The complex role of the triphenylmethyl motif in anticancer compounds

Article abstract of DOI:10.1021/ja8020999

Compounds incorporating the triphenylmethyl motif constitute an emerging family of potent anticancer agents. Although several small molecules containing this pharmacophore have now been identified, the mechanism of cell death induction for some of these c

Full text of DOI:10.1021/ja8020999

Published on Web 07/09/2008
The Complex Role of the Triphenylmethyl Motif in Anticancer
Compounds
Rahul Palchaudhuri, Vitaliy Nesterenko, and Paul J. Hergenrother*
Department of Chemistry, Roger Adams Laboratory, UniVersity of Illinois at
Urbana-Champaign, Urbana, Illinois 61801
Received March 20, 2008; E-mail: hergenro@uiuc.edu
Abstract: Compounds incorporating the triphenylmethyl motif constitute an emerging family of potent
anticancer agents. Although several small molecules containing this pharmacophore have now been
identified, the mechanism of cell death induction for some of these compounds is unknown. In an effort to
define their mechanism of action, and to distinguish subtypes within the group of compounds containing
the triphenylmethyl moiety, we have created novel triphenylmethyl-containing small molecules and have
evaluated them in a battery of biological assays. Here we show that several phosphonate and
phosphonochloridates possessing the triphenylmethyl motif potently induce death of multiple cancer cell
lines in culture. Further assays evaluating the ability to cause cell cycle arrest, inhibit tubulin polymerization,
dissociate mitochondrial-bound hexokinase in cancer cells, and inhibit calcium-dependent potassium ion
channels indicate that triphenylmethyl-containing compounds can be placed into at least four distinct
categories, each with a different mechanism of action.
zole have been linked to its ability to affect intracellular Ca2+
levels thereby inhibiting translation, in addition to inhibition of
Introduction
Approximately one in four deaths in the U.S. are caused by
cancer, and in 2005 cancer overtook heart disease as the top
killer of Americans under age 85.1 Despite dramatic advances
in our understanding of the underlying biology of this disease,
cancer death rates have remained constant (at roughly 200 deaths
for every 100000 people) over the last 30 years,1 highlighting
the need for new therapeutic agents. We recently identified a
family of compounds, the triphenylmethylamides (TPMAs),
which induce apoptotic death in multiple melanoma cell lines.2
Two representative TPMAs, 4BI and 4A, are shown in Figure
1. The TPMAs arrest growth of melanoma cells in the G1-
phase of the cell cycle, substantially reduce cellular levels of
active NFκB, and induce apoptosis. In addition, the TPMAs
have reduced toxicity to normal cells derived from the bone
marrow of healthy human donors, and high doses of these
compounds are well tolerated by mice.2
glycolysis by inducing the detachment of mitochondrial-bound
hexokinase.3,6 Other studies have shown that 3,3-diaryl-1,3-
dihyroindoles also have antiproliferative activity against cancer
cell lines in cell culture,7–9 and this effect may also be a result
of the inhibition of translation mediated by depletion of
intracellular Ca2+ stores.8,9 A clotrimazole derivative lacking
imidazole functionality has also been shown to induce G1-phase
cell cycle arrest and inhibit tumor growth in mouse xenograft
models,10 and even simple triphenylmethane derivatives potently
induce death of cultured cancer cells.5 Finally, a recent
publication highlighted the activity of S-trityl-L-cysteine (STLC,
Figure 1), a compound that showed potent antitumor activity
in a screen against the NCI-60 cell line panel.11 S-trityl-L-
cysteine induces M-phase arrest by targeting the Eg5 kinesin
spindle motor protein involved in bipolar spindle formation and
maintenance.11 As evaluated by the NCI, STLC has shown
anticancer activity in over 20 mouse xenograft models, and is
Other compounds containing the triphenylmethyl motif also
possess anticancer properties (Figure 1). Notably, the antifungal
agent clotrimazole inhibits the growth of cancer cells in culture,3
and has shown efficacy in various mouse models of cancer.3,4
Further studies indicate that clotrimazole arrests cells in the G1-
phase of the cell cycle.5 The anticancer properties of clotrima-
(6) Snajdrova, L.; Xu, A.; Narayanan, N. J. Biol. Chem. 1998, 273, 28032–
9.
(7) Uddin, M. K.; Reignier, S. G.; Coulter, T.; Montalbetti, C.; Granas,
C.; Butcher, S.; Krog-Jensen, C.; Felding, J. Bioorg. Med. Chem. Lett.
2007, 17, 2854–7.
(1) American Cancer Society. In American Cancer Society Statistics
for 2007, see: http://www.cancer.org/docroot/PRO/content/PRO_1_
1_Cancer_Statistics_2007_Presentation.asp.
(8) Natarajan, A.; Fan, Y. H.; Chen, H.; Guo, Y.; Iyasere, J.; Harbinski,
F.; Christ, W. J.; Aktas, H.; Halperin, J. A. J. Med. Chem. 2004, 47,
1882–5.
(2) Dothager, R. S.; Putt, K. S.; Allen, B. J.; Leslie, B. J.; Nesterenko,
V.; Hergenrother, P. J. J. Am. Chem. Soc. 2005, 127, 8686–96.
(3) Benzaquen, L. R.; Brugnara, C.; Byers, H. R.; Gatton-Celli, S.;
Halperin, J. A. Nat. Med. 1995, 1, 534–40.
(9) Natarajan, A.; Guo, Y.; Harbinski, F.; Fan, Y. H.; Chen, H.; Luus,
L.; Diercks, J.; Aktas, H.; Chorev, M.; Halperin, J. A. J. Med. Chem.
2004, 47, 4979–82.
(10) Cao, M. Y.; Lee, Y.; Feng, N. P.; Al-Qawasmeh, R. A.; Viau, S.; Gu,
X. P.; Lau, L.; Jin, H.; Wang, M.; Vassilakos, A.; Wright, J. A.; Young,
A. H. J. Pharmacol. Exp. Ther. 2004, 308, 538–46.
(11) Skoufias, D. A.; DeBonis, S.; Saoudi, Y.; Lebeau, L.; Crevel, I.; Cross,
R.; Wade, R. H.; Hackney, D.; Kozielski, F. J. Biol. Chem. 2006,
281, 17559–69.
(4) Khalid, M. H.; Tokunaga, Y.; Caputy, A. J.; Walters, E. J. Neurosurg.
2005, 103, 79–86.
(5) Al-Qawasmeh, R. A.; Lee, Y.; Cao, M. Y.; Gu, X.; Vassilakos, A.;
Wright, J. A.; Young, A. Bioorg. Med. Chem. Lett. 2004, 14, 347–
50.
9
10274 J. AM. CHEM. SOC. 2008, 130, 10274–10281
10.1021/ja8020999 CCC: $40.75
2008 American Chemical Society

Triphenylmethyl Motif in Anticancer Compounds
A R T I C L E S
Scheme 1. Synthesis of Triphenylmethyl Phosphonates of Class I
and III by the Arbuzov Reaction
Figure 1. Triphenylmethyl-containing compounds with anticancer
properties.
Figure 2. The three new classes of triphenylmethyl derivatives, called
TPMPs, that are described in this manuscript: Class I compounds are
phosphonates without ring substitution on the triphenylmethyl moiety. Class
II compounds are phosphonochloridates without ring substitutions on the
triphenylmethyl moiety. Class III compounds are phosphonates or phospho-
nochloridates with substituents on the triphenylmethyl moiety.
Table 1. The Evaluation of Class I TPMPs against the Human
Melanoma Cell Lines SK-MEL-5 and UACC-62 in Cell Culturea
one of the 171 compounds in the NCI standard agent database
that are of “special interest”.12
The combined data indicates that triphenylmethyl-containing
compounds have considerable potential as anticancer agents.
However, the precise manner by which some compounds of
this class exert their anticancer effect remains unknown. The
goal of our current study was to build a unified picture of the
anticancer mode of action of compounds containing this
functional motif. Interestingly, although the triphenylmethyl
moiety dominates the steric and conformational properties of
the compounds described herein, our work shows that com-
pounds containing this functional group can be placed into at
least four distinct classes based on their mechanism of action
and intracellular targets.
Results
Two of the most potent triphenylmethyl-containing anticancer
compounds described in the literature are STLC and the TPMAs.
Interestingly, these compounds appear to have very different
modes of action. STLC induces M-phase cell cycle arrest,
through the inhibition of the kinesin Eg5.11 The TPMAs,
conversely, induce G1-phase cell cycle arrest;2 the biological
target of the TPMAs is unknown. For this study we have
designed several novel classes and subclasses of triphenylm-
ethyl-containing compounds and have categorized them accord-
ing to their cell cycle arresting properties. Compounds that
arrested cellular growth in a discrete phase of the cell cycle
and that were also potent death inducers were then evaluated
in further biological assays.
Design and Synthesis of Novel Triphenylmethyl-Contain-
ing Compounds. Previous work had shown that the triphenyl-
methyl functional group was critical for activity of the TPMAs,
as diphenylmethyl-containing derivatives are largely inactive
against cultured cancer cells.2 The diphenyl version of STLC
is also inactive.13 The substituent projecting off of the amide
a Biomass was assessed by the sulforhodamine B assay after a 72 h
incubation. Average IC50 values and standard deviations were
determined from at least three independent experiments.
nitrogen of the TPMAs, however, appears to have little influence
on the compound’s anticancer properties. For example, many
different amide derivatives, including amides of phenethylamine,
4-methyoxy-phenylamine, and 2-(3-methyoxy-phenyl)-ethy-
lamine are all potent inducers of apoptosis in cancer cells in
culture.2 With the goal of obtaining a diverse array of triph-
enylmethyl-containing compounds for mechanistic studies, a
series of derivatives were synthesized in which the amide in
the TPMAs was replaced by a phosphonate or phosphonochlo-
ridate. This replacement allows the polarity of the linkage to
the triphenylmethyl motif to be maintained, while permitting
the projection of one additional substituent. These compounds
(12) For NCI standard agent database, see: http://dtp.nci.nih.gov/docs/
cancer/searches/standard_agent_table.html.
(13) Ogo, N.; Oishi, S.; Matsuno, K.; Sawada, J.; Fujii, N.; Asai, A. Bioorg.
Med. Chem. Lett. 2007, 17, 3921–4.
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10275

A R T I C L E S
Palchaudhuri et al.
Scheme 2. Synthesis of p-Hydroxy-Substituted Triphenylmethyl
Phosphonate TPMP-III-8
Scheme 3. Synthesis of Phosphonates of Class I by Treatment of
the Phosphonyldichloride TPMP-II-1 with 2.2 equiv of Alkoxide
Table 2. The Evaluation of Class II TPMPs against the Human
Melanoma Cell Lines SK-MEL-5 and UACC-62 in Cell Culturea
alkoxide 1,2-ethanediol yielded the ring closed phosphonate
TPMP-I-12 (Scheme 5), and dihydroxylation of TPMP-I-10
afforded the phosphonate TPMP-I-11 bearing alkyl groups with
hydroxyl functionality (Scheme 6).
Determination of IC50 Values. All compounds were evaluated
for their ability to induce death in the UACC-62 and SK-MEL-5
cell lines, both human malignant melanoma cell lines. The
UACC-62 cell line has known mutations in the braf, pten, and
cdkn2a genes,15 and expresses Apaf-1 at intermediate levels.16
The SK-MEL-5 cell line is a human melanoma derived from
an auxiliary node of a Caucasian female.17 It has known
mutations in the braf and cdkn2a genes.18 Additionally, there
is little-to-no expression of Apaf-1 in this cell line, making it
resistant to apoptotic stimuli that require apoptosome forma-
tion.16
The 25 TPMPs synthesized and the four compounds from
Figure 1 were evaluated at a range of doses to obtain
dose-response curves from which IC50 values could be
calculated (see Supporting Information for dose response
curves). The IC50 value determinations for active compounds
(IC50 values < 100 µM) were made on at least three different
occasions, in triplicate each time, according to the following
protocol. UACC-62 or SK-MEL-5 cells (3000 cells/well in 96-
well plates) were incubated in the presence of compounds for
72 h at 37 °C. The biomass was quantified by staining with
sulforhodamine B; the results for each compound class are
displayed in Tables 1, 2, and 3. In general, select compounds
in each structural class had low micromolar potency against
these melanoma cell lines; the most potent compounds were
taken into the cell cycle evaluation, described below. The IC50
values for the compounds in Figure 1 were also determined
alongside the TPMPs, and the values are as follows (against
the UACC-62 and SK-MEL-5 cell lines, respectively): 4BI )
11.1 ( 1.1 and 13.1 ( 0.9 µM, 4A ) 13.4 ( 0.7 and 13.8 (
3.0 µM, clotrimazole ) 9.0 ( 1.2 and 11.2 ( 0.6 µM, STLC
) 0.8 ( 0.1 and 0.7 ( 0.3 µM.
a Biomass was assessed by the sulforhodamine B assay after a 72 h
incubation. Average IC50 values and standard deviations were
determined from at least three independent experiments.
can be placed into three structural classes: phosphonates (Figure
2, class I), phosphonochloridates (Figure 2, class II), and ring-
substituted derivatives of class I and II (Figure 2, class III). As
a group these compounds are called triphenylmethyl phosopho-
nates/triphenylmethyl phosphonochloridates (TPMPs, see Figure
2). The list of TPMPs synthesized and described herein is given
in Tables 1–3.
The synthetic routes to these compounds are shown in
Schemes 1-6. Phosphonates bearing dimethyl, diethyl, or dibutyl
groups were synthesized in moderate to good yields (25-98%,
average of 79%) via the Arbuzov reaction between the com-
mercially available trimethyl-, triethyl-, or tributylphosphite and
the triphenylmethyl chlorides (compounds 1-9, Scheme 1).
Reactions were complete after heating to reflux in benzene for
2 hours. Compound TPMP-III-7 was synthesized as previously
described,14 and TPMP-III-8 was created through the dem-
ethylation of the para-methoxy triphenylmethyl phosphonate
TPMP-III-4 (Scheme 2).
Evaluation of Cell Cycle Arrest. As mentioned, TPMAs and
clotrimazole are reported to arrest the growth of cancer cells in
the G1-phase of the cell cycle, while STLC induces arrest in
the M-phase. To determine the cell cycle arresting properties
of the novel phosphorus-containing analogues of the TPMAs,
HL-60 cells (human leukemia cell line) were treated with the
most potent compounds at 20 µM for 12 h. HL-60 cells grow
rapidly and in suspension (not adhered to the culture flask) in
culture, and thus readily lend themselves to cell cycle analysis
experiments. The population of cells in the different phases of
the cell cycle was measured by propidium iodide staining and
flow cytometry. Interestingly, the TPMPs displayed a range of
cell cycle arresting properties. The data is displayed for the
previously known compounds (Figure 3A), the new compounds
Owing to the limited availability of trialkylphosphites,
displacement of the chlorines of triphenylmethyl phosphonyl
dichloride TPMP-II-1 with alkoxides was employed to gain
access to a wider variety of phosphonates (Scheme 3). Treatment
of dichloride TPMP-II-1 with 2.2 equiv of alkoxide yielded
primarily phosphonates of Class I. As shown in Scheme 3, six
Class I TPMPs were synthesized through this route, with yields
ranging from 14-87% (av 59%). The use of 1.3 equiv of
alkoxide allowed access to phosphonochloridates of Class II
and III (Scheme 4, av 51%). Two additional reactions were used
to produce two other TPMPs. Reaction of TPMP-II-1 with the
(14) Iliopulos, M. I.; Wieder, H. Angew. Chem. 1965, 77, 618–19.
9
10276 J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008

Triphenylmethyl Motif in Anticancer Compounds
A R T I C L E S
Table 3. The Evaluation of Class III TPMPs against the Human Melanoma Cell Lines SK-MEL-5 and UACC-62 in Cell Culturea
a Biomass was assessed by the sulforhodamine B assay after a 72 h incubation. Average IC50 values and standard deviations were determined from at
least three independent experiments.
that induce G1-phase cell cycle arrest (Figure 3B), and the new
compounds that induce G2/M-phase cell cycle arrest (Figure
3C). With this identification of several novel TPMPs that
potently kill cancer cells (namely TPMP-II-2, TPMP-III-2,
TPMP-III-4) and induce arrest in either the G1- or the G2/M-
phase of the cell cycle, we set out to compare these novel
compounds with the known compounds (from Figure 1) in
assays examining the effect of compounds on mitosis, mito-
chondrial-bound hexokinase dissociation, and the effect of
compounds on calcium-dependent potassium channels.
Effects on Mitosis and Tubulin Polymerization. STLC has
previously been shown to induce the arrest of cellular growth
in the M-phase of the cell cycle, and cells treated with STLC
display a “monoaster” phenotype when examined by fluores-
cence microscopy staining for tubulin and DNA.11 This
monoaster phenotype is indicative of compounds that target the
kinesin Eg5, and indeed STLC has been shown to be a potent
and reversible Eg5 inhibitor in vitro (IC50 ) 500 nM).11 Figure
4 shows the effect of both STLC and TPMP-III-2 on micro-
tubules as examined by confocal microscopy. Consistent with
data in the literature, cells treated with STLC display the
“monoaster” phenotype. Interestingly, while TPMP-III-2 also
induces arrest in the M-phase of the cell cycle, cells treated
with this compound do not show the monoaster phenotype but
rather possess fragmented mitotic spindles characteristic of cells
treated with antimicrotubule compounds (Figure 4).
in vitro tubulin polymerization. Tubulin polymerization can be
monitored spectrophotometrically by the increase in optical
density at 340 nm. Compounds such as Taxol are known to
stabilize microtublules (and give an increased rate of polym-
erization in this assay), whereas compounds such as nocodazole
prevent tubulin polymerization. As shown by the data in Figure
5, TPMP-III-2 is clearly a compound that prevents the
polymerization of tubulin. The G1-phase arrestor TPMP-I-2
had no effect in this tubulin polymerization assay (Figure 5).
Interestingly, 4,4′-dimethoxytriphenylmethanol (structurally simi-
lar to TPMP-III-2) did not affect tubulin polymerization,
demonstrating the necessity of both the phosphonate moiety and
the para-methoxy functionality (Figure 5). STLC and clotri-
mazole also did not affect tubulin polymerization, while TPMP-
III-1 and TPMP-III-4 inhibited tubulin polymerization to a
modest degree at higher compound concentrations, consistent
with their G2/M-phase arresting properties (See Supporting
Information Figures S1 and S2).
Effect of Compounds on the Detachment of Mitochondrial-
Bound Hexokinase. Clotrimazole and other antifungal azole-
derivatives such as bifonazole, econazole, miconazole, and
ketoconazole have been identified as calmodulin antagonsists
because of their ability to inhibit calmodulin-dependent phos-
phodiesterase activity in vitro.19 Clotrimazole exhibits a greater
than 54-fold selective inhibition of calmodulin-dependent versus
calmodulin-independent phosphodiesterase activity in vitro with
an IC50 value of 18 µM.19 As calmodulin activity plays a central
role in the maintenance of cellular metabolism, calmodulin
antagonists may exert an antiproliferative effect by disrupting
metabolic events such as glycolysis.20 In fact, treatment of
melanoma B16-F10 cells with clotrimazole and bifonazole
causes detachment of the key glycolytic enzymes phosphofruc-
Given this data, we assessed the effect of TPMP-III-2 and
several of the other triphenylmethyl-containing compounds on
(15) The Cancer Genome Project. http://www.sanger.ac.uk/perl/genetics/
CGP/cosmic?action)sample&id)905976.
(16) Soengas, M. S.; Capodieci, P.; Polsky, D.; Mora, J.; Esteller, M.; Opitz-
Araya, X.; McCombie, R.; Herman, J. G.; Gerald, W. L.; Lazebnik,
Y. A.; Cordon-Cardo, C.; Lowe, S. W. Nature 2001, 409, 207–211.
(17) HyperCLDB. http://www.biotech.ist.unige.it/cldb/cl7123.html.
(18) The Cancer Genome Project. http://www.sanger.ac.uk/perl/genetics/
CGP/cosmic?action)sample&id)905956.
(19) Hegemann, L.; Toso, S. M.; Lahijani, K. L.; Webster, G. F.; Uitto, J.
J. InVestig. Dermatol. 1993, 100, 343–346.
(20) Beitner, R. Mol. Genet. Metab. 1998, 64, 161–168.
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10277

A R T I C L E S
Palchaudhuri et al.
Figure 4. The effect of STLC and TPMP-III-2 on microtubule assembly
in HeLa cells after an 18 h incubation in the presence of 20 µM compound.
Cells were fixed and the nuclei and microtubules were stained with
propidium iodide (PI) and a FITC-conjugated antitubulin antibody, respec-
tively. The “monoaster” phenotype is visible for STLC, whereas TPMP-
III-2 induces fragmented mitotic spindles.
detachment is the result of an allosteric interaction rather than
a result of calmodulin antagonistic acitivity.26 This is supported
by the inability of other calmodulin antagonists to induce the
dissociation of mitochondrial-bound hexokinase.22
Given the structural similarity of clotrimazole and the other
triphenylmethyl-containing compounds, it was of interest to
determine whether these compounds could induce detachment
of hexokinase from the mitochondrial outer membrane and
thereby exert their antiproliferative effect by inhibiting glycolysis
and reducing cellular ATP levels. Melanoma B16-F10 cells
were incubated in the presence of modest concentrations of
compound (15 µM) for a short period of time (2 h) after which
the cells were lysed and hexokinase activity of the mitochon-
drial-rich fraction was assessed and compared to vehicle-treated
cells. As shown in Figure 6, treatment of cells with clotrimazole
induced a 71% reduction in mitochondrial-bound hexokinase.
This result is consistent with previous observations.22 Other
triphenylmethyl compounds which exhibit equal or significantly
greater antiproliferative activity than clotrimazole in cell culture
were not able to detach mitochondrial-bound hexokinase as
effectively as clotrimazole under the same conditions (Figure
6), suggesting that these compounds induce death through an
alternate mechanism.
Effect of Compounds on Gardos Channel Inhibition. In ad-
dition to its calmodulin antagonist activity and ability to cause
the detachment of hexokinase from mitochondria, clotrimazole
is also a potent inhibitor of the calcium-dependent potassium
ion channel, also known as the Gardos channel.27 A binding
site model has been proposed for the inhibition of the Gardos
channel by clotrimazole and another triphenylmethane derivative
based on site directed mutagenesis.28 Inhibitors of the Gardos
channel show promise for the treatment of sickle cell disease
Figure 3. Cell cycle arresting properties of compounds containing the
triphenylmethyl group. HL-60 cells (human leukemia) were treated with
the indicated compound at 20 µM for 12 h, and the cell cycle distribution
was examined by propidium iodide staining and flow cytometry. (A) The
cell cycle arresting properties of known triphenylmethyl compounds that
possess anticancer activity. Compound 4BI and clotrimazole induce cell
cycle arrest in the G1-phase whereas STLC induces M arrest. (B) Certain
TPMPs induce G1-phase cell cycle arrest, (C) while others induce arrest in
the G2/M-phase of the cell cycle.
tokinase and aldolase from the cytoskeleton and hexokinase from
the outer mitochondrial membrane prior to affecting cell
viability.21–23 As elevated glycolytic activity is a major hallmark
of cancer, compounds that inhibit glycolysis may prove to be
effective anticancer agents.24,25 It has been proposed that
clotrimazole’s ability to cause mitochondrial-bound hexokinase
(24) Pelicano, H.; Martin, D. S.; Xu, R. H.; Huang, P. Oncogene 2006,
25, 4633–4646.
(25) Xu, R.-h.; Pelicano, H.; Zhou, Y.; Carew, J. S.; Feng, L.; Bhalla, K. N.;
Keating, M. J.; Huang, P. Cancer Res. 2005, 65, 613–621.
(26) Pastorino, J. G.; Shulga, N.; Hoek, J. B. J. Biol. Chem. 2002, 277,
7610–7618.
(21) Glass-Marmor, L.; Beitner, R. Eur. J. Pharmacol. 1997, 328, 241–
248.
(22) Penso, J.; Beitner, R. Eur. J. Pharmacol. 1998, 342, 113–117.
(23) Glass-Marmor, L.; Morgenstern, H.; Beitner, R. Eur. J. Pharmacol.
1996, 313, 265–271.
(27) Brugnara, C.; Armsby, C. C.; Sakamoto, M.; Rifai, N.; Alper, S. L.;
Platt, O. J. Pharm. Exper. Ther. 1995, 273, 266–272.
9
10278 J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008

Triphenylmethyl Motif in Anticancer Compounds
A R T I C L E S
Figure 6. The assessment of mitochondrial-hexokinase activity in murine
melanoma B16-F10 cells after 2 h treatment with compound (15 µM), as
compared to vehicle-treated control cells. Clotrimazole causes a 71%
reduction in mitochondrial-hexokinase activity whereas other triphenylm-
ethyl containing compounds show limited effect (/ ) p values <0.012).
to-modest inhibition with average IC50 values of 1.7 and 13.8
µM, respectively. The microtubule-inhibiting compound TPMP-
III-2, the TPMAs, and STLC were poor inhibitors of the Gardos
channel with IC50 values ranging from 35 to >100 µM. Given
the potent death inducing abilities of these triphenylmethyl
compounds relative to clotrimazole, their generally modest
inhibition of the Gardos channel would suggest that the
inhibition of calcium-dependent potassium channels is not the
primary mechanism by which these compounds exert their
anticancer activity.
Effect of Compounds on Multiple Cell Lines. To investigate
the general anticancer properties of the more potent TPMPs,
compounds TPMP-II-2, TPMP-III-2, and TPMP-III-4 were
tested for their ability to induce death in a variety of cancer
cell lines. As shown in Table 5, these compounds potently
induced death in most of the cancer cell lines tested, with only
the PC-12 rat adrenal cancer cell line being resistant. The
microtubule-inhibiting compound TPMP-III-2 was, in general,
slightly more potent that the other two compounds in the
majority of the cell lines.
Figure 5. The effect of TPMPs (10 µM) on in vitro tubulin polymerization.
As monitored by the optical density at 340 nm, TPMP-III-2 inhibits the
polymerization of tubulin, whereas TPMP-I-2 and 4,4′-dimethoxytriph-
enylmethanol have no effect.
as they prevent red blood cell dehydration that may occur during
oxygen-deoxygenation cycling.29 In fact, several triphenylm-
ethyl-containing compounds with potent Gardos channel inhibi-
tion properties have been previously described, and a triphe-
nylmethylamide is currently in phase III clinical trials for sickle
cell disease.29 Interestingly, halide substitution on the triph-
enylmethyl group appears to be required for potent Gardos
channel inhibition.30
Apart from the relevance of Gardos channel inhibition to
sickle cell disease, the Gardos channel is also implicated in the
proliferation of melanoma.31,32 Blockers of the channels induce
cell cycle arrest by causing membrane depolarization, which
reduces Ca2+ influx that would otherwise promote growth and
proliferation.31,32 It was therefore of interest to assess the
inhibition properties of triphenylmethyl-containing compounds
on Gardos channel activity. As rubidium efflux is mediated by
the Gardos channel,33 human red blood cells were loaded with
86Rb and incubated in the presence of various concentrations
of compound or vehicle. Efflux was initiated by the addition of
calcium and the calcium ionophore (A23187), and dose response
curves were constructed upon analysis of rubidium content in
the supernatant. As shown in Table 4, clotrimazole is a potent
inhibitor of the Gardos channel with an average IC50 value of
0.92 µM, consistent with previously reported results.34 The novel
G1-phase arrestors TPMP-I-2, and TPMP-II-2 exhibited good-
Discussion
We recently reported a class of small molecules, the TPMAs,
which induce arrest of cancer cell lines in culture in the G1-
phase of the cell cycle.2 From our initial analysis, the triph-
enylmethyl functional group was the critical pharmacophore,
as diphenylmethyl derivatives had greatly reduced death-
inducing properties. It was interesting to note other compounds
in the literature that possess a triphenylmethyl motif, most
notably STLC and clotrimazole also appear to be potent
anticancer agents. We thus created a series of novel compounds
containing this functional group, the TPMPs, and found that
several of them induce death in cancer cell lines in culture. These
compounds were then tested in a battery of biological assays,
and the results were compared to the TPMAs, clotrimazole, and
STLC. Interestingly, we have found that these structurally related
compounds have very different mechanisms of action.
(28) Wulff, H.; Gutman, G. A.; Cahalan, M. D.; Chandy, K. G. J. Biol.
Chem. 2001, 276, 32040–32045.
(29) Wulff, H.; Kolski-Andreaco, A.; Sankaranarayanan, A.; Sabatier, J.-
M.; Shakkottai, V. Curr. Med. Chem. 2007, 14, 1437–1457.
(30) Wulff, H.; Miller, M. J.; Hansel, W.; Grissmer, S.; Cahalan, M. D.;
Chandy, K. G. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 8151–8156.
(31) Nilius, B.; Wohlrab, W. J. Physiol. 1992, 445, 537–548.
(32) Lepple-Wienhues, S.; Berweck, M.; Bo¨hmig, C. P.; Leo, B.; Meyling;
Garbe, C.; Wiederholt, M. J. Membr. Biol. 1996, 151, 149–157.
(33) Stocker, J. W.; De Franceschi, L.; McNaughton-Smith, G. A.;
Corrocher, R.; Beuzard, Y.; Brugnara, C. Blood 2003, 101, 2412–
2418.
From analysis of the data presented herein, the replacement
of the TPMA amide with the phosphonate or phosphonochlo-
ridate motif provides many compounds possessing enhanced
anticancer activity. Compounds TPMP-I-2 through -8, TPMP-
I-10, TPMP-I-12, TPMP-II-2, TPMP-II-3, and TPMP-III-1
(34) Gemma, S.; et al. J. Med. Chem. 2008, 51, 1278–94.
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10279

A R T I C L E S
Palchaudhuri et al.
Scheme 4. Synthesis of Phosphonochloridates of Class II and III by Treatment of Phosphonyldichlorides with 1.3 equiv of Alkoxide
Table 4. Comparison of Various Triphenylmethyl-Containing
Compounds and Their IC50 Values for 86Rb Efflux Inhibition in
Human Red Blood Cells
Scheme 5. Synthesis of the Ring-Closed Phosphonate TPMP-I-12
compound
av IC50 (µM)
clotrimazole
TPMP-I-2
TPMP-II-2
TPMP-III-2
4BI
0.92
1.7
13.8
56
35
>100
>100
4A
STLC
functionalities that may irreversibly inhibit the serine protease
(see Supporting Information). Clotrimazole is a potent chymot-
rypsin inhibitor, and it is in fact known to inhibit multiple
enzymes nonspecifically.38 Conversely, several of the TPMPs
(such as TPMP-I-2 and TPMP-III-2) do not show this
promiscuous inhibition (see Supporting Information).
through -6, TPMP-III-9 and TPMP-III-10 all were reasonably
potent versus both melanoma cell lines, with IC50 values ranging
from 2.2 to 26.9 µM. An increase in activity was seen with
compounds that had methyl ether substitutions on the aromatic
rings; for example, TPMP-III-2 and TPMP-III-4 were quite
potent against the melanoma cell lines, with IC50 values ranging
from 2.2 to 5 µM. Compounds containing functional groups
that increased aqueous solubility were not very potent, such as
TPMP-I-11, TPMP-III-7, and TPMP-III-8.
The two phosphonochloridate derivatives created and evalu-
ated were quite effective in killing melanoma cell lines (Table
2). One of these compounds in particular, TPMP-II-2, was quite
potent in the induction of cell growth inhibition, with IC50 values
of 3.3-4.3 µM. To our knowledge this is the first demonstration
of biological activity of triphenylmethyl phosphonochloridates.
Although certain compounds containing the phosphonochlori-
date moiety have been previously shown to rapidly hydrolyze
in water,35,36 our investigations show that compound TPMP-
II-2 is fully stable in aqueous solutions for multiple months
(see Supporting Information Figure S3). This compound is also
stable in both acidic and basic aqueous solutions; no evidence
of degradation is observed when placed in aqueous solutions
of pH 1.5 through pH 11.8 for 20 h (see Supporting Information
Figure S3). Given their potencies, further study of this interesting
class of compounds is warranted.
Because of the size and hydrophobic nature of the triphenyl-
methyl pharmacophore, it might be suspected that anticancer
compounds possessing this functional group share a common
mode of cell death induction. However, we were able to clearly
separate the triphenylmethyl-containing compounds based on
their cell cycle arresting properties. Further assays related to
mitosis, mitochondrial-bound hexokinase detachment, and Gar-
dos channel inhibition, were conducted on the most potent
compounds. From the combined data a unified picture has
emerged, revealing at least four distinct mechanisms of cell
death induced by compounds containing the triphenylmethyl
pharmacophore (Table 6): (1) STLC induces cell cycle arrest
in the M-phase, consistent with its known inhibition of the
kinesin Eg5. STLC has no effect on tubulin polymerization,
nor does it inhibit calcium-dependent potassium ion channels.
(2) Clotrimazole induces cell death through a much different
process. It arrests cells in the G1-phase of the cell cycle, and it
has been shown to cause the translocation of hexokinase from
the mitochondrial outer membrane thus inhibiting glycolysis
resulting in the depletion of ATP. Clotrimazole is also a potent
inhibitor of calcium-dependent potassium ion channels, which
can affect the transition from G1- to S-phase by modulating
intracellular calcium levels.31
Additionally, while certain compounds containing the phospho-
nochloridate moiety have been demonstrated to irreversibly
inhibit serine proteases such as trypsin, chymotrypsin, and
acetylcholine esterase,35–37 TPMP-II-2 is only a modest chy-
motrypsin inhibitor in vitro with activities comparable to
triphenylmethyl-containing compounds that do not possess
On the basis of the data reported herein neither the TPMAs
nor the class I, II, or III TPMPs behave in a manner similar to
STLC or clotrimazole, and we believe these compounds can
be placed into two further mechanistic classes. (3) The third
class of compounds are triphenylmethyl-containing derivatives
that inhibit tubulin polymerization. TPMP-III-2 and TPMP-
(35) Wins, P.; Wilson, I. B. Biochim. Biophys. Acta 1974, 334, 137–145.
(36) Hovanec, J. W.; Lieske, C. N. Biochemistry 1972, 11, 1051–1056.
(37) Bjorkling, F.; Dahl, A.; Patkar, S.; Zundel, M. Bioorg. Med. Chem.
1994, 2, 697–705.
(38) Feng, B. Y.; Simeonov, A.; Jadhav, A.; Babaoglu, K.; Inglese, J.;
Shoichet, B. K.; Austin, C. P. J. Med. Chem. 2007, 50, 2385–2390.
9
10280 J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008

Triphenylmethyl Motif in Anticancer Compounds
A R T I C L E S
Table 5. The Effect of TPMP-II-2, TPMP-III-2, and TPMP-III-4 on Multiple Cancer Cell Lines in Cell Culturea
cell line
cancer type
TPMP-II-2 (IC50 µM)
TPMP-III-2 (IC50 µM)
TPMP-III-4 (IC50 µM)
SK-MEL-5
UACC-62
PC-12
HL-60
U-937
IGROV-1
MCF-7
SK-N-SH
melanoma (human)
melanoma (human)
adrenal (rat)
leukemia (human)
lymphoma (human)
ovarian (human)
breast (human)
3.3 ( 1.6
4.3 ( 1.3
77 ( 25
2.3 ( 0.2
4.5 ( 0.2
1.2 ( 0.1
9.2 ( 0.8
4.0 ( 2.6
2.2 ( 0.6
2.5 ( 0.8
10.6 ( 0.6
1.4 ( 0.4
1.7 ( 0.7
2.9 ( 1.3
5.1 ( 0.2
4.2 ( 0.4
3.9 ( 0.8
5.0 ( 1.0
23 ( 6
2.2 ( 0.3
2.3 ( 0.1
4.3 ( 3.1
6.8 ( 1.9
7.1 ( 4.4
neuroblastoma (human)
a Cells were treated with the compounds for 72 h and death was quantified via the methods described in the Materials and Methods section. Average
IC50 values and standard deviations were determined from at least three independent experiments.
Scheme 6. Synthesis of Triphenylmethyl Phosphonate TPMP-I-11
via the Dihydroxylation of TPMP-I-10
of the triphenylmethyl motif with either cysteine (STLC),
imidazole (chotrimazole), or amides, phosphonates, phospho-
nochloridates (TPMAs/TPMPs). Mechanistically distinct classes
of compounds further arise when substitutions are made on the
aryl rings of TPMAs and TPMPs. Second, a major drawback
of the antitubulin drugs (such as Taxol and the Vinca alkaloids)
is the inability to penetrate the blood-brain-barrier (BBB) due
to active efflux by P-glycoproteins, thus rendering these
compounds less useful for CNS tumors.39 In fact, few com-
pounds that have tubulin as their biological target penetrate the
BBB.39 The lipophilic nature of the triphenylmethyl functional
group and the known BBB penetrance of clotrimazole40 suggest
that derivatives based on TPMP-III-2 may provide the rare
antimitotic that penetrates the BBB. Third, triphenylmethyl
phosphonates and triphenylmethyl phosphonochloridates are
novel and potent anticancer compounds, and the G1-phase
arresting subclass of these agents appears to operate by a unique
mode of action, distinct from clotrimazole. As shown in Table
5, several of these compounds are potent death inducers to a
variety of cancer cell lines in culture.
Table 6. Comparison of Various Anticancer
Triphenylmethyl-Containing Compounds Based on Their Activity in
a Variety of Assays
mitchondrial-
cell
cycle
hexokinase
dissociation dependent
arrest microtubules? ability
K+ channels
Ca2+
-
affects
mechanism
of action
compound
STLC
M
no
weak
no effect Eg5 kinesin
inhibitor
clotrimazole G1 no
strong
inhibitor intracellular
Ca2+ depletion,
translation
inhibition
In summary, we have discovered that phosphonate and
phosphonochloridate analogues of triphenylmethylamides induce
apoptosis in melanoma and other cancer cell lines and arrest
cellular growth in the G1- or M-phases of the cell cycle.
Preclinical studies with the most potent compounds and experi-
ments directed toward the identification of the precise biological
targets of the G1-phase arrestors are underway and will be
reported in due course.
TPMP-III-2 M
inhibitor
weak
weak
weak
weak
no effect antitubulin
no effect unknown
inhibitor unknown
inhibitor unknown
4BI/4A
G1 no
TPMP-I-2 G1 no
TPMP-II-2 G1 no
III-4 inhibit tubulin polymerization in vitro and induce M-phase
arrest of cells in culture. There does appear to be a relationship
between the antimitotic effect and the degree of methoxy
substitution in these compounds. These are the first examples
of triphenylmethyl-containing compounds that are able to inhibit
tubulin polymerization. (4) TPMAs and TPMPs that are G1-
phase arrestors (such as TPMP-I-2, TPMP-III-2) are in a final
category. In fact, TPMAs and TPMPs with no functionality on
the triphenylmethyl moiety all appear to induce cell death in a
similar fashion (causing G1-phase arrest), regardless of whether
they contain an amide, phosphonate, or phosphonochloridate
functional group. Although TPMP-I-2 and TPMP-III-2 and
TPMAs are modest inhibitors of calcium-dependent potassium
ion channels in comparison to clotrimazole, the potency of
Gardos channel inhibition does not correlate with the antipro-
liferative efficacy observed in cell culture. The precise biological
target of these compounds remains unknown at this point.
Several important pieces of information can be gleaned from
this data. First, assessment of these derivatives shows that subtle
changes in chemical composition can dramatically alter the
mechanism of action of a small molecule. In this case, one level
of mechanistic discrimination is made through the connection
Acknowledgment. We thank Professor Timothy Mitchison
(Harvard Medical School) for the gift of tubulin and Professor John
A. Katzenellenbogen and Dr. Kathryn E. Carlson for their guidance
regarding the Gardos channel inhibition assay. We also acknowl-
edge the assistance of the cell flow cytometry and mass spectrometry
facilities at the University of Illinois, Urbana-Champaign. We are
grateful to the American Cancer Society (Grant RSG-07-026-01)
for funding this work.
Supporting Information Available: Complete reference 34;
materials and methods; NMR spectral characterization; IC50
curves for cell culture experiments; Gardos channel inhibition
and chymotrypsin inhibition; supporting figures. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA8020999
(39) Loscher, W.; Potschka, H. Prog. Neurobiol. 2005, 76, 22–76.
(40) Khalid, M. H.; Tokunaga, Y.; Caputy, A. J.; Walters, E. J. Neurosurg.
2005, 103, 79–86.
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10281