Dual Src and Abl inhibitors target wild type Abl and the AblT315I Imatinib-resistant mutant with different mechanisms

Article abstract of DOI:10.1016/j.bmc.2010.04.024

The tyrosine kinase Src and its close homolog Abl, both play important roles in chronic myelogenous leukemia (CML) progression and Imatinib resistance. No clinically approved inhibitors of the drug-resistant AblT315I exist to date. Here, we present a thorough kinetic analysis of two potent dual Src-Abl inhibitors towards wild type Src and Abl, and the AblT315I mutant. Our results show that the most potent compound BO1 shows only a modest loss of potency (fourfold) towards the AblT315I mutant in vitro and was an ATP-competitive inhibitor of wild type Abl but it acted as a non-competitive inhibitor in the case of AblT315I.

Full text of DOI:10.1016/j.bmc.2010.04.024

Bioorganic & Medicinal Chemistry 18 (2010) 3999–4008
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Dual Src and Abl inhibitors target wild type Abl and the AblT315I
Imatinib-resistant mutant with different mechanisms
Emmanuele Crespan a, Marco Radi b, Samantha Zanoli a, Silvia Schenone c, Maurizio Botta b,
,
*
Giovanni Maga a,
*
a Institute of Molecular Genetics IGM-CNR, via Abbiategrasso 207, 27100 Pavia, Italy
b Dept. of Chemistry and Pharmaceutical Technology, University of Siena, via Banchi di Sotto 55, 53100 Siena, Italy
c Dept. of Pharmaceutical Sciences, University of Genova, Viale Benedetto XV 3, 16132 Genova, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 February 2010
Revised 6 April 2010
Accepted 7 April 2010
Available online 18 April 2010
The tyrosine kinase Src and its close homolog Abl, both play important roles in chronic myelogenous leu-
kemia (CML) progression and Imatinib resistance. No clinically approved inhibitors of the drug-resistant
AblT315I exist to date. Here, we present a thorough kinetic analysis of two potent dual Src-Abl inhibitors
towards wild type Src and Abl, and the AblT315I mutant. Our results show that the most potent com-
pound BO1 shows only a modest loss of potency (fourfold) towards the AblT315I mutant in vitro and
was an ATP-competitive inhibitor of wild type Abl but it acted as a non-competitive inhibitor in the case
of AblT315I.
Keywords:
Src
Abl
Ó 2010 Elsevier Ltd. All rights reserved.
Anticancer chemotherapy
Drug resistance
Enzyme kinetics
Tyrosine kinases
1. Introduction
breast, pancreas, lung, and brain cancers. Studies with Src domi-
nant negative mutants suggested that Src kinases play a role in
Despite the past efforts to develop selective targeted therapies
for the treatment of cancer, the aim has recently turned to find
compounds acting on multiple targets in order to face the drug
resistance often connected to the activation of alternative signaling
pathways.1 Multiple-kinase inhibitors currently approved for can-
cer chemotherapy include Lapatinib (which targets HER1 and
HER2 receptorial kinases) as well as Sorafenib, targeting VEGFR,
RAF and PDGFR kinases. Imatinib itself has been shown to act
not only on Bcr-Abl, but also on KIT and PDGFR kinases.2,3 Target-
ing the chronic myelogenous leukemia (CML)-specific Bcr-Abl
tyrosine kinase has proved to be a successful therapeutic approach.
In fact, the non-receptor tyrosine kinase c-Abl, the cellular counter-
part of Bcr-Abl, even though is normally implicated in various
cellular processes, is not an essential enzyme.4 Recent results
established a functional link between Bcr-Abl and the Src family
tyrosine kinases (SFKs), that play an important role in the cellular
adhesion and motility as well as in the growth, progression, and
metastasis of a variety of human malignancies such as colon,
proliferation of Bcr-Abl expressing cell lines1 and overexpression
of Src kinases is implicated in Bcr-Abl-mediated leukemogenesis
and in Imatinib resistance. Abl shares significant sequence homol-
ogy and remarkable structural resemblance in its active state with
Src family members. For this reason, several ATP-competitive
inhibitors targeting the active conformation of the enzyme (differ-
ently from Imatinib that binds the catalytically inactive form)
originally developed as Src inhibitors, showed to be also potent
Abl inhibitors. For example, Dasatinib (Sprycel™), by Bristol Myers
Squibb, is the first dual Src-Abl inhibitor approved by FDA (June 28,
2006) for the treatment of CML patients with resistance or intoler-
ance to prior therapy, including Imatinib.5
Although these second-generation ‘dual’ tyrosine kinase inhibi-
tors have shown to be clinically effective against most of the Imati-
nib-resistant mutants,6 to date poor results have been obtained in
the treatment of the Bcr-Abl T315I mutant and no clinically
approved T315I inhibitors are still on the market.7 Alternative
approaches for the inhibition of the T315I mutant are of great
pharmacological interest and novel inhibitors are still needed.
By a combination of molecular modeling and combinatorial
chemistry techniques, our research group has recently synthesized
two different class of inhibitors: pyrazolo[3,4-d]pyrimidines
(generic structure 1, Fig. 1) and the 1,3,4-thiadiazole derivatives
* Corresponding authors. Tel.: +39 0382546354; fax: +39 0382422286 (G.M.);
tel.: +39 0577234306; fax: +39 0577232188 (M.B.).
E-mail addresses: botta.maurizio@gmail.com (M. Botta), maga@igm.cnr.it (G.
Maga).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.04.024

4000
E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
Figure 2. General reaction pathway for a two substrates–two products [bi–bi]
sequential reaction. Product [ADP] inhibition pathway is also indicated in the
bottom left part. For details see Section 5.
Figure 1. Generic structure of pyrazolo[3,4-d]pyrimidine inhibitors (1), of 1,3,4-
thiadiazole derivatives (2) and of the compounds studied in this work (3 and 4).
and Kpmep. The kinetic mechanism for this pathway is represented
by Eq. (1) (see Section 5). This general kinetic model is valid with-
out any assumption on the relative order of binding of the reac-
tants to the enzyme. In order to precisely determine the reaction
pathway, we derived the four kinetic constants KAs TP, KmATP, Kspep
and Kpmep) for both substrates of the reaction. The variation of the
reaction velocity was measured at fixed concentrations of one sub-
strate, as a function of increasing concentrations of the other, and
analyzed according to the Michaelis–Menten equation. Figure 3A
and B shows the curves obtained for Src by varying the peptide
and ATP substrates, respectively. Figure 3C and D shows the same
experiments in the presence of Abl. The apparent maximal reaction
rates (Vmaxapp) obtained for each substrate were then plotted as a
function of the corresponding substrate concentrations to obtain
the KAmTP and Kpmep values for Src and Abl (Fig. 4A and B). To obtain
the Ks values for each substrate, the variation of the apparent Km
values (Kmapp) values obtained from the plots shown in Figure
3A–D was studied as a function of the varying substrates. As shown
in Figure 4C and D, in the case of Src, the Kmapp for either substrate
was decreased by increasing concentrations of the other substrate.
Fitting of the curves to (Eqs. (2) and (3)) (see Section 5) allowed the
determination of the Kspep and KAs TP values. In the case of Abl, an
opposite behavior was observed, so that the Kmapp values for one
substrate were increased by increasing concentrations of the other
substrate (Fig. 4E and F). The data were fitted to a linear relation-
ship (Eqs. (4) and (5) in Section 5) and the corresponding Kspep
and KAs TP were derived. The opposite trend of Kmapp values variation
observed for Src and Abl did not reflect any particular mechanism,
since it depended only on the particular combinations of the
microscopic rates contained in the Km and Ks values (Fig. 2).
(generic structure 2, Fig. 1). The pyrazolo[3,4-d]pyrimidines are Src
inhibitors endowed with potent antiproliferative and proapoptotic
activity toward A431 (epidermoid) and 8701-BC (breast cancer)
cell lines overexpressing Src.8,9 In addition, these compounds were
able to inhibit proliferation of three Bcr-Abl-positive human leuke-
mia cell lines (K-562, KU-812, and MEG-01), to reduce Bcr-Abl
tyrosine phosphorylation and to promote apoptosis of Bcr-Abl-
expressing cells through directly inhibition of Abl activity.10
A
model for the T315I mutant was used to identify, within our collec-
tion of pyrazolo[3,4-d]pyrimidines, potential inhibitors of the mu-
tated enzyme. The selected compounds resulted active against wt
Bcr-Abl (Imatinib-sensitive) or three of the most common Imati-
nib-resistant mutants T315I, Y253F, and E255 K.11
The 1,3,4-thiadiazole derivatives were identified applying a
computational protocol, based on docking/molecular dynamics
simulations and on a pharmacophore-based database search. This
study led to the identification of a few hit compounds that were
later optimized to give compound BO1 (3) which showed high po-
tency of inhibition in enzymatic assays and a promising biological
profile on Imatinib-sensitive and Bcr-Abl-independent Imatinib-
resistant leukemia cells, being able to reduce the clonogenic
activity (LD50 = 2.2 l
M) of Bcr-Ab l-expressing clones.12–14 Docking
studies demonstrated that compounds 1 and 2 are ATP-competi-
tive inhibitors and bind to the active conformation of the
enzyme.13,15
As a first step towards a better understanding of the molecular
determinants for the dual-inhibitory activity of these two classes of
compounds against Src and Abl, we have preformed a thorough ki-
netic analysis on the two representative dual Src-Abl inhibitors
BO1 (3) and SI178 (4). The mechanism of inhibition of these two
compounds towards wild type Src and Abl and the T315I Abl mu-
tant has been studied. Our results show that these molecules have
the same mechanism of action on wt Src and Abl. Interestingly, the
most potent derivative BO1 (3) is significantly active towards the
T315I mutated Abl, but with a different mechanism with respect
to Abl wild type. These data will be useful for the development
of novel, more potent dual Src-Abl inhibitors.
2.2. Src and Abl tyrosine kinases follow a random bi–bi reaction
mechanism
Having determined all four equilibrium constants for the bind-
ing of the two substrates to each enzymatic form along the reac-
tion pathway (Fig. 2), their values can be compared to sort out
the reaction mechanism. The calculated values are reported in
Table 1. As can be seen, both substrates show comparable dissoci-
ation constants from the free enzyme (Ks values), as well as from
the enzyme–substrate complex (Km values). A difference exists,
however, between Km and Ks values, with the latter generally lower
than the former. Thus, it does not seem that either Src or Abl bind
their substrates in any particular order, since both ATP and the
peptide have equal chances of combining with either the free
enzyme or the enzyme–substrate complex. The general kinetic
model shown in Eq. (1) predicts that, in the case of a random order
of substrate addition, a relationship exists between the four equi-
librium constants (Eq. (6)), so that any of them could be calculated
2. Results
2.1. Kinetic analysis of the reactions catalyzed by Src and Abl
The tyrosine kinases Src and Abl need to bind two substrates
(ATP and the peptide) before the products (ADP and phosphotyro-
sine peptide) can be released. The general reaction pathway for an
enzyme acting on two substrates is schematically drawn in Figure
2. The microscopic rates of each step are indicated as well as their
relationships with the four equilibrium constants KAs TP, KmATP, Kspep

E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
4001
Figure 3. Kinetic analysis of kinase reactions of Abl and Src. Each reaction was performed as described in Section 5. Values are the means of three independent experiments.
Error bars represent S.D. (A) Variation of the reaction velocity of Src as function of Src peptide substrate concentration at different fixed concentrations of ATP. (B) Variation
of the reaction velocity of Src as function of ATP concentration at different fixed concentrations of Src peptide substrate. (C) Variation of the reaction velocity of Abl as function
of Abl peptide substrate concentration at different fixed concentrations of ATP. (D) Variation of the reaction velocity of Abl as function of ATP concentration at different fixed
concentrations of Abl peptide substrate. (E) Variation of the reaction velocity of AblT315I as function of Abl peptide substrate concentration at different fixed concentrations
of ATP. (F) Variation of the reaction velocity of AblT315I as function of ATP concentration at different fixed concentrations of Abl peptide substrate.
from the other three. To verify whether such a relationship holds
for the values obtained, we calculated each of the four constants
through Eq. (6) and compared the theoretical values with the ones
Table 1
experimentally determined. The calculated values for KAs TP, KmATP
,
Equilibrium binding and steady-state kinetic constants for the interaction of Src and
Abl with their substrates
Kpsep and Kmpep were 17
for Src, and 21 M, 19
l
l
M, 25
M, 25
l
l
M, 11
M, and 18 l
l
M, and 28
lM, respectively,
M, respectively, for Abl.
l
Kpmep
,
l
M
Kpsep
, lM
KAmTP
Srca
,
l
M
KAs TP
, lM
These values are in good agreement (less than 1.5-fold difference)
with the experimentally derived ones shown in Table 1, further
supporting a random order of substrate addition for both enzymes.
30
Abla
33
Abl T315Ia
6
5
14
1
23
20
5
3
13
13
25
1
5
12
2
5
2
1
3
3
2.3. Product inhibition studies of Src and Abl reactions
2
16
According to the mechanism depicted in Figure 2, ADP, one of
the products of the reaction, should be able to compete only with
a
Values are the means of three independent experiments S.D.

4002
E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
Figure 4. Determination of the reaction mechanism of Src and Abl. Data analysis was performed as described in the Section 5. Values are the means of three independent
experiments. Error bars represent S.D. (A) Variation of the Vmaxapp values for Src (triangles) and Abl (circles) reactions, determined as shown in Figure 3, as a function of ATP
concentrations. (B) Variation of the Vmaxapp values for Src [triangles] and Abl [circles], determined as shown in Figure 3, as a function of the peptide substrate concentrations.
(C) Variation of the Kmapp values of Src for the peptide substrate determined as shown in Figure 3, as a function of the ATP substrate concentration. (D) Variation of the Kmapp
values of Src for the ATP substrate, determined as shown in Figure 3, as a function of the peptide substrate concentration. (E) Variation of the Kmapp values of Abl for the
peptide substrate determined as shown in Figure 3, as a function of the ATP substrate concentration. (F) Variation of the Kmapp values of Abl for the ATP substrate, determined
as shown in Figure 3, as a function of the peptide substrate concentration.
its parent substrate (i.e., ATP). In order to verify this, inhibition
experiments were carried out by titrating increasing amounts of
ADP in the presence of a fixed subsaturating concentration of one
substrate and varying the concentration of the other substrate.
The resulting mechanism of inhibition with respect to each sub-
strate of the reaction is reported in Table 2, along with the corre-
sponding apparent ADP dissociation constant (Ki). As can be seen,
ADP was a competitive inhibitor with respect to ATP. These results
indicated that, as expected for a random order equilibrium reac-
tion, the product ADP affects only the binding of the parent sub-
strate, increasing the KsATP and KAmTP values, without affecting the
binding of the peptide. Since the KAs TP value is different from KAmTP
(Table 1), a mixed-type mechanism, rather than a purely non-com-
petitive one, was observed when ADP was tested by varying the
peptide substrate (Table 2).
Table 2
Product inhibition of Src and Abl with respect to ADP
KAi DP
Srca
,
l
M
Variable substrate
Type of inhibition
2.4. The T315I mutation of Abl alters the binding of the
substrates to the enzyme
45
56
5
6
ATP
Peptide
Competitive
Mixed non-competitive
Abla
56
100 10
Next, we analyzed the reaction mechanism of the Imatinib-
resistant Abl mutant T315I. Figure 3E and F shows the primary
plots obtained by varying one substrate in the presence of fixed
amount of the other. From these data, the four kinetic constants
5
ATP
Peptide
Competitive
Mixed
a
Values are the means of three independent experiments S.D.

E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
4003
KAs TP, KAmTP, Kspep and Kmpep were derived as described above. The cal-
culated values are reported in Table 1. Again, they indicate a ran-
dom order mechanism of substrate binding. However, contrary to
Abl wt, the KAmTP and Kpmep values of the T315I mutant were lower
than the corresponding Ks values, indicating that either substrate
showed higher affinity for the enzyme bound to the other substrate
than for the free enzyme. In addition, the Michaelis constant for
ATP binding showed absolute values lower than the corresponding
ones for the Abl wt enzyme, suggesting that the T315I mutant had
higher affinity for the ATP substrate. The T315I mutation partially
occludes an hydrophobic pocket located at the rear of the ATP
binding site. In addition, it apparently stabilizes the activation loop
of the enzyme into an ‘active-like’ conformation. Thus, it is
possible that these structural changes alter the conformation of
the enzyme–substrate complexes with respect to the wild type
enzyme.
2.5. Kinetic analysis of the inhibition of wt Src and Abl by
compounds BO1 (3) and SI178 (4)
The kinetic analysis presented above allowed us to determine a
minimal reaction pathway for the tyrosine kinases Src, Abl and
Figure 5. Kinetic analysis of kinase reactions of Src, Abl and AblT315I in the presence of different concentration of BO1. Each reaction was performed as described in Section 5.
Values are the means of three independent experiments. Error bars represent S.D. (A) Variation of the reaction velocity of Src as function of Src peptide substrate
concentration at different fixed concentrations of BO1. (B) Variation of the reaction velocity of Src as function of ATP concentration at different fixed concentrations of BO1. (C)
Variation of the reaction velocity of Abl as function of Abl peptide substrate concentration at different fixed concentrations of BO1. (D) Variation of the reaction velocity of Abl
as function of ATP concentration at different fixed concentrations of BO1. (E) Variation of the reaction velocity of AblT315I as function of Abl peptide substrate concentration
at different fixed concentrations of BO1. (F) Variation of the reaction velocity of AblT315I as function of ATP concentration at different fixed concentrations of BO1.

4004
E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
AblT315I. These informations were essential for the following
investigation of the mechanism of inhibition of two selected com-
pounds BO1 (3) and SI178 (4) which represent the progenitor of
the two classes of inhibitors developed by our group.
We analyzed the reaction velocity as a function of each sub-
strate of the reaction, holding the other at a fixed subsaturating
amount and in the presence of increasing amounts of the inhibitor
to be tested. As an example, Figure 5 shows the primary plots ob-
tained for the compound BO1 with Src (panels A, B) and Abl (panels
C, D).
an example, Figure 6 shows the results of this analysis for the com-
pound BO1. As can be seen, in the case of Src (Fig. 6A and B) the
Kmapp values for the ATP and the peptide substrate were increased
by the inhibitor, whereas the Vmaxapp values were not affected. In
ATP
mapp
the case of Abl, an increase in the K
was observed with no ef-
fects on the Vmaxapp (Fig. 6C), whereas the inhibitor affected both
pep
mapp
the K
(increased) and the Vmaxapp (decreased) values (Fig. 6D).
The same behavior was observed for the inhibitor SI178 (data
not shown). The calculated Ki values as well as the corresponding
inhibitory mechanisms are listed in Table 3. The compound SI178
resulted sevenfold more active towards Src than Abl. Conversely,
BO1 inhibited Abl fivefold more than Src.
The variations of the apparent Vmax and Km values for each sub-
strate were studied as a function of the inhibitor concentration. As
Figure 6. Determination of the mechanism of inhibition of the compound BO1 towards Abl wild type and the AblT315I mutant. Reactions and data analysis were performed
as described in Section 5. Values are the means of three independent experiments. Error bars represent S.D. (A) Variations of the Kmapp (circles) or Vmaxapp (triangles) values
of Src reaction with ATP as the variable substrate determined as shown in Figure 5, as a function of BO1 concentrations. (B) Variations of the Kmapp (circles) or Vmaxapp
(triangles) values of Src reaction with peptide as the variable substrate determined as shown in Figure 5, as a function of BO1 concentrations. (C) As in panel A, but with the
ATP
pep
Abl enzyme. (D) As in panel B, but with the Abl enzyme. (E) Variation of the K
(circles) or K
(triangles) of the AblT315I enzyme, determined as shown in Figure 5, as a
mapp
mapp
ATP
ATP
function of BO1 concentration. (F) Variation of the V
(circles) or V
(triangles) of the AblT315I enzyme, determined as shown in Figure 5, as a function of BO1
mapp
mapp
concentration.

E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
4005
Table 3
Inhibition of Src and Abl by SI178 and BO1 with respect to the reaction substrates
Inhibitor
Ki,
lM
Variable substrate
Type of inhibition
Effects on binding constants
Molecular target
Srca
SI178
Increases KAs TP, Kspep
Increases KAs TP, Kspep
0.1 0.01
ATP
Peptide
ATP
Competitive
Competitive
Competitive
Competitive
Free enzyme
Free enzyme
BO1
0.55 0.03
Peptide
Ki, ðK0iÞb l
M
Inhibitor
Variable substrate
Type of inhibition
Effects on binding constants
Molecular target
Abla
Increases KAs TP
Increases Kpsep
Decreases Vmax
Increases KAs TP
Increases Kpsep
Decreases Vmax
SI178
0.7 0.1 (3 0.5)
0.75 0.05
ATP
Mixed
Free enzyme; [E:ATP] complex
Free enzyme; [E:pep] complex
Peptide
Competitive
BO1
0.1 0.01
ATP
Competitive
Mixed
0.12 0.01 (0.47 0.1)
Peptide
Inhibitor
Ki,
lM
Variable substrate
Type of inhibition
Effects on binding constants
Molecular target
AblT315Ia
BO1
0.4 0.1
0.45 0.1
ATP
Peptide
Non-competitive
Non-competitive
Decreases Vmax
Decreases Vmax
All forms
All forms
a
Values are the means of three independent experiments S.D.
In the case of mixed-type inhibition. See Figure 6B.
b
2.6. The compounds BO1 (3) and SI178 (4) target different
enzymatic forms in the Src versus Abl reaction pathways
the reaction in dependence of both substrates (Fig. 6F). The calcu-
lated Ki values as well as the corresponding inhibitory mechanism
are listed in Table 3. The proposed reaction equilibria are schemat-
ically drawn in Figure 7C. Against the AblT315I mutant, the com-
pound BO1 acted as a purely non-competitive inhibitor with
respect to both the ATP and peptide substrates, indicating that it
was theoretically able to target the enzyme at any point along
the reaction pathway, irrespectively of the presence of bound sub-
strates. Thus, the T315I mutation induced a dramatic change in the
binding mechanism of the BO1 inhibitor, without significantly
affecting its apparent affinity (Ki), which was decreased only four-
fold with respect to the wild type enzyme.
The data summarized in Table 3 allowed to identify the affected
reaction steps and the enzymatic forms targeted by these inhibi-
tors along the reaction pathway (Fig. 2). The proposed mechanism
of action of the inhibitors is summarized in Figure 7. In the case of
Src, both compounds targeted the free enzyme (Fig. 7A). After the
formation of the enzyme–inhibitor complex, neither ATP, nor the
peptide were anymore able to bind to the enzyme. In the case of
Abl, the situation is more complex (Fig. 7B). The inhibition mecha-
nisms listed in Table 3 suggest that BO1 and SI178 target both the
free enzyme and the enzyme–peptide complex, preventing ATP
binding. As reported above, SI178 and BO1 caused an increase of
the Kmapp values for the ATP and peptide substrates, respectively,
as well as a decrease of the corresponding Vmaxapp values, resulting
in a mixed-type inhibition. According to the reaction scheme
(Fig. 7B), this can be explained by the fact that both inhibitors bind
with higher affinity to the free enzyme than to the enzyme–sub-
strate complex (Ki < K0). The Ki and K0 values were derived accord-
3. Discussion
One major limitation in the effectiveness of Abl-targeted therapy
of chronic myeloid leukemia is the development of resistance to-
wards Imatinib by Abl mutants, notably the T315I variant.3,16,17 This
mutant is not effectively targeted by any of the second-generation
Imatinib derivatives, such as dasatinib, nilotinib, bosutinib and
INNO-406.7,18 Mass spectrometry analysis has revealed that, con-
trary to other mutations such as Y253H and E255V, the T315I substi-
tution induces conformational changes in the Abl structure,
particularly in the active site region (aminoacids 287–302) and in
the SH3 linker domain.19 In accordance with this and other observa-
tions,20 our results show that the T315I mutation alters the relative
affinity of the enzyme for its substrates with respect to the wild type
Abl. In addition, we demonstrate, for the first time, that the apo-en-
zyme mutant form shows lower affinity for both ATP and the pep-
tide, than the corresponding binary complexes. These data suggest
that mutant-specific drugs mimicking the natural substrates should
be better designed on the basis of the enzyme–substrate complex
structure, rather than the unliganded form.
i
i
ing to Eq. (7) for a mixed-type mechanism, by studying the Vmaxapp
and Kmapp variations as a function of the inhibitor concentration,
according to Eqs. (8) and (9). The resulting values are listed in Table
3. As can be seen, BO1 showed lower Ki and K0 values than SI178.
i
Thus, the higher potency of BO1 towards Abl with respect to
SI178 depended from its ability to target both the free enzyme
and the enzyme–substrate complex with higher affinity than
SI178.
2.7. The compound BO1 (3) overcomes the T315I resistance
mutation by altering its equilibrium dissociation constants for
the different enzyme–substrate complexes
The concept of multitargeted anticancer therapy is based on the
possibility to simultaneously inhibit different molecular targets
with one compound, in order to maximize the antiproliferative ef-
fects and minimize the development of drug resistance. The clini-
cally used Abl inhibitor, Imatinib, has been shown to target also
The T315I mutation can induce high level (>100-fold) resistance
towards the clinically approved Abl inhibitor Imatinib. Developing
novel effective inhibitors against this mutant is therefore of great
pharmacological interest. Since BO1 was a more potent inhibitor
of Abl wt with respect to SI178, we analyzed its inhibitory mecha-
nism toward the AblT315I mutant. Primary plots are shown in
Figure 5E, F and analysis of the variations of the Kmapp and Vmaxapp
values for both substrates are shown in Figure 6E and F. As can be
seen, no changes could be observed in the Kmapp values for either
substrate (Fig. 6E), whereas the inhibitor decreased the Vmaxapp of
the tyrosine kinases KIT and PDGFRa
4 allowing its use also against
gastrointestinal tumors and not only in chronic myeloid leukemia
patients. The availability of dual Abl and Src inhibitors will undoubt-
fully prove extremely useful in light of the wider range of tumors
whose proliferation depends on the action of these two kinases.21,22

4006
E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
Figure 7. Different mechanisms of action of the compounds SI178 and BO1 towards Src and Abl. (A) Minimal reaction pathway for the inhibition of Src by SI178 and BO1.
Both compounds act as competitive inhibitors with respect to both substrates (ATP, pep), being able to exclusively bind to the free enzyme with the apparent equilibrium
dissociation constant Ki. (B) Minimal reaction pathway for the inhibition of wild type Abl by SI178 and BO1. Both compounds act as competitive inhibitors towards the ATP
substrate only. This gives the possibility to two additional equilibrium binding steps: one with the free enzyme governed by the equilibrium dissociation constant Ki, and one
with the binary complex of the enzyme and the peptide substrate [E:pep], governed by the equilibrium dissociation constant K0i. Symmetrically, the peptide substrate can
bind to either the free enzyme with an apparent equilibrium dissociation constant Ks(pep) or to the enzyme–inhibitor complex, with apparent equilibrium dissociation
constant K0sðpepÞ. (C) Minimal reaction pathway for the inhibition of T315I Abl by BO1. The compound acts as non-competitive inhibitor with respect to both reactions
substrates. Given the random mechanism of the reaction, the BO1 inhibitor can interact with each four enzymatic forms in the reaction pathway (here represented at the
corners of the bottom square of a cubical three-dimensional space). Symmetrically, both substrates can interact with each enzyme–inhibitor complex (placed at the corners of
the top square of the cube). This results in eight equilibrium binding steps. The assumption here is that the Km and Ks values are equal to the corresponding K0 and Ks0 values.
m
The catalytic rate Kcat is the rate of breakdown of the complex between enzyme and substrates ([E:BO1:pep:ATP]) to give the reaction products, whereas the K0cat rate
represents the breakdown of the complex between enzyme, inhibitor and substrates ([E:BO1:pep:ATP]). The model assumes that K0cat << Kcat. For details see text.
Here we present a detailed enzymological characterization of the
mechanism of action of two potent dual Src-Abl inhibitors. Our re-
sults clearly indicate that the selectivity of inhibition of the two en-
zymes depends on the particular form of the enzyme which is
targeted by the inhibitor along the reaction pathway. In particular,
Src inhibitors which are able to target also the Abl–peptide complex
seem more potent than molecules targeting the Abl–ATP complex.
Finally, we show that the most potent derivative, BO1, can overcome
the structural barrier imposed by the drug-resistant Abl mutant
T315I by virtue of its ability to ‘adapt’ its mechanism of action to
the specific enzymatic form of Abl: BO1 in fact was an ATP-compet-
itive inhibitor of wild type Abl (targeting the free enzyme) while the
same compound proved to act as a non-competitive inhibitor with
respect to both the ATP and peptide substrates, in the case of Ab-

E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
4007
lT315I (targeting all enzymatic forms). It is possible that BO1 acts as
an allosteric inhibitor, not physically preventing ATP binding to the
wild type enzyme, but rather inducing a very fast dissociation of the
substrate from the enzyme–inhibitor complex, thus resulting in an
apparent competitive mechanism. In agreement with this hypothe-
sis, our kinetic data (Table 1) suggest that the structural rearrange-
ment produced by the T315I mutation allows a more stable
binding of the ATP substrate to the enzyme–inhibitor complex. The
resulting quaternary complex [E:BO1:pep:ATP], is either catalyti-
cally inactive or breaks down into products at a very reduced rate
(Fig. 7C). The allosteric site responsible for the activity of BO1 on
the AblT315I mutant is still unknown and it is currently under inves-
tigation. Preliminary docking studies were performed with the
known allosteric sites of Abl: (a) the binding site of myristate (PDB
code: 1OPL); (b) the binding site of the recently reported T315I
inhibitor AP24534 (PDB code: 3IK3) which was shown to exploit
both the ATP pocket and the deep allosteric pocket on the back of
the gatekeeper residue. Unfortunately, docking and molecular
dynamicssimulationdidnotprovideenoughconclusiveinformation
to clearly draw a structure–activity relationship. Further experi-
mental data are needed in order to clearly identify the allosteric
pocket targeted by compound BO1. This is relevant, since most of
dual Src-Abl inhibitors described to date are only moderately effec-
tive against this mutant6,23,24 and the identification of kinase inhib-
itors targeting sites other than the ATP cleft has emerged as a
promising therapeutic option25 as exemplified by the recent discov-
ery of an allosteric Bcr-Abl inhibitor GNF-2.26 It has to be noted that
our BO1 compound described here shows only a modest loss of
potency (fourfold) towards the AblT315I mutant in vitro when
compared to Imatinib (20-fold) or GNF-2 (40-fold), thus making it
a very promising lead compound for the development of AblT315I
effective drugs.
VYK] and in the presence of 0.125 pmol of Src and 0.160 pmol of
32P]ATP. Unlabeled ATP was added to reach the final concentra-
tions as indicated in the figure legends. Abl activity was measured
[c-
in a filter-binding assay using an Abl specific peptide substrate (Ab-
tide, Upstate). Reaction conditions were (in a final volume of 10 ll):
25 mM Tris–HCl pH 7.5, 1 mM DTT, 0.012 lM [c- lM
32P]ATP, 0.022
c-Abl. Unlabeled ATP/Mg++ (1:1 M/M) mix was added to reach the fi-
nal ATP concentrations as indicated in the figure legends. Reactions
were incubated 10 min at 30 °C. The samples (9 ll) were spotted on
papercellulose filterswhichwere washedaccordingto themanufac-
turer’s protocol. Filter-bound radioactivity was measured by liquid
scintillation with a Microbeta-Trilux apparatus (Perkin–Elmer).
5.5. Kinetic analysis
The kinetic model for a sequential bi–bi (two substrates–two
products) reaction mechanism, as depicted in Figure 2, is described
by the equation:
v
¼ Vmax=ð1 þ ðKATP=½ATPꢀÞ þ ðKpmep=½pepꢀÞ
m
þ ðKAs TPÅKpmep=½ATPꢀŽpepꢀÞÞ
ð1Þ
where [ATP] and [pep] are the concentrations of the ATP and pep-
tide substrates, respectively, Vmax is the apparent maximal rate of
the reaction and Km and Ks for both substrates are as defined in
Figure 2.
The KsATP and Kspep were derived from the equations:
Kmapp½pepꢀ ¼ Kpep=ð1 þ ½ATPꢀ=KAs TP
Þ
Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
m
Kmapp½ATPꢀ ¼ KATP=ð1 þ ½pepꢀ=Kpsep
m
Kmapp½pepꢀ ¼ KpepÅð1 þ ½ATPꢀ=KAs TP
Þ
m
4. Conclusion
Kmapp½ATPꢀ ¼ KATPÅð1 þ ½pepꢀ=Kpsep
Þ
m
Besides providing an explanation for the antiproliferative activ-
ity of BO1 against Imatinib-resistant leukemia cells, these data rep-
resent a detailed mechanistic background for the development of a
novel class of dual Src-Abl inhibitors.
where Kmapp are the apparent Michaelis constant (Km) values for
one substrate obtained in the presence of varying concentrations
of the other substrate.
The experimentally calculated values were applied to the
relationship:
5. Experimental section
5.1. Chemicals
KAs TPÅKpmep ¼ KAmTPÅKpsep
ð6Þ
Labeled [c-
32P]ATP was from GE Healthcare. All other chemical
reagents were from Merck and Fluka.
The mixed-type inhibition mechanism observed with Abl, was ana-
lyzed according to the equation:
5.2. Chemistry
v
¼ Vmax=ððKs=½SꢀÞÅð1 þ ½Iꢀ=KiÞ þ ð1 þ ½Iꢀ=K0ÞÞ
ð7Þ
i
where Ki is the apparent dissociation constant of the inhibitor for
the free enzyme and K0i is the apparent dissociation constant of
the inhibitor for the enzyme–substrate complex.
Inhibitory constants for the fully non-competitive (Ki) mecha-
nism observed with AblT315I and the mixed-type (non-competi-
tive term K0i) mechanism observed with Abl, were calculated
from the relationship:
The full synthesis and characterization of compound BO1 and
related analogs have been already reported.8–10 The full synthesis
and characterization of compound SI178 is reported in the Supple-
mentary data.
5.3. Enzymes and proteins
Baculovirus-produced recombinant purified his-tagged active
human Src and Abl and the T315I Abl mutant were purchased from
Upstate (Lake Placid, NY).
Ki ¼ ½Iꢀ=ððVmax=VmaxappÞ ꢁ 1Þ
where Vmaxapp are the apparent maximal reaction rates measured in
the presence of the inhibitor.
ð8Þ
The inhibitory constant (Ki) for the competitive mechanism (ob-
served with Src) and mixed-type (competitive term Ki) mechanism
observed with Abl were calculated from the relationship:
5.4. Enzymatic assays
Src activity was measured in a filter-binding assay using a com-
mercial kit (Src Assay Kit, Upstate), according to the manufacturer’s
protocol, using the specific Src peptide substrate [KVEKIGEGTYGV-
Ki ¼ ½Iꢀ=ððKmapp=KmÞ ꢁ 1Þ
ð9Þ

4008
E. Crespan et al. / Bioorg. Med. Chem. 18 (2010) 3999–4008
8. Carraro, F.; Pucci, A.; Naldini, A.; Schenone, S.; Bruno, O.; Ranise, A.; Bondavalli,
F.; Brullo, C.; Fossa, P.; Menozzi, G.; Mosti, L.; Manetti, F.; Botta, M. J. Med. Chem.
2004, 47, 1595.
9. Carraro, F.; Naldini, A.; Pucci, A.; Locatelli, G. A.; Maga, G.; Schenone, S.;
Bruno, O.; Ranise, A.; Bondavalli, F.; Brullo, C.; Fossa, P.; Menozzi, G.; Mosti,
L.; Modugno, M.; Tintori, C.; Manetti, F.; Botta, M. J. Med. Chem. 2006, 49,
1549.
where Kmapp are the apparent Michaelis constants (Km) for the com-
peting substrate measured in the presence of the inhibitor.
Curves were obtained by non-linear least squares computer fit-
ting of the data to the equations above with the program GraphPad
Prism 3.0.
10. Manetti, F.; Pucci, A.; Magnani, M.; Locatelli, G. A.; Brullo, C.; Naldini, A.;
Schenone, S.; Maga, G.; Carraro, F.; Botta, M. ChemMedChem 2007, 2, 343.
11. Santucci, M. A.; Corradi, V.; Mancini, M.; Manetti, F.; Radi, M.; Schenone, S.;
Botta, M. ChemMedChem 2009, 4, 118.
12. Manetti, F.; Locatelli, G. A.; Maga, G.; Schenone, S.; Modugno, M.; Forli, S.;
Corelli, F.; Botta, M. J. Med. Chem. 2006, 49, 327.
13. Radi, M.; Crespan, E.; Botta, G.; Falchi, F.; Maga, G.; Manetti, F.; Corradi, V.;
Mancini, M.; Santucci, M. A.; Schenone, S.; Botta, M. Bioorg. Med. Chem. Lett.
2008, 18, 1207.
14. Manetti, F.; Falchi, F.; Crespan, E.; Schenone, S.; Maga, G.; Botta, M. Bioorg. Med.
Chem. Lett. 2008, 18, 4328.
Acknowledgements
Work in authors’ laboratories was partially supported by an
AIRC-Investigator Grant 2008 and a National Interest Research
Project Grant PRIN_2007_N7KYCY to G.M. E.C. is the recipient of
a FIRC Fellowship 2008–2010. S.Z. is the recipient of a Fondazione
Buzzati-Traverso Fellowship.
15. Falchi, F.; Manetti, F.; Carraro, F.; Naldini, A.; Maga, G.; Crespan, E.; Schenone,
S.; Bruno, O.; Brullo, C.; Botta, M. ChemMedChem 2009, 4, 976.
16. Gora-Tybor, J.; Robak, T. Curr. Med. Chem. 2008, 15, 3036.
17. Schenone, S.; Bruno, O.; Radi, M.; Botta, M. Med. Res. Rev. 2009 doi:10.1002/
med.20175, in press.
18. Tanaka, R.; Kimura, S. Expert Rev. Anticancer Ther. 2008, 8, 1387.
19. Iacob, R. E.; Pene-Dumitrescu, T.; Zhang, J.; Gray, N. S.; Smithgall, T. E.; Engen, J.
R. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 1386.
Supplementary data
Supplementary data (experimental procedures and compound
characterization) associated with this article can be found, in the
online version, at doi:10.1016/j.bmc.2010.04.024.
20. Griswold, I. J.; MacPartlin, M.; Bumm, T.; Goss, V. L.; O’Hare, T.; Lee, K. A.;
Corbin, A. S.; Stoffregen, E. P.; Smith, C.; Johnson, K.; Moseson, E. M.; Wood, L.
J.; Polakiewicz, R. D.; Druker, B. J.; Deininger, M. W. Mol. Cell. Biol. 2006, 26,
6082.
21. Rucci, N.; Susa, M.; Teti, A. Anticancer Agents Med. Chem. 2008, 8, 342.
22. Li, S. Leuk. Lymphoma 2008, 49, 19.
23. Martinelli, G.; Soverini, S.; Rosti, G.; Baccarani, M. Leukemia 2005, 19, 1872.
24. Quintas-Cardama, A.; Kantarjian, H.; Cortes, J. Future Oncol. 2006, 2, 655.
25. Bogoyevitch, M. A.; Fairlie, D. P. Drug Discovery Today 2007, 12, 622.
26. Zhang, J.; Adrián, F. J.; Jahnke, W.; Cowan-Jacob, S. W.; Li, A. G.; Iacob, R. E.; Sim,
T.; Powers, J.; Dierks, C.; Sun, F.; Guo, G. R.; Ding, Q.; Okram, B.; Choi, Y.;
Wojciechowski, A.; Deng, X.; Liu, G.; Fendrich, G.; Strauss, A.; Vajpai, N.;
Grzesiek, S.; Tuntland, T.; Liu, Y.; Bursulaya, B.; Azam, M.; Manley, P. W.; Engen,
J. R.; Daley, G. Q.; Warmuth, M.; Gray, N. S. Nature 2010, 463, 501.
References and notes
1. Warmuth, M.; Damoiseaux, R.; Liu, Y.; Fabbro, D.; Gray, N. Curr. Pharm. Des.
2003, 9, 2043.
2. Grunwald, V.; Soltau, J.; Ivanyi, P.; Rentschler, J.; Reuter, C.; Drevs, J. Onkologie
2009, 32, 129.
3. Piccaluga, P. P.; Paolini, S.; Martinelli, G. Cancer 2007, 110, 1178.
4. Druker, B. J. Adv. Cancer Res. 2004, 91, 1.
5. Konig, H.; Copland, M.; Chu, S.; Jove, R.; Holyoake, T. L.; Bhatia, R. Cancer Res.
2008, 68, 9624.
6. Schenone, S.; Manetti, F.; Botta, M. Mini Rev. Med. Chem. 2007, 7, 191.
7. Noronha, G.; Cao, J.; Chow, C. P.; Dneprovskaia, E.; Fine, R. M.; Hood, J.; Kang, X.;
Klebansky, B.; Lohse, D.; Mak, C. C.; McPherson, A.; Palanki, M. S.; Pathak, V. P.;
Renick, J.; Soll, R.; Zeng, B. Curr. Top. Med. Chem. 2008, 8, 905.