Optimization of novel combi-molecules: Identification of balanced and mixed bcr-abl/DNA targeting properties

Article abstract of DOI:10.1016/j.bmcl.2007.05.067

Steps toward the identification of combi-molecules with strong abl tyrosine kinase (TK) inhibitory property and significant DNA damaging potential are described. The optimized combi-molecule 13a was shown to induce approximately twofold stronger abl TK in

Full text of DOI:10.1016/j.bmcl.2007.05.067

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4248–4253
Optimization of novel combi-molecules: Identification
of balanced and mixed bcr-abl/DNA targeting properties
Zakaria Rachid,b, Athanasia Katsoulas,b, Christopher Williams,a Anne-Laure Larroque,b
James McNameeb and Bertrand J. Jean-Claudeb,*
aChemical Computing Group Inc., 1010 Sherbrooke Street West, Suite 910, Montreal, Que., Canada H3A 2R7
bCancer Drug Research Laboratory, Division of Medical Oncology, Department of Medicine, McGill University/Royal
Victoria Hospital, 687 Pine Avenue West Rm. M-719, Montreal, Que., Canada H3A 1A1
Received 24 February 2007; revised 7 May 2007; accepted 11 May 2007
Available online 24 May 2007
Abstract—Steps toward the identification of combi-molecules with strong abl tyrosine kinase (TK) inhibitory property and signif-
icant DNA damaging potential are described. The optimized combi-molecule 13a was shown to induce approximately twofold
stronger abl TK inhibitory activity than GleevecTM and high levels of DNA damage in chronic myelogenous leukemic cells.
Ó 2007 Elsevier Ltd. All rights reserved.
The combi-targeting concept is a novel approach to
drug design that seeks to synthesize compounds termed
‘combi-molecules’ capable of binding to kinases
involved in cell signaling pathways while remaining
DNA reactive.1–10 This principle was developed on the
premise that cancer cells are endowed with multiple
mechanisms to overcome cytotoxic lesions. By inducing
a tandem DNA damage and blockade of kinases that
control signaling pathways related to the repair of the
lesions in the cells, it was expected that the latter could
be less prone to evade the death pathways. The chronic
myelogenous leukemia (CML) is one of the diseases in
which one such tyrosine kinase bcr-abl is known to be
involved in two major cytoprotective events: (a) induc-
tion or activation of DNA repair proteins (e.g.,
Rad51),11,12 (b) activation of the anti-apoptotic PI3K/
AKT pathway.13,14 Thus, we surmised that combi-
molecules designed to block bcr-abl would on one hand
alleviate anti-apoptotic signaling and on the other inflict
cytotoxic DNA lesions: a strategy designed to induce
enhanced cell-killing in CML cells.
Recently we reported on the synthesis of the first models
to demonstrate the feasibility of the approach by design-
ing two types of molecules: (1) those requiring hydro-
lytic cleavage to generate the DNA alkylating moiety
termed ‘type I combi-molecules’,8–10 (2) those that can
damage DNA without requirement for hydrolytic cleav-
age (type II). Unfortunately, the first probes synthesized
to demonstrate the feasibility of the approach showed
disproportionate bcr-abl-DNA targeting properties.
The compound 2 (Scheme 1), a type II combi-molecule,
while being a potent bcr-abl TK inhibitor (although less
potent than GleevecTM) was a weak DNA damaging
compound.5 The compound 13e (Scheme 2), a type I
combi-molecule, while capable of inducing high levels
of DNA damage was a moderate bcr-abl TK inhibi-
tor.4,7 Thus, to demonstrate the advantages of this novel
approach, a structure optimized to show strong bcr-abl
inhibitory potency and fierce DNA damaging potential
was urgently needed. Here we report on the steps toward
the identification of one such structure 13a, a triazene
with balanced and enhanced bcr-abl-DNA targeting
properties.
We first attempted to improve the DNA damaging po-
tential of the type II combi-molecule 2 (because of its
superior abl TK inhibitory activity) by inserting a CH2
group between its mustard moiety and the benzamide
function (see 6, Scheme 1). This is inspired by the fact
that the electron density at the nitrogen of 2 is depleted
by the benzene ring, thereby reducing its alkylating
Keywords: bcr-abl; CML; Triazene; Inhibitors; Chronic myelogenous
leukemia; Pyridopyrimidine; Gleevec; Combi-molecule.
*
Corresponding author. Tel.: +1 514 934 1934x35841; fax: +1 514 843
1475; e-mail: bertrandj.jean-claude@mcgill.ca
These authors contributed equally to this work.
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.05.067

Z. Rachid et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4248–4253
4249
Cl
O
N
NH
NH
O
N
NH
NH2
N
N
N
N
N
CH3
i
CH3
i
Cl
Cl
Cl
O
1
2
N
Cl
Cl
O
Cl
i
N
NH
NH
O
Cl
N
N
N
CH3
Cl
3
Cl
CH3
Cl
ii
N
NH
NH
O
N
H
N
N
OH
CH3
6
N
NH
NH
O
N
N
N
Cl
CH3
Cl
4
CH3
N
OH
iii
N
NH
NH
O
N
N
N
CH3
N
NH
NH
O
N
N
CH3
Gleevec TM
N
5
CH3
N
Cl
Scheme 1. Reagents and conditions: (i) THF/DIPEA (2 equiv)/0 °C; (ii) DMF/KI (1.1 equiv) at room temperature; (iii) POCl3/heat.
power. Thus, we surmised that inserting a CH2 group in
this part of the molecule would lead to a compound with
not only a strong alkylating power, but also because its
structure straddles that of GleevecTM, a strong bcr-abl
inhibitory property. The synthesis of compounds 2, 5,
and 6 proceeded according to Scheme 1. Briefly, amine
1, synthesized as previously described,15 was treated
with p-{N,N-bis(2-chloroethyl)amino}benzoyl chlo-
ride16 or p-{N,N-bis[(2-chloroethyl)aminomethyl]}ben-
zoyl chloride17 to give compounds 2 or 6. Treatment
of 1 with 4-chloromethylbenzoyl chloride gave 3 which
was stirred in DMF with N-methylethanolamine to pro-
vide 4. Chlorination of amino-alcohol 4 with POCl3
gave the hemi-mustard 5 in good yield. Compound 5
was a poor bcr-abl TK inhibitor and inactive against
K562 cell. Likewise, compound 6 was more than 50-fold
less potent than GleevecTM and virtually inactive against
the CML cell line K562 (Table 1).
ing its mustard moiety to occupy a region in the receptor
where minimum steric clash can occur. In contrast, the
CH2 group of 6 protrudes the mustard moiety into an
orientation where its steric bulk can induce clash with
the pocket. Reducing the size of the mustard by creating
a hemi-mustard 5 (Scheme 1) resulted in a compound
with greater affinity for bcr-abl but with a potency
approximately 75-fold less than that of GleevecTM. Thus
the type II strategy involving the bulky mustard group
was abandoned. However, data acquired from this work
allowed us to better understand the mode of binding of
the combi-molecules, most particularly the anchorage of
the benzamide ring into a narrow hydrophobic pocket
and the requirement for a non-bulky DNA damaging
tail. For the latter property, the triazene tail already
demonstrated to induce high levels of DNA damage in
the pyrido-pyrimidine system could now be selected as
the group of choice.7 However, for enhancing the inter-
action of the abl recognition moiety, the challenge
remained.
Molecular modeling was used to account for the failure
of compound 6, the design of which was simply based
on structural overlap with GleevecTM. As depicted in
Figure 1, the phenyl ring of compound 2 (Scheme 1) is
anchored into a hydrophobic pocket of bcr-abl, allow-
Based upon results obtained for 2, we thought it
of interest to enhance hydrophobic interactions of
the benzamide moiety with the hydrophobic pocket.

4250
Z. Rachid et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4248–4253
N
C
NO2
X=CF3, Cl
X
X
7
i
H
N
N
NH2
N
H3C
ii
ClOC
NO2
HO2C
NO2
1
8
9
X
N
iii
H
N
H
N
H
N
H
N
N
O
N
O
N
H3C
N
H3C
iv
X
11, 12
X
10
N
NH2
N
NO2
11a X=H, 11b X=CF3
12 X=Cl
v
H
N
H
N
N
O
N
H3C
13
X
N
N
N
N
R1
R2
13a X=CF3 R1=H R2=CH3
13b X=CF3 R1=CH3 R2=CH3
13c X=CF3 R1=CH3 R2=CH2OH
13d X=Cl R1=H R2=CH3
13e X=H R1=CH3 R2=CH2OH
Scheme 2. Reagents and conditions: (i) HClconcd/heat; (ii) SOCl2, THF/heat; (iii) pyridine(2 equiv)/CH2Cl2/0 °C; (iv) Fe/ethanol/AcOH/heat;
(v) CH3CN/NOBF4/À5 °C/ether then addition of alkylamine correspondent or 1:30 mixture of aqueous methylamine (40%) and aqueous
formaldehyde (37%).
Table 1. Cytotoxicity and Abl tyrosine kinase inhibitory activities by
different compounds
b
Compound
Inhibition of abl
a
IC50,
lM (K562)
kinase IC50
,
lM
GleevecTM
25
0.04
0.22
0.18
2.46
5
3.11
>10
79.72
>100
0.316
68
6
11b
11a
13a
13b
13c
13d
13e7
0.156
>10
0.028
0.018
0.033
0.180
1.057
0.113
0.052
0.187
9.98
Figure 1. Compounds 2 and 6 in 1IEP (X-ray structure of GleevecTM
shown in yellow).
3.15
a IC50 values were determined by using the c-abl tyrosine kinase
protein. Values are means of two separate experiments.
b IC50 values were obtained from the MTT assay.
ring of the benzamide moiety, and (b) a small 1,2,3-tria-
zene tail carrying one or two methyl groups at the
3-position.
The synthesis of compounds 13a–d proceeded according
to an adaptation of a strategy developed for the synthe-
sis of 13e.7 Briefly, commercially available compound 7
was hydrolyzed under acidic condition to give the acid 8
which was chlorinated with SOCl2 to give the benzoyl
chloride 9. Treatment of 1 with the benzoyl chloride 9
in methylene chloride/pyridine (2 equiv) at 0 °C gave
the nitro compounds 10, which were reduced with
Fe in ethanol to provide the amines 11a–b, and 12.
Interestingly, during the course of this development,
Asaki et al.,18coincidentally addressing the same issue,
demonstrated that compounds with small hydrophobic
substituents, for example, Cl and CF3 on the benzamide
moiety could show enhanced binding affinity when com-
pared with GleevecTM. Thus inspired by our results on 2
and the latter development, we designed combi-mole-
cules containing: (a) a Cl and CF3 group on the phenyl

Z. Rachid et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4248–4253
4251
Diazotization of these amines with nitrosonium tetra-
fluoroborate in acetonitrile followed by the addition of
the desired amine and neutralization with triethylamine
gave triazene conjugates 13a–d.19
atoms were added. The ligand coordinates of GleevecTM
from the 1IEP X-ray structure were used as a template
to manually construct starting structures for the com-
pounds 13 compounds in the pocket of the 1IEP crystal
structure. MMFF94x23,24 partial charges were assigned
to all the atoms in the system, and each ligand structure
Interestingly, abl TK enzyme assays showed that
compound 13d containing a 3-Cl group showed potency
similar to that of 2 but 10-fold better than 13e. The
combi-molecules with the more hydrophobic CF3
group, 13a–c, showed abl TK inhibitory potency not
only superior to that of 13e but also to that of GleevecTM.
Compound 13b, the dimethyl analog of 13a, was the
most potent of the series, which we believe is due to
its more hydrophobic dimethyl function. It should be
noted that the latter compound could not be considered
to be a combi-molecule in vitro since like all dimethyl-
triazenes,20,21 it cannot be converted to 13a without
metabolic activation.
˚
was minimized to a gradient of 0.01 kcal/A mol in the
presence of the receptor using the MMFF94x forcefield
in MOE. The receptor atoms were held fixed throughout
all the minimizations. In all cases, the structures mini-
mized to coordinates very similar to the GleevecTM
X-ray coordinates. Furthermore, the minimizations show
that there is sufficient room for the CF3 substituent in
the hydrophobic subpocket. Experimental evidence for
occupation of this pocket is provided by the 2HIW
PDB crystal structure of complexed abl kinase, which
shows a CF3 substituent on the inhibitor occupying this
subpocket.25 For the optimized combi-molecule 13a,
Poisson–Boltzmann electrostatic maps of the entire abl
pocket26 (Fig. 3) show a 3 kcal/mol hydrophobic isocon-
tour in the subpocket, suggesting that hydrophobic con-
tact with this region will yield an approximate 3 kcal/mol
One of the special features of type I combi-molecules is
their ability to not only block their target on their own
but also to further degrade into another metabolite with
bcr-abl TK inhibitory activity that sustains inhibition of
bcr-abl TK. HPLC analysis showed that more than 70%
of 13a was degraded to 11b as early as 3 h after incuba-
tion in cell culture medium at 37 °C. It should be
remembered here that the previous non-optimized com-
bi-molecule 13e degraded to generate the amine 11a
which was found to be virtually inactive against bcr-
abl TK-expressing cells.4,7 In contrast, compound 11b
is 400-fold more potent than the latter amine and only
threefold less potent than GleevecTM. Molecular model-
ing study suggests that the superior potency of 11b
may be mediated as for the combi-molecule by the abil-
ity of the CF3 group to occupy a hydrophobic subpocket
in the ATP binding site of the receptor (Fig. 2). The
1IEP crystal structure of Abl kinase bound to Gleevec22
was used as a starting point to construct compounds 13
analogs in the Abl pocket. The 1IEP PDB file was
loaded into the MOE software package23; all crystallo-
graphic water molecules were deleted and hydogen
Figure 3. Compound 13a docked in 1IEP abl pocket. Green surfaces
reflect the 3 kcal/mol electrostatic energy isocontour for hydrophobic
contacts within the pocket.
N
F
C
N
3
H
N
O
2
H
N
O
H
N
2
N
N
H
N
HN
HN
N
N
CH
CH
3
3
No CF3 group to
occupy subpocket
CF3 tightly occupies subpocke,
increasing binding energy
11a
11b
Figure 2. Occupation of 1IEP subpocket by ortho CF3 group of 11b.

4252
Z. Rachid et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4248–4253
gain in ligand binding energy. Indeed, this predicted
3 kcal/mol binding energy gain correlates well (Table
2) with the increase in affinity observed between pairs
of compounds that differ only by CF3 substituent [11a/
11b and 13e/13c].
Acknowledgments
We thank the Leukemia and Lymphoma society for
financial support. A. Katsoulas is grateful to the Fonds
´ ´
de la Recherche en Sante du Quebec (FRSQ) for a doc-
toral studentship.
In order to determine the translation of the now
proportionate binary targeting property of the
combi-molecules, we studied their potency against
the bcr-abl expressing CML cells K562. The results
showed that 13a is twofold more potent than
GleevecTM and fivefold more potent than 11b. Further-
more, its DNA damaging potential was demonstrated
by the comet assay. Interestingly, compound 13a
induced significant levels of DNA damage in the
Mo7p210 transfected cells (Fig. 4).
References and notes
1. Qiu, Q.; Domarkas, J.; Banerjee, R.; Merayo, N.; Brahimi,
F.; McNamee, J. P.; Gibbs, B. F.; Jean-Claude, B. J. Clin.
Cancer Res. 2007, 13, 331.
2. Qiu, Q.; Domarkas, J.; Banerjee, R.; Katsoulas, A.;
McNamee, J. P.; Jean-Claude, B. J. Anti-Cancer Drugs
2007, 18, 171.
3. Domarkas, J.; Dudouit, F.; Williams, C.; Qiyu, Q.;
Banerjee, R.; Brahimi, F.; Jean-Claude, B. J. J. Med.
Chem. 2006, 49, 3544.
4. Katsoulas, A.; Rachid, Z.; Brahimi, F.; McNamee, J.;
Jean-Claude, B. J. Leuk. Res. 2005, 29, 693.
5. Katsoulas, A.; Rachid, Z.; Brahimi, F.; McNamee, J.;
Jean-Claude, B. J. Leuk. Res. 2005, 29, 565.
6. Banerjee, R.; Rachid, Z.; McNamee, J.; Jean-Claude, B. J.
J. Med. Chem. 2003, 46, 5546.
7. Rachid, Z.; Katsoulas, A.; Brahimi, F.; Jean-Claude, B. J.
Bioorg. Med. Chem. Lett. 2003, 13, 3297.
8. Matheson, S. L.; McNamee, J. P.; Jean-Claude, B. J.
Cancer Chemother. Pharmacol. 2003, 51, 11.
This study conclusively demonstrated that the potency
of the combi-molecules could be refined by enhancing
the hydrophobicity of the benzamide triazene carrier
with a binding mode similar to that of GleevecTM. The
now augmented abl TK inhibitory potency (activity
superior to GleevecTM) has conferred to these Type I
combi-molecules of the triazene class, strong and bal-
anced bcr-abl/DNA targeting properties. To our knowl-
edge, this is the first report on such a combi-molecule
with balanced bcr-abl/DNA targeting properties. A
complete study on its biological mechanism of action
(blockade of abl phosphorylation in whole cells, analysis
of the DNA damage response pathways p53, p21, induc-
tion of apoptosis (Bax), and comparative analysis with
individual combinations of bcr-abl + DNA damaging
agents will be reported elsewhere.
9. Qiu, Q.; Dudouit, F.; Matheson, S. L.; Brahimi, F.;
Banerjee, R.; McNamee, J. P.; Jean-Claude, B. J. Cancer
Chemother. Pharmacol. 2003, 51, 1.
10. Brahimi, F.; Matheson, S. L.; Dudouit, F.; McNamee, J.
P.; Tari, A. M.; Jean-Claude, B. J. J. Pharmacol. Exp.
Ther. 2002, 303, 238.
11. Slupianek, A.; Schmutte, C.; Tombline, G.; Nieborowska-
Skorska, M.; Hoser, G.; Nowicki, M. O.; Pierce, A. J.;
Fishel, R.; Skorski, T. Mol. Cell 2001, 8, 795.
12. Takeda, N.; Shibuya, M.; Maru, Y. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 203.
13. Kawauchi, K.; Ogasawara, T.; Yasuyama, M.; Ohkawa,
S.-i. Blood Cells Mol. Dis. 2003, 31, 11.
14. Skorski, T.; Bellacosa, A.; Nieborowska-Skorska, M.;
Majewski, M.; Martinez, R.; Choi, J. K.; Trotta, R.;
Wlodarski, P.; Perrotti, D.; Chan, T. O.; Wasik, M. A.;
Tsichlis, P. N.; Calabretta, B. EMBO J. 1997, 16, 6151.
15. Zimmermann, J.; Buchdunger, E.; Mett, H.; Meyer, T.;
Lydon, N. B. Bioorg. Med. Chem. Lett. 1997, 7, 187.
16. Elderfield, R. C.; Liao, T.-K. J. Org. Chem. 1961, 26,
4996.
Table 2. Binding energy gains from –CF3 substitution on benzamide
moiety in GleevecTM analogs
Compound
pair
Inhibition of abl kinase
IC50, lM
DDGa
(kcal/mol)
X = –H
67
1.1
X = CF3
0.156
0.033
11a/11b
13e/13c
3.6
2.1
a Binding energy gains estimated from experimental IC50 values using
27
) .
DDG = Àrt ln(IC50CF3/IC50H
17. Hatono, S.; Yazaki, A.; Yokomoto, M.; Hirao, Y. 87-
307846 259185, 1988, 26 pp.
18. Asaki, T.; Sugiyama, Y.; Hamamoto, T.; Higashioka, M.;
Umehara, M.; Naito, H.; Niwa, T. Bioorg. Med. Chem.
Lett. 2006, 16, 1421.
19. Experimental data for 13a. 1-{N-[4-Methyl-3-(4-pyridin-3-
yl-pyrimidin-2-ylamino)-phenyl]-3-tri fluoromethyl-benz-
amide}-methyl triazene. The amine 11b (100 mg,
0.22 mmol) was dissolved in dry acetonitrile (10 mL) and
after cooling to À5 °C, a solution of nitrosonium tetra-
fluoroborate (50 mg, 0.5 mmol) suspended in acetonitrile
was added. The clear solution was stirred for 1 h at À5 °C
and a mixture of ether (10 mL), water (2 mL), triethyl-
amine (0.43 mL) and aqueous methylamine (40%)
(0.2 mL) was added dropwise. The mixture was subse-
quently stirred at 0 °C for 15 min, the reaction was
extracted with ethyl acetate dried then evaporated and
the precipitate that formed triturated and washed with
Figure 4. Quantitation of DNA damage using the alkaline comet
assay. DNA damage induced by 13a, 11b and Temozolomide (TEM) in
the M07/p210 cell line. Tail moment was used as a parameter for the
detection of DNA damage in M07/p210 cells exposed to the drugs
30 min.

Z. Rachid et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4248–4253
4253
ether to give 13a (50 mg, 42%) as a pure solid. ESI m/z
529.0 (MNa)+; 1H NMR (400 MHz, DMSO-d6) d 11.30
(br s, 1H, N@N–NMeH), 10.33 (s, 1H, NH), 9.27 (s, 1H,
ArH), 8.99 (s, 1H, ArH), 8.67 (d, 1H, J = 4.8 Hz, ArH),
8.51–8.46 (m, 2H, ArH), 8.26 (s, 1H, ArH), 8.18 (d, 1H,
J = 12.4 Hz, ArH), 8.06 (s, 1H, ArH), 7.64 (d, 1H,
J = 11.2, J = 11.2 Hz, ArH), 7.53–7.42 (m, 3H, ArH), 7.21
(d, 1H, J = 11.2, ArH), 3.07 (s, 3H, NHCH3), 2.22 (s, 3H,
CH3); 13C NMR (75 MHz, DMSO-d6) 168.7, 164.5, 162.3,
161.9, 160.2, 152.1, 151.8, 148.9, 138.5, 138.4, 137.7, 135.1,
133.2, 132.9, 131.4, 130.8, 128.5, 126.6, 124.5, 118.1, 118.0,
117.6, 108.2, 31.4, 18.3.
22. Nagar, B.; Bornmann, W. G.; Pellicena, P.; Schindler, T.;
Veach, D. R.; Miller, W. T.; Clarkson, B.; Kuriyan, J.
Cancer Res. 2002, 62, 4236.
23. MOE software (Version 2006.07) available from Chemical
Computing Group Inc., S.S.W., Montreal, Quebec, Can-
ada. www.chemcomp.com.
24. Halgren, T. A. J. Comput. Chem. 1999, 20, 720.
25. Okram, B.; Nagle, A.; Adrian, F. J.; Lee, C.; Ren, P.;
Wang, X.; Sim, T.; Xie, Y.; Wang, X.; Xia, G.; Spraggon,
G.; Warmuth, M.; Liu, Y.; Gray, N. S. Chem. Biol. 2006,
13, 779 .
26. Non-linear Poisson–Boltzmann calculation as imple-
mented in the MOE software package (Version 2006.08)
available from Chemical Computing Group Inc.; 1010
20. Manning, H. W.; Cameron, L. M.; LaFrance, R. J.;
Vaughan, K.; Rajaraman, R. Anti-Cancer Drug Des. 1985,
1, 37.
21. Cameron, L. M.; LaFrance, R. J.; Hemens, C. M.;
Vaughan, K.; Rajaraman, R.; Chubb, D. C.; Goddard,
P. M. Anti-Cancer Drug Des. 1985, 1, 27.
Sherbrooke
www.chemcomp.com.
St.
West,
M.,
Quebec,
Canada
27. The Practice of Medicinal Chemistry, 2nd ed.; Wermuth,
C. G., Ed.; 2003.