Entry into a new class of potent proteasome inhibitors having high antiproliferative activity by structure-based design

Article abstract of DOI:10.1021/jm049660v

Proteasome inhibition is a therapeutic concept of current interest in anticancer research. We report here the design, synthesis, and biological characterization of prototypes of a new class of noncovalent proteasome inhibitors showing high activity in bio

Full text of DOI:10.1021/jm049660v

4810
J . Med. Chem. 2004, 47, 4810-4813
Ta ble 1. IC50 Values (µM) in Enzymatic and Cellular Assays
En tr y in to a New Cla ss of P oten t
P r otea som e In h ibitor s Ha vin g High
An tip r olifer a tive Activity by
Str u ctu r e-Ba sed Design
compd
chyma
PGPHa
trypa
prolb
cellc
1
2
3
4
5
0.9
>20
>20
>20
>20
>20
>20
0.007
0.015
1.0
1.5
0.06
0.02
0.06
>20
>20
0.9
Pascal Furet,* Patricia Imbach,* Maria Noorani,
J uergen Koeppler, Kurt Laumen, Marc Lang,
Vito Guagnano, Peter Fuerst, J ohannes Roesel,
J ohann Zimmermann, and
a
Inhibition of chymotrypsin-like (chym), post-glutamyl-peptide
(PGPH), and trypsin-like (tryp) proteolytic activity of purified
b
human proteasome. Inhibition of proliferation of MDA-MB-435
cells. c Inhibition of chymotrypsin-like activity of the proteasome
in cultured MDA-MB-435 cells.
Carlos Garc´ıa- Echeverr´ıa*
Novartis Pharma AG, Novartis Institutes for Biomedical
Research, CH-4002 Basel, Switzerland
Noncovalent inhibitors of the proteasome seem to
have been investigated less extensively judging from the
paucity of reports on this type of inhibitor.21 In principle,
such inhibitors should be devoid of the inherent draw-
backs associated with the classical reactive warhead
groups (i.e., lack of specificity, excessive reactivity, and
instability), and we decided to explore this noncovalent
strategy in our current drug-discovery activities aimed
at blocking the proteasome-ubiquitin pathway in tumor
cells.
Received May 6, 2004
Abstr a ct: Proteasome inhibition is a therapeutic concept of
current interest in anticancer research. We report here the
design, synthesis, and biological characterization of prototypes
of a new class of noncovalent proteasome inhibitors showing
high activity in biochemical and cellular assays.
The proteasome is a multicatalytic protease complex
that degrades intracellular proteins tagged for destruc-
tion by the covalent attachment of multiple ubiquitin
molecules.1,2 The ubiquitin-proteasome system is in-
volved in the degradation of key components of the
molecular machinery on which rely such important
cellular functions as transcription, cell-cycle progression,
tumor suppression, and apoptosis.3-8 Given the wide
range of substrates and processes that are regulated by
this system, the components of the ubiquitin-protea-
some pathway have become the focus of intensive
biochemical research, especially in the oncology area.
Although this pathway can be blocked at various steps,
most inhibitors have been designed to target the pro-
teolytic activities of the proteasome. Thus, it was shown
that inhibition of proteasomal activity induced cell-cycle
arrest and apoptosis in tumor cells, thereby leading to
antiproliferative effects.9-11 The observation that ma-
lignant cells were more susceptible to the proapoptotic
effects of proteasome inhibition than normal cells raised
the notion of proteasome inhibition as a potential new
approach in cancer therapy.12-15
Most currently available inhibitors of the proteasome
exert their inhibitory action by adduct formation with
the enzyme.16-18 The 20S proteasome, which is the
catalytic core of the proteasome, is an N-terminal
threonine hydrolase.19 The hydroxyl group on the N-
terminal threonine of each â-subunit is prone to react
with compounds possessing functional groups receptive
to nucleophilic attack. 20S proteasome inhibitors of this
type include natural products and synthetic peptides
belonging to the following classes of reactive com-
pounds: epoxyketones, aldehydes, boronic acids, R-
ketoamides, R-ketoaldehydes, and vinyl sulfones.20
We have reported in several publications the discov-
ery and optimization of 2-aminobenzylstatine deriva-
tives that inhibit noncovalently and with high selectivity
the chymotrypsin-like peptidase activity of the human
20S proteasome (1 and 2).22-24 In this class, we could
improve by 2 orders of magnitude the inhibitory activity,
in a biochemical assay, of the initial micromolar hit 1.
However, only modest cellular activity could be achieved
with this compound class (e.g., derivative 2 in Table 1).
Assuming that poor cell penetration was the reason our
2-aminobenzylstatine compounds did not express their
high enzymatic inhibitory activity at the cellular level,
we engaged in efforts to identify alternative inhibitor
scaffolds of reduced size and attenuated peptidic char-
acter. We report herein the discovery by structure-based
design of the first representatives of a new class of
potent, noncovalent 20S proteasome inhibitors that, in
contrast to the 2-aminobenzylstatine inhibitors, show
high activity in cellular assays.
The active site responsible for the chymotrypsin-like
proteolytic activity of the human 20S proteasome is
formed by the association of â-subunits X and HC5.25
Using the X-ray crystal structure of the yeast protea-
some,26 we constructed a homology model of these
â-subunits.23 The model served to establish a binding
mode hypothesis for initial compound 1 and was instru-
mental to the optimization of this class of noncovalent
20S proteasome inhibitors.24 The binding mode is il-
lustrated in Figure 1 with 2, one of the most potent
noncovalent 20S proteasome inhibitors reported to date.
According to this binding model, which is supported
by extensive structure-activity relationships, crucial to
the affinity of the 2-aminobenzylstatine inhibitors for
the X/HC5 active site is a set of hydrogen bond inter-
actions. Four â-sheetlike hydrogen bonds are formed
between the amide bonds flanking the valine residue
of the inhibitor and the main chain of subunit X at
residues Thr 21, Gly 47, and Ala 49. These hydrogen
bonds position the inhibitors in the active site in such
* To whom correspondence should be addressed. For P.F.: phone,
41 61 696 79 90; fax, 41 61 696 15 67; e-mail, pascal.furet@
pharma.novartis.com. For P.I.: phone, 41 61 696 62 56; fax, 41 61 696
62 46; e-mail, patricia.imbach@pharma.novartis.com. For C.G.-E.
phone, 41 61 696 10 94; fax, 41 61 696 62 46; e-mail,
carlos.garcia-echeverria@pharma.novartis.com.
10.1021/jm049660v CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/19/2004

Letters
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 20 4811
F igu r e 1. Model of 2 bound to the proteasome X/HC5
subunits. Hydrogen bonds are shown in magenta.
a way that their C-terminal phenol and statine 2-ami-
nobenzyl moieties fill the enzyme’s S1 and S3 pockets,
respectively, where they form additional hydrophobic
and hydrogen bond interactions. In particular, the
statine 2-amino group establishes a hydrogen bond with
residue Asp 153 of subunit HC5 located on the rim of
the S3 pocket. The tert-leucine side chain and the
lipophilic N-terminal group in the inhibitors also form
favorable interactions with the active site by occupying
small accessory hydrophobic pockets labeled AS1 and
AS2 in Figure 1. In contrast, the two amide bonds
flanking the tert-leucine residue and the statine 4-benzyl
group have no significant interactions with the X/HC5
active site in our binding model. This hypothesis
prompted us to envisage the design of new simplified
proteasome inhibitor scaffolds in which these nonpro-
ductive structural features are removed while main-
taining the above-described favorable interactions. De-
sign by interactive molecular modeling led to several
ideas for realizing this objective. The most appealing
one was a dipeptide scaffold represented by prototype
3.
F igu r e 2. Top: model of designed 3 bound to the proteasome
X/HC5 subunits. Hydrogen bonds are shown in magenta.
Bottom: chemical structures of 2 and 3 showing the cor-
respondence between their respective chemical moieties in
terms of interaction with the proteasome binding site. Match-
ing main chain features are represented in the same color,
while matching side chain features are labeled by the name
of the pocket to which they bind.
Figure 2 shows how the important structural features
of the 2-aminobenzylstatine inhibitors are incorporated
or mimicked in this molecule. As can be seen, the two
crucial amide bonds flanking the valine residue and the
C-terminal phenol group filling the S1 pocket are
preserved whereas the statine moiety is replaced by a
3,4,5-trimethoxy-L-phenylalanine residue. This nonpro-
teinogenic amino acid is able to form exactly the same
interactions with the S3 pocket as the entire 2-ami-
nobenzylstatine moiety. The two other critical pharma-
cophore features, i.e., the tert-leucine side chain and the
hydrophobic N-terminal group, are mimicked in the
prototype compound by a single phenoxy substituted
benzylic N-terminal group. Modeling suggested that the
two phenyl rings of this bulky N-terminal group present
a spatial arrangement adequate to simultaneously fill
the AS1 and AS2 accessory hydrophobic pockets.
Ch a r t 1
Prototypes 4 and 5 (Chart 1), which contrary to 3 do
not incorporate all the favorable binding elements
present in 2, were also envisaged for synthesis to further
probe the validity of our design concept.
The synthesis of 1 and 2 has been reported pre-
viously,22-24 and 3-5 were prepared in a stepwise
procedure by standard solution peptide chemistry. The
general route for the synthesis of these new 20S
proteasome inhibitors is illustrated for 3 in Scheme 1.
HCl‚(S)-Val-(2-hydroxy-4-methoxy)benzylamine (a ) was
obtained by condensation of NR-Boc-L-Val-OH with
2-aminomethyl-5-methoxyphenol followed by NR-Boc
deprotection with 4 N HCl in dioxane.22-24 NR-Boc-3,4,5-
L-trimethoxyphenylalanine (b) was obtained by known

4812 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 20
Letters
Sch em e 1a
cell lines (e.g., IC50 ) 60 nM, MDA-MB-435 cells).
Reduction of the molecular weight and decrease of the
number of amide bonds thus appear to have been an
appropriate strategy to achieve high cellular activity.
In conclusion, we have discovered a new class of
potent, cellularly active inhibitors of the proteasome.
Unlike most previously reported proteasome inhibitors,
the new inhibitors act noncovalently and show high
specificity for the chymotrypsin-like activity of the
enzyme. These unique properties make 3 an interesting
tool for investigations of proteasome function in many
aspects of cellular regulation. In addition, the high
antiproliferative activity obtained is encouraging in our
efforts toward developing an anticancer drug based on
the concept of proteasome inhibition.
a
Ack n ow led gm en t. We thank W. Beck, J . M. Groell,
(i) a , DIEA, TPTU in DMF, 3.5 h, 0 °C, room temp, 65%; (ii)
V. Huy Luu, and R. Wille for technical assistance.
4 N HCl in dioxane, 90 min, room temp, 100%; (iii) 3-phenoxyphen-
ylacetic acid, DIEA, TPTU in DMF, 16 h, room temp, 77%.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and analytical data for all intermediate and final
compounds and description of biochemical and cellular assays.
This material is available free of charge via the Internet at
http://pubs.acs.org.
procedures from 3,4,5-trimethoxybenzaldehyde,27,28 while
the 3,4-dimethoxy analogue, NR-Boc-3,4-L-dimethoxy-
phenylalanine, was from a commercial source. The
phenylalanine derivatives were coupled to a to provide
the Boc-protected dipeptides. After Boc deprotection
under acidic conditions, the N-terminal capping groups
designed by molecular modeling were incorporated into
the corresponding derivatives to afford the target com-
pounds.
Refer en ces
(1) Hochstrasser, M. Ubiquitin, proteasomes and the regulation of
intracellular protein degradation. Curr. Opin. Cell Biol. 1995,
7, 215-223.
(2) Hershko, A.; Ciechanover, The ubiquitin system. Annu. Rev.
Biochem. 1998, 67, 425-479.
(3) Palombella, V. J .; Rando, O. J .; Goldberg, A. L.; Maniatis, T.
The ubiquitin-proteasome pathway is required for processing
the NF-κ B1 precursor protein and the activation of NF-κ Β .
Cell 1994, 78, 773-785.
(4) Pagano, M. Regulation of the cell cycle regulatory proteins by
the ubuquitin pathway. FASEB J . 1997, 11, 1067-1075.
(5) Scheffner, M.; Huibregtse, J . M.; Vierstra, R. D.; Howley, P. M.
The HPV-16 E6 and E6-AP complex functions as a ubiquitin-
protein ligase in the ubiquitination of p53. Cell 1993, 75, 495-
505.
(6) Wojcik, C. Proteasomes in apoptosis: villains or guardians? Cell.
Mol. Life Sci. 1999, 56, 2903-2911.
(7) Maki, C. G.; Huibregtse, J . M.; Howley, P. M. In vivo ubiquity-
lation and proteasome-mediated degradation of p53(1). Cancer
Res. 1996, 56, 2649-2654.
(8) Pagano, M.; Tam, S. W.; Theodoras, A. M.; Beer-Romero, P.; Del
Sal, G.; Chau, V.; Yew, P. R.; Draetta, G. F.; Rolfe, M. A. Role of
the ubiquitin-proteasome pathway in regulating abundance of
the cyclin-dependent kinase inhibitor p27. Science 1995, 269,
682-685.
(9) Kisselev, A. F.; Goldberg, A. L. Proteasome inhibitors: from
research tools to drug candidates. Chem. Biol. 2001, 8, 739-
758.
The designed prototypes 3-5 were tested in biochemi-
cal assays that measure their ability to inhibit the three
types of hydrolytic activities of the human 20S protea-
some. Remarkably, potent and selective inhibition of the
proteasome chymotrypsin-like activity was observed. As
reported in Table 1, 3 turned out to inhibit the chymo-
trypsin-like activity with an IC50 of 15 nM while not
affecting the trypsin-like and post-glutamyl-peptide
hydrolytic activities at a concentration as high as 20 µM.
Furthermore, fully validating the design concept, the
structure-activity relationships observed in the 2-ami-
nobenzylstatine inhibitor class were mirrored in the
novel class, as judged from a comparison of the relative
inhibitory activities of 3-5. We had observed in the
2-aminobenzylstatine class the marked beneficial effect
of filling the AS2 pocket where the main interaction
established by the inhibitors, according to the model,
is aromatic stacking with residue Tyr 133 of subunit
HC5.23,24 Consistently, the phenoxy substituent target-
ing the AS2 pocket present on the N-terminal benzylic
group of 5 causes a 17-fold increase in potency compared
to 4 whose unsubstituted N-terminal group can only
interact with the AS1 pocket. Similarly, as previously
verified in the 2-aminobenzylstatine series, targeting a
residue of the bottom of the S3 pocket29 for hydrogen
bonding by adding a third methoxy substituent on the
central benzylic moiety is beneficial. Compound 3 is
significantly more potent (4-fold) than 5 in inhibiting
the chymotrypsin-like activity of the proteasome.
Thus, overall, the high potency and selectivity of 2
were matched by 3 in biochemical assays. However,
most importantly, in contrast to the 2-aminobenzylsta-
tine derivative, 3 also showed good activity in a cellular
setting. This compound blocks proteasome activity in
cultured cells (IC50 ) 20 nM) and inhibits in a dose-
dependent manner the proliferation of different tumor
(10) Adams, J . Proteasome inhibitors as therapeutic agents. Expert
Opin. Ther. Pat. 2003, 13, 45-57.
(11) Pajonk, F.; McBride, W. H. The proteasome in cancer biology
and treatment. Radiat. Res. 2001, 156, 447-459.
(12) Wojcik, C. Ubiquitin- and proteasome-dependent pathway of
protein degradation as an emerging therapeutic target. Expert
Opin. Ther. Targets 2000, 4, 89-111.
(13) Dou, Q. P.; Nam, S. Pharmacological proteasome inhibitors and
their therapeutic potential. Expert Opin. Ther. Pat. 2000, 10,
1263-1272.
(14) Murray, R. Z.; Norbery, C. Proteasome inhibitors as anti-cancer
agents. Anti-Cancer Drugs 2000, 11, 407-417.
(15) Adams, J . Preclinical and clinical evaluation of proteasome
inhibitor PS-341 for the treatment of cancer. Curr. Opin. Chem.
Biol. 2002, 6, 493-500.
(16) Delcros, J . G.; Floc’h, M. B.; Prigent, C.; Arlot-Bonnemains, Y.
Proteasome inhibitors as therapeutic agents: current and future
strategies. Curr. Med. Chem. 2003, 10, 479-503.
(17) Garcia-Echeverria, C. Recent advances in the identification and
development of 20S proteasome inhibitors. Mini-Rev. Med.
Chem. 2002, 2, 247-259.
(18) Myung, J .; Kim, K. B.; Crews, C. M. The ubiquitin-proteasome
pathway and proteasome inhibitors. Med. Res. Rev. 2001, 21,
245-273.
(19) Bochtler, M.; Ditzel, L.; Groll, M.; Hartmann, C.; Huber, R. The
proteasome. Annu. Rev. Biophys. Biomol. Struct. 1999, 28, 295-
317.

Letters
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 20 4813
(20) One of these covalent 20S proteasome inhibitors, bortezomib
(Millennium Pharmaceuticals, Inc.), was approved in May 2003
by the FDA for the treatment of patients with relapsed or
refractory multiple myeloma.
mization of 2-aminobenzylstatine derivatives: potent and selec-
tive inhibitors of the chymotrypsin-like activity of the human
20S proteasome. Bioorg. Med. Chem. Lett. 2002, 12, 1331-1334.
(25) For the nomenclature of the 20S proteasome subunits, see the
following. Groettrup, M.; Schmidtke, G. Selective proteasome
inhibitors: modulation of antigen presentation? Drug Discovery
Today 1999, 4, 63-71.
(21) Besides the two classes of compounds discussed in the present
work, only two compounds, the HIV inhibitor Ritonavir21a and
the natural product TMC-95A21b (with related analogues21c,d),
have been reported to inhibit the proteasome in a noncovalent
manner. (a) Schmidtke, G.; Holzhu¨tter, H.-G.; Bogyo, M.; Kairies,
N.; Groll, M.; De Giuli, R.; Emch, S.; Groettrup, M. How an
inhibitor of the HIV-I protease modulates proteasome activity.
J . Biol. Chem. 1999, 274, 35734-35740. (b) Groll, M.; Koguchi,
Y.; Huber, R.; Kohno, J . Crystal structure of the 20S protea-
some: TMC-95A complex: a non-covalent proteasome inhibitor.
J . Mol. Biol. 2001, 311, 543-548. (c) Kaiser, M.; Milbradt, A.
G.; Siciliano, C.; Assfalg-Machleidt, I.; Machleidt, W.; Groll, M.;
Renner, C.; Moroder, L. TMC-95A analogues with endocyclic
biphenyl ether group as proteasome inhibitors. Chem. Biodiver-
sity 2004, 1, 161-173. (d) Kaiser, M.; Groll, M.; Renner, C.;
Huber, R.; Moroder, L. The core structure of TMC-95A is a
promising lead for reversible proteasome inhibition. Angew.
Chem., Int. Ed. 2002, 41, 780-783.
(26) Groll, M.; Ditzel, L.; Lo¨we, J .; Stock, D.; Bochtler, M.; Bartunik,
H. D.; Huber, R. Structure of 20S proteasome from yeast at 2.4
Å resolution. Nature 1997, 386, 463-471.
(27) N-Acetyl-3,4,5-trimethoxy-L,D-phenylalanine was prepared from
commercially available 3,4,5-trimethoxybenzaldehyde and N-
acetylglycine according to a literature procedure (Oltz, E. M.;
Bruening, R. C.; Smith, M. J .; Kustin, K.; Nakanishi, K. The
Tunichromes. A Class of Reducing Blood Pigments from Sea
Squirts: Isolation, Structures, and Vanadium Chemistry. J . Am.
Chem. Soc. 1988, 110, 6162-6172). Alternatively, the N-acetyl-
3,4,5-trimethoxy-L,D-phenylalanine methyl ester was obtained
from 3,4,5-trimethoxybenzaldehyde and malonic acid dimethyl
ester following known protocols (see experimental section).
(28) The resolution of the racemic N-acetyl-3,4,5-trimethoxy-L,D-
phenylalanine methyl ester was performed by enzyme-catalyzed
hydrolysis of the L-ester using Alcalase (Novo Nordisk) as
described in the literature. Nestor, J . J ., J r.; Ho, T. L.; Simpson,
R. A.; Horner, B. L.; J ones, G. H.; McRae, G. I.; Vickery, B. H.
Synthesis and Biological Activity of Some Very Hydrophobic
Superagonist Analogues of Luteinizing Hormone-Releasing Hor-
mone. J . Med. Chem. 1982, 25, 795-801. Tyagi, O. D.; Boll, P.
M. Synthesis of (S)-3,4-dihydroxyphenylalanine (L-DOPA) and
unnatural R-amino acids via enzymatic resolution using alcalase.
Indian J . Chem. 1992, 851-854.
(22) Garcia-Echeverria, C.; Imbach, P.; France, D.; Fu¨rst, P.; Lang,
M.; Noorani, M.; Scholz, D.; Zimmermann, J .; Furet, P. A new
structural class of selective and non-covalent inhibitors of the
chymotrypsin-like activity of the 20S proteasome. Bioorg. Med.
Chem. Lett. 2001, 11, 1317-1319.
(23) Furet, P.; Imbach, P.; Fu¨rst, P.; Lang, M.; Noorani, M.; Zim-
mermann, J .; Garcia-Echeverria, C. Modeling of the binding
mode of a non-covalent inhibitor of the 20S proteasome. Ap-
plication to structure-based analogue design. Bioorg. Med. Chem.
Lett. 2001, 11, 1321-1324.
(29) This residue can be Tyr 135 or Ser 157 of subunit HC5. See ref
24.
(24) Furet, P.; Imbach, P.; Fuerst, P.; Lang, M.; Noorani, M.;
Zimmermann, J .; Garcia-Echeverria, C. Structure-based opti-
J M049660V