2,3-Dihydro-1,3-dioxo-1H-isoindole-5-carboxylic acid derivatives: A novel class of small molecule heparanase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2004.03.086

A novel class of 2,3-dihydro-1,3-dioxo-1H-isoindole-5-carboxylic acids are described as inhibitors of the endo-β-glucuronidase heparanase. Several of the compounds, for example, 2-[4-propylamino-5-[5-(4-chloro)phenyl-benzoxazol-2- yl]phenyl]-2,3-dihydro-1,3-dioxo-1H-isoindole-5-carboxylic acid (9c), display potent heparanase inhibitory activity (IC₅₀ 200-500nM) and have high selectivity (>100-fold) over human β-glucuronidase. They also show anti-angiogenic effects. Such compounds should serve as useful biological tools and may provide a basis for the design of novel therapeutic agents.

Full text of DOI:10.1016/j.bmcl.2004.03.086

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3269–3273
2,3-Dihydro-1,3-dioxo-1H-isoindole-5-carboxylic acid derivatives:
a novel class of small molecule heparanase inhibitors
Stephen M. Courtney,a, Philip A. Hay,a,* Richard T. Buck,a Claire S. Colville,a
David W. Porter,a,à David I. C. Scopes,a Faye C. Pollard,a Martin J. Page,b,§
JamesM. Bennett, b Margaret L. Hircock,b Edward A. McKenzie,b,– Colin R. Stubberfieldb
and Paul R. Turnerb
aDepartment of Medicinal Chemistry, Celltech R&D Ltd, 4-10 The Quadrant, Abingdon Science Park, Abingdon,
Oxfordshire, OX14 4YS, UK
bDepartment of Oncology, Celltech R&D Ltd, 4-10 The Quadrant, Abingdon Science Park, Abingdon,
Oxfordshire OX14 4YS, UK
Received 13 January 2004; revised 26 March 2004; accepted 29 March 2004
Abstract—A novel class of 2,3-dihydro-1,3-dioxo-1H-isoindole-5-carboxylic acids are described as inhibitors of the endo-b-glucu-
ronidase heparanase. Several of the compounds, for example, 2-[4-propylamino-5-[5-(4-chloro)phenyl-benzoxazol-2-yl]phenyl]-2,3-
dihydro-1,3-dioxo-1H-isoindole-5-carboxylic acid (9c), display potent heparanase inhibitory activity (IC50 200–500 nM) and have
high selectivity (>100-fold) over human b-glucuronidase. They also show anti-angiogenic effects. Such compounds should serve as
useful biological tools and may provide a basis for the design of novel therapeutic agents.
Ó 2004 Elsevier Ltd. All rights reserved.
Heparanase is an endo-b-glucuronidase that cleaves
heparan sulfate and is implicated in diverse physiologi-
cal and pathological processes.1;2 Expression of hepa-
II clinical trial,s hasasitsprincipal component the
highly sulfated pentasaccharide derivative (1). However,
heparanase inhibition is only one component of the
multifunctional profile of thiscompound, potentially
complicating an understanding of its mode of action.
Iminosugar related heparanase inhibitors of moderate
potency have also been described but showed greater
potency towardsmammalian b-glucuronidase and also
contain potentially labile functionality.9;10 Weak hepa-
ranase inhibitory activity has been reported for an aza-
pseudodisaccharide11 and for the natural product
trachyspic acid.12 We now report on a series of potent
and selective small molecule inhibitors, which may
provide useful biological tools for unravelling the com-
plex biology of heparanase and which may also serve as
basis for the design of novel therapeutic agents.
ranase has long been correlated with the metastatic
3–7
potential of tumour cellsand with angiogeneis.s
In
animal modelstreatment with heparanase inhibitorshas
been shown to markedly reduce the incidence of
metastasis.8 Heparanase thus offers an attractive drug
target but progress in this area has been limited by the
currently available repertoire of inhibitors. The most
advanced inhibitor isPI-88, 8 which iscurrently in Phase
Keywords: Heparanase; b-Glucuronidase; Angiogenesis; Metastasis;
Isoindole-5-carboxylic acid.
* Corresponding author. Tel.: +44-1235-544219; fax: +44-1235-5442-
24; e-mail: phil.hay@celltechgroup.com
Present address: Evotec OAI, 151 Milton Park, Abingdon, Oxford-
shire OX14 4SD, UK.
High throughput screening identified the 2,3-dihydro-
1,3-dioxo-1H-isoindole-5-carboxylic acid (2a) asa small
molecule inhibitor of heparanase (IC50 8 lM) with
modest anti-angiogenic activity (IC50 40 lM). The
molecular weight, lipophilicity and polar surface area of
this compound together with synthetic accessibility of
related structures were deemed acceptable as the basis
à Present address: Novartis Horsham Research Centre, Horsham,
West Sussex RH12 4AB, UK.
§ Present address: OSI Pharmaceuticals, Watlington Road, Oxford
OX4 6LT, UK.
– Present address: School of Biological Sciences, University of
Manchester, Manchester M13 9PT, UK.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.03.086

3270
S. M. Courtney et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3269–3273
OR
OR
OR
OR
O
O
RO
RO
O
RO
RO
O
RO
RO
O
Na2O3PO
RO
O
1
HO2C
N
O
RO
O
O
O
O
3
O
RO
RO
N
OR
R = SO3 Na or H
RO
(1)
(2a)
Table 1. In vitro inhibition of heparanase and anti-angiogenic activity
of (2a–h)
for a programme directed to identifying more potent
and selective heparanase inhibitors. Compounds were
evaluated for their heparanase inhibitory activity13 and
their ability to inhibit angiogenesis.14
R1
O
HO2C
N
O
O
N
Early investigations demonstrated that the carboxylic
acid group wascrucial for the inhibitory activity; its
removal or esterification abolished heparanase inhibi-
tory activity. Furthermore, a 1,4-disposition of the iso-
indole-5-carboxylic acid moiety and the pendant
benzoxazole ring system also resulted in a fall in inhib-
itory activity.
R2
Compound
R1
R2
Heparanase
inhibitiona
IC50 (lM)
Angiogenesis
inhibitiona
IC50 (lM)
2a
2b
2c
2d
2e
2f
H
H
Cl
8.0
1.5
40
ꢀ
H
H
Me
Ph
H
2.5
3.0
>50
5
Efforts were therefore focussed on substitution of both
the central aromatic core and the benzoxazole moiety of
(2a) while retaining the other structural features. Access
to such compounds was via the general route summar-
ised in Scheme 1. The m-nitrobenzoic acids( 3) were first
converted to the o-hydroxybenzamides, cyclisation of
which provided the intermediate benzoxazoles( 4).
Reduction of the nitro group followed by treatment with
1,3,5-benzenetricarboxylic anhydride gave the target
compounds( 2a–h). This short series of compounds,
which included 2-methoxy substitution in the central
phenyl ring led to the selection of the 5-phenyl derivative
(2h) (OGT2492) for further optimisation (Table 1). Al-
though thiscompound reprensted only a lisght
improvement in heparanase inhibition over (2a) it did
display much improved activity in the cell-based anti-
angiogenic assay. It also showed no inhibitory activity
against human b-glucuronidase at 100 lM. In a cor-
roborative assay, OGT2492 inhibited the cleavage of
heterologous FITC-labelled heparan sulfate (as assessed
by size exclusion HPLC) with an IC50 similar to that
obtained via the bFGF protocol.13
H
MeO
MeO
MeO
MeO
>10
3.0
NT
ꢀ
Cl
2g
2h
Me
Ph
>10
3.0
NT
2
ꢀPoor solubility precluded testing; NT: not tested.
a Mean of at least two determinations.
(IC50 0.4 lM) in heparanase inhibitory activity (Table
2). A loss in activity was observed with the introduction
of an ortho-methoxy group (6l) and with a more polar
(e.g., acetamide 6k) group. Translocating the methoxy
substituent on the central phenyl ring to the 4-position
had mixed effectson heparanaes inhibitory activity
depending upon the substitution pattern of the pendant
phenyl ring. Noteworthy wasthe methylenedioxy
derivative (6p) (IC50 0.2 lM), one of the most potent
inhibitors in this series that was also active in the angio-
genesis assay (IC50 3 lM). Although there wasa poor
correlation between anti-angiogenic activity and hepa-
ranase inhibitory activity, particularly striking was the
potent anti-angiogenic activity (IC50 0.75 lM) of the
30,50-difluoro derivative (6i). Monitoring activity against
human b-glucuronidase showed, inter alia, that (6p) and
(6q) showed no inhibition at 100 lM.
In order to explore the influence of substitution on the
pendant phenyl ring the synthetic route was slightly
modified (Scheme 2) to include a boronic acid coupling
step on the central 5-bromobenzoxazole intermediate
(5). Introduction of fluoro or CF3O substitution, for
example, (6b), (6h) and (6q), led to a significant increase
R1
R1
O
R1
O
O
a,b,c
d,e
O2N
N
N
N
HO2C
O2N
CO2H
O
R2
R2
(4)
(2a-h)
Scheme 1. Reagentsand conditions: (a) (COCl) 2, DMF, THF; (b) o-aminophenol, THF; (c) TsOH, toluene, reflux; (d) H2, Pd/C, EtOAc; (e) 1,3,5-
benzenetricarboxylic anhydride, AcOH, reflux.

S. M. Courtney et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3269–3273
3271
MeO
MeO
O
O
MeO
O2N
N
O
a,b,c
d,e,f
O2N
N
HO2C
O
N
CO2H
X
Br
Y
(6a-r)
(5)
Scheme 2. Reagentsand condition:s (a) (COCl) 2, DMF, THF; (b) o-aminophenol, THF; (c) TsOH, toluene, reflux; (d) ArB(OH)2, Pd(PPh3)4,
Na2CO3, dioxane, 100 °C; (e) H2, Pd/C, EtOAc; (f) 1,3,5-benzenetricarboxylic anhydride, AcOH, reflux.
Table 2. In vitro inhibition of heparanase and anti-angiogenic activity
of (6a–r)
systems was straightforward and based upon a range of
nucleophilic displacements on the core intermediate (7)
2
OMe
4
(Scheme 3), followed by arylboronic acid coupling.
From the derivativesevaluated (Table 3) certain SAR
becomes apparent: within a short series of amino sub-
stitutions NHPrn (8j) wasoptimal and notably the
NHMe (8h) waspoorly active. Furthermore, although
(8j) and (8k) demonstrated good heparanase inhibitory
activity, their anti-angiogenic activity fell away mark-
edly. N-Methylation of (8j) to give the corresponding
tertiary amine (8l) resulted in a >20-fold drop in activity.
Heparanase inhibitory activity for the short series of
ethersat the 4-poistion ( 8c–g), including those incor-
porating an additional heteroatom, wasreduced with
increasing bulk of the substituent.
O
N
HO2C
O
N
O
2' X
6'
3'
Y
4'
5'
Compound
R
Phenyl Heparanase Angiogenesis
substitution: inhibitiona inhibitiona
X, Y
IC50 (lM)
IC50 (lM)
6a
6b
6c
6d
6e
6f
2-MeO 30-F
1.0
0.4
>20
15
2-MeO 40-F
2-MeO 30-Cl
0.6
4.5
7
20
2-MeO 40-Cl
With the 4-NHPrn fixed, further optimisation was ex-
plored via introduction of substitutions in the pendant
phenyl ring (Table 4). The 40-F and 40-Cl substituents
gave potent heparanase inhibitors and, significantly,
most of these compounds showed appreciable activity in
the angiogenesis assay. Several of the 4-NHPrn deriva-
tiveswere tested for inhibition of human b -glucuroni-
dase with similar results to earlier analogues, that is (8j),
(9a) and (9c) showed no inhibition at 100 lM. Overall,
4-MeO 40-Cl
2.5
0.8
NT
12
2-MeO 20,40-di-F
2-MeO 30,40-di-F
2-MeO 30,50-di-F
4-MeO 30,50-di-F
2-MeO 30-CN
6g
6h
6i
0.6
0.4
20
>20
0.75
20
0.9
3.0
6j
6k
6l
2-MeO 30-NHCOMe
2-MeO 20-MeO
2-MeO 30-MeO
2-MeO 40-MeO
2-MeO 30,40-OCH2O
4-MeO 30,40-OCH2O
2-MeO 40-OCF3
4-MeO 40-OCF3
9.0
9.0
NT
NT
15
6m
6n
6o
6p
6q
6r
3.0
0.45
0.75
0.2
15
compoundsusch as(
9a) and (9c) displayed the best
>20
3
combination of in vitro activity and solubility.
0.4
20
NT
Some compounds reported in this study display a poor
correlation between heparanase activity and anti-
angiogenic activity. There are several possible reasons
for this: poor solubility of test compounds, high protein
binding, poor cellular uptake or an off-target component
to thisactivity. Certainly, the analoguesreported in
Table 4 showed improved solubility and this might ex-
plain their efficacy in the angiogenesis assay. In sum-
mary, the data presented here describes the first series of
potent, small molecule heparanase inhibitors that show
excellent selectivity over human b-glucuronidase. As
>10
NT: not tested.
a Mean of at least two determinations.
The limited solubility of several members of the com-
poundslisted in Table 2 led usto introduce potentially
solubilising substitutions, for example, O(CH2)2OMe,
O(CH2)2NMe2, NH(CH2)2OMe. Access to these
R
F
O
O
F
N
O
a,b,c
O2N
d,e,f,g
N
HO2C
O
N
O2N
CO2H
X
Br
Y
(7)
(8a-m), (9a-g)
Scheme 3. Reagentsand condition:s (a) (COCl) 2, DMF, THF; (b) o-aminophenol, THF; (c) TsOH, toluene, reflux; (d) NHRR0, NaH, THF;
(e) ArB(OH)2, Pd(PPh3)4, Na2CO3, dioxane, 100 °C; (f) H2, Pd/C, EtOAc; (g) 1,3,5-benzenetricarboxylic anhydride, AcOH, reflux.

3272
S. M. Courtney et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3269–3273
Table 3. In vitro inhibition of heparanase and anti-angiogenic activity of (8a–m)
R1
O
R2
N
HO2C
O
N
O
Compound
R1
R2
Heparanase inhibitiona
Angiogenesis inhibitiona
IC50 (lM)
IC50 (lM)
8a
8b
8c
8d
8e
8f
OEt
O(CH2)2OMe
H
H
3.0
4.0
>20
5
H
H
H
H
H
H
H
H
H
H
H
OEt
OPri
OBut
1.3
3.0
>20
10
>10
5.0
NT
10
O(CH2)2OMe
O(CH2)2NMe2
NHMe
8g
8h
8i
>10
>10
1.0
NT
NT
>20
15
NHEt
NHPrn
NHBun
NMePrn
NH(CH2)2OMe
8j
0.25
0.4
8k
8l
10
6.0
2.0
NT
20
8m
NT: not tested.
a Mean of at least two determinations.
2. Fairbanks, M. B.; Mildner, A. M.; Leone, J. W.; Cavey,
G. S.; Mathews, W. R.; Drong, R. F.; Slightom, J. L.;
Bienkowski, M. J.; Smith, C. W.; Bannow, C. A.;
Heinrikson, R. L. J. Biol. Chem. 1999, 274, 29587.
3. Vlodavsky, I.; Freidmann, Y.; Elkin, M.; Aingorn, H.;
Atzmon, R.; Ishai-Michaeli, R.; Bitan, M.; Pappo, O.;
Peretz, T.; Michal, I.; Spector, L.; Pecker, I. Nat. Med.
1999, 5, 793.
4. Hulett, M. D.; Freeman, C.; Hamdorf, B. J.; Baker, R. T.;
Harris, M. J.; Parish, C. R. Nat. Med. 1999, 5, 803.
5. Toyoshima, M.; Nakajima, M. J. Biol. Chem. 1999, 274,
24153.
6. Vlodavsky, I.; Goldshmidt, O.; Zcharia, E.; Metzger, S.;
Chajek-Shaul, T.; Atzmon, R.; Guatta-Rangini, Z.; Fre-
idmann, Y. Biochimie 2001, 83, 831.
7. Vlodavsky, I.; Freidmann, Y. J. Clin. Invest. 2001, 108,
341.
Table 4. In vitro inhibition of heparanase and anti-angiogenic activity
of (9a–g)
O
NHCH2CH2Me
N
HO2C
O
N
O
2 X
6
3
Y
4
5
Compound
Phenyl
substitution:
X, Y
Heparanase
inhibitiona
IC50 (lM)
Angiogenesis
inhibitiona
IC50 (lM)
9a
9b
9c
9d
9e
9f
40-F
0.5
2.5
0.2
3.0
2.0
2.0
0.8
0.25
0.75
1
30-Cl
40-Cl
30,50-di-F
30-Cl, 40-F
40-CF3
40-OCF3
NT
1
8. Parish, C. R.; Freeman, C.; Brown, K. J.; Francis, D. J.;
Cowden, W. B. Cancer Res. 1999, 59, 3433.
9. Kawase, Y.; Takahashi, M.; Takatsu, T.; Arai, M.;
Nakajima, M.; Tanzawa, K. J. Antibiot. 1996, 49, 61.
10. Nishimura, Y.; Shitara, E.; Adachi, H.; Toyoshima, M.;
Nakajima, M.; Okami, Y.; Takeuchi, T. J. Org. Chem.
2000, 65, 2.
0.2
0.2
9g
NT: not tested.
a Mean of at least two determinations.
11. Takahashi, S.; Kuzuhara, H.; Nakajima, M. Tetrahedron
2001, 57, 6915.
12. Shiozawa, H.; Takahashi, M.; Takatsu, T.; Kinoshita, T.;
Tanzawa, K.; Hosoya, T.; Furuya, K.; Takahashi, S.;
Furihata, K.; Seto, H. J. Antibiot. 1995, 48, 357.
13. Heparanase assay: The assay is based upon the use of the
specific binding of basic fibroblast growth factor (bFGF)
to heparan sulfate. Heparan sulfate can be detected via
binding of bFGF using a horse radish peroxidase-conju-
gated bFGF antibody. Following cleavage of high molec-
ular weight heparan sulfate by heparanase, the smaller
material generated will no longer adhere to the surface of a
96 well plate and hence heparanase activity can be
followed asa reduction in bFGF binding. Nunc Maxisorp
96 well platesare coated for 16 h at rt with 100 lL/well
such they should serve as useful biological tools and
may provide a basis for the design of novel therapeutic
agents.15
References and notes
1. Parish, C. R.; Freeman, C.; Hulett, M. D. Biochim.
Biophys. Acta 2001, 1471, M99.

S. M. Courtney et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3269–3273
3273
0.04 mg/mL heparan sulfate in PBS. The wells are then
aspirated and blocked for 1 h with 200 lL/well 1% BSA–
PBS. Following five washes with 0.01% BSA, 0.05%
Tween20 PBS (wash buffer), 100 lL of recombinant
human basic FGF (90 ng/mL in 0.1% BSA/PBS) is added
per well and the plate isincubated at room temperature
for 1 h. After a further five washes with the wash buffer,
10 lL of test compound (in 10% DMSO) and 90 lL of
human heparanase in 100 mM sodium acetate, 5 mM
CaCl2, pH 5.5 are added to each well and the plate
incubated for 2 h at 37 °C. The wellsare washed again with
wash buffer and 100 lL of bFGF antibody-horse radish
peroxidase conjugate added. The plate is incubated at
room temperature for 1 h and washed again five times with
wash buffer. TMB peroxidase substrate (100 lL) isadded
and the colour allowed to develop for 10 min. The reaction
isstopped with 50 lL 1 M H2SO4 and the colour read at
450 nm on a plate reader.
Willebrand’sFactor, Platelet Endothelial Cell Adheison
Molecule-1 (PECAM-1 or CD31) and Intercellular Adhe-
sion Molecule-2 (ICAM-2). The assay was supplied as
growing cultures at the earliest stage of tubule formation
in a 24 well plate format. It is designed so that test
compoundsand conditioned media can be added to the
cultureswithin individual wellsand the resulting effect on
tubule formation can then be monitored. Positive and
negative test agents, for example, vascular endothelial
growth factor (VEGF) and suramin, are included. On day
1, fresh growth medium, medium plus control agent or
medium plustest compound wasadded to the cellsand the
cultureswere incubated at 37 °C, 5% CO2. Test com-
pounds were dissolved in DMSO and the final concentra-
tion of DMSO in the medium did not exceed 0.1% (v/v).
The specified medium was changed at days 4, 7 and 9 and
the cellswere monitored for growth. On day 11, the cells
were washed with Dulbecco’s phosphate-buffered saline
(PBS) and fixed using 70% ethanol ()20 °C) for 30 min at
room temperature. After fixing, the cellswere washed and
treated with blocking buffer, 1% BSA in PBS. The cells
were stained for PECAM-1 on the same day, following
standard immunohistochemistry procedures using mouse
anti-human CD31 asthe primary antibody and a goat
anti-mouse IgG alkaline phosphate conjugate. Tubule
formation was quantitatively assessed by measuring PE-
CAM-1 positive staining using the image analysis program
‘Matrox inspector’ to evaluate the percentage tubule
staining relative to an untreated control.
14. Angiogenesis assay: A commercial angiogenesis assay for
analysing the angiogenic or anti-angiogenic properties of
test compounds (AngioKit catalogue no ZHA-1000, TCS
CellWorks Ltd, Buckingham, UK) was used. In this assay,
human endothelial cellswere co-cultured with other
human cells in a specifically designed medium. The
endothelial cells initially form small islands within the
culture matrix. They subsequently proliferate and then
enter a migratory phase during, which they move through
the matrix to form threadlike tubule structures. These
gradually join up (by 12–14 days) to form networks of
anatomising tubules, which closely resemble a capillary
bed structure. These tubules stain positive for von
15. All final productswere characteriesd by 1H NMR and
MS.