Design of small molecule ketoamide-based inhibitors of cathepsin K

Article abstract of DOI:10.1016/j.bmcl.2003.11.029

A novel series of ketoamide-based inhibitors of cathepsin K has been identified. Modifications to P² and P³ elements were crucial to enhancing inhibitory activity. Although not optimized, a selected inhibitor was effective in attenua

Full text of DOI:10.1016/j.bmcl.2003.11.029

Bioorganic & Medicinal Chemistry Letters 14 (2004) 719–722
Design of small molecule ketoamide-based inhibitors of
cathepsin K
John G. Catalano,a David N. Deaton,a,* Stacey T. Long,b Robert B. McFadyen,a
Larry R. Miller,c J. Alan Payne,c Kevin J. Wells-Knechtd and Lois L. Wrighte
aDepartment of Medicinal Chemistry, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
bDepartment of World Wide Physical Properties, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
cDepartment of Molecular Pharmacology, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
dDepartment of Research Bioanalysis and Drug Metabolism, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
eDiscovery Research Biology, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
Received 8 August 2003; revised 5 November 2003; accepted 14 November 2003
Abstract—A novel series of ketoamide-based inhibitors of cathepsin K has been identified. Modifications to P2 and P3 elements
were crucial to enhancing inhibitory activity. Although not optimized, a selected inhibitor was effective in attenuating type I col-
lagen hydrolysis in a surrogate assay of bone resorption.
# 2003 Elsevier Ltd. All rights reserved.
Osteoporosis is a skeletal disorder characterized by
enhanced bone resorption relative to bone formation
resulting in decreased bone mass and increased suscept-
ibility to fracture.1 Specialized cells, called osteoclasts,
secrete acid and proteolytic enzymes to remove the
mineral and matrix components of bone. Cathepsin K is
the major proteolytic enzyme implicated in degradation
of bone matrix.2 It is a cysteine protease of the papain
superfamily highly expressed in osteoclasts. As part of a
complex with glycosaminoglycans, cathepsin K effi-
ciently degrades type I collagen, the major component
of bone matrix.3 Small molecule inhibitors of cathepsin
K are efficacious in attenuating bone resorption in
animal models of osteoporosis.4
addition to aldehydes, several other electrophiles
including ketones, a-dicarbonyl derivatives, a-ketohe-
terocycles, and nitriles have been employed in reversible
cysteine protease inhibitors.7 Desiring to explore both S
and S0 enzyme subsites, ketoamides were selected as the
aldehyde replacement. Besides being more electrophilic
than ketones and nitriles,8 a-ketoamide-based inhibitors
have also demonstrated oral bioavailability.9
Three general routes were utilized to synthesize the a-
ketoamides. The first method employed a modified cya-
nohydrin procedure based on the precedent of Harbe-
son and colleagues.10 As depicted in Scheme 1, the
aldehyde 2 was converted to a diastereomeric mixture of
cyanohydrins 3. Subsequent acid catalyzed hydrolysis of
the nitrile with concomitant loss of the carbamate pro-
tecting group followed by reprotection of the primary
As part of a larger program to develop new osteo-
porosis therapies, the discovery of small molecule
cathepsin K inhibitors was pursued. A directed screen
of peptide aldehydes identified calpeptin 1 as a potent
cathepsin K inhibitor (IC50=0.11 nM).5 A truncated
analogue, Boc-Nle-H 2, retained reasonable potency
(IC50=51 nM). Since aldehydes can be metabolically
unstable, replacements for this electrophilic group were
explored.6 Considering the chronic nature of osteo-
porosis therapies, reversible inhibitors were required. In
amine provided the known a-hydroxyacids 4.11,12 Car-
0
bodiimide-based amide bond formation with various P1
amines13 and subsequent oxidation afforded the keto-
amides 5a–k. By early incorporation of the P2 sub-
stituent, the cyanohydrin protocol allows the rapid
0
generation of P1 analogues. It also is amenable to the
* Corresponding author. Tel.:+1-919-483-6270; fax:+1-919-315-0430;
e-mail: david.n.deaton@gsk.com
synthesis of P2 analogues. Furthermore, it employs
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.11.029

720
J. G. Catalano et al. / Bioorg. Med. Chem. Lett. 14 (2004) 719–722
Scheme 1. (a) Acetone cyanohydrin, KCN, nBu4NI, PhMe, H2O; (b)
concd HCl, 100 ꢀC; (c) 1N NaOH, Boc2O, THF; (d) amine, EDC,
HOBt, iPr2NEt, DMF, 40–77% over 4 reactions; (e) TEMPO, 5%
NaOCl, KBr, NaHCO3, CH2Cl2, 42–89%.
Scheme 3. (a) PhCOOH, CH2Cl2, 60%; (b) NaOH, dioxane, H2O,
100 ꢀC, 95%; (c) alcohol, 1.93 M COCl2 in PhMe, pyridine, CH2Cl2,
À20 ꢀC to rt; 13, iP rNEt, dioxane, 22–91%; (d) Dess–Martin period-
2
inane, CH2Cl2, 55–78%.
0
Table 1. Inhibition of human cathepsin K by P1 analogues
Compd
R
IC50
mMa
Compd
R
IC50
mMa
Scheme 2. (a) Pyridine, CH2Cl2, 0 ꢀC; 1.93 M COCl2 in PhMe, 0 ꢀC;
1N HCl, 0 ꢀC; distilled @ 2 Torr, 94%; (b) alcohol, PhMe, 85 ꢀC,
sealed tube, 89–99%; (c) LiOHꢁ H2O, THF, H2O; 1N HCl; (d)
Ph3P¼CCN, DMAP, EDC, CH2Cl2, 46–79% over two reactions; (e)
O3, CH2Cl2, À78 ꢀC; N2; amine, À78 ꢀC to rt; AgNO3, THF, H2O, 19–
45%.
5a
5b
5c
5d
5e
5f
5.9
21
5g
5h
5i
2.9
5
inexpensive reagents and is readily amenable to multi-
gram scale.
18
1.2
The second process applied the acyl cyanophosphorane
oxidative cleavage and amine coupling procedure of
Wasserman to generate the ketoamides.14 As displayed
in Scheme 2, the amino acid ester 6 was transformed
into the isocyanate 7. Subsequent coupling to a P2–P3
alcohol15 followed by hydrolysis of the esters provided
acids 8. Coupling the acids 8 with cyanomethyl-
triphenylphosphonium ylide yielded the phosphoranes
9. In situ generation of acyl nitriles via ozonolysis of the
phosphorus–carbon double b0ond, followed by displace-
ment of cyanide with a P1 amine gave the desired
ketoamides 5l–s and 5x. This procedure0 provided the
best method to quickly vary the P1 substituent,
although chemical yields of the final reaction decreased
with increasing scale.
9.1
5j
4.2
>100
6.3
5k
5l
>100
3.4
a Inhibition of recombinant human cathepsin K activity in a fluores-
cence assay using 10 mM 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.
0
The Passerini reaction was exploited as a third tech-
nique to produce ketoamides.16 As shown in Scheme 3,
the aldehyde 1017 was reacted with the isonitrile 1118À20
and benzoic acid to give the a-acyloxyamides 12.
Hydrolysis of both the ester and carbamate afforded the
b-amino-a-hydroxyamides 13. Various P2–P3 alcohols15
were converted to chloroformates and coupled to the
amines 13 followed by oxidation to afford the keto-
amides 5t–w. The Passerini method allowed the swift
synthesis of compounds with different P2 groups, how-
ever volatile isonitriles can be toxic and smelly.
probing the S1 subsite. Adding simple lipophilic groups
like the methyl, ethyl, or isopropyl moieties in analogues
5b–d did not improve potency. Furthermore, the t-butyl
amide 5e demonstrated that additional steric bulk
adjacent to the amide was detrimental to inhibitory
potency.
Since the cathepsins contain a conserved tryptophan
184Trp in cathepsin K) in their active site, amines linked
(
to phenyl substituents were also employed in attempts
to form p–p stacking interactions with its conserved
indole.21 Unfortunately, the phenyl, benzyl, and phe-
nethyl analogues 5f–h offered no significant potency
advantage over the primary amide 5a. Attempting to
increase the population of the active site rotamer of
analogue 5g, a methyl group was inserted into the
As shown in Table 1, replacement of the aldehyde war-
head of inhibitor 2, by an electrophilic ketoamide group
as in analogue 5a resulted in a 100-fold loss in inhibitory
activity. Attempts were made to regain activity by

J. G. Catalano et al. / Bioorg. Med. Chem. Lett. 14 (2004) 719–722
Table 2. Inhibition of human cathepsin K by P2 analogues
721
Compd
R
IC50
nMa
Compd
R
IC50
nMa
5m
34
20
5s
76
32
5n
5o
5p
5q
5t
5u
5v
5w
Figure 1. Rat calvarial resorption assay with 5u. *2 sided t-test, p=0.03.
1.1
2.5
69
The appendage of a P3 substituent to analogue 5m
should further increase inhibitory activity. Taking a clue
from the benzyl carbamate of calpeptin 1, a phenyl ring
was tethered to the neopentyl carbon of 5m by carbon
linkers of various lengths, giving analogues 5o–r. The
two-atom tether 5o was 30-fold more potent than its
precursor. Since analogue 5n has equivalent activity to
5m, most of this improvement in activity of 5o versus
5m probably arises from binding interactions in the S3
subsite rather than from P2 rotamer stabilization by the
chiral center. The more conformationally flexible three-
atom linker 5q retained equivalent activity to 5o despite
its potentially greater entropic cost on binding. In con-
trast, their corresponding diastereomers 5p and 5r are
substantially less active.22
912b
0.79
0.71
5r
1400c
5x
6.3
a Inhibition of recombinant human cathepsin K activity in a fluores-
cence assay using 10 mM 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.
b This analogue contains ꢂ0.1% of 5o.
The insertion of a P3 substituent into these inhibitors
had conferred a substantial boost in potency, but added
another stereogenic center to the inhibitors resulting in
a more complex synthesis and the need for the chiral
separation of enantiomeric precursor alcohols via
HPLC. To address these concerns a series of achiral P2
substituents were incorporated into the inhibitors 5s–x.
The di-iso-propyl, di-tert-butyl, and dicyclopentyl ana-
logues 5u, 5w, and 5x were significantly more active
than the neopentyl analogue 5m. In addition to occu-
pying the S2 pocket, these P2 residues may pick up some
binding interactions in the active site trough that
accommodates the substrate backbone.
c This analogue contains ꢂ0.1% of 5q.
benzylic carbon to give 5i and 5j. Although 5i, derived
from the R-enantiomer, was apparently more active
than 5j, derived from the S-enantiomer, no significant
gains in potency were produced. Furthermore, di-
substitution of the amide nitrogen was not tolerated. The
N-methyl analogue 5k lost all activity. However, repla-
cement of the phenyl ring, as in thiophene 5l, was tol-
erated and could offer a potential site to alter drug
properties of this class of inhibitor. While this pre-
liminary exploration of the S1 subsite provided no sig-
0
Because the active site of rat cathepsin K differs from
the human enzyme (61Asp!Tyr in S3 and 133Ala!Ser
in S2),23 many human cathepsin K inhibitors are less
potent against the rat ortholog making it difficult to
assess their efficacy in rat osteoporosis models.24 Wish-
ing to evaluate the potential developability of this
potent class of cathepsin K inhibitors, analogue 5u,
which is only 10-fold less active versus the rat enzyme
(rat IC50=20 nM), was selected for further evaluation.
In vitro incubations with rat S9 liver homogenates
rapidly metabolized 5u.25 In spite of this ex vivo pre-
diction of rapid clearance, 5u exhibited a low clearance
and long terminal half-life in rats (t1/2=290 min, Cl=8.7
mL/min/kg, Vss=1900 mL/kg). However, this analogue
was poorly soluble in fasted state simulated intestinal
fluid (0.012 mg/mL)26,27 and exhibited moderate
nificant boost in enzyme inhibitory activity, the a-
methyl benzyl amide moiety of analogue 5i was adopted
0
as a suitable P1 substituent for further modifications in
the ketoamide inhibitor series.
Since the primary determinant of enzyme specificity in
the cathepsins is the S2 pocket, it was perceived that P2
region modifications might prove more fruitful. Cal-
peptin 1 contains an isobutyl group as its P2 substituent.
Since the tert-butyl moiety of 5i does not reach as dee-
ply into the S2 pocket, it was extended by an additional
carbon, as in 5m and 5n, to better mimic this feature of
calpeptin. As shown in Table 2, these modifications
resulted in 35-fold and 60-fold increases in potency over
their precursor 5i.

722
J. G. Catalano et al. / Bioorg. Med. Chem. Lett. 14 (2004) 719–722
membrane permeability in the Madin–Darby canine
kidney cell monolayer assay (PAPP=57ꢃ4 nm/s).28 Not
surprising given the low solubility, 5u had poor oral
bioavailability (F=3%) in rats.
9. Adang, A. E. P.; de Man, A. P. A.; Vogel, G. M. T.;
Grootenhuis, P. D. J.; Smit, M. J.; Peters, C. A. M.; Vis-
ser, A.; Rewinkel, J. B. M.; van Dinther, T.; Lucas, H.;
Kelder, J.; van Aelst, S.; Meuleman, D. G.; van Boeckel,
C. A. A. J. Med. Chem. 2002, 45, 4419.
10. Harbeson, S. L.; Abelleira, S. M.; Akiyama, A.; Barrett,
R., III; Carroll, R. M.; Straub, J. A.; Tkacz, J. N.; Wu,
C.; Musso, G. F. J. Med. Chem. 1994, 37, 2918.
11. Beevers, R.; Carr, M. G.; Jones, P. S.; Jordan, S.; Kay,
P. B.; Lazell, R. C.; Raynham, T. M. Bioorg. Med. Chem.
Lett. 2002, 12, 641.
12. Lubisch, W.; Beckenbach, E.; Bopp, S.; Hofmann, H.-P.;
Kartal, A.; Kaestel, C.; Lindner, T.; Metz-Garrecht, M.;
Reeb, J.; Regner, F.; Vierling, M.; Moeller, A. J. Med.
Chem. 2003, 46, 2404.
This analogues poor oral bioavailability precluded a
chronic pharmacodynamic study in the OVX rat. How-
ever 5u was profiled in the ex vivo rat calvaria resorption
assay, a model often predictive of anti-resorptive activ-
ity.29,30 As shown in Figure 1, treatment with parathyroid
hormone enhanced bone resorption as measured by
deoxypyridinoline (DPD) crosslinks release. Analogue 5u
dose-dependently attenuated this release of DPD cross-
links from type I collagen in the rat calvaria resorption
assay. This efficacy reached statistical significance at the
3000 nM dose (2 sided t-test, p=0.03). Thus, this ketoa-
mide cathepsin K inhibitor acts as an antiresorptive agent
by attenuating type I collagen hydrolysis in bone.
13. All of the amines were commercially available.
14. Wasserman, H. H.; Petersen, A. K. Tetrahedron Lett.
1997, 38, 953.
15. Most alcohols were commercially available. Addition of
phenyl acetylene to pivaldehyde, followed by hydrogena-
tion of the triple bond gave the known racemic alcohol
utilized to make analogues 5o and 5p. The enantiomers
were separated via chiral HPLC. Addition of phenpropyl
magnesium bromide to pivaldehyde gave the racemic
alcohol utilized to make analogues 5q and 5r. The enan-
tiomers were separated via chiral HPLC. Cyclopentyl
magnesium bromide addition to ethyl formate gave the
achiral alcohol utilized to make analogue 5x.
16. Semple, J. E.; Owens, T. D.; Nguyen, K.; Levy, O. E. Org.
Lett. 2000, 2, 2769.
17. Prasad, J. V. N. V.; Rich, D. H. Tetrahedron Lett. 1990,
31, 1803.
18. The known isonitrile 11 was synthesized from (R)-(+)-1-
phenylethylamine, employing the phase transfer catalyzed
modification of the classical Hofmann carbylamine
synthesis.
In summary, this report highlights the discovery of a
potent series of cathepsin K inhibitors. Starting from
the directed screening hit 2, the aldehyde warhead was
replaced with a more drug-like ketoamide electrophile.
The resulting drop in potency was recovered and
enhanced via modifications to the P1 , P2 and P3 sub-
0
stituents. In addition, a representative inhibitor in this
ketoamide series was efficacious in attenuating bone
resorption in a surrogate assay of osteoporosis. Sub-
sequent reports will detail efforts to improve the oral
bioavailability of these cathepsin K ketoamide inhibitors.
Acknowledgements
19. Weber, W. P.; Gokel, G. W.; Ugi, I. K. Angew. Chem.,
Int. Ed. Engl. 1972, 11, 530.
20. Gollnick, K.; Koegler, S.; Maurer, D. J. Org. Chem. 1992,
57, 229.
21. Somoza, J. R.; Palmer, J. T.; Ho, J. D. J. Mol. Biol. 2002,
322, 559.
The authors would like to thank Melissa A. Gomez and
Manon S. Villeneuve for chiral separations of racemic
alcohols utilized to synthesize analogues 5o–r and James
G. Conway for statistical analyses.
22. The residual potency of diastereomers 5p and 5r probably
arises from contamination of their samples with trace
amounts (ꢂ0.1% by HPLC) of the respective active dia-
stereomers 5o and 5q.
References and notes
23. McQueney, M. S.; Feild, J.; Hanning, C. R.; Brun, K.;
Ramachandran, K.; Connor, J.; Drake, F.; Jones, C. S.;
Amegadzie, B. Y. Protein Exp. Purif. 1998, 14, 387.
24. Veber, D. F.; Yamashita, D. S.; Oh, H.-J.; Smith, B. R.;
Salyers, K.; Levy, M.; Lee, C.-P.; Marzulli, A.; Smith, P.;
Tomaszek, T.; Tew, D.; McQueney, M.; Stroup, G. B.;
Lark, M. W.; James, I. E.; Gowen, M. In Peptides for the
New Millennium, Proceedings of the American Peptide
Symposium, 16th, Minneapolis, MN, United States, June
26–July 1, 1999 2000, p 453–455.
25. In rat S9, 9% of analogue 5u remained after 30 min of
incubation.
26. Dressman, J. B.; Amidon, G. L.; Reppas, C.; Shah, V. P.
Pharm. Res. 1998, 15, 11.
27. Kostewicz, E. S.; Brauns, U.; Becker, R.; Dressman, J. B.
Pharm. Res. 2002, 19, 345.
28. Irvine, J. D.; Takahashi, L.; Lockhart, K.; Cheong, J.;
Tolan, J. W.; Selick, H. E.; Grove, J. R. J. Pharm. Sci.
1999, 88, 28.
1. Einhorn, T. A. In Osteoporosis; Marcus, R., Feldman, D.,
Kelsey, J., Eds.; Academic Press Inc: San Diego, Cali-
fornia, 1996; p 3–22.
2. Yamashita, D. S.; Dodds, R. A. Curr. Pharm. Des. 2000,
6, 1.
3. Li, Z.; Hou, W.-S.; Escalante-Torres, C. R.; Gelb, B. D.;
Bromme, D. J. Biol. Chem. 2002, 277, 28669.
4. 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.
5. Catalano, J. G.; Deaton, D. N.; Furfine, E. S.; Hassell, A.
M.; McFadyen, R. B.; Miller, A. B.; Miller, L. R.; Shew-
chuk, L. M.; Willard, D. H., Jr.; Wright, L. L. Bioorg.
Med. Chem. Lett. 2004, 14, 275.
6. Blume, H.; Oelschlager, H. Arzneim. Forsch. 1978, 28, 956.
7. Leung-Toung, R.; Li, W.; Tam, T. F.; Karimian, K. Curr.
Med. Chem. 2002, 9, 979.
8. Jones, D. M.; Atrash, B.; Teger-Nilsson, A.-C.; Gyzan-
der, E.; Deinum, J.; Szelke, M. Lett. Pept. Sci. 1995, 2,
147.
29. Hahn, T. J.; Westbrook, S. L.; Halstead, L. R. Endocrinol.
1984, 114, 1864.
30. Conaway, H. H.; Grigorie, D.; Lerner, U. H. J. Endocri-
nol. 1997, 155, 513.