Pharmacomodulation on Gold-NHC complexes for anticancer applications – is lipophilicity the key point?

Article abstract of DOI:10.1016/j.ejmech.2018.07.070

A series of four new mononuclear cationic gold(I) complexes containing nitrogen functionalized N-heterocyclic carbenes (NHCs) was synthesized and fully characterized by spectroscopic methods. The X-ray structures of three complexes are presented. These lipophilic gold(I) complexes originate from a pharmacomodulation of previously described gold(I)-NHC complexes, by replacing an aliphatic spacer with an aromatic one. The Log P values of the resulting complexes increased by 0.7–1.5, depending on the substituents in comparison to the aliphatic-linker systems. The new series of complexes has been investigated in vitro for their anti-cancer activities in PC-3 (prostate cancer) and T24 (bladder cancer) cell lines and in the non-cancerous MC3T3 (osteoblast) cell line. All tested complexes show high activities against the cancer cell lines with GI₅₀ values lower than 500 nM. One complex (11) has been selected for further investigations. It has been tested in vitro in six cancer cell lines from different origins (prostate, bladder, lung, bone, liver and breast) and two non-cancerous cell lines (osteoblasts, fibroblasts). Moreover, cellular uptake measurements were indicative of a good bioavailability. By various biochemical assays, this complex was found to effectively inhibit the thioredoxin reductase (TrxR) and its cytotoxicity towards prostate PC-3, bladder T24 and liver HepG2 cells was found to be ROS-dependent.

Full text of DOI:10.1016/j.ejmech.2018.07.070

Accepted Manuscript
Pharmacomodulation on Gold-NHC complexes for anticancer applications – is
lipophilicity the key point?
Chen Zhang, Marie-Lise Maddelein, Raymond Wai-Yin Sun, Heinz Gornitzka, Olivier
Cuvillier, Catherine Hemmert
PII:
S0223-5234(18)30635-4
10.1016/j.ejmech.2018.07.070
EJMECH 10605
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 11 June 2018
Revised Date: 24 July 2018
Accepted Date: 29 July 2018
Please cite this article as: C. Zhang, M.-L. Maddelein, R. Wai-Yin Sun, H. Gornitzka, O. Cuvillier, C.
Hemmert, Pharmacomodulation on Gold-NHC complexes for anticancer applications – is lipophilicity the
key point?, European Journal of Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.07.070.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical abstract

ACCEPTED MANUSCRIPT
Pharmacomodulation on Gold-NHC Complexes for Anticancer Applications
–
is Lipophilicity the Key Point?
Dedicated to Professor Dr. Dietmar Stalke on the occasion of his 60th birthday.
Chen Zhang,a Marie-Lise Maddelein,b Raymond Wai-Yin Sun,c Heinz Gornitzka,a* Olivier
Cuvillier,b* Catherine Hemmerta*
a LCC-CNRS, Université de Toulouse, CNRS, UPS, Toulouse, France;
b
Institut de Pharmacologie et de Biologie Structurale, Université de Toulouse, CNRS, UPS,
Toulouse, France;
c Department of Chemistry, The University of Hong Kong, Pokfulam, Hong Kong (P. R.
China).
E-mail addresses: gornitzka@lcc.toulouse.fr, olivier.cuvillier@inserm.fr,
hemmert@lcc.toulouse.fr.
_______________________
1

ACCEPTED MANUSCRIPT
Abstract
A series of four new mononuclear cationic gold(I) complexes containing nitrogen
functionalized N-heterocyclic carbenes (NHCs) was synthesized and fully characterized by
spectroscopic methods. The X-ray structures of three complexes are presented. These
lipophilic gold(I) complexes originate from a pharmacomodulation of previously described
gold(I)-NHC complexes, by replacing an aliphatic spacer with an aromatic one. The Log P
values of the resulting complexes increased by 0.7 to 1.5, depending on the substituents in
comparison to the aliphatic-linker systems. The new series of complexes has been
investigated in vitro for their anti-cancer activities in PC-3 (prostate cancer) and T24 (bladder
cancer) cell lines and in the non-cancerous MC3T3 (osteoblast) cell line. All tested complexes
show high activities against the cancer cell lines with GI50 values lower than 500 nM. One
complex (11) has been selected for further investigations. It has been tested in vitro in six
cancer cell lines from different origins (prostate, bladder, lung, bone, liver and breast) and two
non-cancerous cell lines (osteoblasts, fibroblasts). Moreover, cellular uptake measurements
were indicative of a good bioavailability. By various biochemical assays, this complex was
found to effectively inhibit the thioredoxin reductase (TrxR) and its cytotoxicity towards
prostate PC-3, bladder T24 and liver HepG2 cells was found to be ROS-dependent.
Keywords: gold, N-heterocyclic carbene, cancer, lipophilicity, thioredoxin reductase,
glutathione, ROS.
1.
Introduction
The anticancer efficacy of cisplatin has stimulated the scientific community to focus on the
design of other metal complexes for this therapeutic application [1-7]. With respect to gold(I)
complexes, the perturbation of the redox signaling and other apoptosis-related pathways have
been identified as origins of their cytotoxic activities [8-10]. The design of ligands for the
complexation of gold(I) is guided by conditions for such application: an ideal complex must
be stable enough under physiological condition to reach its targets, lipophilic enough to pass
cell membranes and selective enough to discriminate between cancer and non-cancer cells. A
simple adaptable type of ligands to optimize the systems for these conditions is a key point for
2

ACCEPTED MANUSCRIPT
the design of such complexes. N-heterocyclic carbenes (NHCs) perfectly fit into these
specifications.
There is a particular growing craze for the development of gold NHC complexes as
anticancer or theranostic agents thanks to: 1) the class of easily derivatizable NHC ligands
that are excellent σ-donors, giving stable gold NHC complexes insensitive to biologically
important thiol groups and 2) the pioneering work of Berners-Price and coll. in 2004 [11] on
dinuclear cationic lipophilic AuI-NHC complexes able to induce mitochondrial membrane
permeabilization in isolated rat liver mitochondria. This active research area in bio-inorganic
chemistry includes gold(I) and gold(III), dinuclear, mononuclear, cationic, neutral and
heterobimetallic complexes, and has given rise to a large number of publications during the
last decade [12-24], as illustrated by two reviews published in 2018 [22,23].
According to the Web of Science and Scifinder databases, among about 60 scientific
papers published since 2012 on the in vitro anticancer activities of mononuclear neutral or
cationic NHC gold(I) complexes, 20 of them possess one or several complexes showing
striking cytotoxic activities with IC50 or GI50 ≤ 1 µM towards one or more cancer cell lines
[25-44]. However, only half of these publications mention toxicity of the tested complexes,
which should be first evaluated in vitro, as selectivity is a crucial parameter for subsequent in
vivo studies, even if in vitro results do not always predict in vivo efficacy on murine models.
Among them, several cationic biscarbene Au(I) complexes with relevant cytotoxicity on
human cancer cells (determined by MTT assay) were evaluated in vitro for their selectivity on
non-cancer cells. Complexes [AuI(MeBIm)2][BF4] (BIm: benzimidazol-2-ylidene; IC50 on
ovarian A2780 and SKOV3 cancer cells of 0.54 and 0.75 µM, IC50 on non-cancerous human
embryonic kidney HEK-293T cells of 0.20 µM) [38] and bis(4-arylimidazolylidene) Au(I)
iodide (IC50 on colon HT-29, breast MCF-7 and MDA-MB-231 cancer cells from 0.06 to 0.18
µM, IC50 on normal epithelial kidney RC-124 cells of 0.05 and 0.09 µM) [30] were found to
be toxic. Complexes bis(diethylaminoethylimidazolylidene) Au(I) nitrate (IC50 on colon
prostate PC-3, breast MCF-7, brain U87, lung A549 and liver Hep3B cancer cells from 0.4 to
1 µM, IC50 on human umbilical vein endothelial HUVECs cells > 5µM) [43] and Au(I)
bisNHC containing quinoline or methylbipyridine functionalized groups (IC50 on ovarian
A2780 and cisplatin resistant A2780cis, and liver HepG2 and HepAD38 cancer cells from
0.11 to 0.78 µM, IC50 on Madin-Darby canine kidney epithelial (MDCK) from 2.8 to 3.0 µM)
[25] show selectivity with restricted effects on normal cells. Finally, complex
[AuI(EtBIm)2][I] (IC50 on pancreatic Bxpc3, Miapaca2, Panc1 and ASPC1 cancer cells from
3

ACCEPTED MANUSCRIPT
0.06 to 0.28 µM, IC50 on human neonatal foreskin fibroblasts HFF cells > 20 µM) [37]
display a high selectivity towards cancer cells. In other reports, the toxicity was evaluated in
vitro on peripheral blood mononuclear cells [33], ex vivo in healthy rat kidney tissues [26], or
in vivo on zebrafish lung xenograft model [27] or mouse melanoma or renal xenograft models
[42,29].
It is difficult to highlight a lead gold(I) complex based on structure-activity relationship
(SAR) despite numerous structural modifications on the carrier NHCs, i.e variation of 1,3-N-
substituents on the imidazole core, modification of the skeleton of the azolium ring,
introduction of biological entities or organometallic moieties. However, it is well-established
that delocalized lipophilic cations (DLCs) often show strong antiproliferative potency and are
able to selectively accumulate in tumor cell mitochondria. Due to the electric gradient
between membranes' inner and outer layers, DLCs can penetrate the hydrophobic barrier of
the cellular membranes and accumulate in mitochondria. Furthermore, a difference of around
60 mV between the mitochondrial membrane potentials of cancerous and healthy cells induce
a selective accumulation in tumor cell mitochondria [45,46]. We focused our research on
bioactive gold(I) NHCs complexes for parasitic diseases [47-50] and anticancer applications
[25,43,51]. In this context, we have recently demonstrated that some mononuclear lipophilic
cationic Au(I) bis(NHC) complexes containing aliphatic or aromatic functionalized arms
display in vitro high and selective cytotoxicity towards several human cancer cell models.
One class of our ligands was characterized by a nitrogen-containing group connected by an
aliphatic linker to the NHC skeleton [43]. Herein we present the synthesis and biological
studies of an optimized family of complexes. Our pharmaco-modulation consisted in the
replacement of the aliphatic linker by an aromatic one in order to increase the lipophilicity of
our complexes. The cytotoxicity and selectivity of the complexes have been studied on PC-3
and T24 cancer cell lines and on MC3T3 non-cancerous cell line. Moreover, one complex has
been selected for cytotoxicity tests on six cancer and two non-cancerous cell lines and for
clonogenic assays. Cellular uptake, TrxR inhibition and ROS formation have been evaluated
for the same complex.
2.
Results and discussion
2.1. Chemistry
4

ACCEPTED MANUSCRIPT
The substituted imidazoles 1-4 were readily prepared by a copper-catalyzed N-arylation of
imidazole with 1) 4-bromo-N,N-dimethylaniline or 4-bromo-N,N-diethylaniline in the
presence of K2CO3 and CuSO4 at 205 °C for 10 h in the case of 1 and 2 respectively, and 2)
with 1-(4-chlorophenyl)-1H-pyrrole or 4-(4-bromophenyl)morpholine in the presence of
K3PO4 and CuI at 160 °C for 3 days in the case of 3 and 4 respectively. The four
corresponding proligands 5-8 were obtained by reacting 1-arylimidazoles with benzyl
chloride in dry acetonitrile at 80 °C for 10 h and were isolated after purification by column
chromatography with yields spanning from 76 to 95%. They were classically characterized by
13
1H and C NMR spectroscopy, mass spectrometry and elemental analysis. The most notable
features in the 1H and 13C spectra of the imidazolium salts are the resonances for imidazolium
protons (H2) located at 8.33-10.36 ppm, the upfield value being attributed to the proligand 6
and the corresponding imidazolium carbons (C2) in the range of 135.2-136.1 ppm.
The target gold(I) complexes 9-12 were synthetized according the transmetalation
pathway. Firstly, the silver precursor complexes were prepared by deprotonation of the
imidazolium salts 5-8, with one equivalent of the mild base Ag2O in dry methanol at room
temperature. An ion exchange with AgNO3 dissolved in acetonitrile was performed, in order
to avoid the formation of insoluble compounds or ionic species of type [Ag(NHC)2][AgBr2].
The carbene transfer reaction was thus carried out in situ, by adding one half equivalent of
Au(SMe2)Cl, with respect to the ligand. Gold(I) complexes 9-12 were isolated after
purification by column chromatography as yellow powders with relatively good yields (40-
62%); all the synthesized complexes are highly stable towards air and moisture and even in
solution. NMR spectroscopy unequivocally demonstrates the formation of the gold(I)
13
complexes with the absence of the proton resonance of the acidic imidazolium and the C
NMR spectra show the resonance for the carbene carbon atoms at 182.7 to 182.8 ppm.
Protons 4 and 5 of the imidazole rings in the complexes were shifted upfield when compared
to their corresponding proligands. The elemental analysis of the gold(I) complexes correspond
to the general [AuL2][NO3] formula for 9-12, attesting of their purity and the HRMS exhibit
- +
the classical peak corresponding to the cationic fragment [M – NO3 ] for 9-12.
5

ACCEPTED MANUSCRIPT
Scheme 1. Synthesis of imidazolium salts 5-8 and gold(I) complexes 9-12.
2.2. Crystal structures of complexes 9-11
Crystals of 9 have been obtained by gas phase diffusion of diethylether in an acetonitrile
solution of 9.
Figure 1: Molecular structure of 9 in the solid state. The cationic part is depicted as a 50%
level thermal ellipsoid plot. H-atoms and non-coordinating anion are omitted for clarity.
Selected bond distances [Å] and angles [°]: Au1-C1 2.012(7), Au1-C19 2.013(8), C1-Au1-
C19 177.4(4).
The coordination sphere of the gold(I) cation (see Figure 1) shows a classical linear geometry
and the two NHC ring systems are slightly twisted with a torsion angle of around 8°. The two
6

ACCEPTED MANUSCRIPT
benzyl groups are in trans position in relation to the C-Au-C axis and to the NHC-Au-NHC
plane. It is remarkable that the geometry around nitrogen atom N3 is not planar as reflected in
the total sum of the C-N-C angles, which is 350.45°. The crystal packing of 9 presented in
-
Figure 2 shows the arrangement of the NO3 anions in channels in direction b.
Figure 2: Crystal packing of 9 with view along b.
Crystals of 10 have been obtained by slow evaporation of a methanol solution of 10.
Figure 3: Molecular structure of 10 in the solid state. The cationic part is depicted as a 50%
level thermal ellipsoid plot. For clarity only one of two independent cations is shown and H-
atoms, non-coordinating anions and non-coordinating MeOH are omitted. Selected bond
distances [Å] and angles [°]: Au1-C1 2.019(3), Au1-C21 2.024(3), C1-Au1-C21 176.92(9).
In the case of 10 the asymmetric unit contains two cations, two anions and one molecule of
MeOH. As visualized in Figure 3, in this complex, the benzyl groups are in trans position
7

ACCEPTED MANUSCRIPT
relative to the C-Au-C axis, but they are on the same side of the NHC-Au-NHC plane.
Moreover, the two NHC planes of a cation are coplanar (Figure 4).
Figure 4: Dimeric arrangement of 10 with view along the C-Au-C axe on the left side and
perpendicular to this on the right side.
In the solid state, 10 forms dimers with Au-Au distances of 3.45 Å (Au1-Au1) and 3.64 Å
(Au2-Au2), respectively. One is in the range for aurophilic interactions, the other one not.
Also in this case, the anions are located in channels following direction b.
Figure 5: Crystal packing of 10 with view along b.
Crystals of 11 have been obtained by slow evaporation of an acetonitrile solution of 11.
8

ACCEPTED MANUSCRIPT
Figure 6: Molecular structure of 11 in the solid state. The cationic part is depicted as a 30%
level thermal ellipsoid plot. For clarity H-atoms, non-coordinating anions and disorders of one
benzyl group are omitted. Selected bond distances [Å] and angles [°]: Au1-C1 2.024(16),
Au1-C21 2.018(14), C1-Au1-C21 171.9(6), Au1-Au1 3.309(2).
In 11, the groups show the same arrangement as in 10, which means that the benzyl groups
are in trans position relative to the C-Au-C axis and on the same side relative to the best plane
formed by the NHC ring systems and the gold cation. In this case, a deformation of the linear
geometry of the gold cation is reflected by the C-Au-C angle of 171.9°, this has also an effect
on the co-planarity of the NHC rings, which are twisted by about 10°. This deformation can
be explained by an aurophilic interaction (see Figure 6) with an Au-Au distance of 3.309 Å.
Figure 7: Dimeric arrangement of 11 in the solid state. Left side shows the view
perpendicular to the Au-Au bond and the on the right the view along the Au-Au bond is
depicted.
The dimer shows a rotation of 36.2° of the two cations around the Au-Au bond showing a
slightly crossed configuration of the two cations relatively to each other (see Figure 7).
Moreover, the anios are not located in channels or layers (see Figure 8).
9

ACCEPTED MANUSCRIPT
Figure 8: Packing of 11 with view along a.
A common feature of the structures presented here is the trans position of the substituents
relative to the NHC-Au-NHC unit in the solid state. This conformation is probably not
maintained in solution where almost free rotation of the NHC ligands around the C-Au-C axis
is allowed.
2.3. Antiproliferative activity
We first evaluated the in vitro cytotoxicity of the cationic gold(I) bis(NHC) complexes 9 to
12 and proligands 7 and 8 on the representative PC-3 prostate and T24 bladder cancer cell
lines using the colorimetric MTT assay. The results are summarized in Table 1. As shown in
Figures 9 and 10, the four complexes 9-12 induced a pronounced loss of PC-3 and T24 cancer
cell viability with GI50 values ranging from 30 to 250 nM. The two complexes 9 and 10
exhibited the strongest effects with GI50 values of 50 and 30 nM on PC-3 cells, respectively,
and 60 nM for both complexes on T24 cells. Interestingly, among all tested compounds, the
complex 11 was the only one clearly displaying a dose-dependent effect (see Supporting
Information Figure S1). As predicted, the ligand precursors 7 and 8 showed no cytotoxic
effects (GI50 > 5 µM on the two tested cancer cell lines). We used the anti-arthritic drug
auranofin as a reference for gold complexes given its potential repurposing in cancer as a pro-
oxidant therapy with ongoing clinical trials in leukemia and lung cancer alone or in a
combined modality. Auranofin clearly exhibited a weaker efficacy profile in our prostate and
bladder cell models (GI50 = 1.05 and 0.84 µM on PC-3 and T24, respectively) as compared to
complexes 10-14 (Figures 9 and 10). The MC3T3 osteoblast precursor cell line derived from
mouse calveria was used as non-cancerous cell model. Whereas complexes 9, 10 and 12
10

ACCEPTED MANUSCRIPT
showed low selectivity with GI50 comprised between 110 and 350 nM against osteoblasts and
auronofin was not selective, MC3T3 cells showed a marked insensitivity to complex 11 with a
GI50 value of 900 nM and a selectivity index of 3.6 (see Table 1 and Figure 11). We have
determined the lipophilicity of the gold complexes 9-12 using a shake-flask method. [52] The
corresponding Log P values are 1.256 ± 0.047 for 9, 0.926 ± 0.030 for 10, 1.033 ± 0.060 for
11 and 0.501 ± 0.024 for 12. These results reflect that for the complexes with high
lipophilicity in this study, no linear correlation can be observed between the lipophilicity and
the cytotoxicity of these complexes.
PC-3
(prostate cancer)
150
9
10
11
100
12
7
8
50
0
auranofin
Concentration (µM)
Figure 9: Cell viability of PC-3 prostate cancer cells treated for 72 h with increasing
concentrations of gold complexes 9 to 12, proligands 7 and 8 and auranofin.
T24
(bladder cancer)
150
9
10
11
100
12
7
8
50
0
auranofin
Concentration (µM)
Figure 10: Cell viability of T24 bladder cancer cells treated for 72 h with increasing
concentrations of gold complexes 9 to 12, proligands 7 and 8 and auranofin.
11

ACCEPTED MANUSCRIPT
MC3T3
(osteoblast)
150
100
50
9
10
11
12
auranofin
0
Concentration (µM)
Figure 11: Cell viability of MC3T3 osteoblastic cancer cells treated for 72 h with increasing
concentrations of gold complexes 9 to 12 and auranofin.
Table 1. GI50 values (growth inhibitory activity) of proligands 7 and 8 and complexes 9-12
after 72 h of treatment.
Compound
GI50(nM)a
T24
PC-3
MC3T3
> 5
> 5
> 5
> 5
-
7
-
8
0.05
0.03
0.25
0.16
1.05
0.06
0.06
0.25
0.14
0.84
0.12
0.11
0.90
0.35
1.05
9
10
11
12
Auranofin
aThe GI50 values represent the concentration of
compound causing 50% inhibition of cell growth.
Mean of three independent experiments.
Regarding SAR, we previously reported that lipophilic mononuclear cationic gold(I)
bis(NHC) complexes containing amino groups on the NHC scaffold were highly cytotoxic
and selective (see Table 2). For example, a cationic gold(I) bis(NHC) complex (A) involving
a 1-[2-(diethylamino)ethyl] group on one side and a mesitylene group on the other side of the
NHC ligands, with a Log P = 1.13, showed GI50 values of 400 nM on PC-3 prostate cancer
cells and >5 µM on normal primary human umbilical vein endothelial cells (HUVECs) [43].
The corresponding benzyl containing complex (B) had a Log P of 0.3 and showed GI50 values
12

ACCEPTED MANUSCRIPT
of 1440 nM on PC-3 prostate cancer cells. By replacing the aliphatic linker of this complex by
an aromatic one (10), the Log P value increased to 0.9 and the GI50 value on PC-3 prostate
cancer cells decreased to 30 nM (Table 2). In the case of 11, the log P was about 1 and the
GI50 value on HepG2 cells was 640 nM, while the corresponding aliphatic linked pyrrole
system (C) [25] showed a log P of 0.3 and a GI50 value on HepG2 cells of 3100 nM (Table 2).
Taken together, these outcomes highlight that an increased lipophilicity of the complexes
improves cytotoxicity, but sometimes leads to the detriment of the selectivity towards the
cancer cells.
Table 2. Comparative antiproliferative activity versus lipophilicity of gold(I) bisNHC
complexes.
Complex N-functionalized Non functionalized Cell line
GI50
(nM)
Log P
1.13
group
group
ARef. 43
BRef.43
Mes
PC-3
PC-3
400
Bn
1440
0.26
Bn
Bn
Bn
PC-3
30
0.93
0.29
1.03
10
CRef. 25
11
HepG2
HepG2
3100
640
Accordingly, complex 11 was selected as the most effective and selective compound for
further in vitro screening on a representative panel of hepatocarcinoma, osteosarcoma and
breast cancer in addition to prostate and bladder cancer cell models. Two non-cancerous cell
lines (MC3T3 osteoblasts and NIH3T3 fibroblasts) were used to evaluate its selectivity. Cells
were exposed to increasing concentrations of complex 11 (0.01 to 5 µM) for 72 h (Figure 12).
The resulting GI50 (growth inhibitory activity) values, corresponding to the concentration of
compound causing 50% decrease in cell growth, were in the range of 120 to 250 nM for all
studied cancer cell lines with the exception of HepG2 cells (GI50 of 640 nM) (Table 3). It
should be pointed out that hepatocarcinoma is notoriously refractory to chemotherapy and
sorafenib, the standard of care, displays a high IC50 value of 6.4 µM on the HepG2 cell line
13

ACCEPTED MANUSCRIPT
[53]. In contrast, the efficacy of complex 11 towards MC3T3 osteoblasts and NIH3T3
fibroblasts was low with GI50 values of 900 nM and 1.3 µM, respectively. These results
suggest that complex 11 is highly active against multiple cancer cell lineages and is selective
with limited effects on non-cancerous cells. Moreover, the kinetic effects of gold complex 11
on PC-3 and T24 cancer cell lines showed in nearly all cases a linear effect with time (see
Supporting Information Figure S1).
Concentration of 11 (µM)
150
0.1
0.25
0.5
100
1
5
50
0
Tumor cell line
Non tumor cell line
Figure 12: Cell viability of PC-3, T24, HepG2, A549, U-2 OS, MCF-7, MC3T3 and NIH3T3
treated for 72 h with increasing concentrations of gold complex 11.
Table 3. GI50 values (growth inhibitory activity) of complex 11 after 72 h of treatment.
a
Origin
Cell line
PC-3
GI50
Prostate cancer
Bladder cancer
Liver cancer
Lung cancer
Osteosarcoma
Breast cancer
Osteoblast
250 nM
240 nM
640 nM
230 nM
240 nM
120 nM
900 nM
T24
HepG2
A549
U-2 OS
MCF-7
MC3T3
Fibroblast
NIH3T3 1300 nM
aThe GI50 values represent the concentration of
complex 11 causing 50% inhibition of cell growth.
Mean of three independent experiments.
14

ACCEPTED MANUSCRIPT
Clonogenic assay or colony formation assay is an in vitro cell survival assay based on the
ability of a single cell to form a colony (>50 cells) [54]. It is the method of choice to
determine the effectiveness of radiotherapy or cytotoxic agents. As shown in Figure 13, no
visible colony of PC-3 or T24 cells was observed at concentrations of 0.25 and 0.5 µM, when
compared to controls (Figure 13). This assay demonstrates the ability of complex 11 to
completely abolish the capacity of prostate and bladder cells to proliferate at low
concentrations.
Figure 13. Colony formation in PC-3 (top) and T24 (bottom) cells treated with increasing
concentrations of complex 11 for 24 h and stained with crystal violet after 3 weeks of
cultivation.
2.4. Cellular uptake
As complex 11 belongs to the class of DLCs, it is expected to easily penetrate the cell
membrane and efficiently accumulate in mitochondria of the tumor cells. Indeed, carcinoma
cells have an elevated mitochondrial membrane potential (∆Ψm) compared to normal cells so
it is possible to exploit this phenomenon as driving force to deliver DLCs just inside the
cancer cell [45]. Relationship between high cytotoxicity and intracellular accumulation has
been reported by previous studies. We therefore evaluated the cellular uptake of complex 11
in PC-3 cells using a method based on gold quantification, measured by inductively coupled
plasma atomic emission spectroscopy (ICP-AES). The measurements were performed under
serum-free cell culture medium. For this purpose, cells were incubated with a 2 µM solution
of complex 11 for up to 4 h at 37 °C and 5% CO2, and the gold and protein levels of the cells
were determined by ICP-AES and Bradford method, respectively. An acute increase of
cellular gold content occurred upon treatment with complex 11 culminating to 0.087 µg/mg
15

ACCEPTED MANUSCRIPT
protein after 4 h of treatment, which corresponds to 0.46 nmole of gold per mg of protein
(Figure 14). This result suggests the rapid cellular accumulation of the gold bis(NHC)
complex and a good cellular bioavailability, in agreement with the antiproliferative activity of
this complex.
0.15
0.10
0.05
0.00
incubation (h)
Figure 14. Gold uptake by PC-3 cells exposed to 2 µM of complex 11.
2.5. Inhibition of mammalian TrxR
Gold complexes have been reported to generate reactive oxygen species (ROS), which may
be due to their capability to inhibit the TrxR, one of the antioxidant enzymes present in the
mammalian cells together with catalase, glutathione peroxidase and superoxide dismutase that
carry out the removal of ROS from the cell. Inhibition of TrxR could lead to apoptosis
through a mitochondrial pathway [34,55]. However, if gold displays a high affinity towards
thiol and selenol groups, the inhibition potency of TrxR is influenced by the N-substituted
groups of the NHC ligands and cannot be always correlated to the in vitro cytotoxicity. The
potential of complex 11 to inhibit TrxR was studied in vitro on purified rat liver TrxR (see
Supporting Information Figure S1). In addition, inhibition of reducing agents present in cell
extracts has been tested in vitro. Both tests used a previously reported assay based on the
NADPH dependent reduction of the sulfide bond of the substrate 5,5’-dithiobis-2-nitrobenzoic
acid (DTNB) by the isolate enzyme [56]. Complex 11 was able to inhibit isolated TrxR in the
low micromolar range with an IC50 value of 2.2 µM. In order to evaluate the inhibition of
cellular reducing agents, PC-3 cells were exposed to complex 11 in graded concentrations for
1 h and after treatment, the cells were monitored colorimetrically for their ability to reduce
16

ACCEPTED MANUSCRIPT
DTNB. The IC50 value for cellular reducing agents was 7 µM (see Supporting Information
Figure S2).
By comparison with cationic Au(I) bisNHC complexes reported by other groups, Berners-
Price and co-workers synthetized lipophilic cationic gold(I) bisNHC complexes with high
cytotoxicity against breast MDA-MB-231 and MDA-MB-468 cancer cells lines and the most
efficient complex involving iPr groups on the NHC ligands induced at 5 µM a 50% inhibition
of the TrxR activity [46]. Recently, a novel series of C4-substituted cationic gold(I) bis(NHC)
complexes designed by Ott and co-workers displayed high cytotoxicity against several cancer
cells, whereas they just showed moderate to low inhibitory activities towards isolated
mammalian TrxR enzyme with IC50 values ranging from 16.3 to 127.8 µM [30]. Arambula et
al. designed cationic gold(I) complexes containing ferrocenylated bis(NHC) ligands, which
were able to reduce the cellular TrxR activity in the range of 55 to 60% in A549 lung cancer
cells treated with 2.5 µM of complexes [34].
2.6. Effects on intracellular ROS levels
In order to confirm whether ROS played a role in complex 11-induced cytotoxicity, we
used N-acetylcysteine (NAC) and reduced glutathione (GSH), that neutralize ROS, and study
their effects on cell viability. PC-3, T24 and HepG2 cells were pretreated with NAC or GSH
at different concentrations (2 mM, 5 mM and 10 mM) for 1 h before the addition of complex
11. The results showed that both NAC and GSH substantially dose-dependently reduced ROS
accumulation and cytotoxicity of complex 11 in cancer cells (Figure 15), suggesting that
cytotoxicity induced by complex 11 is ROS-dependent. Furthermore, the effect of gold
complex 11 on intracellular ROS levels was detected by fluorometric assays with 2’,7’–
dichlorofluorescein diacetate (DCFDA), a fluorogenic dye that measures hydroxyl, peroxyl
and other ROS activities within the cell. After diffusion into the cells, DCFDA is deacetylated
by cellular esterases to a non-fluorescent compound, which is later oxidized by ROS into
2’,7’–dichlorofluorescin (DCF). DCF is a highly fluorescent compound which can be detected
by fluorescence spectroscopy with maximum excitation and emission spectra of 485 nm and
535 nm, respectively [57]. As shown in Fig 16, PC-3, T24 and HepG2 cells treated with
auronofin (5 µM) and gold complex 11 (1 µM and 5 µM) significantly triggered substantial
generation of ROS in cells within 6 h. Collectively, all these results suggest that ROS play a
critical role in complex 11-induced cell death.
17

ACCEPTED MANUSCRIPT
PC-3
T24
NAC
100
80
60
40
20
0
100
NAC
GSH
GSH
80
**
**
*
60
40
20
0
**
*
**
**
*
**
**
*
**
Concentration of NAC and GSH (mM)
HepG2
Concentration of NAC and GSH (mM)
NAC
GSH
100
80
60
40
20
0
**
**
**
***
**
*
Concentration of NAC and GSH (mM)
Figure 15. Impact of N-Acetyl-L-cysteine (NAC) and reduced glutathione (GSH) on the
cytotoxicity of complex 11 (1 µM). PC-3, T24 and HepG2 cells were treated with complex 11
(1 µM) for 24 h in the absence or presence of different concentrations of NAC and GSH. Cell
viability was measured by MTT assay. Data are presented as means ± SEM of three
independent experiments. *p<0.05, **p<0.005, ***p<0.001, compared with the cell viability
of complex 11 alone.
PC-3
T24
200
150
100
200
150
100
0h
1h
3h
6h
0h
1h
3h
6h
**
***
*
*
*
**
**
**
*
**
**
*
*
***
18

ACCEPTED MANUSCRIPT
HepG2
200
150
100
0h
1h
3h
*
6h
**
**
*
*
*
Figure 16. Induction of ROS by gold complexes (auranofin and complex 11) on PC-3, T24
and HepG2 cells after different times of treatment. *p<0.05, **p<0.005, ***p<0.001,
compared with ROS generation at 0 h.
3.
Conclusion
Pharmacomodulation has been made for four gold-NHC complexes bearing a nitrogen-
atom-containing group on one NHC-side arm in order to increase the lipophilicity of the
complexes. For this purpose, an aliphatic linker of a family of complexes, published before,
has been replaced by an aromatic one. The Log P value of the resulting complexes increased
by 0.7 to 1.5, depending on the substituents in comparison to the aliphatic-linker systems.
With the increase of the lipophilicity, cytotoxicity is rising leading to GI50 values lower than
250 nM for complexes 9 to 12 on the tested cancer cells, but in the same way, cytotoxicity
against non-cancer cells also increased. Deeper ongoing studies have been performed on
complex 11. Clonogenic assays confirmed the MTT results and bioavailability has been
shown by a fast cellular uptake of the gold complexes. TrxR inhibition has been evidenced,
which is in line with the literature on anticancer gold-NHC complexes. Moreover,
mechanistic investigations led to conclusion that the cell death induced by these gold
complexes was ROS dependent. Further pharmacomodulations in order to increase the
selectivity between non-cancerous and cancer cells without loss of high cytotoxicity on cancer
cells are under investigation in our group.
4.
Experimental section
4.1. Materials
19

ACCEPTED MANUSCRIPT
All complexation reactions were performed under an inert atmosphere of dry nitrogen by
using standard vacuum line and Schlenk tube techniques. 4-(4-Bromophenyl)morpholine was
prepared according a modified literature procedure [58]. All other reagents were used as
received from commercial suppliers. Reactions involving silver compounds were performed
with the exclusion of light. Human prostate cancer PC-3 and lung carcinoma A549 cell lines
were obtained from DSMZ (Braunschweig, Germany). Human bladder cancer T24, human
osteosarcoma U-2 OS, mouse osteoblast MC3T3, murine fibroblast NIH3T3, human breast
cancer MCF-7 and human hepatocarcinoma HepG2 cells were from ATCC-LGC Standards
(Molsheim, France). All the cell culture medium, fetal bovine serum (FBS) and phosphate-
buffered saline (PBS) were purchased from Thermo Fisher Scientific. N-Acetyl-L-cysteine
(NAC), reduced Glutathione (GSH) and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) were obtained from Sigma-Aldrich.
4.2. Instrumentation
1H (400 MHz) and 13C NMR spectra (100 MHz) were recorded at 298 K on a Bruker
AV400 spectrometer in CDCl3, DMSO-d6 and CD3OD as solvents. Elemental analyses were
carried out by the “Service de Microanalyse du Laboratoire de Chimie de Coordination
(Toulouse)”. HRMS were performed with a Thermo Finnigan MAT 95 XL spectrometer
using electrospray ionization (ESI) by the “Service de Spectrométrie de Masse de Chimie
UPS-CNRS (Toulouse)”. The absorbance for MTT assay was measured using a Promega
E7031 microplate reader. ICP-AES analysis were performed with a Thermo Scientific iCAP
6300 DUO spectrometer by the “Service Analyses du Laboratoire de Chimie de Coordination
(Toulouse)”.
4.3. Synthesis of proligands 5-8 and complexes 9-12
The following picture describes the numbering of H (1H NMR) and C (13C NMR). These
notations are used in the following experimental section.
20

ACCEPTED MANUSCRIPT
4.3.1. 3-Benzyl-1-[4-(dimethylamino)phenyl]-1H-imidazol-3-ium chloride (5)
Imidazole (3.0 g, 44 mmol), 4-bromo-N,N-dimethylaniline (4.4 g, 22 mmol), K2CO3 (3.04
g, 22 mmol) and a catalytic amount of CuSO4 were stirred in a closed pressure flask at 205 °C
for 10 h. After cooling to room temperature, the crude product was extracted by
dichloromethane, filtered and evaporated to dryness. The resulting residue was purified by
column chromatography on silica gel with a mixture of dichloromethane and methanol (30:1)
as eluent to give 4-(1H-imidazol-1-yl)-N,N-dimethylaniline (1) as a yellow solid (3.7 g, 90%
1
yield). H NMR (400 MHz, CDCl3): δ 7.79 (s, 1H, H2), 7.28-7.23 (m, 4H, H4/H5/H8), 6.78
(d, 2H, J = 8.8 Hz, H9), 3.02 (s, 6H, H11). A mixture of 1 (400 mg, 2.13 mmol) and benzyl
chloride (296 mg, 2.34 mmol) in acetonitrile were stirred in a closed pressure flask at 80 °C
for 10 h. After cooling down to room temperature, the solution was evaporated to dryness.
The residue was purified by column chromatography on silica gel with a mixture of
dichloromethane and methanol (10:1) as eluent to give 5 as a highly hydroscopic yellow solid
(512 mg, 76% yield). Anal. Calcd. For C18H20ClN3: C, 68.89; H, 6.42; N, 13.39. Found: C,
1
68.67; H, 6.38; N, 13.48. H NMR (400 MHz, DMSO-d6): δ 10.02 (s, 1H, H2), 8.22(s, 1H,
H5), 8.02(s, 1H, H4), 7.59(d, 2H, J = 9.2 Hz, H8), 7.57-7.54 (m, 2H, benzyl), 7.46-7.40 (m,
3H, benzyl), 6.86 (d, 2H, J = 9.2 Hz, H9), 5.52 (s, 2H, H6), 2.97 (s, 6H, H11). 13C NMR (100
MHz, DMSO-d6): δ 151.3(1C, C10), 135.2 (1C, C2), 135.1 (1C, benzyl), 129.4 (2C, benzyl),
129.2 (1C, benzyl), 129.0 (2C, benzyl), 124.0 (1C, C7), 123.3 (1C, C5), 123.1 (2C, C8), 122.1
(1C, C4), 112.7 (2C, C9), 52.1 (2C, C11), 49.0 (1C, C6). HRMS (ES+): calcd. for C18H20N3
278.1657; found 278.1658.
4.3.2. 3-Benzyl-1-[4-(diethylamino)phenyl]-1H-imidazol-3-ium chloride (6)
Imidazole (1 g, 14.7 mmol), 4-bromo-N,N-diethylaniline (1.67 g, 7.34 mmol), K2CO3 (2.01
g, 7.34 mmol) and a catalytic amount of CuSO4 were stirred in a closed pressure flask at 205
°C for 10 h. After cooling to room temperature, the crude product was extracted by
dichloromethane, filtered and evaporated to dryness. The resulting residue was purified by
column chromatography on silica gel with a mixture of dichloromethane and methanol (30:1)
as eluent to give N,N-diethyl-4-(1H-imidazol-1-yl)aniline (2) as a brown solid (1.43 g, 90%
yield). 1H NMR (400 MHz, CDCl3): δ 7.73(s, 1H, H2), 7.21(d, 2H, J = 9.2 Hz, H8), 7.18-7.17
(m, 2H, H4/H5), 6.72(d, 2H, J = 9.2 Hz, H9), 3.40 (q, 4H, J = 7.2 Hz, H11), 1.21(t, 6H, J =
21

ACCEPTED MANUSCRIPT
7.2 Hz, H12). A mixture of 2 (300 mg, 1.39 mmol) and benzyl chloride (194 mg, 1.53 mmol)
in acetonitrile were stirred in a closed pressure flask at 80 °C for 10 h. After cooling down to
room temperature, the solution was evaporated to dryness. The residue was purified by
column chromatography on silica gel with a mixture of dichloromethane and methanol (10:1)
as eluent to give 6 as a highly hydroscopic yellow solid (443 mg, 93% yield). Anal. Calcd.
1
For C20H24ClN3: C, 70.26; H, 7.08; N, 12.29. Found: C, 70.47; H, 7.28; N, 12.48. H NMR
(400 MHz, DMSO-d6): δ 9.98 (s, 1H, H2), 8.19 (s, 1H, H5), 8.00 (s, 1H, H4), 7.55-7.51 (m,
4H, benzyl/H8), 7.47-7.40 (m, 3H, benzyl), 6.81 (d, 2H, J = 9.2 Hz, H9), 5.50 (s, 2H, H6),
3.40 (q, 4H, J = 7.2 Hz, H11), 1.11(t, 6H, J= 7.2 Hz, H12). 13C NMR (100 MHz, DMSO-d6):
δ 148.5 (1C, C10), 135.2 (1C, C2), 135.1 (1C,benzyl), 129.4 (2C, benzyl), 129.2 (1C, benzyl),
128.9 (2C, benzyl), 123.6 (2C, C8), 123.3 (1C, C7), 123.1 (1C, C5), 122.2 (1C, C4), 112.0
(2C, C9), 52.6 (1C, C6), 44.3 (2C, C11, 12.7 (2C, C12). HRMS (ES+): calcd. for C20H24N3
306.1970; found 306.1967.
4.3.3. Benzyl-1-(1H-pyrrol-1-yl)-1H-imidazol-3-ium chloride (7)
Imidazole (268.5 mg, 3.94 mmol), 1-(4-chlorophenyl)-1H-pyrrole (500 mg, 2.81 mmol),
K3PO4 (1.19 g, 5.62 mmol) and CuI (107 mg, 0.56 mmol) in DMF were placed to a closed
pressure flask, then stirred at 160 °C for 3 days. After cooling to room temperature, distilled
water was added to the reaction and the crude was extracted with ethyl acetate. The organic
layer was washed with brine and dried over Na2SO4. After filtration and evaporation of the
solvent under reduced pressure, the residue was purified by column chromatography on silica
gel with a mixture of dichloromethane and methanol (30:1) as eluent to give 1-(1H-pyrrol-1-
yl)-1H-imidazole (3) as a yellow solid (300 mg, 51% yield). 1H NMR (400 MHz, DMSO-d6):
δ 8.33 (s, 1H, H2), 7.82 (s, 1H, H5), 7.76-7.71 (m, 4H, H8/H9), 7.45 (t, 2H, J = 2.4 Hz, H11),
7.16 (s, 1H, H4), 6.30 (t, 2H, J = 2.4 Hz, H12). A mixture of (200 mg, 0.96 mmol) of 3 and
benzyl chloride (133 mg, 1.05 mmol) in acetonitrile was stirred in a closed pressure flask at
80 °C for 10 h. After cooling down to room temperature, the solution was evaporated to
dryness. The residue was purified by column chromatography on silica gel with a mixture of
dichloromethane and methanol (10:1) as eluent to give 7 as a yellow solid (306 mg, 95%
yield). Anal. Calcd. For C20H18ClN3: C, 71.53; H, 5.40; N, 12.51. Found: C, 71.57; H, 5.38;
N, 12.48. 1H NMR (400 MHz, DMSO-d6): δ 10.36 (s, 1H, H2), 8.41 (s, 1H, H5), 8.10 (s, 1H,
H4), 7.94-7.89 (m, 4H, H8/H9), 7.59-7.43 (m, 7H, benzyl/H11), 6.33 (s, 2H, H12), 5.56 (s,
2H, H6). 13C NMR (100 MHz, DMSO-d6): δ 140.9 (1C, C10), 136.1 (1C, C2), 135.0 (1C,
22

ACCEPTED MANUSCRIPT
benzyl), 131.9 (1C, C7), 129.4 (2C, benzyl), 129.3 (1C, benzyl), 129.0 (2C, benzyl), 123.7
(2C, C8), 123.6 (1C, C5), 122.1 (1C, C4), 120.4 (2C, C9), 119.6 (2C, C11), 112.7 (2C, C12),
52.8 (1C, C6). HRMS (ES+): calcd. for C20H18N3 300.1501; found 300.1500.
4.3.4. 3-Benzyl-1-[4-(morpholin-4-yl)phenyl]-1H-imidazol-3-ium chloride (8)
To a mixture of 4-phenylmorpholine (816 mg, 5 mmol) and DMSO (429 mg, 5.5 mmol) in
ethyl acetate (20 mL) was added hydrobromic acid (48 %, 1.85 g, 5.5 mmol) at 60 °C under
air. The mixture was stirred at 60 °C, then potassium carbonate (1 g) was added to the
solution and the mixture were stirred for another 0.5 h. After cooling down to room
temperature and evaporation of the solvent under vacuum, the residue was purified by column
chromatography on silica gel with a mixture of hexane and ethyl acetate (10:1) as eluent to
give 4-(4-bromophenyl)morpholine as a white solid (1.03 g, 85% yield). 1H NMR (400 MHz,
CDCl3): δ 7.37 (d, 2H, J = 8.8 Hz), 6.78 (d, 2H, J = 8.8 Hz), 3.86 (t, 4H, J = 4.8 Hz), 3.12 (t,
4H, J = 4.8 Hz). Imidazole (275.6 mg, 4.05 mmol), 4-(4-bromophenyl)morpholine (700 mg,
2.89 mmol), K3PO4 (1.2 g, 5.65 mmol) and CuI (110 mg, 0.56 mmol) in DMF were placed to
a closed pressure flask, then stirred at 160 °C for 3 days. After cooling to room temperature,
distilled water was added to the reaction and the crude was extracted with ethyl acetate. The
organic layer was washed with brine and dried over Na2SO4. After filtration and evaporation
of the solvent under reduced pressure, the residue was purified by column chromatography on
silica gel with a mixture of dichloromethane and methanol (30:1) as eluent to give 4-[4-(1H-
1
imidazol-1-yl)phenyl]morpholine (4) as a yellow solid (200 mg, 30% yield). H NMR (400
MHz, DMSO-d6): δ 8.13 (s, 1H, H2), 7.66 (s, 1H, H5), 7.76-7.71 (d, 2H, J = 8.8 Hz, H8),
7.45 (m, 3H, H4/H9), 3.75 (t, 4H, J = 4.8 Hz, H12), 3.14 (t, 4H, J = 4.8 Hz, H11). HRMS
(ES+): calcd. for C13H16N3O 230.1293; found 230.1293. A mixture of 4 (150 mg, 0.65 mmol)
and benzyl chloride (121.5 mg, 0.96 mmol) in acetonitrile was stirred in a closed pressure
flask at 80 °C for 10 h. After cooling down to room temperature, the solution was evaporated
to dryness. The residue was purified by column chromatography on silica gel with a mixture
of dichloromethane and methanol (10:1) as eluent to give 8 as a highly hydroscopic yellow
solid (200 mg, 91 % yield). Anal. Calcd. For C20H22ClN3O: C, 67.50; H, 6.23; N, 11.81.
1
Found: C, 67.57; H, 6.38; N, 11.58. H NMR (400 MHz, DMSO-d6): δ 10.15 (s, 1H, H2),
8.27 (s, 1H, H5), 8.04 (s, 1H, H4), 7.65 (d, 2H, J = 8.8 Hz, H8), 7.55 (m, 2H, benzyl), 7.47-
7.40 (m, 3H, benzyl), 7.15 (d, 2H, J = 8.8 Hz, H9), 5.52 (s, 2H, H6), 3.76 (t, 4H, J = 4.8 Hz,
H12), 3.22 (t, 4H, J = 4.8 Hz, H11). 13C NMR (100 MHz, DMSO-d6): δ 152.1 (1C, C10),
23

ACCEPTED MANUSCRIPT
135.4 (1C, C2), 135.2 (1C, benzyl), 129.4 (2C, benzyl), 129.2 (1C, benzyl), 129.0 (2C,
benzyl), 126.4 (1C, C7), 123.4 (1C, C5), 123.1 (2C, H8), 122.1 (1C, C4), 115.6 (2C, H9),
66.4 (2C, C12), 52.6 (1C, C6), 48.2 (2C, C11). HRMS (ES+): calcd. for C20H22N3O 320.1763;
found 320.1761.
4.3.5. Complex 9
In a Schlenk tube, 5 (100 mg, 0.32 mmol) was suspended in 10 mL of dry methanol and
Ag2O (74 mg, 0.32 mmol) was added. The mixture was stirred at room temperature for 3 h
under nitrogen in the dark. A solution of AgNO3 (27 mg, 0.16 mmol) in 2 mL of acetonitrile
was added and the mixture was stirred for 10 h. Finally, Au(SMe2)Cl (47 mg, 0.15 mmol) was
added and after stirring for 1 h, the mixture was filtered through a pad of celite. The solvent
was evaporated under reduced pressure and the residue was purified by column
chromatography on silica gel with a mixture of dichloromethane and methanol (20:1) as
eluent to give complex 9 as a yellow solid (52 mg, 40% yield). Anal. Calcd. For
1
C36H38AuN7O3: C, 53.14; H, 4.71; N, 12.05. Found: C, 53.20; H, 4.58; N, 12.13. H NMR
(400 MHz, CD3OD): δ 7.49 (d, 2H, J = 2.0 Hz, H5), 7.45 (d, 4H, J = 9.2 Hz, H8), 7.43 (d, 2H,
J = 2.0 Hz, H4), 7.34-7.31 (m, 6H, benzyl), 7.22-7.19 (m, 4H, benzyl), 6.67 (d, 4H, J = 9.2
13
Hz, H9), 5.28 (s, 4H, H6), 2.87 (s, 12H, H11). C NMR (100 MHz, CD3OD): δ 182.8 (2C,
C2), 150.9 (2C, C10), 136.0 (2C, benzyl), 128.6 (4C, benzyl), 128.1 (2C, C7), 128.1 (2C,
benzyl), 127.5 (4C, benzyl), 125.5 (4C, C8), 123.1 (2C, C5), 121.6 (2C, C4), 111.9 (4C, C9),
54.1 (2C, C6), 39.1 (4C, C11). HRMS (ES+): calcd. for C36H38AuN6 751.2823; found
751.2817.
4.3.6. Complex 10
In a Schlenk tube, 6 (140 mg, 0.41 mmol) was suspended in 10 mL of dry methanol and
Ag2O (94 mg, 0.41 mmol) was added. The mixture was stirred at room temperature for 3 h
under nitrogen in the dark. A solution of AgNO3 (35 mg, 0.21 mmol) in 2 mL of acetonitrile
was added and the mixture was stirred for 10 h. Finally, Au(SMe2)Cl (60 mg, 0.21 mmol) was
added and after stirring for 1 h, the mixture was filtered through a pad of celite. The solvent
was evaporated under reduced pressure and the residue was purified by column
chromatography on silica gel with a mixture of dichloromethane and methanol (20:1) as
eluent to give complex 10 as a yellow solid (110 mg, 62% yield). Anal. Calcd. For
1
C40H46AuN7O3: C, 55.23; H, 5.33; N, 11.27. Found: C, 55.16; H, 5.34; N, 11.25. H NMR
24

ACCEPTED MANUSCRIPT
(400 MHz, CD3OD): δ 7.48 (d, 2H, J = 2.0 Hz, H5), 7.46-7.42 (m, 6H, H4/H8), 7.34-7.30 (m,
6H, benzyl), 7.24-7.19 (m, 4H, benzyl), 6.64 (d, 4H, J = 9.2 Hz, H9), 5.27 (s, 4H, H6), 3.30
13
(q, 8H, J = 7.2 Hz, H11), 1.07 (t, 12H, J = 7.2 Hz, H12). C NMR (100 MHz, CD3OD): δ
182.7 (2C, C2), 148.1 (2C, C10), 136.0 (2C, benzyl), 128.6 (4C, benzyl), 128.1 (2C, benzyl),
127.7 (4C, benzyl), 127.2 (2C, C7), 125.7 (4C, C8), 123.0 (2C, C5), 121.5 (2C, C4), 111.2
(4C, C9), 54.1 (2C, C6), 44.0 (4C, C11), 11.3 (4C, C12). HRMS (ES+): calcd. for
C40H46AuN6 807.3450; found 807.3447.
4.3.7. Complex 11
In a Schlenk tube, 7 (200 mg, 0.60 mmol) was suspended in 10 mL of dry methanol and
Ag2O (138 mg, 0.60 mmol) was added. The mixture was stirred at room temperature for 3 h
under nitrogen in the dark. A solution of AgNO3 (50 mg, 0.30 mmol) in 2 mL of acetonitrile
was added and the mixture was stirred for 10 h. Finally, Au(SMe2)Cl (88 mg, 0.30 mmol) was
added and after stirring for 1 h, the mixture was filtered through a pad of celite. The solvent
was evaporated under reduced pressure and the residue was purified by column
chromatography on silica gel with a mixture of dichloromethane and methanol (20:1) as
eluent to give a yellow solid (150 mg, 60 % yield). Anal. Calcd. For C40H34AuN7O3: C, 56.01;
1
H, 4.00; N, 11.43. Found: C, 56.12; H, 3.89; N, 11.52. H NMR (400 MHz, CD3CN): δ 7.70
(d, 4H, J = 8.8 Hz, H8), 7.50 (d, 2H, J = 2.0 Hz, H5), 7.45 (d, 4H, J = 8.8 Hz, H9), 7.40 (d,
2H, J = 2.0 Hz, H4), 7.32-7.28 (m, 6H, benzyl), 7.21-7.19 (m, 4H, benzyl), 7.09 (t, 4H, J =
2.4 Hz, H11), 6.34 (t, 4H, J = 2.4 Hz, H12), 5.32 (s, 4H, H6). 13C NMR (100 MHz, CD3CN):
δ 182.7 (2C, C2), 140.7 (2C, C10), 136.1 (2C, benzyl), 136.0 (2C, C7), 128.9 (4C, benzyl),
128.4 (2C, benzyl), 127.5 (4C, benzyl), 126.5 (4C, C8), 123.3 (2C, C5), 122.5 (2C, C4), 120.2
(4C, C9), 119.0 (4C, C11), 111.1 (4C, C12), 54.4 (2C, C6). HRMS (ES+): calcd. for
C40H34AuN6 795.2510; found 795.2515.
4.3.8. Complex 12
In a Schlenk tube, 8 (108 mg, 0.30 mmol) was suspended in 10 mL of dry methanol and
Ag2O (70 mg, 0.30 mmol) was added. The mixture was stirred at room temperature for 3 h
under nitrogen in the dark. A solution of AgNO3 (26 mg, 0.15 mmol) in 2 mL of acetonitrile
was added and the mixture was stirred for 10 h. Finally, Au(SMe2)Cl (45 mg, 0.30 mmol) was
added and after stirring for 1 h, the mixture was filtered through a pad of celite. The solvent
was evaporated under reduced pressure and the residue was purified by column
25

ACCEPTED MANUSCRIPT
chromatography on silica gel with a mixture of dichloromethane and methanol (20:1) as
eluent to give a yellow solid (70 mg, 52% yield). Anal. Calcd. For C40H42AuN7O5: C, 53.51;
H, 4.72; N, 10.92. Found: C, 53.65; H, 4.70; N, 10.87. 1H NMR (400 MHz, CD3OD): δ 7.54-
7.51 (m, 6H, H5/H8), 7.46 (d, 2H, J = 2.0 Hz, H4), 7.36-7.32 (m, 6H, benzyl), 7.23-7.20 (m,
4H, benzyl), 6.92 (d, 4H, J = 9.2 Hz, H9), 5.28 (s, 4H, H6), 3.78 (t, 8H, J = 4.8 Hz, H12),
13
3.03 (t, 8H, J = 4.8 Hz, H11). C NMR (100 MHz, CD3OD): δ 182.7 (2C, C2), 151.8 (2C,
C10), 135.9 (2C, benzyl), 130.9 (2C, C7), 128.7 (4C, benzyl), 128.2 (2C, benzyl), 127.6 (4C,
benzyl), 125.6 (4C, C8), 123.0 (2C, C5), 121.8 (2C, C4), 115.2 (4C, C9), 66.3 (4C, C12), 54.2
(2C, C6), 48.3 (4C, C11). HRMS (ES+): calcd. for C40H42AuN6O2 835.3035; found 835.3055.
4.4. Crystallographic data for 9-12
All data were collected at low temperature using oil-coated shock-cooled crystals on a
Bruker-AXS APEX II diffractometer with MoKα radiation (λ = 0.71073 Å). The structures
were solved by direct methods [59] and all non hydrogen atoms were refined anisotropically
using the least-squares method on F2 [60].
9: C36H38AuN7O3, Mr = 813.7, crystal size = 0.40 x 0.10 x 0.02 mm3, monoclinic, space
group Cc, a = 17.300(1) Å, b = 8.384(1) Å, c = 23.168(2) Å, β = 101.883(2)°, V = 3288.2(3)
Å3, Z = 4, 31546 reflections collected, 10626 unique reflections (Rint = 0.0511), R1 = 0.0457,
wR2 = 0.1095 [I>2σ(I)], R1 = 0.0535, wR2 = 0.1141 (all data), Flack x refined to 0.001(7),
residual electron density = 1.553 e Å-3.
10: C40.5H48AuN7O3.5, Mr = 885.82, crystal size = 0.40 x 0.40 x 0.30 mm3, triclinic, space
group P 1, a = 12.989(1) Å, b = 13.120(1) Å, c = 22.637(2) Å, α = 89.475(2)°, β =
84.108(2)°, γ = 79.203(2)°, V = 3769.0(4) Å3, Z = 4, 108465 reflections collected, 22101
unique reflections (Rint = 0.0298), R1 = 0.0249, wR2 = 0.0448 [I>2σ(I)], R1 = 0.0384, wR2 =
0.0486 (all data), residual electron density = 1.408 e Å-3.
11: C40H34AuN7O3, Mr = 857.71, crystal size = 0.50 x 0.30 x 0.05 mm3, monoclinic, space
group C2/c, a = 15.119(2) Å, b = 14.404(2) Å, c = 32.326(4) Å, β = 101.172(3)°, V =
6906.1(13) Å3, Z = 8, 37635 reflections collected, 4910 unique reflections (Rint = 0.0713), R1
= 0.0730, wR2 = 0.1436 [I>2σ(I)], R1 = 0.1020, wR2 = 0.1538 (all data), residual electron
density = 3.517 e Å-3.
CCDC-1846349 to 1846351 contain the supplementary crystallographic data. These data
can be obtained free of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
26

ACCEPTED MANUSCRIPT
4.5. Measurement of lipophilicity
The octanol-water partition coefficients (P) of 9 to 12 were determined using a shake-flask
method [52]. Water (100 mL, distilled after milli-Q purification) and n-octanol (100 mL,
Sigma-Aldrich, ACS spectrophotometric grade, ≥ 99%) were shaken together for 24 h to
allow saturation of both phases. Stock solutions of the four compounds (50 µM) were
prepared in the aqueous phase and aliquots (1 mL) of each of these stock solutions were then
added to an equal volume of the n-octanol phase. The resultant biphasic solutions were mixed
for 1 h and then centrifuged (3000 × g, 5 min) (Eppendorf Centrifuges 5415R and 5804R) to
separate the phases. The concentrations of the complexes in the organic and aqueous phases
were then determined using UV absorbance spectroscopy (260 nm). Log P was defined as the
logarithm of the ratio of the concentrations of the studied complex in the organic and aqueous
phases (Log P = Log {[Au(org)]/[Au(aq)]}; values reported are the means of at least two
separate determinations).
4.6. Cell culture and growth inhibitory assay (MTT assay)
PC-3, HepG2 and A549 cells were cultured in RPMI 1640; T24, MCF-7, U-2 OS and
NIH3T3 in DMEM; and MC3T3 in MEM medium containing 10% fetal bovine serum and
o
1% antibiotics (100 U/mL penicillin and 100 µg/mL streotimycin) at 37 C in 5% CO2
humidified
incubators.
The
MTT
reagent
(3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide) was used to determine cell death as originally described by
Mosmann [61] and modified by Cuvillier et al. [62]. Briefly, cells were seeded 5000 cells/well
in 24-well plates for all kinds of cells and allowed to attach overnight. All of the complexes
were dissolved in DMSO. The concentration of the complexes was calculated according to the
elemental composition of the complexes determined by the elemental analyses. Media in the
presence of the tested complexes were added and serially diluted to various concentrations
(from 5 µM to 0.01 µM). The maximum concentration of DMSO in media did not exceed 0.5
% (v/v). After 24 h, 48 h or 72 h of treatment, cells were incubated at 37 °C and 5% CO2 with
25 µL MTT solution (5 mg/mL; Sigma-Aldrich) in 24-well plates for approximately 3 h. After
solubilization with 500 µL of lysis buffer (DMSO), formazan was quantified by
spectrophotometry with a microplate reader at 570 nm absorbance. The GI50 values
corresponding to the concentration that caused 50% inhibition of cell proliferation were
calculated from dose-response curves obtained by nonlinear regression analysis (Prism 5,
27

ACCEPTED MANUSCRIPT
Graphpad Software). All the results were calculated from data obtained in three independent
experiments.
4.7. Clonogenic assays
PC-3 and T24 cells were counted and seeded (500 cells/well) to 6-well plates. Cells were
allowed to attach overnight and then treated with different concentrations of complex 11 (0.01
µM，0.1 µM, 0.25 µM, 0.5 µM, 1 µM, 5 µM). After incubation for additional 6 days for T24
and 8 days for PC-3, the cells were washed with PBS (2 mL), fixed with methanol (1 mL) for
20 min and stained with 0.5% crystal violet (2 mL) for 30 min. The relative survival was
calculated from the number of single cells that formed colonies of more than 50 cells.
4.8. Cellular uptake studies
The cellular metal uptake was determined as described according to a previously described
method [55]. Briefly, for the cellular uptake into PC-3 human prostate cancer cells, the cells
were cultured until at least 80% confluence in 75 cm2 flasks. Stock solution of complex 11 in
sterile DMSO was prepared and diluted with cell culture medium to a final concentration of 2
µM (freshly prepared, final DMSO concentration of 0.1 % v/v). The cell culture medium was
replaced by the fresh medium containing the gold complexes, and the flasks were incubated at
°
37 C and 5% CO2 for 1, 2 and 4 h. After the desired incubation period, the cellular uptake
was stopped by removing the cell culture medium. The cells were washed with PBS (10 mL),
and then the pellets were collected after 10 min of trypsinization (0.05% trypsin-EDTA
solution (10 mL) and centrifugation (room temperature; 1,500 rpm, 5 min). Cells were
resuspended in Mili-Q water (1 mL) and lysed by sonication for 30 s. The protein contents of
lysates were determined by the Bradford method. 100 µL of samples were digested in 300 µL
of 70 % HNO3 at 70 °C for 2 h and room temperature overnight, then diluted 1:100 in
ultrapure water for inductively coupled plasma atomic emission spectroscopy (ICP-AES)
analysis. Results were calculated from the data obtained from three independent experiments
and are expressed as microgram of gold per milligram of protein.
4.9. Inhibition of mammalian TrxR
To determine the inhibition of mammalian TrxR, an established microplate reader-based
assay was performed with minor modifications [63]. For this purpose, commercially available
rat liver TrxR (from Sigma-Aldrich) was used and diluted with distilled water to achieve 3.5
28