Discovery of a Dihydroisoquinolinone Derivative (NVP-CGM097): A Highly Potent and Selective MDM2 Inhibitor Undergoing Phase 1 Clinical Trials in p53wt Tumors

Article abstract of DOI:10.1021/acs.jmedchem.5b00810

(Figure Presented). As a result of our efforts to discover novel p53:MDM2 protein-protein interaction inhibitors useful for treating cancer, the potent and selective MDM2 inhibitor NVP-CGM097 (1) with an excellent in vivo profile was selected as a clinical candidate and is currently in phase 1 clinical development. This article provides an overview of the discovery of this new clinical p53:MDM2 inhibitor. The following aspects are addressed: mechanism of action, scientific rationale, binding mode, medicinal chemistry, pharmacokinetic and pharmacodynamic properties, and in vivo pharmacology/toxicology in preclinical species.

Full text of DOI:10.1021/acs.jmedchem.5b00810

Page 1 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of a dihydroisoquinolinone derivative
(NVPꢀCGM097) – a highly potent and selective
MDM2 inhibitor undergoing phase 1 clinical trials in
p53wt tumors
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Philipp Holzer*, Keiichi Masuya†, Pascal Furet, Joerg Kallen, Therese Valat-Stachyra,
Stéphane Ferretti, Joerg Berghausen, Michèle Bouisset-Leonard, Nicole Buschmann, Carole
Pissot-Soldermann, Caroline Rynn, Stephan Ruetz, Stefan Stutz, Patrick Chène, Sébastien Jeay,
Francois Gessier
Novartis Institutes for BioMedical Research, 4002 Basel, Switzerland
KEYWORDS PPI, proteinꢀprotein interaction inhibitor, NVPꢀCGM097, p53, MDM2, HDM2,
clinical trial
ABSTRACT As a result of our efforts to discover novel p53ꢀMDM2 proteinꢀprotein interaction
inhibitors useful for treating cancer, the potent and selective MDM2 inhibitor NVPꢀCGM097 (1)
with an excellent in vivo profile was selected as a clinical candidate and is currently in phase 1
clinical development. This article provides an overview of the discovery of this new clinical p53ꢀ
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 40
1
2
3
4
5
6
7
8
9
MDM2 inhibitor. The following aspects are addressed: mechanism of action, scientific rationale,
binding mode, medicinal chemistry, pharmacokinetic and pharmacodynamic properties and in
vivo pharmacology/toxicology in preclinical species.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Introduction
The p53 protein, also known as the “guardian of the human genome”,1 is a transcription factor
that controls the cellular response to DNA damage or other stress stimuli through the induction
of cellꢀcycle arrest, DNA repair, apoptosis or senescence.2 Significant drug discovery efforts
have been directed to block the proteinꢀprotein interaction between p53 and MDM2 (also known
as HDM2, mouse double minute 2 homolog), which is an E3 ubiquitin ligase that targets p53 for
degradation by the proteasome.3 Mice lacking p53 develop normally but are prone to the
development of a variety of tumors.4 TP53, the gene encoding the p53 protein, is mutated or
deleted in nearly 50% of human cancers, rendering p53 nonfunctional as a tumor suppressor.5
Although TP53 retains wildꢀtype status in the remaining 50% of human cancers, its function is
effectively regulated by a variety of complex mechanisms.6 In tumor cells p53 is frequently
inactivated, often via overꢀexpression of MDM2 that binds p53 with high affinity, thereby
promoting its degradation as well as masking its ability to activate transcription. Whereas p53
activation underlies the activity of many DNAꢀdamaging chemotherapeutics in response to
cellular stresses, a more attractive therapeutic strategy would be to competitively bind MDM2
and restore p53 levels to a threshold to induce apoptosis.7
Proteinꢀprotein interactions (PPIs) often span large surface areas (>1600 Å2) and typically lack
deep, wellꢀdefined binding pockets. When Kussie et al. published the first Xꢀray structure of the
ACS Paragon Plus Environment
2

Page 3 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Nꢀterminal p53[17ꢀ125] binding domain of MDM2 and its analog from Xenopus laevis (71
percent identity to human MDM2) bound to a 15ꢀresidue peptide analog of p53 in 1996, first
insights into the structural features and the amino acid side chains involved in the intracellular
PPI were obtained.8 The crystal structure showed that the interface relies on the steric
complementarity between the MDM2 cleft and the hydrophobic face of the αꢀhelix of p53, and
revealed the particular importance of a triad of p53 amino acids. The sideꢀchains of Pheꢀ19, Trpꢀ
23 and Leuꢀ26 have extended conformations and pack closely to each other, making sequential
and extensive van der Waals contacts with the surface of MDM2. These observations led
Novartis scientists to publish pioneering work in 2000, where the critical roles of these residues
in molecular recognition were demonstrated in experiments performed with phage display
peptide libraries.9 In this work peptideꢀscanning investigations identified an 8ꢀmer peptide with a
key 6ꢀchlorotryptophan residue which gave a 63ꢀfold enhancement in activity relative to the
corresponding tryptophan analog and inhibited fullꢀlength p53 binding to Glutathione Sꢀ
transferases (GST) tagged MDM2 with an IC50 value of 5 nM. Although the initial interest of
polypeptides to target p53/MDM2 was limited due to their poor membrane permeability and
physiological stability, the pharmacophore model that emerged from this work propelled the
identification of both peptideꢀlike and nonꢀpeptidic inhibitors for this target.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Parallel to the p53/MDM2 PPI disruptors obtained by screening proprietary compound
collections,10 in silico approaches have also been used by several groups to assist in the search
for hits for this target. Using our acquired internal knowledge, the “central valine concept” was
established using molecular modeling to orient in 3D the key pharmacophores via a central
connecting fragment. This led to the discovery of two new p53ꢀMDM2 inhibitor scaffolds based
on a 3ꢀimidazolyl indole11 and a tetrasubstituted imidazole core structure.12 Parallel to these
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 40
1
2
3
4
5
6
7
8
9
efforts and seeking to extend chemical space, a large scale knowledgeꢀbased virtual screening
exercise was undertaken.13 Medicinal chemistry exploration of one of the virtual screening hits
allowed us to identify dihydroisoquinolinone derivatives as an additional new class of p53ꢀ
MDM2 inhibitors.14 Optimization of this new series of inhibitors has culminated in compound 1
(NVPꢀCGM097), a compound currently being evaluated in phase 1 clinical trials.15,16 Several
other small molecule MDM2 antagonists have also recently progressed into clinical trials and a
compilation of known chemical matter has recently been published.17 In this paper, we provide
an overview of the scientific efforts resulting in the discovery of 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Development of the dihydroisoquinolinone series using X-ray structural information14,18
Starting from a virtual screening hit, we identified a new class of inhibitors of the p53ꢀMDM2
interaction having an isoquinolinone core structure.14 An initial optimization effort guided by
crystal structure information afforded compound 2, a potent MDM2 inhibitor with an IC50 of 8
nM in the TRꢀFRET biochemical assay (Figure 1 and Table 1).19 Compound 2 demonstrated
significant activity in the SJSAꢀ1 cell proliferation assay (IC50 of 3.86 ꢀM) and was therefore an
interesting starting point for further development. We succeeded in rapidly generating an Xꢀray
coꢀcrystal structure of 2 with MDM2 (Figure 1), which was key to guiding our optimization
strategy. It confirmed the binding mode for this class of compounds and the preference of the C1
stereo center, bearing the paraꢀchlorophenyl group, for the (S) configuration. This observation
aligned well with data from the TRꢀFRET biochemical assay with a measured IC50 of 2.3 nM for
the pure C1ꢀ(S)ꢀstereoisomer 3 compared to an IC50 of 1.17 ꢀM for the C1ꢀ(R)ꢀstereoisomer 4.
The coꢀcrystal structure showed that all three critical binding pockets (Leuꢀ26, Trpꢀ23, and Pheꢀ
ACS Paragon Plus Environment
4

Page 5 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
19) of MDM2 were occupied by compound 2. Consistent with the previously observed SAR, the
C7ꢀ(R)ꢀsecꢀbutoxy and the C1 pꢀchlorophenyl residues of 2 bind efficiently in the Leuꢀ26 and
Trpꢀ23 binding pockets of MDM2 respectively, suggesting limited options to form additional
favorable interactions with MDM2 at these positions.14 In contrast, the analysis of the Xꢀray coꢀ
crystal structure suggested some opportunities to improve the interaction with the Pheꢀ19 pocket,
occupied by the 4ꢀpyridinyl moiety of 2 which is surrounded by residues Tyrꢀ67, Metꢀ62 and
Glnꢀ72 of MDM2, according to the Xꢀray structure (Figure 1). The 4ꢀpyridine ring of 2 adopts
an almost parallel orientation with respect to the phenol ring of Tyrꢀ67, consistent with the
original design concept where an aromatic stacking interaction with Tyrꢀ67 was sought.14
Supported by molecular modeling, we then investigated other binding possibilities within the
Pheꢀ19 pocket of MDM2, in particular van der Waals contacts with Metꢀ62 and Hꢀbond
interactions with Glnꢀ72. The replacement of the 4ꢀpyridinyl moiety of 2 by cyclic amines or
cycloalkyl motifs substituted by amino functions, were modeled and suggested a unique
opportunity to interact with both residues at the same time. The piperidine derivative 5 and the
transꢀcyclohexyl amine derivative 6 were the most promising analogs with a 2 – 3 fold gain of
potency over 2 in both TRꢀFRET biochemical and SJSAꢀ1 cellular proliferation assays.
Furthermore, dialkylation of analog 6 was well tolerated and led to compounds with slightly
improved potencies, as illustrated by the dimethylamine analog 7. In addition to good potency in
the SJSAꢀ1 proliferation assay (IC50 of 0.94 ꢀM), its C1ꢀ(S)ꢀstereoisomer 8 possessed interesting
in vitro ADME properties such as high solubility in our highꢀthroughput equilibrium assay (>1
mM at pH 6.8), high permeability in our PAMPA assay (logPe of ꢀ4.4) and a intrinsically low
metabolic clearance in rat liver microsomes (CLint of 5.6 ꢀl.minꢀ1.mgꢀ1 and t1/2 of 246 min). The
only liabilities identified were timeꢀdependent inhibition (TDI) of cytochrome P450 3A (kobs of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 40
1
2
3
4
0.042 minꢀ1) and inhibition of the hERG receptor (IC50 of 1.3 ꢀM in the hERG radioligand
5
binding assay).
6
7
8
9
To further characterize our benchmark compound 8, we performed a rat pharmacokinetic (PK)
experiment (Table 4). Not surprisingly, compound 8 exhibited typical PK parameters for a
compound bearing a basic tertiary amine, meaning a very high volume of distribution (Vss of
23.9 L.kgꢀ1) associated with a long halfꢀtime (t1/2 of 19.1 h). Whereas the low to medium in vivo
clearance in rat (CL of 15 ml.minꢀ1.kgꢀ1) was consistent with measured in vitro data, the
relatively low oral bioavailability was disappointing and suggested an absorption issue not
anticipated based on its in vitro solubility / permeability balance. Further characterization of the
permeability profile of 8 in the Cacoꢀ2 cell permeability assay revealed an asymmetric recovery
of the compound (AꢀB recovery below 30%, a Papp AꢀB of 0.08 (10ꢀ6cm.sꢀ1) and a Papp BꢀA of
0.12 (10ꢀ6cm.sꢀ1)). As the Cacoꢀ2 cell permeability assay can predict human oral absorption and
identify passively transported compounds, these initial data suggest that the PAMPA assay had
overestimated the permeability of our compound and the Cacoꢀ2 data better explain the low oral
absorption of compound 8. Close analogs of 8 were obtained by varying the isobutyl group
occupying the leucine pocket of MDM2. Switching to an achiral isopropoxy group notably
resulted in the identification of 9 with similar cellular potency while simplifying the structure of
our molecules by removing the stereocenter. In conclusion, the profiling of compound 8 and the
identification of 9 highlighted a certain number of challenges which directed our focus to: a)
reducing the basicity to improve permeability and cell potency and b) exploring possibilities to
reduce the lipophilicity of the compounds and at the same time potentially increasing the fu
(fraction unbound) via reduction of the plasma protein binding (PPB). The influence of
molecular properties on the likelihood of success for compounds moving into development was
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
6

Page 7 of 40
Journal of Medicinal Chemistry
1
2
initiated by Lipinski and his “Rule of Five”.21 But more recently, the attention has focused on the
understanding of the physicochemical properties on in vivo toxicological outcomes.22
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Modulating the basicity in the dihydroisoquinolinone series18
The replacement of the basic dimethylamine on the cyclohexyl side chain by amides, for
example a compound where the 4ꢀaminocyclohexyl is extended to propionamide (compound 10)
showed an IC50 > 30 ꢀM in the hERG radioligand binding assay, however displayed clearly
reduced potency in the biochemical TRꢀFRET assay with an IC50 of 20 nM (Table 1). As the
basic dimethylamine appeared to be favorable for potency, we synthesized additional analogs
which maintained the basic group. According to our SAR knowledge and the geometry of the
Pheꢀ19 binding pocket of MDM2, we focused our investigations on cyclic analogs, which were
designed to nicely occupy the Pheꢀ19 binding region as well as modulating the basicity of the
remaining basic center. We rapidly identified very potent compounds such as the imidazolidinꢀ4ꢀ
one 11 and the piperazinꢀ2ꢀone 12, see Table 2, both showing similar or slightly improved
potency compared to 9 in the SJSAꢀ1 proliferation assay with IC50’s of 0.60 and 0.35 ꢀM
respectively. As expected, the basicity of the tertiary amine of compound 12 (calculated pKa of
8.05 ± 0.59 using MoKa20 was significantly reduced compared to the dimethylamine analog 8
(calculated pKa of 10.4 ± 0.59) leading to a ~ 10ꢀfold reduction of the hERG activity (IC50 of 16
ꢀM for compound 12 vs 1.3 ꢀM for compound 8).
The reduced basicity of the tertiary amine of compound 12 had favorable effects on its PK
properties (Table 4). In comparison with compound 8, compound 12 had a much lower volume
of distribution (Vss of 8.0 L.kgꢀ1) and a reduced halfꢀlife (t1/2 of 9.3 h) while its clearance
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 40
1
2
remained moderate (CL of 12 ml.minꢀ1.kgꢀ1) in rats. More importantly, the measured absorption
was quicker and more complete reaching a maximum at 4 hours and resulting in a ~ 3 fold higher
oral exposure and oral bioavailability than for 8. The improved absorption profile of compound
12, where a bioavailability of 49% was reached, was achieved despite the permeability of
compound 12 being significantly lower than for 8 or 9 as assessed in the PAMPA assay (Table
2). Further Nꢀalkylation of the piperazinone group with the aim of increasing permeability led to
the discovery of compound 1. The additional Nꢀmethyl substituent significantly increased
permeability as assessed in our PAMPA assay (logPe of ꢀ4.6), while maintaining very good
cellular potency in the SJSAꢀ1 proliferation assay (IC50 of 0.35 ꢀM) and good metabolic stability
in monkey and human liver microsomes (Table 3). These overall good properties of compound 1
translated to favorable PK properties in rats (see Table 5) where the compound exhibited a low
clearance (CL of 7 ml.minꢀ1.kgꢀ1) and moderate volume of distribution (Vss of 6.4 L.kgꢀ1),
accompanied by a ~ 2ꢀ3 fold increase in oral exposure and bioavailability compared to 12.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As stated earlier, while efforts were focused on reducing the overall basicity of our
compounds, we also explored in parallel possibilities to reduce lipophilicity. As we succeeded on
the first part and the basicity of our compounds was reduced significantly, the lipophilicity to
some extend increased. In our attempts to improve the overall physicoꢀchemical properties of the
compounds, especially with regard to their solubility at neutral pH, positions in the molecules
were probed for accommodating polarity. One opportunity was to replace the central phenyl ring
with a nitrogenꢀcontaining heterocycle, such as pyridine. This idea was supported by modeling
as the introduced polarity was expected to point towards the solventꢀexposed region and not to
be involved in any key interactions. A series of derivatives was synthesized and the results are
summarized in Table 3. Through the introduction of the nitrogen in the phenyl ring, the
ACS Paragon Plus Environment
8

Page 9 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
measured logP could be reduced by at least one log unit. Thus, 1 has a measured logP of 5.8
whereas compounds 15 and 16 have logP values of 4.8 and 4.6, respectively. In parallel, the
calculated polar surface area increases from 65.6 Å2 for 1 up to 91.3 Å2 for compound 17. While
most of the compounds retained their biochemical potency, confirming the tolerance of polarity
in this binding region, they clearly showed reduced liver microsomal stabilities in human and
monkey species, compared to 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 1 binding to MDM2 is species dependent
It was interesting to observe that the dihydroisoquinolinone series demonstrated species
specific effects with respect to MDM2 binding. A comparison of the potency of compound 1 and
Nutlinꢀ3a10 on MDM2 from different species (human, dog, mouse and rat proteins) (Figure 2A)
revealed only minimal deviations between species for Nutlinꢀ3a whereas significant variation
was observed with 1. This compound is 16ꢀfold more potent on human than dog MDM2 and 51ꢀ
and 37ꢀfold more potent on human than mouse and rat MDM2, respectively.
A careful analysis of the peptide sequence of MDM2 in the p53 binding pocket identified two
amino acid differences: Leuꢀ54 and Leuꢀ57 in the human sequence are switched to isoleucines in
rodent species, see Figure 2B. The additional methyl groups in Ileꢀ54 and ꢀ57 result in slightly
larger side chains which protrude into the binding pocket. This leads to a steric clash (see Figure
2C) resulting in a weaker binding affinity and a loss in potency in rodents. In dog, one of the
leucins (Leu54) is replaced by isoleucin which results in a 16ꢀfold reduced binding affinity. As
the monkey to human sequences are the same, we concluded that the required toxicity studies to
assess the onꢀtarget toxicity would need to be evaluated in nonꢀhuman primate (NHP). As a
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 40
1
2
3
4
5
6
7
8
9
consequence, the measured monkey liver microsomal stability became very important as well as
the monkey in vivo PK. Comparing the monkey liver microsomal stability of compounds 1 and
14, the low in vitro clearance of 1 translated well into the in vivo situation (CL 4 ml.minꢀ1.kgꢀ1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Pharmacokinetic profiles of compound 1 in preclinical species
The nonꢀcompartmental PK parameters of 1 estimated in blood of OFꢀ1 mouse, Spragueꢀ
Dawley rat, Beagle dog and Cynomolgus monkey after a single oral (p.o.) or intravenous (i.v.)
dose are shown in Table 5. After i.v. administration, the total blood clearance (CL) of 1 was 5
ml.minꢀ1.kgꢀ1 for mouse, 7 ml.minꢀ1.kgꢀ1 for rat, 3 ml.minꢀ1.kgꢀ1 for dog and 4 ml.minꢀ1.kgꢀ1 for
monkey. Based on the respective hepatic blood flows, 1 showed a consistent low total blood CL
in all species (5 to 10% of hepatic blood flow). The estimated volume of distribution at steadyꢀ
state (Vss) was in the range of 2 to 6 L.kgꢀ1 in all species indicating moderate tissue distribution.
The apparent terminal halfꢀlife (t1/2) was long in rodents and monkey (6 to 12 h), but was
comparatively longer in dogs (20 h). After oral dosing, the compound was well absorbed with
Tmax occurring between 1 and 4.5 h in all species tested. The oral bioavailability (%F) was high
in mouse, rat and dog, and moderate in monkey.
With its combination of high potency and excellent physicochemical and pharmacokinetic
properties, compound 1 was selected for further development towards a clinical candidate.
Binding mode – X-ray of 1 bound to MDM2
ACS Paragon Plus Environment
10

Page 11 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
As illustrated in Figure 3, The dihydroisoquinolinone scaffold of compound 1 occupies the
middle of the binding site and provides the appropriate exit vectors to reach the three critical
binding pockets of the p53 residues Leuꢀ26, Trpꢀ23, and Pheꢀ19. The isopropyl ether at C7 and
the methyl ether at C6 fill the Leuꢀ26 pocket and the ether oxygens make waterꢀmediated Hꢀ
bond interactions with the OH of Tyrꢀ100 and the carbonyl of Glnꢀ24, while the C1 4ꢀ
chlorophenyl group reaches deep into the Trpꢀ23 binding cavity. The C3 carbonyl function of the
dihydroisoquinolinone scaffold is engaged in a waterꢀmediated Hꢀbond interaction with the
carbonyl oxygen of Pheꢀ55 and therefore contributes to the overall affinity of the molecule. More
importantly, it induces a conformational constraint to the Nꢀaryl sideꢀchain next to it, providing
the right torsion angle to ideally enter the Pheꢀ19 binding region. The front of the pocket is
closed as Pheꢀ55 swings up and displays a faceꢀtoꢀedge interaction with the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
dihydroisoquinolinone core. The movement of Pheꢀ55 is specific to this scaffold. The dialkylated
aniline functionality affords the proper exit vector and projects the methyl residue deep into a
small hydrophobic cavity on one side, and the transꢀcyclohexyl moiety on the other side to nicely
occupy the central part of the Pheꢀ19 binding region. Finally, the Nꢀmethyl piperazinone motif
binds towards the exit of the Pheꢀ19(p53) binding cavity in a waterꢀrich region and nestles
perfectly between the protein walls, contributing significantly to the potency of 1.
Chemical synthesis of compound 1
Compound 1 and related analogs described in this paper were synthesized according to the
convergent route presented in Scheme 1. The asymmetric synthesis of the key (S)ꢀ
isoquinolinone intermediate 23 was accomplished in 5 linear steps starting with the alkylation of
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 40
1
2
3
4
5
6
7
8
9
the commercially available (4ꢀhydroxyꢀ3ꢀmethoxyꢀphenyl)ꢀacetic acid ethyl ester. Intermediate
18 was selectively formylated para to the methoxy group with dichloro(methoxy)methane to give
the aldehyde 19 in good yield. Condensation of the aldehyde functionality with the (S)ꢀ2ꢀ
methylpropaneꢀ2ꢀsulfinamide following the published reaction conditions from Davis et al.23
delivered the enantiopure (S)ꢀsulfinylimine 20, which was used as a chiral auxiliary for the
enantioselective introduction of the pꢀchlorophenyl moiety. Classical nucleophiles such as
Grignard or organolithium derivatives reacted much faster on the ethyl ester of 20 leading to the
formation of sideꢀproducts only. Following a methodology developed by Inoue et al.,24 the
desired sulfinylamine 22 was finally obtained with a diastereoselectivity of 96:4 by the addition
of the arylstannane 21 to the (S)ꢀsulfinylimine 20 using the rhodium complex
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
[Rh(cod)(MeCN)2]BF4 as a catalyst. Acidic cleavage of the sulfoxamine group generated the
corresponding primary amine which reacted with the neighboring ester under basic conditions to
deliver the key isoquinolinone intermediate 23 in nearly quantitative yield and with an
enantiomeric excess of 92%. The preparation of the isoquinolinone 23 as a pure (S)ꢀstereoisomer
constituted an important boost for the preparation of the whole series by avoiding the time
consuming chiral separation of all final compounds of interest.
The side chain of 1 was synthesized in six linear steps following the sequence shown in
Scheme 1. Reductive amination of the commercially available (4ꢀformylꢀcyclohexyl)ꢀcarbamic
acid tertꢀbutyl ester with 4ꢀiodoaniline using NaBH(OAc)3 as a reductant led to the intermediate
24. The same reaction conditions were employed in the subsequent step but using formaldehyde
to produce the Nꢀmethyl analog 25. Acidic cleavage of the Bocꢀprotecting group led to the free
amine 26, which was selectively monoꢀalkylated with methyl 2ꢀbromoacetate at low temperature
to give the amino ester 27. The remaining free position on the amino functionality was further
ACS Paragon Plus Environment
12

Page 13 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
alkylated with methylꢀ(2ꢀoxoꢀethyl)ꢀcarbamic acid tertꢀbutyl ester using similar reductive
amination conditions as described above to give the intermediate 28. Acidic cleavage of the Bocꢀ
protecting group of 28 led to the corresponding Nꢀmethyl amine which cyclized with the
neighboring ester under basic conditions to deliver the iodo analog 29 in good yields. Finally,
both components 23 and 29 were combined under crossꢀcoupling reaction conditions developed
by Buchwald et al.,25 using copper(I) iodide as a catalyst and (+/ꢀ)ꢀtransꢀ1,2ꢀdiaminocyclohexane
as a ligand, to provide 1 with greater than 98% chemical purity and an enantiomeric excess of
92%. In the process of further characterization of compound 1, a bisulfate salt has been
identified. Recrystallization allowed compound 1 to be isolated with improved enantiomeric
purity with ee >99%. In addition, the crystalline form provides improvements in solubility and
stability compared to the free base amorphous form.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Profile and potency of 1, selectivity among other PPI
The binding of compound 1 was characterized by several biochemical and biophysical
methods. The data for 1 is provided in Table 6A and compared to Nutlinꢀ3a. Compound 1 binds
to human MDM2 with an IC50 of 1.7 nM and shows high selectivity over MDM4 (IC50 = 2000
nM). Compound 1 is about 4ꢀtimes more potent than Nutlinꢀ3a (IC50 = 8.0 nM).
To understand the thermodynamic parameters of the interactions in solution, isothermal
titration calorimetry (ITC) was carried out using 1 and human MDM2 (using the construct
MDM2[14ꢀ111;L33E]). The thermodynamic signature is shown in Table 6B. Overall, both the
ꢁH and –TꢁS contributions are favorable (i.e. negative), which correlates with the favorable van
der Waals interactions made by compound 1 which do not lead to a strong rigidification of
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 40
1
2
3
4
5
6
7
8
9
MDM2. Finally, as shown in Table 6C, a thermal shift assay, also called Differential Scanning
Fluorimetry (DSF), was done to measure the thermal stability increase of human MDM2 upon
binding of 1 to the protein. For compound 1, thermal stabilization with a ꢁT of 25°C was
observed, which correlates with the low nM affinity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 1 selectivity was evaluated in a number of biochemical assays to assess its activity
against multiple binding partners, including the p53:MDM4 interaction (MDM4 protein being
closely related to MDM2). In TRꢀFRET assays, as described in Table 6D, compound 1 was
shown to be selective for the p53:MDM2 interaction compared to the p53:MDM4 interaction
(1’176ꢀfold selectivity) and the RasꢀRaf interaction (3’000ꢀfold selectivity). In addition, 1
showed no significant activity against Bclꢀ2:Bak, Bclꢀ2:Bad, Mclꢀ1:Bak, Mclꢀ1:NOXA,
XIAP:BIR3 and cꢀIAP:BIR3 proteinꢀprotein interactions.
Activity of compound 1 in cell-based models
Potency of p53:MDM2 interaction inhibitors can be evaluated in cells with wildꢀtype p53.
Disruption of the p53:MDM2 interaction should release p53 from MDM2, leading not only to a
stabilization but also to a nuclear translocation of p53. Compound 1ꢀinduced p53 nuclear
translocation can be evaluated in p53 wildꢀtype cells using a mechanistic p53 reꢀdistribution
assay (based on the GRIP® technology) and can be quantified with a highꢀcontent imaging
platform. As shown in Table 7, compound 1 was able to significantly reꢀdistribute wildꢀtype p53
into the cell nucleus with an IC50 of 0.224 ꢂM, demonstrating its ability to inhibit the
p53:MDM2 interaction in living cells. In addition, compound 1 activity against the p53:MDM2
interaction was assessed in proliferation assays using either wildꢀtype p53 or p53 null cells. Two
ACS Paragon Plus Environment
14

Page 15 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pairs of cell lines were used to assess compound 1 p53ꢀdependent antiꢀproliferative effects: 1) an
isogenic pair of HCT116 cell lines either expressing wildꢀtype p53 or knockedꢀout for the p53
gene and 2) a nonꢀisogenic pair of osteosarcoma cell lines either endogenously expressing wildꢀ
type p53 and amplified for MDM2 (SJSAꢀ1 cells) or null for p53 (SAOSꢀ2 cells). As indicated in
Table 7, compound 1 significantly inhibited the proliferation of cells expressing wildꢀtype p53,
while sparing the p53 null cells with a 35 to 58ꢀfold difference. Overall, these results indicate
that 1 inhibits the p53:MDM2 interaction in cells, leading to p53 nuclear translocation that
results in cell growth inhibition, in a p53ꢀdependent manner.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In vivo anti-tumor activity of compound 1
Compound 1 was able to inhibit the interaction between p53 and MDM2 and reactivate the p53
pathway in vivo in a MDM2ꢀ amplified SJSAꢀ1 human tumor model, as judged by elevation of
p21 mRNA levels, a pharmacodynamic (PD) indicator for p53 activity.26 As shown in Figure 5,
p21 mRNA levels were found to increase concomitantly with levels of compound 1 in tumorꢀ
bearing rats dosed at 30 mg/kg. The PD response was biꢀphasic and prolonged up to 24 h.
Additional p53 target genes such as MDM2 and PUMA mRNA levels were assessed in the
tumor samples as well and showed a similar behavior.
Daily treatment with 1 doseꢀdependently and significantly inhibited SJSAꢀ1 tumor growth in
rats (Figure 6). It promoted stable disease at 20 mg/kg which was associated with a plasma
AUC0ꢀ24 of 163 ꢂM.h (Table 8). The treatment dose at 30 mg/kg, dosed daily, induced >85%
tumor regression. The 3qw (treatment on Monday, Wednesday, and Friday) treatment doseꢀ
dependently and significantly affected SJSAꢀ1 tumor growth. It induced a T/C of 25% at 30
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 40
1
2
3
4
5
6
7
8
9
mg/kg and full tumor regression at 70 mg/kg which was associated with a plasma AUC0ꢀ24 of 616
ꢂM.h and a plasma AUC0ꢀ48 of 1045 ꢂM.h (Table 8). 3qw treatment was well tolerated at all the
doses and no histopathology findings related to the compound were made. For the daily
treatment, histopathology was performed for the 30 mg/kg dose and no histopathology findings
related to the compound were made either. The fact that the doses were well tolerated was not a
surprise to us, considering the human MDM2 specificity of 1, as described above. In summary, 1
shows doseꢀproportional increases in exposure and a clear PK/PD relationship, resulting in tumor
growth inhibition effects at wellꢀtolerated doses in tumorꢀbearing rats. Dosing schedules have
been investigated and indicate that 1 inhibits tumor growth in rats, not only on a daily dosing
schedule but also when animals are treated 3ꢀtimes a week.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 1 in toxicity studies
Following repeatedꢀdose oral administration, the bone marrow, the lymphoid organs, the GI
tract and the testes (organs with a high degree of cell turnover) were identified as target organs of
toxicity in the monkey. In the rat, compound 1ꢀrelated microscopic findings were present in the
testes, heart (males only) and adrenal gland (both gender). It should be noticed that in rats, the
lesion in the testes was a subtle structural change affecting the tubular epithelium (no effect on
proliferative cells), in contrast to the monkey where germinal cells were affected. Thus, actively
dividing cells in the rat testes were not affected in contrast to the monkey testes.
The changes observed in the 4ꢀweek GLP studies were consistent with the nonꢀGLP 2ꢀweek
DRF studies and were treatment durationꢀ and doseꢀdependent. All changes showed complete or
partial reversibility after 4 weeks. Based on the similar potency of compound 1 against the
ACS Paragon Plus Environment
16

Page 17 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
p53/MDM2 proteinꢀprotein interaction in humans and Cynomolgus monkey, the latter is
considered a good predictor of pharmacological and toxicological responses in humans. In
contrast, the potency towards p53/MDM2 proteinꢀprotein interaction is 63ꢀfold lower in rat than
in human. Although monkeys are expected to more accurately predict human pharmacological
response to 1 and any resulting onꢀtarget toxicity, the rat may also predict offꢀtarget effects not
observed in monkeys.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conclusion
A combination of a good understanding of the target and its binding pocket, virtual screening,
Xꢀray crystallography, molecular modeling and iterative medicinal chemistry resulted in the
discovery of the dihydroisoquinolinone series. Within this novel series, compound 1 was
identified as a potent and selective human MDM2 inhibitor with excellent in vitro and in vivo
characteristics. Compound 1 is currently undergoing phase 1 clinical trials in p53wt tumors.
AUTHOR INFORMATION
Corresponding Author
* P.H.: (phone, office) +41 61 696 2087; (email) philipp.holzer@novartis.com
Present Addresses
† PeptiDream Inc., 4ꢀ6ꢀ1 Komaba, Meguroꢀku, Tokyo, 153ꢀ8904, Japan
Notes
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 40
1
2
3
4
The authors declare no competing financial interest.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACKNOWLEDGMENT
The authors would like to acknowledge all colleagues being involved in the discovery of NVPꢀ
CGM097. Especially we would like to thank the following associates for their technical
assistance: Andreas Boos, JeanꢀMarc Gröll, Emilie Joly, Alexandra Löffler, Julien Lorber, Pierre
Nimsgern, Julien Scesa, Sylvie TeixeiraꢀFouchard, Rainer Tschan and Werner Vetterli. We
thank Robert Mah and Gregory Hollingworth for their careful review of the manuscript.
ABBREVIATIONS
3qw, three times per week dosing; AcOH, acetic acid; AUC, area under the curve; CL, clearance;
CMC, carboxymethyl cellulose ; CYP, cytochrome P450; DCM, dichloromethane; DMF, N,Nꢀ
diemethylformamide; DMSO, dimethylsulfoxide; dr, diastereoselectivity ratio; DSF, differential
scanning fluorimetry; ee, enantiomeric excess; er, enantioselectivity ratio; EtOAc, ethyl acetate;
EtOH, ethanol; GI50, concentration causing 50% growth inhibition; aq. HCHO, aqueous
formaldehyde; HDM2, human analog of MDM2; HPMC, hydroxypropyl methyl cellulose; fu,
fraction unbound; HTS, highꢀthroughput screening; IC50, half maximal inhibitory concentration;
ITC, isothermal titration calorimetry; iv, intravenous; K2CO3, potassium carbonate; LM, liver
microsome; MeOH, methanol; MDM2, murine double minute 2; NHP, nonꢀhuman primate;
NMR, nuclear magnetic resonance; PD, pharmacodynamic; PK, pharmacokinetics; PSA, polar
surface area; q24h, once a day dosing; qRTꢀPCR, reverse transcription polymerase chain
reaction; RT, room temperature; SAR, structure − activity relationship; SD, standard deviation;
ACS Paragon Plus Environment
18

Page 19 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
SEM, standard error of mean; TDI, time dependent inhibition; TFA, trifluoroacetic acid; THF,
tetrahydrofuran; TLC, thin layer chromatography; TRꢀFRET, time resolved fluorescence
resonance energy transfer; VSS, volume of distribution.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Supporting Information Available: Synthesis procedures of compound 1 has been provided. This
material is available free of charge via the Internet at http://pubs.acs.org.
REFERENCES
1. Lane, D.P.; p53, Guardian of the genom. Nature 1992, 358, 15–16.
2. Harris, S. L.; Levine, A. J.; The p53 pathway: positive and negative feedback loops.
Oncogene 2005, 24, 2899ꢀ2908.
3. Michael, D.; Oren, M.; The p53ꢀMdm2 module and the ubiquitin system. Semin. Cancer
Biol. 2003, 13, 49ꢀ58.
4. Kemp, C. J.; Donehower, L. A.; Bradley, A.; Balmain, A.; Reduction of p53 gene dosage
does not increase initiation or promotion but enhances malignant progression of chemically
induced skin tumors. Cell 1993, 74, 813−822.
5. Feki, A.; IrmingerꢀFinger, I.; Mutational spectrum of mutations in primary breast and
ovarian tumors. Crit. Rev. Oncol. Hematol. 2004, 52, 103−116.
6. Kruse, J.ꢀP.; Gu, W.; Modes of p53 regulation. Cell 2009, 137, 609ꢀ622.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 40
1
2
3
4
7. Khoo, K. H.; Verma, C. S.; Lane, D. P.: Drugging the p53 pathway: understanding the route
5
6
to clinical efficacy. Nature Rev. Drug Discov. 2014, 13, 217ꢀ236.
7
8
9
8. Kussie, P. H.; Gorina, S.; Marechal, V.; Elenbaas, B.; Moreau, J.; Levine, A. J.; Pavletich,
N. P.; Structure of the MDM2 oncoprotein bound to the p53 tumor supressor transactivation
domain. Science 1996, 274, 948ꢀ953.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9. GarcíaꢀEcheverría, C.; Chène, P.; Blommers, M.; Furet, P.; Discovery of potent antagonists
of the interaction between human double minute 2 and tumor suppressor p53. J. Med. Chem.
2000, 43, 3205ꢀ3208.
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.; In vivo activation of the p53
pathway by smallꢀmolecule antagonists of MDM2. Science 2004, 303, 844ꢀ848.
11. Furet, P.; Chene, P.; De Pover, A.; Valat, T.; Hergovich Lisztwan, J.; Kallen, J.; Masuya,
K.; The central valine concept provides an entry in a new class of non peptide inhibitors of the
p53ꢀMDM2 interaction. Bioorg. Med. Chem. Lett. 2012, 22, 3498ꢀ3502.
12. Vaupel, A.; Bold, G.; De Pover, A.; StachyraꢀValat, T.; Hergovich Lisztwan, J.; Kallen, J.;
Masuya, K.; Furet, P.; Tetraꢀsubstituted imidazoles as a new class of inhibitors of the p53ꢀ
MDM2 interaction. Bioorg. Med. Chem. Lett. 2014, 24, 2110–2114.
13. Jacoby. E.; Boettcher, A.; Mayr, L.; Brown, N.; Jenkins, J.; Kallen, J.; Engeloch, C.;
Schopfer, U.; Furet, P.; Masuya, K.; Knowledgeꢀbased virtual screening: application to the
MDM4/p53 proteinꢀprotein interaction. Methods Mol. Biol. 2009, 575, 173ꢀ194.
ACS Paragon Plus Environment
20

Page 21 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
14. Gessier, F.; Kallen, J.; Jacoby, E.; Chène, P.; StachyraꢀValat, T.; Ruetz, S.; Jeay, S.;
Holzer, P.; Masuya, K.; Furet, P.; Discovery of dihydroisoquinolinone derivatives as novel
inhibitors of the p53ꢀMDM2 interaction with a distinct binding mode. Bioorg. Med. Chem. Lett.,
2015, DOI: http://dx.doi.org/10.1016/j.bmcl.2015.06.058.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
15. Jeay, S.; Gaulis, S.; Ferretti, S.; Bitter, H.; Ito, M.; Valat, T.; Murakami, M.; Ruetz, S.;
Guthy, D.A.; Rynn, C.; Jensen, M. R.; Wiesmann, M.; Kallen, J.; Furet, P.; Gessier, F.; Holzer,
P.; Masuya, K.; Würthner, J.; Halilovic, E.; Hofmann, F.; Sellers, W. R.; Graus Porta, D.; A
distinct p53 target gene set predicts for response to the selective p53ꢀHDM2 inhibitor NVPꢀ
CGM097. eLife 2015; DOI: 10.7554/eLife.06498.
16. Collingwood, S.P.; Ratcliffe, A.J.; Pryde, D.; Porter, R.; Recent disclosures of clinical
candidates: highlights from the Society for Medicines Research symposium, held December 4,
2014 – National Heart & Lung Institute, London, UK. Drug Future, 2015, 40, 81ꢀ91.
17. Neochoritis, C.; EstradaꢀOrtiz, N.; Khoury, K.; Dömling, A.; p53–MDM2 and MDMX
antagonists. Annu. Rep. Med. Chem., 2014, 49, 167ꢀ187.
18. Berghausen, J.; Buschmann, N.; Furet, P.; Gessier, F.; Hergovich Lisztwan, J.; Holzer, P.;
Jacoby, E.; Kallen, J.; Masuya, K.; PissotꢀSoldermann, C.; Ren, H.; Stutz, S.; Substituted
isoquinolinones and quinazolinones; WO 2011076786 A1.
19. Routinely, inhibitory effects of our compounds on p53ꢀMDM2 or p53ꢀMDM4 binding
were measured by a TimeꢀResolved Fluorescence Resonance Energy Transfer (TRꢀFRET) assay
and expressed as the concentration causing 50% inhibition (IC50). Compounds at different
concentrations were added to 0.1 nM Europiumꢀlabeled streptavidin (Perkin Elmer), 0.1 nM
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 40
1
2
3
4
5
6
biotinylated MDM2 or MDM4 p53 binding domain and 10 nM Cy5ꢀp53 aa18ꢀ26 (unpublished
results).
7
8
9
20. Milletti, F.; Storchi, L.; Sforna, G.; Cruciani, G.; New and original pKa prediction method
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
using grid molecular interaction fields. J. Chem. Inf. Model., 2007, 47, 2172–2181.
21. Lipinski, C.A.; Lombardo, F.; Dominy, B.W.; Feeney, P.J.; Experimental and
computational approaches to estimate solubility and permeability in drug discovery and
development settings. Adv Drug Deliv Rev., 1997, 23, 3ꢀ25.
22. Hughes, J.D.; Blagg, J.; Price, D.A.; Bailey, S.; Decrescenzo, G.A.; Devraj, R.V.;
Ellsworth, E.; Fobian, Y.M.; Gibbs, M.E.; Gilles, R.W.; Greene, N.; Huang, E.; KriegerꢀBurke,
T.; Loesel, J.; Wager, T.; Whiteley, L.; Zhang, Y.; Physiochemical drug properties associated
with in vivo toxicological outcomes. Bioorg. Med. Chem. Lett., 2008, 18, 4872ꢀ4875.
23. Davis, F. A.; Mohanty, P. K.; Asymmetric synthesis of the protoberberine alkaloid (S)ꢀ(ꢀ)ꢀ
xylopinine using enantiopure sulfinimines. J. Org. Chem., 2002, 67, 1290ꢀ1296.
24. Oi, S.; Moro, M.; Fukurhara, H.; Kawanishi, T.; Inoue; Y.; Rhodiumꢀcatalyzed addition of
arylstannanes to carbonꢀheteroatom double bond. Tetrahedron, 2003, 59, 4351ꢀ4361.
25. Klapars, A.; Huang, X.; Buchwald, S. L.; A general and efficient copper catalyst for the
amidation of aryl halides. J. Am. Chem. Soc., 2002, 124, 7421ꢀ7428.
26. elꢀDeiry, W. S.; Tokino, T.; Velculescu V.E.; Levy, D. B.; Parsons, R.; Trent, J. M.; Lin,
D.; Mercer, W. E.; Kinzler, K. W.; Vogelstein, B.; WAF1, a potential mediator of p53 tumor
supression. Cell, 1993, 75, 817ꢀ825.
ACS Paragon Plus Environment
22

Page 23 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 1. Xꢀray cocrystal structure of compound 2 bound to human MDM2 at 1.67 Å resolution
(PDB code 4ZYI). MDM2 binding pockets are labeled in yellow font by p53 side chain. The
stick representation is used for the MDM2 residues proximal to the 4ꢀpyridine ring of 2. Water
molecules are not represented.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 40
1
2
3
4
Table 1. Early SAR showing the activities of compounds 2 to 10 in the TRꢀFRET biochemical
5
6
assay and in the SJSAꢀ1 cell proliferation assay.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
SJSAꢀ1 cell
proliferation
TRꢀFRET
assay IC50
Compound C1ꢀchirality
R1
R2
(ꢀM)a with SD
GI50 (ꢀM)
2
3
4
(R/S)
(S)
0.008 b
3.86 b
0.0023
±0.0004
1.21 ± 0.27
29.58 b
1.171
±0.338
(R)
5
6
7
(R/S)
(R/S)
(R/S)
0.004 b
1.09 b
0.003
±0.001
2.85 ± 1.69
1.13 ± 0.25
0.003 b
0.0019
±0.0001
8
9
(S)
(S)
0.94 ± 0.015
0.71 ± 0.23
0.0011
±0.0002
ACS Paragon Plus Environment
24

Page 25 of 40
Journal of Medicinal Chemistry
1
2
3
4
5
6
0.020
±0.002
10
(S)
4.96 ± 3.45
7
8
9
a
Values are expressed as the mean ± standard deviation of at least three independent
experiments. b Data provided from single experiment.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 40
1
2
3
4
Table 2. Summary of compounds 1 and 9ꢀ12 with modified basicity on the side chain.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
passive
SJSAꢀ1 cell
proliferation
GI50 (ꢀM)a
TRꢀFRET
assay IC50
(ꢀM)a
permeability
logPe
Compound
Side chain R
[10ꢀ6 cm.sꢀ1]
ꢀ4.4
0.0011
±0.0002
9
0.71 ± 0.23
0.0017
±0.0003
11
0.60 ± 0.036
ꢀ5.5
ꢀ5.2
ꢀ4.6
0.0016
±0.0002
12
0.35 ± 0.17
0.0017
±0.0001
1
0.35 ± 0.19
a
Values are expressed as the mean ± standard deviation of at least three independent
experiments.
ACS Paragon Plus Environment
26

Page 27 of 40
Journal of Medicinal Chemistry
1
2
3
4
Table 3. Summary of selected ADME properties of dihydroisoquinolinone analogs.
5
6
7
8
9
10
11
12
13
14
15
16
17
Compound
Side chain R
TRꢀFRET
assay IC50
(ꢀM)a
CYP3A4
inhibition in
human LM IC50 Partition
(ꢀM)
Octanol/
Water
passive
permeabi stability in
metabolic
metabolic
stability in
human
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
monkey
lity logPe
Coefficie
nt logP
LMb: CLint
LMb: CLint
[10ꢀ6 cm.
sꢀ1]
(ꢁL.minꢀ1.
mgꢀ1)
(ꢁL.minꢀ1.
mgꢀ1)
0.0017
±0.0001
1
9
17.0
9.4
5.8
ꢀ4.6
ꢀ4.4
ꢀ5.2
26.6
n.d.
52.3
4.6
0.0011
±0.0002
>4.2c
5.3
0.0016
±0.0002
12
11.3
79.7
3.4
13
14
0.0114d
5.0
>6.3c
4.4
n.d.
ꢀ4.6
n.d.
78.3
0.0025
±0.0012
13.2
224.1
156.7
0.0020
±0.0005
15
>20
4.8
ꢀ4.7
278.6
150.8
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 40
1
2
3
4
5
6
7
8
0.0076
±0.0008
16
17
>20
>20
4.6
4.6
ꢀ4.5
ꢀ4.3
355.4
420
155.7
9
0.0041
±0.0007
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
220
a
Values are expressed as the mean ± standard deviation of at least three independent
b
c
d
experiments. LM = liver microsomes. logP data point could not be determined accurately.
Data provided from single experiment.
ACS Paragon Plus Environment
28

Page 29 of 40
Journal of Medicinal Chemistry
1
2
3
4
Table 4. Intravenous (i.v.) and oral (p.o.) pharmacokinetic profiles of compounds 8 and 12 in
5
6
ratsa.
7
8
9
CL
AUC (i.v.)
(nmol.h.Lꢀ
)
PPB
t1/2term
(h)
Vss
AUC (p.o.)
(nmol.h.Lꢀ1)
Cmax
tmax
(rat %)b
(ml.minꢀ
1.kgꢀ1)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound
%F
(L.kgꢀ1)
(nmol.Lꢀ1) (h)
1
7.0
±
1.2
19.1 ±
3.1
23.9 ±
3.0
1574 ±
128
18 ±
5.0
8
15.0 ± 1.0
12.0 ± 1.0
297 ± 116
17.6 ± 5.0
>99
>99
4.0 49.0
±
1.4 15.0
9.3 ±
1.9
8.0 ±
1.8
2154 ±
207
74.6 ±
22.7
12
1018 ± 346
±
a) dose: 1 mg/kg i.v. as a solution in NMP:PEG200 (10:90) and 3 mg/kg p.o. as a suspension in
CMC:Water:Tween (0.5:99:0.5). AUC and Cmax data are dose normalized to 1 mg.kgꢀ1 and data
are expressed as the mean ± standard deviation from four rats. b) plasma protein binding
measured by rapid equilibrium dialysis (RED) assay using 100% plasma.
ACS Paragon Plus Environment
29

Journal of Medicinal Chemistry
Page 30 of 40
1
2
3
4
5
6
7
8
9
Figure 2. A: Measured Ki’s for comparison of the potency of compound 1 and Nutlinꢀ3a was
evaluated on MDM2 from different species (human, dog, mouse and rat proteins). B:
Comparison of the protein sequence in the p53 binding pocket of MDM2 identified two
differences in the amino acid sequence. C: Interactions of 1 with amino acids 54 and 57 of
MDM2 results in a steric clash and reduced affinity to the binding pocket.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
Ki ± SD in nM (Ki ratio vs human)
1
Nutlinꢀ3a
Human MDM2
Dog MDM2
1.3 ± 0.1 (1)
20.5 ± 1.7 (16)
65.9 ± 1.0 (51)
47.4 ± 10.4 (37)
6.4 ± 1.3 (1)
10.5 ± 2.0 (2)
19.9 ± 1.4 (3)
12.6 ± 0.7 (2)
Mouse MDM2
Rat MDM2
Ki, calculated from the TRꢀFRET IC50 using the ChengꢀPrusoff equation
ACS Paragon Plus Environment
30