5-Aminopyrimidin-2-ylnitriles as Cathepsin K inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.01.110

A series of pyrimidine nitrile inhibitors of Cathepsin K with reduced glutathione reactivity has been identified and Molecular Core Matching (MoCoM) has been used to quantify the effect of an amino substituent at C5.

Full text of DOI:10.1016/j.bmcl.2009.01.110

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1658–1661
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
5-Aminopyrimidin-2-ylnitriles as Cathepsin K inhibitors
a,
b
b
a
Andrew D. Morley a, , Peter W. Kenny , Brenda Burton , Robert A. Heald , Philip A. MacFaul ,
Julia Mullett b, Ken Page a, Soraya S. Porres b, Lyn Rosenbrier Ribeiro a, Phil Smith b, Stuart Ward b,
Tina J. Wilkinson b
*
*
a AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
b Argenta Discovery, 8-9 Spire Green Centre, Harlow, Essex CM19 5TR, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of pyrimidine nitrile inhibitors of Cathepsin K with reduced glutathione reactivity has been iden-
Received 7 January 2009
Revised 29 January 2009
Accepted 30 January 2009
Available online 6 February 2009
tified and Molecular Core Matching (MoCoM) has been used to quantify the effect of an amino substituent
at C5.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Cathepsin K
Cheminformatics
Covalent inhibitor
Covalent adduct
Cysteine protease
Electrophilicity
Glutathione
Leatherface
Matched molecular pair analysis
MMPA
MoCoM
Molecular core matching
Molecule editor
SMARTS
The lysosomal cysteine protease Cathepsin K1,2 is highly ex-
pressed in osteoclasts and has been implicated as a key driver of
collagen breakdown. Cathepsin K is seen as an attractive target
for intervention in the treatment of both osteoporosis and
osteoarthritis.2,3
High throughput screening against Cathepsin S led to the iden-
tification of the triazine scaffold (1; X = N) initially as a singleton
hit.4 Subsequent SAR evaluation demonstrated that the template
could be replaced with pyrimidine (1; X = CH).4 Further evaluation
highlighted the template as a start point for broader inhibition of
the Cathepsin enzyme family. Concurrently, the template has been
the focus of research by other groups.5–8 The carbon atom of the ni-
trile binds covalently, although reversibly, to the cysteine residue
in the active site. Whilst this interaction makes an important con-
tribution to binding, it cannot be the sole basis for inhibition be-
cause excessively reactive electrophiles present significant safety
issues.9 Selectivity with respect to other Cathepsins is also impor-
tant and cannot be achieved unless molecular recognition ele-
ments other than the nitrile are exploited.10
R2
X
N
R1
N
CN
1
Relationships between nitrile reactivity and structure were ex-
plored using Molecular Core Matching (MoCoM). This cheminfor-
matic method complements approaches based on energies
calculated using density functional theory.11 Reactivity to glutathi-
one, quantified as half life (t1/2) was used as an indicator of inher-
ent electrophilicity.12,13 The rationale for applying MoCoM to these
* Corresponding authors. Tel.: +44 1509 644772 (A.D. Morley); tel.: +44 1625
514396 (P.W. Kenny).
E-mail
addresses:
andy.morley@astrazeneca.com
(A.D.
Morley),
pwk.pub.2008@gmail.com (P.W. Kenny).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.01.110

A. D. Morley et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1658–1661
1659
systems is that variable parts of molecules are insulated electron-
ically from the reactive centers and not expected to contribute to
differences in reactivity. MoCoM of the heteroaromatic nitriles
was performed using the OELeatherface14,15 molecular editor to
trim16 substitution of the nitrogen atoms at C4 and C6 to methyl
or phenyl. The resulting molecular cores were matched as canoni-
cal17,18 (unique) SMILES19 strings.
comparisons of mean values of properties. For example, MoCoM
can be used to define series in structurally diverse data sets for
exploration of correlations between solubility and lipophilicity.
The effect of 5-amino substituent on nitrile reactivity was also
explored (Fig. 1) using Matched Molecular Pair Analysis15,22,23
(MMPA). On average, substitution at C5 with amino results in a
fourfold (0.62 log units) increase in t1/2. This technique addresses
the question more directly than MoCoM because structures are
identical except for the specific difference that is probed. However,
the small number of exactly matched pairs does mean that less
data is used in the analysis.
The results presented in Table 1 show the beneficial effect of a
5-amino substituent on reactivity. Comparing mean values of
log(t1/2) for the Core1/Core2 and Core3/Core4 pairs suggests that
substitution at C5 with an amino group leads to an increase in
t1/2 of 0.6–0.7 log units. The range of 0.8 log units in t1/2 measured
for the compounds specified by Core5 reflects the sensitivity of
reactivity to the substitution of the phenyl ring. The range and
standard deviation in a property provide useful checks on the
validity of the assumption that different structures with the same
core are equivalent.
The C5 amino substituent was adopted in lead generation be-
cause it appeared to provide the best opportunity to achieve an
adequate window between Cathepsin K inhibition and reaction
with other cysteine thiols.
O
O
MoCoM is related to the RECAP20 and Scaffold Tree21 methods.
The power of the approach lies in being able to partition molecules
arbitrarily into core and peripheral regions in a user specified man-
ner. The method represents a general approach to associating
structurally related molecules. Data analysis is not restricted to
N
N
MeHN
HN
H2N
HN
N
N
N
7
CN
N
CN
Table 1
Analysis of nitrile reactivity using molecular core matching (MoCoM) for neutral
pyrimidine and triazine nitriles (keyed to 1). The glutathione reactivity assay is
described in Ref. 12
OMe
OMe
b
Core
X
R1
R2
Na
Mean log(t1/2
)
SEc
Core1
Core2
Core3
Core4
Core5d
Core6
Core7
Core8
Core9
Core10
Core11
Core12
CNH2
CH
CNH2
CH
CNH2
CH
CNH2
CNHMe
COMe
CBr
NHMe
NHMe
NMe2
NMe2
NMe2
NMe2
NMe2
NMe2
NMe2
NHMe
NHMe
NMe2
NMe2
NMe2
NMe2
NMe2
NHPh
NHPh
OPh
NHPh
NHPh
NMe2
NMe2
NHPh
27
3
9
4
9
1
2
1
3
1
1
1
3.38
2.67
3.26
2.56
3.00
2.12
2.98
2.91
2.31
2.19
< 2.08
2.07
0.03
0.16
0.03
0.05
0.10
6
The binding of 6 to Cathepsin K has been modelled (Fig. 2).24
The aromatic ring sits in the lipophilic S2 pocket and the morpho-
line substituent is substantially exposed to solvent. Synthesis of
analogues of 6 is summarized in Scheme 1.
Nitration of 2 and subsequent conversion to the corresponding
dichloro analogue 3 is straightforward, and can be carried out on
multi-gram scale. Introduction of the first amine substituent can
be achieved rapidly and selectively at room temperature, with 4
being isolated in high yield. The change in electronic properties
then reduces the overall reactivity of the molecule, and subse-
quently more forcing conditions are necessary to incorporate the
second amino substituent, to generate 5. Cyanide addition has to
0.14
0.05
CNO2
N
a
Number of compounds with core.
Mean log(half life in minutes) for compounds with core.
Standard error in mean.
b
c
d
The log(t1/2) values for this core range from 2.60 to 3.41.
Figure 1. Matched molecular pair analysis of effect of 5-amino substituent on
reactivity of pyrimidin-2-yl nitriles. Solid red lines indicate mean values for the two
quantities being plotted and broken red lines mark the 95% confidence interval for
the mean difference (0.62) in reactivity.
Figure 2. Potential binding mode of adduct of
molecular surface of protein. The covalent bond is formed between the nitrile
carbon atom and the sulfur atom of the catalytic cysteine (hidden by surface).
6 with Cathepsin K showing

1660
A. D. Morley et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1658–1661
Table 2
be undertaken prior to reduction of the nitro group, in order to uti-
lize its beneficial electron withdrawing effect.
Effect of P2 substituent on potency and stability
Alkylation of the 5-amino group was considered as an approach
to exploiting the S3 pocket. The N-methyl analogue 7 shows equiv-
alent Cathepsin K activity to 6. However, the glutathione reactivity
of 7 (t1/2 = 800 min) is greater that that of 6 (t1/2 = 2500 min). It ap-
pears that the additional bulk of the N-methyl substituent compro-
mises the ability of the substituents at C4, C5, and C6 to release
electrons to the heteroaromatic ring. Based on this result, the focus
of the work shifted to optimization of the P1 and P2 substituents.
The S2 pocket of Cathepsin K is a well-defined hydrophobic cav-
ity and the relevant SAR is summarized in Table 2. The glutathione
stability of compounds with alkylamino R1 substituents ap-
proached that of Nilvadipine which was used as a reactivity bench-
mark.12 Compound 8, with an anilino R1 substituent is both less
potent and less stable than its cyclohexylamino analogue 11. To
some extent increasing the size of the alkylamino R1 substituent
leads to an increase in potency as reflected by the 40-fold differ-
ence in IC50 values measured for 9 and 10. Although compounds
such as 15 show excellent potency, they are too lipophilic to be
of interest, as the poor solubility and high metabolic turnover of
these motifs present significant challenges to the optimization pro-
cess. The S2 pocket appears intolerant of ligand polarity as the tet-
rahydrofuran substituent 16 shows ꢀ25-fold loss of potency
compared to 10. Further expansion of the SAR demonstrated that
cyclic amines were also tolerated in this region (17,18). Small,
branched lipophilic substituents such as isobutylamino appeared
to give a more balanced profile with greater scope to achieve the
desired overall parameters for progression.
O
N
N
CN
N
H2N
R1
a
b
Compd
R1
IC50 (nM)
GSH t1/2 (min)
8
9
Anilino
110
560
13
23
18
17
300
2
340
60
22
1200
3200
2000
2000
2700
3600
3300
2900
1900
2000
1700
Isopropylamino
Cyclopentylamino
Cyclohexylamino
Isobutylamino
Neopentylamino
sec-Butylamino
Norbornan-2-ylamino
Tetrahydrofuran-3-ylamino
Pyrrolidin-1-yl
10
11
12
13
14
15
16
17
18
1-Piperidyl
a
Values are means of at least three experiments.
The glutathione reactivity assay is described in Ref. 12.
b
Table 3
Effect of P1 substituent on potency and stability
R2
N
The opportunity to modify the physicochemical and DMPK
parameters of the series is most easily achieved through modifica-
tion of R2. As can be seen in Table 3, a wide variety of substitution
is tolerated. Glutathione reactivity and selectivity, with respect to
other Cathepsins, can be controlled by variation of this substituent.
The presence of a basic centre in R2 leads to an increase in Cathep-
sin K potency and a decrease in GSH t1/2. Conversely, introduction
of an acid results in an increase in stability but loss of potency.
These differences may reflect electrostatic interactions between
ionized centres in the substituent and the anionic carboxylate
groups of glutathione.
A more detailed evaluation of key compounds was undertaken
(Table 4). Basic functionality in R2 (19) leads to increases in
Cathepsin K, L, and B potency, solubility, and free fraction but a
decrease in GSH t1/2. Better selectivity with respect to other
Cathepsins was observed for 19 than for 12 or 13 despite the lower
CN
N
H2N
HN
a
b
Compd
R2
IC50 (nM)
GSH t1/2 (min)
12
19
20
21
22
23
24
Morpholino
Piperazin-1-yl
18
6
24
9
14
32
62
2700
440
2-Methoxyethylamino
4-Acetylpiperazin-1-yl
6000
1800
2600
3000
>9999
(1,1-Dioxothiolan-3-yl)amino
3-Hydroxyazetidin-1-yl
1H-Tetrazol-5-ylmethylamino
a
Values are means of at least three experiments
The glutathione reactivity assay is described in Ref. 12.
b
OH
Cl
OH
Cl
(c)
O2N
HO
O2N
R1
(b)
O2N
Cl
(a)
N
N
N
N
HO
S
N
2
N
S
S
N
N
S
N
63%
62%
R2
3
4
R3
R4
R4
R3
N
R4
N
N
R3
(e)
(d)
N
N
(f)
O2N
O2N
R1
N
N
H2N
R1
R1
N
S
N
5
N
CN
N
CN
N
R2
R2
R2
Scheme 1. Reagents and conditions: (a) Fuming nitric acid; (b) POCl3 N,N-diethylaniline; (c) R1R2NH, THF room temp.; (d) R3R4NH, THF, reflux; (e) (i) MCPBA, CH2Cl2; (ii)
NaCN, DMSO; (f) H2, 10% Pd/C, EtOAc.

A. D. Morley et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1658–1661
1661
Table 4
References and notes
Profiles for Cathepsin K inhibitors.
1. Grabowska, U. B.; Chambers, T. J.; Shiro, M. Curr. Opin. Drug. Discov. Devel. 2005,
8, 619.
2. Salminen-Mankonen, H. J.; Morko, J.; Vuorio, E. Curr. Drug Targets 2007, 8, 315.
3. Sun, S. Expert Opin. Ther. Targets 2008, 12, 239.
R2
N
CN
4. Bailey, A.; Pairaudeau, G.; Patel, A.; Thom, S. WO 000819, 2004.
5. Altmann, E.; Cowan-Jacob, S. W.; Missbach, M. J. Med. Chem. 2004, 47, 5833.
6. Teno, N.; Miyake, T.; Ehara, T.; Irie, O.; Sakaki, J.; Ohmori, O.; Gunji, H.;
Matsuura, N.; Masuya, K.; Hitomi, Y.; Nonomura, K.; Horiuchi, M.; Gohda, K.;
Iwasaki, A.; Umemura, I.; Tada, S.; Kometani, M.; Iwasaki, G.; Cowan-Jacob, S.
W.; Missbach, M.; Lattmann, R.; Betschart, C. Bioorg. Med. Chem. Lett. 2007, 17,
6096.
7. Teno, T.; Irie, O.; Miyake, T.; Gohda, K.; Horiuchi, M.; Tada, S.; Nonomura, K.;
Kometani, M.; Iwasaki, G.; Betschart, C. Bioorg. Med. Chem. Lett. 2008, 18, 2599.
8. Teno, N.; Masuya, K.; Ehara, T.; Kosaka, T.; Miyake, T.; Irie, O.; Hitomi, Y.;
Matsuura, N.; Umemura, I.; Iwasaki, G.; Fukaya, H.; Toriyama, K.; Uchiyama, N.;
Nonomura, K.; Sugiyama, I.; Kometani, M. J. Med. Chem. 2008, 51, 5459.
9. Evans, D. C.; Watt, A. P.; Nicoll-Griffith, D. A.; Baillie, T. A. Chem. Res. Toxicol.
2004, 17, 3.
N
H2N
R1
Parameter
12
19
13
R1
Isobutylamino
Morpholino
18
Isobutylamino
Piperazin-1-yl
6
Neopentylamino
Morpholino
17
R2
CatK IC50 nM
CatS IC50 nM
CatL IC50 nM
CatB IC50 nM
GSH t1/2 min
logD
560
890
1400
6100
440
0.7
1000
2
47
910
2000
>10,000
2700
2.8
>10,000
>10,000
3600
3
a
10. Felbor, U.; Kessler, B.; Mothes, W.; Ploegh, H. L.; Goebel, H. H.; Bronson, R. T.;
Olsen, B. R. Proc. Nat. Acad. Sci. 2002, 99, 7883.
11. Oballa, R. M.; Truchon, J.-F.; Bayly, C. I.; Chauret, N.; Day, S.; Crane, S.;
Berthelette, C. Bioorg. Med. Chem. Lett. 2007, 17, 998.
Solubility
l
M
67
27
16
25
PAMPA 106 cm/s
Human % free
NTe
5.3
12. Morley, A. D.; Crawford, J.; MacFaul, P. A. Bioorg. Med. Chem. Lett., in press. doi:
10.1016/j.bmcl.2008.12.105.
hERG IC50
Ames
l
M
>100
52
70
Àve
Àve
NTe
13. Guengerich, F. P.; MacDonald, J. S. Chem. Res. Toxicol. 2007, 20, 344.
14. OELeatherface is a version of the Leatherface molecular editor (Ref. 15) written
by Hongming Chen (AstraZeneca Mölndal) that uses the SMARTS parser
supplied by OpenEye Scientific Software (Santa Fe, NM) to perform
substructural matches.
15. Kenny, P. W.; Sadowski, J. Structure Modification in Chemical Databases. In
Chemoinformatics in Drug Discovery; Oprea, T. I., Ed.; Wiley-VCH Verlag GmbH:
Weinheim, 2005; p 271.
16. In the trim modes of Leatherface and OELeatherface, the complete molecular
editing sequence is repeated until the molecule remains unchanged. This
allows removal of structurally complex substituents to be specified using a
small number of simple substructural definitions (e.g., atom with only one
non-hydrogen connection).
CYP IC50
l
Mb
>10 (5:5)
>10 (5:5)
>10 (5:5)
Clint Rat Hepsc
<2
11
13
5.1
3.1
70%
4.7
<2
14
5
3
36%
19
18
10
3
2.7
42%
Clint Human Micsd
Rat Cl ml/min/kg
Rat t1/2
h
Rat Vdss l/kg
Rat bioavailability
a
The glutathione reactivity assay is described in Ref. 12.
b
c
Inhibition of cytochrome P450 isoforms: 1A2, 2C9, 2C19, 2D6, and 3A4.
Intrinsic clearance, rat hepatocytes (
l/min/106 cells).
l/min/mg).
l
d
e
Intrinsic clearance, human microsomes (
l
17. Weininger, D.; Weininger, A.; Weininger, J. L. J. Chem. Inf. Comput. Sci. 1989, 29,
Not tested.
97.
18. Algorithms for writing canonical SMILES generate the same string for
a
specific structure regardless of the ordering of atoms in that structure.
This allows identity of molecules to be established by character string
matching.
GSH t1/2 measured for 19. This observation provides evidence that
the R1 and R2 substituents make significant contributions to free
energy of binding to Cathepsin K.
19. Weininger, D. J. Chem. Inf. Comput. Sci. 1988, 28, 31.
20. Lewell, X. Q.; Judd, D. B.; Watson, S. P.; Hann, M. M. J. Chem. Inf. Comput. Sci.
1998, 38, 511.
21. Schuffenhauer, A.; Ertl, P.; Roggo, S.; Wetzel, S.; Koch, M. A.; Waldmann, H. J.
Chem. Inf. Model 2007, 47, 47.
22. Birch, A. M.; Kenny, P. W.; Simpson, I.; Whittamore, P. R. O. Bioorg. Med. Chem.
Lett. 2009, 19, 850.
23. Leach, A. G.; Jones, H. D.; Cosgrove, D. A.; Kenny, P. W.; Ruston, L.; MacFaul, P.;
Wood, J. M.; Colclough, N.; Law, B. J. Med. Chem. 2006, 49, 6672.
Good physicochemical and in vivo pharmacokinetic profiles
were observed for these compounds. Acceptable half-lives were
achieved with modest volumes and the compounds were readily
bioavailable. Despite the presence of the aniline-like amino group,
neither 12 nor 19 showed activity in the Ames test either in the
presence or absence of metabolizing enzymes. The good pharma-
cokinetic properties of compounds 12, 13, and 19 combined with
the eightfold dynamic range in GSH t1/2 suggest value as tools with
which to investigate the link between glutathione reactivity and
covalent adduct formation.
24. Model of binding mode the covalent adduct of
6 with Cathepsin K was
generated from RSB (home.rcsb.org) Protein Data Bank (www.rcsb.org/pdb/
home/home.do) entry with reference code 1U9X (Ref. 5) using the Maestro
molecular modelling program. This model was energy-minimized using
MacroModel (OPLS 2005 force field; water solvent model). Both Maestro and
MacroModel were licensed from Schrödinger, LLC (www.schrodinger.com).