Substituted 4-amino-1H-pyrazolo[3,4-d]pyrimidines as multi-targeted inhibitors of insulin-like growth factor-1 receptor (IGF1R) and members of ErbB-family receptor kinases

Article abstract of DOI:10.1016/j.bmcl.2010.08.052

This Letter describes the lead discovery, optimization, and biological characterization of a series of substituted 4-amino-1H-pyrazolo[3,4-d] pyrimidines as potent inhibitors of IGF1R, EGFR, and ErbB2. The leading compound 11 showed an IGF1R IC₅₀ of 12 nM, an EGFR (L858R) IC₅₀ of 31 nM, and an ErbB2 IC₅₀ of 11 nM, potent activity in cellular functional and anti-proliferation assays, as well as activity in an in vivo pharmacodynamic assay.

Full text of DOI:10.1016/j.bmcl.2010.08.052

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6067–6071
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Substituted 4-amino-1H-pyrazolo[3,4-d]pyrimidines as multi-targeted
inhibitors of insulin-like growth factor-1 receptor (IGF1R) and members
of ErbB-family receptor kinases
⇑
Gary T. Wang , Robert A. Mantei, Robert D. Hubbard, Julie L. Wilsbacher, Qian Zhang, Lora Tucker,
Xiaoming Hu, Peter Kovar, Eric F. Johnson, Donald J. Osterling, Jennifer Bouska, Jieyi Wang,
Steven K. Davidsen, Randy L. Bell, George S. Sheppard
Global Pharmaceutical Research and Development, Abbott Laboratories, Abbott Park, IL 60064, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
This Letter describes the lead discovery, optimization, and biological characterization of a series of substi-
tuted 4-amino-1H-pyrazolo[3,4-d]pyrimidines as potent inhibitors of IGF1R, EGFR, and ErbB2. The lead-
ing compound 11 showed an IGF1R IC50 of 12 nM, an EGFR (L858R) IC50 of 31 nM, and an ErbB2 IC50 of
11 nM, potent activity in cellular functional and anti-proliferation assays, as well as activity in an
in vivo pharmacodynamic assay.
Received 30 June 2010
Revised 9 August 2010
Accepted 11 August 2010
Available online 13 August 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Multi-targeted kinase inhibitors
IGF1R
EGFR
ErbB2
Insulin-like growth factor-1 receptor (IGF1R) and members of
the ErbB family (e.g., EGFR, ErbB2) are receptor tyrosine kinases
that play significant roles in cancer and have been pursued as drug
targets for cancer therapy.1,2 Several inhibitors of the ErbB-family
signaling have gained regulatory approval for treatment of various
cancers, including trastuzumab (Herceptin™), gefitinib (Iressa™),
erlotinib (Tarceva™), and lapatinib (Tykerb™), while several inhib-
itors of IGF1R are currently undergoing human clinical trials as
anti-cancer agents.3 There is substantial evidence for the existence
of reciprocal cross-talk between the IGF1R axis and ErbB axes of
cellular signaling. Thus, acquired resistance to the EGFR inhibitors
gefitinib and erlotinib or the ErbB2 inhibitor trastuzumab has been
linked to activation of IGF1R signaling through induced overex-
pression of IGF1R or loss of IGF binding proteins.4–6 Conversely,
resistance to the IGF1R inhibitor BMS-536924 has been linked to
ErbB2 signaling.7 Additionally, synergy between IGF1R inhibition
and ErbB inhibition has been demonstrated both in vitro in anti-
proliferation assays and in vivo in xenograft models using small
molecules, antibodies, and siRNA.8–11 These literature data
strongly suggest that molecules targeting both IGF1R and ErbB-
family members may offer superior anti-tumor efficacy and de-
layed onset of resistance in comparison with agents that target
either IGF1R or ErbB-family members alone.
Recently, we have disclosed a series of inhibitors targeting
both IGF1R and ErbB-family members based on 4-amino-1H-pyr-
azolo[3,4-d]pyrimidines bearing 1H-benzo[d]imidazol-5-yl moie-
ties at the C3 position.12 These compounds were derived from
high throughput screening hits.13 Also uncovered in this high
throughput screen were 4-amino-1H-pyrazolo[3,4-d]pyrimidine
and 4-amino-7H-pyrrolo[2,3-d]pyrimidine bearing 4-(benzo[d]
oxazol-2-ylamino)phenyl moieties, compounds 1 and 2 (Table
1). These molecules exhibited modest activity against IGF1R
and weak activity against EGFR and ErbB2. Molecular modeling
suggested that the 4-amino-pyrazolopyrimidine would form the
key H-bonding interaction with the hinge residues of the target
kinases, while the benzoxazole head group would occupy the
hydrophobic back pocket of the typical DFG-out conformation
of these kinases and the cyclohexyl-piperazine tail group would
extend toward the solvent front. Herein, we describe our efforts
to optimize these leads. Our goal was to simultaneously improve
the potency against IGF1R, EGFR and/or ErbB2, while improving
drug-like properties in order to identify candidates suitable for
clinical development.
As depicted in Scheme 1, the synthesis of the target molecules
started with the condensation of 4-bromo-isothiocyanates 3 with
ortho-hydroxy-anilines 4 to afford the 4-(benzo[d]oxazol-2-ylami-
no)phenyl bromides 5 in good yield. These bromides were con-
verted to the corresponding pinacol boronates
standard conditions. Reductive amination of the previously
6 under the
⇑
Corresponding author. Tel.: +1 847 937 2489; fax: +1 847 937 8378.
E-mail address: gary.t.wang@abbott.com (G.T. Wang).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.08.052

6068
G. T. Wang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6067–6071
Table 1
tion at this position (9d) gave rise to over 10-fold enhancement in
Pyrazolopyrimidine and pyrrolopyrimidine screening hits
IGF1R potency and major improvement in EGFR potency. The
ErbB2 potency of 9d also showed major improvement in compari-
son with compounds 1 and 2. Computer modeling of both IGF1R
and EGFR with compound 9d did not provide conclusive rationale
for this observation. Moving the chlorine substitution from C7 (9e)
to C5 (9f) of benzoxazole resulted in loss of potencies. Similarly,
substitution of the benzoxazole at both C7 and C5 positions (9g
and 9h) also reduced potency against all three target enzymes.
Thus, 7-chloro-benzoxazole was found to be optimal based on this
study.
N
N
HN
HN
F
F
O
O
Br
NH2
N
NH2
N
N
N
N
N
N
With the preferred benzoxazole substituent identified, we
expended a significant effort optimizing the tail groups at the
N1-position of the pyrazolopyrimidine core. A variety of linking
moieties and polar capping groups were examined, but saturated
six-membered ring systems were found to be preferred (data not
shown). Synthesis of this sub-class of compounds is described in
Schemes 2 and 3. Suzuki coupling of 7 with 7-chloro-N-(4-
(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)benzo[d]oxa-
zol-2-amine (6, R1 = R2 = R4 = H, R3 = Cl) gave ketone 10. Reductive
amination of 10 gave separable cis and trans amines, including
compounds 11–17. Alternatively, compound 20 was obtained
via Mitsunobu reaction of 18 with t-butyl 4-hydroxy-piperidine-
1-carboxylate, followed by Suzuki coupling and removal of the
Boc-protecting group. Reductive alkylation of 20 furnished com-
pounds 21–23.
The enzyme inhibitory activities of compounds 11–17 and 21–
23 are summarized in Table 3. In comparison with compound 9d
(Table 2), simply changing the substitution geometry of the cyclo-
hexyl spacer from cis (9d) to trans (11) resulted in sixfold improve-
ment in IGF1R potency, 12-fold improvement in EGFR potency and
26-fold improvement in ErbB2 potency. This strong preference for
trans geometry was also evident from the comparison of 13 with
14 and 15 with 16. Changing the methyl substituent on the
terminal piperazine nitrogen of 11 to other groups generally
N
N
N
N
2
1
Compd
Enzyme IC50 (nM)15
EGFR (L858R)
IGF1R
ErbB2
1
2
160
94
2270
1450
4030
>62,500
described ketone 7 gave a mixture of cis/trans isomers of 8 in
approximately 1:1 ratio,14 which can be separated by column chro-
matography or HPLC. Suzuki coupling of 8 with boronates 6 under
microwave conditions gave the target compounds in fair to good
yields.
We first focused our attention on the structure–activity rela-
tionships (SAR) of the substitution on the benzoxazole moiety by
holding the N1-substitution of the pyrazolopyrimidine core con-
stant (Table 2). With reference to compound 9a which has an
unsubstituted benzoxazole head group, introduction of cyano
(9b) or carbomethoxy (9c) at the C7-position of the benzoxazole
moiety resulted in modest improvement in potency against IGF1R
but decrease in potency against EGFR. However, a chloro-substitu-
R1
R1
R4
N
R4
H2N
HO
H
N
N
C
S
c
a, b
+
O
Br
R3
Br
R3
R2
3
R2
4
5
R1
N
R4
H
N
O
O
R3
B
O
R2
6
N
R4
R1 HN
NH2
N
O
I
NH2
N
R3
I
N
NH2
N
N
R2
e
N
d
N
N
N
N
N
N
N
R'
R
O
N R'
R
7
8
9a-h, 24-28
Scheme 1. Reagents and conditions: (a) THF, rt, overnight; (b) LiOH, H2O2, 50–80%; (c) bis(pinacolato)diboron, Pd(dppf)2, dppf, KOAc, dioxane, 80 °C, 40–70%; (d) RR0NH,
HOAc, CH2Cl2, NaBH(OAc)3, 50–75%; (e) 6, Pd(PPh3)4, CSF, DME, MeOH, 150 °C, microwave, 40–75%.

G. T. Wang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6067–6071
6069
Table 2
We briefly investigated the SAR of substitution on the linking
phenyl ring (Table 4). A fluorine placed on this phenyl ring ortho
to the 2-amino-benzoxazole head group gave compound 24 that
was essentially equipotent to the parent compound 11 against all
three target enzymes. A chlorine group placed on this phenyl ring
at the position ortho to the 2-amino-benzoxazole head group (25)
reduced the activity against IGF1R fourfold, whereas compound 27
with the Cl ortho to the pyrazolopyrimidine core was equipotent to
11 against IGF1R. The potency of 25 and 27 against EGFR and ErbB2
was reduced by 2- to 4-fold relative to 11. Substituting the linking
phenyl ring with a methyl group (26 and 28) resulted in greater
loss of activities against all three target enzymes compared to par-
ent analog 11.
SAR of benzoxazole substitution
N
R4
R1 HN
O
R3
NH2
N
N
N
N
N
In addition to measuring the kinase inhibition activities, se-
lected compounds were screened in functional cellular assays for
their ability to inhibit the cellular phosphorylation of the target
receptors. The assays were carried out in ELISA format using Mia-
PaCa cells, a human pancreatic carcinoma cell line which expresses
both IGF1R and EGFR, and BT474, a human breast tumor cell line
which highly expresses ErbB2. The cellular activities of compounds
11 and 24 are shown in Table 5. Compound 11 and close analog 24
showed potent inhibition of IGF1R in MiaPaCa cells and ErbB2
phosphorylation in BT474 cells, but were about 10-fold less potent
in the inhibition of EGFR phosphorylation. The poor potency of the
N
9
Compd
R1
R3
R4
Enzyme IC50 (nM)15
EGFR (L858R)
IGF1R
ErbB2
9a
9b
9c
9d
9e
9f
H
H
H
H
F
F
H
H
H
–CN
–CO2Me
Cl
Cl
H
Cl
Cl
H
H
H
H
H
Cl
810
113
179
77
1660
2810
3590
393
NT
NT
NT
200
1350
NT
5230
NT
72
852
415
348
584
1470
1130
639
9g
9h
CH3
Cl
inhibition of EGFR phosphorylation (EC50 >1 lM) was consistent
for all the analogs tested in this series. Since MiaPaCa cells express
the wild-type EGFR, whereas our EGFR enzyme screening assay
was carried out employing EGFR harboring the L858R activating
mutation, we determined the IC50 of compound 11 against the
wild-type EGFR, to be 2 nM versus 31 nM against EGFR L858R
(Table 3). This result seemingly ruled out the possibility that differ-
NT: not tested.
resulted in reduced potency (13, 15, 17). Replacing the cyclohexyl
spacer with a piperidine moiety also led to decrease in potency
(21–23), particularly against on IGF1R and EGFR potencies.
N
N
HN
HN
O
O
NH2
Cl
NH2
N
I
Cl
NH2
N
a
N
b
N
N
N
N
N
N
N
N
N
N R'
O
O
R
7
10
11-17
Scheme 2. Reagents and conditions: (a) 6 (R1 = R2 = R4 = H, R3 = Cl), Pd(PPh3)4, CSF, DME, MeOH, 150 °C, microwave, 56%; (b) RR0NH, HOAc, CH2Cl2, NaBH(OAc)3, 30–40%.
N
HN
N
O
HN
NH2
N
I
O
Cl
NH2
N
NH2
N
I
N
Cl
NH2
N
N
b, c
N
a
d
N
N
N
N
N
N
N
H
N
N
N
N
Boc
R'
NH
R
18
20
19
21-23
Scheme 3. Reagents and conditions: (a) t-Butyl 4-hydroxy-piperidine-1-carboxylate, Ph3P, DIAD, THF, 58%; (b) 6 (R1 = R2 = R4 = H, R3 = Cl), Pd(PPh3)4, CSF, DME, MeOH, 150 °C,
microwave, 73%; (c) TFA, CH2Cl2, ꢀ95%; (d) RR0C(O), HOAc, CH2Cl2, NaBH(OAc)3, 41–90%.

6070
G. T. Wang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6067–6071
Table 3
Table 5
SAR of the amine tail group
Cellular activities of selected compounds
Compd
Inhibition of receptor phosphorylation IC50 (nM)a
N
HN
pIGF1R
pEGFR
pErbB2
O
11
24
Erlotinib
Lapatinib
PQIP
207
87
2300
2880
51
433
>30,000
340
362
6730
140
>30,000
Cl
NH2
N
>30,000
>30,000
65
N
N
N
a
ELISA assays of receptor auto-phosphorylation, in MiaPaCa cells for pIGF1R and
R
pEGFR and BT474 cells for pErbB2.
Compd
R
Enzyme IC50 (nM)15
EGFR (L858R)
IGF1R
12.5
42.0
45.1
174
ErbB2
7.5
activities of the known EGFR inhibitor erlotinib,16 EGFR/ErbB2
inhibitor lapatinib,16 and IGF1R inhibitor PQIP17 were also mea-
sured using the same cell lines (Table 5).
Compound 11 was chosen for further characterization. The
pharmacokinetic properties of 11 are summarized in Table 6. In
the rat, this compound showed a relatively long t1/2 when dosed
intravenously (iv) as well as a late tmax when dosed orally, but
the oral Cmax was low and the oral bioavailability was a modest
30%. In mouse, the clearance further increased, both the iv t1/2
and the oral tmax significantly decreased, and the oral bioavailabil-
ity decreased to 20%.
Compound 11 was also evaluated in an in vivo pharmacody-
namic (PD) model.18 In this study, female C57BL/6 mice were
dosed with the test compounds or vehicle control. One hour later,
the mice were given IGF-1 intravenously (0.5 mg/kg). The animals
were sacrificed 10 min later, and drug levels in plasma and other
tissues was determined. The IGF1R phosphorylation status in the
homogenized lung tissue was analyzed by Western blot, along
with the phosphorylated Akt (p-Akt) status. As shown in Figure
1, intravenous dosing of compound 11 resulted ꢀ50% inhibition
of pIGF1R at 30 mg/kg and nearly complete inhibition at 60 mg/
kg. Significant inhibition of p-Akt was also observed.
11
12
13
14
31.5
1020
53.7
322
N
N
N
N
N
N
O
N
N
N
334
9.8
62
15
16
17
72.4
424
126
108
21.5
397
21.5
O
N
N
N
N
874
O
1840
N
O
21
22
108
115
99
97
14.5
15.1
N
N
N
N
N
N
23
283
264
13.8
O
Table 6
Pharmacokinetic data of compound 11
Table 4
SAR of the substitution on the linking phenyl ring
Units
Mouse
Rat
IV DNAUCa
IV t1/2
l
h
g-h/mL/mg/kg
0.94
2.4
1.26
8.4
N
R1 HN
O
IV CLp
L/h/kg
1.91
0.19
3.0
0.91
0.38
8.0
PO DNAUCa
PO tmax
PO Cmax
F
l
h
l
%
g-h/mL/mg/kg
Cl
NH2
0.17b
20
0.16c
30
R2
M
N
N
a
b
c
N
N
Dose-normalized area under the curve.
Dosed at 10 mpk.
Dosed at 5 mpk.
N
IGF-1, IV, 0.5 mpk
N
24-28
mg/kg Compd 11
Compd
R1
R2
Enzyme IC50 (nM)
EGFR (L858R)
0
3
10
60
30
IGF1R
ErbB2
p-IGF1R
24
25
26
27
28
F
H
H
H
Cl
7.5
49
196
4.1
22
37
63
156
86
16
29
NT
15.4
478
Cl
CH3
H
IGF1R
p-Akt
H
CH3
530
Akt
ence in the sensitivity of different EGFR constructs to this class of
compounds could account for the observed lack of potency in EGFR
phosphorylation inhibition. For comparison, the functional cellular
Figure 1. In vivo pharmacodynamic assay: inhibition of IGF1R and Akt phosphor-
ylation by compound 11 in C57BL/6 mice.

G. T. Wang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6067–6071
6071
Figure 2. Heat map showing the inhibitory activity of compound 11 against 80 kinases. Pink: IC50 <10 nM; Red: IC50 10–100 nM; Yellow: IC50 100–1000 nM; Gray: IC50
>1000 nM.
6. Guix, M.; Faber, A. C.; Wang, S. E.; Olivares, M. G.; Song, Y.; Qu, S.; Rinehart, C.
The kinase selectivity of compound 11 was assessed against a
panel of 80 kinases and the results are shown in a heat map
R.; Seidel, B.; Yee, D.; Arteaga, C. L.; Engelman, J. A. J. Clin. Invest. 2008, 118,
2609.
(Fig. 2). Besides the intended receptor tyrosine kinases (IGF1R,
IR, EGFR, ErbB2), compound 11 also potently inhibits several of
the Src family of non-receptor tyrosine kinases including Fyn,
Lyn, and Lck. This was not surprising in consideration of litera-
ture reports on related compounds as Src family inhibitors.19
Unexpectedly, the compound also inhibits the Aurora family of
serine–threonine kinases, being particularly potent against Aur-
ora B.
In summary, we have identified a series of substituted 4-amino-
1H-pyrazolo[3,4-d]pyrimidines as potent inhibitors of IGF1R, EGFR,
and ErbB2 receptor kinases. Many analogs showed potent and bal-
anced activity against all three target enzymes, as exemplified by
the leading compound 11 (IGF1R IC50 of 12 nM, EGFR (L858R)
IC50 of 31 nM, and ErbB2 IC50 of 11 nM). In cellular functional as-
says, these compounds potently inhibit the phosphorylation of
7. Haluska, P.; Carboni, J. M.; TenEyck, C.; Attar, R. M.; Hou, X.; Yu, C.; Sagar, M.;
Wong, T. W.; Gottardis, M. M.; Erlichman, C. Mol. Cancer Ther. 2008, 7, 2589.
8. Wilsbacher, J. L.; Zhang, Q.; Tucker, L. A.; Hubbard, R. D.; Sheppard, G. S.;
Bamaung, N. Y.; Fidanze, S. D.; Wang, G. T.; Hu, X.; Davidsen, S. K.; Bell, R. L.;
Wang, J. J. Biol. Chem. 2008, 283, 23712.
9. Charkraborty, A. K.; Liang, K.; DiGiovanna, M. P. Cancer Res. 2008, 68, 1538.
10. Lu, D.; Zhang, H.; Koo, H.; Tonra, J.; Balderes, P.; Prewett, M.; Corcoran, E.;
Mangalampalli, V.; Bassi, R.; Anselma, D.; Patel, D.; Kang, X.; Ludwig, D. L.;
Hicklin, D. J.; Witte, L.; Zhu, Z. J. Biol. Chem. 2005, 280, 19665.
11. Goetsch, L.; Gonzalez, A.; Leger, O.; Beck, A.; Pauwels, P. J.; Haeuw, J. F.; Corvaia,
N. Intl. J. Cancer 2005, 113, 316.
12. Hubbard, R. D.; Bamaung, N. Y.; Fidanze, S. D.; Erickson, S. A.; Palazzo, F.;
Wilsbacher, J. L.; Zhang, Q.; Tucker, L. A.; Hu, X.; Kovar, P.; Osterling, D. J.;
Johnson, E. F.; Bouska, J.; Wang, J.; Davidsen, S. K.; Bell, R. L.; Sheppard, G. S.
Bioorg. Med. Chem. Lett. 2009, 19, 1718.
13. Hubbard, R. D.; Bamaung, N. Y.; Palazzo, F.; Zhang, Q.; Kovar, P.; Osterling, D. J.;
Hu, X.; Wilsbacher, J. L.; Johnson, E. F.; Bouska, J.; Wang, J.; Bell, R. L.; Davidsen,
S. K.; Sheppard, G. S. Bioorg. Med. Chem. Lett. 2007, 17, 5406.
14. (a) Hirst, G. C.; Calderwood, D.; Wishart, N.; Rafferty, P.; Ritter, K.; Arnold, L. D.;
Friedman, M. M. WO 01/19829.; (b) Borhan, D.; Calderwood, D.; Dixon, R. W.;
Hirst, G. C.; Hrnciar, P.; Loew, A.; Leung, A.; Ritter, K. WO 03/020880.
15. The kinase inhibitory activities of the test compounds were measured using a
homogeneous time-resolved fluorescence assay, as described in Ref. 12.
16. Lackey, K. Curr. Top. Med. Chem. 2006, 6, 435.
IGF1R and ErbB2, but not EGFR (EC50 >1 lM). Compound 11 also
showed activity in an in vivo pharmacodynamic assay of IGF1R
inhibition.
17. Ji, Q.; Mulvihill, M. J.; Rosenfeld-Franklin, M.; Cooke, A.; Feng, L.; Mak, G.;
O’Connor, M.; Yao, Y.; Pirritt, C.; Buck, E.; Eyzaguirre, A.; Arnold, L. D.; Gibson,
N. W.; Pachter, J. A. Mol. Cancer Ther. 2007, 6, 2158.
18. Garcia-Echeverria, C.; Pearson, M. A.; Marti, A.; Meyer, T.; Mestan, J.;
Zimmermann, J.; Gao, J.; Brueggen, J.; Capraro, H.-G.; Cozens, R.; Evans, D. B.;
Fabbro, D.; Furet, P.; Porta, D. G.; Liebetanz, J.; Martiny-Baron, G.; Reutz, S.;
Hofmann, F. Cancer Cell 2004, 5, 231.
19. (a) Burchat, A.; Borhani, D. W.; Calderwood, D. J.; Hirst, G. C.; Li, B.; Stachlewitz,
R. F. Bioorg. Med. Chem. Lett. 2006, 16, 118; (b) Bohani, D. W.; Calderwood, D. J.;
Friedman, M. M.; Hirst, G. C.; Li, B.; Leung, K. W.; McRae, B.; Ratnofsky, S.;
Ritter, K.; Waegell, W. Bioorg. Med. Chem. Lett. 2004, 14, 2613.
References and notes
1. (a) Pollak, M. Nat. Rev. Cancer 2008, 8, 915; (b) Sachdev, D.; Yee, D. Mol. Cancer
Ther. 2007, 6, 1.
2. Hynes, N.; Lane, H. Nat. Rev. Cancer 2005, 5, 341.
3. Li, R.; Pourpak, A.; Morris, S. W. J. Med. Chem. 2009, 52, 4981.
4. Jones, H. E.; Goddard, L.; Gee, J. M. W.; Hiscox, S.; Rubini, M.; Barrow, D.;
Knowlden, J. M.; Williams, S.; Walkeling, A. E.; Nicholson, R. I. Endocr. Relat.
Cancer 2004, 11, 793.
5. Nahta, R.; Yuan, L. Y.; Du, Y.; Esteva, F. J. Mol. Cancer Ther. 2007, 6, 667.