Design, synthesis, cytoselective toxicity, structure-activity relationships, and pharmacophore of thiazolidinone derivatives targeting drug-resistant lung cancer cells

Article abstract of DOI:10.1021/jm7012024

Ten cytoselective compounds have been identified from 372 thiazolidinone analogues by applying iterative library approaches. These compounds selectively killed both non-small cell lung cancer cell line H460 and its paclitaxel-resistant variant H460_taxR at an IC₅₀ between 0.21 and 2.93 μM while showing much less toxicity to normal human fibroblasts at concentrations up to 195 μM. Structure-activity relationship studies revealed that (1) the nitrogen atom on the 4-thiazolidinone ring (ring B in Figure 1) cannot be substituted, (2) several substitutions on ring A are tolerated at various positions, and (3) the substitution on ring C is restricted to the -NMe₂ group at the 4-position. A pharmacophore derived from active molecules suggested that two hydrogen bond acceptors and three hydrophobic regions were common features. Activities against P-gp-overexpressing and paclitaxel-resistant cell line H460_taxR and modeling using a previously validated P-gp substrate pharmacophore suggested that active compounds were not likely P-gp substrates.

Full text of DOI:10.1021/jm7012024

1242
J. Med. Chem. 2008, 51, 1242–1251
Design, Synthesis, Cytoselective Toxicity, Structure–Activity Relationships, and Pharmacophore
of Thiazolidinone Derivatives Targeting Drug-Resistant Lung Cancer Cells
,O
Hongyu Zhou,†,‡ Shuhong Wu,§ Shumei Zhai,‡ Aifeng Liu,‡ Ying Sun,‡ Rongshi Li,‡, Ying Zhang,‡ Sean Ekins,#,
Peter W. Swaan,O Bingliang Fang,§ Bin Zhang,‡ and Bing Yan*,†,‡
Department of Chemical Biology and Therapeutics, St. Jude Children’s Research Hospital, Memphis, Tennessee 38105, School of
Pharmaceutical Sciences, Shandong UniVersity, Jinan, Shandong, China, Department of Thoracic and CardioVascular Surgery, UniVersity of
Texas M. D. Anderson Cancer Center, Houston, Texas 77030, Collaborations in Chemistry, 601 Runnymede AVenue,
Jenkintown, PennsylVania 19046, Department of Pharmacology, Robert Wood Johnson Medical School, UniVersity of Medicine and Dentistry
of New Jersey, Piscataway, New Jersey 08854, and Department of Pharmaceutical Sciences, UniVersity of Maryland, Maryland 21201
ReceiVed September 24, 2007
Ten cytoselective compounds have been identified from 372 thiazolidinone analogues by applying iterative
library approaches. These compounds selectively killed both non-small cell lung cancer cell line H460 and
its paclitaxel-resistant variant H460taxR at an IC50 between 0.21 and 2.93 µM while showing much less
toxicity to normal human fibroblasts at concentrations up to 195 µM. Structure–activity relationship studies
revealed that (1) the nitrogen atom on the 4-thiazolidinone ring (ring B in Figure 1) cannot be substituted,
(2) several substitutions on ring A are tolerated at various positions, and (3) the substitution on ring C is
restricted to the -NMe2 group at the 4-position. A pharmacophore derived from active molecules suggested
that two hydrogen bond acceptors and three hydrophobic regions were common features. Activities against
P-gp-overexpressing and paclitaxel-resistant cell line H460taxR and modeling using a previously validated
P-gp substrate pharmacophore suggested that active compounds were not likely P-gp substrates.
Introduction
Lung cancer is the number one cause of cancer-related deaths
worldwide and in the United States (∼160,000 deaths annually).
Despite advances in cancer research, the overall five-year
survival of lung cancer patients remains a dismal 15% in contrast
to most other solid organ tumors according to the American
Association for Cancer Research (http://www.aacr.org/home/
Figure 1. Early hit: MMPT, R1 ) H, R2 ) 4-Me [5-[(4-methylphe-
nyl)methylene-2-(phenylamino)-4(5H)-thiazolone].
public--media/for-the-media/fact-sheets/organ-site-fact-sheets/
lung-cancer.aspx). More than 80% of bronchogenic malignancy
is from non-small cell lung cancer (NSCLC).a Among other
factors, inherent and acquired resistance to treatment1 and the
dose-limiting toxicity caused by the narrow therapeutic window
of many cancer drugs2 are recognized as major obstacles for
effective cancer therapy. Multidrug resistance (MDR) is a
phenotype of cross-resistance to multiple drugs with diverse
chemical structures. One of the well-documented MDR mech-
anisms is the overexpression of the MDR-1 gene that encodes
the transmembrane, ATP-dependent drug efflux transporter
P-glycoprotein (P-gp) in response to chemotherapy.3–5 P-gp
prevents the intracellular accumulation of many cancer drugs
by increasing their efflux out of cancer cells as well as through
hepatic, renal, or intestinal clearance pathways.4 Attempts to
coadminister P-gp modulators or inhibitors to increase cellular
availability by blocking the actions of P-gp have met with
limited success.4,6–8 Therefore, a more promising approach lies
in the design and discovery of novel compounds that are not
substrates of P-gp and are effective against drug-resistant cancer
while at the same time exhibiting minimal toxicity to normal
cellular functions.
Thiazolidinone derivatives have been investigated for a range
of pharmacologic indications such as anti-inflammatory,9 anti-
microbial,10 antiproliferative,11,12 antiviral,13 anticonvulsant,14,15
antifungal,16 and antibacterial17 activities but their anticancer
effects have been less widely documented.18 In an early random
screening of commercial compounds, a compound with a
4-thiazolidinone core structure was found to selectively kill drug-
resistant cancer cells and induce apoptosis19 (Figure 1). How-
ever, the optimal pharmacophore structure and the structure-
–activity relationship for the cytoselective toxicity have not to
date been explored, and the expected therapeutic window is
small. To search for more selective and novel thiazolidinone
compounds with a wider therapeutic window that could reveal
the structure–activity relationship for the cytoselective anticancer
activity, we designed and synthesized iterative focused thiazo-
lidinone libraries. These consecutive libraries were screened
against paclitaxel-sensitive and -resistant NSCLC cell lines H460
and H460taxR, respectively, using a fast-dividing normal human
fibroblast (NHFB) line as a general cell line for measuring
cytotoxicity against normal cells. Here we report our hit-follow-
up approaches that led to the discovery of novel thiazolidinone-
related compounds that were highly toxic to NSCLC H460 cells
* Author to whom correspondence should be addressed (telephone
+9014952797; fax +9014955715; e-mail bing.yan@stjude.org).
†
St. Jude Children’s Research Hospital.
Shandong University.
University of Texas M. D. Anderson Cancer Center.
‡
§
Present address: Pharmaron Inc. Beijing, China.
Collaborations in Chemistry.
#
University of Medicine and Dentistry of New Jersey.
University of Maryland.
NSCLC, non-small cell lung cancer; MDR, multidrug resistance; P-gp,
O
a
P-glycoprotein; NHFB, normal human fibroblast; LC/MS, liquid chro-
matograpy/mass spectrometry; ESI-MS, electrospray ionization mass spec-
trometry; HRMS, high-resolution mass spectrometry; NMR, nuclear
magnetic resonance; FTIR, Fourier transform infrared; TLC, thin layer
chromatography.
10.1021/jm7012024 CCC: $40.75
2008 American Chemical Society
Published on Web 02/08/2008

CytoselectiVe Anticancer Thiazolidinone Analogues
Chart 1. Building Blocks for Library I
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1243
thiazolidinones26–28 and their analogues, such as thiazolines,29,30
thiohydantoins,31,32 and thiazolidinediones,33,34 were reported
using both solid- and solution-phase strategies, there has been,
to our knowledge, no report on the synthesis of thiazolidinones
with an imino group at the 3-position. Our synthetic route to
the thiazolidinone libraries is shown in Scheme 1. Due to the
shortage of commercial isothiocyanates, the synthesis route
starting from 3 was not used in library synthesis. The construc-
tion of primary thioureas 5 was achieved by the reaction of
aniline 1 with ammonium thiocyanate, in the presence of acid.
Thioureas 5 reacted with ethyl 2-chloroacetate to produce
thiazolidinones 7 as precipitate, which was filtered and washed
with absolute ethanol to give the product with a purity of ∼90%
and a yield of ∼75%.
To prepare library I (Scheme 1), the final step of the reaction
was carried out in piperidine and absolute ethanol at 60 °C.
About 95% of the products formed as precipitate, and they were
purified by simple filtration and washing with ethanol. About
5% of the products that did not precipitate were purified by
normal phase column chromatography. A total of 185 of 216
designed compounds were successfully prepared and character-
ized by LC/MS/UV214 with an average purity of 91% with the
unreacted starting materials as the major impurities. Selected
compounds were further characterized by HRMS, 1H NMR, and
FTIR (see chromatograms and spectra in Supporting Informa-
tion). The compound structure, purity, and yield are summarized
in Supporting Information Table 1.
To explore the effect of nitrogen substitution on anticancer
activity, we designed library II, in which the nitrogen in the
4-thaizolidinone ring was substituted (Scheme 2). Isothiocyan-
ates 2 reacted with amines to yield N-substituted thioureas 11.
They then reacted with ethyl 2-chloroacetate 6 to give N-
substituted thiazolidinones 12 with a purity of ∼90% and a yield
of ∼70%. After the coupling reaction with aromatic aldehydes,
a total of 17 compounds were synthesized and purified. The
average purity of this library was 92%, as determined by LC/
MS/UV214 with the unreacted starting materials as the major
impurities. The structure, yield, and purity of these compounds
are summarized in Supporting Information Table 2.
and their paclitaxel-resistant variant H460taxR, yet with much
less toxicity to NHFBs. Furthermore, pharmacophore modeling
revealed unique features and structural requirements for active
compounds.
Results and Discussion
Early cytotoxicity studies of commercial compounds19,20
identified a hit that contained two substituted phenyl moieties
on the 4-thiazolidinone core, suggesting that the three-ring
system might be required for the anticancer activity. The
discovery of novel and more potent lead compounds with a
wider therapeutic window would require a thorough exploration
of all the combinations of diverse functional groups on each of
the three rings. Therefore, we kept the three-ring feature and
designed focused combinatorial libraries with maximal diversity
on substitution groups on the three rings.
Design of Focused Combinatorial Libraries. The diversity
in building blocks eventually determines the chemical space
coverage of a library. We selected aniline and aromatic aldehyde
building blocks using the following criteria: they must (1) have
structural diversity as determined by calculated physicochemical
properties of the virtual product, (2) form products that obey
Lipinski’s “rule of five”,21 and (3) generate products with
synthetic feasibility. A total of 18 anilines and 20 aldehydes
were first used to generate a virtual library of 360 compounds.
Using Lipinski’s rule of five selection criteria, 120 compounds
were eliminated. Chemical feasibility and building block valida-
tion studies eliminated another 24 compounds. The final library
consisted of 216 compounds (Chart 1). The diverse molecular
properties of the virtual products (Figure 2) indicated that they
were theoretically “drug-like” albeit on the basis of the limited
criteria above.
Cytoselective Anticancer Toxicity Assays. We used a panel
of drug-sensitive (H460) and drug-resistant (H460taxR) cells and
NHFBs to screen the libraries I and II. Screening was carried
out in three stages: the primary screening, the secondary
confirmation, and the dose–response determinations. In the
primary screening of library I, compounds at a concentration
of 10 µM were tested against the cell panel. Compound effects
were monitored by observing cell numbers and cell morphology
changes in 96-well plates at 24 and 48 h after compound
addition. Compounds that exhibited toxicity to both cancer cell
lines but not to normal cells were selected for the secondary
confirmation assays. In the secondary screening, compounds at
the same concentration as in the primary screening were
screened in triplicate. As a result, eleven compounds were
identified as potent agents for inducing cytoselective toxicity.
Dose–response studies showed that these compounds selectively
killed or inhibited H460 (IC50 at 0.28-14.79 µM) and H460taxR
(IC50 at 0.42-12.88 µM) cells in a dose-dependent manner and
showed less toxicity to NHFBs (35->100 µM; Figure 3).
Structural analysis of these compounds revealed that R1 could
tolerate functional groups such as Cl- and Me- at the 2- or
4-position or CF3- at the 3-position. However, R2 tended to
accept only functional groups such as -NMe2 and -benzene
at the 4-position or -OH group at the 2-position.
Synthesis, Purification, and Characterization of Libraries.
Syntheses of thiazolidinones are well documented in the
literature.9,10,16,22–25 Although combinatorial syntheses of

1244 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Zhou et al.
Figure 2. Calculated molecular properties for the focused thiazolidinone library I.
Scheme 1. Synthesis of Route for Libraries I and IIIa
a Reagents and conditions: (i) NH4SCN (2), 6 N HCl, 80 °; (ii) NH3 ·H2O
(4), rt; (iii) ethyl chloroacetate (6); NaOAc, EtOH, 60 °C; (iv) aldehydes
(8), piperidine, EtOH, 60 °C.
Figure 3. Dose-dependent cytotoxicity assays for compounds from
primary screening in cell lines H460, H460taxR, and NHFB.
selected functional groups from screening library I and leave
the nitrogen unsubstituted in the design of the early lead
optimization library.
Scheme 2. Synthetic Route for Library IIa
Preparation and Screening of the Early Lead Optimiza-
tion Library. The early lead optimization library III was
designed to have R1 groups such as -Cl, -Me, -OH, and -CF3
at all possible positions and have R2 groups such as -NMe2,
-benzene, and others at the 4-position or -OH at the 2-position
(Chart 3). Other functional groups were also incorporated to
test for the possible enhancement of activity or solubility.
Important goals for this library were to select better lead
compounds and to explore the structure–activity relationship
of the active compounds. The library was synthesized according
to Scheme 1, and all compounds were purified by either
recrystallization or chromatography. We enforced two stringent
requirements for the early lead optimization library III: (1) all
compounds in the library must be obtained to allow exploration
of the full chemical space and all substituent combinations and
(2) all compounds must have a high purity to ensure unambigu-
ous biological data. We synthesized, purified, and obtained all
170 designed compounds, and the average purity of the library
was 95%, shown by LC/MS/UV214. The single-concentration
(10 µM) primary screening identified 40 hits. Using a lower
concentration (5 µM), 10 compounds were confirmed and
selected for dose–response studies, which showed that the 50%
inhibitory concentration (IC50) for H460 and H460taxR cells was
a Reagents and conditions: (i) R1, -NH2, (10), rt; (ii) ethyl chloroacetate
(6), NaOAc, EtOH, 60 °C; (iii) aldehydes (8), piperideine, EtOH, 60 °C.
To test the effect of nitrogen substitution on the cytoselective
toxicity of the compounds, we screened library II, which did
not yield any compounds with cytoselective toxicity. Com-
pounds in this library fell into two categories: those that showed
no toxicity, and those that were toxic to all cell lines tested.
Several compounds in this group, when the nitrogen was not
substituted, were highly active in selectively killing cancer cells
(see I-7 in Chart 2, for example). Therefore, nitrogen substitution
blocked the cytoselective toxicity of the compounds. From this
finding, we decided to explore the optimal combination of the

CytoselectiVe Anticancer Thiazolidinone Analogues
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1245
Chart 2. Active Compounds from the Screening of Libraries I and II
Chart 3. Building Blocks for Library III
289, and III-324- showed no sign of 50% cell inhibition at
concentrations much higher than 100 µM (Figure 4).
Time-Dependent Compound Effects on Cell Prolifera-
tion and Morphology. To examine the time-dependent response
of the cells to these compounds, we studied H460 and H460taxR
cells and NHFBs for their time-dependent morphology changes
and cell proliferation in the presence of compounds I-7-, I-27,
I-47, and I-67 at concentrations of 1 and 2 µM for up to 48 h.
Changes in cell morphology were observed using a phase-
contrast microscope. Figure 5 shows that compound I-7
selectively killed both H460 and H460taxR cells in a time-
dependent manner while having a much less toxic effect on
NHFBs. Results for compounds I-27, I-47, and I-67 are in the
as low as 300 nM. Except for compound I-25 (IC50 for NHFB
is 195 µM), NHFBs did not reach 50% cell inhibition at the
highest compound concentration used (100 µM) for active
compounds. Higher compound concentrations could not be used
due to the solubility limitation and the limited allowable DMSO
(1% for cells used here) in cell cultures. We tentatively took
>100 µM as the IC50 of the compounds tested against NHFB
cells. It was noted that compounds such as I-27-, I-47-, III-
Figure 4. Dose-dependent cell viability for selected compounds from
early lead optimization screening in cell lines H460, H460taxR, and
NHFB.

1246 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Zhou et al.
Figure 5. Time-dependent morphology changes of H460, H460taxR, and NHFB cells in the presence of compound I-7 at 1 µM. The magnification
is 200×.
Table 1. P-gp Pharmacophore Modeling for Group 1
very little 2D structural similarity to the two molecules used in
P-gp substrate pharmacophore model building and generally
seem to be smaller than molecules previously assessed.
P-gp
features
compd no.
R1
R2
substrate fit mapped
I-18
I-36
I-48
I-69
I-94
I-98
H
4-SMe
9H-fluoren-2-yl
2-Cl
4-formyl
2,4,5-trimethoxy
4-SMe
no mapa
Structure–Activity Relationship and Pharmacophore
Modeling. The cytoselective anticancer activities of compounds
are summarized in Table 3. In general, these compounds were
potent cytoselective anticancer agents that selectively kill drug-
resistant cancer cells but not normal cells. IC50 values for cancer
cell killing were all <2.93 µM, and the IC50 for normal NHFB
cells were all >100 µM (195 µM for compound I-25). The use
of a compound concentration >100 µM was effectively
prohibited for most compounds due to solubility limitation and
the allowable amount of DMSO in cell cultures.
2-Me
4-Me
2-Cl
4-Cl
4-Cl
2.64b
no map
no map
3.86
4H, 1HBAc
4H, 1HBA
3H, 2HBA
4H, 1HBA
3H, 2HBA
1.75
I-114
I-118
I-126
I-132
I-196
I-213
I-233
2-OMe
2-OMe
3-NO2
3-NO2
4-OH
2,4,5-trimethoxy
4-SMe
2-OH
4-naphthalen-2-yl 3.21
9H-fluoren-2-yl
3.62
0.716
no map
3H, 2HBA
4H, 2HBA
3H, 2HBA
3H, 2HBA
0.702
2.6
2.61
4-phenoxy
4-ethoxycarbonyl 4-COOH
4-COOH
Substitution groups were dominantly electron-donating groups
at the 2- or 4-position for both rings A and C. R1 allowed groups
such as -Me and -Cl at both the 2- and 4-positions. However,
R2 was restricted to only the -NMe2 group at the 4-position.
Compounds with nitrogen substitutions on the 4-thiazolidinone
ring B did not show any cytoselective toxicity. We, therefore,
tentatively concluded that nitrogen substitution blocked such
activity. This nitrogen and the R2 on rings B and C are
stereochemically confined and are likely to be close to the target
binding site. In contrast, R1 may be remote from the active site
(Figure 7). Because R1 is much more tolerant to structural
variations, the future optimization of solubility, absorption,
distribution, metabolism, excretion, toxicity, and pharmacoki-
netic properties can be tested through the modifications of this
site. Using all of the active compounds from Table 3, a common
feature pharmacophore was developed that indicated that two
hydrogen bond acceptors and three hydrophobic regions were
common across all of these compounds (Figure 7).
a If two or more features were missed, the compound was designated
“no map”. b Higher values suggested better map. c H, hydrophobe; HBA,
hydrogen bond acceptor.
Table 2. P-gp Pharmacophore Modeling Results for Group 2
compd no.
R1
R2
P-gp substrate fit features mapped
I-7
I-25
I-27
I-47
I-67
I-86
I-87
I-166
I-187
III-289
III-324
H
4-NMe2
4-OMe
4-NMe2
4-NMe2
4-NMe2
2-OH
4-NMe2
2-OH
4-NMe2
4-NHAc
no mapa
no map
no map
no map
no map
2-Me
2-Me
4-Me
2-Cl
4-Cl
4-Cl
3-CF3
4-OH
3-Cl
1.88b
1
3H, 2HBAc
3H, 2HBA
no map
no map
3.24
3H, 2HBA
3H, 2HBA
4-OMe 4-NMe2
1.4
a If two or more features were missed, the compound was designated
“no map”. b Higher values suggested better map. c H, hydrophobe; HBA,
hydrogen bond acceptor.
Anticancer Compounds Are Not Michael Acceptors. The
stability of anticancer compounds in vivo is important for the
potential clinical applications of these compounds. However,
the R,ꢀ-unsaturated-ketone-like structure of the active com-
pounds raised a concern that they might undergo Michael
reactions in ViVo with nucleophiles. To test this possibility, we
carried out aqueous reactions between two nucleophiles, L-
glutathione reduced and L-glutathione reduced ethyl ester, and
several thiazolidinone analogues (Scheme 3) under basic condi-
tions. The starting materials, products, and reaction progression
were monitored by LC/MS. The lack of reaction products after
12 h indicated that these anticancer compounds were not likely
to be deactivated by nucleophiles in ViVo through Michael
addition.
Supporting Information. These findings confirmed the lack of
toxicity to normal cells and the unique cytoselective anticancer
effect of this class of compounds.
P-gp Substrate and Inhibitor Pharmacophore Modeling.
We selected structures of 13 compounds that killed H460, but
not drug-resistant H460taxR (group 1, see Table 1) and 11
compounds that killed both cancer cells, but not normal cells
(group 2, see Table 2) for P-gp pharmacophore modeling studies.
P-gp pharmacophore models that had been previously generated
were used to assess the fit of molecules.35,36 From group 1, 9
of 13 molecules (70%) were found to map to the P-gp substrate
pharmacophore (Table 1 and Figure 6). In contrast, only 4 of
11 molecules (36%) in group 2 mapped to the P-gp substrate
pharmacophore (Table 2). This suggests that the molecules in
group 2 are less likely to be P-gp substrates than those in group
1. It may be worth noting that the molecules evaluated bear
Conclusion. To search for effective and selective chemo-
therapeutic agents against drug-resistant lung cancer, we
prepared iterative focused thiazolidinone libraries containing 372

CytoselectiVe Anticancer Thiazolidinone Analogues
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1247
Figure 6. P-gp substrates mapped to the substrate pharmacophore.36 Green spheres represent hydrogen bond acceptors, and cyan features represent
hydrophobic features.
Table 3. Cytoselective Anticancer Activity
IC50(µM)
compd no. (purity
LC/MS/UV214
)
R1
R2
H460
H460taxR
NHFB
I-7(95%)
H
4-NMe2
4-OMe
4-NMe2
4-NMe2
4-NMe2
2-OH
4-NMe2
4-NMe2
4-NHAc
4-NMe2
0.50 ( 0.15
1.78 ( 0.60
0.65 ( 0.25
0.94 ( 0.20
0.73 ( 0.02
1.85 ( 0.65
2.89 ( 0.03
1.32 ( 0.25
1.19 ( 0.02
2.52 ( 0.12
0.21 ( 0.03
1.70 ( 0.50
0.54 ( 0.10
0.82 ( 0.07
0.88 ( 0.01
1.18 ( 0.03
2.93 ( 0.12
1.08 ( 0.02
1.25 ( 0.10
1.60 ( 0.10
>100
195
I-25(95%)
I-27(94%)
I-47(96%)
I-67(93%)
I-86(90%)
I-87(85%)
I-187(93%)
2-Me
2-Me
4-Me
2-Cl
4-Cl
4-Cl
4-OH
3-Cl
4-OMe
>100
>100
>100
>100
>100
>100
>100
>100
III-289(95%)
III-324(93%)
compounds with an average purity of 90–95%. By screening
the compounds against drug-sensitive and drug-resistant NSCLC
cell lines H460 and H460taxR using NHFBs as a toxicity control,
a series of potent compounds that had a much wider expected
therapeutic window for their cytoselective toxicity for drug-
resistant cancer cells has been identified. We found a unique
structure–activity relationship for the cytoselective anticancer
activity of the potent compounds. The nitrogen substitution
blocked cytoselective anticancer activity. Although the R1 site
was lenient to substitution, the R2 site required an -NMe2 group
at the 4-position for optimal activity. Because H460taxR ex-
pressed excessive amounts of P-gp proteins, these anticancer
compounds discovered were evidently not P-gp substrates on
the basis of their cytotoxicity, and this conclusion was supported
by a P-gp substrate pharmacophore assessment. A separate
pharmacophore for the most active compounds showed a
common arrangement of two hydrogen bond acceptors and three
hydrophobic regions (Figure 7). This anticancer compound
pharmacophore may be useful in further database searching35
to scaffold-hop in finding novel molecules that conform to the
overall shape, volume, and molecular feature distribution
outlined in this study and with potential for fewer interactions
with P-gp. Further studies to determine the mechanism of these
anticancer compounds will be essential. It is interesting to
note the pharmacophore discovered in this study was shared
by some nonpeptidic inhibitors of a protease, ubiquitin isopep-
tidase that were discovered using a simple pharmacophore-based
search of the NCI database.37,38 These inhibitors with IC50 values
in the low tens of micromolar caused cell death independent of
the tumor suppressor p53. The ubiquitin isopeptidase inhibitors
shikoccin, dibenzylideneacetone, and curcumin as well as the
more recently described punaglandins from coral indicate that
an accessible R,ꢀ-unsaturated ketone is essential for ubiquitin
isopeptidase activity.39 This raised an interesting possibility that
Figure 7. Anticancer molecule pharmacophore generated with the 10
molecules from Table 3 showing the four most active (I-7, I-27, I-47,
I-67) aligned. Green spheres represent hydrogen bond acceptors, cyan
features represent hydrophobic features. The outer translucent shape is
the van der Waals surface of all the molecules.

1248 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Scheme 3. Test for Possible Michael Reaction in ViVo
Zhou et al.
thiourea and ethyl chloroacetate according to the procedure
described above: yield, 72.3%, yellow powder; 1H NMR (400 MHz,
MeOD), δ 7.69 (s, 1H), 7.13 (m, 3H), 4.02 (d, 2H, J ) 39.1 Hz);
ESI-MS, m/z 211.5 (M + 1).
molecules in this study could be targeting ubiquitin isopeptidase
or other known targets.
Experimental Section
2-(2-Hydroxyphenyl)imino-thiazolidin-4-one. 2-(2-Hydroxy-
phenyl)imino-thiazolidin-4-one was prepared from 1-(2-hydroxy-
phenyl)thiourea and ethyl chloroacetate as described above: yield,
The chemical reagents were purchased from Acros Organics
(Geel, Belgium) and used without further purification. IR spectra
were recorded on a Nicolet 380 FTIR spectrophotometer. LC/MS
was performed on a Waters system equipped with a Waters 2795
separation module, a Waters 2996 PDA detector, and a Micromass
ZQ detector. A C18 column (2.0 µM, 2.0 × 50 mm) was used for
the separation. The eluent was a mixture of methanol and water
containing 0.05% trifluoroacetic acid with a linear gradient from
50:50 v/v methanol/H2O to 100% methanol over 6.5 min at a 1.0
mL/min flow rate. UV detection was at 214 nm. Mass spectra were
recorded in positive ion mode using electrospray ionization. NMR
spectra were recorded on a Bruker 400 MHz NMR spectrometer
using MeOD as solvent. Parallel synthesis was done on a Zhicheng
ZHMY-113H parallel synthesizer (Shanghai, China). Molecular
properties were calculated using TSAR software from Accelrys (San
Diego, CA) and Pipeline Pilot from SciTegic (San Diego, CA).
High-resolution mass spectrometry analysis was performed by the
Mass Spectroscopy Laboratory of the University of Illinois at
Urbana–Champaign.
Arylthioureas (5). General Procedure. To a magnetically stirred
solution of primary anilines 1 (10.0 mmol) in 10 mL of 6 N HCl
was added NH4SCN 2 (1.14 g, 15.0 mmol). After being heated to
80 °C, the mixture became clear. The mixture was heated for
another 12 h until abundant precipitate appeared. The mixture was
cooled to room temperature, filtered, washed with water and
anhydrous petroleum ether, and dried under vacuum to produce
the corresponding thiourea 3 at high purity and in high yield. This
compound was used directly without further purification. For a
typical compound, 2-chlorophenyl thiourea, 1.77 g of white powder
was obtained in a 95.4% yield with an HPLC purity of >95%.
ESI-MS, m/z 187.0.
2-Arylimino-thiazolidin-4-ones (7). General Procedure. To a
stirred suspension of primary thiourea 5 (6.0 mmol) and
anhydrous sodium acetate (2.479 g, 30 mmol) in 20 mL of
absolute ethanol was added 1.28 mL of ethyl chloroacetate 6
(12.0 mmol). The mixture was heated at 60 °C for 6 h. After
being cooled to room temperature, a precipitate formed. The
precipitate was filtered and washed with ethanol to produce the
crude product. The filtrate was concentrated under reduced
pressure and extracted between water and EtOAc. The organic
layer was concentrated to yield more product. The combined
crude product was recrystallized from EtOAc to produce
2-arylimino-thiazolidin-4-ones 7, which was used directly for
subsequent reactions without further purification.
1
62.5%, black powder; H NMR (400 MHz, MeOD), δ 7.35, 7.04
(dd, 1H, J ) 7.7 Hz), 7.19 (m, 1H), 6.66 (m, 2H), 4.00 (d, 2H, J
) 32.4 Hz); ESI-MS, m/z 209.5 (M + 1).
2-Arylimino-5-arylidene-thiazolidin-4-ones (Libraries I and
III). General Procedure. 2-Arylimino-thiazolidin-4-ones 7 (0.5
mmol) and benzaldehyde 8 (0.6 mmol) were dissolved in 3 mL of
absolute ethanol. Next, 50 µL of piperidine (0.5 mmol) was added
to the mixture, and the mixture was stirred for 12 h at 60 °C until
a precipitate formed. The mixture was cooled to room temperature,
and the precipitate was filtered and washed with petroleum ether
and absolute ethanol to yield 2-arylimino-5-arylidene-thiazolidin-
4-one at purity of >90%.
Reaction mixtures that did not form precipitates were extracted
with EtOAc. The organic layer was washed with brine, dried by
anhydrous Na2SO4, and concentrated under reduced pressure. The
crude compound was purified by silica gel column chromatography
and eluted with 20% EtOAc in petroleum ether.
2-(2-Chlorophenylimino)-5-(4-hydroxy-3-methoxyben-
zylidene)thiazolidin-4-one (I-063). I-063 was prepared from 2-(2-
chlorophenyl)imino-thiazolidin-4-one and 4-hydroxy-3-methoxy-
benzaldehyde (AL003) according to the procedure described above:
yield, 33.26%, yellow powder; 1H NMR (400 MHz, MeOD), δ
7.51 (s, 1H), 7.37 (d, 1H, J ) 8.0 Hz), 7.21 (d, 1H, J ) 7.5 Hz),
7.08 (s, 1H), 6.92 (t, 3H, J ) 29.2 Hz), 6.74 (d, 1H, J ) 8.2 Hz),
3.77 (d, 3H, J ) 38.2 Hz), 1.91, 1.18 (ss, 1H); ESI-MS, m/z 361.7
(M + 1). HR-MS calculated for C17H13ClN2O3S (M + H)+:
361.0414; found 361.3403.
2-(2-Chlorophenylimino)-5-(4-methoxybenzylidene)thiazo-
lidin-4-one (I-065). I-065 was prepared from 2-(2-chlorophe-
nyl)imino-thiazolidin-4-one and 4-methoxybenzaldehyde (AL005)
according to the procedure described above: yield, 73.9%, yellow
powder; 1H NMR (400 MHz, MeOD), δ 7.53 (s, 1H), 7.38 (d, 1H,
J ) 8.0 Hz), 7.32 (d, 2H, J ) 8.4 Hz), 7.23 (t, 1H, J ) 7.6 Hz),
7.09 (t, 1H, J ) 7.6 Hz), 6.98 (d, 1H, J ) 7.4 Hz), 6.89 (d, 2H, J
) 8.4 Hz), 3.72 (s, 3H), 1.91, 1.19 (ss, 1H); ESI-MS, m/z 345.6
(M + 1).
2-(2-Chlorophenylimino)-5-(2-naphthalenyl)methylene-thia-
zolidin-4-one (I-072). I-072 was prepared from 2-(2-chlorophe-
nyl)imino-thiazolidin-4-one and 2-naphthaldehyde (AL012) ac-
cording to the procedure described above: yield, 99.0%, orange
powder; 1H NMR (400 MHz, MeOD), δ 8.30 (s, 1H), 8.04 (d, 1H,
J ) 8.1 Hz), 7.82 (d, 2H, J ) 4.1 Hz), 7.49 (m, 3H), 7.37 (dd, 2H,
J ) 7.8 Hz), 7.19 (t, 1H, J ) 7.4 Hz), 7.06 (t, 1H, J ) 7.6 Hz),
6.96 (d, 1H, J ) 7.9 Hz), 1.91, 1.18 (ss, 1H); ESI-MS, m/z 365.6
(M + 1). HR-MS calculated for C20H13ClN2OS (M + H)+:
365.0515; found 365.0516.
2-(3-Trifluoromethyl-phenylimino)-5-(2-hydroxy-ben-
zylidene)thiazolidin-4-one (I-166). I-166 was prepared from 2-(3-
trifluoromethyl-phenyl)imino-thiazolidin-4-one and 2-hydroxy-
benzaldehyde (AL006) according to the procedure described above:
2-p-Tolylimino-thiazolidin-4-one. 2-p-Tolylimino-thiazolidin-
4-one was prepared from 1-p-tolylthiourea and ethyl chloroacetate
according to the procedure described above: yield, 91.7%, green
powder; 1H NMR (400 MHz, MeOD), δ 7.56 (d, 1H, J ) 8.2 Hz),
7.22 (dd, 2H, J ) 8.1 Hz), 7.10 (d, 1H, J ) 8.0 Hz), 4.00 (d, 2H,
J ) 35.6 Hz), 2.35 (d, 3H, J ) 11.4 Hz); ESI-MS, m/z 208.3 (M
+ 1).
2-p-Fluorophenylimino-thiazolidin-4-one. 2-p-Fluorophenyl-
imino-thiazolidin-4-one was prepared from 1-(4-fluorophenyl)-

CytoselectiVe Anticancer Thiazolidinone Analogues
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1249
1
yield, 74.1%, yellow powder; 1H NMR (400 MHz, MeOD), δ 8.12
(d, 1H, J ) 20.3 Hz), 7.93 (d, 1H, J ) 44.3 Hz), 7.48 (s, 1H), 7.38
(d, 1H, J ) 7.5 Hz), 7.16 (d, 3H, J ) 32.1 Hz), 6.81 (d, 2H, J )
33.4 Hz), 1.91, 1.19 (ss, 1H); ESI-MS, m/z 365.7 (M + 1).
2-(3-Trifluoromethyl-phenylimino)-5-(2-naphthalenyl)meth-
ylene-thiazolidin-4-one (I-172). I-172 was prepared from 2-(3-
trifluoromethyl-phenyl)imino-thiazolidin-4-one and 2-naphthalde-
hyde (AL012) according to the procedure described above: yield,
32.3%, yellow powder; 1H NMR (400 MHz, DMSO), δ 12.61,
11.83 (ss, 1H), 8.21 (m, 2H), 7.99 (m, 3H), 7.58 (m, 5H), 7.26 (s,
1H); ESI-MS, m/z 399.7 (M + 1). HR-MS calculated for
C21H13F3N2OS (M + H)+: 399.0779; found 399.0776.
2-(4-Ethoxycarbonyl-phenylimino)-5-(4-methoxy-ben-
zylidene)thiazolidin-4-one (I-225). I-225 was prepared from 2-(4-
ethoxycarbonyl-phenyl)imino-thiazolidin-4-one and 4-methoxy-
benzaldehyde (AL005) according to the procedure described above:
yield, 95.7%, yellow powder; 1H NMR (600 M Hz, DMSO), δ
12.47, 11.82 (ss, 1H), 8.02–7.74 (m, 3H), 7.61 (M, 1H), 7.47 (d,
1H, J ) 7.78 Hz), 7.16 (s, 2H), 6.98 (d, 1H, J ) 7.71 Hz), 4.32(q,
2H, J ) 7.08 Hz), 3.81 (t, 3H), 1.32 (t, 3H, J ) 7.10 Hz); ESI-
MS, m/z 383.5 (M + 1).
2-(3-Hydroxy-phenylimino)-5-(4-hydroxy-3-methoxyben-
zylidene)thiazolidin-4-one (III-272). III-272 was prepared from
2-(3-hydroxy-phenyl)imino-thiazolidin-4-one and 4-hydroxy-3-
methoxybenzaldehyde (AL003) according to the procedure
described above: yield, 72.7%, yellow powder; 1H NMR (400
MHz, MeOD), δ 7.61 (d, 1H, J ) 36.4 Hz), 7.48 (s, 2H), 7.37
(d, 1H, J ) 8.1 Hz), 7.20 (t, 1H, J ) 7.8 Hz), 6.97 (m, 3H),
6.81 (s, 1H), 3.76 (d, 3H, J ) 18.4 Hz), 2.28 (s, 3H); ESI-MS,
m/z 325.7 (M + 1).
yellow powder; H NMR (400 MHz, MeOD), δ 8.33 (d, 1H, J )
54.2 Hz), 8.08 (dd, 1H, J ) 8.5, 24.5 Hz), 7.89 (m, 2H), 7.54 (m,
4H), 7.20 (m, 1H), 6.69 (dd, 1H, J ) 7.8, 16.5 Hz), 6.53 (d, 1H,
J ) 11.5 Hz), 3.71 (d, 3H, J ) 22.5 Hz); ESI-MS, m/z 361.7 (M
+ 1).
2-(2-Methoxy-phenylimino)-5-(4-hydroxy-benzylidene)thiazo-
lidin-4-one (III-375). III-375 was prepared from 2-(2-methoxy-
phenyl)imino-thiazolidin-4-one and 4-hydroxy-benzaldehyde (AL021)
according to the procedure described above: yield, 17.7%, yellow
1
powder; H NMR (400 MHz, MeOD), δ 9.66 (m, 1H), 8.08 (s,
1H), 7.65 (m, 1H), 7.50 (s, 1H), 7.37 (s, 1H), 7.16 (d, 2H, J )
40.9 Hz), 7.01 (s, 1H), 6.90 (s, 1H), 6.77 (d, 2H, J ) 26.4 Hz),
3.79 (d, 3H, J ) 30.4 Hz), 1.91, 1.18 (ss, 1H); ESI-MS, m/z 327.7
(M + 1).
2-(3-Trifluoromethyl-phenylimino)-5-(4-hydroxy-ben-
zylidene)thiazolidin-4-one (III-376). III-376 was prepared from
2-(3-trifluoromethyl-phenyl)imino-thiazolidin-4-one and 4-hydroxy-
benzaldehyde (AL021) according to the procedure described above:
yield, 21.9%, yellow powder; 1H NMR (400 MHz, MeOD), δ 7.99
(m, 1H), 7.66 (s, 1H), 7.49 (t, 2H, J ) 7.8 Hz), 7.38 (d, 2H, J )
7.7 Hz), 7.22 (m, 2H), 6.78 (d, 2H, J ) 19.7 Hz), 2.05, 1.18 (ss,
1H); ESI-MS, m/z 375.6 (M + 1). HR-MS calculated for
C17H11F3N2O2S (M + H)+: 365.0572; found 365.0573.
N-Substituted Arylthioureas (11). General Procedure. To a
stirred solution of phenyl isothiocyanate 2 (1.0 mmol) in 5 mL of
anhydrous THF was added dropwise a solution of the substituted
amine 10 (2.0 mmol) in 2 mL of THF. After being stirred at room
temperature for 10 min, the reaction was heated to reflux in an oil
bath for 30 min. The solvent was removed under reduced pressure.
The residue was dissolved in EtOAc and extracted with 1 N HCl.
The combined organic layer was washed with brine, dried over
anhydrous Na2SO4, and concentrated under reduced pressure. The
crude product, which had a purity of >90%, was used directly for
the next step without further purification. For a typical compound,
phenylthiourea was obtained at a yield of 85%, with an HPLC purity
of >94%. ESI-MS, m/z 297.5 (M + 1).
N-Substituted 2-Arylimino-thiazolidin-4-ones (12). General
Procedure. N-Substituted primary thiourea 11 (6.0 mmol), anhy-
drous sodium acetate (2.479 g, 30 mmol), and ethyl chloroacetate
6 (1.28 mL, 12.0 mmol) were dissolved in 20 mL of absolute
ethanol. The reaction mixture was heated at 60 °C for 6 h until
thin layer chromatography (TLC) showed the disappearance of the
N-substituted thiourea. The mixture was concentrated in a vacuum,
and the residue was dissolved in EtOAc. The solution was extracted
with water and brine. The organic layer was dried over anhydrous
Na2SO4 and concentrated in a vacuum to produce crude product.
The crude product was crystallized from EtOAc to yield N-
substituted 2-arylimino-thiazolidin-4-one 12, which was used
directly for the next step reactions without further purification.
3-Phenethyl-2-phenylimino-thiazolidin-4-one. 3-Phenethyl-2-
phenylimino-thiazolidin-4-one was prepared from isothiocyanato-
benzene, 2-phenylethanamine, and ethyl 2-chloroacetate according
2-(3-Hydroxy-phenylimino)-5-(2-hydroxy-benzylidene)thiazo-
lidin-4-one (III-273). III-273 was prepared from 2-(3-hydroxy-
phenyl)imino-thiazolidin-4-one and 2-hydroxy-benzaldehyde (AL006)
according to the procedure described above: yield, 65.2%, yellow
1
powder; H NMR (400 MHz, MeOD), δ 7.34 (d, 1H, J ) 55.8
Hz), 6.76 (m, 1H), 6.50 (m, 3H), 6.24 (t, 1H, J ) 7.8 Hz), 6.10
(m, 3H), 1.56 (s, 3H); ESI-MS, m/z 311.7 (M + 1). HR-MS
calculated for C17H14N2O2S (M + H)+: 311.0854; found 311.0857.
2-(3-Hydroxy-phenylimino)-5-(4-dimethylamino-benzylidene)-
thiazolidin-4-one (III-274). III-274 was prepared from 2-(3-
hydroxy-phenyl)imino-thiazolidin-4-one and 4-dimethylamino-
benzaldehyde (AL007) according to the procedure described above:
yield, 94.1%, yellow powder; 1H NMR (400 MHz, MeOD) δ, 8.02
(s, 1H), 7.60 (s, 1H), 7.46 (s, 1H), 7.38 (s, 1H), 7.27 (s, 1H), 7.20
(t, 1H, J ) 7.8 Hz), 6.95 (d, 1H, J ) 7.3 Hz), 6.70 (m, 3H), 2.94
(d, 6H, J ) 8.2 Hz), 2.27 (d, 3H, J ) 7.8 Hz); ESI-MS, m/z 338.7
(M + 1).
2-(3-Hydroxy-phenylimino)-5-(4-methylthio-benzylidene)thi-
azolidin-4-one (III-278). III-278 was prepared from 2-(3-hydroxy-
phenyl)imino-thiazolidin-4-one and 4-methylthio-benzaldehyde
(AL018) according to the procedure described above: yield, 89.4%,
yellow powder; 1H NMR (400 MHz, MeOD), δ 7.65 (s, 1H), 7.52
(d, 1H, J ) 15.3 Hz), 7.44 (d, 1H, J ) 8.6 Hz), 7.30 (dd, 2H, J )
8.0 Hz), 7.19 (dd, 2H, J ) 5.1 Hz), 6.95 (m, 1H), 6.82 (s, 1H),
2.42 (m, 3H), 2.27 (d, 2H, J ) 6.3 Hz); ESI-MS, m/z 341.6 (M +
1). HR-MS calculated for C18H16N2OS2 (M + H)+: 341.0782; found
341.0783.
2-(3-Methoxy-phenylimino)-5-(4-hydroxy-3-methoxyben-
zylidene)thiazolidin-4-one (III-311). III-311 was prepared from
2-(3-methoxy-phenyl)imino-thiazolidin-4-one and 4-hydroxy-3-
methoxybenzaldehyde (AL003) according to the procedure de-
scribed above: yield, 34.8%, yellow powder; 1H NMR (400 MHz,
MeOD), δ 7.62 (s, 1H), 7.46 (d, 1H, J ) 23.8 Hz), 7.18 (s, 2H),
7.04 (s, 1H), 6.93 (s, 1H), 6.79 (d, 1H, J ) 21.4 Hz), 6.62 (d, 2H,
J ) 45.0 Hz), 3.76 (m, 6H), 1.91, 1.18 (ss, 1H); ESI-MS, m/z 357.7
(M + 1). HR-MS calculated for C18H16N2O4S (M + H)+: 357.0909;
found 357.0907.
1
to the procedure described above: white powder; H NMR (300
MHz, CDCl3), δ 7.37–7.23 (m, 7H), 7.14 (t, 2H, J ) 7.5 Hz), 6.91
(d, 2H, J ) 5.4 Hz), 4.11 (t, 2H, J ) 7.6 Hz), 3.75 (s, 2H), 3.04
(t, 2H, J ) 7.6 Hz); ESI-MS, m/z 297.5 (M + 1).
N-Substituted 2-Arylimino-5-arylidene-thiazolidin-4-ones (II).
General Procedure. N-Substituted 2-arylimino-thiazolidin-4-ones
12 (0.5 mmol) and benzaldehyde 8 (0.6 mmol) were dissolved in
3 mL of absolute ethanol. Next, 50 µL of piperidine (0.5 mmol)
was added to this mixture, and the mixture was stirred for 12 h at
60 °C until precipitation formation. After the mixture was cooled
to room temperature, the precipitation was filtered and washed
with petroleum ether and absolute ethanol to yield N-subsituted
2-arylimino-5-arylidene-thiazolidin-4-one at a purity of >90%.
Reaction mixtures that did not yield precipitates were concen-
trated under reduced pressure to remove ethanol solvent. The residue
was dissolved in EtOAc and extracted with 1 N HCl and brine.
After being dried by anhydrous Na2SO4 and concentrated in a
vacuum, the crude compound was purified by silica gel column
chromatography and eluted with EtOAc and petroleum ether.
2-(3-Methoxy-phenylimino)-5-(2-naphthalenyl)methylene-
thiazolidin-4-one (III-316). III-316 was prepared from 2-(3-
methoxy-phenyl)imino-thiazolidin-4-one and 2-naphthaldehyde
(AL012) according to the procedure described above: yield, 34.8%,

1250 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Zhou et al.
5-(4-Dimethamino-benzylidene)-2-phenylimino-3-(2-hydro-
ethyl)-thiazolidin-4-one (II-18). II-18 was prepared from 3-(2-
hydroxyethyl)-2-phenylimino-thiazolidin-4-one and 4-dimethamino-
benzaldehyde (AL007) according to the procedure described above:
yield, 71.0%, yellow powder; 1H NMR (400 MHz, MeOD), δ 7.52
(s, 1H), 7.26 (m, 4H), 7.08 (t, 1H, J ) 7.5 Hz), 6.91 (t, 2H, J )
9.6 Hz), 6.64 (d, 2H, J ) 9.0 Hz), 4.02 (t, 2H, J ) 6.0 Hz), 3.78
(t, 2H, J ) 5.9 Hz), 2.92 (d, 6H, J ) 10.0 Hz); ESI-MS, m/z 368.7
(M + 1). HR-MS calculated for C20H21N3O2S (M + H)+: 368.1433;
found 368.1447.
5-(4-Hydroxy-3-methoxybenzylidene)-2-phenylimino-3-(2-hy-
droethyl)-thiazolidin-4-one (II-20). II-18 was prepared from 3-(2-
hydroxyethyl)-2-phenylimino-thiazolidin-4-one and 4-hydroxy-3-
methoxybenzaldehyde (AL003) according to the procedure described
above: yield, 70.0%, yellow powder; 1H NMR (400 MHz, MeOD),
δ 8.05 (s, 1H), 7.40 (t, 2H, J ) 7.9 Hz), 7.19 (t, 1H, J ) 7.5 Hz),
7.05 (d, 2H, J ) 8.4 Hz), 6.93 (s, 1H), 6.73 (s, 1H), 4.15 (t, 2H, J
) 5.9 Hz), 3.89 (m, 8H), 3.71 (s, 3H); ESI-MS, m/z 415.8 (M +
1). HR-MS calculated for C21H22N2O5S (M + H)+: 415.1328: found
415.1324.
5-(2,4,5-Trimethoxy-benzylidene)-2-phenylimino-3-(3-meth-
oxy-propyl-thiazolidin-4-one (II-25). II-25 was prepared from
3-(3-methoxy-propyl)-2-phenylimino-thiazolidin-4-one and 2,4,5-
trimethoxy-benzaldehyde (AL014) according to the procedure
described above: yield, 77.0%, yellow powder; 1H NMR (400 MHz,
MeOD), δ 7.91 (s, 1H), 7.27 (t, 2H, J ) 7.7 Hz), 7.07 (t, 1H, J )
7.5 Hz), 6.92 (d, 2H, J ) 8.4 Hz), 6.78 (s, 1H), 6.61 (d, 1H, J )
13.2 Hz), 3.96 (t, 2H, J ) 6.9 Hz), 3.79 (d, 6H, J ) 4.5 Hz), 3.58
(s, 3H), 3.41 (t, 2H, J ) 6.0 Hz), 3.23 (s, 3H), 1.94 (p, 2H, J ) 6.5
Hz); ESI-MS, m/z 443.7 (M + 1). HR-MS calculated for
C23H26N2O5S (M +H)+: 443.1641; found 443.1648.
General Procedure for the Library Synthesis. Using a 10 ×
10 library as an example, 10 intermediates 7 or 12 (5 mmol) were
dissolved in 10 mL of absolute ethanol. Ultrasonic and heating baths
were used to make sure they were all dissolved. Each solution was
divided into 10 parts. Next, 10 different benzylaldehydes (6 mmol)
were dissolved in 10 mL of absolute ethanol and divided into 10
parts following the same procedure. Each of the 100 reaction vessels
contained a solution of intermediate 7 or 12 (0.5 mmol) and different
benzylaldehydes (0.6 mmol) in 2 mL of absolute ethanol. After
piperidine was added (50 µL, 0.5 mmol), all reaction vessels were
put into a Zhicheng ZHMY-113H parallel synthesizer and shaken
overnight. The conversion of the reactions was monitored by HPLC/
MS. After reaction completion, the reaction vessels were cooled
to room temperature, and the solutions were simultaneously filtered
and washed with absolute ethanol repeatedly to yield the target
products. All of the residues were dried in a vacuum at 60 °C
overnight and analyzed by an HPLC/MS system.
Time-Dependent Compound Effects on Cell Proliferation
and Cell Morphology. To examine the time-dependent response
of the cells to these compounds, compounds I-7, I-27, I-47, and
I-67 were dissolved in DMSO to a concentration of 15 mM and
stored at 4 °C. The cytotoxicity of compounds was studied in 24-
well plates. H460, H460taxR, and NHFB cells (3 × 104 cells in 1000
µL of culture medium/well) were seeded in 24-well flat-bottom
plates and treated the next day with the compounds at 1 and 2 µM,
respectively, for up to 48 h. An equal volume of DMSO (0.1%)
had no effect on cell viability and was used as a control. The cell
morphology changes were observed under an Olympus I X 71
phase-contrast microscope (Center Valley, PA).
P-gp Substrate Pharmacophore Modeling. Briefly, a P-gp
substrate pharmacophore36 derived from a common feature HIPHOP
alignment of verapamil and digoxin was used as described
previously.35,36 All molecules evaluated in this study were imported
as sdf files, and up to 255 conformations were generated with the
Best conformer generation method, allowing a maximum energy
difference of 20 kcal/mol with Discovery Studio 1.7 Catalyst
(Accelrys, San Diego, CA). Molecules were mapped to the P-gp
substrate pharmacophore using rigid fit and allowing a maximum
of two feature misses. The larger the fit value, the closer the
molecule maps to the feature centroids.
Anticancer Compound Pharmacophore Modeling. All of the
active molecules in Table 3 were imported as sdf files, and up to
255 conformations were generated with the Best conformer
generation method, allowing a maximum energy difference of 20
kcal/mol. Molecules I-7, I-27, I-47, and I-67 were annotated as
more active than the others, and a HIPHOP alignment was
performed36 after the selection of hydrogen bond acceptor, donor,
and hydrophobic features. The most active molecules were then
aligned on the final pharmacophore, and a translucent van der Waals
surface shape was added.
Acknowledgment. We thank Hong Ruan and Anang Shelat
for cheminformatics assistance and Ms. Sulian Gao for technical
assistance. This work was supported by the American Lebanese
Syrian Associated Charities (ALSAC), St. Jude Children’s
Research Hospital, and Shandong University.
Supporting Information Available: Details of library charac-
terizations, including purity, yield, LC/MS, and NMR data, and
cell morphology changes. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Green, D. R.; Evan, G. I. A matter of life and death. Cancer Cell
2002, 1, 19–30.
(2) Kaelin, W. G. The concept of synthetic lethality in the context of
anticancer therapy. Nat. ReV. Cancer 2005, 5, 689–698.
(3) Fojo, A. T.; Ueda, K.; Slamon, D. J.; Poplack, D. G.; Gottesman,
M. M.; Pastan, I. Expression of a multidrug-resistance gene in human-
tumors and tissues. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 265–269.
(4) Gottesman, M. M.; Fojo, T.; Bates, S. E. Multidrug resistance in cancer:
role of ATP-dependent transporters. Nat. ReV. Cancer 2002, 2, 48–
58.
(5) Kartner, N.; Riordan, J. R.; Ling, V. Cell-surface-glycoprotein
associated with multidrug resistance in mammalian-cell lines. Science
1983, 221, 1285–1288.
(6) Ferry, D. R.; Traunecker, H.; Kerr, D. J. Clinical trials of P-
glycoprotein reversal in solid tumours. Eur. J. Cancer 1996, 32A,
1070–1081.
(7) Samuels, B. L.; Hollis, D. R.; Rosner, G. L.; Trump, D. L.; Shapiro,
C. L.; Vogelzang, N. J.; Schilsky, R. L. Modulation of vinblastine
resistance in metastatic renal cell carcinoma with cyclosporine a or
tamoxifen: a cancer and leukemia group B study. Clin. Cancer Res.
1997, 3, 1977–1984.
(8) Smith, A. J.; Mayer, U.; Schinkel, A. H.; Borst, P. Availability of
PSC833, a substrate and inhibitor of P-glycoproteins, in various
concentrations of serum. J. Natl. Cancer Inst. 1998, 90, 1161–1166.
(9) Ottana, R.; Maccari, R.; Barreca, M. L.; Bruno, G.; Rotondo, A.; Rossi,
A.; Chiricosta, G.; Di Paola, R.; Sautebin, L.; Cuzzocrea, S.; Vigorita,
M. G. 5-Arylidene-2-imino-4-thiazolidinones: design and synthesis of
novel anti-inflammatory agents. Bioorg. Med. Chem. 2005, 13, 4243–
4252.
Cells and Culture Conditions. The human NSCLC H460 cell
line and paclitaxel-resistant and P-gp-overexpressing H460taxR cell
lines were routinely propagated in monolayer culture in RPMI 1640
medium supplemented with 10% heat-inactivated fetal calf serum,
100 units/mL penicillin, and 100 mg/mL streptomycin. NHFBs were
maintained in Dulbecco’s modified Eagle’s medium with the same
supplements. All cells were maintained in the presence of 5% CO2
at 37 °C.
Cytotoxicity Assays. The inhibitory effects of thiazolidinone
analogues on cell growth were determined by the sulforhodamine
B assay. Cells [(2–8) × 103 cells in 100 µL of culture medium/
well] were seeded in 96-well flat-bottom plates and treated the next
day with the drugs at the indicated concentrations. The IC50 was
determined by sulforhodamine B assay 72 h after treatment, as
described previously.40,41 DMSO solvent was used to dissolve
compounds, and its volume was <1% in the assay medium. An
equal volume of DMSO was used as a control. The experiments
were performed at least three times for each cell line. Cell viability
was calculated using the following formula: cell viability ) 100 ×
Atreatment/Acontrol (percentage). The IC50 was determined by the
SigmaPlot 10.0 program (Systat Software, Inc., San Jose, CA) using
a four-parameter logistic function for the sigmoid dose–response
curves.

CytoselectiVe Anticancer Thiazolidinone Analogues
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1251
(10) Vicini, P.; Geronikaki, A.; Anastasia, K.; Incerti, M.; Zani, F. Synthesis
and antimicrobial activity of novel 2-thiazolylimino-5-arylidene-4-
thiazolidinones. Bioorg. Med. Chem. 2006, 14, 3859–3864.
(11) Gududuru, V.; Hurh, E.; Dalton, J. T.; Miller, D. D. Synthesis and
antiprofilerative activity of 2-arly-4-oxo-thiazolidin-3-yl-amides for
prostate cancer. Bioorg. Med. Chem. Lett. 2004, 14, 5289–5293.
(12) Ottana, R.; Carotti, S.; Maccari, R.; Landini, I.; Chiricosta, G.; Caciagli,
B.; Vigorita, M. G.; Mini, E. In vitro antiproliferative activity against
human colon cancer cell lines of representative 4-thiazolidinones. Part
I. Bioorg. Med. Chem. Lett. 2005, 15, 3930–3933.
(13) Elbarbary, A. A.; Khodair, A. I.; Pedersen, E. B.; Nielsen, C. Synthesis
and evaluation of antiviral activity of 2′-deoxyuridines with 5-meth-
ylene-2-thiohydantoin substituents in the 5-position. Monatsh. Chem.
1994, 125, 593–598.
(14) Rydzik, E.; Szadowska, A.; Kaminska, A. Synthesis of benzylidene
derivatives of 3-ortho-chlorophenylhydantoin, 3-meta-chlorophenyl-
hydantoin, and 3-para-chlorophenylhydantoin and investigation of their
anticonvulsant action. Acta Pol. Pharm. 1984, 41, 459–464.
(15) Troutman, H. D.; Long, L. M. The synthesis of 2,3-disubstituted-4-
thiazolidones. J. Am. Chem. Soc. 1948, 70, 3436–3439.
(16) Liu, H. L.; Li, Z. C.; Anthonsen, T. Synthesis and fungicidal activity
of 2-imino-3-(4-arylthiazol-2-yl)-thiazolidin-4-ones and their 5-arylidene
derivatives. Molecules 2000, 5, 1055–1061.
(17) Kalaiya, S. B.; Parikh, A. R. Studies on 4-thiazolidinones-preparation
and antimicrobial activity of R-phenyl-2-(2′,5′-dihydroxy-3′-bro-
mophenyl)-3-aryl-5-carboxymethyl-4-thiazolidinones, R-methyl-2-(2′-
hydroxy-3′-nitro-5′-benzyl)-3-aryl-5-carboxymethyl-4-thiazolidino-
nes and R-methyl-2-(2′-hydroxy-3′-nitro-5′-chlorophenyl)-3-aryl-5-
carboxymethyl-4-thizaolidinones. J. Indian Chem. Soc. 1987, 64, 172–
175.
(18) Dayam, R.; Aiello, F.; Deng, J. X.; Wu, Y.; Garofalo, A.; Chen, X. Y.;
Neamati, N. Discovery of small molecule integrin Rvꢀ3 antagonists
as novel anticancer agents. J. Med. Chem. 2006, 49, 4526–4534.
(19) Teraishi, F.; Wu, S. H.; Sasaki, J.; Zhang, L. D.; Zhu, H. B.; Davis,
J. J.; Fang, B. L. P-glycoprotein-independent apoptosis induction by
a novel synthetic compound, MMPT [5-[(4-methylphenyl)methylene]-
2-(phenylamino)-4 (5H)-thiazolone]. J. Pharmacol. Exp. Ther. 2005,
314, 355–362.
(20) Wu, S.; Guo, W.; Teraishi, F.; Pang, J.; Kaluarachchi, K.; Zhang, L.;
Davis, J.; Dong, F.; Yan, B.; Fang, B. Anticancer activity of
5-benzylidene-2-phenylimino-1, 3-thiazolidin-4-one (BPT) analogs.
Med. Chem. 2006, 2, 597–605.
(21) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J.
Experimental and computational approaches to estimate solubility and
permeability in drug discovery and development settings. AdV. Drug
DeliVery ReV. 2001, 46, 3–26.
iminothiazolidinones containing the benzo[b]thiophene moiety. J. Serb.
Chem. Soc. 2005, 70, 153–161.
(26) Blanchet, J.; Zhu, J. P. Reeve’s synthesis of 2-imino-4-thiazolidinone
from alkyl (aryl) trichloromethylcarbinol revisited, a three-component
process from aldehyde, chloroform and thiourea. Tetrahedron Lett.
2004, 45, 4449–4452.
(27) Holmes, C. P.; Chinn, J. P.; Look, G. C.; Gordon, E. M.; Gallop,
M. A. Strategies for combinatorial organic-synthesis-solution and
polymer-supported synthesis of 4-thiazolidinones and 4-metathiaza-
nones derived from amino-acids. J. Org. Chem. 1995, 60, 7328–7333.
(28) Srivastava, T.; Haq, W.; Katti, S. B. Carbodiimide mediated synthesis
of 4-thiazolidinones by one-pot three-component condensation. Tet-
rahedron 2002, 58, 7619–7624.
(29) Bae, S.; Hahn, H. G.; Dal Nam, K. Syntheses of 1,3-imidazoline-2-
thione and 2-phenylimino-1,3-thiazoline combinatorial libraries through
different sequences of the same components. J. Comb. Chem. 2005,
7, 826–836.
(30) Bae, S.; Hahn, H. G.; Dal Nam, K.; Mah, H. Solid-phase synthesis of
fungitoxic 2-imino-1,3-thiazolines. J. Comb. Chem. 2005, 7, 7–9.
(31) Nefzi, A.; Giulianotti, M.; Truong, L.; Rattan, S.; Ostresh, J. M.;
Houghten, R. A. Solid-phase synthesis of linear ureas tethered to
hydantoins and thiohydantoins. J. Comb. Chem. 2002, 4, 175–178.
(32) Sim, M. M.; Ganesan, A. Solution-phase synthesis of a combinatorial
thiohydantoin library. J. Org. Chem. 1997, 62, 3230–3235.
(33) Brummond, K. M.; Lu, J. Solid-phase synthesis of BRL 49653. J.
Org. Chem. 1999, 64, 1723–1726.
(34) Tomkinson, N. C. O.; Sefler, A. M.; Plunket, K. D.; Blanchard, S. G.;
Parks, D. J.; Willson, T. M. Solid-phase synthesis of hybrid thiazo-
lidinedione-fatty acid PPAR γ ligands. Bioorg. Med. Chem. Lett. 1997,
7, 2491–2496.
(35) Chang, C.; Bahadduri, P. M.; Polli, J. E.; Swaan, P. W.; Ekins, S.
Rapid identification of P-glycoprotein substrates and inhibitors. Drug
Metab. Dispos. 2006, 34, 1976–1984.
(36) Ekins, S.; Kim, R. B.; Leake, B. F.; Dantzig, A. H.; Schuetz, E. G.;
Lan, L. B.; Yasuda, K.; Shepard, R. L.; Winter, M. A.; Schuetz, J. D.;
Wikel, J. H.; Wrighton, S. A. Application of three-dimensional
quantitative structure-activity relationships of P-glycoprotein inhibitors
and substrates. Mol. Pharmacol. 2002, 61, 974–981.
(37) Mullally, J. E.; Fitzpatrick, F. A. Pharmacophore model for novel
inhibitors of ubiquitin isopeptidases that induce p53-independent cell
death. Mol. Pharmacol. 2002, 62, 351–358.
(38) Mullally, J. E.; Moos, P. J.; Edes, K.; Fitzpatrick, F. A. Cyclopentenone
prostaglandins of the J series inhibit the ubiquitin isopeptidase activity
of the proteasome pathway. J. Biol. Chem. 2001, 276, 30366–30373.
(39) Verbitski, S. M.; Mullally, J. E.; Fitzpatrick, F. A.; Ireland, C. M.
Punaglandins, chlorinated prostaglandins, function as potent Michael
receptors to inhibit ubiquitin isopeptidase activity. J. Med. Chem. 2004,
47, 2062–2070.
(40) Monks, A.; Scudiero, D.; Skehan, P.; Shoemaker, R.; Paull, K.; Vistica,
D.; Hose, C.; Langley, J.; Cronise, P.; Vaigrowolff, A.; Graygoodrich,
M.; Campbell, H.; Mayo, J.; Boyd, M. Feasibility of a high-flux
anticancer drug screen using a diverse panel of cultured duman tumor-
cell lines. J. Natl. Cancer Inst. 1991, 83, 757–766.
(41) Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.; McMahon, J.;
Vistica, D.; Warren, J. T.; Bokesch, H.; Kenney, S.; Boyd, M. R. New
colorimetric cytotoxicity assay for anticancer-drug screening. J. Natl.
Cancer Inst. 1990, 82, 1107–1112.
(22) St. Laurent, D. R.; Gao, Q.; Wu, D.; Serrano-Wu, M. H. Regioselective
synthesis of 3-(heteroaryl)-iminothiazolidin-4-ones. Tetrahedron Lett.
2004, 45, 1907–1910.
(23) Kasmi-Mir, S.; Djafri, A.; Paquin, L.; Hamelin, J.; Rahmouni, M. One-
pot synthesis of 5-arylidene-2-imino-4-thiazolidinones under micro-
wave irradiation. Molecules 2006, 11, 597–602.
(24) El’tsov, O. S.; Mokrushin, V. S.; Bel’skaya, N. P.; Kozlova, N. M.
Synthesis and transformations of (imidazolylimino)thiazolidinones.
Russ. Chem. Bull. 2003, 52, 461–466.
(25) Kachhadia, V. V.; Patel, M. R.; Joshi, H. S. Heterocyclic systems
containing S/N regioselective nucleophilic competition: facile syn-
thesis, antitubercular and antimicrobial activity of thiohydantoins and
JM7012024