Arylaminoethyl carbamates as a novel series of potent and selective cathepsin S inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.07.032

We report a novel series of noncovalent inhibitors of cathepsin S. The synthesis of the peptidomimetic scaffold is described and structure-activity relationships of P3, P1, and P1′ subunits are discussed. Lead optimization to a non-peptidic scaffold has r

Full text of DOI:10.1016/j.bmcl.2006.07.032

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5107–5111
Arylaminoethyl carbamates as a novel series of potent and
selective cathepsin S inhibitors
David C. Tully,* Hong Liu, Arnab K. Chatterjee, Phil B. Alper, Jennifer A. Williams,
Michael J. Roberts, Daniel Mutnick, David H. Woodmansee, Thomas Hollenbeck,
Perry Gordon, Jonathan Chang, Tove Tuntland, Christine Tumanut, Jun Li,
Jennifer L. Harris and Donald S. Karanewsky
Genomics Institute of the Novartis Research Foundation, 10675 John J. Hopkins Dr., San Diego, CA 92121, USA
Received 18 May 2006; revised 8 July 2006; accepted 11 July 2006
Available online 28 July 2006
Abstract—We report a novel series of noncovalent inhibitors of cathepsin S. The synthesis of the peptidomimetic scaffold is
described and structure–activity relationships of P3, P1, and P10 subunits are discussed. Lead optimization to a non-peptidic scaffold
has resulted in a new class of potent, highly selective, and orally bioavailable cathepsin S inhibitors.
ꢀ 2006 Elsevier Ltd. All rights reserved.
P1'
R
Cathepsin S (Cat S) is a lysosomal cysteine protease that
is expressed in antigen presenting cells such as macro-
phages, dendritic cells, and B cells. Cat S plays a critical
role in the targeted processing of the invariant chain (Ii),
a chaperone protein for the major histocompatibility
class II complex (MHC-II).1–3 Proteolytic removal of
Ii from the MHC-II binding groove is a prerequisite
for productive antigen loading onto the MHC-II com-
plex,4 and consequently, inhibition of Cat S attenuates
antigen presentation to CD4+ T-cells. The resulting
immunosuppression is specific for CD4+ T-cells, making
cathepsin S an attractive therapeutic target for the mod-
ulation and regulation of immune hyperresponsive dis-
eases such as myasthenia gravis, multiple sclerosis, and
rheumatoid arthritis.5–7
P3
P1
R
O
O
H
N
N
N
N
F
H
O
1a R,R = H
Cat S Ki = 0.087 μM Cat S Ki = 0.011 μM
Cat K Ki = 5.130 μM Cat K Ki = 16.9 μM
Cat L Ki = 0.512 μM Cat L Ki = 0.810 μM
1b R,R = CH3
P2
O
O
O
N
N
2
N
H
F
Cat S Ki = 0.662 μM
Cat K Ki = 8.79 μM
Cat L Ki = >30.0 μM
O
We have recently reported on a series of arylaminoethyl
amides including compounds such as 1a and 1b (Fig. 1),
which lack an electrophilic covalent warhead and act as
reversible, competitive inhibitors of cathepsin S.8–11
Compound 1a (Ki = 0.087 lM) has only a narrow win-
dow of selectivity over closely related cathepsin L, while
1b and many of its closely related analogs exhibit poor
pharmacokinetic (PK) properties in rats.11 Lead optimi-
3
Cat S Ki = 0.260 μM
Cat K Ki = >30.0 μM
Cat L Ki = >30.0 μM
O
O
N
N
O
N
F
H
O
Figure 1.
zation efforts have focused primarily on improving the
PK properties by reducing the peptidic nature of the ser-
ies while retaining the potency and selectivity of our ear-
lier, more peptidic chemotypes.
Keywords: Cathepsin; Cysteine protease inhibitor; Non-covalent inhib-
itor; Peptidomimetic.
*
Corresponding author. Tel.: +1 8588121559; fax: +1 8588121648;
e-mail: dtully@gnf.org
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.07.032

5108
D. C. Tully et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5107–5111
F
Medicinal chemistry optimization efforts have been aid-
ed by structural data that we and others have reported
involving co-crystal structures of active-site inhibitors
bound to cathepsin S.10–14 In peptidic Cat S inhibitors
such as arylaminoethyl amides10,11 or dipeptide nitr-
iles,13 the P2 amide-NH forms a hydrogen bond with
the Cat S backbone carbonyl of Gly69. In order to
determine whether this interaction was essential for
binding to the protein, we synthesized the carbamate
analog of urea 1a. Lactic acid analog 2, in which the
P2 NH has been replaced by oxygen,15 was not as active
toward Cat S (Ki = 0.662 lM), but the corresponding
drop-off in Cat L inhibitory activity was much more
pronounced (>60·). Initial efforts to exploit this diver-
gence resulted in compound 3 (Cat S Ki = 0.260 lM),
in which the connectivity of the main chain has been
directly transposed from 2. The Cat S inhibitory activity
was retained, and more importantly the selectivity over
cathepsins K and L was improved to well over 100-fold.
Furthermore, the kinetics of inhibition was shown to be
fully reversible as the enzymatic activity was fully
restored following dilution and dialysis of the enzyme–
inhibitor complex.16
F
R1
NAr =
NAr
N
N
H2N
b
c
e
7a-e
F
OCF3
F
N
N
N
H
a
d
Figure 2. N-Aryl diamines (Ref. 16).
and 1b. This becomes more apparent with an increasing
size of the P1 alkyl substituent, as the hydroxymethyl 9
is nearly equipotent with 3, while the benzyloxymethyl
10 analog (Cat S Ki = 0.008 lM) gains over 30-fold in
inhibitory activity over compound 3. Moreover, com-
pound 10 is highly selective (>1000-fold) over both
cathepsins K and L—a major contrast to the arylamino-
ethyl amides described previously which typically
exhibit a decrease in selectivity over Cat L with increas-
ing Cat S potency that is derived from larger P1 alkyl
substituents.9–11
The a-hydroxyacids 4 were prepared from their respec-
tive lactic acid or amino acid precursors. The synthesis
of carbamate 3 and related analogs with the P3 morpho-
line amide is shown in Scheme 1. Morpholine amides 5,
prepared by standard peptide coupling conditions using
PyBOP, were activated as the nitrophenyl carbonates 6,
and then condensed with the desired N-aryl diamines
7a–e (Fig. 2).17
Replacement of the 5-fluoroindoline moiety at the prime
side with a 4-trifluoromethoxyaniline led to a slight
improvement (ꢀ2.5·) in potency (11 vs. 3), while larger
alkyl groups in P1 cooperatively contributed to an addi-
tional boost in the potency of cathepsin S inhibitors (12,
13, and 14). The cyclohexylmethyl group has previously
been shown to be a preferred P2 group for cathepsin S
inhibitors of varied chemotypes, in particular arylami-
noethyl amides8–11 and dipeptide nitriles,13 however,
other hydrophobic groups are often well tolerated.5 In
contrast, neither the benzyl (15) nor the tert-butylmethyl
(16) P2 moieties are well tolerated by Cat S in the con-
text of this current carbamate series. The well-defined
hydrophobic S2 pocket of Cat S is known from reported
The SAR of a representative set of examples is illustrat-
ed in Table 1. It is clear that an alkyl group in the P1 po-
sition of this arylaminoethyl carbamate series is required
for sub-micromolar activity as demonstrated by com-
pound 8 (Cat S Ki = 2.95 lM). This is in contrast to
the arylaminoethyl amide series described in previous re-
ports which were typically very potent even without a P1
substituent.8–11 Presumably a different mode of cross-
talk between P1 and the other subsites predominates
in the case of this non-peptidic carbamate scaffold in
such a manner that the conformation assumed by the
inhibitor allows for a more favorable interaction (partic-
ularly P1–S1 and P10–S10) with the enzyme than exists
with the peptidic scaffold for compounds such as 1a
X-ray structural data to be a comparatively deep pocket
10–14
(versus Cat K and L) with a narrow entrance.
It is
possible, in this case, that the trajectory of suboptimal
P2 groups emanating from a non-peptidic backbone
leads to hydrophobic clash with the narrow entrance
of the S2 pocket.
In light of this SAR, it appeared that the cyclohexyl-
methyl was the optimal choice for the P2 substituent
both in terms of potency and selectivity regardless of
the P1 and P10 moieties. Continuing with the optimiza-
tion of the P10 moiety led to the 2,2-dimethyl-5-fluoro-
R2
R2
O
indoline 17, which showed
a
moderate (3·)
a
b
N
improvement in potency over the unsubstituted 5-flu-
oroindoline 3. A more significant improvement was seen
with the 3,3-gem-dimethyl indoline analog 18, which is
considerably more potent than the unsubstituted indo-
line 3 (>10·), and especially noteworthy considering it
did not require a P1 substituent larger than a methyl
group for this activity. This is further exemplified with
compound 19, which is equipotent with 18 despite bear-
ing a larger cyclopropyl group in P1. Moreover, the
selectivities of indolines 17, 18, and 19 over cathepsins
K and L are greater than three orders of magnitude.
Compound 20 further illustrates the influence that the
HO
OH
OH
R2
O
O
N
5
4
NO2
R2
O
O
R1
O
O
N
NAr
c
O
N
O
O
H
O
O
8-21
6
Scheme 1. Reagents and conditions: (a) morpholine, PyBOP, CH2Cl2,
0 ꢁC to rt, 74–89%; (b) 4-nitrophenylchloroformate, pyridine, CH2Cl2,
rt, 62–91%; (c) 7a–d, DIEA, DMF, rt, 48–89%.

D. C. Tully et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5107–5111
Table 1. Inhibition of cathepsin S, K, and La
5109
R2
R1
O
O
N
NAr
O
N
H
O
Compound
R2
R1
NAr
Cat S Ki (lM)
Cat K Ki (lM)
Cat L Ki (lM)
N
3
–CH3
0.226
2.95
>30
>30
>30
F
N
N
N
8
–H
>30
>30
F
F
F
9
–CH2OH
–CH2OCH2Ph
–CH3
0.143
0.008
0.100
0.016
0.008
0.018
>100
10
11
12
13
14
15
16
8.71
>30
17.7
H
N
>100
OCF3
OCF3
OCF3
OCF3
OCF3
OCF3
H
N
–CH2CH2CH3
–CH(CH3)2
–CH2CH2Ph
–CH3
1.39
1.18
3.97
>100
0.355
5.47
H
N
6.03
9.79
H
N
H
N
12.6
>100
H
N
–CH3
1.05
>100
>100
>100
>30
17
18
19
–CH3
–CH3
0.080
0.019
0.019
>100
>100
>30
N
N
F
F
N
F
20
–CH3
0.122
>30
>30
N
F
prime side indoline has on binding, as the subtle varia-
tion to the 3,3-spiro-cyclopropyl indoline resulted in a
6-fold drop-off in potency.
line P10 group was considered to confer the optimal
properties for both potency and selectivity. The synthe-
sis of these P3 analogs is shown in Scheme 2. Protection
of a-hydroxyacid 4a as the benzyl ester 21 was followed
by conversion to the para-nitrophenylcarbonate 22. The
activated carbonate 22 was condensed with diamine 7d,
Exploration of the SAR in P3 was done with direct ana-
logs of compound 18, as the 5-fluoro-3,3-dimethylindo-

5110
D. C. Tully et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5107–5111
NO2
a
b
c
O
O
O
O
O
HO
HO
OH
21
OH
OH
O
O
O
O
d
22
4a
O
O
e,f
P3
N
O
N
O
N
O
N
F
H
F
H
O
O
24-33
23
Scheme 2. Reagents and conditions: (a) H2, 1000 psi, 5% Rh/C, 10% aq EtOH, 50 ꢁC, 18 h, 90%; (b) PhCH2Br, Et3N, DMF, rt, 87%; (c) 4-
nitrophenylchloroformate, pyridine, THF, 60 ꢁC, 83%; (d) 7d, DIEA, THF, 57%; (e) H2 (40 psi), 5% Pd/C, EtOH, 94%; (f) P3 amine, HATU, DIEA,
DMF, 63–87%.
Table 2. Inhibition of cathepsin S by P3 analogsa
debenzylated under hydrogenolysis conditions, and then
reacted with the desired P3 amine using standard pep-
tide coupling conditions to afford compounds 24–33.
O
P3
N
The SAR of a selected set of P3 analogs is shown in
Table 2. All of the compounds listed in Table 2 were
inactive against both cathepsins K and L in the concen-
tration range tested (Ki > 100 lM), so only the cathepsin
S inhibition constants are reported. Modification of the
morpholine ring by addition of methyl groups (24) or
replacement with a linear amide (25) was deleterious,
leading to a 10-fold or more drop-off in activity. Like-
wise, the simpler dimethylamide (26) and azetidine
amide (27) were roughly equipotent with 25. However,
increasing the ring size to the pyrrolidine (28), piperidine
(29), and thiomorpholine (30) returned most of the
inhibitory activity as compared to compound 18. A
six-membered ring is clearly preferred in the hydropho-
bic S3 pocket, with the enzyme being particularly sensi-
tive to subtle changes in polarity of this P3 group.
Although we have seen empirically from recent SAR
involving a more peptidic scaffold that the S3 pocket
can accommodate much larger hydrophobic P3 groups,8
it is interesting to note that a further increase in the
polarity and steric bulk at the 4-position was detri-
mental to the Cat S inhibitory activity, as was exempli-
fied with sulfone 31 and substituted piperazines 32
and 33.
O
N
H
F
O
Compound
P3
Cat S Ki (lM)
18
0.019
O
O
N
24
25
26
27
28
29
30
31
32
0.297
0.172
0.227
0.217
0.056
0.039
0.023
0.422
0.986
N
Me
MeO
N
Me2N
N
N
N
The pharmacokinetic profiles of carbamates 17 and 18
were determined in male Wistar rats, and the data are
presented in Table 3 alongside the arylaminoethyl amide
analog 1b for direct comparison of the two scaffold
types. Despite its similarly high clearance (Cl = 52 mL/
min/kg), carbamate 18 exhibited an increased terminal
half-life (t1/2 = 100 min), higher steady state volume of
distribution (Vss = 4.10 L/kg), and good oral bioavail-
ability (F = 42%), which was an order of magnitude
improvement over its amide analog 1b (F = 4%). The
high clearance for each of these compounds was sugges-
tive of extensive first-pass metabolism. A detailed
metabolite ID study following 30 min incubation in rat
liver microsomes confirmed the primary metabolic prod-
S
N
O
O
S
N
O
N
N
H3C
H3C
O
S
33
1.08
N
N
O

D. C. Tully et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5107–5111
Table 3. Pharmacokinetic data for selected analogsa
5111
Compound
Single iv dose (3 mg/kg)
Cl (mL/min/kg)
Single po dose (10 mg/kg)
t1/2 (min)
Vss (L/kg)
AUC (min lg/mL)
Cmax (nM)
T1/2 (min)
F (%)
1b
17
18
43
82
52
52
52
1.76
3.10
4.10
8
44
81
118
513
774
51
62
69
4
23
42
100
a Pharmacokinetic data in male Wistar rats, where values are means of three individual experiments.
7. Nakagawa, T. Y.; Brissette, W. H.; Lira, P. D.; Griffiths,
R. J.; Petrushova, N.; Stock, J.; McNeish, J. D.; Eastman,
S. E.; Howard, E. D.; Clarke, S. R. M.; Rosloniec, E. F.;
Elliott, E. A.; Rudensky, A. Y. Immunity 1999, 10, 207.
8. Liu, H.; Tully, D. C.; Epple, R.; Bursulaya, B.; Li, J.;
Harris, J. L.; Williams, J. A.; Russo, R.; Tumanut, C.;
Roberts, M. J.; Alper, P. B.; He, Y.; Karanewsky, D. S.
Bioorg. Med. Chem. Lett. 2005, 15, 4979.
9. Alper, P. B.; Liu, H.; Chatterjee, A. K.; Nguyen, K. T.;
Tully, D. C.; Tumanut, C.; Li, J.; Harris, J. L.; Tuntland,
T.; Chang, J.; Gordon, P.; Hollenbeck, T.; Karanewsky,
D. S. Bioorg. Med. Chem. Lett. 2006, 16, 1486.
10. Tully, D. C.; Liu, H.; Alper, P. B.; Chatterjee, A. K.;
Epple, R.; Roberts, M. J.; Williams, J. A.; Nguyen, K. T.;
Woodmansee, D. H.; Tumanut, C.; Li, J.; Spraggon, G.;
Harris, J. L.; Chang, J.; Tuntland, T.; Karanewsky, D. S.
Bioorg. Med. Chem. Lett. 2006, 16, 1975.
ucts arose from hydroxylation of the cyclohexyl P2 moi-
ety and oxidation/dehydration on the indoline ring.
In summary, this report details the optimization of
arylaminoethyl amides 1a and 1b by transforming the
peptidic main chain into a carbamate scaffold, resulting
in potent and highly selective cathepsin S inhibitors such
as 10 and 18. Moreover, this transformation resulted in
a significant improvement in oral bioavailability, as
compound 18 demonstrated a 10-fold enhancement over
earlier lead 1b. Subsequent reports will detail the logical
progression of the medicinal chemistry efforts to further
improve upon the pharmacokinetic and physicochemi-
cal properties of this series of cathepsin S inhibitors.
11. Tully, D. C.; Liu, H.; Chatterjee, A. K.; Alper, P. B.;
Epple, R.; Williams, J. A.; Roberts, M. J.; Woodmansee,
D. H.; Masick, B. T.; Tumanut, C.; Li, J.; Spraggon, G.;
Hornsby, M.; Chang, J.; Tuntland, T.; Hollenbeck, T.;
Gordon, P.; Harris, J. L.; Karanewsky, D. S. Bioorg. Med.
Chem. Lett. preceding paper.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2006.07.032.
12. Pauly, T. A.; Sulea, T.; Ammirati, M.; Sivaraman, J.;
Danley, D. E.; Griffor, M. C.; Kamath, A. V.; Wang, I.
K.; Laird, E. R.; Seddon, A. P.; Menard, R.; Cygler, M.;
Rath, V. L. Biochemistry 2003, 42, 3203.
13. Ward, Y. D.; Thomson, D. S.; Frye, L. L.; Cywin, C. L.;
Morwick, T.; Emmanuel, M. J.; Zindell, R.; McNeil, D.;
Bekkali, Y.; Giradot, M.; Hrapchak, M.; DeTuri, M.;
Crane, K.; White, D.; Pav, S.; Wang, Y.; Hao, M.-H.;
Grygon, C. A.; Labadia, M. E.; Freeman, D. M.;
Davidson, W.; Hopkins, J. L.; Brown, M. L.; Spero, D.
M. J. Med. Chem. 2002, 45, 5471.
14. Mcgrath, M. E.; Palmer, J. T.; Bromme, D.; Somoza, J. R.
Protein Sci. 1998, 7, 1294.
15. Synthesis of compound 2: (a) 4a, 7b, HATU, DIEA,
DCM, 57%; (b) 4-morpholinecarbonyl chloride, NaH,
DMF, rt, 42%.
References and notes
1. (a) Riese, R. J.; Mitchell, R. N.; Villadangos, J. A.; Shi, G.
P.; Palmer, J. T.; Karp, E. R.; De Sanctis, G. T.; Ploegh,
H. L.; Chapman, H. A. J. Clin. Invest. 1998, 101, 2351; (b)
Villandangos, J. A.; Bryant, R. A. R.; Deussing, J.;
´
Driessen, C.; Lennon-Dumenil, A.-M.; Riese, R. J.; Roth,
W.; Saftig, P.; Shi, G. P.; Chapman, H. A.; Peters, C.;
Ploegh, H. L. Immunol. Rev. 1999, 172, 109; (c) Shi, G. P.;
Villandangos, J. A.; Dranoff, F.; Small, C.; Gu, L.; Haley,
K. J.; Riese, R.; Ploegh, H. L.; Chapman, H. A. Immunity
1999, 10, 197.
2. (a) Nakagawa, T. Y.; Rudensky, A. Y. Immunol. Rev.
1999, 172, 121; (b) Honey, K.; Rudensky, A. Y. Nat.
Immunol. 2003, 3, 472.
3. Beers, C.; Burich, A.; Kleijmeer, M. J.; Griffith, J. M.;
Wong, P.; Rudensky, A. Y. J. Immunol. 2005, 174, 1205.
4. (a) Riese, R. J.; Wolf, P. R.; Bro¨mme, D.; Natkin, L. R.;
Villadangos, J. A.; Ploegh, H. L.; Chapman, H. A.
Immunity 1996, 4, 357; (b) Driessen, C.; Bryant, R. A.
16. Experimental details of the enzymatic assays and the
dilution-dialysis experiment with compound 18 are includ-
ed in the Supplementary material.
17. Diamines 7a–e were prepared as previously described in
Refs. 9–11, and the spiro-cyclopropyl indoline in diamine
7e was synthesized as indicated below:
´
R.; Lennon-Dumenil, A.-M.; Villadangos, J. A.; Bryant,
O
P. W.; Shi, G.-P.; Chapman, H. A.; Ploegh, H. L. J. Cell
Biol. 1999, 147, 775.
F
F
F
O
O
c,d,e
a,b
5. (a) Thurmond, R. L.; Sun, S.; Karlsson, L.; Edwards, J. P.
Curr. Opin. Invest. Drugs 2005, 6, 473; (b) Leroy, V.;
Thurairatnam, S. Expert Opin. Ther. Patents 2004, 14, 301.
6. Yang, H.; Kala, M.; Scott, B. G.; Goluszko, E.; Chapman,
H. A.; Christadoss, P. J. Immunol. 2005, 174, 1729.
N
N
N
H
PMB
H
(a) NaH, PMBCl, DMF, 0 ꢁC, 82%; (b) H2NNH2, EtOH,
reflux, 90%; (c) NaH, 1,2-dibromoethane, DMF, 0 ꢁC,
61%; (d) TFA, 60 ꢁC, 75%; (e) LAH, 0 ꢁC, 60%.