An in vitro comparative assessment with a series of new triphenyltin(IV) 2-/4-[(E)-2-(aryl)-1-diazenyl]benzoates endowed with anticancer activities: Structural modifications, analysis of efficacy and cytotoxicity involving human tumor cell lines

Article abstract of DOI:10.1016/j.jinorgbio.2011.10.008

Four new triphenyltin(IV) complexes of composition Ph₃SnLH (where LH = 2-/4-[(E)-2-(aryl)-1-diazenyl]benzoate) (1-4) were synthesized and characterized by spectroscopic (¹H, ¹³C and ¹¹⁹Sn NMR, IR, ¹¹⁹Sn M?ssbauer) techniques in combination with elemental analysis. The ¹¹⁹Sn NMR spectroscopic data indicate a tetrahedral coordination geometry in non-coordinating solvents. The crystal structures of three complexes, Ph₃SnL¹H (1), Ph₃SnL³H (3), Ph₃SnL⁴H (4), were determined. All display an essentially tetrahedral geometry with angles ranging from 93.50(8) to 124.5(2)°; ¹¹⁹Sn M?ssbauer spectral data support this assignment. The cytotoxicity studies were performed with complexes 1-4, along with a previously reported complex (5) in vitro across a panel of human tumor cell lines viz., A498, EVSA-T, H226, IGROV, M19 MEL, MCF-7 and WIDR. The screening results were compared with the results from other related triphenyltin(IV) complexes (6-7) and tributyltin(IV) complexes (8-11) having 2-/4-[(E)-2-(aryl)-1-diazenyl]benzoates framework. In general, the complexes exhibit stronger cytotoxic activity. The results obtained for 1-3 are also comparable to those of its o-analogs i.e. 4-7, except 5, but the advantage is the former set of complexes demonstrated two folds more cytotoxic activity for the cell line MCF-7 with ID₅₀ values in the range 41-53 ng/ml. Undoubtedly, the cytotoxic results of complexes 1-3 are far superior to CDDP, 5-FU and ETO, and related tributyltin(IV) complexes 8-11. The quantitative structure-activity relationship (QSAR) studies for the cytotoxicity of triphenyltin(IV) complexes 1-7 and tributyltin(IV) complexes 8-11 is also discussed against a panel of human tumor cell lines.

Full text of DOI:10.1016/j.jinorgbio.2011.10.008

Journal of Inorganic Biochemistry 107 (2012) 119–128
Contents lists available at SciVerse ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
An in vitro comparative assessment with a series of new triphenyltin(IV)
2-/4-[(E)-2-(aryl)-1-diazenyl]benzoates endowed with anticancer activities:
Structural modifications, analysis of efficacy and cytotoxicity involving human tumor
cell lines
a
b
b
c
Tushar S. Basu Baul a, , Anup Paul , Lorenzo Pellerito , Michelangelo Scopelliti , Andrew Duthie ,
⁎
Dick de Vos d, Rajeshwar P. Verma e,1, Ulli Englert f
Department of Chemistry, North-Eastern Hill University, NEHU Permanent Campus, Umshing, Shillong 793 022, India
Dipartimento di Chimica “Stanislao Cannizzaro”, Università degli Studi di Palermo, Viale delle Scienze, Parco D'Orleans II, Edificio 17, 90128 Palermo, Italy
School of Life & Environmental Science, Deakin University, Geelong Victoria 3217, Australia
Pharmachemie BV, PO Box 552, 2003 RN Haarlem, The Netherlands
Department of Chemistry, Pomona College, 645 North College Avenue, Claremont, CA 91711, USA
Institut für Anorganische Chemie, RWTH Aachen University, 52056 Aachen, Germany
a
b
c
d
e
f
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 May 2011
Received in revised form 10 September 2011
Accepted 17 October 2011
Available online 25 October 2011
Four new triphenyltin(IV) complexes of composition Ph3SnLH (where LH=2-/4-[(E)-2-(aryl)-1-diazenyl]
benzoate) (1–4) were synthesized and characterized by spectroscopic (1H, 13C and 119Sn NMR, IR, 119Sn Möss-
bauer) techniques in combination with elemental analysis. The 119Sn NMR spectroscopic data indicate a tetra-
hedral coordination geometry in non-coordinating solvents. The crystal structures of three complexes,
Ph3SnL1H (1), Ph3SnL3H (3), Ph3SnL4H (4), were determined. All display an essentially tetrahedral geometry
with angles ranging from 93.50(8) to 124.5(2)°; 119Sn Mössbauer spectral data support this assignment. The
cytotoxicity studies were performed with complexes 1–4, along with a previously reported complex (5) in
vitro across a panel of human tumor cell lines viz., A498, EVSA-T, H226, IGROV, M19 MEL, MCF-7 and
WIDR. The screening results were compared with the results from other related triphenyltin(IV) complexes
(6–7) and tributyltin(IV) complexes (8–11) having 2-/4-[(E)-2-(aryl)-1-diazenyl]benzoates framework. In
general, the complexes exhibit stronger cytotoxic activity. The results obtained for 1–3 are also comparable
to those of its o-analogs i.e. 4–7, except 5, but the advantage is the former set of complexes demonstrated two
folds more cytotoxic activity for the cell line MCF-7 with ID50 values in the range 41–53 ng/ml. Undoubtedly,
the cytotoxic results of complexes 1–3 are far superior to CDDP, 5-FU and ETO, and related tributyltin(IV) com-
plexes 8–11. The quantitative structure-activity relationship (QSAR) studies for the cytotoxicity of triphenyltin
(IV) complexes 1–7 and tributyltin(IV) complexes 8–11 is also discussed against a panel of human tumor cell lines.
© 2011 Elsevier Inc. All rights reserved.
Keywords:
Anti-cancer drugs
Triphenyltin(IV) benzoates
Triphenyltin(IV) 2-/4-[(E)-2-(aryl)-1-diaze-
nyl]benzoates
Cell lines
QSAR
1. Introduction
revealed in their remarkable therapeutic potential reflected in recent
research reports [6–17]. Consequently, a large number of organotin(IV)
Organotin(IV) compounds are a widely studied class of metal-based
anti-tumor drugs and their intensive investigation has led to the
discovery of compounds with excellent in vitro anti-tumor activity, but
in many cases disappointingly low in vivo potency or high in vivo
toxicity [1–3]. It is well established that organotin(IV) compounds are
very important in cancer chemotherapy because of their apoptosis-
inducing character [4,5]. The design of improved organotin(IV) anti-
tumor agents occupies a significant place in cancer chemotherapy, as
carboxylates have been investigated for their anti-tumor potential.
Among organotin(IV) carboxylates, triorganotin(IV) carboxylates
are quite well known for exceptionally high in vitro anti-tumor
activities, e.g., triphenyltin(IV) -benzoates, -salicylates [18], -3,6-
dioxaheptanoate, -3,6,9-trioxadecanoate [19], -4-carboxybenzo-15-
crown-5, -4-carboxybenzo-18-crown-6 [19,20], -steroidcarboxylate
[21], -terebate [22,23,24] and -aminoacetates (Schiff bases) [25,26].
From these examples, it is clear that the compounds can be developed
with high in vitro antitumor activity and sufficient water solubility. The
most important point remains the activity. The organotin(IV)
compounds containing the diazenyl group show not only high in vitro
antitumor activity, but also displayed interesting interactions with
various enzymes (see below). In the present study, attempts have been
made to improve the water solubility by the systematic study of various
⁎
Corresponding author. Tel.: +91 364 2722626; fax: +91 364 2721000.
E-mail addresses: basubaul@nehu.ac.in, basubaul@hotmail.com (T.S. Basu Baul),
Rajeshwar.Verma@fda.hhs.gov, rajeshwarv@yahoo.com (R.P. Verma).
1
Present address (Rajeshwar P. Verma): U.S. Food and Drug Administration, Center
for Food Safety and Applied Nutrition, OFAS/OCAC, 5100 Paint Branch Parkway, College
Park, MD 20740, USA.
0162-0134/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2011.10.008

120
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
structural changes. In general, organotin(IV) compounds are dissolved in
DMSO and diluted with test medium prior to the in vitro testing. The
limited solubility needs further improvement in a way comparable to
cisplatin which shows limited water solubility too.
(Hi Media) were used without further purification. The solvents
used in the reactions were of AR grade and were dried using standard
procedures. Toluene was distilled from sodium benzophenone ketyl.
In view of the remarkable activity of the triphenyltin(IV) carboxylates,
triphenyltin(IV) carboxylates containing the 2-[(E)-2-(3-formyl-4-hydro-
xyphenyl)-1-diazenyl]benzoate and 2-[(E)-2-(4-hydroxy-5-methylphe-
nyl)-1-diazenyl]benzoate skeletons have recently been investigated,
showing encouraging cytotoxic activity across a panel of cell lines [27].
As a result of these promising cytotoxic activities, the mechanistic role
of the compounds was investigated to determine the influence of the
azo group nitrogen. Docking studies were performed with some of the
key enzymes, such as ribonucleotide reductase, thymidylate synthase,
thymidylate phosphorylase and topoisomerase II, which take part in the
synthesis of raw materials for DNA and its replication [28]. The docking
studies indicated that the azo group nitrogen atoms and formyl, carbonyl
and ester oxygen atoms in the ligand moiety play an important role. They
exhibit hydrogen bonding interactions with the active site of amino acids
of the aforementioned enzymes. The higher activity was attributed to the
presence of the azo group nitrogen atoms in the molecules of triphenyltin
(IV) complexes [27]. As a continuation of our previous work in this area,
we report some new triphenyltin(IV) complexes, Ph3SnL1−4H (1–4), of
related systems where the ligand skeletal framework has been modified
(Scheme 1) in an attempt to improve the dissolution properties and
thereby influence cytotoxicity. The carboxylate ligands selected herein
have variations in the position of the carboxylate functionality in the
diazo part and also have variations of the nuclear substituents in the cou-
pling moieties of the molecule. The newly synthesized complexes (1–4)
were characterized by spectroscopic (1H, 13C and 119Sn NMR, IR, 119Sn Möss-
bauer) techniques. Complete characterization was accomplished from the
crystal structure determination of some representative complexes Ph3-
SnL1H (1), Ph3SnL3H (3) and Ph3SnL4H (4). The newly synthesized triphe-
nyltin(IV) complexes (1–4) and one previously reported triphenyltin(IV)
complexes Ph3SnL5H (5) [29] were tested across a panel of human tumor
cell lines consisting of A498 (renal cancer), EVSA-T (mammary cancer),
H226 (non-small-cell lung cancer), IGROV (ovarian cancer), M19 MEL
(melanoma), MCF-7 (mammary cancer) and WIDR (colon cancer) and
the results were compared with analogous Ph3SnL6H (6), Ph3SnL7H (7)
[27] and related tributyltin(IV) complexes (8–11) (see Scheme 1 for com-
plex description).
2.2. Physical measurements
Carbon, hydrogen and nitrogen analyses were performed with a
Perkin Elmer 2400 series II instrument. IR spectra in the range
4000–400 cm–1 were obtained on a Perkin Elmer Spectrum BX series
FT-IR spectrophotometer with samples investigated as KBr disks. The
1H and 13C spectra were recorded on a Bruker AMX 400 spectrometer
and measured at 400.13 and 100.62 MHz, respectively, while 119Sn
NMR spectra were recorded on a Jeol GX 270 spectrometer and mea-
1
sured at 100.75 MHz. The H, 13C and 119Sn chemical shifts were refer-
enced to Me4Si, CDCl3 and Me4Sn set at 0.00, 77.0 and 0.00 ppm,
respectively. The 119Sn Mössbauer spectra (Table 1) were recorded at
liquid nitrogen temperature with a conventional instrument in trans-
mission mode, constituted by
a multichannel analyzer (TAKES
Mod.269, Ponteranica, Italy) and the following Wissenschaftliche Elek-
tronik system (MWE, München, Germany): MR250 driving unit, FG2
digital function generator and MA250 velocity transducer, moving at
linear velocity, constant acceleration, in a triangular waveform. The
organotin(IV) samples (1–4) were maintained at liquid nitrogen tem-
perature in a Cryo NDR-1258-MD liquid nitrogen cryostat (Cryo Indus-
tries of America, Inc. Atkinson, NH, USA) with a Cryo sample holder. The
77.3 0.1 K temperature was controlled with an Oxford Instruments
ITC 502 temperature controller (Oxford, UK). The multichannel cali-
bration was performed with an enriched iron foil (α57Fe, 4 μm
thick, RITVERC GmbH, St. Petersburg, Russia), at room temperature,
by using a 57Co/Rh source (10 mCi, RITVERC GmbH, St. Petersburg,
Russia), while the zero point of the Doppler velocity scale was deter-
mined, at room temperature, through absorption spectra of natural
CaSnO3 (
119Sn=0.5 mg cm–2) and a Ba119SnO3 source (10 mCi, RIT-
VERC GmbH). The obtained 5·105 count spectra were interpreted
by means of non-linear least square analysis as a sum of Lorentzian
doublets, to obtain the isomer shift, δ 0.03 mm s–1, the nuclear
quadrupole splitting, |Δexp
|
0.03 mm s–1 and the average full
width at half height, Γav
,
0.03 mm s–1
.
2.3. Synthesis of ligands and complexes
2. Experimental
Ligands 4-[(E)-2-(4-hydroxy-3-methylphenyl)-1-diazenyl]benzoic
acid (L2HH′) [31], 4-[(E)-(5-tert-butyl-2-hydroxyphenyl)diazenyl]ben-
zoic acid (L3HH′) [31,32], 2-[(E)-(5-tert-butyl-2-hydroxyphenyl)diaze-
nyl]benzoic acid (L4HH′) [31,33], and complex 5 [29] were prepared by
the methods described in our earlier reports and purities were estab-
lished from melting point, elemental analysis and 1H NMR
spectroscopy.
2.1. Materials
Ph3SnOH was prepared from Ph3SnCl (Fluka) by following the lit-
erature method [30]. (Ph3Sn)2O (Fluka), 2-hydroxybenzaldehyde
(Sisco), 2-methylphenol, 4-methylphenol, 4-tert-butylphenol
(Merck), anthranilic acid (Spectrochem), and 4-aminobenzoic acid
7
Scheme 1. Structures and numbering protocol of triphenyltin(IV) complexes Ph3SnL1− H (1–7).

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
121
1
Table 1
1608 ν(OCO)as. H-NMR (DMSO-d6); δ H: 9.60 [brs, 1H, OH], 8.95 [dd,
2.5 Hz, 8.0 Hz, 2H, H2/6], 7.88 [m, 6H, Sn-Pho], 7.77 [dd, 2.5 Hz,
8.0 Hz, 2H, H3/5], 7.68 [d, 2.5 Hz, 1H, H6′], 7.61 [dd, 2.5, 8 Hz, 1H,
H2′], 7.40 [m, 9H, Sn-Phm/p], 6.88 [d, 8 Hz, 1H, H5′], 2.26 [s, 3H, CH3]
ppm. 13C-NMR (DMSO-d6); δ C: 168.5 [CO2], 141.7 [Sn-Phi], 135.5 [Sn-
Pho], 127.8 [Sn-Php], 127.2 [Sn-Phm], 15.3 [CH3], other carbons: 158.6,
153.3, 144.6, 129.6, 124.1, 123.9, 122.8, 120.7, 113.8 ppm. 119Sn-NMR
(CDCl3): −105.6 ppm.
Experimentala and calculated Mössbauer parameters, including C-Sn-O angles for the
triphenyltin(IV) complexes 1–4.
b
Complex
δ
|Δexp
|
Γav
Δcalcd.
C-Sn-O angle
1
2
3
4
1.17
1.19
1.16
1.16
2.27
2.24
2.44
2.40
0.76
0.84
0.99
0.96
−2.22
−2.22
−2.22
−2.22
109.07
109.31
107.63
107.98
119
a
Conditions: 77 K, 0.50–0.60 mg cm−2 Sn. δ, Δ, and Γ values are in mm s−1 and
angles are in degrees, and the errors limits are 0.03 mm s−1 and 13°, respectively.
2.3.4. Synthesis of Ph3SnL3H (3)
b
Complex 3 was prepared by reacting Ph3SnOH with L3HH′ using
the procedure described for 2 in anhydrous toluene. Dried residue
was extracted in chloroform, concentrated, precipitated with hexane
and filtered. The crude product was dried and recrystallized with ac-
etone, affording orange crystalline material in 56% yield. M.p.: 130–
132 °C. Anal. Calc. for C35H32N2O3Sn: C, 64.92; H, 4.98; N, 4.33%.
Calculated assuming a regular tetrahedral structure. Used partial quadrupole splitting
(p.q.s.) values (mm s-1): {Ph}=−1.26 [51], {COO−
}unid =−0.15[51].
2.3.1. Synthesis of 4-[(E)-2-(2-hydroxy-5-methylphenyl)-1-diazenyl]
benzoic acid (L1HH′)
Ligand L1HH′ was prepared by reacting p-carboxybenzenediazonium
chloride with 4-methylphenol in alkaline solution under cold condi-
tions, following the method described for its o-analog L5HH′ [29].
Several recrystallizations from hot methanol yielded an orange
crystalline product in 60% yield; M.p.: 260–261 °C. Anal. Calc. for
Found. C, 65.08; H, 4.85; N, 4.35%. IR absorption (cm–1): 1621
1
ν(OCO)as. H-NMR (CDCl3): δ H: 12.5 [brs, 1H, OH], 8.26 [dd, 2.5 Hz,
8.0 Hz, 2H, H2/6], 7.87 [m, 3H, H3/5/6′], 7.78 [m, 6H, Sn-Pho], 7.45
[m, 10H, Sn-Phm/p/H4′], 6.91 [d, 8 Hz, 1H, H3′], 1.31 [s, 9H, CH3]
ppm. 13C-NMR (CDCl3); δ C: 171.6 [CO2], 138.1 [Sn-Phi], 136.8 [Sn-
Pho], 131.8 [Sn-Php], 128.8 [Sn-Phm], 30.3 [CH3], other carbons:
153.0, 150.7, 142.4, 136.7, 130.0, 129.9, 121.6, 117.8 ppm. 119Sn-NMR
(CDCl3): −104.6 ppm.
C
14H12N2O3: C, 65.62; H, 4.72; N, 10.93%. Found. C, 65.28; H, 4.65; N,
10.85%. IR absorption (cm–1): 1680 ν(OCO)as. 1H-NMR (DMSO-d6): δ
H: 8.19 [dd, 2.5 Hz, 8.0 Hz, 2H, H2/6], 7.91 [dd, 2.5 Hz, 8.0 Hz, 2H, H3/
5], 7.75 [d, 2.5 Hz, 1H, H6′], 7.20 [dd, 2.5, 8 Hz, 1H, H4′], 6.93 [d, 8 Hz,
1H, H3′], 2.39 [s, 3H, CH3] ppm. Signal for carboxylic and phenolic
protons were exchanged due to presence of water in the solvent. 13C-
NMR (DMSO-d6); δ C: 167.1 [CO2], 19.7 [CH3], other carbons: 152.6,
150.1, 136.7, 134.6, 132.3, 130.4, 128.9, 121.4, 117.4 ppm.
2.3.5. Synthesis of Ph3SnL4H (4)
Complex 4 was prepared by reacting Ph3SnOH with L4HH′ by fol-
lowing an analogous procedure described for complex 3. The product
was recrystallized from chloroform to afford orange crystals in 76%
yield. M.p.: 138–140 °C. Anal. Calc. for C35H32N2O3Sn: C, 64.92; H,
4.98; N, 4.33%. Found. C, 65.18; H, 4.90; N, 4.35%. IR absorption (cm–1):
1606 ν(OCO)as 1H-NMR (CDCl3): δ H: 12.7 [brs, 1H, OH], 8.22 [d, 8 Hz,
1H, H3], 7.84 [m, 8H, Sn-Pho/H6/H6′], 7.56 [t, 8 Hz, 1H, H5], 7.42
[m, 11H, Sn-Phm/p/H4/H4′], 6.96 [d, 8 Hz, 1H, H3′], 1.37 [s, 9H, CH3]
ppm. 13C-NMR (CDCl3); δ C: 171.3 [CO2], 137.1 [Sn-Phi], 136.0
[Sn-Pho], 129.0 [Sn-Php], 127.7 [Sn-Phm], 30.3 [CH3], other carbons:
149.5, 149.4, 140.9, 136.3, 131.7, 131.6, 130.0, 128.9, 128.7, 127.1,
126.1, 117.2, 114.9 ppm. 119Sn-NMR (CDCl3): −104.2 ppm.
2.3.2. Synthesis of Ph3SnL1H (1)
Complex 1 was synthesized by reacting L1HH′ (0.30 g, 1.17 mmol)
and (Ph3Sn)2O (0.42 g, 0.59 mmol) in anhydrous toluene (40 ml) in a
100 ml flask equipped with a Dean-Stark moisture trap and water-
cooled condenser. The reaction mixture was refluxed for 9 h and fil-
tered while hot. The filtrate was collected and the volatiles removed
using a rotary evaporator. The residue was dried in vacuo, washed
with hexane (3×5 ml), extracted into benzene and filtered. The fil-
trate was concentrated and petroleum ether added to precipitate
the crude product, which was collected by filtration and dried in
vacuo (m.p.: 152–154 °C). Recrystallization from benzene gave maroon
crystals of the desired product in 55% (0.40 g) yield. M.p.: 168–170 °C.
Anal. Calc. for C32H26N2O3Sn: C, 63.48; H, 4.33; N, 4.63%. Found. C,
64.28; H, 4.45; N, 4.70%. IR absorption (cm–1): 1624 ν(OCO)as. 1H-NMR
(CDCl3): δ H: 12.6 [brs, 1H, OH], 8.26 [dd, 2.5 Hz, 8.0 Hz, 2H, H2/6],
7.87 [dd, 2.5 Hz, 8.0 Hz, 2H, H3/5], 7.84 [m, 6H, Sn-Pho], 7.72 [d, 2.5 Hz,
1H, H6′], 7.46 [m, 9H, Sn-Phm/p], 7.14 [dd, 2.5, 8 Hz, 1H, H4′], 6.91 [d,
8 Hz, 1H, H3′], 2.35 [s, 3H, CH3] ppm. 13C-NMR (CDCl3); δ C: 172.2
[CO2], 138.6 [Sn-Phi], 137.4 [Sn-Pho], 130.7 [Sn-Php], 129.5 [Sn-Phm],
20.7 [CH3], other carbons: 153.6, 151.2, 135.6, 133.7, 133.0, 132.3,
129.8, 122.3, 118.5 ppm.119Sn-NMR (CDCl3): −104.6 ppm.
2.4. Experimental protocol and cytotoxicity tests
The in vitro cytotoxicity testing of triphenyltin(IV) complexes 1–5
was performed using the SRB test for estimation of cell viability. Ex-
perimental protocols for 6 and 7 have been reported [27] and the cy-
totoxic results are included here for convenience of discussion. The
cell lines WIDR colon carcinoma, M19 MEL melanoma, A498 renal
cell carcinoma, IGROV ovarian carcinoma and H226 non-small-cell
lung cancer belong to the currently used anticancer screening panel
of the NCI, USA [34]. The breast carcinoma MCF7 cell line is estrogen
receptor (ER)+/progesterone receptor (PgR)+and the breast carci-
noma cell line EVSA-T is (ER)-/(Pgr)-. Prior to the experiments, a my-
coplasma test was carried out on all cell lines and found to be
negative. All cell lines were maintained in a continuous logarithmic
culture in RPMI 1640 medium with HEPES and phenol red. The medi-
um was supplemented with 10% FCS, penicillin 100 μg/ml and strep-
tomycin 100 μg/ml. The cells were mildly trypsinized for passage
and for use in the experiments. RPMI and FCS were obtained from
Life Technologies or from Gibco (Paisley, Scotland). SRB, DMSO, pen-
icillin and streptomycin were obtained from Sigma (St. Louis MO,
USA), TCA and acetic acid from Baker BV (Deventer, NL) and PBS
from NPBI BV (Emmer-Compascuum, NL).
2.3.3. Synthesis of Ph3SnL2H (2)
Ph3SnOH (0.43 g, 1.17 mmol) was dissolved in anhydrous ben-
zene and added to an anhydrous methanol solution containing L2HH′
(0.30 g, 1.17 mmol). The reaction mixture was refluxed for 6 h and
the water formed during the reaction being removed azeotropically
using a Dean-Stark apparatus. The reaction mixture was filtered
while hot and the filtrate was evaporated to dryness using a rotary
evaporator. The dried mass was washed thoroughly with hexane,
extracted into anhydrous benzene and filtered. The filtrate was con-
centrated slowly on a hot plate, cooled to room temperature and
solid precipitated with hexane. The crude product was collected by fil-
tration, washed several times with hot hexane and recrystallized
using benzene to afford yellow crystalline material in 45% (0.33 g)
yield. M.p.: 148–150 °C. Anal. Calc. for C32H26N2O3Sn: C, 63.48; H,
4.33; N, 4.63%. Found. C, 63.50; H, 4.45; N, 4.45%. IR absorption (cm–1):
The test complexes 1–5 and reference compounds were dissolved
to a concentration of 250,000 ng/ml in full medium, by dilution of a
stock solution which contained 5 mg of complexes 1–5/ml in DMSO.
(Note: Complex 1 did not dissolve well in DMSO and hence needed
stirring at 37 °C and then at 50 °C, but it still did not dissolve

122
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
completely and therefore a suspension of the test complex 1 in DMSO
was used to make the dilutions in medium). The compounds were
subsequently diluted to a final concentration of 250,000 ng/ml in
full medium. Cytotoxicity was estimated by the microculture sulfor-
hodamine B (SRB) test [35].
[log 1/ID50 (obsd) — log 1/ID50 (pred)]>2 s, where s is the standard
deviation [41,44]. Each QSAR model includes 95% confidence limits
for each term in parentheses. Statistical diagnostics (r2, q2, s, Q, and
F) and internal validation (cross validation and Y-randomization)
tests have validated all the QSAR models. Due to the small data sets,
we did not perform the external validation test.
The experiment was started on day 0. On day 0, 1500–2000 cells
per well were seeded into 96-wells flat-bottomed micro-titer plates
(Cellstar, Greiner Bio-one). The plates were pre-incubated overnight
at 37 °C, 5% CO2 to allow the cells to adhere to the bottom. On day
2, a three-fold dilution sequence of ten steps was made in full medi-
um, starting with the 250,000 ng/ml stock solution. Every dilution
was used in quadruplicate by adding 50 μl to a column of four wells.
This procedure results in the highest concentration of 62,500 ng/ml
being present in column 12. Column 2 was used for the blank.
Medium was added to column 1 to diminish interfering evapora-
tion. On day 7, the incubation was terminated. Subsequently, the
cells were fixed with 10% trichloroacetic acid in PBS and placed at
4 °C for an hour. After three washings with tap water, the cells were
stained for at least 15 min with 0.4% SRB dissolved in 1% acetic acid.
After staining, the cells were washed with 1% acetic acid to remove
the unbound stain. The plates were air-dried and the bound stain
was dissolved in 150 μl 10 mM Tris-base. The absorbance was mea-
sured at 540 nm using an automated microplate reader (Labsystems
Multiskan MS). The data were used for construction of concentration-
response curves and determination of the ID50 values by use of Deltasoft
3 software. ID50 is the dose in ng/ml that causes 50% inhibition of the
tumor cells.
2.6. X-ray crystallography
Crystals of the triphenyltin(IV) complexes Ph3SnL1H (1), Ph3SnL3H
(3) and Ph3SnL4H (4) suitable for single crystal X-ray structure deter-
mination were obtained from slow evaporation of benzene, acetone
and chloroform solutions of the complexes, respectively. Intensity
data were collected with graphite-monochromated MoKα radiation
(λ=0.71073 Å) on a Bruker D8 goniometer equipped with a SMART
APEX CCD detector. Frames were collected in ω-scan mode and inte-
grated with the help of the program SAINT [45]. Crystal data, data col-
lection parameters and convergence results are listed in Table 2.
Absorption corrections based on a multi-scan approach [46] were ap-
plied to the data sets before averaging over symmetry-related reflec-
tions. The structures were solved by direct methods with the help of
the program SHELXS-97 [47] and refined on reflection intensities F2
using SHELXL-97 [47]. In the final least-squares refinements, non-
hydrogen atoms were refined with anisotropic displacement parameters
and hydrogen atoms were placed in idealized positions and included as
riding on the corresponding atoms.
The variability of the in vitro cytotoxicity test depends on the cell
lines used and the serum applied. With the same batch of cell lines
and the same batch of serum the inter-experimental CV (coefficient
of variation) is 1–11% depending on the cell line and the intra-
experimental CV is 2–4%. These values may be higher in the other
batches of cell lines and/or serum.
The quality of the intensity data sets varied considerably: only
room temperature data were available for 1. 18 distance restraints
and 656 rigid-bond restraints were used during refinement, and reli-
ability factors were rather high. In the case of 3, low temperature (130
(2) K) data were available; in each of the two symmetrically indepen-
dent molecules one of the phenyl rings attached to the metal was dis-
ordered over two orientations;
a total of 144 restraints were
2.5. Quantitative structure-activity relationship (QSAR) methods
employed in the final structure model, and partially occupied atom
QSAR models were developed by the C-QSAR program [36] using
multi-regression analyses (MRA). This program avoids the collineari-
ty problem by auto-selection of descriptors based on permutation and
correlation matrices among the descriptors. Details about this pro-
gram and its use in the development of QSAR models can be seen in
Table 2
Crystal data, data collection parameters and convergence results for complexes 1, 3 and 4.
1
3
4
Empirical formula
Formula weight
Crystal size (mm)
Crystal morphology
Temperature (K)
Crystal system
Space group
a (Å)
b (Å)
c (Å)
α (°)
β (°)
C
32H26N2O3Sn
C35H32N2O3Sn
647.32
C35H32N2O3Sn
647.32
previous publications [37,38]. The in vitro cytotoxicity data (ID50
;
605.24
ng/ml) of organotin(IV) compounds 1–20 along with some standard
drug molecules against a panel of seven human tumor cell lines are
summarized in Table 5. In QSAR analysis, we often like to convert
the concentration ID50 of the compound into an activity parameter
“A” using the following equation A=−log ID50 =log 1/ID50 in
molar concentration. This transformation represents that the more
potent compound always has a higher “activity” and vice versa
[39,40]. This is the reason the in vitro cytotoxicity data (ID50; ng/ml)
(see Table 5) was converted into an activity parameter log 1/ID50
(mol l–1) using the following equation: log 1/ID50 (mol l–1)=6–log
[ID50 (ng/ml)/MW] and then the QSAR models were developed.
Clog P is the calculated partition coefficient of a compound in octa-
nol/water system and is a measure of its hydrophobicity, while CMR
is the calculated molar refractivity for the whole molecule [36,41].
In all QSAR models, n is the number of data points, r2 is the square
of the correlation coefficient, q2 is the cross-validated r2, s is the stan-
dard deviation, Q is the quality factor, and F is the Fischer ratio. The
cross-validated r2 (q2) is obtained by using the leave-one-out (LOO)
procedure [42]. In each QSAR model, the value of F in parenthesis re-
fers to their literature value at 95% level [43]. The modeling was taken
to be optimal when Q reached a maximum together with F, even if
slightly non-optimal F values have normally been accepted. A com-
pound was assigned as an outlier on the basis of deviation between
observed and predicted activities from the respective QSAR model
0.30×0.15×0.15 0.42×0.23×0.03 0.28×0.18×0.04
Rod
293(2)
Triclinic
P1
11.580(2)
12.288(3)
20.158(4)
98.265(5)
94.693(4)
92.866(4)
2823.5(10)
4
Platelet
130(2)
Triclinic
Platelet
110(2)
Triclinic
ꢀ
ꢀ
ꢀ
P1
P1
9.6709(10)
14.5077(15)
21.438(2)
92.255(2)
95.299(2)
92.960
2988.0(5)
4
1.439
9.8449(6)
10.4548(7)
14.9268(10)
97.600(2)
92.305(2)
104.383(2)
1470.86(17)
2
γ (°)
V (Å3)
Z
Dx (g cm–3
)
1.424
0.939
0.765,0.872
1.462
0.907
0.782, 0.964
μ (mm–1
)
0.893
0.705, 0.973
Transmission
factors (min, max)
θ range (°)
Reflections measured
Independent reflections; 10,478; 0.0692
Rint
1.68–25.57
16,377
2.1–28.59
31,236
14,959; 0.0632
2.28–30.56
21,927
8347; 0.0612
Reflections with I>2σ(I) 10,478
14,959
737
144
0.081
0.110
1.284
0.97, –1.31
8347
373
0
0.038
0.048
1.00
Number of parameters
Number of restraints
R(F) [I>2σ(I)reflns]
wR(F2) (all data)
GOF(F2)
687
674
0.070
0.212
1.05
Δρmax, min (e, Å–3
)
0.67, –0.57
1.39, –0.82

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
123
Fig. 1. Displacement ellipsoid plot (30% probability) showing one of the two indepen-
dent molecules in the crystal of Ph3SnL1H (1) with the atom-labeling scheme.
Hydrogen atoms have been omitted for clarity.
Fig. 3. Displacement ellipsoid plot (50% probability) of a molecule in the crystal of Ph3-
SnL4H (4) with the atom-labeling scheme. Hydrogen atoms have been omitted for
clarity.
sites were assigned isotropic displacement parameters. Refinement of
4 was straightforward.
data can usually give information on the more or less covalent nature of
the bonds formed by tin with other and different atoms through determi-
nation of the isomer shift values, δ, and also an insight into the probable
structure of the complexes, both in the solid state or in frozen solution,
by the determination of experimental nuclear quadrupole splittings,
|Δexp|. The experimental Mössbauer spectra show, in all the cases, a single
resonant doublet with full width at half height broader than that of the
Ba119SnO3 source (~1 mm s–1) (see Γav in Table 1), which is consistent
with the occurrence of only one absorbing species. The experimental
Mössbauer parameters are reported in Table 1, together with calculated
Δ according to the point-charge model formalism applied to the idealized
tetrahedral structure [51–56]. The difference between experimental
and calculated values does not exceed the limit of tolerance of the
method ( 0.4 mm s-1) [55], and thus the experimental |Δexp| data
can be explained satisfactorily by assuming a slightly distorted tetra-
hedral geometry [56]. As a consequence, taking into account the pro-
posed tetracoordination of tin(IV) and using the Parish equation, the
calculated C–Sn–OCO angles for the complexes 1–4 lie in the range
107.6–109.3° [56] (see Table 1).
3. Results and discussion
3.1. Synthesis and spectroscopy
Ligands L1HH′-L5HH′ were prepared by reacting the appropriate
ortho-/para-carboxybenzenediazonium chloride with 2-methylphenol,
4-methylphenol or 4-tert-butylphenol in alkaline solution under cold
conditions [27,29,31,32]. Reactions of the ligand with (Ph3Sn)2O or
Ph3SnOH in a suitable solvent (see experimental section for details)
after proper work-up afforded triphenyltin(IV) complexes of composi-
tion Ph3SnLH with yields greater than 45%. The compounds are colored
solids that are air stable and soluble in all common organic solvents.
Analytical purities of the complexes were established by elemental an-
alyses and multinuclear NMR (1H, 13C and 119Sn) spectroscopic data.
The IR spectra of the triphenyltin(IV) complexes displayed a strong
sharp band at around 1620 cm−1 (for 1 and 3) and 1605 cm−1 (for 2
and 4) that has been assigned to the asymmetric carboxylate [ν(OCO)as]
stretching vibration, in accord with our earlier reports [27,29,31,48]. The
assignment of the symmetric [ν(OCO)sym] stretching vibration band
could not be made owing to the complex spectral pattern. All complexes
3.2. Crystal structures
1
displayed H and 13C signals due to ligand and Sn-Ph skeletons. The ob-
1
served H- and 13C-NMR signals are given in the experimental section.
The crystal structures of three of the triphenyltin complexes (1, 3
and 4) have been determined. Crystal data, data collection parameters
and refinement results are given in Table 2. All three complexes crys-
The 1H-NMR integration values were completely consistent with
1
the formulation of the products. The H and 13C chemical shift assign-
ꢀ
ment of the phenyltin moiety is straightforward from the multiplicity pat-
tallize in the triclinic space group P1. The crystal structures of 1 and 3
13
119
terns, resonance intensities and also by examining the nJ( C- 117Sn)
coupling constants [29,48,49]. In CDCl3, the triphenyltin(IV) complexes
(1–4) exhibit a single sharp 119Sn resonance at around −105 ppm, sug-
gesting that the Sn-atom in the complexes are isostructural in solution
where the tin atom is four-coordinate [27,29,49,50]. 119Sn Mössbauer
feature two independent molecules in the asymmetric unit. Views of
the molecules of 1, 3 and 4 are shown in Figs. 1–3. With respect to
the metal coordination, the structures conform to the same motif. Se-
lected geometric parameters are given in Table 3. As indicated by the
119Sn Mössbauer spectroscopy, the molecules are monomeric in the
solid state. The tin atom is four coordinate with a distorted tetrahedral
geometry defined by a C3O donor set. The tetrahedral angles for the
five molecules range between 93.50(8) and 124.5(2)°. Similar tetra-
hedral geometries about the Sn atom were observed in Ph3Sn
[O2CC6H4(N=NC6H3-2-OH-5-Me)-o] and its acetone solvated com-
plex [57,58], Ph3Sn{O2CC6H3-p-OH[N=N(C6H4-X)]} X=H, 2-Me,
3-Me, 4-Me, 4-OMe, 4-Cl [50,59,60] and Bu3Sn[O2CC6H4
(N=NC6H3-4-OH-5-Me)-p] [61]. The fifth potential donor atom, O
(2), remains at a significantly longer distance from the tin center
than O(1), with values between 2.564(2) and 2.862(7) Å. In agree-
ment with the different metal-oxygen distances, C(1)-O(1) bonds are
consistently longer than the C(1)-O(2) distances (Table 3). The second-
ary Sn-O(2) interaction causes a slight distortion of the tetrahedral pri-
mary coordination sphere but should not be regarded as a strong
coordinative bond: The bond angles around the Sn atom in 1, 3 and 4
are more consistent with a tetrahedral environment than with a trigonal
bipyramidal or square-pyramidal five-coordinate arrangement, and the
/
Fig. 2. Displacement ellipsoid plot (50% probability) showing one of the two indepen-
dent molecules in the crystal of Ph3SnL3H (3) with the atom-labeling scheme.
Hydrogen atoms and an alternative minority conformation for one of the phenyl
rings have been omitted for clarity.

124
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
Table 4
Hydrogen bonding geometry (Å, º) for 1, 3 and 4.
Compound
D-H···A
D-H
H···A
D···A
D-H···A
1 Molecule 1a
Molecule 2a
O(3)-H(3O)···N(1)
O(3A)-H(3P)···N(1A)
O(3)-H(3O)···N(1)
O(3A)-H(3P)···N(1A)
O(3)-H(3O)···O(3)i
O(3)-H(3O)···N(1)
O(3)-H(3O)···O(2)
0.82
0.82
0.84
0.84
0.84
0.84
0.84
1.87
1.90
1.92
1.90
2.54
1.90
2.12
2.569(11)
2.600(10)
2.628(7)
2.608(6)
2.843(6)
2.598(3)
2.777(2)
143
142
142
141
102
140
135
3 Molecule 1a
Molecule 2a
b
a
4
b
“i” refers to atoms from the next symmetrically-related molecule (symmetry code:
-x,2-y,-z).
a
Intramolecular H-bond.
Intermolecular H-bond (longer and weaker).
b
Fig. 4. Plot of observed log 1/ID50(M19
QSAR 5 (Table 6).
versus predicted log 1/ID50(M19
from
MEL)
Such delicate bonding interactions are regulated by the geometrical
features, size and presence of the donor atoms in the molecule. In
view of this, some new triphenyltin(IV) 4-[(E)-2-(aryl)-1-diazenyl]
benzoates (1–3) have been prepared by controlling and confining
the parameters described above. These complexes were tested for
their cytotoxic potential, along with two more triphenyltin(IV) 2-
[(E)-2-(aryl)-1-diazenyl]benzoates (4–5), across the panel of human
tumor cell lines A498, EVSA-T, H226, IGROV, M19 MEL, MCF-7 and
WIDR. The results of the in vitro cytotoxicity tests performed with tri-
phenyltin(IV) complexes (1–5) are summarized in Table 5 and the
screening results are compared with the results from other related tri-
phenyltin(IV) complexes (6–7) and tributyltin(IV) complexes (8–11)
having 2-/4-[(E)-2-(aryl)-1-diazenyl]benzoates framework. Subse-
quently, the results were evaluated with the triphenyltin(IV)- 2-{[(2Z)-
(3-hydroxy-1-methyl-2-butenylidene)]amino}-4-methyl-pentanoate,
2-{[(E)-1-(2-hydroxyphenyl)alkylidene]amino}-4-methyl-pentanoates,
2-{[(2Z)-(3-hydroxy-1-methyl-2-butenylidene)]amino}phenylpropio-
nate, 2-{[(E)-1-(2-hydroxyphenyl)alkylidene]amino}phenylpropionate
(12–17) [25,26], terebate, steroidcarboxylate and benzocrowncarboxy-
late (18–20) [23], which have shown promising activity. When looking
into the antitumor test results of triorganotin(IV) derivatives, the ques-
tion of hydrolysis yielding e.g. triphenyltin hydroxide as an active species
may arise. The potent and similar in vitro activity of triphenyltin(IV) ben-
zoates suggested triphenyltin hydroxide as the common intermediate
responsible for cytotoxicity [63] and this hypothesis was also proposed
for a series of dioxastannolanes. Later, this proposal was abandoned be-
cause of the observed resistance in the cell lines which was not seen with
triphenyltin hydroxide [64]. Organotin(IV) carboxylates proved to be
stable at least for days in water [65]. In our test system, either ethanol
or DMSO solution of the organotin(IV) compound is diluted with test
medium. In case of compounds insoluble in aqueous media, the test pro-
cedure is discontinued because of improper outcomes. Consequently, no
reliable test results could be obtained for triaryllead hydroxides and also
for some stannoxanes. This was also confirmed by others e.g. for triphe-
nyltin hydroxide [66,67]. Some poorly water soluble stannoxanes were
tested and showed high cytotoxicity [68–72]. Several series of organo-
tin(IV) compounds containing diazenyl group were studied for some
years and no signs of hydrolysis were observed during spectroscopic
measurements and in vitro testing protocols. These compounds invari-
ably displayed different ID50 values and different reactivities towards en-
zymes. The organotin(IV) compounds studied may form intermediates
in test medium containing tumor cells as with many cytotoxic com-
pounds, but there was no sign of a common intermediate.
MEL)
same hypothesis was drawn from the 119Sn Mössbauer data (vide supra).
Compounds 1, 3 and 4 represent typical van der Waals crystals without
strong directional intermolceular contacts (Table 4). In all three com-
pounds, the hydroxy oxygen atom O(3) acts as the donor and N(1) as
the acceptor of an intramolecular hydrogen bond with graph set symbol
S(6) [62]. As a result of the 1, 2 substitution in the aromatic ring C(2)-C
(7), molecules of 4 adopt a more compact conformation than those in 1
and 3. The packing diagrams for 1, 3 and 4 can be viewed in Figs. S1–S3.
3.3. Cytotoxicity studies
Triphenyltin(IV) 2-[(E)-2-(aryl)-1-diazenyl]benzoates, e.g. 6–7,
are an important class of compounds as they have shown both effec-
tive docking results, particularly the hydrogen bonding interactions
through the azo group nitrogen atoms, formyl, carbonyl and ester ox-
ygen atoms with various key enzymes, and cytotoxic potential [27].
Table 3
Selected geometric parameters (Å, °)a for the triphenyltin(IV) complexes 1, 3 and 4.
1
3
4
Sn(1)–O(1)
2.053(5)
2.079(4)
[2.072(4)]
2.655(4)
[2.775(4)]
2.140(6)
[2.136(6)]
2.109(7)
[2.135(6)]
2.109(7)
[2.115(7)]
1.326(7)
[1.318(7)]
1.237(8)
[1.220(7)]
54.57(15)
[52.22(15)]
100.5(2)
[93.9(2)]
111.3(2)
[109.3(2)]
111.1(2)
[104.4(2)]
105.0(2)
[112.3(2)]
109.4(2)
[107.7(2)]
117.9(2)
[124.5(2)]
104.4(4)
[107.9(4)]
2.094(2)
[Sn(1a)–O(1a)]
Sn(1)–O(2)
[2.058(6)]
2.749(6)
2.564(2)
2.135(2)
2.140(2)
2.126(2)
1.304(3)
1.239(3)
54.93(6)
111.53(8)
93.50(8)
112.27(8)
107.12(9)
118.06(9)
111.52(9)
103.29(14)
[Sn(1a)-O(2a)]
Sn(1)–C(15)
[Sn(1a)–C(15a)]
Sn(1)–C(21)
[Sn(1a)–C(21a)]
Sn(1)–C(27)
[Sn(1a)–C(27a)]
O(1)–C(1)
[O(1a)–C(1a)]
O(2)–C(1)
[O(2a)–C(1a)]
O(1)–Sn(1)–O(2)
[O(1a)–Sn(1a)–O(2a)]
O(1)–Sn(1)–C(15)
[O(1a)–Sn(1a)–C(15a)]
O(1)–Sn(1)–C(21)
[O(1a)–Sn(1a)–C(21a)]
O(1)–Sn(1)–C(27)
[O(1a)–Sn(1a)–C(27a)]
C(15)–Sn(1)–C(21)
[C(15a)–Sn(1a)–C(21a)]
C(15)–Sn(1)–C(27)
[C(15a)–Sn(1a)–C(27a)]
C(21)–Sn(1)–C(27)
[C(21a)–Sn(1a)–C(27a)]
C(1)–O(1)–Sn(1)
[C(1a)–O(1a)–Sn(1a)]
[2.862(7)]
2.067(10)
[2.085(11)]
2.107(10)
[2.069(10)]
2.042(10)
[2.074(11)]
1.321(10)
[1.309(11)]
1.221(10)
[1.217(11)]
52.7(2)
[50.1(2)]
95.0(3)
[94.4(4)]
111.3(3)
[108.4(3)]
109.2(3)
[110.1(3)]
109.1(4)
The triphenyltin(IV) complexes of the present investigation de-
serve specific mention. In complexes (4–7), the triphenyltin(IV) car-
boxylate is o-positioned in the diazo-forming moiety, while in
compounds 1–3 it is p-positioned (see Scheme 1). The cytotoxicity
data (ID50) for the test complexes (1–3) are of the same order of mag-
nitude and the change of ligand substitution does not influence the
cytotoxic activity significantly. The results of 1–3 are also comparable
to that of its o-analogs, i.e. 4–7, except 5. Pair wise comparison of the
[112.1(5)]
110.0(4)
[111.9(5)]
119.5(4)
[117.4(5)]
107.9(6)
[112.1(7)]
a
Values in square brackets with the atom label extension ‘a’ refer to the second in-
dependent molecule in the asymmetric unit.

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
125
Table 5
merit over their o-analogs (4–6) for better activity, particularly
against MCF-7 cell line. On the other hand, although the ID50 values
for the test complexes (1–4) of the present investigation are also sim-
ilar to that of the triphenyltin(IV) complexes of Schiff bases derived
from L-leucine and phenylalanine [25,26], their advantage lies in the
stability. It should be mentioned here that the triphenyltin(IV) 2/4-
[(E)-2-(aryl)-1-diazenyl]benzoates (1–7) are stable for a significantly
longer period in both the solid-state and in solution than triphenyltin
(IV) complexes containing amino acetate skeletons [25,26]. Thus, the
cytotoxicity data of complexes 1–3, together with better solubility, is
suggestive of promising candidates for further in vitro and in vivo
studies after appropriate modifications. The cytotoxicity data of com-
plexes 1–3 further indicate that structural variation of the L-R skele-
tons does not influence the activity. However, attempts were made
to provide additional precise information from the quantitative
structure-activity relationship (see below).
In vitro ID50 values (ng/ml) of test complexes 1–5, reported organotin(IV) complexes
(6–20) and standard drugs using cell viability tests in seven human tumor cell linesa.
Test complexb
Cell lines
A498 EVSA-T H226 IGROV M19
MEL
MCF-7 WIDR
Ph3SnL1H (1)
103
103
101
101
162
101
103
182
177
176
162
43
41
35
43
97
41
49
101
27
100
97
35
49
31
81
102
101
102
102
148
104
101
163
167
165
148
56
107
107
110
111
214
109
101
239
269
253
214
90
100
98
53
43
41
79
113
92
78
102
100
105
106
106
104
95
Ph3SnL2H (2)
Ph3SnL3H (3)
101
103
118
103
104
125
127
126
118
42
Ph3SnL4H (4)
Ph3SnL5H (5)c [29]
Ph3SnL6H (6)d [27]
Ph3SnL7H (7)d [27]
Bu3SnL2H (8)e [61]
Bu3SnL3H (9)e [61]
Bu3SnL4H (10)e [61]
Bu3SnL5H (11)e [61]
118
112
120
113
36
106
105
105
106
33
[Ph3SnL3H]n (12)f [26] 96
[Ph3SnL4H]n (13)f [26] 104
[Ph3SnL5H]n (14)f [26] 39
[Ph3SnL6H]n.nCCl4 (15)f 105
[25]
111
38
105
99
46
101
75
36
102
76
42
34
31
3.4. QSAR studies
111
106
From the cytotoxicity data [log 1/ID50 (mol l–1)] of organotin(IV)
compounds 1–11 in Table 6, the following QSAR models 1–7 were de-
veloped:QSAR for the cytotoxicity of compounds 1–11 against A498
human tumor cell line (Table 6)
[Ph3SnL7H]n (16)f [25] 120
100
96
b3
b3
b2
8
115
108
39
61
b2
199
3269 169
340
2287
105
106
19
18
b2
60
130
112
42
51
b2
115
110
17
110
109
17
[Ph3SnL8H]n (17)f [25] 113
Ph3SnR1 (18)g [23]
42
65
b2
90
Ph3SnR2 (19)g [23]
16
19
Ph3SnR3 (20)g [23]
2.9
b2
DOX
CDDP
5-FU
MTX
ETO
16
10
11
log1=ID50ðA498Þ ¼ −0:07ðꢀ0:05Þ Clog P þ 0:10ðꢀ0:04ÞCMR þ 5:47ðꢀ0:81Þ
n ¼ 11; r2 ¼ 0:826; s ¼ 0:065; q2 ¼ 0:738; Q ¼ 13:985; F2;8 ¼ 18:989ð4:459Þ
2253 422
558
442
23
505
b3.2
699
750
18
2594
b3.2
967
225
b3.2
150
b3.2
143
37
475
5
297
7
ð1Þ
1314 317
b3.2 b3.2
3934 580
b3.2 b3.2
TAX
QSAR for the cytotoxicity of compounds 1–11 against EVSA-T human
tumor cell line (Table 6)
a
Abbreviation: DOX, doxorubicin; CDDP, cisplatin; 5-FU, 5-fluorouracil; MTX, metho-
trexate; ETO, etoposide and TAX, paclitaxel.
b
Standard drug reference values are cited immediately after the test complexes under
log1=ID50ðEVSA−TÞ ¼ −0:11ðꢀ0:05Þ Clog P þ 0:17ðꢀ0:04ÞCMR þ 5:15ðꢀ0:68Þ
n ¼ 9; r2 ¼ 0:959; s ¼ 0:051; q2 ¼ 0:899; Q ¼ 19:196; F2;6 ¼ 70:171ð5:143Þ
identical conditions.
c
For reported triphenyltin(IV) complex 5, ID50 values (ng/ml) are now included. Refer
to Scheme 1 for ligand skeleton.
ð2Þ
d
ID50 values (ng/ml) were taken from ref. 27 for the reported triphenyltin(IV) com-
plexes (6 and 7) for comparison. Refer to Scheme 1 for ligand skeletons.
outliers: compounds 5 and 9
e
ID50 values (ng/ml) were taken from ref. 61 for the reported tributyltin(IV) complexes
(8–11) for comparison. Refer to Scheme 1 for ligand skeletons.
f
Reported triphenyltin(IV) complexes (12–17) have been included for comparison; see refs. 25 and
QSAR for the cytotoxicity of compounds 1–11 against H226 human
tumor cell line (Table 6)
3
26: LH is a carboxylate residue where L H, 2-{[(2Z)-(3-hydroxy-1-methyl-2-butenylidene)]amino}-
4-methyl-pentanoate; L4H, 2-{[(E)-1-(2-hydroxyphenyl)- methylidene]amino}-4-meth-
yl-pentanoate; L5H, 2-{[(E)-1-(2-hydroxyphenyl)-ethylidene]amino}-4-methyl-pentano-
ate; L6H, 2-{[(2Z)-(3-hydroxy-1-methyl-2-butenylidene)]amino}phenylpropionate; L7H,
2-{[(E)-1-(2-hydroxyphenyl)methylidene]- amino}phenylpropionate; L8H, 2-{[(E)-1-(2-
hydroxyphenyl)ethylidene]amino}-phenylpropionate.
log1=ID50ðH226Þ ¼ −0:06ðꢀ0:04Þ Clog P þ 0:09ðꢀ0:04ÞCMR þ 5:70ðꢀ0:67Þ
n ¼ 11; r2 ¼ 0:841; s ¼ 0:053; q2 ¼ 0:755; Q ¼ 17:302; F2;8 ¼ 21:157ð4:459Þ
ð3Þ
g
Reported triphenyltin(IV) complexes (18–20) have been included for comparison; see
ref. 23: R is a carboxylate residue where R1=−terebate, R2 =−steroidcarboxylate, R3=
−benzocrowncarboxylate.
QSAR for the cytotoxicity of compounds 1–11 against IGROV human
tumor cell line (Table 6)
log1=ID50ðIGROVÞ ¼ −0:12ðꢀ0:08Þ Clog P þ 0:14ðꢀ0:07ÞCMR þ 5:22ðꢀ1:23Þ
n ¼ 11; r2 ¼ 0:807; s ¼ 0:098; q2 ¼ 0:707; Q ¼ 9:163; F2;8 ¼ 16:725ð4:459Þ
complexes, wherein the ligand nuclear substituents are held constant,
was also necessary to judge the influence of the p- and o- derivatives.
For pair 1 and 5, the p- derivative 1 is found to be more active, while
for pairs 3 and 4, and 2 and 6, both p- and o- derivatives displayed
comparable activity. Upon closer inspection of the ID50 values of 1
and 5, one can see that 1 is two folds more cytotoxic than 5 for the
cell line MCF-7 and similar observations were noted for pairs 3 and
4, and 2 and 6. As noted, 5 is an exception; its lower cytotoxicity
might be the result of internal co-ordination of OH with Sn. This can
make the Sn less attractive for co-ordination with DNA or sugar moi-
eties. The larger t-Bu group might impede the internal co-ordination.
In a manner similar to 4–7, a comparison is possible for Ph3SnLH (2–
4) and Bu3SnLH (8–11) since each set of complexes contains a com-
mon ligand frame work. Triphenyltin(IV) complexes 2–4 were
found to be more active than corresponding tributyltin(IV) com-
plexes 8–11. In general, the cytotoxic results of complexes 1–3 are
undoubtedly far superior to CDDP, 5-FU and ETO, and related tributyl-
tin(IV) complexes 8–11. Triphenyltin(IV) complexes 1–3 deserve
ð4Þ
QSAR for the cytotoxicity of compounds 1–11 against M19 MEL
human tumor cell line (Table 6)
log1=ID50ðM19MELÞ ¼ −0:03ðꢀ0:02Þ Clog P þ 0:05ðꢀ0:02ÞCMR þ 6:13ðꢀ0:30Þ
n ¼ 11; r2 ¼ 0:887; s ¼ 0:024; q2 ¼ 0:823; Q ¼ 39:25; F2;8 ¼ 31:398ð4:459Þ
ð5Þ
QSAR for the cytotoxicity of compounds 1–11 against MCF-7 human
tumor cell line (Table 6)
log1=ID50ðMCF−7Þ ¼ −0:14ðꢀ0:05Þ Clog P þ 0:18ðꢀ0:04ÞCMR þ 5:10ðꢀ0:69Þ
n ¼ 7; r2 ¼ 0:983; s ¼ 0:039; q2 ¼ 0:947; Q ¼ 25:410; F2;4 ¼ 115:647ð6:944Þ
ð6Þ
outliers: compounds 4–7

126
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
Table 6
descriptors (Clog P and CMR) is not sufficient enough and may need
the additional descriptor(s), but the additional descriptor(s) cannot
be considered due to the very small data set. Lastly, we kept this
equation only for the comparison point of view.
Comparison between observed and predicted log 1/ID50 (mol l–1), Clog P, and CMR of
1–11.
No. Compound log 1/ID50
log 1/ID50
log 1/ID50
log 1/ID50
(A498)
(EVSA-T)
(H226)
(IGROV)
The presence of Clog P term with negative coefficient in QSARs
1–6 suggests that the cytotoxicity of compounds 1–11 decreases
with increasing hydrophobicity against six cancer cell lines (e.g.
A498, EVSA-T, H226, IGROV, M19 MEL, and MCF-7). On the contrary,
the cytotoxicity of these compounds increases with increasing their
steric/polarizability (CMR) against all the seven cancer cell lines as
evident by the positive coefficient of the CMR term. Although there
are high statistics associated with QSARs 1–6 (r2 =0.807–0.983),
the C log P of compounds 1–11 with high values (6.37–8.81) are not
great enough to establish the upper limit of the activity, these com-
pounds may have higher C log P values than that of the optimum. It
suggests that QSARs 1–6 may represent only the second half of the
parabolic/bilinear model in terms of C log P, which may be the
cause for the presence of negative C log P term in QSARs 1–6. Thus,
more compounds with lower C log P (C log Pb6.37) values will be
needed to establish the upper C log P limit either for the development
of a parabolic or bilinear QSAR model.
On comparison among QSAR models 1–7, one can suggest that the
mechanism for the cytotoxic activity of compounds 1–11 against
A498, EVSA-T, H226, IGROV, M19 MEL, and MCF-7 cancer cell lines
is almost the same and depends directly on the hydrophobic and
molar refractivity parameters of the compound. It must be noted
here that QSARs 1 and 3–5 explain 82.3–88.2% of the variance in the
data sets without any outlier, but QSARs 2 and 6 explain 95.9% and
98.3% of the variance in the data sets with the help of 2 and 4 outliers,
respectively. Thus QSARs 2 and 6 are not exactly the same to that of
QSARs 1 and 3–5. Now, it can be concluded that the mechanism for
the cytotoxic activity of compounds 1–11 against A498, H226,
IGROV, and M19 MEL cancer cell lines is almost the same, but some-
what different from that of against EVSA-T and MCF-7 cancer cell
lines. These cancer cell lines may have some different mechanism in
addition to the common one. On the other hand, QSAR 7 explains
90.3% of the variance in the data set using only one CMR descriptor
and one outlier. There is no room to accommodate the hydrophobic
term in this QSAR model, which suggests that the cytotoxic activity
of compounds 1–11 against WIDR cancer cell line may have some dif-
ferent mechanism in comparison to that of the other six cancer cell
lines such as A498, EVSA-T, H226, IGROV, M19 MEL, and MCF-7. The
similar mechanism for the cytotoxic activity of compounds 1–11
against A498, H226, IGROV, and M19 MEL cancer cell lines can also
be demonstrated by very high mutual correlations (r2 =0.948–
0.996) in activity–activity relationships i.e. the direct comparison
among the cytotoxic activity of compounds 1–11 against A498,
H226, IGROV, and M19 MEL cancer cell lines as shown in Table 7
and Figs. 5, S4–S8.
(Eq. (1))
(Eq. (2))
(Eq. (3))
(Eq. (4))
Obsd. Pred. Obsd. Pred. Obsd. Pred. Obsd.
Pred.
1
2
3
Ph3SnL1H
Ph3SnL2H
Ph3SnL3H
Ph3SnL4H
Ph3SnL5H
Ph3SnL6H
Ph3SnL7H
Bu3SnL2H
Bu3SnL3H
Bu3SnL4H
Bu3SnL5H
6.77
6.77
6.81
6.81
6.57
6.78
6.78
6.48
6.52
6.52
6.53
6.73
6.74
6.79
6.79
6.73
6.74
6.77
6.49
6.54
6.54
6.48
7.15
7.17
7.27
7.18
6.80
7.17
7.10
6.73
7.34
6.77
6.75
7.13
7.14
7.21
7.21
7.13
7.14
7.19
6.73
6.81
6.81
6.73
6.77
6.78
6.80
6.80
6.61
6.76
6.79
6.52
6.55
6.55
6.57
6.74
6.75
6.78
6.78
6.74
6.75
6.77
6.53
6.57
6.57
6.53
6.75
6.75
6.77
6.77
6.45
6.74
6.79
6.36
6.34
6.37
6.41
6.70
6.71
6.73
6.73
6.70
6.71
6.75
6.35
6.38
6.38
6.35
4
5a
6
7
8
9a
10
11
No. Compound log 1/ID50
log 1/ID50
log 1/ID50
Clog P CMR
(M19 MEL)
(MCF-7)
(WIDR)
(Eq. (5))
(Eq. (6))
(Eq. (7))
Obsd. Pred. Obsd. Pred. Obsd. Pred.
1
2
3
Ph3SnL1H
Ph3SnL2H
Ph3SnL3H
Ph3SnL4H
Ph3SnL5H
Ph3SnL6H
6.78
6.79
6.81
6.80
6.71
6.77
6.77
6.64
6.67
6.67
6.66
6.76
6.77
6.80
6.80
6.76
6.77
6.78
6.65
6.67
6.67
6.64
7.06
7.15
7.20
6.91
6.73
6.82
6.90
6.66
6.72
6.69
6.68
7.11
7.12
7.17
7.17
7.11
7.12
7.18
6.66
6.72
6.72
6.66
6.77
6.78
6.79
6.79
6.76
6.76
6.81
6.71
6.75
6.75
6.71
6.76
6.76
6.80
6.80
6.76
6.76
6.76
6.72
6.75
6.75
6.72
6.79
6.74
8.11
8.11
6.79
6.74
6.37
7.44
8.81
8.81
7.49
16.63
16.63
18.02
18.02
16.63
16.63
16.67
14.66
16.06
16.06
14.66
4b
5b
6b
7b,c Ph3SnL7H
8
9
10
11
Bu3SnL2H
Bu3SnL3H
Bu3SnL4H
Bu3SnL5H
a
Not included in the derivation of QSAR 2.
Not included in the derivation of QSAR 6.
Not included in the derivation of QSAR 7.
b
c
QSAR for the cytotoxicity of compounds 1–11 against WIDR human
tumor cell line (Table 6)
log1=ID50ðWIDRÞ ¼ 0:024ðꢀ0:006Þ CMR þ 6:371ðꢀ0:104Þ
n ¼ 10; r2 ¼ 0:903; s ¼ 0:009; q2 ¼ 0:856; Q ¼ 105:556; F1;8 ¼ 74:474ð5:318Þ
ð7Þ
outlier: compound 7
QSAR models 1–6 are well defined by two descriptors; Clog P and
CMR (Fig. 4 of QSAR 5 represents the best example). But for the QSAR
7, the cytotoxic activity is well correlated with only the CMR term and
there is no role of hydrophobic parameter. On comparing the statisti-
cal contribution of Clog P and CMR descriptors, it has been found that
CMR is a more important parameter than that of Clog P. CMR explains
65%, 77%, 63%, 54%, 72%, 72%, and 90%, respectively, of the variance in
the data of QSAR models 1–7, while Clog P accounts for only 18%, 19%,
21%, 27%, 17%, and 27%, respectively, of the variance in the data of
QSAR models 1–6. Thus CMR (steric/polarizability factor) plays a
major role for the cytotoxicity of organotin(IV) compounds 1–11
against a panel of seven human tumor cell lines studied.
Compounds were deemed to be outliers in QSARs (Eqs. (2), (6),
and (7)) on the basis of deviations in their activities (obsd –
pred>2×s). To assess the effects of excluding outlier(s), QSAR
models were examined thoroughly before and after the removal of
a compound. Although QSAR 6 is a very good model with respect to
their statistics, we could not consider it as a predictive model. This
is because the QSAR 6 was developed for a very small data set using
two descriptors and also after removing four outliers. A large number
of outliers for this QSAR model may suggest the combination of used
3.4.1. Validation of QSAR models
Two steps, statistical diagnostics and internal validation, were
used to validate QSAR models 1–7. In statistical diagnostics, QSARs
1–7 were filtered through the following conditions: (i) n/p≥4 (ex-
cept QSAR 6) (ii) r2 >0.6 (iii) q2 >0.5 (iv) r2 – q2 b0.3 (v) F>F(lit) at
95% level (vi) high Q value, and (vii) low s value [43,73-78]. On the
other hand, the internal validation was carried out by using cross-
validation (q2 >0.5) [74,77]. Since there is not sufficient difference
in activities of compounds, there is probability of a chance correlation
in the QSAR models. To examine this problem, we performed a Y-
randomization test. According to this test, if a strong correlation re-
mains between the selected descriptor(s) and the randomly permut-
ed response, the significance of the proposed QSAR is suspect
[44,73,77,78]. In the Y-randomization test, the poor values of r2 and
q2 in Table 8 ensure the robustness of the QSAR models 1–7 and
also the lack of over fitting. Due to the small data sets, the external

T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
127
irrespective of their nuclear substituents in the coupling moiety.
From QSAR models 1–7 it can be suggested that the cytotoxicity of
compounds 1–11 against the four different cancer cell lines (A498,
H226, IGROV, and M19 MEL) are due to a very similar mechanism,
but the same is not true against the EVSA-T, MCF-7, and WIDR cancer
cell lines. These three cancer cell lines may have some different mech-
anism in addition to the common one. The triphenyltin(IV) 2/4-[(E)-
2-(aryl)-1-diazenyl]benzoates (1–4) may therefore be explored fur-
ther by in vivo testing in animal models to develop them as anticancer
chemotherapeutics. Further work in this area is underway.
Acknowledgments
The financial support of the Department of Science and Technology,
New Delhi, India (Grant No. SR/S1/IC-03/2005,TSBB), the council of Scien-
tific and Industrial Research, New Delhi, India (Grant No. 09/347/(0197)/
2011/EMR I, for an award of a Senior Research Fellowship, AP), of the Uni-
versità degli Studi di Palermo, Italy (Grants ORPA079E5M and
ORPA0737W2) and the University Grants Commission, New Delhi, India
through SAP-DSA, Phase-III, are gratefully acknowledged. The in vitro cy-
totoxicity experiments were carried out by Ms. P. F. van Cuijk in the Lab-
oratory of Translational Pharmacology, Department of Medical Oncology,
Erasmus Medical Center, Rotterdam, The Netherlands, under the supervi-
sion of Dr. E. A. C. Wiemer and Prof. Dr. G. Stoter.
Fig. 5. Plot of observed log 1/ID50(A498) versus observed log 1/ID50(H226) (Table 7).
Table 7
Statistical data for the correlations among the cytotoxic activities of compounds 1–11
(n=11) against four cancer cell lines e.g. A498, H226, IGROV, and M19 MEL.
Correlations
r2
q2
s
Slope
log 1/ID50(A498) vs log 1/ID50(H226)
log 1/ID50(A498) vs log 1/ID50(IGROV)
log 1/ID50(A498) vs log 1/ID50(M19 MEL)
log 1/ID50(H226) vs log 1/ID50(IGROV)
log 1/ID50(H226) vs log 1/ID50(M19 MEL)
0.996 0.994 0.010
0.983 0.972 0.019
0.976 0.968 0.023
0.990 0.984 0.013
0.981 0.972 0.018
−1.04 ( 0.39)
2.14 ( 0.46)
−7.69 ( 1.70)
2.75 ( 0.30)
Appendix A. Supplementary material
Crystallographic data for complexes 1, 3 and 4 reported in this
paper have been deposited with the Cambridge Crystallographic
Data Centre (CCDC) as supplementary publication no. CCDC-
822371–CCDC-822373. Copies of the data can be obtained free of
charge from the CCDC (12 Union Road, Cambridge CB2 1EZ, UK;
Tel.: +44-1223-336408; Fax: +44-1223-336003; e-mail: depos-
it@ccdc.cam.ac.uk; Web site: http://www.ccdc.cam.ac.uk).
The following information (Figs. S1–S8) are available as supplemen-
tary materials. Fig S1: Diagram (PLATON) of a unit cell in compound 1.
The short intramolecular hydrogen bonds have been drawn as dashed
lines. H atoms bonded to C have been omitted; Fig. S2: Diagram (PLA-
TON) of a unit cell in compound 3. The short intramolecular hydrogen
bonds have been drawn as dashed lines. H atoms bonded to C have
been omitted; Fig. S3: Diagram (PLATON) of a unit cell in compound
4. The short intramolecular hydrogen bonds have been drawn as dashed
lines. H atoms bonded to C have been omitted; Fig. S4: Plot of observed
log 1/ID50(A498) versus observed log 1/ID50(IGROV) (Table 7); Fig. S5:
Plot of observed log 1/ID50(A498) versus observed log 1/ID50(M19
MEL) (Table 7); Fig. S6: Plot of observed log 1/ID50(H226) versus ob-
served log 1/ID50(IGROV) (Table 7); Fig. S7: Plot of observed log 1/
ID50(H226) versus observed log 1/ID50(M19 MEL) (Table 7); Fig. S8:
Plot of observed log 1/ID50(IGROV) versus observed log 1/ID50(M19
MEL) (Table 7).
−5.76 ( 1.32)
log 1/ID50(IGROV) vs log 1/ID50(M19 MEL) 0.948 0.924 0.048 −13.83 ( 3.59)
validation test was not considered. Although QSAR models 1–7 have
been passed through all the necessary validation tests, we did not
consider these models as the predictive QSAR models due to the pres-
ence of low activity range (b1 log unit) in the data sets. To construct
the predictive QSAR models, we need to synthesize additional orga-
notin(IV) compounds similar to compounds 1–11 with lower C log
P (C log Pb6.37) and higher/similar CMR values. It is interesting to
note that there is no mutual correlation between C log P and CMR
of compounds 1–11 (C log P vs CMR; r=0.025), providing us enough
space to manipulate the structure of organotin(IV) compounds 1–11
with respect to their varying C log P and CMR in order to obtain the
compound with desired activity.
4. Conclusions and outlook
We reported here the syntheses, characterization and cytotoxic
activities of some new triphenyltin(IV) 2/4-[(E)-2-(aryl)-1-diazenyl]
benzoates (1–4). Our studies have demonstrated the potential of de-
veloping compounds 1–4 as effective cytotoxic agents on all the cell
lines studied, particularly MCF-7 with an ID50 range of 41–53 ng/ml
in vitro. Complexes 1–4 appear to display similar cytotoxicity,
Supplementary data to this article can be found online at doi:10.
1016/j.jinorgbio.2011.10.008.
Table 8
Y-randomization test data for QSARs 1–7.
QSAR
no.
NOR-1a
NOR-2
NOR-3
NOR-4
NOR-5
r2
q2
r2
q2
r2
q2
r2
q2
r2
q2
1
2
3
4
5
6
7
0.320
0.271
0.366
0.385
0.372
0.391
0.242
−0.216
−0.452
−0.113
−0.064
−0.089
−0.347
−0.224
0.247
0.203
0.265
0.310
0.205
0.232
0.028
−0.314
−0.478
−0.311
−0.220
−0.423
−0.944
−0.407
0.562
0.648
0.591
0.576
0.642
0.694
0.311
0.270
0.203
0.315
0.306
0.406
−0.405
0.006
0.342
0.390
0.373
0.405
0.378
0.648
0.052
−0.183
−0.180
−0.155
−0.114
−0.115
−0.147
−0.438
0.244
0.115
0.266
0.328
0.221
0.245
0.197
−0.328
−1.370
−0.288
−0.180
−0.342
−0.886
−0.151
a
NOR=number of Y-randomization.

128
T.S. Basu Baul et al. / Journal of Inorganic Biochemistry 107 (2012) 119–128
[39] R.P. Verma, C. Hansch, Chem. Rev. 109 (2009) 213–235.
References
[40] R.P. Verma, C. Hansch, Chem. Rev. 111 (2011) 2865–2899.
[41] C. Selassie, R.P. Verma, History of quantitative structure-activity relationships, in:
D.J. Abraham, D.P. Rotella (Eds.), Burger's Medicinal Chemistry, 7th edition, Drug
Discovery and Development, vol. 1, John Wiley, New York, USA, 2010, pp. 1–95,
2010.
[42] R.D. Cramer III, J.D. Bunce, D.E. Patterson, I.E. Frank, Quant. Struct.-Act. Relat. 7
(1988) 18–25.
[43] C.A. Bennett, N.L. Franklin, Statistical Analysis in Chemistry and the Chemical Industry,
John Wiley, New York, 1967, pp. 708–709.
[44] C.D. Selassie, S. Kapur, R.P. Verma, M. Rosario, J. Med. Chem. 48 (2005)
7234–7242.
[45] SAINT + V. 6.45, Program for Reduction of Data Collected on Bruker CCD Area
Detector Diffractometer, Bruker, Madison, Wisconsin, USA, 2003.
[46] SADABS V. 2004/1, Program for the Absorption of Area Detector Data, Bruker,
Madison, Wisconsin, USA, 2004.
[47] G.M. Sheldrick, Acta Crystallogr. A64 (2008) 112–122.
[48] T.S. Basu Baul, S.M. Pyke, K.K. Sarma, E.R.T. Tiekink, Main Group Met. Chem. 19
(1996) 807–814.
[49] J. Holeček, M. Nádvorník, K. Handliř, A. Lyčka, J. Organomet. Chem. 241 (1983)
177–184.
[1] V.L. Narayanan, M. Nasr, K.D. Paul, in: M. Gielen (Ed.), Tin-based Antitumour
Drugs, Springer-Verlag, Berlin, 1990, pp. 201–217.
[2] A.J. Crowe, Metal Based Antitumour Drugs, in: M. Gielen (Ed.), vol. 1, Freund, London,
1988, pp. 103–149.
[3] M. Gielen, P. Lelieveld, D. de Vos, R. Willem, in: M. Gielen (Ed.), vol. 2, Freund, Tel
Aviv, 1992, pp 29–54.
[4] C. Pellerito, P. D'Agati, T. Fiore, C. Mansueto, V. Mansueto, G. Stocco, L. Nagy, L. Pellerito,
J. Inorg. Biochem. 99 (2005) 1294–1305.
[5] F. Cima, L. Ballarin, Appl. Organomet. Chem. 13 (1999) 697–703.
[6] M.N. Shuaibu, H. Kanbara, T. Yanagi, A. Ichinose, D.A. Ameh, J.J. Bonire, A.J. Nok,
Jpn. Parasitol. Res. 91 (2003) 5–11.
[7] C.-R. Jan, B.-P. Jiann, Y.-C. Lu, H.-T. Chang, W. Su, W.-C. Chen, C.-C. Yu, J.-K. Huang,
Life Sci. 70 (2002) 1337–1345.
[8] M.P. Samuel, D. de Vos, D. Raveendra, J.A.R.P. Sarma, S. Roy, Bioorg. Med. Chem.
Lett. 12 (2002) 61–64.
[9] C.E. Carraher Jr., A. Battin, K.R. Shahi, M.R. Roner, J. Inorg. Organomet. Polym. 17
(2007) 631–639.
[10] G. Barot, K.R. Shahi, M.R. Roner, C.E. Carraher Jr., J. Inorg. Organomet. Polym. 17
(2007) 595–603.
[11] M. Roner, C. Carraher, T. Sabir, K. Shahi, J. Roehr, K. Bassett, Polym. Mater. Sci. Eng.
95 (2006) 525–527.
[12] C. Carraher, Y. Ashida, G. Battin, Polym. Mater. Sci. Eng. 95 (2006) 556–558.
[13] C. Carraher, T. Sabir, M. Roner, K. Shahi, R. Bleicher, J. Roehr, K. Bassett, J. Inorg.
Organomet. Polym. 16 (2006) 249–257.
[14] M. Roner, C. Carraher, J. Roehr, K. Bassett, J. Polym. Mater. 23 (2006) 153–159.
[15] C. Carraher, D. Siegmann-Louda, Macromolecules containing metal and metal-like
elements, Biomedical Applications, vol. 3, Wiley, Hoboken, NJ, 2004.
[16] R. Doucette, D. Siegmann-Louda, C. Carraher, A. Cardoso, Polym. Mater. Sci. Eng.
91 (2004) 564–566.
[17] P.J. Blower, Annu. Rep. Prog. Chem. A 100 (2004) 633–658.
[18] M. Gielen, R. Willem, M. Biesemans, M. Bouălam, A. El Khloufi, D. de Vos, Appl.
Organomet. Chem. 6 (1992) 287–291.
[19] M. Kemmer, M. Gielen, M. Biesemans, D. de Vos, R. Willem, Metal-Based Drugs 5
(1998) 189–196.
[20] M. Kemmer, L. Ghys, M. Gielen, M. Biesemans, E.R.T. Tiekink, R. Willem, J. Organomet.
Chem. 582 (1999) 195–203.
[21] M. Gielen, P. Lelieveld, D. de Vos, H. Pan, R. Willem, M. Biesemans, H.H. Fiebig,
Inorg. Chim. Acta 196 (1992) 115–117.
[50] T.S. Basu Baul, S. Dhar, S.M. Pyke, E.R.T. Tiekink, E. Rivarola, R. Butcher, F.E. Smith,
J. Organomet. Chem. 633 (2001) 7–17.
[51] G.M. Bancroft, R.H. Platt, Adv. Inorg. Chem. Rad. 15 (1972) 59–258.
[52] R.L. Collins, J.C. Travis, in: I.J. Gruverman (Ed.), The Electric Field Gradient
Tensor: Mössbauer Effect Methodology, vol. 3, Plenum Press, New York, 1967,
pp. 123–161.
[53] G.M. Bancroft, V.G. Kumar Das, T.K. Sham, M.G. Clark, J. Chem. Soc. Dalton Trans.
(1976) 643–654.
[54] T.K. Sham, G.M. Bancroft, Inorg. Chem. 14 (1975) 2281–2283.
[55] M.G. Clark, A.G. Maddock, R.H. Platt, J. Chem. Soc. Dalton Trans. (1972) 281–290.
[56] R.V. Parish, in: G.J. Long (Ed.), Structure and Bonding in Tin Compound: Mössbauer
Spectroscopy Applied to Inorganic Chemistry, vol. 1, Plenum Press, New York,
1984, pp. 544–548.
[57] P.J. Harrison, K. Lambert, T.J. King, B. Majee, J. Chem. Soc. Dalton Trans. (1983)
363–369.
[58] T.S. Basu Baul, E.R.T. Tiekink, Z. Kristallogr. (NCS) 211 (1996) 489–490.
[59] T.S. Basu Baul, S. Dhar, N. Kharbani, S.M. Pyke, R. Butcher, F.E. Smith, Main Group
Met. Chem. 22 (1999) 413–421.
[60] T.S. Basu Baul, W. Rynjah, E. Rivarola, A. Linden, J. Organomet. Chem. 690 (2005)
613–621.
[22] M. Bouălam, M. Gielen, A. El Khloufi, D. de Vos, R. Willem, Eur. Pat., Publ. 538 517,
Appl 91/202, 746.3-, 22.10.91; Chem. Abstr. 119 (1993) 117548b.
[23] M. Gielen, E.R.T. Tiekink (Eds.), Metallotherapeutic Drug and Metal-based Diagnostic
Agents: 50Sn Tin Compounds and their Therapeutic Potential, Wiley, Chichester,
England, 2005, pp. 421–439, (and references therein).
[61] T.S. Basu Baul, A. Paul, L. Pellerito, M. Scopelliti, C. Pellerito, P. Singh, P. Verma, A.
Duthie, D. de Vos, R.P. Verma, U. Englert, J. Inorg. Biochem. 104 (2010) 950–966.
[62] M.C. Etter, J.C. MacDonald, J. Bernstein, Acta Crystallogr. B46 (1990) 256–262.
[63] C.J. Tranter, S.B. Price, J. Cutts, P.G. Parsons, G. Rintoul, D.J. Young, Main Group
Chem. 1 (1995) 165–167.
[64] S.C. Ng, P.G. Parsons, K.Y. Sim, C.J. Tranter, R.H. White, Appl. Organometal. Chem.
11 (1997) 577–578.
[65] M. Gielen, E.R.T. Tiekink, Tin compounds and their therapeutic potential, in: M.
Gielen, E.R.T. Tiekink (Eds.), Metallotherapeutic Drugs and Metal-Based Diag-
nostic Agents: The Use of Metals in Medicine, John Wiley & Sons, London, 2005,
pp. 421–439.
[66] R.A. LeBoeuf, G.A. Kerkaert, M.J. Aardema, D.P. Gibson, R. Brauninger, R.J. Isfort,
Mut. Res. 356 (1996) 85–127.
[67] R.J. Isfort, G.A. Kerckaert, R.A. LeBoeuf, Mut. Res. 356 (1996) 11–63.
[68] A. Meriem, R. Willem, M. Biesemans, B. Mahieu, D. de Vos, P. Lelieveld, M. Gielen,
Appl. Organometal. Chem. 5 (1991) 195–201.
[69] M. Gielen, P. Lelieveld, D. de Vos, R. Willem, Metal-Based Anti-Tumour Drugs 2
(1992) 29–54.
[70] H. Dalil, M. Biesemans, R. Willem, M. Gielen, Main Group Met. Chem. 21 (1998)
741–745.
[24] M. Gielen, R. Willem, H. Dalil, D. de Vos, C.M. Kuiper, G.J. Peters, Metal-Based
Drugs 5 (1998) 83–90.
[25] T.S. Basu Baul, C. Masharing, G. Ruisi, R. Jirásko, M. Holčapek, D. de Vos, D.
Wolstenholme, A. Linden, J. Organomet. Chem. 692 (2007) 4849–4862.
[26] T.S. Basu Baul, S. Basu, D. de Vos, A. Linden, Invest. New Drugs 27 (2009) 419–431.
[27] T.S. Basu Baul, A. Paul, L. Pellerito, M. Scopelliti, P. Singh, P. Verma, D. de Vos, Invest.
New Drugs 28 (2010) 587–599.
[28] P. Singh, A. Bhardwaj, Mini Rev. Med. Chem. 8 (2008) 388–398.
[29] R. Willem, I. Verbruggen, M. Gielen, M. Bieseman, B. Mahieu, T.S. Basu Baul, E.R.T.
Tiekink, Organometallics 17 (1998) 5758–5766.
[30] B. Kushlefsky, I. Simmons, A. Ross, Inorg. Chem. 2 (1963) 187–189.
[31] T.S. Basu Baul, A. Paul, L. Pellerito, M. Scopelliti, P. Singh, P. Verma, A. Duthie, D. de
Vos, E.R.T. Tiekink, Invest. New Drugs 29 (2011) 285–299.
[32] T.S. Basu Baul, A. Paul, E.R.T. Tiekink, Acta Crystallogr. E66 (2010) o540.
[33] T.S. Basu Baul, A. Paul, H.D. Arman, E.R.T. Tiekink, Acta Crystallogr. E64 (2008)
o2125.
[34] M.R. Boyd, Status of the NCI preclinical antitumor drug discovery screen, in: V.T.
DeVita Jr., S. Hellman, S.A. Rosenberg (Eds.), Cancer: Principles and Practice of
Oncology Updates, vol. 3, J.P. Lippincott, Philadelphia, 1989, pp. 1–12.
[35] Y.P. Keepers, P.R. Pizao, G.J. Peters, J.V. Ark-Otte, B. Winograd, H.M. Pinedo, Eur. J.
Cancer 27 (1991) 897–900.
[71] S.W. Ng, J.M. Hook, M. Gielen, Appl. Organometal. Chem. 14 (2000) 1–7.
[72] C. Camacho-Camacho, D. de Vos, B. Mahieu, M. Gielen, M. Kemmer, M. Biesemans,
R. Willem, Main Group Met. Chem. 23 (2000) 433–438.
[73] C. Hansch, R.P. Verma, ChemMedChem 2 (2007) 1807–1813.
[74] A. Golbraikh, A. Tropsha, J. Mol. Graph Modl. 20 (2002) 269–276.
[75] L. Eriksson, J. Jaworska, A.P. Worth, M.T.D. Cronin, R.M. McDowell, P. Gramatica,
Environ. Health Prospect. 111 (2003) 1361–1375.
[36] C-QSAR Program, BioByte Corp., 201 W. 4th St., Suite 204, Claremont, CA 91711,
USA, www.biobyte.com.
[37] C. Hansch, D. Hoekman, A. Leo, D. Weininger, C.D. Selassie, Chem. Rev. 102 (2002)
783–812.
[76] L. Pogliani, Chem. Rev. 100 (2000) 3827–3858.
[77] A. Tropsha, P. Gramatica, V.K. Gombar, QSAR Comb. Sci. 22 (2003) 69–77.
[78] S. Wold, L. Eriksson, Statistical validation of QSAR results, in: H. van de Waterbeemd
(Ed.), Validation Tools, Chemometrics Methods in Molecular Design, Methods
and Principles in Medicinal Chemistry, vol. 2, VCH, Weinheim, Germany, 1995,
pp. 309–318.
[38] R.P. Verma, C. Hansch, Development of QSAR models using C-QSAR program: a re-
gression program that has dual database of over 21,000 QSAR models freely avail-
able at: Nat. Protoc. (2007) (Nature Publishing Group), Freely available at: http://
www.nature.com/protocolexchange/protocols/181, doi:10.1038/nprot.2007.125.