A novel organobismuth compound, 1-[(2-di-p-tolylbismuthanophenyl)diazenyl] pyrrolidine, induces apoptosis in the human acute promyelocytic leukemia cell line NB4 via reactive oxygen species

Article abstract of DOI:10.1016/j.jinorgbio.2012.09.009

A novel organobismuth compound, 1-[(2-di-p-tolylbismuthanophenyl)diazenyl] pyrrolidine (4), which has 1-(phenyldiazenyl)pyrrolidine (1) substituent in a benzene ring of tri(p-tolyl)bismuthane (2), was synthesized and tested for biological activity toward

Full text of DOI:10.1016/j.jinorgbio.2012.09.009

Journal of Inorganic Biochemistry 117 (2012) 77–84
Contents lists available at SciVerse ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
A novel organobismuth compound, 1-[(2-di-p-tolylbismuthanophenyl)diazenyl]
pyrrolidine, induces apoptosis in the human acute promyelocytic leukemia cell
line NB4 via reactive oxygen species
Kengo Onishi a, Mizuho Douke a, Taisuke Nakamura a, Youta Ochiai a, Naoki Kakusawa b, Shuji Yasuike b,
,
⁎
Jyoji Kurita b, Chika Yamamoto b, Masatoshi Kawahata c, Kentaro Yamaguchi c, , Tatsuo Yagura
⁎
a,⁎⁎
a
Department of Bioscience, Faculty of Science and Technology, Kwansei Gakuin University, 2‐1, Gakuen, Sanda, 669‐1337, Japan
Faculty of Pharmaceutical Science, Hokuriku University, Kanagawa-machi, Kanazawa 920‐1181, Japan
Faculty of Pharmaceutical Sciences at Kagawa Campus, Tokushima Bunri University, Shido, Sanuki 769‐2193, Japan
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 May 2012
Received in revised form 4 September 2012
Accepted 4 September 2012
Available online 12 September 2012
A novel organobismuth compound, 1-[(2-di-p-tolylbismuthanophenyl)diazenyl]pyrrolidine (4), which has
1-(phenyldiazenyl)pyrrolidine (1) substituent in a benzene ring of tri(p-tolyl)bismuthane (2), was synthe-
sized and tested for biological activity toward human tumor cell lines. 4 had a potent anti-proliferative effect
on human cancer cell lines, although both 1 and 2 exhibited only weak activity. The sensitivity of leukemic
cell lines to 4 was relatively high; IC50 values for the human leukemia cell line NB4 and cervical cancer cell
line HeLa were 0.88 μM and 5.36 μM, respectively. Treatment of NB4 cells with 4 induced apoptosis, loss of
mitochondrial membrane potential (ΔΨmt) and the generation of cellular reactive oxygen species (ROS). 1
and 2 did not induce apoptosis and had only a marginal effect on ΔΨmt and the generation of ROS.
N-acetyl cysteine (NAC) reduced the generation of ROS and conferred protection against 4-induced apoptosis,
Keywords:
Organobismuth
Triazene
Human cancer cell
Apoptosis
ROS
indicating
a role for oxidative stress. 4 did not inhibit the polymerization of tubulin in vitro.
1-[2-(di-p-tolylstibanophenyl)diazenyl]pyrrolidine (3), which has the same chemical structure as 4 but
contains antimony in place of bismuth, did not show any cytotoxic activity. The results suggest that the
conjugated structure of the diazenylpyrrolidine moiety and bismuth center are key to the bioactivity of 4.
© 2012 Elsevier Inc. All rights reserved.
1. Introduction
Over the last several years, our laboratory has engaged in the study
of chemically synthesized bismuth-based organometallic compounds
Alkylating agents can be used in cancer chemotherapy though they
are carcinogenic. Triazenes possessing cytotoxic alkylating effects
have been known to show anticancer activity [1–8]. Clinical studies
show that triazenes exhibit anti-neoplastic activity in patients with
acute leukemia [9–11]. Through metabolic activation, dacarbazin
(DTIC), a triazene, generates a cytotoxic monomethyl triazene species
that alkylates DNA [3,4,12–14]. Temozolomide, a tetranitrogen
heterocycle that spontaneously decomposes into a monomethyl
metabolite without metabolic activation, showed significant clini-
cal activity against high-grade glioma [15]. On the other hand, a
series of synthesized chemical compounds with a triazene moiety
exhibits antimalarial activity [16]. These studies suggest the triazene
moiety to be crucial to interaction with many cellular targets.
with anticancer activity and has shown that their antiproliferative ac-
tivity varies dependent on chemical structure [17,18]. These studies
should lead to the development of new organobismuth compounds
with stronger and more selective biological activity for future applica-
tions in medicine.
In the present study, as a part of a search for new organometalloid
compounds with excellent anticancer activities, we synthesized a
triarylbismuth(III) compound with a diazenyl pyrrolidine (a triazene
structure) substitute at the ortho position, based on tri(p-tolyl)
bismuthane which showed only low biological activity [see review,
19], and explored its cytotoxic activity against human acute
promyelocytic leukemia (APL) cell line, NB4. The purpose was to deter-
mine whether the substitute could enhance the biological activity or
provide insight into how organobismuth compounds exert their
antiproliferative activity toward cancer cells. We found that the cytotox-
ic activity of tri(p-tolyl)bismuthane 2 was markedly enhanced by the in-
troduction of a diazenyl pyrrolidine moiety. The present results might
open the way to the design of a new type of therapeutic agent based
on organobismuth compounds.
⁎
Corresponding authors.
⁎⁎ Corresponding author. Tel./fax:+81 79 565 8473.
E-mail addresses: s-yasuike@hokuriku-u.ac.jp (S. Yasuike),
yamaguchi@kph.bunri-u.ac.jp (K. Yamaguchi), tyagura@kwansei.ac.jp (T. Yagura).
0162-0134/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jinorgbio.2012.09.009

78
K. Onishi et al. / Journal of Inorganic Biochemistry 117 (2012) 77–84
2. Experimental
4H, 2×CH2), 2.30 (s, 6H, 2×Tol), 3.47 (br, 2H, CH2), 3.80 (br, 2H,
CH2), 7.05 (t, 1H, J=6.9 Hz, Ar\H), 7.14 (d, 4H, J=7.5 Hz, Ar\H),
7.32 (t, 1H, J=6.9 Hz, Ar\H), 7.60 (d, 4H, J=7.5 Hz, Ar\H), 7.62
(d, 1H, J=6.9 Hz, Ar\H), and 7.66 (d, 1H, J=6.9 Hz, Ar\H). 13C
NMR (CDCl3) δ: 21.5 (q), 23.7 (br), 46.8 (br), 51.5 (br), 116.6 (d),
127.9 (d), 128.6 (d), 130.9 (d), 136.7 (s), 137.9 (d), 138.1 (d), 152.2
(s), 153.0 (s), and 154.3 (s). Electron ionization-mass spectrometry
(EI-MS) m/z: 474 (M+-pTol), anal. (%) calcd. for C24H26BiN3: C,
50.98; H, 4.63; N, 7.43; found: C, 50.93;H, 4.58; N, 7.40.
2.1. Synthesis: general procedures and materials
Melting points were taken on a Yanagimoto micro melting point
hot-stage apparatus (MP‐S3) and are not corrected. 1H NMR
(tetramethylsilane: δ: 0.00 as an internal standard) and 13C NMR
(CDCl3: δ: 77.00 as an internal standard) spectra were recorded on a
JEOL JNM‐ECA400 (400 MHz and 100 MHz) in CDCl3 unless other-
wise stated. Mass spectra (MS) were obtained on a JEOL JMS-DX300
(70 eV, 300 μA). All chromatographic separations were accomplished
with Silica Gel 60 N (Kanto Chemical Co., Inc.). Thin-layer chromatog-
raphy (TLC) was performed with Macherey–Nagel Pre-coated TLC
plates Sil G25 UV254. 1-(Phenyldiazenyl)pyrrolidine (1) [20] and
tri(p-tolyl)bismuthane (2) [21] were prepared according to reported
procedures.
2.2. X-ray crystal structural determinations of 3 and 4
X-ray data was collected on a Bruker APEX II diffractometer with a
CCD detector with graphite monochromated radiation (MoKα, λ=
0.71073 Å) at 120 K. Empirical absorption correction was applied to
the collected reflections with SADABS and the space group was deter-
mined using XPREP. The structure was solved by the direct methods
using SHELXS-97 and refined on F2 by full-matrix least-squares
using the SHELXL-97 program package [22]. All hydrogen atoms
were included in calculated positions and treated as riding atoms by
using default SHELXL parameters (Table 1).
2.1.1. Synthesis of 1-[2-(di-p-tolylstibanophenyl)diazenyl]pyrrolidine
(3)
To a solution of 1-[(2-iodophenyl)diazenyl]pyrrolidine (903 mg,
3 mmol) in tetrahydrofuran (30 ml) was n-BuLi (3.1 ml, 5.1
mmmol, 1.65 mol/l in hexane) at −78 °C under an argon atmosphere
added, and the solution was stirred for 20 min at the same tempera-
ture. A solution of p-Tol2SbBr [prepared from tri(p-tolyl)stibane
(1.58 g, 4 mmol) and tribromoantimony (722 mg, 2 mmol)] in
ether (20 ml) was then added dropwise with stirring over a 15 min
period at −78 °C. After an additional 2 h of stirring at the same tem-
perature, the mixture was allowed to warm slowly to room tempera-
ture, then stirred for another 18 h. Ether (50 ml) and water (50 ml)
were added to the reaction mixture with stirring at 0 °C. The resulting
organic layer was separated and the aqueous layer was extracted
with ether (50 ml×2). The combined organic layer was washed
with brine, dried over anhydrous MgSO4, and evaporated in vacuo.
The residue was subjected to column chromatography on silica gel
with a mixture of hexane and dichloromethane (5:3) to give 3 as col-
orless prisms with hexane and ether (5:2). 3: 582 mg, 40% yield, mp
127–130 °C. 1H NMR (CDCl3) δ 1.90 (m, 4H, 2×CH2), 2.31 (s, 6H,
2×Tol), 3.38 (br, 2H, CH2), 3.79 (br, 2H, CH2), 6.69 (t, 1H, J=7.3Hz,
Ar\H), 7.07 (d, 1H, J=7.3 Hz, Ar\H), 7.08 (d, 4H, J=7.8 Hz,
Ar\H), 7.30 (t, 1H, J=7.3 Hz, Ar\H), 7.31 (d, 4H, J=7.8 Hz,
Ar\H), and 7.51 (d, 1H, J=7.3 Hz, Ar\H). 13C NMR (CDCl3) δ: 21.4
(q), 23.5 (br), 46.5 (br), 51.7 (br), 116.2 (d), 125.7 (d), 129.3 (d),
129.4 (d), 135.6 (d), 135.7 (s), 136.1 (s), 136.5 (d), 137.5 (s), and
154.1 (s). Electron ionization-mass spectrometry (EI-MS) m/z: 477
(M+), anal. (%) calcd. for C24H26N3Sb: C, 60.27; H, 5.48; N, 8.79;
found: C, 60.35; H, 5.43; N, 8.74.
2.3. Reagents
The organometallic compounds (1–4) used in this study are
shown in Fig. 1. They were resolved in DMSO at 50 mM and stocked
at −80 °C prior to use. Hydroxyphenyl fluorescein (HPF) was pur-
chased from Sekisui Medical (Tokyo, Japan); the WST-8 assay kit
from Kishida (Osaka, Japan); and Annexin V-HiLyte Fluor™ 488
from AnaSpec International (San Jose, CA, USA). All other reagents
were from Sigma or Wako Chemicals (Osaka, Japan) and of the
highest quality commercially available.
2.4. Cells and culture
All the human cancer cell lines, except NB4 and TIG cell lines, used
in this study were obtained from the American Type culture Collection
(Manassas, VA, U.S.A.). TIG-3 was obtained from Japanese Collection of
Research Bioresource (Osaka, Japan) and NB4 was kindly provided by
Dr. Junko Kado (Saitama Cancer Center Research Institute, Saitama,
Japan). K562, NB4 and HL60 cells were cultured in standard culture
conditions (37 °C, 95% air: 5% CO2) in RPMI1640 medium (Nissui;
Tokyo, Japan) supplemented with 10% fetal bovine serum (Gibco),
Table 1
Crystal data and structure refinement parameters for compounds 3 and 4.
Parameters
3 (M=Sb)
4 (M=Bi)
2.1.2.
Synthesis
of
1-[2-(di-p-tolylbismuthanophenyl)diazenyl]
pyrrolidine (4)
Empirical formula
Formula weight
Temperature (K)
Crystal size (mm)
Crystal system
Space group
a (Å)
b (Å)
c (Å)
α (°)
β (°)
C24H26N3Sb
478.23
120
0.18×0.13×0.10
Orthorhombic
Pbca
17.5856(12)
9.1101(6)
27.2491(18)
90
90
90
4365.5(5)
8
1.455
C24H26N3Bi
565.46
120
0.21×0.10×0.10
Monoclinic
P21/c
17.255(2)
5.8701(7)
21.926(3)
90
102.908(2)
90
2164.7(4)
4
1.735
1096
To a solution of 1-[(2-iodophenyl)diazenyl]pyrrolidine (903 mg,
3 mmol) in tetrahydrofuran (30 ml) was n-BuLi (3.2 ml, 5.3
mmmol, 1.65 mol/l in hexane) at −78 °C under an argon atmosphere
added, and the solution was stirred for 20 min at the same tempera-
ture.
A suspension of p-Tol2BiCl [prepared from tri(p-tolyl)
bismuthane (964 mg, 2 mmol) and trichlorobismuthane (315 mg,
1 mmol)] in ether (15 ml) was then added dropwise with stirring
over a 10-min period at −78 °C. After an additional 2 h of stirring
at the same temperature, the mixture was allowed to warm slowly
to room temperature, then stirred for another 16 h. Ether (50 ml)
and water (50 ml) were added to the reaction mixture with stirring
at 0 °C. The resulting organic layer was separated and the aqueous
layer was extracted with ether (30 ml×2). The combined organic
layer was washed with brine, dried over anhydrous MgSO4, and evap-
orated in vacuo. The resulting solids were recrystallized from a mix-
ture of hexane and ether (5:1) to give 4 as colorless prisms
(808 mg, 48% yield), mp 136–139 °C. 1H NMR (CDCl3) δ: 1.93 (m,
γ (°)
V (A3)
Z
Density (Mg/m3)
F(000)
1936
20,448
4540
0.0231
1.017
0.0224, 0.0512
Reflections collected
Independent reflections
R(int)
12,537
4935
0.0444
1.036
Goodness-of-fit on F2
R1, wR2 [I>2σ(I)]
0.0317, 0.0648

K. Onishi et al. / Journal of Inorganic Biochemistry 117 (2012) 77–84
79
Fig. 1. Molecular structure of 3 (Sb) and 4 (Bi). (Hydrogens are omitted.)
2 mM L-glutamine and 1% penicillin–streptomycin. Other human can-
cer cell lines were cultured in DMEM (Nissui) supplemented with 10%
fetal bovine serum (Gibco), 2 mM L-glutamine and 1% penicillin–
streptomycin at 37 °C under 5% CO2.
propidium iodide. Cells were analyzed with a Becton-Dickinson
FACSCalibur flow cytometer.
2.8. Determination of intracellular ROS production
2.5. Quantitation of metals accumulated in cells
ROS were detected using hydroxyphenyl fluorescein (HPF). HPF
fluoresces upon reacting with highly reactive oxygen species (hROS)
[25]. Briefly, NB4 cells (1×105/ml)were pretreated with 10 μM HPF
for 30 min. Then the compound was added and the culture was con-
tinued for 1 h at 37 °C. After incubation, the cells were pelleted by
centrifugation at 2500 rpm for 3 min, resuspended in 500 μl of
phosphate-buffered saline (PBS) and analyzed with the flow
cytometer.
Cells were cultured in 100-mm dishes and treated with the
organobismuth compound or organoantimony compound for 3 h.
The monolayer was washed two times with phosphate-buffered sa-
line (PBS) (0.5 ml of SDS buffer (50 mM Tris, 2% SDS, 10% glycerol,
pH 6.8) was added to each dish, and the cells were lysed and collected
into microtubes. The tubes were incubated for 10 min at 95 °C. The
lysate of 100 μl was digested at 130 °C for 2 days with nitric acid–
H2O2 in an aluminum dry-block bath. The samples were diluted
with 0.1 N nitric acid and the antimony or bismuth concentration
was measured by inductively coupled plasma mass spectrometry
(inductively coupled plasma mass spectrometry (ICP-MS), ELAN
DRC II, PerkinElmer, MA USA).
2.9. Assay of microtubule assembly in vitro
The assembly of porcine tubulin was monitored using porcine
brain tubulin purified as described earlier [24]. Tubulin was
resuspended on ice in ice-cold tubulin-polymerization buffer
(80 mM piperazine-1, 4-bis (2-ethane sulfonic acid (PIPES) pH 6.9,
2 mM MgSO4, 1 mM ethylene glycol tetra acetic acid (EGTA),
1 mM Mg-GTP, and 1.5% (v/v) glycerol) and 100 μl (250 μg) was
pipetted into the designated wells of a half area 96-well plate
pre-warmed to 25 °C in 0.20 mM of each compound. The assay was
conducted at 25 °C, and tubulin polymerization was monitored at
340 nm in a Multiscan micro plate reader (Thermo Labsystems,
Franklin, MA, USA). The increase in absorbance was measured at
340 nm at 25 °C and recorded every 30 s for 30 min.
2.6. Analysis of inhibition of cell growth
The cytotoxity of organometallic compounds in human cancer cell
lines was determined by the WST-8 assay [23] using a modified meth-
od according to the manufacturer's instructions. In all assays, the
compounds were dissolved in DMSO and diluted in sterile culture
medium immediately before addition to the cell culture, and the con-
centration of DMSO did not exceed 0.1% (v/v). The cells (1×105/ ml)
were seeded in 96-well plates. After 12 h, various concentrations of
1–3 or 4 (0.001 μM–10 μM in case for 1–3; 0.001 μM–20 μM for 4)
were added. After 12 h treatment, WST-8 was added and the plates
were incubated for an additional 4 h. The medium was discarded
and the formazane blue that formed in the cells was dissolved with
100 μl of DMSO. The optical density was measured at 570 nm.
2.7. Apoptosis quantified by FACS analysis
Apoptosis was assessed by annexin-V staining and propidium io-
dide exclusion using methods described previously [23]. After treat-
ment with the indicated concentrations of compounds for 12 h, cells
were harvested and cell pellets were resuspended in a staining solu-
tion containing annexin V-HiLyte Fluor 488 and then stained with
Scheme 1. Synthesis of the organoanitimony 3 and organobismuth 4.

80
K. Onishi et al. / Journal of Inorganic Biochemistry 117 (2012) 77–84
Fig. 2. Chemical structure of the compounds.
2.10. Assessment of mitochondrial membrane potential (ΔΨmt
)
bond (1.25 Å). On the other hand, N2\N3 bonds [3 (M_Sb):
1.315(2) Å, 4 (M_Bi): 1.318(7) Å] and N1\C4 bonds [3 (M_Sb):
1.420(3) Å, 4 (M_Bi): 1.426(7) Å] are shorter than the typical single
N\N (1.45 Å) and N\C bonds (1.47 Å) [27], respectively. These devi-
ation of N_N, N\N and N\C bond lengths indicate a delocalization
of the π and n electrons on triazene group. Intramolecular interaction
between pnictogen elements (in particular Sb, Bi) and nitrogen atoms
has attracted interest [28]. We also have reported that triarylstibane
and bismuthane derivatives frequently interact with nearly nitrogen,
producing Sb, Bi⋯N intramolecular coordination [29–31]. In the case
of 3 and 4 having triazene moiety, intramolecular interaction was absent
between antimony/bismuth and nitrogen (N2) in solid-state: the spatial
distances of Sb\N2 and Bi\N2 are 4.235(2) Å (3: Sb) and 4.257(5) Å
(4: Bi) which corresponds to 113% and 108% of the sum of the van der
Waals radii of both elements (N\Sb: 3.74 Å, N\Bi: 3.94 Å) [27]. The cen-
tral antimony and bismuth atoms exhibit a pyramidal structure, and the
three M\C bond lengths and the bond angles (C\M\C) surrounding
of antimony and bismuth were similar to trivalent organoantimony and
bismuth such as triphenylstibane (2.143–2.169 Å, 95.1–98.0°) [32] and
triphenylbismuthane (2.237–2.273 Å, 92.7–94.7°) [33].
NB4 cells (1×105/ml) were incubated in the absence or presence
of the organobismuth compounds. The cells were harvested at 6 h
post-treatment, pelleted by centrifugation at 2500 rpm for 3 min,
resuspended in 500 μl of rhodamine 123 (20 μg/ml) and incubated
for 15 min at 37 °C. After washing with PBS, cells were resuspended
in PBS and analyzed with the flow cytometer.
2.11. Statistical analysis
All the results were expressed as the mean S.D. for data obtained
from at least three separate experiments. Data were analyzed for sta-
tistical significance by the paired t test, and Pb0.05 was considered to
indicate statistically significant differences.
3. Results
3.1. Synthesis and molecular structure of organoantimony and
organobisumuth compounds 3, 4
The synthesis of novel organoantimony and organobisumuth com-
pounds having a triazene moiety, 1-[2-(di-p-tolylstibanophenyl)
diazenyl]pyrrolidine (3) and 1-[2-(di-p-tolylbismuthanophenyl)
diazenyl]pyrrolidine (4), is shown in Scheme 1. 1-[(2-iodophenyl)
diazenyl]pyrrolidine [26] was treated with n-BuLi in anhydrous THF
at −78 °C under an argon atomosphere, followed by trapping with
an electrophilic metal reagent [p-Tol2SbBr and p-Tol2BiCl] to give 3
and 4, respectively, in 40 and 48% yields.
The molecular structures of 3 and 4 are confirmed by their ele-
mental analyses and spectral techniques (1H and 13C NMR, MS). Crys-
tal structures of 3 and 4 are shown in Fig. 2, and selected bond lengths
and angles are listed in Table 2. The N_N double bonds in
diazoamino group (–N_N\Nb) of 3 and 4 adopt a trans configura-
tion, and the N1_N2 bond distances [3 (M_Sb): 1.278(2) Å, 4
(M_Bi): 1.260(7) Å] are larger than the typical value for N_N
3.2. Cytotoxic activity
Cytotoxic effects of the compounds depicted in Fig. 2 on NB4 cells
were examined using the WST‐8 assay. As shown in Fig. 3, 4 in which
a phenylmethyl was changed to a phenyldiazenyl exhibited strong
antiproliferative activity; the viability of NB4 cells was decreased at
a concentration of more than 0.1 μM. On the other hand, compounds
1–3 exerted only a weak anti-proliferative effect up to a concentra-
tion of 10 μM. The cytotoxic effect of 1–3 could not be determined
above 10 μM, because limited solubility resulted in the precipitation
of the compound which had an apparent effect on cell viability. Com-
pound 2, tri(p-tolyl)bismuthane, showed only a weak cytotoxic
Table 2
Selected bond distances (Å) and angles (°) for triazene compounds 3 and 4.
3 (M_Sb)
4 (M_Bi)
N1\N2
N2\N3
N1\C4
Sb\C1
Sb\C2
Sb\C3
Sb\N1
Sb\N2
Sb\N3
1.278(2)
1.315(2)
1.420(3)
2.151(2)
2.150(2)
2.163(2)
2.989(2)
4.235(2)
4.757(2)
95.38(7)
96.28(7)
95.96(7)
N1\N2
N2\N3
N1\C4
Bi\C1
Bi\C2
Bi\C3
Bi\N1
Bi\N2
Bi\N3
1.260(7)
1.318(7)
1.426(7)
2.240(6)
2.263(5)
2.274(6)
3.027(5)
4.257(5)
4.736(5)
95.1(2)
Fig. 3. Cell viability assay. Viability of NB4 cells after exposure for 12 h at 37 °C to in-
creasing concentrations of 1 (▲), 2 (■), 3 (◆), or 4 (×). Viability was determined by
WST-8 assay. Each point is the mean S.D. for three different experiments performed
in triplicate.
C1\Sb\C2
C1\Sb\C3
C2\Sb\C3
C1\Bi\C2
C1\Bi\C3
C2\Bi\C3
92.3(2)
91.98(19)

K. Onishi et al. / Journal of Inorganic Biochemistry 117 (2012) 77–84
81
activity toward NB4 cells. This result is consistent with previous tox-
A
300
250
200
150
100
50
NB4
icological results for this compound including low cytotoxicity in vas-
cular endothelial cells [19,34]. Arata et al. also reported that 30 μM of
triphenylbismuth was needed to achieve clear cytotoxic effect on rat
thymocytes [35]. Unexpectedly, 3, the compound containing antimo-
ny in place of bismuth, showed no cytotoxic activity up until 10 μM
either.
Next, 50% growth inhibition (IC50) values of 4 for various human
cancer cell lines and normal fibroblast cells (TIG) were determined
at concentrations of 0.001 μM–20 μM (Table 3). Leukemia cell lines
were relatively sensitive compared to other cell lines. Among the leu-
kemia cell lines tested, NB4 cells were the most sensitive. The IC50 of 4
(0.88 μM) for NB4 cells was low compared to that of the heterocyclic
organobismuth(III) (bi-chlorodibenzo[c,f ][1,5]thiabismocine) com-
pound that exhibited potent bioactivity against human cancer cell
lines (IC50 =0.05–0.25 μM) [18].
0
0
12
24
Time (h)
B
300
250
200
150
100
50
K562
We next examined the time course of cytotoxic effect of 4 on NB4,
K562, and HeLa cells, respectively, at each IC50 concentration (Fig. 4).
The viability of control culture increased during the culture because
of cell proliferation. On the other hand, the treatment with 4 inhibited
the proliferation of all the cell lines examined, and further reduced
their viability time-dependently.
3.3. Accumulation of the compounds in NB4 cells
As shown in Fig. 3, the toxicity study indicated the antimony com-
pound 3 to show no activity, although 4 was active. To explain this
difference in cytotoxic activity, the accumulation of metals in treated
cells was examined. After treatment with 4 or 3 at 0.88 μM for 1 h,
the amount of bismuth or antimony accumulated in NB4 cells was de-
termined using luminescence FT-IR. The amount of bismuth was
53.22 2.74 fmol/μg DNA, but that of antimony in the 3-treated
cells was only 0.75 0.49 fmol/μg DNA. Tests were performed using
other organometallic compounds and the results indicated that the
bismuth compounds were more cytotoxic and caused more accumu-
lation in cells than the corresponding antimony compounds. The rea-
son for this difference is unknown.
0
0
12
24
Time (h)
C
300
250
200
150
100
50
HeLa
3.4. Apoptosis induction
To verify whether cell damage is attributable to the induction of
apoptosis, NB4 cells were treated for 12 h with the indicated
concentrations of compound and then the ratio of apoptotic cells
was determined by estimating FITC-labeled annexin-V binding to
phosphatidylserine, as diagnostic of apoptosis [36] (Fig. 5). As
shown in Fig. 5A and B, the ratio of apoptotic NB4 cells treated with
1 or 2 did not increase compared with the non-drug control culture.
On the other hand, 4 began to induce apoptosis at around its IC50
0
0h
12h
24h
Time (h)
Fig. 4. Time course of cytotoxic effect of 4. NB4 (A), K562 (B) and HeLa (C) cells were
treated for the indicated times with each IC50 value of 4 (NB4, 0.88 μM; K562, 3.33 μM;
HeLa, 5.36 μM). Cell viability was examined using WST-8 assay. Cell treatments were
as follows: ○ (dashed line), control; □, treated with IC50 value for each cell line. Each
point is the mean S.D. for three different experiments performed in triplicate.
Table 3
and there was a concentration-dependent increase in the population
of cells in apoptosis (Fig. 5C). We also examined the induction of ap-
optosis by counting cells that exhibited typical apoptotic morpholog-
ical changes (i.e., condensed nuclei and chromatin fragmentation)
under the fluorescence microscope and obtained the similar results
with those as the annexin-V labeling procedure.
Cytotoxic activity of triazene compound 4 in several human cancer cell lines.
Human tumor type
Leukemia
Cell line
IC50 (μM)a
NB4
K562
HL60
MDA-MB-435S
DLD-1
MIA paca
A549
MG-63
HT1080
HeLa
0.880 0.102b
3.331 0.932
1.445 0.034
3.344 0.574
3.501 1.106
2.589 0.011
3.344 0.574
4.830 0.510
8.478 0.297
5.360 0.534
7.716 1.323
Breast cancer
Colon cancer
Pancreatic cancer
Lung cancer
Osteosarcoma
Fibrosarcoma
Cervical cancer
Normal human cell fibroblasts
3.5. Compound 4 increases in ROS levels and reduces mitochondrial
membrane potential disruption(ΔΨmt
)
TIG-3
In an attempt to gain insight into the molecular basis of the cyto-
toxic effect of 4, we first examined whether ROS levels were increased.
ROS has been shown to play important roles in the initiation and
a
IC50 was obtained through the 4 Parameter Logistic nonlinear regression model.
, standard deviation of three independent experiments.
b

82
K. Onishi et al. / Journal of Inorganic Biochemistry 117 (2012) 77–84
control
1 (10µM)
2 (10µM)
4 (5µM)
A
apo.84.82%
apo.27.52%
apo.16.81%
apo.15.27%
AnnexinV
AnnexinV
AnnexinV
AnnexinV
B
C
Fig. 5. Induction of apoptosis. Representative annexin-V/PI bivariate dot-blot figures displaying apoptotic NB4 cells after treatment with 1, 2 or 4 for 12 h at the concentration
depicted under the figure (A). Apoptotic cells were detected by the annexin-V and propidium iodide (PI) method. Annexin-V and PI stainings were measured by flow cytometry.
X axis, the fluorescent intensity of Annexin-V. Y axis, the fluorescent intensity of PI. The upper right quadrant shows later apoptotic and necrotic cells and the lower right quadrant,
early apoptotic cells. Total percentages of the Annexin-V-positive, PI-negative cells plus both positive cells, are shown in each panel as a percentage of apoptotic cells. Percentage of
apoptotic cells after 12 h treatment with increased concentrations of 1 (◆) or 2 (■) (B), or 4 (▲) (C). Rate of apoptotic cells quantified by flow cytometry as shown in (A). Each point
is the mean S.D. for three different experiments performed in triplicate.
execution of apoptosis [18,37]. ROS production was evaluated with a
fluorescent probe, 2-[6-(4′-hydroxy)phenoxy-3H-xanthen-3-on-
9-yl]benzoic acid (HPF). HPF detects highly reactive oxygen species
(hROS) such as the hydroxyl radical (OH) and reactive intermediates
of peroxidase and is resistant to light-induced autoxidation, unlike
DCFH-DA [38]. As shown in Fig. 6, the treatment of NB4 cells with
4 increased ROS levels (Pb0.05 vs. control). However the ROS
level was not significantly changed by the 12 h treatment with 1
or 2. Our previous study has shown that caspase activation, a key
process in apoptosis, followed a loss of ΔΨmt [18]. Thus, we next ex-
amined changes in ΔΨmt after the 12 h treatment with bismuth
compounds. As summarized in Fig. 7, 4 caused a marked loss of
ΔΨmt as compared with 1 or 2 in treated cells. Taken together,
these results support the idea that 4 induces apoptosis in NB4 cells
by a ROS-mediated pathway, as we have previously demonstrated
in the case of a heterocyclic organobismuth (III) compound [18].
Fig. 6. ROS production in NB4 cells treated with organobismuth compounds. An in-
crease in intracellular peroxide production in NB4 cells treated with 1, 2, or 4 was de-
termined by flow cytometric analysis using DCHF-DA as a fluorescent dye. Cells were
treated with a concentration of two-fold the IC50 of each compound for 12 h. The
ROS generated are indicated as a relative value against the control ROS production.
Data are expressed as % and are the mean S.D. for three different experiments
performed in triplicate. **Pb0.01 versus control.
Fig. 7. Mitochondrial membrane potential disruption. Cells were cultured with differ-
ent organobismuth compounds at a concentration of two-fold the IC50 value for 12 h,
then stained with rhodamine 123 and analyzed immediately by flow cytometry. The
number indicates the percentage of cells with reduced ΔΨmt. Data are expressed as %
and are the mean S.D. for three different experiments performed in triplicate.
*Pb0.05, ***Pb0.001 versus control.

K. Onishi et al. / Journal of Inorganic Biochemistry 117 (2012) 77–84
83
3.6. Suppression of 4-induced ROS production and apoptosis by
cotreatment with antioxidant, N-acetyl-cysteine (NAC)
Next, we designed experiments to determine whether 4-induced
ROS generation and apoptotic cell death were attenuated by NAC,
an antioxidant [39]. The present results showed that pretreatment
with NAC conferred significant protection against 4-induced ROS pro-
duction and apoptosis in NB4 cells (Fig. 8).
3.7. Inhibition of tubulin polymerization in vitro
We have previously shown that heterocyclic organobismuth(III)
targets tubulin resulting in depolymerization in vitro and destroys
Fig. 9. Inhibition of tubulin polymerization in vitro. Organobismuth compounds
inhibited tubulin assembly in vitro. MAP-rich tubulin in a reaction buffer was incubated
A
200
at 25 °C in the presence of DMSO (control, 0.01% ), nocodazole (20 μM) (■), 1 (▲)
◆
and 4 (×) (25 μM, each). The polymerization of tubulin was determined by measuring
180
160
the increase in absorbance over time at 340 nm.
the microtubular network of HeLa cells [24]. Microtubules function
in a number of cellular processes and numerous anti-tumor drugs
have been known to target microtubules [40]. Many heavy metal
compounds have also been shown to damage microtubules [41].
Then, we examined the effect of three compounds on the assembly
of tubulin subunits into microtubules in vitro. As shown in Fig. 9, 1
and 4 did not affect tubulin polymerization, whereas nocodazole, a
positive control drug, inhibited it. In accordance with these results,
neither of these two compounds disrupted the cellular microtubule
network of HeLa cells treated with their IC50 concentrations for 12 h
(data not shown).
140
*
120
100
80
60
40
20
0
4. Discussion
control
NAC
compound4
compound4+NAC
We describe the synthesis of a novel organobismuth (III) with
the diazenyl pyrrolidine
4 and revealed that the tri(p-tolyl)
bismuthane 2 can be modified from an inactive to highly toxic com-
pound against human cancer cell lines by the introduction of a
diazenyl pyrrolidine at the ortho position of one of the benzene
rings. Bismuth compounds, especially pentavalent organobismuth
compounds, have been known to serve as an oxidant under mild
conditions in organic chemistry [42]. However, the mechanism of
action of the organobismuth compound in cells is not fully resolved.
Inorganic bismuth(III) compounds, like bismuth subsalicylate, were
reported to make a complex with cysteine or glutathione (GSH) in
vitro [43,44]. As has been shown, GSH is used as a reducing sub-
strate for the function of glutathione peroxidase in cells [45]; it
can be supposed that the change in cellular GSH content affected
the intracellular redox environment and resulted in the elevation
of ROS levels [18]. However, our preliminary experiments with
heterocyclic organobismuth(III) using an in vitro assay system
with GSH as substrate did not show any evidence of the direct in-
hibition of glutathione peroxidase activity. In addition, we have re-
vealed that one specific biological target of heterocyclic
organobismuth(III) is tubulin [24]. However, as shown in Fig. 9, 4
is not a tubulin-disrupter. Thus, an unknown target of 4 should
exist to trigger apoptosis. In this respect, the diazenylpyrrolidine
moiety would be important to the bioactivity of 4 because com-
pounds with the triazene structure are known to have antimalarial
activity [16]. Several drugs for cancer undergo metabolic conver-
sion and their metabolites are significantly more toxic to cancer
cells [46,47]. Triazene compounds also are known to activate
through a metabolic pathway [14,48]. Although 4 shows high a de-
gree of stability during synthesis and purification, some metabolic
activation cannot be excluded.
B
40
35
30
25
20
15
10
5
**
0
control
NAC
compound 4
compound4+NAC
Fig. 8. NAC protected against 4-induced ROS production and apoptosis. NB4 cells were
cultured in the absence or presence of 10 mM NAC for 2 h, and then, exposed to 1.8 μM
(2×IC50) 4 alone or in combination of NAC for 12 h. ROS production was determined
by flow cytometric analysis using DCHF-DA (A). Apoptotic cells were detected by the
annexin-V and propidium iodide (PI) method (B). In panels, results are mean S.D.
for three different experiments performed in triplicate. *Pb0.05, **Pb0.01 versus com-
pound 4.

84
K. Onishi et al. / Journal of Inorganic Biochemistry 117 (2012) 77–84
[13] J.L. Skibba, D.D. Beal, G. Ramirez, G.T. Bryan, Cancer Res. 30 (1970) 147–150.
5. Conclusion
[14] J.M. Reid, M.J. Kuffel, J.K. Miller, R. Rios, M.M. Ames, Clin. Cancer Res. 5 (1999)
2192–2197.
In conclusion, introduction of the triazene structure (–N_N\NR2)
into a benzene ring of tri(p-tolyl)bismuthane gave potent cytotoxic
activity against human tumor cell lines through the induction of apo-
ptosis. The results of this study should aid the development of
organometallics-based drugs for leukemia because compounds with
multiple intracellular targets designed using the pharmacophore
model would be more effective against drug-resistant tumors.
[15] M. Bower, E.S. Newlands, N.M. Bleehen, M. Brada, R.J. Begent, Cancer Chemother.
Pharmacol. 40 (1997) 484–488.
[16] K. Nishiwaki, A. Okamoto, K. Matsuo, Y. Kawaguchi, Y. Hayase, K. Ohba, Bioorg.
Med. Chem. 15 (2007) 2856–2859.
[17] T. Kotani, D. Nagai, K. Asahi, H. Suzuki, F. Yamao, N. Kataoka, T. Yagura,
Antimicrob. Agents Chemother. 49 (2005) 2729–2734.
[18] K. Iuchi, Y. Hatano, T. Yagura, Biochem. Pharmacol. 76 (2008) 974–986.
[19] T. Yagura, Res. Adv. Antimicrob. Chemother. 6 (2006) 19–32.
[20] N.I. Foster, Synthesis (1980) 572–573.
[21] S. Combes, J.-F. Finet, Synth. Commun. 26 (1996) 4569–4575.
[22] G.M. Sheldrick, Acta Crystallogr. A64 (2008) 112–122.
[23] M. Ishiyama, Y. Miyazono, K. Sasamoto, Y. Ohkura, K. Ueno, Talanta 44 (1997)
1299–1305.
Acknowledgments
[24] K. Iuchi, K. Akagi, T. Yagura, J. Pharmacol. Sci. 109 (2009) 573–582.
[25] K. Setsukinai, Y. Urano, K. Kakinuma, H.J. Majima, T. Nagano, J. Biol. Chem. 278
(2003) 3170–3175.
[26] M.L. Bell, R.C. Chiechi, C.A. Johnson, D.B. Kimball, A.J. Matzger, W.B. Wan, T.J.R.
Weakley, M.M. Haley, Tetrahedron 57 (2001) 3507–3520.
[27] J. Emsley, The Elements, Third ed. Clarendon Press, Oxford, 1998.
[28] K.-y. Akiba, Chemistry of Hypervalent Compounds, VCH, New York, 1999.
[29] T. Tokunaga, H. Seki, S. Yasuike, M. Ikoma, J. Kurita, K. Yamaguchi, Tetrahedron 56
(2000) 8833–8839.
Partial financial support for this work was provided by a grant-in
aid for Scientific Research (J. K.) from the Ministry Education, Sciences,
Sports and Culture of Japan, and Special Research Found from
Hokuriku University.
Appendix A. Supplementary data
[30] S. Yasuike, Y. Kishi, S. Kawara, K. Yamaguchi, J. Kurita, J. Organomet. Chem. 691
(2006) 2213–2220.
[31] M. Kawahata, S. Yasuike, I. Kinebuchi, K. Yamaguchi, J. Kurita, Acta Crystallogr.
E67 (2011) m25.
[32] E.A. Adams, J.W. Kolis, W.T. Pennington, Acta Crystallogr. C46 (1990) 917–919.
[33] P.G. Jones, A. Blaschette, D. Henschel, A. Weitze, Z. Kristallogr. 210 (1995)
377–378.
[34] Y. Fujiwara, M. Mitani, S. Yasuike, J. Kurita, T. Kaji, J. Health Sci. 51 (2005)
333–340.
Supplementary data are available from the CCDC, 12 Union Road,
Cambridge CB2 1EZ, U.K., e-mail: deposit@ccdc.cam.ac.uk, on request,
quoting the deposition nos. CCDC 882185 (3) and 882186 (4)
respectively.
References
[1] Y.F. Shealy, C.A. Krauth, Nature 210 (1966) 208–209.
[2] S.K. Carter, M.A. Friedman, Eur. J. Cancer 8 (1972) 85–92.
[3] R.C.S. Audette, T.A. Connors, H.G. Mandel, K. Merai, W.C.J. Ross, Biochem.
Pharmacol. 22 (1973) 1855–1864.
[4] F.A. Schmid, D.J. Hutchison, Cancer Res. 34 (1974) 1671–1675.
[5] S.M. Lee, N. Thatcher, D. Crowther, G.P. Margison, Br. J. Cancer 69 (1994)
452–456.
[6] B.J. Jean-Claude, A. Mustafa, N.D. Cetateanu, Z. Damian, J. De Marte, R. Yen, D.
Vasilescu, T.H. Chan, B. Leyland-Jones, Br. J. Cancer 76 (1997) 467–473.
[7] S.L. Matheson, J. Mcnamee, B.J. Jean-Claude, J. Pharmacol. Exp. Ther. 296 (2001)
832–840.
[35] T. Arata, Y. Oyama, K. Tabaru, M. Satoh, H. Hayashi, S. Ishida, Y. Okano, Environ.
Toxicol. 17 (2002) 472–477.
[36] M. Vanengeland, L.J.W. Nieland, F.C.S. Ramaekers, B. Schutte, C.P.M.
Reutelingsperger, Cytometry 31 (1998) 1–9.
[37] C. Fleury, B. Mignotte, J.L. Vayssiere, Biochimie 84 (2002) 131–141.
[38] J.A. Royall, H. Ischiropoulos, Arch. Biochem. Biophys. 302 (1993) 348–355.
[39] I.A. Cotgreave, Adv. Pharmacol. 38 (1997) 205–227.
[40] M.A. Jordan, L. Wilson, Nat. Rev. Cancer 4 (2004) 253–265.
[41] I.N. Chou, Biomed. Environ. Sci. 2 (1989) 358–365.
[42] In: H. Suzuki, Y. Matano (Eds.), Organobismuth Chemistry, Elsevier, Amsterdam,
The Netherlands, 2001.
[8] T. Cimbora-Zovko, A. Brozovic, I. Piantanida, G. Fritz, A. Virag, B. Alic, V. Majce, M.
Kocevar, S. Polanc, M. Osmak, Eur. J. Med. Chem. 46 (2011) 2971–2983.
[9] A. Franchi, G. Papa, S. D'Atri, D. Piccioni, M. Masi, E. Bonmassar, Haematologica 77
(1992) 146–150.
[10] K. Seiter, D. Liu, T. Loughran, A. Siddiqui, P. Baskind, T. Ahmed, J. Clin. Oncol. 20
(2002) 3249–3253.
[11] M. Turriziani, P. Caporaso, L. Bonmassar, F. Buccisano, S. Amadori, A. Venditti, M.
Cantonetti, S. D'Atri, E. Bonmassar, Pharmacol. Res. 53 (2006) 317–323.
[12] R. Preussman, H. Druckrey, S. Ivankovic, A. von Hodenberg, Ann. N. Y. Acad. Sci.
163 (1969) 697–716.
[43] N. Burford, M.D. Eelman, D.E. Mahony, M. Morash, Chem. Commun. 7 (2003)
146–147.
[44] P.J. Sadler, H. Sun, H. Li, Chem. Eur. J. 2 (1996) 701–708.
[45] I.A. Cotgreave, P. Moldeus, R. Brattsand, A. Hallberg, C.M. Andersson, L. Engman,
Biochem. Pharmacol. 43 (1992) 793–802.
[46] H.V. Aposhian, Annu. Rev. Pharmacol. Toxicol. 37 (1997) 397–419.
[47] C. Rodriguez-Antona, M. Ingelma-Sundberg, Oncogene 25 (2006) 1679–1691.
[48] B.L. Pool, J. Cancer Res. Clin. Oncol. 93 (1979) 221–231.