Structure-activity requirements for the antiproliferative effect of troglitazone derivatives mediated by depletion of intracellular calcium

Article abstract of DOI:10.1016/j.bmcl.2004.02.087

Depletion of calcium from the endoplasmic reticulum has shown to affect protein synthesis and cell proliferation. The anticancer effect of troglitazone was reported to be mediated by depletion of intracellular calcium stores resulting in inhibition of translation initiation. The unsaturated form of troglitazone displays similar anticancer properties in vitro. In this letter, we report our findings on the minimum structural requirements for both compounds to retain their calcium release and antiproliferative activities.

Full text of DOI:10.1016/j.bmcl.2004.02.087

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2547–2550
Structure–activity requirements for the antiproliferative effect of
troglitazone derivatives mediated by depletion of intracellular
calcium
Yun-Hua Fan,a Han Chen,a Amarnath Natarajan,a Yuhong Guo,a Fred Harbinski,a
Julia Iyasere,a William Christ,a Huseyin Aktasa and Jose A. Halperina,b,
*
aLaboratory for Translational Research, Harvard Medical School, Cambridge, MA 02139, USA
bDepartment of Medicine, Brigham and Women’s Hospital, Boston, MA 02115, USA
Received 7 January 2004; revised 16 February 2004; accepted 26 February 2004
Abstract—Depletion of calcium from the endoplasmic reticulum has shown to affect protein synthesis and cell proliferation. The
anticancer effect of troglitazone was reported to be mediated by depletion of intracellular calcium stores resulting in inhibition of
translation initiation. The unsaturated form of troglitazone displays similar anticancer properties in vitro. In this letter, we report
our findings on the minimum structural requirements for both compounds to retain their calcium release and antiproliferative
activities.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
leading to the phosphorylation of the a subunit of eIF2
(eIF2a) and inhibition of translation initiation. Both
CLT and TZD inhibit growth of cancer cells in vitro at
micromolar concentrations, and also tumor growth in
experimental cancer models.
Calcium ions play a critical role in key biological pro-
cesses.1–3 The major intracellular calcium store is the
endoplasmic reticulum (ER). Depletion of ER Ca2þ
stores promote ER stress4–6 and that leads to the phos-
phorylation of eukaryotic initiation factor 2 (eIF2) and
thereby to inhibition of translation initiation.7–11
Translation initiation plays an important role in regu-
lation of cell proliferation and malignant transformation
and is therefore an attractive target for the development
of mechanism-specific anticancer drugs.
Troglitazone (TRO), an agonist ligand for the nuclear
receptor peroxisome proliferator-activated receptor-c
(PPAR-c), improves insulin sensitivity, induces terminal
differentiation of normal pre-adipocytes and human
liposarcoma cells in vitro15 and also exhibits anticancer
activity in other cancer cell types.13;16 We have previ-
ously shown that the anticancer activity of TRO is
independent of PPAR-c and mediated by Ca2þ release
mediated inhibition of translation initiation.13;17
Recently, we have observed that the unsaturated form of
TRO (1b), and ciglitazone release Ca2þ from intracel-
lular stores and inhibit cell growth in the lung cancer cell
line, A549, in a manner similar to TRO (1a), while
rosiglitazone, a more potent member of the TZD family,
does not release Ca2þ nor inhibit cell proliferation (Fig.
1). Ciglitazone and rosiglitazone differ from TRO only
in the substitution on the 40-hydroxy. We have, there-
fore, investigated the minimum requirement on the
40-hydroxy substitution of the TZD type compounds
needed to maintain their translation initiation inhibitory
activity (Fig. 2).
We have previously demonstrated that eicosapentaenoic
acid (EPA),12 thiazolidinediones (TZD),13 and clotrim-
azole (CLT),14 exhibit anticancer properties mediated by
inhibition of translation initiation. These compounds
cause partial and sustained depletion of intracellular
Ca2þ stores because they simultaneously release Ca2þ
from internal stores and block store-operated-calcium
channels (SOC) that open to refill the internal Ca2þ
stores. This partial depletion of intracellular Ca2þ stores
causes activation of eIF2 kinases (PKR or/and PERK)
Keywords: Troglitazone; Intracellular calcium; Thiazolidinediones.
* Corresponding author. Tel.: +1-6176216104; fax: +1-6176216148;
e-mail: yunhuafan@yahoo.com
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.02.087

2548
Y.-H. Fan et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2547–2550
Figure 1.
Figure 2.
In the past we have observed a good correlation among
intracellular Ca2þ release, phosphorylation of eIF2a,
and inhibition of cell proliferation.12–14 Therefore, in this
communication we decided to use these three bioassays
to measure the activities of our compounds and to
determine their mechanisms of action.
Scheme 2. (a) t-Butyldimethylsilyl chloride, imidazole, DMF; (b)
LAH, rt, 3 h (75.9%, two steps); (c) 4-fluorobenzaldehyde, KtOBu,
DMF, 80 ꢁC, 8 h; (d) 2,4-thiazolidinedione, AcOH, piperidine, toluene,
reflux, 4 h (37%, two steps); (e) HCl, MeOH, 15 min; (f) CoCl2, DMG
(84%).
Compounds 2–1318 were all synthesized in a similar
fashion as outlined in Scheme 1. The reaction conditions
were not optimized for the individual compounds. Dif-
ferent linear and cyclic alkanols were O-alkylated with
4-fluorobenzaldehyde using potassium tert-butoxide as
base in DMSO at 80 ꢁC for 8 h. The desired crude
O-alkylated aldehydes underwent Knoevenagel con-
densation with 2,4-thiazolidinedione in refluxing toluene
for 3–4 h. The 5-benzylidene thiazolidinedione deriva-
tives (b) were purified by recrystallization from aceto-
nitrile. The 5-benzylidene thiazolidinediones was then
reduced with CoCl2 and NaBH4 to yield the 5-benzyl
derivatives (a). The final products (a) were purified by
silica gel flash chromatography. Compounds 1a and 1b
were synthesized in six steps with moderate yield as
shown in Scheme 2.19;20 All the compounds were dried
under vacuum overnight before use. All the compounds
had greater than 95% purity as determined by 1H NMR
and LC–MS.
the phosphorylation assay, cells were treated with drugs
for 2 h. Inhibition of cell growth was determined at
concentrations ranging from 0.5 to 100 lM. In the cell
growth assay, the cells were treated with drug for 5 days.
The intracellular Ca2þ release was assayed using fluo-4
loaded cells,21 and phosphorylation of eIF2a was
determined by western blotting extracts of treated cells
with anti-phospho-serine 51 eIF2a-specific antibodies.14
Inhibition of cell proliferation was measured using the
sulforhodamine B (SRB) assay.22 The results are sum-
marized in Table 1.
2. Results and discussion
The parent compounds (1a and 1b) are the most potent
compounds in the series in terms of their Ca2þ release
and antiproliferative activities. The unsaturated thiazo-
lidinediones are generally more potent than their satu-
rated counterparts in the cell proliferative assay.
Interestingly, Reddy et al.23 have observed a similar
trend in the euglycemic and hypolipidemic activities for
some of the unsaturated thiazolidinediones. The reason
for the enhanced potency of the unsaturated thiazolid-
inediones is still not clear. However, this result suggests
that the exocyclic double bond may improve the anti-
proliferative activities of these compounds, although it is
not required.
Mechanism related biological activities were measured
by testing all compounds for intracellular Ca2þ release
and phosphorylation of eIF2a at 80 and 15 lM,
respectively. Different concentrations were used to
compensate for the different durations of the assays. In
the Ca2þ release assay cells were treated for 10 min. In
Some of the structurally simpler analogs in this series
still caused the release of Ca2þ, phosphorylation of
eIF2a, and inhibition of cell growth. Compounds 12a,b
and 13a,b have completely lost the ability to induce the
release of Ca2þ from intracellular stores suggesting that
the substitution on the 40-hydroxyl group is essential for
Scheme 1. (a) KtOBu, DMF, 80 ꢁC; (b) 2,4-thiazolidinedione, AcOH,
piperidine, toluene, reflux; (c) CoCl2, DMG.

Y.-H. Fan et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2547–2550
2549
Table 1. Lipophilicity and structure–activity relationships of the troglitazone-derived series (TRO)
Compoundsa
R
c Log Pb
Ca2þc
eIF2-Pd
IC50 (lM)e
Yield (%)f
Mp (ꢁC)g
1a (Tro)
1b
2a
4.46
5.31
2.29
2.02
3.94
3.66
3.38
3.21
2.81
2.54
2.26
1.99
3.76
3.49
3.36
3.09
2.96
2.69
2.43
2.16
1.91
1.63
1.32
1.05
1.99
1.71
3.67
3.40
5.07
+
+
)
+
+
+
+
+
+
+
)
)
+
+
+
)
)
+
+
)
+
15 2.1
1.2
84
24
50
49
52
41
45
33
46
39
48
32
54
30
46
38
44
32
50
51
35
62
43
66
39
58
46
63
––
179.6–182.1
205.5–208.2
125.4–127.3
184.6–187.4
122.2–124.1
190.4–192.8
101.7–102.9
155.3–158.1
98.6–101.2
177.7–179.4
86.9–89.2
5
>100
Tetrahydropyranylmethyl
Tetrahydropyranylmethyl
Cyclohexylmethyl
Cyclohexylmethyl
Cyclopentylmethyl
Cyclopentylmethyl
Cyclobutylmethyl
Cyclobutylmethyl
Cyclopropylmethyl
Cyclopropylmethyl
tert-Amyl
2b
3a
22 2.1
26 1.9
14 1.7
25 3.6
10 1.0
>100
3b
4a
+
+
+
)
+
)
4b
5a
5b
6a
13 1.6
>100
6b
7a
23 4.5
23 1.6
15 0.9
25 3.3
16 2.8
45 3.4
189.7–191.3
160.6–161.9
182.6–184.1
87.7–89.5
+
+
)
7b
8a
tert-Amyl
iso-Butyl
8b
9a
iso-Butyl
n-Propyl
142.7–145.2
81.4–83.4
9b
n-Propyl
n-Ethyl
7
0.9
186.3–188.4
109.5–110.9
145.4–148.0
111.5–113.6
216.5–217.9
140.7–144.0
301.9–303.3
78.5–80.1
10a
10b
11a
11b
12a
12b
13a
13b
14a
14b
Ciglitazoneh
)
)
)
)
)
)
)
)
)
)
+
>100
27 2.6
>100
n-Ethyl
Methyl
)
)
Methyl
H
23 1.2
>100
H
)
)
)
)
)
+
20 1.0
>100
>100
(Without OH)
(Without OH)
Benzyl
122.9–124.3
123.4–125.6
223.7–225.2
––
>100
15 0.9
20 2.4
Benzyl
a In compounds 1a–13a the thiazolidine ring is substituted by a benzyl moiety, and in compounds 1b–13b the thiazolidine ring is substituted by a
benzylidene moiety.
b c Log P’s are calculated log P values generated by CHEMDRAW program.
c Release of Ca2þ from intracellular stores was observed using fluo-4 loaded cells. The +, ), and
notations indicate release, no release, and
insignificant release, respectively. The experiment was done with clotrimazole, a known Ca2þ releaser, as a positive control.
d Phosphorylation of eIF2a was determined by western blotting. The +, ), and notations indicate phosphorylation, no phosphorylation, and
insignificant phosphorylation, respectively. The experiment was done with Tro and DMSO as the positive and negative controls, respectively.
e IC50 was measured using sulforhodamine B assay. The values indicate the concentration needed to inhibit 50% of cell proliferation. The experiment
was done in triplicate.
f The yields for the b series are the overall yield. The yields for the a series are the yield of the reduction step from the correspond compounds from b
series.
g The melting points were determined on a Mel-Tempꢂ apparatus from Electrothermal with a thermocouple thermometer from Barnant.
h Ciglitazone was purchased from Sigma, St. Louis, MO.
the monitored activities. The size of the substituents at
the 40-hydroxyl group also seems to be important for
their activities. In most cases the potencies improve as
the O-alkyl substituents become longer and bulkier. The
cyclopropyl analogs (6a and 6b) in the cycloalkyl series
and compounds 9a,b–13a,b with fewer than four car-
bons in the linear alkyl chain series completely lost their
ability to deplete Ca2þ from the intracellular stores.
stores and inhibited SOC channels. In the saturated
series a, only compounds with c log P greater than 2.7
induce the release of Ca2þ; in the unsaturated series b,
with the exception of 2b, only compounds with c log P
greater than 2.5 induce the release of Ca2þ. However,
this observation holds true only in the aliphatic substi-
tuted compounds. When the substituent was a benzyl
group, both compounds (14a and 14b) lost their ability
to release Ca2þ
.
Contrary to our prediction, compound 2a substituted by
the tetrahydropyranyl moiety was devoid of any intra-
cellular Ca2þ releasing activity, possibly explained by the
lower lipophilicity of the compound. Fischer et al.24
have observed a similar phenomenon in the rat baso-
philic leukemia cell (RBL-2H3) in which compounds
with high lipophilicity released Ca2þ from intracellular
Drug-induced phosphorylation of eIF2a suggests inhi-
bition of translation initiation. In most cases, com-
pounds that release Ca2þ also phosphorylate eIF2a and
inhibit cell proliferation, and compounds that do not
release Ca2þ also fail to cause phosphorylation of eIF2a
and inhibition of cell proliferation. A few compounds

2550
Y.-H. Fan et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2547–2550
17. Debrock, G.; Vanhentenrijk, V.; Sciot, R.; Debiec-Rych-
ter, M.; Oyen, R.; Oosterom, A. V. Brit. J. Cancer 2003,
89, 1409.
(5a and 8a), which cause some release of intracellular
Ca2þ without phosphorylating eIF2a, as well as those
compounds that inhibit cell growth without releasing
Ca2þ from intracellular stores are likely to affect cell
proliferation by a mechanism(s) other than inhibition of
translation initiation.
18. All the compounds were prepared following the outlined
procedure preparation of 2b: Potassium tert-butoxide (1 g,
8.9 mmol) in 7.5 mL of DMF was added to a solution of
tetrahydropyran-2-methanol (1 mL, 8.4 mmol) at ambient
temperature followed by the addition of 4-fluorobenzal-
dehyde (0.95 mL, 8.9 mmol). The reaction mixture turned
dark as the benzaldehyde was added. The final mixture
was heated to 80 ꢁC under N2 for 8 h. The mixture was
concentrated in vacuo then dissolved in toluene and
washed with water and brine. The crude product was then
concentrated to dryness, and was taken to the next step
without further purification. The entire crude product was
dissolved in fresh toluene (8 mL) then 2,4-thiazolidien-
dione (1 g, 8.5 mmol) was added followed by catalytic
amount of piperidine (0.08 mL) and glacial acetic acid
(0.08 mL). The final mixture was refluxed for 4 h. The
product (precipitated out upon cooling) was filtered and
washed with acetonitrile and recrystallized with hot
acetonitrile. The product was obtained as yellow needles
(mp 184.6–187.4 ꢁC, 1.35 g 49% yield): 1H NMR (DMSO-
d6, 500 mHz) d 12.51 (s, 1H), 7.74 (s, 1H), 7.53 (d,
J ¼ 8:5 Hz, 2H), 7.09 (d, J ¼ 8:5 Hz, 2H), 3.97 (t,
J ¼ 3 Hz, 2H), 3.87 (m, 1H), 3.63 (m, 1H), 3.36 (m, 1H),
1.81 (m, 1H), 1.63 (m, 1H), 1.43–1.52 (m, 1H), 1.31 (m,
1H); MS (APCIþ): m=z 319.80 (M+H).
These observations suggest that the combination of
Ca2þ release and eIF2a phosphorylation assays allow us
to distinguish between antiproliferative compounds that
act via Ca2þ induced inhibition of translation initiation
and those that employ other mechanisms. In summary,
we were able to identify the minimal structural
requirement for TRO as Ca2þ depleting translation ini-
tiation inhibitors.
Acknowledgements
The authors wish to thank Drs. Daniel C. Tosteson and
Magdalena Tosteson for their continuous support. In
addition, we thank Dr. Michael Chorev for his com-
ments and discussion. These studies were supported in
part by NIH National Cooperative Drug Discovery
Group (NCDDG) grant U19 CA87427 and NIH
CA78411.
Preparation of 2a: Compound 2b (0.13 g, 0.4 mmol),
CoCl2 (0.067 mg, 0.28 lmol) and trace amount of DMG
were dissolved in a mixture of solution composed of H2O/
THF/1 N NaOH (0.44:0.26:0.28 mL) at ambient tempera-
ture followed by the addition of NaBH4 (23.6 mg,
0.62 mmol) in 0.2 N NaOH (0.36 mL). Reaction progress
was accompanied by change of color from deep-purple to
yellow. A few drops of acetic acid were added to the
mixture when the reaction turned yellow (the reaction
should turn purple if not completed). The reaction was
monitored by TLC and terminated 3 h after completion.
The reaction was quenched by acetone (1 mL) then
allowed to stir for 15 min. The product was extracted into
CH2Cl2 (3 · 15 mL) and then concentrated in vacuo. The
crude product was purified by chromatography on a silica
gel column with a hexane/TBME (3:1) solution. A white
powder was obtained (62 mg, 50%, mp 125.4–127.3 ꢁC):
1H NMR (DMSO-d6, 500 mHz) d 11.99 (s, 1H), 7.12 (d,
J ¼ 8 Hz, 2H), 6.85 (d, J ¼ 8 Hz, 2H), 4.85 (dd, J ¼ 9:5,
4 Hz, 1H), 3.82–3.88 (m, 3H), 3.59 (m, 1H), 3.36 (m, 1H),
3.28 (dd, J ¼ 14, 4 Hz, 1H), 3.04 (dd, J ¼ 14, 9.5 Hz, 1H),
1.80 (m, 1H), 1.62 (m, 1H), 1.42–1.51 (m, 1H), 1.29 (m,
1H); MS (APCIþ): m=z 321.94 (M+H).
References and notes
1. Berridge, M. J.; Lipp, P.; Bootman, M. D. Nat. Rev. 2000,
1, 11.
2. Koch, G. L. E. BioEssays 1990, 12, 527.
3. Li, J. Frontiers of Biotechnology and Pharmaceuticals 2002,
3, 110.
4. Soboloff, J.; Berger, S. A. J. Biol. Chem. 2002, 277, 13812.
5. Brostrom, M. A.; Brostrom, C. O. Cell Calcium 2003, 34, 345.
6. Harper, J. L.; Daly, J. W. Drug Dev. Res. 1999, 4, 107.
7. Kimball, S. R.; Clemens, M. J.; Tileray, V. J.; Wek, R. C.;
Horetsky, R. L.; Jefferson, L. S. Biochem. Biophys. Res.
Commun. 2001, 280, 293–300.
8. Alcazar, A.; del la Vega, C. M.; Bazan, E.; Fando, J. L.;
Salinas, M. J. Neurochem. 1997, 69, 1703.
9. Prostko, C. R.; Dholakia, J. N.; Brostrom, M. A.;
Brostrom, C. O. J. Biol. Chem. 1995, 270, 6211.
10. Alcazar, A.; Bazan, E.; Ribera, J.; Salinas, M. Neurosci.
Lett. 1995, 201, 215.
11. Prostko, C. R.; Brostrom, M. A.; Brostrom, C. O. Mol.
Cell. Biochem. 1993, 127/128, 255.
12. Palakurthi, S. S.; Fluckiger, R.; Aktas, H.; Changolkar,
A. K.; Shahsafaei, Al.; Harneit, S.; Kilic, E.; Halperin,
J. A. Cancer Res. 2000, 60, 2919.
13. Palakurthi, S. S.; Aktas, H.; Grubissich, L. M.; Morten-
sen, R. M.; Halperin, J. A. Cancer Res. 2001, 61, 6213.
14. Aktas, H.; Fluckiger, R.; Acosta, J. A.; Savage, J. M.;
Palakurthi, S. S.; Halperin, J. A. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 8280.
15. Demetri, G. D.; Fletcher, C. D. M.; Mueller, E.; Sarraf,
P.; Naujoks, R.; Campbell, N.; Spiegelman, B. M.; Singer,
S. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 3951.
16. Sugimura, A.; Kiriyama, Y.; Nochi, H.; Tsuchiya, H.;
Tamoto, K.; Sakurada, Y.; Ui, M.; Tokumitsu, Y.
Biochem. Biophys. Res. Commun. 1999, 261, 833.
19. Egusa, K.; Fukase, K.; Kusumoto, S. Synlett 1997, 675.
20. Carpenter, D. E.; Imbordino, R. J.; Maloney, M. T.;
Moeslein, J. A.; Reder, M. R.; Scott, A. Org. Process Res.
Dev. 2002, 6, 721.
21. Clementi, E.; Scheer, H.; Zacchetti, D.; Fasolato, C.;
Pozzan, T.; Meldolesi, J. J. Biol. Chem. 1992, 267, 2164.
22. Monks, A.; Scudiero, D.; Skehan, P.; Shoemaker, R.;
Paull, K.; Vistica, D.; Hose, C.; Langley, J.; Cronise, P.;
Vaigro-Wolff, A.; Gray-Goodrich, M.; Campbell, H.;
Mayo, J.; Boyd, M. J. Natl. Cancer Inst. 1991, 83, 757.
23. Reddy, K. A.; Lohray, B. B.; Bhushan, V.; Reddy, A. S.;
Rao Mamidi, N. V. S.; Reddy, P. P.; Saibaba, V.; Reddy,
N. J.; Suryaprakash, A.; Misra, P.; Vikramadithyan, R.
K.; Rajagopalan, R. J. Med. Chem. 1999, 42, 3265.
24. Fischer, M. J. E.; Paulussen, J. J. C.; Tollenaere, J. P.; De
Mol, N. J.; Janssen, L. H. M. Eur. J. Pharmacol. 1998,
350, 353.