Identification of the first non-peptidic small molecule inhibitor of the c-Abl/14-3-3 protein-protein interactions able to drive sensitive and Imatinib-resistant leukemia cells to apoptosis

Article abstract of DOI:10.1016/j.bmcl.2010.08.019

An in silico structure-based ligand design approach resulted in the identification of the first non-peptidic small molecule able to inhibit protein-protein interactions between 14-3-3 and c-Abl. This compound shows an anti-proliferative effect on human leukemia cells either sensitive or resistant to Imatinib, in consequence of the T315I mutation. It also mediates c-Abl release from 14-3-3 in a way similar to that found in response to Imatinib treatment.

Full text of DOI:10.1016/j.bmcl.2010.08.019

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6133–6137
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of the first non-peptidic small molecule inhibitor
of the c-Abl/14-3-3 protein–protein interactions able to drive sensitive
and Imatinib-resistant leukemia cells to apoptosis
a,
b
a
b
b
Valentina Corradi , Manuela Mancini , Fabrizio Manetti , Sara Petta , Maria Alessandra Santucci ,
Maurizio Botta
a
a,
⇑
Dipartimento Farmaco Chimico Tecnologico, Università degli Studi di Siena, Via Aldo Moro 2, I-53100 Siena, Italy
Dipartimento di Ematologia e Scienze Oncologiche ‘‘Lorenzo e Ariosto Seràgnoli”, Università di Bologna, Via Massarenti 9, I-40138 Bologna, Italy
b
a r t i c l e i n f o
a b s t r a c t
Article history:
An in silico structure-based ligand design approach resulted in the identification of the first non-peptidic
small molecule able to inhibit protein–protein interactions between 14-3-3 and c-Abl. This compound
shows an anti-proliferative effect on human leukemia cells either sensitive or resistant to Imatinib, in
consequence of the T315I mutation. It also mediates c-Abl release from 14-3-3 in a way similar to that
found in response to Imatinib treatment.
Received 28 June 2010
Revised 3 August 2010
Accepted 4 August 2010
Available online 8 August 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
1
4-3-3 Proteins
Leukemia
Drug discovery
Molecular modeling
Non-peptidic compounds
Protein–protein interactions
The constitutively activated tyrosine kinase (TK) activity of the
Abl moiety of the p210 Bcr-Abl fusion protein is the causative
event of chronic myeloid leukemia (CML). Accordingly, TK inhibi-
tors (Fig. 1) are currently used in CML, and, among them, Imatinib
mesylate (1) represents the frontline therapy as it has revolution-
ized the treatment of this disease, improving the long-term sur-
vival of CML patients.2
However, despite the generally positive response to the clinical
therapy, resistance to 1 emerged as a major problem in the treat-
ment of CML, particularly in more advanced stages. Resistance to
alternative and/or complementary therapeutic strategies to over-
come the drug resistance in CML, we focused on 14-3-3 proteins,
which play a critical role in sub-cellular redistribution of pro-apop-
totic proteins, such as Abl. In fact, binding of Abl to 14-3-3 is
responsible for its prevalent cytoplasmic localization,6 while dis-
ruption of the Abl/14-3-3 complex upon DNA damage leads Abl
to nuclear translocation and activation of apoptosis.6 Moreover,
we have recently proved that p210 Bcr-Abl precludes post-transla-
tional modifications of 14-3-3 (namely, its phosphorylation at
1
Ser186 for the
r isoform), preventing Abl nuclear translocation
7
1
is most often driven by Bcr-Abl point mutations (such as T315I,
and reducing the apoptosis activation in CML cells. All these re-
sults suggest that, in addition to the direct inhibition of the Bcr-
Abl oncoprotein, targeting 14-3-3 with molecules able to affect
its interactions with c-Abl (the residual c-Abl not rearranged with
Bcr) may potentiate the current therapies for CML and be also
effective against resistance to 1 due to the T315I mutation.8
Accordingly, pharmaceutical research is now much more inter-
ested in discovery and developing small molecule inhibitors of pro-
tein–protein interactions that could provide novel opportunities
for therapy. However, no small molecule non-peptidic inhibitors
of 14-3-3 have been reported yet. The sole inhibitor of 14-3-3
known so far is a 20 amino acid polypeptide (thereafter referred
to as R18) that binds the amphipatic groove of the protein through
F317L, F359V) that can directly impair the binding of 1 or affect
the conformation of the kinase domain required for the drug bind-
ing (Y253F, E255K in the ATP-binding loop; H396R in the activa-
3
tion loop; M351T, E355G, F359V in the catalytic domain). The
new generation of Bcr-Abl inhibitors, such as Dasatinib and Niloti-
nib (2 and 3, Fig. 1), overcomes the resistance associated with the
great majority of Bcr-Abl mutations, with the only exception of
4
T315I. Moreover, compounds 5–9 (Fig. 1) are also known to inhi-
bit T315I mutants resistant to 1.5 To explore the potentiality of
⇑
Corresponding author. Tel.: +39 0577 234306; fax: +39 0577 234333.
E-mail address: botta@unisi.it (M. Botta).
Present address: Department of Biological Sciences, University of Calgary, 2500
University Dr. NW, Calgary, AB, Canada T2N-1N4.
9
a WLDL sequence. R18 is able to mask the region of the protein re-
0
960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.08.019

6
134
V. Corradi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6133–6137
Figure 1. Schematic representation of current available inhibitors of Abl (1–8) and BV02 (9, discovered by a structure-based drug design approach). Compounds 5–8 are
inhibitors of the T315I mutant form of Bcr-Abl resistant to 1.
quired to bind client proteins, thus promoting Abl translocation to
the nucleus. In the attempt to identify compounds lacking in
However, because of the size of the peptide ligand, the pharma-
cophoric model was characterized by a high complexity in terms of
the number of pharmacophoric features (Table 1S of Supplemen-
tary data). As a consequence, it was simplified by keeping hydro-
gen bond interactions with distances and angles stabilized to
default values during molecular dynamics runs (Table 2S of Sup-
plementary data). The final pharmacophoric model was used as a
three-dimensional query to screen in silico a database of about
200,000 commercially available compounds, resulting in 99 entries
that were further pruned to 87 by application of the Lipinski’s rule
of five. Selected compounds were then docked by means of GOLD
6
drawbacks of peptides, small non-peptidic molecules possibly able
to disrupt the interactions between 14-3-3 and Abl have been
searched by us by means of a computational protocol based on
structure-based pharmacophore modeling, in silico virtual screen-
ing and molecular docking simulations. Structure-based pharma-
cophores are useful tools for virtual screening procedures
because they codify important information derived from target–li-
gand interactions. Moreover, their combination with docking-
based virtual screening procedure is known to significantly im-
1
0
13
prove the probability of identifying putative hit candidates.
software into the 14-3-3 binding site, and ligand–target interac-
The three-dimensional structure of the complex between 14-3-
isoform and the MARSH-phosphoS-YPA nonapeptide (protein
data bank entry 1YWT) was used to generate a map of the major
tions were evaluated by a consensus scoring approach based on
1
3
14
3r
both ChemScore and XScore scoring functions. Moreover, the
best docked conformation of each ligand was checked for its ability
to map at least three pharmacophoric features of the model, as
well as GRID minima corresponding to the regions of most favor-
1
1
interactions between the protein and its ligand by means of the
1
2
software LIGANDSCOUT
.

V. Corradi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6133–6137
6135
able interaction between the protein and five different chemical
probes, as found by GRID software.15 As a result, 14 compounds
that were able to satisfy such constraints were purchased and sub-
mitted to biological tests.
24 h) mediates c-Abl release from 14-3-3 in a way similar to that
found in response to treatment with 1, through events encompass-
ing 14-3-3
r phosphorylation by c-Jun N-terminal kinase
6
(Fig. 3A). As a consequence, c-Abl translocated to the nuclear com-
partment, where it is expected to trigger p73- and p53-dependent
pro-apoptotic signals. c-Abl nuclear import concerned the whole
length protein (145 kDa) and the p120 kDa fragment generated by
interaction with caspase 9 (Fig. 3B). The significant increase of cy-
clin-dependent kinase inhibitor p27
Preliminary cell proliferation assays conducted in Ba/F3 cells
transduced with the wt Bcr-Abl construct led to the identification
of BV02 (9, Fig. 1) as the most proficient compound in terms of
cytotoxic potential (data not shown). BV02 cytotoxicity was as-
sayed in clonogenic assays, an in vitro technique used to quantify
the drug impact on cell reproductive integrity.16 Further experi-
ments showed that BV02 induced a significant dose-dependent
reduction of reproductive integrity of Ba/F3 cells expressing both
wt and T315I Bcr-Abl protein, with LD50 values of 1.04 ± 0.09 and
1
9
Kip1
in response to BV02 sug-
gests that a common mechanism for the nuclear import of
proteins involved in cell proliferation and survival could be based
on their release from 14-3-3 (Fig. 3B). Moreover, BV02-induced c-
Abl translocation to mitochondrial membranes in the full-length
and caspase-induced cleaved isoforms (mainly the 60 kDa fragment,
Fig. 3C). At this level, c-Abl integration was associated with a signif-
icant increase of caspase 8 (18 kDa) and caspase 9 (35 kDa) cleaved
fragments, a critical event for their dissipation in response to BV02
1
.47 ± 0.12
l
M, respectively (Fig. 2A).17 The cytotoxicity of BV02
(5 lM for 24 h) in both cell types was, at least partly, driven by
the induction of apoptosis, measured by cytofluorimetric analysis
of annexin V and propidium iodide uptake (Fig. 2B).17 BV02 had
a similar anti-proliferative effect on early hematopoietic progeni-
2
0
(Fig. 3C). All these events were elicited by BV02 also in Ba/F3 cell
line expressing the T315I Bcr-Abl mutation (Fig. 4A–C).
+
tors (CD34 ) isolated from CML patients who developed resistance
1
7
to 1 in consequence of T315I Bcr-Abl mutation (Fig. 2C). Taken
together, these results showed that the inhibition of 14-3-3 upon
BV02 in vitro treatment is effective against Bcr-Abl-expressing
cells either sensitive or resistant to 1.
Finally, Figure 5A and 5B describes the theoretical binding mode
of BV02 found with our docking calculations. In particular, the car-
boxylic group at the C5 of BV02 terephthalic moiety corresponds to
the phosphate group of the phosphopeptide used to generate the
pharmacophoric model. Its binding pocket is lined by K49, R56,
R129 and Y130. Hydrogen bond interactions also occur between
K122 and N175 with the carbonylic group at the C3 of the pyrazole
ring and the NH group of the amide moiety, respectively. More-
over, the hydrophobic interactions involving the aromatic ring of
the Tyr residue of the phosphopeptide are mimicked by the two
methyl groups of the pyrazole ring of BV02.
To elucidate the mechanism driving apoptotic death induction in
response to BV02, we investigated whether the compound owns the
ability of releasing c-Abl from 14-3-3r and promoting its relocation
at sub-cellular compartments, that is, the nucleus and mitochon-
drial membranes. Western blot or immunoprecipitation (IP)/immu-
1
8
noblotting technique was used to perform protein analysis. In Ba/
F3 cell line transduced with the wt Bcr-Abl construct, BV02 (5 M for
l
Figure 2. Effects of BV02 on Bcr-Abl-expressing cells. (A) Activity of BV02 toward reproductive integrity of Ba/F3 cells expressing the wt Bcr-Abl sensitive to 1 (upper panels)
+
or the T315I mutation resistant to 1 (lower panels). (C) Activity of BV02 toward CD34 cells isolated from bone marrow samples of 3 (CML) patients in blast crisis who
developed in vivo resistance to 1. (B) Cytofluorimetric analysis of annexin V and propidium iodode uptake was used to measure apoptosis induction in response to BV02
(
5
lM for 24 h) in Ba/F3 cells expressing the wt or the T315-mutated Bcr-Abl protein. Results shown in sections A and B represent the mean values ± SD of three individual
experiments.

6
136
V. Corradi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6133–6137
Figure 3. Effects of BV02 on c-Abl sub-cellular relocation in wt Bcr-Abl-expressing Ba/F3 cells. Protein expression and interactions in cytoplasmic (A) and nuclear (B)
compartments and in mitochondrial membranes (C) were analyzed by means of Western blot or IP/immunoblotting of either untreated or after 24 h exposure to 1 M 1 or 5
M BV02. b-Actin, histone H1 and cytochrome c (Cyt c) levels served as controls for loading of cytoplasmic, nuclear and mitochondrial membrane proteins, respectively.
l
l
Identical results were obtained in two additional experiments. Signal intensities in single blots were measured by a dedicated software to assess the statistical significance of
differences elicited by either 1 and BV02. See Ref. 18 for further details.
Figure 4. Effects of BV02 on c-Abl sub-cellular relocation in Ba/F3 cells expressing T315I Bcr-Abl mutation. See legend to Figure 3 and Ref. 18 for details.
Figure 5. (A) Binding mode of BV02 (cyan sticks) on the structure of 14-3-3r. (B) Superimposition of BV02 and the phosphopeptide (magenta sticks) used to generate the
pharmacophoric model. Some of the residues of the peptide are not shown for clarity.
In summary, BV02 represents the first non-peptidic inhibitor of
4-3-3 proteins, discovered by application of a virtual screening
be used in Bcr-Abl-associated diseases to overcome the resistance
to the traditional drugs.
1
approach. In addition to lack the drawbacks of peptidic drugs,
BV02 belongs to the first generation of 14-3-3 inhibitors that could
Finally, our results further support the idea that in silico virtual
screening is a valuable tool for the identification of new hit/lead

V. Corradi et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6133–6137
10. Polgar, T.; Keseru, G. M. J. Med. Chem. 2005, 48, 3749.
6137
compounds and it should be routinely added to the protocols for
1
1. Wilker, E. W.; Grant, R. A.; Artim, S. C.; Yaffe, M. B. J. Biol. Chem. 2005, 280,
drug discovery. In fact, application of sequential filters (pharmaco-
phores, molecular docking and dynamics) to databases of commer-
cially available compounds allows the researcher to avoid the time
and money consuming step of synthesis and/or to apply testing
procedures less expensive than a high throughput screening.
18891.
1
2. Langer, T.; Wolber, G. Pure Appl. Chem. 2004, 76, 991.
13. Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.; Taylor, R. D. Proteins
003, 52, 609.
2
1
1
1
4. Wang, R.; Lai, L.; Wang, S. J. Comput. Aided Mol. Des. 2002, 16, 11.
5. Grid 22, Molecular Discovery Ltd., Pinner, Middlesex, UK.
6. In vitro clonogenic assay technique: this technique allows to calculate the
lethal dose 50 (LD50) through the reduction of number of colonies (defined as
aggregates containing more than 50 cells generated in 0.9% methylcellulose,
addition with 30% fetal calf serum) in response to scalar doses of drug.
Acknowledgment
This work was in part supported by Fondazione Monte dei
Paschi di Siena.
17. We confirmed BV02 anti-proliferative effect on early hematopoietic
+
progenitors (CD34 ) isolated from CML patients who developed resistance to
1
in consequence of T315I Bcr-Abl mutation. Using indirect immuno-magnetic
+
technique, CD34 cells were isolated from mononuclear cell fractions of bone
marrow samples purchased at Miltenyi Biotec (Miltenyi Biotec, Bergisch
Gladbach, Germany). Their content, measured by cytofluorimetric analysis of
CD34 expression, was greater than 85% in all three cases. Moreover, high-
performance liquid chromatography (D-HPLC) and sequencing let identify Bcr-
Abl point mutation associated with in vivo resistance to 1.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.08.019.
1
8. General method for Western blot or immunoprecipitation (IP)/immunoblotting
7
technique: lysates through sonication from 2 ꢀ 10 whole cells, naked nuclei
References and notes
(
isolated in low salt/detergent buffer: 10 mM NaCl, 5 mM MgCl
2
, 10 mM
0
phosphate buffer and 0.1% NP40) or mitochondrial membranes (isolated by 30
centrifugation at 14,000 rpm in 2 mL lysis buffer: 20 mM Hepes pH 7.5, 1.5 mM
1.
2.
3.
McWhirter, J. R.; Galasso, D. L.; Wang, J. Y. Mol. Cell. Biol. 1993, 13, 7587.
Druker, B. J. Blood 2008, 112, 4808.
Weisberg, E.; Manley, P. W.; Cowan-Jacob, S. W.; Hochhaus, A.; Griffin, J. D. Nat.
Rev. Cancer 2007, 7, 345.
MgCl , 10 mM KCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol and 0.1%
2
PMS) were resolved in SDS–PAGE, transferred in nitrocellulose membranes,
and labeled with primary and secondary antibodies. b-Actin, histone H1 and
cytochrome c were used as internal controls for cytoplasmic, nuclear and
mitochondrial membrane protein loading. Signal intensities from three
repeated experiments were quantified by a GS-700 Imaging densitometer
from BioRad, equipped with a dedicated software (Launch VisionWorks LS,
Upland, CA).
4.
5.
6.
Branford, S.; Melo, J. V.; Hughes, T. P. Blood 2009, 114, 5426.
Schenone, S.; Brullo, C.; Botta, M. Curr. Med. Chem. 2010, 17, 1220.
Yoshida, K.; Yamaguchi, T.; Natsume, T.; Kufe, D.; Miki, Y. Nat. Cell Biol. 2005, 7,
2
78.
7.
8.
9.
Mancini, M.; Veljkovic, N.; Corradi, V.; Zuffa, E.; Corrado, P.; Pagnotta, E.;
Martinelli, G.; Barbieri, E.; Santucci, M. A. Traffic 2009, 10, 637.
Dong, S.; Kang, S.; Lonial, S.; Khoury, H. J.; Viallet, J.; Chen, J. Leukemia 2008, 22,
19. Barilà, D.; Rufini, A.; Condò, I.; Ventura, N.; Dorey, K.; Superti-Furga, G. Mol. Cell.
Biol. 2003, 23, 2790.
20. Raina, D.; Pandey, P.; Ahmad, R.; Bharti, A.; Ren, J.; Kharbanda, S. J. Biol. Chem.
2005, 280, 11147.
5
72.
Wang, B.; Yang, H.; Liu, Y. C.; Jelinek, T.; Zhang, L.; Ruoslahti, E.; Fu, H.
Biochemistry 1999, 38, 12499.