17) Jimenez, J. M.; Davis, C.; Boyall, D.; Fraysse, D.; Knegtel, R.;
Settimo, L.; Young, S.; Bolton, C.; Chiu, P.; Curnock, A.; Rasmussen,
R.; Tanner, A.; Ager, I. Structure-based optimization of amino-
pyridines as PKCθ inhibitors. Bioorg. Med. Chem. Lett. 2012, 22,
4645−4649.
(18) Schubert, D.; Maier, B.; Morawietz, L.; Krenn, V.; Kamradt, T.
Immunization with glucose-6-phosphate isomerase induces T cell-
dependent peripheral polyarthritis in genetically unaltered mice. J.
Immunol. 2004, 172, 4503−4509.
(19) Matsumoto, I.; Zhang, H.; Yasukochi, T.; Iwanami, K.; Tanaka,
Y.; Inoue, A.; Goto, D.; Ito, S.; Tsutsumi, A.; Sumida, T. Therapeutic
effects of antibodies to tumor necrosis factor-α, interleukin-6 and
cytotoxic T-lymphocyte antigen 4 immunoglobulin in mice with
glucose-6-phosphate isomerase induced arthritis. Arthritis Res. Ther.
2008, 10, R66.
346
dx.doi.org/10.1021/jm5013006 | J. Med. Chem. 2015, 58, 333−346