(1 R,2 R)-N-(1-cyanocyclopropyl)-2-(6-methoxy-1,3,4,5-tetrahydropyrido[4,3- b]indole-2-carbonyl)cyclohexanecarboxamide (AZD4996): A potent and highly selective cathepsin k inhibitor for the treatment of osteoarthritis

Article abstract of DOI:10.1021/jm3007257

Directed screening of nitrile compounds revealed 3 as a highly potent cathepsin K inhibitor but with cathepsin S activity and very poor stability to microsomes. Synthesis of compounds with reduced molecular complexity, such as 7, revealed key SAR and demonstrated that baseline physical properties and in vitro stability were in fact excellent for this series. The tricycle carboline P3 unit was discovered by hypothesis-based design using existing structural information. Optimization using small substituents, knowledge from matched molecular pairs, and control of lipophilicity yielded compounds very close to the desired profile, of which 34 (AZD4996) was selected on the basis of pharmacokinetic profile.

Full text of DOI:10.1021/jm3007257

Article
pubs.acs.org/jmc
(1R,2R)-N-(1-Cyanocyclopropyl)-2-(6-methoxy-1,3,4,5-
tetrahydropyrido[4,3-b]indole-2-carbonyl)cyclohexanecarboxamide
(AZD4996): A Potent and Highly Selective Cathepsin K Inhibitor for
the Treatment of Osteoarthritis
Alexander G. Dossetter,*,† Howard Beeley,† Jonathan Bowyer,† Calum R. Cook,† James J. Crawford,†
Jonathan E. Finlayson,† Nicola M. Heron,† Christine Heyes,† Adrian J. Highton,† Julian A. Hudson,†
Anja Jestel,‡ Peter W. Kenny,† Stephan Krapp,‡ Scott Martin,† Philip A. MacFaul,† Thomas M. McGuire,†
Pablo Morentin Gutierrez,† Andrew D. Morley,† Jeffrey J. Morris,† Ken M. Page,†
Lyn Rosenbrier Ribeiro,† Helen Sawney,† Stefan Steinbacher,‡ Caroline Smith,† and Madeleine Vickers†
†AstraZeneca R&D, Mereside, Alderley Park, Macclesfield, Cheshire, SK10 4TG, U.K.
‡Proteros Biostructures, Am Klopferspitz 19, D-82152 Martinsried, Germany
S
* Supporting Information
ABSTRACT: Directed screening of nitrile compounds revealed 3 as a highly potent cathepsin K inhibitor but with cathepsin S
activity and very poor stability to microsomes. Synthesis of compounds with reduced molecular complexity, such as 7, revealed
key SAR and demonstrated that baseline physical properties and in vitro stability were in fact excellent for this series. The tricycle
carboline P3 unit was discovered by hypothesis-based design using existing structural information. Optimization using small
substituents, knowledge from matched molecular pairs, and control of lipophilicity yielded compounds very close to the desired
profile, of which 34 (AZD4996) was selected on the basis of pharmacokinetic profile.
bone resorption by degradation of type I cartilage.6 Cat K
knockout mice exhibit osteopetrosis (abnormally dense bone)
INTRODUCTION
Osteoarthritis (OA) is made up of a group of degenerative
disorders characterized by joint pain and loss of function in the
absence of chronic autoimmune or autoinflammatory mecha-
■
of the long bones and vertebrae and abnormal joint
morphology,7 indicating the potential role of this enzyme in
nisms.1 With a global aging population, coupled with other risk
key bone pathologies associated with abnormal bone turnover
factors such as obesity,2 the disease represents a growing
such as osteoporosis (OP), osteoarthritis (OA), and metastatic
bone disease (MBD).8−10As well as a role in bone disorders
and OA, more recent findings have shown potential for the
treatment of diabetes and obesity.11
socioeconomic burden.3 Pharmacological treatment is largely
confined to symptomatic pain relief with intraarticular cortico-
steroid injection, analgesics, and NSAIDs,4 to the point where,
in severely advanced disease, joint replacement surgery is often
performed, but is not always successful in alleviating pain and
restoring function.5 Articular cartilage breakdown is a
prominent feature of joint degeneration and has consequently
received significant attention from the pharmaceutical industry,
where focusing on identifying targets, such as overexpressed
proteases has led to drug development programs.
Several pharmaceutical companies have trialed Cat K
inhibitors in these disease areas; Novartis has completed
phase II studies in OP and OA with balicatib (Figure 1).12 The
phase II OP trial reported positive outcomes on bone mineral
density measures. Merck is studying odanacatib (MK0822) in a
phase III trial for OP and has abandoned plans for a phase II in
Cathepsin K (Cat K), a 24 kDa lysosomal cysteine protease,
is expressed at high levels in osteoclasts and plays a key role in
Received: March 14, 2012
Published: June 28, 2012
© 2012 American Chemical Society
6363
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
Figure 1. Literature nitrile cathepsin K inhibitors.
OA.13−15 GSK (relacatib)16 conducted phase I trials for both
OP and OA several years ago but switched to MBD for phase II
during 2006.17
Drug-Hunting Approach. A directed screen of nitrile-
containing compounds, from the corporate collection, identi-
fied 3 and 4, which were originally designed as inhibitiors of
cathepsin S (Cat S), as starting points for this work (Figure
2).18 Both 3 (Cat K IC50 0.002 μM; Cat S IC50 0.076 μM) and
the contribution that the core portion of the molecule makes to
affinity. Second, it was desirable to provide multiple scaffolds
with hopefully high ligand efficiency (LE) that could readily be
elaborated synthetically. The number of compounds in the
design was further reduced, selecting those tertiary amides that
would be rigid in structure and thereby most likely to make
good surface contact with the protein surface, which is known
to be fairly open in nature.22,23 With this initial understanding
of the SAR, these hopefully efficient scaffolds would allow
exploration of the nitrile end of the molecules and then finally
optimization to quality candidates.
Chemistry. The resolution of ( )-trans-cyclohexyl-1,2-
dicarboxylic acid to yield (R,R)-5 was initially carried out
using (R)-1-phenethylamine by the method of Berkessel,24 and
subsequent recrystallizations gave a high enantiomeric excess
(>99%). We found heating in neat acetic anhydride was a better
method of dehydrating 5 into anhydride 6, as this only required
concentration to dryness to provide material used directly in
the next step. Ring-opening of the anhydride at ambient
temperature with a series of secondary amines (7a−13a, 18a−
39a) yielded in situ the acids 7b−13b, 18b− 39b) (Scheme 1).
Final compounds (7−39) were synthesized by a general
procedure of amide formation with the hydrochloride salt of
aminoacetonitrile either with O-(7-azabenzotriazol-1-yl)-
N,N,N′,N′-tetramethyluronium hexafluorophosphate (HATU)
or benzotriazolyloxy-tris(pyrrolidino)phosphonium hexafluor-
ophosphate (PYBOP) coupling reagents and diisopropylethyl-
amine (DIPEA) at ambient temperature in overall yields of 7−
78%. The carbolines used to synthesize final compounds were
made by Fischer indole synthesis25 from a suitable hydrazine 40
and 4-piperidone hydrochloride 41. Simple heating of the two
together in ethanol/concentrated HCl, in general, yielded the
carboline in good yield (for example 34a, R6 = OCH3 in 55%,
Scheme 2). With electron-poor hydrazines, more forcing
conditions from trifluoroboron etherate complex (BF3·OEt2)
in acetic acid at 150 °C were used. The exception to this was
Figure 2. Screening hits from cathepsin S program.
4 (Cat K IC50 0.003 μM; Cat S IC50 0.258 μM) were attractive,
potent neutral inhibitors of Cat K with some selectivity over
Cat S. Cathepsins are present in the lysosome, an acidic
compartment in cells, and as such basic compounds have been
shown to accumulate to high concentration (lysosotropism),
resulting in pan cathepsin inhibition; hence, these neutral
cyclohexyls are attractive from this point of view.19,20 However,
3 and 4 both had poor metabolic stability in rodent
pharmacokinetic (PK) studies, with clearances exceeding liver
blood flow. The objective of the current study was to optimize
these nitrile inhibitors by increasing Cat K selectivity and
reduce metabolism to produce a PK profile compatible with
oral dosing and an acceptable half-life.
We elected to start with an unsubstituted nitrile-amide
covalent binder and selected simple tertiary amides with low
molecular complexity.21 There were two reasons for adopting
this approach to library design. First, it was desirable to quantify
a
Scheme 1. Synthesis of Cyclohexyl-1,2-diamides 7−39
a
Reagents and conditions: (a) Ac2O, 80 °C, 1 h, concentrate. (b) CH2Cl2, RT, 1 h, concentrate. (c) DMF or CH2Cl2, DIPEA (5 equiv), HATU (1.1
equiv), 13c−17c HCl salts (1.0 equiv), RT, 12 h, 12−78% yield.
6364
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of Carbolines via Fischer Synthesis
RESULTS AND DISCUSSION
■
Enzyme inhibitory activity for seven compounds from this first
library is reported in Table 1 with measured physicochemical
properties and in vitro microsomal stability. The IC50 value of
0.257 μM measured for the prototypical N,N-dimethyl amide 7
shows that the core substructure does indeed make a significant
contribution to affinity and thus provides a useful reference
point. The LE of 0.53 and ligand lipophilicity efficiency (LLE =
−log10(IC50) − log D7.4) > 6 for 7 are both high, and the
compound shows some good selectivity with respect to other
cathepsins (Cat S, 14-fold). Compound 7 is polar (log D7.4
<
0.5) and, as such, has good aqueous solubility and high stability
in both human (HLM) and rat liver microsomes (RLM),
revealing the cyclohexyl diamide core as an excellent chemical
starting point.27−30 Compared to 7, partially and fully saturated
rings with connecting aromatic rings yield positive gains in
enzyme inhibition, which was in line with a concomitant rise in
lipophicity, suggesting hydrophobic and favorable ligand van de
Waals interactions were responsible, and these were efficient
(only slight reduction in LE on going from 7 to 13). The
change from 8 to 9 is interesting, suggesting an improved fit
into the P3 groove by the partially saturated cyclohexenyl ring.
The aromatic ring of 10 adds 9-fold in Cat K potency,
suggesting only partial contact with the protein surface, but
gratifyingly selectivity is improved (Cat S, 250 fold). (N,N)-
Piperazines inspired by the original hits (3, 4) were included in
the library. Compound 11 is sufficiently basic to protonate at
normal physiological pH, and the observation that it is
a
Reagents and conditions: (a) (i) EtOH, concd HCl(aq), 80 °C, 3 h,
29−71% yield; or (ii) AcOH, BF3.OEt2, 80 °C, 3 h, 44−68% yield. (b)
(i) PPA, 80 °C; (ii) HCO2H, NH4(HCO2), Pd(OAc)2, 40%.
the synthesis of 2,3,4,5-tetrahydro-1H-pyrido[4,3-b]-7-azain-
dole 35a where an N-benzyl-protected 4-piperidone 39 with
pyridine-1-hydrazine 38 and prolonged heating in polyphos-
phoric acid was required. Transfer hydrogenation using
homogeneous conditions with palladium acetate, formic acid,
and ammonium formate at reflux yielded 39a in 40% overall
yield on small scale only.26 With these tricycles in hand, a
similar procedure was followed to yield final SAR compounds
by the same ring-opening method and coupling to the acid.
Table 1. SAR for Tertiary Amide Variations
a
b
c
Binding affinity for cathepsin versus FRET substrate; mean of greater than n = 4 tests, unless where stated. Mean of n = 2 tests. kJ mol−1 Da1−.
d
e
In vitro human liver microsomal turnover mean of at least n = 2 tests (μL/min/mg). In vitro rat liver microsomal turnover mean of at least n = 2
tests (μL/min/mg); values in parentheses denote rat liver hepatocyte Clint mean n = 2 (μL/min/106 cells).
6365
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
Table 2. SAR for Acetonitrile Substitutions
a
b
c
Binding affinity Cat K versus FRET substrate; mean of greater than n = 4 tests, unless where stated. Mean of n = 2 tests. In vitro human liver
d
microsomal turnover mean of at least n = 2 tests (μL/min/mg). In vitro rat liver microsomal turnover mean of at least n = 2 tests (μL/min/mg).
Scheme 3. Metabolites Found after in Vitro Incubation of 14 with Rat Hepatocytes
essentially equipotent with 8 suggests that solvation of the
cation is not compromised by binding to these proteins.
Compound 9 again indicates that addition of a phenyl group
increases potency to yield a lead similar to 10 but with lower
selectivity (Cat S, 147 fold). The tricyclic carboline 13 stood
out from this first library by having high enzyme inhibition
(close to the tight binding limit for the assay31) and higher
selectivity (Cat S, >3300; Cat B, >1400) despite compromised
aqueous solubility and in vitro rat clearance.
The effects of substitution on the acetonitrile portion of the
molecule were explored with compounds 14 through 17 to
further develop the series from 13 (Table 2). Compound 14 is
as potent as 13, about 2-fold less stable in the in vitro HLM but
45-fold more soluble in water. This solubility difference is
especially noteworthy because the latter is more lipophilic than
the former. Enantiomers 15 and 16 are equipotent but over 10-
fold less potent against Cat K than either 13 or 14. One
interpretation of this observation is that 15 and 16 incur greater
energy costs in adopting their bound positions in the protein.
Similarly, the extra substitution of a methyl group on 15 had
increased solubility. Increasing the bulk of the alkyl substituent
appears to have an adverse effect on potency as shown by 17
Cat K IC50 is ∼800 fold less active.
Further testing of compound 14 found that high in vitro rat
liver microsome turnover translated into high hepatocyte
clearance and so explained the near liver blood flow clearance
and low bioavailability; however, we did not rule out absorption
as an additional contributing factor (Table 4). Thus, we
examined 14 in iv and po dog pharmacokinetics and were
encouraged to find that low in vitro translated into low in vivo
clearance and pleasingly the bioavailability was high at 82%.
While 14 is a good compound for further consideration the low
6366
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
bioavailability in rodents left uncertainties about the human
fraction absorbed so we sought to optimize clearance in rodents
while maintaining a good dog PK. Metabolite identification
from in vitro rat hepatocyte incubation of 14 highlighted that
an oxidation event occurred on the cyclohexyl ring (14-met-2),
as has been reported for structurally related compounds
(Scheme 3).22 However, major metabolites were found from
oxidation on the carboline ring and subsequent conjugation
(14-met-1, -4, -5, -6, -7), suggesting that substitution on this
ring may lead to reduced metabolism.
The structure of a cocrystal of 14 with Cat K revealed a
covalent bond between the nitrile of 14 and the active site C25
as the primary interaction, with the primary amide group
forming hydrogen bonds to N161 and G66, as has been
observed in related inhibitor complexes (Figure 3A).32 The
acid moeity of D61 was measured in this static structure at 3.95
Å from the carboline N−H (R5), suggesting a possible
hydrogen bond; however, both groups are exposed to solvent
so this is unlikely to contribute significantly to binding free
energy. The fluoro atom at R8 does not appear to make much
surface contact, and the structure suggests space to expand
along the protein surface. From these findings we decided to
explore SAR at R5, R6, R7, and R8, of the carboline, leaving R9
unsubstituted to maintain the interaction with Y67, monitoring
the effects on potency and aiming to reduce metabolism using
the in vitro RLM assay (Table 3). With LE and LLE already
high for 14, we focused on small substituents or heteroatom
substitution to combine SAR exploration and potential
metabolic blocks, while keeping lipophilicity under control to
retain aqueous solubility. The structural prototype 18 was
synthesized first and was nearly equipotent to 14, with almost
millimolar aqueous solubility. A fluorine scan (19 to 21 and
including 14) revealed that binding to Cat K compromised
fluorination at positions R7 and R9 although this substituent
can be accommodated without loss of potency at positions R6
and R8. Of these four compounds, 20, the least potent, was the
only compound with a significant increase in RLM stability.
Addition of a methyl at 22 (or N−H of carboline) resulted in a
small drop in activity consistent with the hypothesis that the
N−H did not form a H-bond to D61. Despite the initial
hypothesis, increasing the size of the group at R8 resulted in a
decrease in Cat K inhibition (23 to 27); we speculate that the
bulky groups have moved the tricycle out of the groove such
that the R9 is further from Y67. The target protein is more
tolerant of larger substituents at R6 (29−38), although none of
these led to a significant increase in potency relative to 18.
From the structure of 14 bound in Cat K, these groups would
point out to solvent and should make an impact on inhibition:
chloro 29, methyl 30, bromo 31, and CF3 33 are all less active,
despite being more lipophilic, and did not alter RLM clearence.
Lipophilicity is generally associated with poor ADMET
characteristics so we included polar substituents into our
compound design.33 The CH3SO2 substituent, which often
improves the physicochemical properties of compounds34 and
HLM stability,35,36 was introduced at three positions (27, 28,
and 38). Although this led to significantly increased stability in
both HLM and RLM assays, the accompanying loss of potency
was unacceptably large. Similarly, the 6-CN compound 32 was
predicted to increase stability and yielded a significant
reduction in RLM (80 μL/min/mg) and a small reduction in
measured log D7.4 and was similar in Cat K potency. However,
the 6-OCH3 (34, IC50 < 0.001 μM) and 6-OCF3 (36, IC50
0.002 μM) were more stable in HLM and RLM assays despite
the increase in lipophilicity, and comparing to 6-SCH3 35,
which showed a small reduction in activity, and 6-OCH2CH3,
we observed that the alkoxy substituent had an electronic effect
that maintained high Cat K activity and increased stability.
Finally, we synthesized 6-aza-carboline 39 following the SAR
that the 6 position tolerated substitution and allowed reduction
in lipophilicity. As hypothesized, the compound has improved
stability in both RLM and HLM stability but without any
reduction of Cat K potency.
Figure 3. Compound 34 bound to cathepsin K (PDB 4DMY). (A)
Identification of key H-bonding residues. (B) Surface of protein to
illustrate the fit of the cyclohexyl and carboline moieties.
Table 4 has comparison data for 14, 34, and 39 obtained
from in vitro assay and in vivo rat and dog PK experiments,
including scaling to human predicted data.37,38 Plasma protein
binding was not an issue because the fraction unbound
exceeded 10% in rat, dog, and human plasma. The increase
in the in vitro stability for 34 translated into improved
cyclohexane ring of a related inhibitor has been shown to
occupy the relatively shallow S2 pocket of the enzyme as does
14 (seen more clearly in Figure 3B).22,23 The tricyclic carboline
fits into a relatively flat open S3 groove, bound on one side by a
glycine-rich loop and Y67 on the other side, such that an
aromatic C−H (R9) is placed against it. The solvent-exposed
6367
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
Table 3. SAR for Carboline Substitutions
a
Cat K IC50
log
D7.4
aqueous solubility, pH 7.4
HLM
μL/min/mg
RLM
μL/min/mg
c
d
compd
R5
R6
R7
R8
(μM)
(μM)
18
19
20
14
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
H
H
F
H
H
F
H
H
H
F
0.001
<0.001
0.004
<0.001
0.010
0.003
0.003
0.041
2.7
3.0
3.0
3.0
3.2
2.8
3.7
2.4
4.2
4.1
−
1.5
3.6
3.0
3.1
2.5
3.8
2.7
−
910
470
540
580
128
300
130
822
74
26
55
36
41
57
170
87
29
33
−
14
<2.0
46
25
43
34
26
16
79
19
13
4.0
12
143
311
89
H
H
H
H
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
>250
322
>350
>350
(20)
82
H
F
Cl
CH3O
b
CF3
0.042
b
(CH3)2CH
0.050
84
−
(3.5)
b
CH3SO2
0.078
193
902
210
510
140
400
150
620
−
120
300
1600
1500
b
CH3SO2
H
H
H
H
H
H
H
H
H
H
H
0.034
7.2
H
H
H
H
H
H
H
H
H
H
0.002
0.003
0.004
226
179
164
80
CH3
Br
b
NC
0.002
b
CF3
0.005
116
67
CH3O
CH3S
CF3O
CH3CH2O
CH3SO2
<0.001
0.004
0.002
<0.001
0.017
0.001
303
53
4.0
3.1
1.7
2.1
62
23
25
a
Binding affinity Cat K versus FRET substrate; mean of greater than n = 4 tests, unless where stated. Compounds tested at least >4 μM for Cat S,
b
c
>2.9 μM for Cat B and as inactive for Cat L. Mean of n = 2 tests. In vitro human liver microsomal turnover mean of at least n = 2 tests (μL/min/
d
mg). In vitro rat liver microsomal turnover mean of at least n = 2 tests (μL/min/mg); values in parentheses for rat liver hepatocyte Clint mean n = 2
(μL/min/106 cells).
Table 4. Serum Protein Binding and Pharmacokinetic Data
a
compd species protein binding (% free) in vitro hepatocyte Clint (μL/min/106 cells) Clp (mL/min/kg) Vdss (L/kg) iv half-life (h) bioavailability (%)
b
14
34
39
rat
16
13
14
17
22
17
42
38
33
126 28
6.7 (n = 1)
<2.5 (n = 2)
23 16
63 7.8
10 4.9
2.3 0.3
1.2 0.5
[1.7]
0.6 0.2
1.8 0.2
[3.7]
3.8 2.1
82 21
[40]
c
dog
d
human
rat
[<5.3]
b
c
53 3.4
10 2.4
2.3 0.3
1.7 0.4
[2.0]
1.0 0.08
1.5 0.06
[2.2]
28 12
65 5.1
[53]
dog
2.5 2.3
5.3 4.6
9.9 7.9
3.1 1.0
<2.0 (n = 6)
d
human
rat
[8.7]
b
32 9.7
1.6 0.6
0.9 0.1
[1.0]
1.8 0.8
1.4 0.4
[2.1]
12 3.7
59
c
dog
7.2 4.8
d
human
[<5.4]
[42]
a
All results are a mean at least n = 2 experiments. All pharmacokinetic experiments reported as single dosed compounds in at least two experiments
b
involving two animals in each. Alderley Park Han Wistar male rats, dosed to fed animals, po 2.0 mg/kg as a suspension in 5% DMSO/95% HPMC/
TWEEN, iv dosed as solution in 40% DMA/water at 2.0 mg/kg. Beagle dog, dosed to fasted animals, po 1.0 mg/kg as a suspension in 5% DMSO/
95% HPMC/TWEEN, iv dosed as solution in 10% DMSO/90% Sorensons at 1.0 mg/kg. Numbers inside brackets are predicted values from scaling
c
d
to human from rat and dog results; see Supporting Information.
bioavailability in rat (28%), which was maintained in dogs
(65%). The in vitro human hepatic clearance of 39 was
consistently below the limit of detection, rat hepatic clearance
was low at 9.9 μL/min/106 cells, and this led to a prediction of
low clearance in vivo; however, we observed a CL of 32 mL/
min/kg. Metabolite identification of rat urine after oral dosing
found a significant number of metabolites and parent
compound (ca. 50% of total clearance).39 When scaled to
human, this extra component is likely to cause an approximate
30% increase in the human clearance.40 The consequent
decrease in human half-life made it unlikely that the compound
would be a suitable clinical candidate. For compound 34,
neither in vitro human hepatic Clint (predicted low clearance)
nor rodent renal elimination data (not detectable) suggested
6368
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
Scheme 4. Metabolites Found after in Vitro Incubation of 34 with Rat Hepatocytes
Table 5. SAR for Carboline Substitutions
RLM
(μL/min/mg)
Clp
bioavailability
(%)
a
b
compd Cat K IC50 (μM) log D7.4 aqueous solubility, pH 7.4 (μM) protein binding (% free)
(mL/min/kg)
c
18
44
0.004
0.025
2.7
3.0
910
580
23
12
143
46 7.9
8.2 5.1
>350
52 (n = 1)
0.0 (n = 1)
a
b
c
All results are a mean of at least n = 2 experiments. In vitro rat liver microsomal turnover mean of at least n = 2 tests (μL/min/mg). All
pharmacokinetic experiments reported as single dosed compounds in at least two experiments involving two animals in each. Alderley Park Han
Wistar male rats, dosed to fed animals, po 2.0 mg/kg as a suspension in 5% DMSO/95% HPMC/TWEEN, iv dosed as solution in 40% DMA/water
at 2.0 mg/kg.
that human clearance would be a risk; this plus the consistency
in oral absorption observed across species for the compound
confirmed it as a more likely clinical candidate.
protein fragment of type 1 collagen, as a measure of Cat K
inhibition.45−47 Compound 34 tested in a dog Cat K enzyme
FRET assay had a mean inhibition IC50 of 0.8 nM (n = 7), the
same as human enzyme inhibition. Blood level concentrations
of 34 and CTX-1 were measured as a function of time in this
PK/PD study at three doses (0.1, 0.3, and 2.0 mg/kg), the
results of which are summarized in Figure 4. A dose-dependent
increase in the area under the curve was observed (Figure 4a)
and concomitant reduction in [CTX-1] which appeared very
similar at the three doses studied (Figure 4b). This clearly
demonstrated target engagement, rapid onset of action, and the
in vivo potency of 34; a maximum effect level has been reached
with all three doses. These data suggest that inhibition of Cat K
need only occur for a short period of time to achieve maximum
inhibition.
To understand the effect of the OCH3 group at R6 on
pharmacokinetic properties, metabolites of 34 post in vitro
incubation with rat haptocytes were examined. This showed a
switch of major metabolite to oxidation of the cyclohexyl ring
(34-met-3) and reduced oxidation of the carboline aromatic
(34-met-4) (Scheme 4) when compared to 14 (Scheme 3). A
new metabolic route by oxidative demethylation and
conjugation was found (34-met-1 and -2) although they were
minor metabolites. This change in profile resulted in an overall
reduction in rate of turnover, perhaps by a change in
recognition by the Cyp450s involved. We also considered
that the effect of the OCH3 was to partially mask the N−H
donor of the carboline group. Reduction of the number of
hydrogen bond donors is an adopted approach for improving
pharmacokinetics and, in particular, absorption.41−44 As such,
compound 44 would test this hypothesis by complete removal
of the donor and could be compared with the unfunctionalized
carboline 18 (Table 5). Surprisingly, 44 had high in vitro rat
microsomal turnover and no bioavailability; in comparison, 18
had low bioavailability and high clearance, marking it hard to
establish the role of the H-bond donor of the carbolines from a
pharmokinetic point of view.
CONCLUSION
■
Directed screening of nitrile compounds revealed 3 and 4 as
highly potent cathepsin K (Cat K) inhibitors but with Cat S
activity and very poor stability to microsomes. Reducing
molecular complexity led to compounds such as 7 that revealed
key SAR and demonstrated that baseline physical properties
and in vitro stability were in fact excellent for this series. The
tricycle carboline P3 unit was discovered by hypothesis-based
design using existing structural information. Optimization using
small substituents, knowledge from matched molecular pairs,
and control of lipophilicity yielded compounds very close to the
In Vivo Pharmacology. The in vivo effects of 34 were
studied in dogs using levels of C-telopeptide (CTX-1), a
6369
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
short convergent route, compound 34 (AZD4996) was selected
as a clinical candidate.
EXPERIMENTAL SECTION
■
All solvents and chemicals used were reagent grade. Anhydrous
solvents tetrahydrofuran (THF) and dimethoxyethane (DME) were
purchased from Aldrich. Purity and characterization of compounds
were established by a combination of low resolution mass spectra (LC-
MS), obtained using a Waters liquid chromatography mass
spectrometry system, where purity was determined by UV absorption
at a wavelength of 254 nm and the mass ion determined by
electrospray ionization (Micromass instrument), and NMR analytical
1
techniques. All test compounds were >95% pure. H NMR spectra
were recorded using a AVIAN AV400 FT spectrometer or via a Flow
NMR process using an AVANCE 500 FT spectrometer. DMSO-d6 or
CDCl3 solvent was used, with the data expressed as chemical shifts in
ppm from internal standard TMS on the δ scale. Coupling constants
(J) values are reported in hertz (Hz). Splitting patterns are indicated as
follows: s, singlet; d, doublet; t, triplet; m, multiplet; br, broad peak.
1
Compounds 7, 11, 12, and 14− 27 when measured by H NMR at
ambient temperature showed differential rotameric peaks around the
tertiary amide group. In these cases, assignments for protons which
were measured distinctly have been expressed as fractions (e.g., 0.5H).
Higher temperature 1H NMR spectra have been recorded on
compounds 13, 14, and 39 for comparison and clear assignment.
The reverse phase column used was a 4.6 mm × 50 mm Phenomenex
Synergi Max-RP 80 Å, and the solvent system was water containing
0.1% formic acid and acetonitrile unless otherwise stated. A typical run
was 5.5 min with a 4.0 min gradient from 0% to 95% acetonitrile.
Purification by column chromatography was typically performed using
silica gel (Merck 7734 grade) and solvent mixtures, and gradients are
described herein. Purification by reverse phase high performance
chromatography was typically performed using a Perkin-Elmer
instrument using UV detection at 254 nm and a C18 1500 × 21.2
mm Phenomenex column 100 Å. Acidic conditions (0.1% to 0.5%
formic acid) or basic conditions (ammonia to pH 10) were used with
gradiant solvent mixtures of acetonitrile and water. SCX columns were
supplied from International Sorbent Technology and used as directed
in this specific case .
General Procedure for the Preparation of 7, 12−14, 34, and
39. An appropriate amine (1.0 equiv) was added to (3aR,7aR)-
hexahydroisobenzofuran-1,3-dione (6) (1.0 equiv) and DIPEA (3.0
equiv) in CH2Cl2 to a concentration of 0.1 M. The resulting solution
was stirred at 20 °C for 18 h and then either aminoacetonitrile
hydrochloride (13c) or 14c (3.0 equiv) was added followed by HATU
(1.1 equiv) and further DIPEA (5.0 equiv). The resulting suspension
was stirred at 20 °C for 72 h. The solution was diluted with CH2Cl2
(equal volume), partitioned with 50% brine (equal volume), dried
(Na2SO4), concentrated in vacuo, and adsorbed onto silica. Flash
chromatography (silica, 0−100% EtOAc in isohexane) yielded the
desired compound, usually as a solid, or alternatively preparative
HPLC was used where stated.
Figure 4. Dog PK/PD experiment for compound 34. Each point was
mean n = 2 with the same blood sample taken at a given time after
compound dosing split to measure [34] and [CTX-1]. (a) Total
○
plasma [34] from the dog PK experiment: , PK[34]/nM at 0.1 mg/
◆
kg uid; +, PK[34]/nM at 0.3 mg/kg uid/nM; , PK[34]/nM at 2.0
○
,
mg/kg uid. (b) ELISA-measured reduction [CTX-1] from baseline:
[CTX-1] as % baseline 0.1 mg/kg uid; +, [CTX-1] as % baseline 0.3
◆
mg/kg uid; , [CTX-1] as % baseline 2.0 mg/kg uid.
(1R,2R)-N-(Cyanomethyl)-N,N-dimethylcyclohexane-1,2-di-
carboxamide (7). Dimethylamine (2 M in THF, 0.65 mL, 1.30
mmol) was utilized with 1-aminoacetonitrile hydrochloride (192 mg,
2.08 mmol) (13c) to yield 7, after purification via prep HPLC, as a
white powder (192 mg, 61%). HRMS (ES+) for C12H20O2N3 (M+ +
H): calcd 238.1550; found, 238.1551. 1H NMR (400 MHz, CDCl3) δ
1.21−1.48 (m, 4H), 1.48−1.67 (m, 1H), 1.67−2.00 (m, 4H), 2.70
(ddd, J = 12.5, 10.6, 3.6 Hz, 1H), 2.82−2.96 (m, 4H), 3.08 (s, 3H),
3.99 (dd, J = 17.4, 5.9 Hz, 1H), 4.10 (dd, J = 17.4, 5.9 Hz, 1H), 7.09
(s, 1H).
(1R,2R)-N-(Cyanomethyl)-2-(4-phenylpiperazine-1-
carbonyl)cyclohexanecarboxamide (12). N-(4-Phenyl)piperazine
(162 mL, ∼1.0 mmol) was utilized with 1-aminoacetonitrile
hydrochloride (29.0 mg, 0.52 mmol) (13c) to yield 12, after
purification via prep HPLC, as a solid gum (80.3 mg, 52%). HRMS
(ES+) for C20H27O2N4 (M+ + H): calcd 355.2129; found, 355.2130.
1H NMR (500 MHz, DMSO) δ 1.11−1.50 (m, 4H), 1.62−1.92 (m,
desired profile, such as 39. We were surprised to find an OCH3
at R6 (34) had reduced metabolic turnover and thus improved
pharmacokinetics in rodents, as well as the benefit of a simple
high-yielding synthesis. The OCH3 group also retained Cat K
inhibition, although the mechanism for this and the increased
stability are not understood. Selectivity was also maximized
with only Cat B inhibition observed at 2.97 μM (>3000 fold)
and hERG ionworks screening inactive (>100 μM). Comparing
to balicatib, synthesized and tested in our hands (Cat K 0.005
μM, Cat B 4.75 μM), and bdanacatib (Cat K <0.001 μM, Cat S
0.183 μM), we found our compounds as potent but with higher
selectivity. Our in vivo pharmacology experiment in dogs
demonstrated target engagement and a maximum inhibition of
activity at 0.1 mg/kg, showing the potential of this compound
as a clinical candidate. Considering all of these factors and the
6370
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
4H), 2.58 (t, J = 16.0 Hz, 1H), 2.94 (t, J = 16.0 Hz, 1H), 3.01−3.27
(m, 4H), 3.59−3.70 (m, 4H), 4.05 (d, J = 5.6 Hz, 2H), 6.81 (t, J = 7.2
Hz, 1H), 6.94 (d, J = 7.2 Hz, 2H), 7.23 (t, J = 7.2 Hz, 2H), 8.49 (t, J =
5.6 Hz, 1H).
for 2 h, cooled to RT, and concentrated in vacuo. The residue was
partitioned between 2 N NaOH (aq) (10 mL) and CH2Cl2 (2 × 30
mL), and the combined organic fractions were dried (Na2SO4),
concentrated in vacuo, and adsorbed onto silica for purification by
flash column chromatography (0−10% MeOH/CH2Cl2). This gave 45
as a pale yellow gum which was used crude in the next reaction (7.70
( 1 R , 2 R ) - N - ( C y a n o m e t h y l ) - 2 - ( 8 - fl u o r o - 1 , 3 , 4 , 5 -
t e t r a h y d r o p y r i d o [ 4 , 3 - b ] i n d o l e - 2 - c a r b o n y l ) -
cyclohexanecarboxamide (13). 8-Fluoro-1,3,4,5-tetrahydropyrido-
[4,3-b]indole 13a (162 mg, 0.87 mmol) was utilized with 1-
aminoacetonitrile hydrochloride (69.0 mg, 0.75 mmol) (13c) to
yield 13, after purification via prep HPLC, as a solid gum (134 mg,
57%). HRMS (ES+) for C21H24O2N4F (M+ + H): calcd 383.1878;
found, 383.1880. 1H NMR (500 MHz, DMSO) δ 1.13−1.50 (m, 4H),
1.62−1.92 (m, 4H), 2.58 (t, J = 10.4 Hz, 1H), 2.75−3.09 (m, 4H),
3.48−3.75 (m, 3H), 4.00−4.16 (m, 2H), 6.88 (d, J = 3.8 Hz, 2H), 6.97
1
g, 100% yield). H NMR (400 MHz, DMSO) δ 2.34 (d, J = 5.6 Hz,
2H), 2.62−2.36 (m, 6H), 3.52 (s, 2H), 6.68 (ddd, J = 7.1, 4.9, 0.8 Hz,
1H), 7.04 (s, 1H), 7.51−7.06 (m, 5H), 7.59−7.51 (m, 1H), 8.05 (dd, J
= 4.9, 1.1 Hz, 1H), 9.47 (s, 1H).
N-Benzyl-2,3,4,5-tetrahydro-1-benzylpyrido[4,3-b]-7-azain-
dole (46). Polyphosphoric acid (60 g) was added to 45 (7.69 g, 27.5
mmol) and the mixture stirred gently at 150 °C for 24 h. The mixture
was cooled to RT, and ice (50 g) was added to break up the gum. The
reaction mixture was made basified with 2 M NaOH (aq) and
extracted with EtOAc (3 × 300 mL). The combined organic extracts
were treated with brine (90 mL), dried (Na2SO4), concentrated in
vacuo, and adsorbed onto silica for purification by flash column
chromatography (0−10% MeOH/CH2Cl2). The mustard colored
solid obtained (3.80 g) was triturated with a small volume of CH2Cl2,
filtered, and dried to furnish the desired compound as a sand colored
solid 57 (3.00 g, 42% yield). LCMS (+ve ESI): tR = 0.97 min, 264 (M
1
(dt, J = 13.7, 5.9 Hz, 2H), 8.48 (t, J = 5.5 Hz, 1H). H NMR (500
MHz, DMSO) δ 1.18−1.48 (m, 4H), 1.62−1.91 (m, 4H), 2.79 (t, J =
11.7 Hz, 1H), 2.75−2.93 (m, 2H), 3.03(t, J = 11.7 Hz, 1H), 3.80 (s,
1H), 3.87−4.06 (m, 2H), 4.64 (t, J = 16.6 Hz, 2H), 6.83 (td, J = 9.2,
2.4 Hz, 1H), 7.20 (d, J = 7.0 Hz, 1H), 7.25 (dd, J = 8.7, 4.5 Hz, 1H),
8.08 (s, 1H), 10.64 (s, 1H).
(1R,2R)-N-(1-Cyanocyclopropyl)-2-[(8-fluoro-1,3,4,5-tetrahy-
d r o - 2 H - p y r i d o [ 4 , 3 - b ] i n d o l - 2 - y l ) c a r b o n y l ] -
cyclohexanecarboxamide (14). 13a (798 mg, 4.20 mmol) was
utilized with 14c (585 mg, 5.04 mmol) to yield 14, after purification
silica chromatography, as a solid gum (212 mg, 12%). HRMS (ES+)
1
+ H)+. H NMR (400 MHz, DMSO) δ 2.80 (m, 4H), 3.57 (s, 2H),
3.73 (s, 2H), 6.94 (dd, J = 7.8, 4.7 Hz, 1H), 7.21−7.42 (m, 5H), 7.67
(dd, J = 7.8, 1.2 Hz, 1H), 8.07 (dd, J = 4.7, 1.5 Hz, 1H), 11.32 (s, 1H).
2,3,4,5-Tetrahydro-1-benzylpyrido[4,3-b]-7-azaindole (39a).
46 (2.90 g, 11.0 mmol), NH4CO2H(s) (2.78 g, 44.0 mmol), and 20%
PdOH on carbon (290 mg) were suspended in EtOH (200 mL) and
stirred under reflux. After 1 h, more NH4CO2H(s) (695 mg, 11.0
mmol) was added and refluxing continued for 1 h. The catalyst was
filtered off through Celite and washed with a small volume of CH2Cl2,
and the combined filtrate was concentrated in vacuo and dried under
vacuum to furnish the desired compound as off-white solid 39a (1.90
1
for C23H26O2N4F (M+ + H): calcd 409.2034; found, 409.2036. H
NMR (400 MHz, CDCl3) δ 0.77−0.94 (m, 1H), 0.95−1.54 (m, 7H),
1.75−1.91 (m, 4H), 2.61 (“t”, J = 11.4 Hz, 1H), 2.68−2.94 (m, 3H),
3.64−3.80 (m, 1H), 3.88−4.03 (m, 0.5H), 4.20−4.33 (m, 0.5H), 4.60
(d, J = 15.7 Hz, 0.5H), 4.69 (s, 1H), 4.89 (d, J = 15.8 Hz, 0.5H), 6.53
(s, 0.5H), 6.61 (s, 0.5H), 6.89 (td, J = 11.3, 2.4 Hz, 1H), 7.08 (td, J =
9.4, 2.3 Hz, 1H), 7.20 (td, J = 8.8, 2.3 Hz, 1H), 7.93 (s, 1H). 1H NMR
(500 MHz, DMSO) δ 0.88 (s, 1H), 1.01 (s, 1H), 1.19−1.45 (m, 6H),
1.70−1.85 (m, 4H), 2.79−2.90 (m, 2H), 3.01 (td, J = 11.9, 3.6 Hz,
1H), 3.81 (s, 1H), 3.88 (dt, J = 11.3, 5.7 Hz, 1H), 4.63 (s, 2H), 6.82
(dt, J = 9.2, 2.6 Hz, 1H), 7.22 (d, J = 9.2 Hz, 1H), 7.26 (dd, J = 8.8, 4.6
Hz, 1H), 8.26 (s, 1H), 10.65 (s, 1H).
1
g, 100% yield). H NMR (400 MHz, DMSO) δ 2.46−2.55 (m, 2H),
2.67 (t, J = 5.6 Hz, 2H), 3.02 (t, J = 5.6 Hz, 2H), 3.83 (s, 2H), 6.95
(dd, J = 7.7, 4.7 Hz, 1H), 7.70 (dd, J = 7.7, 1.4 Hz, 1H), 8.07 (dd, J =
4.7, 1.4 Hz, 1H), 11.25 (s, 1H).
6-Methoxy-2,3,4,5-tetrahydro-1H-pyrido[4,3-b]indole (34a).
Piperidin-4-one hydrochloride (41) (4.66 g, 34.5 mmol) was added to
a suspension of (2-methoxyphenyl)hydrazine hydrochloride (6.00 g,
34.4 mmol) in ethanol (80 mL) and concd HCl (6.0 mL), and the
resulting solution was heated at 80 °C for 3 h. The reaction was
allowed to cool to room temperature overnight, and the resulting
precipitated solid was filtered, washed with chilled ethanol (20 mL),
and dried under vacuum to give a salt. This was stirred in water (100
mL) and made basic with 2 M NaOH(aq) (∼5 mL), and the resulting
solid was filtered, azeotroped with toluene (30 mL, rotary evaporator),
and dried under vacuum to yield 34a (3.84 g, 55%) as an off-white
solid. LCMS (+ve ESI): tR = 0.85 min, 203.3 (M + H)+. 1H NMR (400
MHz, DMSO) 2.63−2.66 (m, 2H), 2.99−303 (m, 2H), 3.83 (‘s’, 2H),
3.89 (s, 3H), 6.59 (d, J = 7.5 Hz, 1H), 6.83 (“t”, J = 7.8 Hz, 1H), 6.93
(d, J = 7.8 Hz, 1H), 10.72 (s, 1H).
(1R,2R)-N-(1-Cyanocyclopropyl)-2-[1,3,4,5-tetrahydro-1H-
p y r i d o [ 4 , 3 - b ] - 7 - a z a i n d o l - 2 - y l ) c a r b o n y l ] -
cyclohexanecarboxamide (39). 39a (200 mg, 1.15 mmol) was
utilized with 14c (177 mg, 1.20 mmol) to furnish 39, after purification,
as an off-white solid (40.0 mg, 9% yield). HRMS (ES+) for
1
C22H26O2N5 (M+ + H): calcd 392.2081; found, 392.2084. H NMR
(400 MHz, DMSO, 303 K) δ 0.72 (s, 1H), 0.69−0.80 (dd, J = 41.8,
7.8 Hz, 0.5H), 0.93−1.30 (m, 1H), 1.15−1.45 (m, 6H), 1.62−1.85 (m,
4H), 2.67−2.75 (m, 0.5H), 2.78−2.84 (m, 0.5H), 2.95−3.05 (m, 1H),
3.69−3.82 (m, 0.5H), 3.84−3.94 (m, 1.5H), 4.52 (d, J = 15.8 Hz,
0.5H), 4.65 (d, J = 15.8 Hz, 0.5H), 4.75 (s, 1H), 7.02 (“ddd”, J = 16.6,
7.7, 4.8 Hz, 1H), 7.83 (d, J = 7.6 Hz, 0.5H), 7.93 (d, J = 7.6 Hz, 0.5H),
8.13 (dd, J = 8.5, 4.4 Hz, 1H), 8.71 (d, J = 3.7 Hz, 1H), 11.44 (s, 1H).
1H NMR (500 MHz, DMSO, 373 K) δ 0.84−0.86 (m, 1H), 0.94 (d, J
= 62.4 Hz, 1H), 1.16−1.53 (m, 6H), 1.70−1.92 (m, 4H), 2.81 (m,
1H), 3.02 (td, J = 11.8, 3.5 Hz, 1H), 3.83 (b, 1H), 3.89 (dt, J = 13.3,
5.6 Hz, 1H), 4.66 (s, 2H), 6.98 (dd, J = 7.2, 4.7 Hz, 1H), 7.80 (d, J =
7.2 Hz, 1H), 8.12 (dd, J = 4.7, 1.5 Hz, 1H), 8.27 (s, 1H), 11.04 (s,
1H).
Assay for the Identification of Cat K Inhibitors. QFRET
Technology (quenched fluorescent resonance energy transfer) was
used to measure the inhibition by test compounds of Cat K-mediated
cleavage of the synthetic peptide Z-Phe-Arg-AMC. Compounds were
screened at twelve concentrations (0.0003−10 μM), on two separate
occasions, and the mean IC50 values reported.
A concentration of 0.5 nM [final] rhuman Cat K (prepared in-
house) in phosphate buffer was added to a 384-well black microtiter
plate containing investigative compounds. The enzyme and compound
were preincubated at room temperature for 30 min before the addition
of 50 μM [final] Z-Phe-Arg-AMC synthetic substrate in sodium
acetate/EDTA buffer at pH 5.5. The plates were covered and
incubated for 1 h at room temperature and protected from light.
Following the incubation, the reaction was stopped with 7.5% [final]
(1R,2R)-N-(1-Cyanocyclopropyl)-2-[(6-methoxy-1,3,4,5-tetra-
h y d r o - 2 H - p y r i d o [ 4 , 3 - b ] i n d o l - 2 - y l ) c a r b o n y l ] -
cyclohexanecarboxamide (34). 34a (3.84 g, 19.0 mmol) was
utilized with 1-aminocyclopropanecarbonitrile hydrochloride 14c
(6.76 g, 57.0 mmol) to yield 34 after purification via flash silica
chromatography and crystallization slurry stirring over 48 h (6.27 g,
78%) as a white powder. HRMS (ES+) for C24H29O3N4 (M+ + H):
1
calcd 421.2234; found, 421.2233. H NMR (400 MHz, DMSO) δ
0.73−0.83 (m, 1H), 0.87−0.91 (m, 0.5H), 0.87−1.03 (m, 1H), 1.17−
1.40 (m, 6H), 1.61−1.82 (m, 5H), 2.63−2.70 (m, 1H), 2.75−2.82 (m,
0.5H), 2.84−3.05 (m, 1.5H), 3.64−3.72 (m, 0.5H), 3.94−3.73 (m,
5H), 4.71 (s, 1H), 6.64 ("t", J = 8.2 Hz, 1H), 6.90 (qn, J = 7.8 Hz, 1H),
6.97 (d, J = 7.8 Hz, 0.5H), 7.10 (d, J = 7.8 Hz, 0.5H), 8.64 (s, 1H),
10.91 (s, 1H).
1-Benzylpiperidin-4-one Pyridin-2-yl Hydrazone (45). 2-
Hydrazinopyridine dihydrochloride (5.00 g, 27.5 mmol) and 1-
benzylpiperidin-4-one (6.18 g, 27.5 mmol) were suspended in EtOH
(70 mL). AcOH (2 mL) was added and the mixture stirred at reflux
6371
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
acetic acid. Relative fluorescence was measured using the Ultra plate
reader at a wavelength of 360 nm excitation and 425 nm emission.
Data was corrected for background fluorescence (minimum controls
without enzyme). This data was used to plot inhibition curves and
calculate IC50 values by nonlinear regression using a variable slope,
offset=zero model in Origin 7.5 analysis package. Reproducibility of
data was assessed using a quality control statistical analysis package
whereby internal variability of the assayed indicated a repeat testing (n
= 3) if pIC50 SD was >0.345.
Assay for the Identification of Cat L Inhibitors. In a similar
manner using the same QFRET technology compounds were tested
for inhibition of cathepsin L-mediated cleavage of a synthetic peptide
Z-Phe-Arg-AMC (10 μM final concentration) using 0.1 nM [final] r-
human Cat L (prepared in-house). A potassium phosphate buffer was
used to maintain pH at 6.4.
ASSOCIATED CONTENT
■
S
* Supporting Information
1
Details of synthesis of final compounds and intermediates; H
NMR for compounds 13, 14, and 39 at 373 K; rat and human
hepatic metabolite identification for 14 and 34, and rat in vivo
urinalysis metabolite identification for 39. Table of data: dog
PK/PD [CTX-1] for 34; baseline [CTX-1] levels over 48 h for
comparison of the magnitude of effect; dose to man prediction
calculations; crystallization coordinates for Cat K/34 cocom-
plex. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +44 (0) 1625 517673. Fax: +44 (0) 1625 516667. E-
mail: al.dossetter@astrazeneca.com.
Assay for the Identification of Cat S Inhibitors. In a similar
manner using the same QFRET technology, compounds were tested
for inhibition of cathepsin S-mediated cleavage of a synthetic peptide
Z-Val-Val-Arg-AMC (100 μM final concentration) using 0.5 nM
[final] r-human Cat S (prepared in-house). A potassium phosphate
buffer was used to maintain pH at 6.4.
Notes
The authors declare no competing financial interest.
Assay for the Identification of Cat B Inhibitors. In a similar
manner using the same QFRET technology, compounds were tested
for inhibition of cathepsin B-mediated cleavage of a synthetic peptide
Z-Arg-Arg-AMC (100 μM final concentration) using 0.5 nM [final] r-
human Cat B (prepared in-house). A potassium phosphate buffer was
used to maintain pH at 6.4.
Assay for the Identification of Cat K inhibitors in Dog. In a
similar manner using the same QFRET technology, compounds were
tested for inhibition of cathepsin K-mediated cleavage of a synthetic
peptide Z-Phe-Arg-AMC (15 μM final concentration) using 0.5 nM
[final] r-dog Cat K (prepared in-house). A sodium acetate buffer was
used to maintain pH at 5.5.
Dog Beagle Blood Bone Remodeling via ELISA Detection of
CTK-1. All pharmacodynamic work was carried out on a small colony
of 6 male dogs (Alderley Park Beagle), at the age of 12−14 months,
weighing 14−17 kg. All studies were started at the same time of day to
standardize results against the normal 24 h circadian rhythm. Dogs
were typically dosed with test compound by gavage (po), and blood
samples for both PK and PD analysis were taken at predetermined
time points via the jugular vein. Formulation of test compounds was in
5% DMSO + 95% HPMC/Tween, in a typical dosing volume of 5
mL/kg. PD sampling (serum): blood samples were allowed to clot for
a minimum of 1 h, at room temperature, prior to being spun at 2000
rpm for 15 min. The serum was then removed and placed into
appropriate tubing and stored at −20 °C until analysis, after which the
samples were stored at −80 °C. Bone resorption: samples were
analyzed for CTX-1 using a Nordic Bioscience Serum CrossLaps
ELISA kit (4CRL4000). ELISA was carried out according to protocol,
while data were analyzed using Molecular Devices Softmax Pro
software installed on a Molecular Devices Versamax tunable plate
reader. PK sampling (whole blood): 200 μL of whole blood was placed
into the appropriate well of a 96-well plate, containing 200 μL of
distilled water. Analysis was by LC-MS against standard concentration
curves for compound 1, after protein precipitation with cold CH3CN
and centrifugation.
ABBREVIATIONS USED
■
OA, osteoarthritis; OP, osteoporosis; MBD, metastatic bone
disease; Cat K, cathepsin K; Cat S, cathepsin S; cat B, cathepsin
B; Cat L, cathepsin L; HATU, O-(7-azabenzotriazol-1-yl)-
N,N,N′,N′-tetramethyluronium hexafluorophosphate; PYBOP,
benzotriazolyloxy-tris(pyrrolidino)phosphonium hexafluoro-
phosphate; DIPEA, diisopropylethylamine; LLE, ligand lip-
ophilicity efficiency; HLM, human liver microsomes; RLM, rat
liver microsomes; CTX-1, C-telopeptide fragment of type 1
collagen; QFRET, quenched fluorescent resonance energy
transfer technology; CCP4, collaborative computational proj-
ect, number 4
REFERENCES
■
(1) Hunter, D. J.; Felson, D. T. Osteoarthritis. BMJ [Br. Med. J.]
2006, 332, 639−642.
(2) Lawrence, R. C.; Felson, D. T.; Helmick, C. G.; Arnold, L. M.;
Choi, H.; Deyo, R. A.; Gabriel, S.; Hirsch, R.; Hochberg, M. C.;
Hunder, G. G.; Jordan, J. M.; Katz, J. N.; Kremers, H. M.; Wolfe, F.
National Arthritis Data Workgroup Estimates of the prevalence of
arthritis and other rheumatic conditions in the United States. Part II.
Arthritis Rheum. 2008, 58, 26−35.
(3) WHO Disease and Injury country estimates http://www.who.
int/healthinfo/global_burden_disease/estimates_country/en/index.
html.
(4) Felson, D. T.; Chaisson, C. E.; Hill, C. L.; Totterman, S. M.;
Gale, M. E.; Skinner, K. M.; Kazis, L.; Gale, D. R. The association of
bone marrow lesions with pain in knee osteoarthritis. Ann. Intern. Med.
2001, 134, 541−549.
(5) Alzahrani, K.; Gandhi, R.; Debeer, J.; Petruccelli, D.; Mahomed,
N. Prevalence of clinically significant improvement following total
knee replacement. J. Rheumatol. 2011, 38, 753−759.
(6) Teitelbaum, S. L. Bone resorption by osteoclasts. Science
(Washington, D. C.) 2000, 289, 1504−1508.
(7) Gowen, M.; Lazner, F.; Dodds, R.; Kapadia, R.; Feild, J.; Tavaria,
M.; Bertoncello, I.; Drake, F.; Zavarselk, S.; Tellis, I.; Hertzog, P.;
Debouck, C.; Kola, I. Cathepsin K knockout mice develop
osteopetrosis due to a deficit in matrix degradation but not
demineralization. J. Bone Miner. Res. 1999, 14, 1654−1663.
(8) Gupta, S.; Singh, R. K.; Dastidar, S.; Ray, A. Cysteine cathepsin S
as an immunomodulatory target: present and future trends. Expert
Opin. Ther. Targets 2008, 12, 291−299.
Crystallography. Diffraction data for compound 14 were collected
at 100 K on a beamline PX at the Swiss Light Source (SLS, Villigen,
Switzerland). The structure was solved by molecular replacement and
refined to a final resolution of 2.33 Å and an R-factor of 25.0% using
the Collaborative Computational Project, Number 4 CCP4 (http://
www.ccp4.ac.uk) CCP4 and Coot (http://www.ysbl.york.ac.uk/
∼emsley/coot) software packages. This structure has been deposited
in the RSB (home.rcsb.org) Protein Data Bank (www.rcsb.org/pdb/
home/home.do) with reference code 4DMY.
(9) Stoch, S. A.; Wagner, J. A. Cathepsin K inhibitors: a novel target
for osteoporosis therapy. Clin. Pharmacol. Ther. 2008, 83, 172−176.
6372
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
(10) Yasuda, Y.; Kaleta, J.; Broemme, D. The role of cathepsins in
osteoporosis and arthritis: rationale for the design of new therapeutics.
Adv. Drug Delivery Rev. 2005, 57, 973−993.
(11) Yang, M.; Sun, J.; Zhang, T.; Liu, J.; Zhang, J.; Shi, M. A.;
Darakhshan, F.; Guerre-Millo, M.; Clement, K.; Gelb, B. D.; Dolgnov,
G.; Shi, G. Deficiency and Inhibition of Cathepsin K Reduce Body
Weight Gain and Increase Glucose Metabolism in Mice. Arterioscler.,
Thromb., Vasc. Biol. 2008, 28, 2202−2208.
(12) Vasiljeva, O.; Reinheckel, T.; Peters, C.; Turk, D.; Turk, V.;
Turk, B. Emerging roles of cysteine cathepsins in disease and their
potential as drug targets. Curr. Pharm. Des. 2007, 13, 387−403.
(13) Gauthier, J. Y.; Chauret, N.; Cromlish, W.; Desmarais, S.;
Duong, L. T.; Falgueyret, J.; Kimmel, D. B.; Lamontagne, S.; Leger, S.;
LeRiche, T.; Li, C. S.; Masse, F.; McKay, D. J.; Nicoll-Griffith, D. A.;
Oballa, R. M.; Palmer, J. T.; Percival, M. D.; Riendeau, D.; Robichaud,
J.; Rodan, G. A.; Rodan, S. B.; Seto, C.; Therien, M.; Truong, V.;
Venuti, M. C.; Wesolowski, G.; Young, R. N.; Zamboni, R.; Black, W.
C. The discovery of odanacatib (MK-0822), a selective inhibitor of
cathepsin K. Bioorg. Med. Chem. Lett. 2008, 18, 923−928.
(14) Stoch, S. A.; Zajic, S.; Stone, J.; Miller, D. L.; Van, D., K.;
Gutierrez, M. J.; De, D., M.; Liu, L.; Liu, Q.; Scott, B. B.; Panebianco,
D.; Jin, B.; Duong, L. T.; Gottesdiener, K.; Wagner, J. A. Effect of the
Cathepsin K Inhibitor Odanacatib on Bone Resorption Biomarkers in
Healthy Postmenopausal Women: Two Double-Blind, Randomized,
Placebo-Controlled Phase I Studies. Clin. Pharmacol. Ther. (N. Y., NY,
U. S.) 2009, 86, 175−182.
(15) Eisman, J. A.; Bone, H. G.; Hosking, D. J.; McClung, M. R.;
Reid, I. R.; Rizzoli, R.; Resch, H.; Verbruggen, N.; Hustad, C. M.; Da,
S. C.; Petrovic, R.; Santora, A. C.; Ince, B. A.; Lombardi, A. Odanacatib
in the treatment of postmenopausal women with low bone mineral
density: three-year continued therapy and resolution of effect. J. Bone
Miner. Res. 2011, 26, 242−251.
(16) Yamashita, D. S.; Marquis, R. W.; Xie, R.; Nidamarthy, S. D.;
Oh, H.; Jeong, J. U.; Erhard, K. F.; Ward, K. W.; Roethke, T. J.; Smith,
B. R.; Cheng, H.; Geng, X.; Lin, F.; Offen, P. H.; Wang, B.; Nevins, N.;
Head, M. S.; Haltiwanger, R. C.; Narducci, S.; Amy, A.; Liable-Sands,
L. M.; Zhao, B.; Smith, W. W.; Janson, C. A.; Gao, E.; Tomaszek, T.;
McQueney, M.; James, I. E.; Gress, C. J.; Zembryki, D. L.; Lark, M.
W.; Veber, D. F. Structure−Activity Relationships of 5-, 6-, and 7-
Methyl-Substituted Azepan-3-one Cathepsin K Inhibitors. J. Med.
Chem. 2006, 49, 1597−1612.
(17) Kumar, S.; Dare, L.; Vasko-Moser, J. A.; James, I. E.; Blake, S.
M.; Rickard, D. J.; Hwang, S.; Tomaszek, T.; Yamashita, D. S.;
Marquis, R. W.; Oh, H.; Jeong, J. U.; Veber, D. F.; Gowen, M.; Lark,
M. W.; Stroup, G. A highly potent inhibitor of cathepsin K (relacatib)
reduces biomarkers of bone resorption both in vitro and in an acute
model of elevated bone turnover in vivo in monkeys. Bone (San Diego,
CA, U. S.) 2007, 40, 122−131.
(18) Bailey, A.; Pairaudeau, G.; Patel, A.; Thom, S.
WO2004000825A1, 2003.
(19) Black, W. C.; Percival, M. D. The consequences of
lysosomotropism on the design of selective cathepsin K inhibitors.
ChemBioChem 2006, 7, 1525−1535.
(20) Falgueyret, J.; Desmarais, S.; Oballa, R.; Black, W. C.; Cromlish,
W.; Khougaz, K.; Lamontagne, S.; Masse, F.; Riendeau, D.; Toulmond,
S.; Percival, M. D. Lysosomotropism of Basic Cathepsin K Inhibitors
Contributes to Increased Cellular Potencies against Off-Target
Cathepsins and Reduced Functional Selectivity. J. Med. Chem. 2005,
48, 7535−7543.
(21) Hann, M. M.; Leach, A. R.; Harper, G. Molecular Complexity
and Its Impact on the Probability of Finding Leads for Drug Discovery.
J. Chem. Inf. Comput. Sci. 2001, 41, 856−864.
(22) Crane, S. N.; Black, W. C.; Palmer, J. T.; Davis, D. E.; Setti, E.;
Robichaud, J.; Paquet, J.; Oballa, R. M.; Bayly, C. I.; McKay, D. J.;
Somoza, J. R.; Chauret, N.; Seto, C.; Scheigetz, J.; Wesolowski, G.;
Masse, F.; Desmarais, S.; Ouellet, M. β-Substituted Cyclohexanecar-
boxamide: A Nonpeptidic Framework for the Design of Potent
Inhibitors of Cathepsin K. J. Med. Chem. 2006, 49, 1066−1079.
(23) Robichaud, J.; Bayly, C. I.; Black, W. C.; Desmarais, S.; Leger, S.;
Masse, F.; McKay, D. J.; Oballa, R. M.; Paquet, J.; Percival, M. D.;
Truchon, J.; Wesolowski, G.; Crane, S. N. β-Substituted cyclo-
hexanecarboxamide cathepsin K inhibitors: Modification of the 1,2-
disubstituted aromatic core. Bioorg. Med. Chem. Lett. 2007, 17, 3146−
3151.
(24) Berkessel, A.; Glaubitz, K.; Lex, J. Enantiomerically pure β-
amino acids: a convenient access to both enantiomers of trans-2-
aminocyclohexanecarboxylic acid. Eur. J. Org. Chem. 2002, 2948−2952.
(25) Harbert, C. A.; Plattner, J. J.; Welch, W. M.; Weissman, A.; Koe,
B. K. Neuroleptic activity in 5-aryltetrahydro-γ-carbolines. J. Med.
Chem. 1980, 23, 635−643.
(26) Ram, S.; Spicer, L. D. Debenzylation of N-benzylamino
derivatives by catalytic transfer hydrogenation with ammonium
formate. Synth. Commun. 1987, 17, 415−418.
(27) Waring, M. J. Lipophilicity in drug discovery. Expert Opin. Drug
Discovery 2010, 5, 235−248.
(28) Andersson, S.; Armstrong, A.; Bjoere, A.; Bowker, S.; Chapman,
S.; Davies, R.; Donald, C.; Egner, B.; Elebring, T.; Holmqvist, S.;
Inghardt, T.; Johannesson, P.; Johansson, M.; Johnstone, C.; Kemmitt,
P.; Kihlberg, J.; Korsgren, P.; Lemurell, M.; Moore, J.; Pettersson, J. A.;
Pointon, H.; Ponten, F.; Schofield, P.; Selmi, N.; Whittamore, P.
Making medicinal chemistry more effective-application of Lean Sigma
to improve processes, speed and quality. Drug Discovery Today 2009,
14, 598−604.
(29) Teague, S. J.; Davis, A. M.; Leeson, P. D.; Oprea, T. The design
of leadlike combinatorial libraries. Angew. Chem., Int. Ed. 1999, 38,
3743−3748.
(30) Oprea, T. I.; Davis, A. M.; Teague, S. J.; Leeson, P. D. Is There a
Difference between Leads and Drugs? A Historical Perspective. J.
Chem. Inf. Comput. Sci. 2001, 41, 1308−1315.
(31) Williams, J. W.; Morrison, J. F. The kinetics of reversible tight-
binding inhibition. Methods Enzymol. 1979, 63, 437−467.
(32) Li, C. S.; Deschenes, D.; Desmarais, S.; Falgueyret, J.; Gauthier,
J. Y.; Kimmel, D. B.; Leger, S.; Masse, F.; McGrath, M. E.; McKay, D.
J.; Percival, M. D.; Riendeau, D.; Rodan, S. B.; Therien, M.; Truong,
V.; Wesolowski, G.; Zamboni, R.; Black, W. C. Identification of a
potent and selective non-basic cathepsin K inhibitor. Bioorg. Med.
Chem. Lett. 2006, 16, 1985−1989.
(33) Johnson, T. W.; Dress, K. R.; Edwards, M. Using the Golden
Triangle to optimize clearance and oral absorption. Bioorg. Med. Chem.
Lett. 2009, 19, 5560−5564.
(34) Leach, A. G.; Jones, H. D.; Cosgrove, D. A.; Kenny, P. W.;
Ruston, L.; MacFaul, P.; Wood, J. M.; Colclough, N.; Law, B. Matched
Molecular Pairs as a Guide in the Optimization of Pharmaceutical
Properties; a Study of Aqueous Solubility, Plasma Protein Binding and
Oral Exposure. J. Med. Chem. 2006, 49, 6672−6682.
(35) Dossetter, A. G. A statistical analysis of in vitro human
microsomal metabolic stability of small phenyl group substituents,
leading to improved design sets for parallel SAR exploration of a
chemical series. Bioorg. Med. Chem. 2010, 18, 4405−4414.
(36) Lewis, M. L.; Cucurull-Sanchez, L. Structural pairwise
comparisons of HLM stability of phenyl derivatives: Introduction of
the Pfizer metabolism index (PMI) and metabolism-lipophilicity
efficiency (MLE). J. Comput.-Aided Mol. Des. 2009, 23, 97−103.
(37) Henderson, J. D.; Olson, R. D.; Ravis, W. R. Pharmacokinetic
calculator program for generation of initial parameter estimates from a
three-compartment infusion model. J. Pharmacol. Methods 1985, 14,
13−24.
(38) McGinnity, D. F.; Collington, J; Austin, R. P.; Riley, R. J.
Evaluation of Human Pharmacokinetics, Therapeutic Dose and
Exposure Predictions Using Marketed Oral Drugs. Curr. Drug Metab.
2007, 8, 463−479.
(39) Sadick, M.; Attenberger, U.; Kraenzlin, B.; Kayed, H.;
Schoenberg, S. O.; Gretz, N.; Schock-Kusch, D. Two non-invasive
GFR-estimation methods in rat models of polycystic kidney disease:
3.0 T dynamic contrast-enhanced MRI and optical imaging. Nephrol.,
Dial., Transplant. 2011, 26, 3101−3108.
6373
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374

Journal of Medicinal Chemistry
Article
(40) Rowland, M.; Tozer, T. N. In Clinical Pharmacokinetics; Concepts
and Application; 3rd ed., Chapter 11, p 171.
(41) Kuhn, B.; Mohr, P.; Stahl, M. Intramolecular Hydrogen Bonding
in Medicinal Chemistry. J. Med. Chem. 2010, 53, 2601−2611.
(42) Ashwood, V. A.; Field, M. J.; Horwell, D. C.; Julien-Larose, C.;
Lewthwaite, R. A.; McCleary, S.; Pritchard, M. C.; Raphy, J.; Singh, L.
Utilization of an Intramolecular Hydrogen Bond To Increase the CNS
Penetration of an NK1 Receptor Antagonist. J. Med. Chem. 2001, 44,
2276−2285.
(43) Sasaki, S.; Cho, N.; Nara, Y.; Harada, M.; Endo, S.; Suzuki, N.;
Furuya, S.; Fujino, M. Discovery of a Thieno[2,3-d]pyrimidine-2,4-
dione Bearing a p-Methoxyureidophenyl Moiety at the 6-Position: A
Highly Potent and Orally Bioavailable Nonpeptide Antagonist for the
Human Luteinizing Hormone-Releasing Hormone Receptor. J. Med.
Chem. 2003, 46, 113−124.
(44) Rezai, T.; Bock, J. E.; Zhou, M. V.; Kalyanaraman, C.; Lokey, R.
S.; Jacobson, M. P. Conformational Flexibility, Internal Hydrogen
Bonding, and Passive Membrane Permeability: Successful in Silico
Prediction of the Relative Permeabilities of Cyclic Peptides. J. Am.
Chem. Soc. 2006, 128, 14073−14080.
(45) Garnero, P.; Ferreras, M.; Karsdal, M.; Nicamhlaoibh, R.; Risteli,
J.; Borel, O.; Qvist, P.; Delmas, P. D.; Foged, N. T.; Delaisse, J. M. The
type I collagen fragments ICTP and CTX reveal distinct enzymatic
pathways of bone collagen degradation. J. Bone Miner. Res. 2003, 18,
859−867.
(46) Herrmann, M.; Seibel, M. The amino- and carboxyterminal
cross-linked telopeptides of collagen type I, NTX-I and CTX-I: A
comparative review. Clin. Chim. Acta 2008, 393, 57−75.
(47) Okuno, S.; Inaba, M.; Kitatani, K.; Ishimura, E.; Yamakawa, T.;
Nishizawa, Y. Serum levels of C-terminal telopeptide of type I
collagen: A useful new marker of cortical bone loss in hemodialysis
patients. Osteoporosis Int. 2005, 16, 501−509.
6374
dx.doi.org/10.1021/jm3007257 | J. Med. Chem. 2012, 55, 6363−6374