Discovery of hybrid Hsp90 inhibitors and their anti-neoplastic effects against gefitinib-resistant non-small cell lung cancer (NSCLC)

Article abstract of DOI:10.1016/j.bmcl.2013.11.034

Heat shock protein 90 (Hsp90) represents an attractive cancer therapeutic target due to its role in the stabilization and maturation of many oncogenic proteins. We have designed a series of hybrid Hsp90 inhibitors by connecting the resorcinol ring of VER-

Full text of DOI:10.1016/j.bmcl.2013.11.034

Bioorganic & Medicinal Chemistry Letters 24 (2014) 224–227
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of hybrid Hsp90 inhibitors and their anti-neoplastic effects
against gefitinib-resistant non-small cell lung cancer (NSCLC)
Chul-Ho Jeong a,b, Hee Baek Park b, Won Jun Jang a, Su Hyun Jung b, Young Ho Seo a,b,
⇑
a College of Pharmacy, Keimyung University, Daegu 704-701, South Korea
b Institute for New Drug Development, Keimyung University, Daegu 704-701, South Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 July 2013
Revised 29 October 2013
Accepted 15 November 2013
Available online 22 November 2013
Heat shock protein 90 (Hsp90) represents an attractive cancer therapeutic target due to its role in the sta-
bilization and maturation of many oncogenic proteins. We have designed a series of hybrid Hsp90 inhib-
itors by connecting the resorcinol ring of VER-49009 (2) and the trimethoxyphenyl ring of PU3 (3) using
structure-based approach. Subsequent testing established that compound 1f inhibited gefitinib-resistant
H1975 cell proliferation, brought about the degradation of Hsp90 client proteins including EGFR, Met,
Her2 and Akt and induced the expression of Hsp70. The design, synthesis, and evaluation of 1f are
described herein.
Keywords:
Hsp90
Lung cancer
Gefitinib resistance
Inhibitor
Ó 2013 Elsevier Ltd. All rights reserved.
Rational design
Conventional drug design embraces the ‘one gene, one drug, one
disease’ philosophy. Over the past two decades, several targeted
cancer drugs, including Gleevec, Iressa and Herceptin have been
discovered to eradicate tumors in more specific ways and reduce
the harmful nonspecific side effects of chemotherapeutics. How-
ever, this notion is being challenged by the occurrence of drug resis-
tance. The tumor cells outsmart single-targeted drugs to escape
from their destiny by mutating targeted proteins, down-regulating
death signals, or up-regulating survival pathways. In this regard, it
is being recognized that single-target drugs can be problematic and
multi-target drugs have emerged as a new paradigm to overcome
the resistance in drug discovery.1,2 Alternatively, to elucidate a sin-
gle protein, so called ‘nodal’ protein that integrates multiple signal-
ing pathways and discover an inhibitor against nodal proteins may
be best suited to overcome the genetic and molecular heterogeneity
of progressive disease through simultaneously interrupting multi-
ple mechanisms of tumor maintenance.
Heat shock protein 90 (Hsp90) is a cancer nodal protein and has
become an attractive therapeutic target in cancer research. Hsp90
is ATP dependent molecular chaperone that is responsible for the
stabilization and maturation of their substrate proteins, referred
to as ‘client’ proteins. Disruption of Hsp90 chaperone activity in-
duces client proteins degradation via the ubiquitin–proteasome
pathway, which can ultimately lead to cell death. Many Hsp90
MMP2 play significant roles in six essential hallmarks of a cancer
cell.3–5 More interestingly, Hsp90 is constitutively expressed at
2–10 fold higher levels in tumor cells compared to their normal
counterparts and Hsp90 inhibitors demonstrate selective anti-pro-
liferative effects toward cancer cells as compared to normal cells,
due to the greater dependence of tumor cells on Hsp90’s chaperon-
ing function against oncogenic stressors in the hostile hypoxic, aci-
dotic and nutrient-deprived microenvironment.6,7
The natural product geldanamycin was first identified as an
Hsp90 inhibitor in 1994.8 Since then, a number of natural products
and synthetic small molecules that target Hsp90 have been discov-
ered for the treatment of cancerous diseases, which include radic-
icol9, VER-49009,10 and PU311 (Fig. 1). Despite of these advances,
none of Hsp90 inhibitors are clinically approved as an anti-cancer
chemotherapy until now, and there still remains a need for the dis-
covery of a novel class of small molecule inhibitors against Hsp90.
Here, we report the design, synthesis, and anti-cancer effects of a
new class of Hsp90 inhibitors.
Structural analysis of Hsp90 revealed that ATP-binding pocket
of Hsp90 consisted of a hydrophilic region and a hydrophobic re-
gion (Fig. 2). Co-crystal structure of VER-4900910 (2) bound to
the N-terminal ATP-binding region of Hsp90 demonstrated that
the resorcinol ring of 2 positioned in the hydrophilic region of
the pocket. The hydrophilic region of the pocket consisted of
Asp93 and Asp54 residues, which typically interact with the ade-
nine ring. The crystal structure of PU311 (3) indicated that the
3,4,5-trimethoxyphenyl ring of PU3 (3) was located in the opposite
client proteins, including Her2, Met, Cdk4, Akt, HIF-1
a and
⇑
Corresponding author.
E-mail address: seoyho@kmu.ac.kr (Y.H. Seo).
orientation of the resorcinol ring of VER-49009 (2). The
p-rich
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.11.034

C.-H. Jeong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 224–227
225
H
O
O
O
O
H
N
Cl
Cl
Br
N
SO2Cl2
49%
CONHEt
OMe
H
Cl
99%
HO
OH
HO
OH
AllylO
O
OAllyl
O
Cl
O
5
6
7
HO
OH
O
R'1
R'4
Cl
AllylO
OHC
R'2
R'1
VER-49009 (2)
HO
OH
KOH
H2O-EtOH
R'2
R'3
7
+
Radicicol (4)
O
NH2
R'3
Allyl
O
N
N
9a g
-
Cl
8a-g
R'4
N
N
OMe
OMe
HO
OH
OMe
OMe
nBu
1f
MeO
PU3 (3)
OMe
Figure 1. Structures of known Hsp90 inhibitors and 1f.
Scheme 1. Claisen–Schmidt aldol condensation of acetophenone 7 with aldehyde
8a–g.
hand, Claisen–Schmidt aldol condensation of ketone 7 with the
corresponding aromatic aldehydes 8a–g was carried out in the
presence of KOH in MeOH–H2O. The condensation reaction suc-
cessfully provided enones 9a–g in 42–80% yield. Finally, removal
of allyl-protecting groups, using PdCl2(PPh3)2 and ammonium for-
mate under microwave irradiation produced the resulting 3-
chloro-2,4-dihydroxychalcones (1a–g) (Scheme 2).20
Figure 2. X-ray crystal structure of apo-Hsp90 (PDB code: 1UYM).
To investigate the effects of newly synthesized compounds (1a–
g) for Hsp90 inhibition, we first screened the efficacy of these com-
pounds by measuring anti-proliferative effects against H1975 cell
line. H1975 is a gefitinib-resistant non-small cell lung cancer cell
line and its resistance is mediated by ‘gatekeeper’ the mutation
T790M-EGFR in combination with L858R,15–17 The resistance is
also related to Met amplification, compensating for the loss of
EGFR signals.18 To test anti-proliferative activities of compounds
(1a–g) against gefitinib-resistant H1975, we treated H1975 cells
with various concentrations of compounds (1a–g) and measured
amino acids of Phe138, Trp162, and Tyr139 formed the hydropho-
bic region of the pocket and the trimethoxyphenyl ring of PU3 (3)
had p–p interaction with those residues. With the aim of maximiz-
ing interactions in the ATP binding pocket of Hsp90, we intended
to design a molecule by hybridizing the resorcinol ring of VER-
49009 (2) and the 3,4,5-trimethoxyphenyl ring of PU3 (3). To con-
nect the 3,4,5-trimethoxyphenyl ring to the resorcinol ring, we
decided to use a chalcone scaffold as a core template. Chalcones
are abundant natural products in edible plants such as green tea
and exhibit a wide spectrum of biological activities including
anti-tumor activities.12,13 Accordingly, chalcones are an important
class of molecules and speculated as promising candidates as anti-
cancer agents. Besides, our recent study has also demonstrated
that a natural product, licochalcone A disrupts Hsp90 chaperoning
function.14 With a hybrid inhibitor embedded in the chalcone scaf-
fold, we envision that the resorcinol ring of the inhibitor would
make a hydrogen bond with Asp93 involved in ATP binding pocket
and the hydrophobic trimethoxyphenyl group of the inhibitor
O
Cl
PdCl2(PPh3)2
R1
R4
9a-g
R2
R3
NH4HCO2
THF, µw
HO
OH
1a-g
would project into the p
-rich lipophilic cavity of the pocket.15,16
The synthesis of compounds (1a–g) began with the preparation
of 6-chloro-2,4-dihydroxyacetophenone (6) (Scheme 1). Treatment
of 2,4-dihydroxyacetophenone (5) with sulfuryl chloride provided
a chlorinated product 6 as well as its undesired regio-isomer of
3-chloro-2,4-dihydroxyacetophenone in 1:1 molar ratio. After
carefully being resolved by silica gel chromatography, the protec-
tion reaction of 2,4-dihydroxyl groups of compound (6) was carried
out. Initial studies to protect 2,4-dihydroxyl groups of 6 with TBSCl
or MOMCl using combinations of various bases (TEA, K2CO3, and
DBU) and solvents (CH2Cl2 and DMF) were not fruitful but only
to produce a single protected adduct. Consequently, compound 6
was protected with allyl bromide in the presence of K2CO3 to fur-
nish the allyl-protected ketone 7 in 99% yield. With ketone 7 in
Scheme 2. Removal of allyl protecting groups.

226
C.-H. Jeong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 224–227
Figure 3. (a) Comparative effects of 1d, 1e, and 1f on cell viability of H1975. Cells
were treated for 3 days at the indicated concentrations of each compound. Cell
viability was measured by MTS assay. (b) Anti-proliferative effect of 1f on H1975
cells. Cell proliferation was determined at 1, 2, and 3 days using MTS assay at the
indicated concentrations of each compound. Data are presented as mean SD
(n = 4).
Figure 5. Dose-dependent effect of compound 1f on cellular biomarkers of Hsp90
inhibition and their downstream signaling. H1975 Cells were treated for 24 h with
the indicated concentration of 1f and the expressions of cellular biomarker proteins
of Hsp90 inhibition and their downstream signaling were analyzed by Western blot.
Geldanamycin (GA, 1
lM) and DMSO (D) were employed, respectively, as positive
and negative controls.
treated with increasing concentrations of compound 1f for 24 h,
and then the expression levels of Hsp90’s client proteins and the
downstream signaling proteins were assessed. As shown in Fig-
ure 5, compound 1f resulted in dose-dependent downregulation
of Hsp90’s client proteins including EGFR, Her2, Met and Akt and
up-regulation of Hsp70. In contract, the expression level of Erk1/
2 was not depleted, in that Erk1/2 is not a client protein of
Hsp90. Moreover, the phosphorylation status of EGFR was reduced
according to the increasing concentrations of compound 1f. Consis-
tently, downstream signaling pathway including the phosphoryla-
tion of Akt and Erk1/2 was effectively abolished. Collectively, these
results indicate that compound 1f circumvents gefitinib-resistance
by the blockade of EGFR and Met signaling through Hsp90 inhibi-
tion, in that gefitinib-resistance in NSCLC is associated with the
activation of signaling pathways by the mutation of EGFR T790M
and the amplification and activation of Met tyrosine kinase recep-
tor.15–18
To investigate the binding pose of 1f in the Hsp90 active site,
molecular docking studies were performed using the human-
Hsp90 crystal structure (PDB code: 1UYM). Hybrid inhibitor 1f
was docked with the 3D coordinates of the Hsp90’s N-terminal do-
main using Autodock 4.2 (Molecular Graphics Laboratory). In silico
modeling demonstrated that that hybrid inhibitor 1f bound
N-terminal ATP binding pocket of Hsp90 as expected (Fig. 6). The
resorcinol ring of 1f binds to the hydrophilic region of the pocket
Figure 4. Effects of compounds 1a–g on cellular biomarkers of Hsp90 inhibition.
H1975 cells were treated for 24 h with the indicated compound (40
expression of the Hsp90’s client proteins was analyzed by Western blot. Geldana-
mycin (GA, 1 M) and DMSO (D) were employed, respectively, as positive and
lM) and the
l
negative controls.
the viability using MTS assay. The result revealed that compounds
1d, 1e and 1f effectively inhibited cell proliferation of H1975 by
dose-dependent manners (Fig. 3), whereas others did not (data
not shown).
To further determine whether the observed viability was re-
lated to Hsp90 inhibition, compounds (1a–g) were incubated with
H1975 cells and screened by analyzing the expression level of
Hsp90’s clients, EGFR, Her2, Met and Akt along with Hsp70
(Fig. 4). The molecular hallmark of Hsp90 inhibition includes the
proteosomal degradation of Hsp90 client proteins and the tran-
scriptional upregulation of Hsp70.11 As expected, compound 1f
among others revealed a robust degradation of EGFR, Her2, Met
and Akt and up-regulated cochaperone Hsp70, suggesting that
compound 1f targets the Hsp90 protein folding machinery.
To precisely determine the degradation of Hsp90’s client pro-
teins and their downstream signaling by 1f, H1975 cells were

C.-H. Jeong et al. / Bioorg. Med. Chem. Lett. 24 (2014) 224–227
227
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2013.1
1.034.
References and notes
1. Petrelli, A.; Giordano, S. Curr. Med. Chem. 2008, 15, 422.
2. Boran, A. D.; Iyengar, R. Curr. Opin. Drug Discovery Dev. 2010, 13, 297.
3. Mahalingam, D.; Swords, R.; Carew, J. S.; Nawrocki, S. T.; Bhalla, K.; Giles, F. J. Br.
J. Cancer 2009, 100, 1523.
4. Hanahan, D.; Weinberg, R. A. Cell 2011, 144, 646.
5. Hanahan, D.; Weinberg, R. A. Cell 2000, 100, 57.
6. Ferrarini, M.; Heltai, S.; Zocchi, M. R.; Rugarli, C. Int. J. Cancer 1992, 51, 613.
7. Whitesell, L.; Lindquist, S. L. Nat. Rev. Cancer 2005, 5, 761.
8. Whitesell, L.; Mimnaugh, E. G.; De Costa, B.; Myers, C. E.; Neckers, L. M. Proc.
Nat. Acad. Sci. U.S.A. 1994, 91, 8324.
9. Sharma, S. V.; Agatsuma, T.; Nakano, H. Oncogene 1998, 16, 2639.
10. Dymock, B. W.; Barril, X.; Brough, P. A.; Cansfield, J. E.; Massey, A.; McDonald,
E.; Hubbard, R. E.; Surgenor, A.; Roughley, S. D.; Webb, P.; Workman, P.;
Wright, L.; Drysdale, M. J. J. Med. Chem. 2005, 48, 4212.
Figure 6. Molecular docking model of 1f with Hsp90 (PDB code: 1UYM). The
carbon, oxygen, hydrogen atoms of 1f are illustrated in blue, red, and gray,
respectively. The side chains of binding site are colored by atom types (carbon,
gray; nitrogen, blue; oxygen, red) and labeled with their residue name. Hydrogen
bonds are shown in dashed red lines.
11. Wright, L.; Barril, X.; Dymock, B.; Sheridan, L.; Surgenor, A.; Beswick, M.;
Drysdale, M.; Collier, A.; Massey, A.; Davies, N.; Fink, A.; Fromont, C.; Aherne,
W.; Boxall, K.; Sharp, S.; Workman, P.; Hubbard, R. E. Chem. Biol. 2004, 11, 775.
12. Batovska, D. I.; Todorova, I. T. Curr. Clin. Pharmacol. 2010, 5, 1.
13. Dimmock, J. R.; Elias, D. W.; Beazely, M. A.; Kandepu, N. M. Curr. Med. Chem.
1999, 6, 1125.
and the 3,4,5-trimethoxyphenyl ring of 1f binds to
p-rich hydro-
phobic region of the pocket. The hydroxyl groups at C2 and C4 of
the resorcinol ring formed hydrogen-bonding with Thr184 and
Asp93, respectively in the hydrophilic region of the pocket, while
the chlorine atom of the resorcinol ring formed Van der Waals con-
tacts with Leu48 and Val186. On the other hand, the phenyl group
14. Seo, Y. H. Bull. Korean Chem. Soc. 1917, 2013, 34.
15. Sordella, R.; Bell, D. W.; Haber, D. A.; Settleman, J. Science 2004, 305, 1163.
16. Paez, J. G.; Janne, P. A.; Lee, J. C.; Tracy, S.; Greulich, H.; Gabriel, S.; Herman, P.;
Kaye, F. J.; Lindeman, N.; Boggon, T. J.; Naoki, K.; Sasaki, H.; Fujii, Y.; Eck, M. J.;
Sellers, W. R.; Johnson, B. E.; Meyerson, M. Science 2004, 304, 1497.
17. Kobayashi, S.; Boggon, T. J.; Dayaram, T.; Janne, P. A.; Kocher, O.; Meyerson, M.;
Johnson, B. E.; Eck, M. J.; Tenen, D. G.; Halmos, B. N. Engl. J. Med. 2005, 352, 786.
18. Tang, Z.; Du, R.; Jiang, S.; Wu, C.; Barkauskas, D. S.; Richey, J.; Molter, J.; Lam,
M.; Flask, C.; Gerson, S.; Dowlati, A.; Liu, L.; Lee, Z.; Halmos, B.; Wang, Y.; Kern,
J. A.; Ma, P. C. Br. J. Cancer 2008, 99, 911.
19. Sharp, S. Y.; Prodromou, C.; Boxall, K.; Powers, M. V.; Holmes, J. L.; Box, G.;
Matthews, T. P.; Cheung, K. M.; Kalusa, A.; James, K.; Hayes, A.; Hardcastle, A.;
Dymock, B.; Brough, P. A.; Barril, X.; Cansfield, J. E.; Wright, L.; Surgenor, A.;
Foloppe, N.; Hubbard, R. E.; Aherne, W.; Pearl, L.; Jones, K.; McDonald, E.;
Raynaud, F.; Eccles, S.; Drysdale, M.; Workman, P. Mol. Cancer Ther. 2007, 6,
1198.
of the 3,4,5-trimethoxyphenyl ring made
p–p interaction with
Phe138 and formed Van der Waals contact with Leu107 in the
hydrophobic region of the pocket. Collectively, the estimated bind-
ing energy (DGb) and inhibition constants (Ki) using the Lamarck-
ian genetic algorithm result in 7.91 kcal/mol and 1.59
respectively.
In conclusion, a series of hybrid Hsp90 inhibitors were ratio-
nally designed, synthesized, and evaluated against gefitinib-
resistant non-small cell lung cancer cell line (H1975). Compound
1f appears to overcome gefitinib resistance and inhibit H1975 cell
proliferation. The hybrid inhibitor 1f exerts a modest inhibitory
lM,
20. General procedure for preparing compounds (9a–g and 1a–g), as exemplified
for compound 9f and 1f.
Procedure for the synthesis of compound 9f: A mixture of compound 7 (0.1 g,
0.37 mmol), 3,4,5-trimethoxybenzaldehyde (0.16 g, 0.93 mmol), KOH (4.0 g) in
2 mL of water and 10 mL of ethanol was stirred at rt for 12 h. The mixture was
neutralized with 6 N HCl to pH 6 and then extracted with ethyl acetate. The
organic layer was washed with saturated NaHCO3 solution three times, dried
over Na2SO4, and concentrated under reduced pressure, and purified by
column chromatography (20% ethyl acetate in hexane) to afford compound 9f
(0.21 g, 63.7%) as yellow solid. 1H NMR (400 MHz, CDCl3) d 7.77 (s, 1H), 7.55 (d,
J = 15.6 Hz, 1H), 7.44 (d, J = 15.6 Hz, 1H), 6.76 (s, 2H), 6.46 (s, 1H), 6.08–5.59 (m,
2H), 5.49–5.38 (m, 2H), 5.32–5.21 (m, 2H), 4.60 (dd, J = 5.2 Hz, J = 18.0 Hz, 4H),
3.83 (s, 9H). ESI MS (m/e) = 445 [M+1]+.
Procedure for the synthesis of compound 1f: The resulting compound 9f was
stirred under microwave irradiation (Biotage Initiator) for 30 min at 120 °C in
the presence of PdCl2(PPh3)2 (10 mg) and ammonium formate (150 mg) in
4 mL of THF. The reaction mixture was diluted with ethyl acetate. The organic
layer was washed with water, dried over Na2SO4, concentrated under reduced
pressure, and purified by column chromatography (20% ethyl acetate in
hexane) to afford compound 1f (0.12 g, 69.7%) as yellow solid. 1H NMR
(400 MHz, CDCl3) d 13.20 (s, 1H), 7.09 (s, 1H), 7.83 (d, J = 15.2 Hz, 1H), 7.35 (d,
J = 15.6 Hz, 1H), 6.88 (s, 2H), 6.65 (s, 1H), 6.20 (s, 1H), 3.95 (s, 6H), 3.91 (s, 3H).
13C NMR (100 MHz, CDCl3) d 191.1, 164.1, 160.0, 153.4, 145.2, 140.5, 130.9,
130.1, 119.1, 113.9, 112.3, 106.0, 104.1, 60.9, 56.2. ESI MS (m/e) = 365 [M+1]+.
activity (48
cancer cells, considering that known Hsp90 inhibitors, VER-
49009 and PU3 are reported to have 1–5 M and 50 M of GI50
lM of GI50) against the growth of gefitinib-resistant
l
l
in human cancer cell lines, respectively.11,19 Consequently, 1f man-
ifests considerable degradation of Hsp90’s client proteins and
induction of Hsp70, supporting that 1f targets the Hps90 chaper-
one machinery. Currently, our efforts are directed toward explor-
ing further structure–activity relationships and investigating the
complete mechanism, biological profile and safety of the hybrid
inhibitor. The result will be reported in due course.
Acknowledgment
The present research has been conducted by the Settlement
Research Grant (to Y.H. Seo) of Keimyung University in 2011.