Isoindolinone-based inhibitors of the MDM2-p53 protein-protein interaction

Article abstract of DOI:10.1016/j.bmcl.2004.12.061

A series of 2-N-alkyl-3-aryl-3-alkoxyisoindolinones has been synthesised and evaluated as inhibitors of the MDM2-p53 interaction. The most potent compound, 3-(4-chlorophenyl)-3-(4-hydroxy-3,5-dimethoxybenzyloxy)-2-propyl-2,3- dihydroisoindol-1-one (NU8231), exhibited an IC₅₀ of 5.3 ± 0.9 μM in an ELISA assay, and induced p53-dependent gene transcription in a dose-dependent manner, in the SJSA human sarcoma cell line.

Full text of DOI:10.1016/j.bmcl.2004.12.061

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1515–1520
Isoindolinone-based inhibitors of the MDM2–p53
protein–protein interaction
Ian R. Hardcastle,a,* Shafiq U. Ahmed,b Helen Atkins,b A. Hilary Calvert,b
Nicola J. Curtin,b Gillian Farnie,b Bernard T. Golding,a Roger J. Griffin,a
Sabrina Guyenne,a Claire Hutton,b Per Ka¨llblad,c Stuart J. Kemp,a Martin S. Kitching,a
David R. Newell,b Stefano Norbedo,a Julian S. Northen,a Rebecca J. Reid,a
K. Saravanan,a Henriette M. G. Willemsc and John Lunecb,*
¨
aNorthern Institute for Cancer Research, School of Natural Sciences—Chemistry, Bedson Building,
University of Newcastle upon Tyne, Newcastle upon Tyne, NE1 7RU, UK
bNorthern Institute for Cancer Research, Paul O’Gorman Building, Medical School, Framlington Place,
University of Newcastle upon Tyne, Newcastle upon Tyne, NE2 4HH, UK
cDe Novo Pharmaceuticals, Compass House, Vision Park, Histon, Cambridge CB4 9ZR, UK
Received 22 November 2004; revised 17 December 2004; accepted 21 December 2004
Available online 30 January 2005
Abstract—A series of 2-N-alkyl-3-aryl-3-alkoxyisoindolinones has been synthesised and evaluated as inhibitors of the MDM2–p53
interaction. The most potent compound, 3-(4-chlorophenyl)-3-(4-hydroxy-3,5-dimethoxybenzyloxy)-2-propyl-2,3-dihydroisoindol-
1-one (NU8231), exhibited an IC50 of 5.3 0.9 lM in an ELISA assay, and induced p53-dependent gene transcription in a dose-
dependent manner, in the SJSA human sarcoma cell line.
ꢀ 2005 Elsevier Ltd. All rights reserved.
The p53 tumour suppressor acts as a Ôguardian of the
genomeÕ by reacting to cellular stress, such as hypoxia
and DNA damage. Such stresses increase cellular levels
of p53 and activate its transcriptional function to drive
the expression of a number of genes that govern progres-
sion through the cell cycle, the initiation of DNA repair,
and programmed cell death.1,2 The activity of p53 is
tightly regulated by the MDM2 protein, the gene for
which is itself regulated by p53. MDM2 binds to the
p53 transactivation domain, blocking the transcriptional
activity of p53 and ubiquitylating the MDM2–p53 com-
plex to target it for proteosomal destruction. In normal
cells the balance between active p53 and inactive
MDM2-bound p53 is maintained by this negative feed-
back loop.3,4 The X-ray crystal structure of MDM2
bound to a p53 peptide corresponding to the transacti-
vation loop, reveals a hydrophobic pocket on the sur-
face of MDM2, into which the Phe19, Trp23 and
Leu26 residues of p53 bind.5 Inactivation of p53
through mutation is frequently found in a wide range
of sporadic cancers. However, around 7% of human tu-
mours show evidence of amplification and overexpres-
sion of the MDM2 gene resulting in suppression of
functional p53, promoting transformation and uncon-
trolled tumour growth.6 Inhibitors of the MDM2–p53
binding interaction would be expected to restore normal
p53 activity in MDM2 overexpressing cells and thus ex-
ert an anti-tumour effect.7 A number of inhibitors of the
MDM2–p53 interaction have been reported including
potent peptide inhibitors,8 the natural product chloro-
fusin,9 and small molecules including the recently
described ÔnutlinsÕ.10,11
Here we describe inhibitors of the MDM2–p53 interac-
tion, based on an isoindolinone scaffold. Preliminary
screening studies, using an in vitro p53–MDM2 binding
assay, identified compounds 1a and 2a,b as modest
inhibitors of the p53–MDM2 interaction (IC50
Keywords: Cancer; MDM2; p53; Protein–protein interactions.
*
Corresponding authors. Tel.: +44 191 222 6645; fax: +44 191 222
8591 (I.R.H.); tel.: +44 191 246 4420 (J.L.); e-mail addresses:
i.r.hardcastle@ncl.ac.uk; john.lunec@ncl.ac.uk
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.12.061

1516
I. R. Hardcastle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1515–1520
ꢀ 200 lM). These compounds also displayed growth
inhibitory activity in the NCI 60 cell-line screen and,
importantly, were COMPARE negative with respect to
any known class of anti-tumour agents.12 In this paper,
we report a programme of focused library synthesis,
incorporating in silico ligand design, resulting in the dis-
covery of novel inhibitors of the MDM2–p53
interaction.
(1YCR)5 using easyDock.13 A single plausible binding
mode was chosen for each compound, and the position
of the isoindolinone scaffold was preserved during the
following virtual screen of new substituents. The binding
interaction between ligand and receptor was explored
using a simulated annealing optimisation of an empirical
free-energy function using the program Skelgen.14 Re-
agents able to form at least one additional H-bond with
target residues of MDM2 were selected as Ôvirtual hitsÕ
and suggested for synthesis, a selection of which were
synthesised. Comparison of the inhibitory activity of
these synthesised Ôvirtual hitsÕ with a set of isoindoli-
nones bearing randomly selected substituents revealed
no significant difference in activity between the two sets
(data not shown).
R2
R2
HN
O
Ar
R1
Ar
R1
N
N
O
O
1a: R1 = CH2Ph; R2 = n-Pr; Ar = Ph 2a: R1 = CH2Ph; R2 = n-Pr; Ar = Ph
2b: R1 = R2 = n-Pr; Ar = Ph
At this point, the binding mode determination was revis-
ited with the six most active compounds synthesised
(Table 1). The six lead compounds were docked into
the MDM2 crystal structure 1YCR, using the programs
easyDock13 and GOLD.15 It was impossible to distin-
guish a single preferred binding mode from the large
pool of docking solutions. Previously, it has been shown
that the probability of predicting the experimental bind-
ing mode correctly increases significantly when multiple
binding modes are considered.16 Therefore, a total of 24
(6 per compound · 2 per stereoisomer · 2 per docking
program) high scoring, unique binding modes were se-
lected as starting points for a second round of virtual
screening. Again, reagents able to form additional
hydrogen bonds with the protein were suggested for
synthesis.
Using the published structure of the MDM2–p53 bind-
ing site,5 we have employed computational methods,
and focussed library synthesis based on the isoindolinone
template, to develop compounds with improved inhibi-
tory activity. These studies have resulted in the identifica-
tion of a number of inhibitors (Table 1) with increased
potency over the preliminary compounds (1a and 2a,b).
Briefly, the determination of single, low-energy binding
modes for 1a and 2b was attempted by docking the com-
pounds into the published crystal structure of MDM2
Table 1. Seed compounds used in the second round of binding mode determinations
R2
X
Ar
N
R1
O
X
Ar
R1
R2
IC50 (lM)
O
NH
n-Pr
27
3
OSEM
O
NH
O
n-Pr
66
8
MeO
OH
Ph
Ph
n-Pr
90 39
O
HO
NH
85
OH
O
Ph
Ph
92 11
HO
NH
70
8
OMe

I. R. Hardcastle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1515–1520
1517
In order to validate this approach 57 hit compounds
were selected, including substituents unique to each
binding mode. The majority of these compounds were
synthesised and assayed for inhibition of MDM2–p53
binding using the ELISA format assay. A number of
compounds that displayed improved activity were iden-
tified including 2e,f and 2g (Table 2).
enhanced chemiluminescence (ECLTM, Amersham Bio-
sciences) using the oxidation of the diacylhydrazide sub-
strate, luminol, to generate a quantifiable light signal.
The luminol substrate together with enhancer was auto-
matically injected into each well and the relative lumi-
nescence units (RLU) measured over a 30 s interval
using a Berthold MicroLumat-Plus LB 96V microplate
luminometer. The percentage MDM2 inhibition at a gi-
ven concentration was calculated as the (RLU detected
in the compound treated sample Ä RLU of DMSO con-
trols) · 100. The IC50 was calculated using a plot of
%MDM2 inhibition versus concentration and is the
average of three independent experiments. The results
are presented in Table 2.
Consideration of these results and others from addi-
tional isoindolinones synthesised as part of random
libraries, enabled the design of a combinatorial array
of compounds bearing the N-2 and C-3 substitutents
that appeared to confer improved activity. The substitu-
ents chosen were: Ar = phenyl, 4-(2-trimethylsilylet-
hoxy)- methoxyphenyl and 4-chlorophenyl; R1 =
n-propyl, benz- yl and 2-acetamidoethyl; R2 = 4-t-
butylbenzyloxy, 3,5-dimethoxy-4-hydroxybenzyloxy, 2-
(2-pyridyl)ethoxy and 3-hydroxypropoxy.
In comparison with the lead compounds bearing an
unsubstituted phenyl group at the C3 position (2d,e
and g), none of the newly synthesised compounds dis-
played improved potency, with only the syringic alcohol
derivative 2k and the 2-(2-pyridyl)ethoxy derivative 2m
showing inhibition comparable with the 4-t-butylbenzyl-
oxy derivative 2d. In the 3-(4-chlorophenyl) series, the
N-propyl substituted 3-hydroxypropoxy derivative 2r
was equipotent with the lead N-benzyl compound 2c.
In contrast, the N-propyl substituted 2-(2-pyridyl)-
ethoxy derivative 2s was significantly less potent than
the lead N-benzyl compound 2f. The reverse trend was
observed for the N-propyl syringic alcohol derivative
2q, which was significantly more potent than the N-pro-
pyl derivative 2o. Interestingly, for the syringic alcohol
derivatives, the 4-chloro substituent was favourable in
the N-propyl series (2k and 2q) but resulted in a loss
of potency in the N-benzyl series (2i and 2q). In the N-
ethylacetamido series, the 3-(4-chlorophenyl) derivatives
(2v and 2w) were significantly less potent than the lead
2g. In the light of these disappointing results, and the
difficulties encounted with the synthesis of these deriva-
tives, this series was abandoned. Similarly, in the 4-(2-
trimethylsilylethoxy)methoxyphenyl series, none of the
compounds synthesised (2v–z) displayed improved
potency compared with the lead 1b and the series was
not completed.
Compounds 2a–z were prepared according to Methods
A, B and C (Scheme 1). The appropriate benzoylbenzoic
acids (3) were converted into the w-acid chlorides (4),
under Vilsmeier conditions, and then condensed with
the R1-primary amine to give the 4-hydroxyisoindol-
inone (5). Compound 5 was converted into the chloride
and subsequently reacted with R2-alcohol in the pres-
ence of base (Et3N or K2CO3) to give 2d,e,g–m,o–r
and 2u,v (Method A).17 Alternatively, compound 5
was converted into the unstable chloride and trapped
with benzylmercaptan to give the stable thioether (6).
This was activated to nucleophilic displacement on
treatment with N-iodosuccinimide (NIS) in the presence
of catalytic camphorsulfonic acid (CSA) and reacted in
situ with the appropriate R2-alcohol to give 2a,f,n,s
and t. Compounds 2w–z were prepared according to
Method C (Scheme 1). Directed-ortho-metallation of n-
propylbenzamide (7 R1 = nPr) and reaction with the
appropriate benzoate ester afforded the hydroxyisoind-
olinone 5, which was converted into the target isoindoli-
none as for Method A. The final compounds 2h–z were
isolated and tested as racemic mixtures.18
Compounds were assayed for inhibition of the MDM2–
p53 interaction using a 96-well plate binding assay (ELI-
SA) with a luminometric detection end-point. Briefly,
96-well plates were coated with streptavidin followed
by biotinylated IP3 peptide (b-IP3: Biotin-Met-
Pro-Arg-Phe[19]-Met-Asp-Tyr-Trp-Glu-Gly-Leu[26]-Asn-
NH2).19 Control experiments consisted of both 5%
DMSO carrier alone as a negative control and 100 nM
active peptide (AP-B: Ac-Phe[19]-Met-Aib-Pmp-6-Cl-
Trp-Glu-Ac3-Leu[26]-NH2) as a positive control peptide
The increased potency observed for the 4-chlorophenyl
compound 2q, is consistent with the predicted binding
mode for the parent 2g, which is seen bound to
MDM2 with the phenyl ring occupying the tryptophan
binding pocket, the N-propylisoindolinone in contact
with a broad, shallow, hydrophobic cleft and the pheno-
lic OH of the syringic alcohol making an H-bond to the
backbone of Tyr100 on MDM2 (Fig. 1). The impor-
tance of the tryptophan binding pocket for affinity has
been demonstrated previously by the potent activity of
the AP peptide,8 and the nutlin series.10,20 Experiments
to confirm the binding mode of 2q are ongoing.
antagonist of the MDM2–p53 interaction (IC50
=
5 nM).8 Compounds and controls were pre-incubated
at 20 ꢁC for 20 min with MDM2, before transfer of
the MDM2-compound mixture to the b-IP3 streptavidin
plates and incubation at 4 ꢁC for 90 min. After washing
to remove unbound MDM2, each well was incubated at
20 ꢁC with a buffered solution of primary anti-MDM2
antibody (Ab-5, Calbiochem), then washed and incu-
bated at 20 ꢁC with a solution of secondary horseradish
peroxidase (HRP) conjugated antibody (Dako), and
washed again. The HRP activity was measured by
The most potent compound identified, 2q (NU8231;
IC50 = 5.3 0.9 lM), was selected for further evalua-
tion. SJSA cells, in which the MDM2 gene is amplified,
were treated with increasing concentrations (5, 10 and
20 lM) of 2q. Cells were lysed at 6 h and protein ex-
tracts analysed by Western immunoblotting for p53,
p21WAF1 and actin (Fig. 2). There was a dose-dependent

1518
I. R. Hardcastle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1515–1520
Table 2. Inhibition of the MDM2–p53 binding interaction by isoindolinones
Compound
Method
B
Ar
R1
R2
IC50 (lM)
15.9 0.8
2c
Cl
OH
2d
2e
2f
A
A
B
Ph
Ph
92 11
14 0.3
H
N
O
N
26.2 4.2
Cl
OMe
OH
2g
2h
2i
A
A
A
A
A
A
A
B
Ph
Ph
Ph
Ph
Ph
Ph
Ph
17.9 0.3
245
OMe
OH
N
206 30
>500
2j
n-Pr
n-Pr
n-Pr
n-Pr
OMe
OH
2k
2l
82
8
OMe
>500
OH
N
2m
2n
2o
2p
2q
2r
2s
2t
100 14
99 18
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
OMe
OH
A
A
A
A
B
42
8
OMe
n-Pr
n-Pr
n-Pr
n-Pr
187 38
5.3 0.9
16.4 1.6
OMe
OH
OMe
OH
N
57
6
H
N
O
O
B
91.4 0.4
OMe
OH
H
N
2u
A
76
4
OMe
OMe
OH
2v
A
257 34
OSEM
OMe

I. R. Hardcastle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1515–1520
1519
Table 2 (continued)
Compound
2w
Method
C
Ar
R1
R2
IC50 (lM)
n-Pr
464 31
OSEM
OSEM
OMe
OH
2x
C
n-Pr
118 24
OMe
2y
2z
C
C
n-Pr
n-Pr
476 24
312 22
OSEM
OSEM
OH
N
Method A
R2
O
Cl
O
HO
Ar
Ar
N R1
Ar
N R1
a
c
b
Ar
COOH
O
O
O
O
3
4
2
5
e
g
d
Method C
Method B
H
O
Cl
O
N
R1
S
Ar
N R1
f
O
7
6
Scheme 1. Reagents and conditions: Method A: (a) SOCl2, cat DMF, THF; (b) R1NH2, THF; (c) (i) SOCl2, cat DMF, THF; (ii) R2OH, THF, Et3N
or K2CO3. Method B: (d) (i) SOCl2, cat DMF, THF; (ii) PhCH2SH, THF; (e) NIS, cat CSA, THF, R2OH. Method C: (f) R1–NH2, (g) s-BuLi,
ArCOOEt, THF.
Figure 1. A—Model of the low energy binding mode of 2g bound to MDM2. B—Pharmacophore model of 2q bound to MDM2.
increase in MDM2 and p21, consistent with p53 activa-
tion. No change was observed for p53 levels or the actin
controls.
tein–protein binding interaction. The most potent com-
pound 2q has an IC50 of 5.3 0.9 lM in a cell-free
binding assay (ELISA) and shows dose-dependent
induction of MDM2 and p21 when used to treat an in-
tact MDM2 amplified human sarcoma cell line. Further
development of these compounds is in progress.
In summary, we have discovered interesting structurally-
novel isoindolinone antagonists of the MDM2–p53 pro-

1520
I. R. Hardcastle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1515–1520
5. Kussie, P. H.; Gorina, S.; Marechal, V.; Elenbaas, B.;
Moreau, J.; Levine, A. J.; Pavletich, N. P. Science 1996,
274, 948–953.
6. Oliner, J. D.; Kinzler, K. W.; Meltzer, P. S.; George, D.
L.; Vogelstein, B. Nature 1992, 358, 80–83.
7. Chene, P. Nat. Rev. Cancer 2003, 3, 102–109.
8. Garcia-Echeverria, C.; Chene, P.; Blommers, M. J. J.;
Furet, P. J. Med. Chem. 2000, 43, 3205–3208.
9. Duncan, S. J.; Gruschow, S.; Williams, D. H.; McNicolas,
C.; Purewal, R.; Hajek, M.; Gerlitz, M.; Martin, S.;
Wrigley, S. K.; Moore, M. J. Am. Chem. Soc. 2001, 123,
554–560.
10. Vassilev, L. T.; Vu, B. T.; Graves, B.; Carvajal, D.;
Podlaski, F.; Filipovic, Z.; Kong, N.; Kammlott, U.;
Lukacs, C.; Klein, C.; Fotouhi, N.; Liu, E. A. Science
2004, 303, 844–848.
11. Zhao, J. H.; Wang, M. J.; Chen, J.; Luo, A. P.; Wang, X.
Q.; Wu, M.; Yin, D. L.; Liu, Z. H. Cancer Lett. 2002, 183,
69–77.
Figure 2. Western blot from SJSA cells treated with 2q (U: untreated;
D: DMSO control; 5, 10 and 20 lM compound 2q).
12. Paull, K. D.; Shoemaker, R. H.; Hodes, L.; Monks, A.;
Scudiero, D. A.; Rubinstein, L.; Plowman, J.; Boyd, M. R.
J.N.C.I. 1989, 81, 1088–1092.
Acknowledgements
13. Mancera, R. L.; Ka¨llblad, P.; Todorov, N. P. J. Comput.
Chem. 2004, 25, 858–864.
14. Stahl, M.; Todorov, N. P.; James, T.; Mauser, H.; Boehm,
H. J.; Dean, P. M. J. Comput. Aided Mol. Des. 2002, 16,
459–478.
The authors thank Cancer Research UK, EPSRC and
BBSRC for financial support. The authors are grateful
to Dr Edward Sausville and Dr. Robert Schultz of the
DTP, National Cancer Institute, Washington, USA,
for facilitating the analysis of compounds in the NCI
cell line panel.
15. Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor,
R. J. Mol. Biol. 1997, 267, 727–748.
16. Ka¨llblad, P.; Mancera, R. L.; Todorov, N. P. J. Med.
Chem. 2004, 47, 3334–3337.
References and notes
17. Kitching, M. S.; Clegg, W.; Elsegood, M. R. J.; Griffin, R.
J.; Golding, B. T. Synlett 1999, 997–999.
1
18. All compounds gave satisfactory H and 13C NMR, IR,
LCMS and HRMS or combustion analysis.
1. Lane, D. P. Nature 1992, 358, 15–16.
2. Vousden, K. H.; Lu, X. Nat. Rev. Cancer 2002, 2, 594–
604.
3. Momand, J.; Zambetti, G. P.; Olson, D. C.; George, D.;
Levine, A. Cell 1992, 69, 1237–1245.
19. Bottger, V.; Bottger, A.; Howard, S. F.; Picksley, S. M.;
Chene, P.; Garcia-Echeverria, C.; Hochkeppel, H. K.;
Lane, D. P. Oncogene 1996, 13, 2141–2147.
20. Fry, D. C.; Emerson, S. D.; Palmeb, S.; Vua, B. T.; Liua,
C.-M.; Podlaskia, F. J. Biomol. NMR 2004, 30, 163–173.
4. Fuchs, S. Y.; Adler, V.; Buschmann, T.; Wu, X. W.;
Ronai, Z. Oncogene 1998, 17, 2543–2547.