Novel and potent cyclic cyanamide-based cathepsin K inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.02.033

Starting from a PDE IV inhibitor hit derived from high throughput screening of the compound collection, a key pyrrolidine cyanamide pharmacophore was identified. Modifications of the pyrrolidine ring produced enhancements in cathepsin K inhibition. An X-r

Full text of DOI:10.1016/j.bmcl.2005.02.033

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1815–1819
Novel and potent cyclic cyanamide-based cathepsin K inhibitors
David N. Deaton,a,* Anne M. Hassell,b Robert B. McFadyen,a
Aaron B. Miller,b Larry R. Miller,c Lisa M. Shewchuk,b Francis X. Tavares,a
Derril H. Willard, Jr.,d,z and Lois L. Wrighte
aDepartment of Medicinal Chemistry, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
bDiscovery Research Computational, Analytical, and Structural Sciences, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
cDepartment of Molecular Pharmacology, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
dDepartment of Gene Expression and Protein Purification, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
eDiscovery Research Biology, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
Received 31 January 2005; revised 9 February 2005; accepted 11 February 2005
Abstract—Starting from a PDE IV inhibitor hit derived from high throughput screening of the compound collection, a key pyrrol-
idine cyanamide pharmacophore was identified. Modifications of the pyrrolidine ring produced enhancements in cathepsin K inhi-
bition. An X-ray co-crystal structure of a cyanamide with cathepsin K confirmed the mode of inhibition.
Ó 2005 Elsevier Ltd. All rights reserved.
Skeletal fragility, resulting from the loss of bone mineral
density and integrity, is the leading cause of non-trau-
matic fracture.1 Typically, this phenotype results from
an imbalance in the normally tightly coupled processes
of bone formation and bone resorption that maintain
skeletal homeostasis. Bone is resorbed by osteoclasts,
which secrete protons and proteases that degrade the
mineral and protein components of bone. The cysteine
protease cathepsin K is highly expressed in osteoclasts
and rapidly hydrolyzes type I collagen, the major com-
ponent of bone matrix, as a complex with glycosamino-
glycans.2 A rare human deficiency in cathepsin K is the
cause of pycnodysostosis, an autosomal recessive trait
characterized by short stature, abnormal bone, and
tooth development, increased bone mineral density,
and increased bone fragility.3 In light of the key role
played by cathepsin K in bone resorption, pharmaceuti-
cal companies have pursued the development of cathep-
sin K inhibitors for the treatment of osteoporosis,4 and
small molecule inhibitors of cathepsin K have proven
efficacious in attenuating bone resorption in animal
models of osteoporosis.5
As part of a larger program to develop novel cathepsin
K inhibitors, researchers from these laboratories re-
cently reported the discovery of aldehyde-based cathep-
sin K inhibitors.6,7 Starting from a focused screening hit,
Boc–Nle–H 1 (IC50 = 51 nM), potent inhibitors like 2
(IC50 = 2.7 nM) and 3 (IC50 = 0.13 nM) were developed
by optimizing P2–P3 substituents. In addition to this fo-
cused screen approach to hit discovery, a high through-
put screen of the GlaxoWellcome compound collection
was also employed, resulting in the identification of
the PDE IV inhibitor 48 as a reasonably active cathepsin
K inhibitor (IC50 = 430 nM).9 This cyanamide-contain-
ing compound proved to be a reversible and substrate-
competitive inhibitor. Surmising that the cyanamide
was acting as an active site electrophile for the 25Cys
thiol, the commercially available truncated N-pyrrol-
idinecarbonitrile 5 (IC50 = 2100 nM) was also tested
and found to be a micromolar inhibitor of cathepsin
K. Pyrrolidine 5 was chosen as a key pharmacophore
for further optimization with the goal of appending sub-
stituents to the a- or b-position to interact with the S2
subsite, the deepest, most pronounced hydrophobic
binding pocket of the cathepsin K active site. Subse-
quent to this work being completed, Merck/Celera
researchers also reported the discovery of cyanamide-
based cathepsin K inhibitors from a HTS hit.10,11 The
results of efforts to design and prepare potent cathepsin
K inhibitors derived from cyanamide 5 are detailed in
this report.
Keywords: Cathepsin K; Cyanamide; Cysteine protease inhibitor;
Warhead.
*
Corresponding author. Tel.: +1 919 483 6270; fax: +1 919 315 0430;
e-mail: david.n.deaton@gsk.com
z Deceased.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.02.033

1816
D. N. Deaton et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1815–1819
H
N
H
H
N
O
O
O
N
O
O
O
H2N
O
O
O
a, b
a, c
NH
NH
N
R1
O
O
nBu
O
nBu
O
O
O
1
2
10a R1 = Bn
10b R1 = tBu
6k
6l
O
O
H
N
O
H
N
d, c
O
N
N
O
nBu
CN
CN
H
3
4
5
N
OH
O
N
NH
H
O
6n
9b
The cyanamides 7a–n were synthesized from the corre-
sponding amines 6a–n as shown in Scheme 1. Treatment
of the amine with cyanogen bromide in the presence of a
base produced the desired cyanamides in reasonable
yields.
Scheme 3. Reagents and conditions: (a) 9a or 9b, 1.93 M COCl2 in
PhMe, pyridine, CH2Cl2, ꢀ20 °C to rt; 10a or 10b, i-Pr2NEt, THF, 59–
88%; (b) H2/Pd–C, MeOH, 98%; (c) HCl, EtOAc, 99%; (d) Cbz–Leu,
EDC, HOBt, NMM, 71%.
The amines 6a–o utilized for this coupling were either
commercially available (6a–d), known in the literature
(6e,f, and h),12–14 or synthesized as depicted in Schemes
2–4 (6g, i–n).
O
NH2
O
N
HN
a, b
N
CF3
O
H
As shown in Scheme 2, the known acid 8a15 and com-
mercially available acid 8b were converted to esters of
known alcohol 9a7 via the method of Fischer. Then
the carbamates were cleaved to provide the amines 6i
and j. The amine 6g was also obtained from 8b. Reac-
tion of its acid chloride, obtained from treatment of 8b
with oxalyl chloride, with tert-butylmethylamine in the
presence of triethylamine afforded the amide. Catalytic
hydrogenation of the benzyl carbamate as before yielded
the amine 6g.
11
6m
Scheme 4. Reagents and conditions: (a) 9b, 1.93 M COCl2 in PhMe,
pyridine, CH2Cl2,ꢀ20 °C to rt; 11, i-Pr2NEt, THF, 95%; (b) K2CO3,
MeOH, H2O, 89%.
9a was converted to its corresponding chloroformate
with phosgene, and then coupled to the amine 10a to
give a bis-carbamate. Subsequent hydrogenolysis of
the benzylcarbamate afforded amine 6k. The amine 6l
was synthesized from the commercially available amine
10b. As before, conversion of known alcohol 9b7 to its
chloroformate and coupling to amine 10b provided the
bis-carbamate. Then, the cyclic amine was deprotected
under acidic conditions to provide amine 6l. Further-
more, amine 10b was coupled to commercially available
cbz-protected leucine via a carbodiimide method to give
the amide. Acid catalyzed deprotection of the cyclic
amine, as before, provided the amine 6n.
The amine 6k was synthesized from the commercially
available amine 10a as illustrated in Scheme 3. Alcohol
R2
N
R2
a
NH
R1
CN
R1
6a-6n
7a-7n
The reversed pyrrolidine amine 6m was synthesized as
shown in Scheme 4. Coupling of the chloroformate de-
rived from alcohol 9b with the commercially available
amine 11 afforded the carbamate. Then, base catalyzed
hydrolysis of the trifluoromethyl acetamide provided
amine 6m.
Scheme 1. Reagents and conditions: (a) BrCN, K2CO3, MeCN, 40–
78%.
()
N
()n
NH
a, b
c, b
n Ph
O
HO
O
Because the tert-butyl P2 moiety had provided reason-
able potency to aldehyde inhibitor 1, it was appended
to the a- and b-positions of cyanamide 5 to provide
the cyanamides 7a–d. As shown in Table 1, the a-series
(S)-proline derived cyanamide 7a (IC50 = 60 nM) was
35-fold more potent than the starting cyanamide 5
(IC50 = 2100 nM). Its (R)-proline derived enantiomer
7b (IC50 = 300 nM) was also more potent than analog
5, but 5-fold less potent than the (S)-enantiomer 7a, sug-
gesting that the inhibitor interacts with the enzyme in a
stereospecific manner. Gratifyingly, the b-series (S)-
cyanamide 7c (IC50 = 81 nM) was equipotent with the
O
O
O
8a n = 1
8b n = 2
6i n = 1
6j n = 2
OH
N
N
H
O
9a
6g
Scheme 2. Reagents and conditions: (a) 8a or b, 9a, TsOH, PhH, "#,
61–93%; (b) H2/Pd–C, MeOH, 77–97%; (c) 8b, (COCl)2, PhMe,
t-BuNHMe, NEt3, CH2Cl2, 83%.

D. N. Deaton et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1815–1819
1817
Table 1. Inhibition of human cathepsin K
that of 7c. Apparently, the effect of the stereocenter in
the b-position is much smaller than the a-position.
#
Inhibitor
IC50 nMa
To further explore the influence of the stereocenter adja-
cent to the pyrrolidine nitrogen, analog 7e was synthe-
sized, in which replacement of the carbon atom with
nitrogen places the presumed P2 substituent more in
plane with the pyrrolidine ring due to the sp2 character
of the C–N carbamate bond. This change resulted in a
cathepsin K inhibitor that was equipotent with (R)-pro-
line derived analog 7b, confirming the importance of the
stereocenter. Since amides are chemically more robust
than esters, the amides 7f and g were prepared. Unfortu-
nately, the tert-butyl amide 7f (IC50 = 8100 nM) was
>100-fold less active that the ester 7a, and the N-methyl
analog 7g exhibited no inhibition at concentrations up
to 13,000 nM.
O
O
7a
60
N
CN
O
O
7b
7c
7d
7e
7f
300
81
N
CN
H
N
O
O
CN
CN
N
N
H
N
O
O
200
O
O
N
Desiring to explore if the P2–P3 SAR from the aldehyde-
based inhibitors could be further transferred to the
cyanamide series, the P2–P3 substituent from aldehyde
2 was incorporated into cyanamide 5 to provide cyan-
amide 7i (IC50 = 1.8 nM). The addition of a P3 substitu-
ent resulted in a 30-fold boost in inhibitory activity
versus cathepsin K. A comparable 18-fold increase was
previously realized in the aldehyde inhibitors.7
380
N
CN
H
N
8100
>13,000
>13,000
1.8
N
CN
O
N
O
7g
7h
7i
N
CN
Changing the size of the cyanamide ring should alter the
position of the P2–P3 moiety relative to the cyanamide
warhead, possibly allowing more favorable inter-
action with the S2 or S3 subsites. To explore this hypoth-
esis, piperidine analog 7h and azetidine analog 7j were
synthesized. Whereas the piperidine 7h was completely
inactive at concentrations up to 13,000 nM, a greater
than 200-fold reduction in potency from 7a,16 a compar-
ison of the azetidine 7j (IC50 = 0.048 nM) with 7i
(IC50 = 1.8 nM) reveals that the four-membered ring
is optimal for inhibitory activity versus cathepsin K.
Analog 7j is >30-fold more active than 7i. Indepen-
dently, Merck/Celera researchers have also disclosed
that the N-azetidinecarbonitrile warhead is a more ac-
tive cathepsin K inhibitor than the five-membered N-
pyrrolidinecarbonitrile electrophile.10 They attribute
the increased potency of the four-membered analog to
increased reactivity of the azetidinecarbonitrile war-
head, supporting this assertion with data showing a loss
of cathepsin K inhibitory potency of the azetidinecarbo-
nitrile relative to the pyrrolidinecarbonitrile in the pres-
ence of added glutathione at pH = 7.0. Since the basicity
of the parent cyclic amines are essentially identical
(azetidine pKBH + = 11.29, pyrrolidine pKBH + =
11.27),17 the increased reactivity of the azetidine-derived
inhibitors must arise from reduced steric impediments to
nucleophilic attack as opposed to any electronic effect.
O
O
N
CN
O
N
CN
O
O
O
O
N
CN
7j
0.048
O
O
O
H
N
7k
7l
12
CN
CN
N
H
N
8.9
N
CN
NH
O
O
N
7m
7n
200
19
O
H
N
O
N
H
CN
N
O
a Inhibition of recombinant human cathepsin K activity in a fluores-
cence assay using 10 lM Cbz-Phe-Arg-AMC as substrate in 100 mM
NaOAc, 10 mM DTT, 120 mM NaCl, pH = 5.5. The IC50 values are
the mean of two or three inhibition assays, individual data points in
each experiment were within a 3-fold range of each other.
Incorporation of P2–P3 aldehyde potency enhancing
groups7 at the b-position of the parent cyanamide 5 also
produced significant enhancements in inhibitory activ-
ity. As shown in Table 1, the P2–P3 analog 7k
(IC50 = 12 nM) was 6-fold more active than the corre-
sponding tert-butyl analog 7c (IC50 = 81 nM), although,
less potent than the corresponding a-series analog 7i
(IC50 = 1.8 nM). An attempt to further improve activity
a-series analog 7a. Interestingly, the b-series (R)-enan-
tiomer 7d (IC50 = 200 nM) exhibited an inhibitory po-
tency versus cathepsin K not statistically different from

1818
D. N. Deaton et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1815–1819
in the b-series, by incorporating the more potent P2–P3
substituent7 from aldehyde 3 was not successful, the
cyanamide 7l (IC50 = 8.9 nM) not being significantly
more active than 7k. Reversal of the substituents on
the 3-aminopyrrolidine as in analog 7m (IC50 = 200 nM)
resulted in a ꢁ10-fold loss in activity.
As shown in Table 2, these analogs exhibit moderate
selectivity versus the closely related endopeptidase
cathepsin L with selectivity ranging from 6- to 70-fold.
In contrast, these cyanamides are more selective versus
the exopeptidases cathepsin B (B/K = 56–500) and
cathepsin H (H/K = 320–830). These cyanamides were
not tested versus cathepsin S, another closely related
endopeptidase.
In order to better understand the mechanism of action
of these reversible inhibitors, an X-ray co-crystal struc-
ture of cathepsin K with the leucine-derived inhibitor
7n (IC50 = 19 nM) was solved. This is the first published
X-ray co-crystal structure of an inhibitor containing a
cyanamide warhead, and the active site is shown in Fig-
ure 1. The cyanamide moiety of the inhibitor and the ac-
tive site 25Cys of the enzyme form a covalent isothiourea
intermediate, consistent with the reversible nature of
these inhibitors. This structure confirms the 13C NMR
experiments of the Merck researchers with their cyan-
amide-based inhibitors and papain.10 This mechanism
of inhibition is similar to the reaction of cysteine prote-
ases with nitriles to form thioimidate esters. The nitro-
gen of the carbon–nitrogen double bond points into
the oxy-anion hole, and is stabilized by hydrogen bonds
to the side chain carbonyl of 19Gln and the backbone
NH of 25Cys. One additional hydrogen bond between
the peptide backbone recognition site of the enzyme
and the amide carbonyl further stabilizes the inhibitor.
Thus, the amide carbonyl accepts a hydrogen bond from
the backbone NH of 66Gly.
Figure 1. Active site of the X-ray co-crystal structure of compound 7n
complexed with cathepsin K. The cathepsin K carbons are colored
magenta with inhibitor 7n carbons colored green. The semi-transpar-
ent white surface represents the molecular surface, while hydrogen
bonds are depicted as yellow dashed lines. The coordinates have been
deposited in the Brookhaven Protein Data Bank, accession number
1YK7. This figure was generated using PYMOL version 0.97 (Delano
Scientific, www.pymol.org).
ized with the S1 wall formed from 23Gly, 24Ser, 64Gly,
and 65Gly. Further substitution of the pyrrolidine
ring might provide hydrophobic interactions with the
S1 subsite.
In summary, starting from a high throughput screening
hit 4, a key pharmacophore 5 was identified. Addition of
a P2 group to the pyrrolidine led to a >10-fold increase
in potency as in analog 7a. Subsequent elaboration with
P2–P3 moieties derived from aldehyde-based cathepsin
K inhibitors produced further enhancements of inhibi-
tory activity as exemplified by analogs 7i and k. Further
manipulation of the cyanamide ring size resulted in the
discovery of the picomolar cathepsin K inhibitor 7j.
These inhibitors exhibit modest selectivity versus other
cathepsin endopeptidases, but their high potency argues
for further work to enhance selectivity and other drug
properties.
Besides these hydrogen bond stabilizing interactions
with the protein, the P2 isobutyl forms lipophilic interac-
tions with the S2 pocket composed of 67Tyr,
68Met, 134Ala, 163Ala, and 209Leu. Moreover, the P3 phe-
nyl interacts with the S3 subsite. No interaction is real-
Table 2. Cathepsin B, H, and L inhibition and selectivity
References and notes
#
Cat K
IC50 nM
Cat B
IC50 nMa
Cat H
IC50 nMb
Cat L
IC50 nMc
1. Einhorn, T. A. In Osteoporosis; Marcus, R., Feldman, D.,
Kelsey, J., Eds.; Academic: San Diego, CA, 1996; p 3.
2. Li, Z.; Hou, W.-S.; Escalante-Torres, C. R.; Gelb, B. D.;
Bromme, D. J. Biol. Chem. 2002, 277, 28669.
3. Gelb, B. D.; Shi, G.-P.; Chapman, H. A.; Desnick, R. J.
Science 1996, 273, 1236.
7i
1.8
0.048
12
310
2.4
680
580
40
130
0.71
81
7j
7k
8300
a Inhibition of recombinant human cathepsin B activity in a fluores-
cence assay using 10 lM Cbz-Phe-Arg-AMC as substrate in 100 mM
NaOAc, 10 mM DTT, 120 mM NaCl, pH = 5.5. The IC50 values are
the mean of two or three inhibition assays, individual data points in
each experiment were within a 2-fold range of each other.
b Inhibition of recombinant human cathepsin H activity in a fluores-
cence assay using 50 lM L-Arg-b-naphthalamide as substrate in
100 mM NaOAc, 10 mM DTT, 120 mM NaCl, pH = 5.5.
4. Deaton, D. N.; Kumar, S. Prog. Med. Chem. 2004, 42,
245.
5. Stroup, G. B.; Lark, M. W.; Veber, D. F.; Bhattacharyya,
A.; Blake, S.; Dare, L. C.; Erhard, K. F.; Hoffman, S. J.;
James, I. E.; Marquis, R. W.; Ru, Y.; Vasko-Moser, J. A.;
Smith, B. R.; Tomaszek, T.; Gowen, M. J. Bone Miner.
Res. 2001, 16, 1739.
c Inhibition of recombinant human cathepsin L activity in a fluores-
cence assay using 5 lM Cbz-Phe-Arg-AMC as substrate in 100 mM
NaOAc, 10 mM DTT, 120 mM NaCl, pH = 5.5.
6. Catalano, J. G.; Deaton, D. N.; Furfine, E. S.; Hassell, A.
M.; McFadyen, R. B.; Miller, A. B.; Miller, L. R.;

D. N. Deaton et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1815–1819
1819
Shewchuk, L. M.; Willard, D. H.; Wright, L. L. Bioorg.
Med. Chem. Lett. 2004, 14, 275.
10. Falgueyret, J.-P.; Oballa, R. M.; Okamoto, O.; Wesolow-
ski, G.; Aubin, Y.; Rydzewski, R. M.; Prasit, P.;
Riendeau, D.; Rodan, S. B.; Percival, M. D. J. Med.
Chem. 2001, 44, 94.
11. Rydzewski, R. M.; Bryant, C.; Oballa, R.; Wesolowski,
G.; Rodan, S. B.; Bass, K. E.; Wong, D. H. Bioorg. Med.
Chem. 2002, 10, 3277.
7. Boros, E. E.; Deaton, D. N.; Hassell, A. M.; McFadyen,
R. B.; Miller, A. B.; Miller, L. R.; Paulick, M. G.;
Shewchuk, L. M.; Thompson, J. B.; Willard, D. H.;
Wright, L. L. Bioorg. Med. Chem. Lett. 2004, 14,
3425.
8. Feldman, P. L.; Brackeen, M. F.; Cowan, D. J.; Marron,
B. E.; Schoenen, F. J.; Stafford, J. A.; Suh, E. M.;
Domanico, P. L.; Rose, D.; Leesnitzer, M. A.; Brawley, E.
S.; Strickland, A. B.; Verghese, M. W.; Connolly, K. M.;
Bateman-Fite, R.; Noel, L. S.; Sekut, L.; Stimpson, S. A.
J. Med. Chem. 1995, 38, 1505.
12. Dutta, A. S.; Morley, J. S. J. Chem. Soc., Perkin Trans. 1
1975, 1712.
13. OÕNeil, I. A.; Miller, N. D.; Barkley, J. V.; Low, C. M. R.;
Kalindjian, S. B. Synlett 1995, 619.
14. Egbertson, M.; Danishefsky, S. J. J. Org. Chem. 1989, 54,
11.
9. Barrett, D. G.; Clay, W. C.; Deaton, D. N.; Hassell, A.
M.; Hoffman, C. R.; Jurgensen, C. H.; Long, S. T.;
McFadyen, R. B.; Miller, A. B.; Miller, L. R.; Payne, J.
A.; Shewchuk, L. M.; Tavares, F. X.; Wells-Knecht, K. J.;
Willard, D. H., Jr.; Wright, L. L. In Third International
Conference on Cysteine Proteinases and their Inhibitors;
Dolinar, M., Turk, B., Turk, V., Eds.; Garamond d. o. o.,
Ljubljana, Slovenia: Portoroz, Slovenia, 2002; p 95.
15. Matsuura, F.; Hamada, Y.; Shioiri, T. Tetrahedron 1993,
49, 8211.
16. A simpler analog containing no P2 moiety, commercially
available N-piperidinecarbonitrile, had also been inactive
at 10,000 nM in the high throughput screen of the
GlaxoWellcome compound collection.
17. Ohwada, T.; Hirao, H.; Ogawa, A. J. Org. Chem. 2004, 69,
7486.