The Anticancer Activity of Organotelluranes: Potential Role in Integrin Inactivation

Article abstract of DOI:10.1002/cbic.201500614

Organic Te^IV compounds (organotelluranes) differing in their labile ligands exhibited anti-integrin activities in vitro and anti-metastatic properties in vivo. They underwent ligand substitution with l-cysteine, as a thiol model compound. Unlike inorganic Te^IV compounds, the organotelluranes did not form a stable complex with cysteine, but rather immediately oxidized it. The organotelluranes inhibited integrin functions, such as adhesion, migration, and metalloproteinase secretion mediation in B16F10 murine melanoma cells. In comparison, a reduced derivative with no labile ligand inhibited adhesion of B16F10 cells to a significantly lower extent, thus pointing to the importance of the labile ligands of the Te^IV atom. One of the organotelluranes inhibited circulating cancer cells in vivo, possibly by integrin inhibition. Our results extend the current knowledge on the reactivity and mechanism of organotelluranes with different labile ligands and highlight their clinical potential.

Full text of DOI:10.1002/cbic.201500614

DOI: 10.1002/cbic.201500614
Full Papers
The Anticancer Activity of Organotelluranes: Potential
Role in Integrin Inactivation
Alon Silberman,[a, b, c] Yona Kalechman,[b] Shira Hirsch,[a] Ziv Erlich,[b] Benjamin Sredni,*[b] and
Amnon Albeck*[a]
Organic TeIV compounds (organotelluranes) differing in their
labile ligands exhibited anti-integrin activities in vitro and anti-
metastatic properties in vivo. They underwent ligand substitu-
tion with l-cysteine, as a thiol model compound. Unlike inor-
ganic TeIV compounds, the organotelluranes did not form a
stable complex with cysteine, but rather immediately oxidized
it. The organotelluranes inhibited integrin functions, such as
adhesion, migration, and metalloproteinase secretion media-
tion in B16F10 murine melanoma cells. In comparison, a re-
duced derivative with no labile ligand inhibited adhesion of
B16F10 cells to a significantly lower extent, thus pointing to
the importance of the labile ligands of the TeIV atom. One of
the organotelluranes inhibited circulating cancer cells in vivo,
possibly by integrin inhibition. Our results extend the current
knowledge on the reactivity and mechanism of organotellur-
anes with different labile ligands and highlight their clinical po-
tential.
Introduction
Tumor metastasis is a dynamic process involving a number of
complex interactions between tumor cells and their environ-
ment.[1] The progression of a tumor and its metastasis depend
on factors such as cellular adhesion molecules, proteins of the
extracellular matrix (ECM), and proteases (e.g., matrix metallo-
proteinases, MMPs). Specifically, integrins play important roles
during metastasis.[2]
bind FN.[3] Thus, FN is involved in many biological processes,
such as tumor formation and metastases.
Although a wealth of evidence[4] shows that “inside-out” sig-
naling (activation of the ligand binding function) can control
integrin activation, it has been postulated that specific integrin
function, at least, might be directly affected by redox rear-
rangements within the cysteine-rich domain of the extracellu-
lar integrin regions. Thus, a disulfide bond-reshuffling mecha-
nism in which resting and active integrins differ in the number
and positions of unpaired cysteine residues has been pro-
posed.[5] We found that integrins are key targets for inorganic
tellurium compounds. Based on the unique TeIV–thiol chemis-
try,[6] we have shown that the inorganic TeIV compounds am-
monium [trichloro(dioxoethylene-O,O’-)tellurate] (AS-101) and
octa-O-bis-(R,R)-tartarate ditellurane (SAS; Scheme 1), synthe-
sized and studied by us,[7] interact with specific thiol-contain-
ing integrins by redox modulation, thereby affecting their con-
formation and inhibiting their physiological activity. This phe-
nomenon was specifically relevant to integrin a4b1 (very late
antigen 4; VLA-4),[8] inhibition of which has consequences both
in vitro and in vivo. B16F10 melanoma cells abundantly ex-
press VLA-4,[9] which interacts with VCAM-1 (vascular cell adhe-
sion molecule 1) to enhance transendothelial migration of mel-
anoma cells.[10]
Integrins are heterodimeric membrane glycoproteins com-
prised of two noncovalent subunits (a and b) that promote
cell adhesion and migration on the surrounding ECM. Integrins
expressed on tumor cells contribute to the metastatic process
by increasing tumor cell migration, invasion, proliferation, and
survival. Adhesion of a tumor cell to the ECM by its integrins is
often regulated by the expression and secretion patterns of
various ECM ligands. For example, the fibrillar cell-adhesion
protein fibronectin (FN) is a ligand for various integrin recep-
tors; at least ten different receptors have been recognized to
[a] Dr. A. Silberman, S. Hirsch, Prof. A. Albeck
The Julius Spokojny Bioorganic Chemistry Laboratory
Department of Chemistry, Bar-Ilan University
Ramat-Gan 5290002 (Israel)
E-mail: amnon.albeck@biu.ac.il
[b] Dr. A. Silberman, Dr. Y. Kalechman, Z. Erlich, Prof. B. Sredni
C.A.I.R. Institute, The Safdi› AIDS and Immunology Research Center
The Mina & Everard Goodman Faculty of Life Sciences, Bar-Ilan University
Ramat-Gan 5290002 (Israel)
However, prior to redox modulation of integrins by either
AS-101 or SAS, ligand substitution reactions occur: substitution
of labile ligands covalently bound to the tellurium atom (i.e.,
good leaving groups) with cell-membrane cysteines.[11,12] Based
on our knowledge regarding the different labile ligands re-
sponsible for the reactivity of the inorganic TeIV compounds
AS-101 and SAS, we hypothesized that synthetic TeIV organotel-
lurane compounds bearing labile ligand moieties similar to
E-mail: benjamin.sredni@biu.ac.il
[c] Dr. A. Silberman
Present address: Department of Biological Regulation
The Weizmann Institute of Science
Rehovot 7610001 (Israel)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201500614.
ChemBioChem 2016, 17, 918 – 927
918
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ed. This led us to assess whether organotelluranes have anti-in-
tegrin activity in cancer and to elucidate the role of the labile
ligands bound to TeIV. Also, we addressed several mechanistic
questions regarding the interaction between organotelluranes
and cysteines in vitro, and investigated the clinical potential of
one organotellurane compound in vivo.
Results
Chemistry
Design and synthesis of organotelluranes: A few organotellur-
anes of the form TeAr2X2 (1, 2, and 3, Scheme 1) were designed
such that each shared the same core structure but had only
one type of labile ligand (leaving group). The chlorides and
carboxylates in 1 and 3 mimic the leaving groups of the highly
reactive compounds AS-101 and SAS, respectively (Scheme 1).
The oxide moiety (or its hydrate form) of 2 imitates a possible
hydrolyzed oxide product of AS-101 and/or SAS, as we previ-
ously reported.[12] The corresponding TeII analogue
4
Scheme 1. Inorganic tellurium compounds AS-101 and SAS, and the synthet-
(Scheme 1), which bears no labile ligand moiety, was used to
evaluate the effect of the leaving group on integrin inhibition
and to study the mechanism of the reaction between our syn-
thetic organotelluranes and l-cysteine. In order to identify the
specific effect of the leaving group and to exclude any effects
of the interactions between the organotellurane side chain
and the target protein, we used the same methoxybenzene
aryl moiety in all four compounds. All were synthesized by
known procedures and were obtained in reasonable overall
yields (Scheme 2).[19–22] Briefly, tellurium tetrachloride (5) was
dissolved in 6 equivalents of anisole, and the mixture was re-
fluxed for 6 h under dry argon, thereby allowing a Friedel–
Crafts tellurization reaction to occur by substituting two chlor-
ines for two aromatic anisole rings. Solvent evaporation and
hot filtration in boiling acetonitrile afforded 1 in 47% yield as
ic organotellurium compounds.
those of AS-101 and SAS are likely to exhibit biological activity
in vitro, possibly associated with integrins.
Organotelluranes, a sub-family of organotellurium com-
pounds, have at least one TeÀC covalent bond. They have
been studied as antioxidants,[13] as inducers of the Ca2+-depen-
dent MPTP (mitochondrial permeability transition pore) open-
ing,[13] as anti-leishmanial agents,[14] and as antiepileptogenic
agents.[15]
Previously, we evaluated a few inorganic TeIV compounds as
specific inhibitors of the cysteine proteases papain and cathe-
psin B.[12] We showed that ligand substitution takes place be-
tween the labile ligands of AS-101 or SAS upon reaction with
four equivalents of l-cysteine (per one tellurium atom), thereby
resulting in a stable Te–Cys4 complex. Since then, organotellur-
ane compounds have also been demonstrated to inhibit cys-
teine proteases.[16] Cunha et al. evaluated different organotel-
luranes as inhibitors of the cysteine proteases cathepsins B, L,
S, and K;[17] however, the authors suggested that the different
reactivity patterns were attributable to a combination of the
tellurium labile groups and the organic side chain, which was
manipulated to fit the cathepsin allosteric sites. Moreover,
whether these biological activities allow the formation of
a stable Te–Cys complex, as seen for inorganic TeIV compounds,
is unknown.
As cathepsins are involved in malignancy progression, anti-
tumor activity was also expected for organotelluranes. Indeed,
a new organotellurane molecule, RT-04, had cell-killing effects
in the human promyelocytic leukemia cell line HL60,[18] by trig-
gering apoptosis (DNA fragmentation and caspase-3, -6, and -9
activation).
Nevertheless, the role of the labile ligand moieties of orga-
notelluranes has been poorly studied, and to the best of our
knowledge, no structure–activity relationship (SAR) experi-
ments with integrins (the suggested target) have been report-
Scheme 2. Synthetic pathways for organotellurium compounds 1–4. a) ani-
sole, reflux, 6 h, 47%; b) sodium ascorbate, H2O/MeOH, RT, 27 h, 93%;
c) NaOH (2m), reflux, 3 h, 58%; d) Ac2O, CHCl3, RT, 90 min, 65%.
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
919
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
pure crystals. Next, 1 was refluxed in NaOH (2m) for 3 h. Etha-
nol was then added under reflux, thereby resulting in full dis-
solution of the reaction mixture, which was then cooled to
48C to allow formation of pure crystals of 2 in 58% yield. For
the synthesis of 3, 2 was dissolved in chloroform and treated
with one equivalent of acetic anhydride at room temperature
for 90 min. Next, the unreacted acetic anhydride was evaporat-
ed under vacuum, and the crude product was crystallized from
hexane/chloroform (3:1) to afford pure crystals of 3 in 65%
yield. For the synthesis of 4, 1 was dissolved in acetone, and
the mixture was added to a suspension of sodium ascorbate in
MeOH/H2O (8:2), and the resulting clear reaction mixture was
stirred for 27 h. The product was extracted with dichlorome-
thane, dried over CaCl2, and evaporated under vacuum to af-
ford 4 in 93% yield.
Biological evaluation
In vitro assays
VLA-4 integrin is a target for organotelluranes 1–3: Adhe-
sion screening assays. In order to evaluate the biological activi-
ties of our synthetic organotelluranes and to assess whether
integrins are also a target for organotelluranes, we used adhe-
sion assays. Adhesion of integrin-expressing cells to ligand-
NMR studies: We characterized the chemical reaction be-
tween the synthetic organotelluranes and l-cysteine (a thiol
model compound) by applying 125Te and 13C NMR to follow the
reaction products. Upon addition of two equivalents of l-cys-
teine, 1, 2, and 3 were immediately and completely reduced,
in the NMR tube, to yield the corresponding diorganotelluride
4 (Scheme 3B; resonates at 639 ppm, see Figure S4C in the
Scheme 3. Reactions of tellurium compounds with l-cysteine. A) The ligand
substitution reaction of the inorganic tellurium compounds AS-101 and SAS
with l-cysteine leads to the formation of a Te–Cys4 complex, as previously
reported.[12] B) Suggested scheme of reactions between organotelluranes of
the form TeAr2X2 and l-cysteine, thereby leading to the reduced organotel-
lurium(II) product.
Supporting Information). 13C NMR data support the formation
of disulfide cystine, the product of the oxidation of two cys-
teine molecules, in comparison to a blank (not shown). To our
surprise, no complex of the form TeAr2XmCys2Àm was detected,
thus suggesting a different mechanism of reaction and/or ki-
netics for organotelluranes in comparison to the inorganic TeIV
compounds AS-101 and SAS.[12]
Figure 1. Adhesion of synthetic organotellurium compounds in vitro. A) Cells
were seeded in a 96-well plate onto an FN- or BSA-coated surface, together
with organotellurane or the diorganotelluride (0, 1.2, 2.4, or 6.0 mm) for 1 h.
The cells were then washed, and the attached cells were subjected to an
XTT assay (450 nm). Data are meanÆSD (n=3); *p<0.05 in comparison to
1% DMSO control (FN). B) Fold change of the adherence for 6 mm 1–3 in
comparison to 4; *p<0.05. C) Representative images of the adherence
assay. Cells were seeded in a 6-well plate onto an FN- or BSA-coated surface,
together with 1 (1.2, 2.4, and 6.0 mm) for 1 h. The cells were then washed,
and the cells remaining attached were photographed with ”10 magnifica-
tion.
Organotellurane 1 was further reacted with two equivalents
of a different thiol compound, N-acetylcysteine. 125Te NMR of
the reaction solution revealed, in addition to the reduced di-
organotelluride 4, a new signal at 921 ppm, thus indicating the
formation of a TeIV complex of the form TeAr2Xm(N-acetylcys-
teine)2Àm (Figure S5).
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
920
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
coated plates is an established method to study integrin func-
tion, and we have previously applied this for the study of vari-
ous integrins and ligands.[8] First, we quantified B16F10 cells
adhered to either FN or to bovine serum albumin (BSA, a non-
specific ligand), after treatment with 1, 2, 3, or 4 (1.2, 2.4, and
6.0 mm). At all concentrations, 1–3 were biologically active (17–
83% adhesion inhibition), and exhibited dose-dependent reac-
tivity (Figure 1A). As expected, 4 inhibited adhesion of B16F10
cells to FN to a significantly lower extent (34% at 6 mm) than
1, 2, and 3 (80–83% at 6 mm; Figure 1A). Next, we assessed
whether our organotelluranes specifically target integrin, by
quantifying B16F10 cells adhered to VCAM-1 (Figure 2) after
organotelluranes and cell-membrane cysteines and to confirm
integrins as a target for organotelluranes.
In order to support our adhesion findings and to exclude cy-
totoxic effects on B16F10 cells, we incubated 1 (2.4 and 6 mm)
with cells for 1 h and assessed the results with propidium
iodide (PI). FACS analysis revealed that 1 is non-toxic at either
of these concentrations (Figure 3A; positive control: etopo-
side). In order to confirm this result, we incubated cells with
1 (6 mm) for a few hours, and quantified the number of living
cells by using the XTT colorimetric assay. No toxic effects were
observed for 1 (Figure 3B; positive control: cycloheximide).
Taken together, these results suggest that the inhibition of ad-
hesion by 1 can be attributed to specific inhibition of VLA-4
rather than to cytotoxicity.
Next, we tested the ability of 1 to inhibit the migration of
B16F10 cells; migration can also be induced by integrins.[23] Or-
ganotellurane 1 significantly inhibited migration only when
the cells were incubated in the presence of FN (67–76%, Fig-
ure 4A), whereas cells that were incubated with the nonspecif-
ic ligand BSA (as a negative control) did not demonstrate any
migration. These results suggest a possible ligand-dependent
activation of B16F10 integrins and their further inhibition, in
accordance with the adhesion screening assay results (Figure 1
and Figure 2). Further to our observation that 1–3 do not form
a stable TeAr2XmCys2Àm complexes when treated with l-cysteine
(but are immediately reduced), we wished to address the bio-
chemical interaction between organotelluranes and cell-mem-
brane cysteines. For this, we applied the impermeable cysteine
modifier pCMBS,[24] which irreversibly binds cysteine thios, and
followed its ability to inhibit cell migration. No impaired migra-
tory activity was observed, compared with 1-treated cells (at
2.4 and 6 mm, Figure 4A), thus pointing to a different mecha-
nism of interaction between pCMBS and cell-membrane cys-
teine thiols.
Next, we wished to evaluate the clinical potential of organo-
telluranes. For this, we first assessed the anti-invasionary prop-
erties of 1. Specifically, we followed the enzymatic activities of
secreted proMMP-2 and -9, both of which are involved in
metastasis[25] and regulated by VLA-4. Interestingly, 1 highly
suppressed the secretion of proMMP-9 and significantly sup-
pressed the secretion of proMMP-2 (Figure 4B).
Figure 2. Synthetic organotellurium compounds inhibit VCAM-1 in vitro: ad-
hesion assay. A) Cells were seeded in a 96-well plate onto a VCAM-1- or BSA-
coated surface, together with organotelluranes or diorganotelluride (0, 1.2,
2.4, or 6.0 mm) for 1 h. The cells were then washed, and the attached cells
were subjected to an XTT assay (450 nm). Data are meanÆSD (n=3).
*p<0.05 in comparison to 1% DMSO control; n.s.: not significant. B) Fold
change of the adherence assay for 6 mm 1–3 in comparison to 4; *p<0.05.
In vivo activity
We wished to test the ability of 1 to inhibit metastasis in vivo.
First, we evaluated the toxicity of 1 by applying an acute toxic-
ity model. Organotellurane 1 was injected intraperitoneally (IP)
into C57Bl/6 male mice every other day, for a total of three
weeks per dose. Because of the limited solubility of 1 in the
vehicle, the highest administered dose was 5.4 mgkgÀ1. En-
couragingly, no severe weight loss (<10%) was observed at
any of the administered doses following each three-week ex-
periment (Table 1). Most importantly, not even mild behavioral
clinical symptoms pointing to deterioration in the condition of
the mice were observed at any dose.
treatment with 1, 2, 3, or 4 (1.2, 2.4, and 6.0 mm): 1, 2, and 3 in-
hibited adhesion (6–62% inhibition, Figure 2A), whereas 4 did
not cause significant inhibition (5% at 6 mm, Figure 2). Further-
more, B16F10 cells did not adhere to the a4b7 ligand
MadCAM (not shown). These data imply that 1–3 specifically
inhibit the activity of the a4b1 integrin, which is abundantly
expressed on B16F10 melanoma cells; they do not inhibit non-
specifically other a4 integrins (e.g., a4b7) that are not ex-
pressed on these cells.
In vitro toxicity, cell migration, and zymography. In vitro bio-
logical studies with 1 aimed to study the interaction between
Finally, as a proof of concept, we studied the effect of 1 on
the migratory properties of B16F10 murine melanoma cells in
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
921
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 3. Organotellurane 1 is non-toxic in vitro. A) Cells were cultured to 60% confluence. Upon medium replacement, cells were incubated with either
1 (2.4 or 6 mm) or DMSO (1%) for 1 h, or overnight with etoposide (40 mm). Floating cells were collected together with trypsinized cells, centrifuged at 394g
and suspended in 1 mL of PBS with 1 mL of propidium iodide (PI). PI-positive and -negative cells were quantified by flow cytometry (representative images
from two experiments). B) Cells were incubated for 6 h in a 96-well plate with either 1 (6.0 mm) or controls, DMSO (1%) or cycloheximide (5 mgmLÀ1). The cells
were subjected to an XTT assay (450 nm). Data are meanÆSD (n=3); *p<0.05 in comparison to cycloheximide.
vivo, by using an experimental metastasis model.[26] In C57Bl/6
mice injected with B16F10 cells into the tail vein and IP treated
with 1 (0.25, 0.9, or 1.8 mgkgÀ1) or vehicle (1%DMSO, control),
1 inhibited the formation of liver metastases. Quantification re-
vealed that 1 significantly inhibited infiltration of B16F10 cells
into the liver of C57Bl/6 mice (55% at 0.9 mgkgÀ1; Figure 5A).
Representative images of livers were taken from both treated
and untreated mice (Figure 5B).
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
922
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 4. Organotellurane 1 inhibits cell migration and proMMPs secretion,
in vitro. A) Cells were incubated for 30 min at 378C with 1 (0, 1.2, 2.4, or
6.0 mm) or pCMBS (300 mm), in the presence of mobilized FN and MnCl2.
Cells incubated with mobilized BSA served as a negative control. Next, the
cells were subjected to a migration assay. The migrated cells were collected
the following day and quantified. Data are meanÆSD (n=3). *p<0.05 in
comparison to control. B) Cells were incubated for 72 h on FN-coated plates
with 0.1% BSA-containing medium with 1 (0, 2.4, or 6.0 mm). The condi-
tioned medium was collected and subjected to zymography. A representa-
tive image and quantification (densitometric analysis) of proMMP-2 and -9
on a zymography gel (n=3); *p<0.05 in comparison to control
Figure 5. Organotellurane 1 inhibits circulating B16F10 cells in vivo.
A) B16F10 cells were injected into the tail vein of C57Bl/6 mice, which were
divided into three groups (N=10–14). Each group was treated with a differ-
ent dose of 1 (0.25, 0.9, or 1.8 mgkgÀ1 or 1% DMSO as control), injected IP
for three consecutive days and then every other day. Following 18 days, the
mice were sacrificed and subjected to liver analysis. The number of nodules
(metastatic lesions) was quantified. B) Representative images of livers from
untreated (1% DMSO, right) and treated (0.9 mgkgÀ1, left) mice. Data are
meanÆSD (N=10–14 mice). *p<0.05.
Table 1. Average weights of mice in acute toxicity assay with 1.
Administered dose
[mgkgÀ1
Weight
[g]
Administered dose
[mgkgÀ1
Weight
[g]
]
]
chemistry behind these reactions is different for organotellur-
anes and whether a stable and detectable complex of the
form TeAr2XmCys2Àm can also be formed when reacting organo-
telluranes of the form TeAr2X2 with l-cysteine, was unknown.
To our surprise, the chemical interaction between the orga-
notelluranes and l-cysteine was substantially different to that
for the inorganic tellurium compounds. Both AS-101 and SAS
yielded the stable ligand-substituted complex TeCys4
(1807 ppm, 125Te NMR) when separately reacted with four
equivalents of l-cysteine (Scheme 3A).[11,12] These results are in
accordance with recent experiments where we identified selec-
tive affinity of AS-101 towards vicinal thiols and possible for-
mation of an S-Te-S intermediate.[8] In contrast, upon reacting
our organotelluranes with l-cysteine, an immediate redox reac-
tion yielded the TeII derivative 4 (Scheme 3B), which was also
independently synthesized and characterized (Scheme 2). The
fact that no stable complex of the form TeAr2XmCys2Àm (X=Cl,
CH3COO, OH) was detected excludes the formation of a stable
S-Te-S intermediate. Our results are also in agreement with
a study that demonstrated that the number of labile groups in
each organotellurane compound corresponds to the number
of consumed cysteines.[17] Thus, an organotellurane composed
(1% DMSO)
0.90
23.81Æ0.83
22.55Æ0.61
22.78Æ0.26
4.00
5.40
23.55Æ0.67
22.27Æ1.69
2.25
Discussion
Over the last few years, the inorganic biologically active TeIV
molecules AS-101 and SAS were shown to exhibit diverse bio-
logical activities.[11,27,28] Many were attributed to the pivotal tel-
lurium(IV) atom and its high affinity for thiols.[6]
Following our previous observations that the anti-metastatic
activities of AS-101[8] and SAS are by integrin inhibition, we
wished to evaluate the potential of organotelluranes as anti-
metastatic agents, and to address several mechanistic ques-
tions regarding their interactions with thiols. Here, we synthe-
sized and characterized organotelluranes 1–3 and diorganotel-
luride 4, which lacks the leaving group moieties.
The inorganic compounds AS-101 and SAS are highly re-
active towards l-cysteine, and yield stable, NMR-detectable
complexes of TeCys4 (Scheme 3A).[11,12] However, whether the
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
923
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
of two chlorines consumes only two cysteine equivalents, as
was also observed for 1–3, which rapidly consumed two cys-
teine equivalents and were reduced to the corresponding TeII
analogue 4 (with no trace of starting material). In contrast, re-
acting 1 (as a representative) in the NMR tube with two equiv-
alents of N-acetylcysteine resulted in the formation of a new
TeIV complex at 921 ppm (Figure S5), in addition to 4 and the
starting material. This suggests that even in the protein-free
“clean” NMR environment, the chemistry between organotel-
luranes and thiols can be controlled by stereoelectronic fac-
tors.
We selected 1 for subsequent in vitro and in vivo experi-
ments aimed at studying the chemistry between organotellur-
anes and cell-membrane thiols, and to confirm integrins as
a target for organotelluranes. As cell migration and tissue inva-
sion are essential for metastasis, we evaluated 1 for its ability
to inhibit the migratory activity of the highly metastatic
B16F10 cells and to inhibit proMMP secretion, two key meta-
stasis activities that are known to be regulated by integrin re-
ceptors such as VLA-4..[30,31] Thus, having demonstrated its abil-
ity to inhibit adhesion (Figures 1 and 2) in a specific manner
rather than by cytotoxic effects (Figure 3), we expected that
1 would inhibit the migratory activity of B16F10 cells. Indeed,
1 significantly inhibited the migratory activity only of B16F10
cells (6–76%, Figure 4A) that were incubated in the presence
of FN. In comparison, no migration activity was observed for
cells that were incubated in the presence of the nonspecific
ligand BSA, thus pointing to specific inhibition by 1, possibly
at FN-activated integrins. Importantly, incubation of the cells
with the impermeable cysteine modifier pCMBS did not have
any effect on the migratory activity of B16F10 cells. This sug-
gests that in order to inhibit the migratory activity of B16F10
cells, oxidation of cysteines and formation of disulfides are
necessary. This further corroborates our NMR experiments
(Scheme 3), in which we observed a redox-based mechanism
for 1–3 when reacting with l-cysteine.
Integrins also contribute to the invasion properties of tumor
cells by regulating the localization and activity of matrix-de-
grading proteases, such as MMP-2 and -9.[32] Cultured melano-
ma cells have been shown to produce at least MMP-1, -2, -3,
and -9, and their activity correlates with melanoma invasion.[33]
Indeed, our cultured B16F10 murine melanoma cells were
found to secrete high levels of proMMP-2 and to a lesser
extent proMMP-9, in agreement with previous findings (Fig-
ure 4B, untreated controls). Inhibited secretion of proMMP-9
and proMMP-2 (Figure 4B) by 1 suggests a further mechanistic
link for the inactivation of VLA-4 and possibly other FN-activat-
ed integrin receptors.
In order to assess the anti-cancer activity of our synthetic or-
ganotelluranes and to determine whether the cysteine-rich
domain of extracellular integrin regions is a possible target for
organotelluranes, we conducted adhesion screening experi-
ments in which we followed the ability of 1–3 to inhibit the
adhesion of murine melanoma B16F10 cells to the abundant
integrin ligand FN. In addition, in order to evaluate the impor-
tance of the labile ligand moiety to integrin inhibition, we also
studied the diorganotelluride 4, which bears no labile ligands.
All three organotellurane compounds inhibited adhesion of
B16F10 cells to FN (Figure 1A) in the low-micromolar range
(partial inhibition (17–56%) at 1.2 mm; almost complete inhibi-
tion (80–83%) at 6 mm). In comparison, 4 inhibited adhesion of
B16F10 cells to FN to a significantly lower extent at 6 mm
(33%, Figure 1A and B). Although less probable, we cannot
exclude the possibility of substitution reactions between 4 (at
its methoxy phenyl carbons) and integrin membrane thiols.
However, other studies suggest that diorganotellurides work
by other mechanisms.[29]
In order to assess whether our synthetic organotelluranes
specifically target integrin, we quantified B16F10 cells adhered
to VCAM-1 (Figure 2) after treatment with 1, 2, 3, or 4. Deriva-
tives 1, 2, and 3 inhibited adhesion (6–62% inhibition, Fig-
ure 2A), thus suggesting high specificity toward VLA-4, where-
as 4 did not significantly inhibit adherence (5% at 6 mm, Fig-
ure 2A and B).
These results point again to the important role of the labile
ligand of the organotelluranes in relation to integrin inhibition
and confirm our previous observations (Figure 1B).
In light of the ability of 1 to inhibit in vitro cell adhesion to
VCAM-1, migration, and MMP secretion, we wished to assess
its activity in vivo. However, because preclinical toxicology
studies play an important role in drug development, and be-
cause of the scarcity of data regarding the toxicity of tellurium
compounds (and specifically of organotelluranes), we wished
to evaluate the acute toxic effects of 1 on C57Bl/6 mice. No
severe changes (<10%) in the average weights of the mice
were observed in response to administered doses of 1
(Table 1), neither was there any acute toxic clinical symptom.
Therefore, we conclude that intraperitoneal administration of
1 at 0.9–5.4 mgkgÀ1 is safe in vivo. As a proof of concept for
the ability of 1 to inhibit the infiltration of B16F10 cells into
organs/tissues (a process that is known to involve integrins),[10]
we applied an experimental liver metastasis model. Hepatic
metastases are most frequent, and once diagnosed they are
often associated with poor prognosis.[34] We showed that 1 sig-
nificantly inhibited the formation of liver metastases in C57Bl/6
mice that were directly injected with B16F10 melanoma cells
through the tail vein (55% inhibition at 0.9 mgkgÀ1; Fig-
B16F10 cells did not adhere to the a4b7 ligand MadCAM
(not shown). These data imply that 1–3 specifically inhibit the
activity of the a4b1 integrin, abundantly expressed on B16F10
melanoma cells and do not merely inhibit nonspecifically other
a4 integrins (e.g., a4b7) that are not expressed on these cells.
In light of these results, we suggest that 1–3 participate in
ligand substitution reactions with cell-membrane integrins,
specifically VLA-4 (Figure 2). We conclude that the different
labile ligands do indeed play a significant role in the biological
activity of organotelluranes. However, we cannot exclude the
possibility that all the analogues were partly hydrolyzed in the
presence of aqueous biological solutions to the same active
metabolite. Nevertheless, after establishing the role of the
labile ligands, the similar range of activity of the synthetic or-
ganotelluranes might be a good reference point for further
manipulation of the aryl side chain for the development of
more-potent TeIV organotelluranes.
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
924
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
cooled, left at 48C for 30 min, then left to precipitate at À208C
overnight. The collected product was obtained in 47% yield.
1H NMR ([D6]DMSO): d=7.87 (d, J=9.0 Hz, 4H; H-3), 7.11 (d, J=
9.0 Hz, 4H; H-2), 3.81 (s, 6H; CH3); 13C NMR ([D6]DMSO): d=161.0
(q, C1), 136.1 (CH, C3), 128.2 (q, C4), 114.8 (CH, C2), 55.4 (CH3); 125Te
NMR ([D6]DMSO): d=994.8; elemental analysis calcd (%) for
C14H14Cl2O2Te : C 40.74, H 3.42; found: C 40.44, H 3.37; HRMS (CI+):
m/z calcd: 378.9730; found: 378.9735 ([M+-Cl], 100).
ure 5A), thus validating 1 as an in vivo bio-active compound,
possibly by the inactivation of integrins.
Conclusion
We have studied the anti-cancer activities of organotelluranes.
Our NMR experiments indicate an immediate redox reaction
between organotelluranes (at their labile ligands) and l-cys-
teine. These findings are substantially different from the previ-
ous reported behavior of inorganic tellurium compounds. This
study demonstrates for the first time, to the best of our knowl-
edge, that VLA-4 is a target for organotellurane compounds.
The labile ligands of organotelluranes play a significant role in
cancer cell adhesion to the abundant integrin ligand FN and to
the specific VLA-4 integrin ligand VCAM-1.
4,4’-Tellurinylbis(methoxybenzene) (2): Organotellurane 1 (0.5 g,
1.21 mmol) was stirred in NaOH (2m, 2 mL) in a 10 mL flask at
reflux for 3 h. Then, ethanol (1 mL) was slowly added under reflux,
thereby resulting in full dissolution of the reaction mixture, which
was then cooled to RT, and then to 48C to allow crystallization.
The crystals were washed with a small amount of H2O/EtOH (2:1)
to afford purified crystals of the product in 58% yield. 1H NMR
(CD3OD): d=7.78 (d, J=7.6 Hz, 4H; H-3), 7.10 (d, J=7.6 Hz, 4H; H-
2), 3.84 (s, 6H; CH3); 13C NMR (CD3OD): d=163.6 (q, C1), 135.2 (CH,
C3), 126.2 (q, C4), 116.3 (CH, C2), 56.0 (CH3); 125Te NMR ([D6]DMSO):
d=1209.2; elemental analysis calcd (%) for C14H14O3Te : C 46.98, H
3.94; found: C 46.63, H 3.98; HRMS (CI+): m/z calcd: 344.0056;
found: 344.0025 ([M+-O], 100).
The migration assay supports our NMR findings and points
to a different mechanism of interaction between 1 and cell-
membrane cysteines, compared to the well-characterized inter-
action between pCMBS and cell-membrane cysteines. Organo-
tellurane 1 was further established as a potential non-toxic and
anti-metastatic TeIV-based drug for the treatment of metastatic
melanoma.
Bis(4-methoxyphenyl)-l4-tellanediyl diacetate (3): Organotellur-
ane 2 (0.14 g, 0.38 mmol) was dissolved in CHCl3 (5.5 mL) for 1 h.
Then, Ac2O (0.043 g, 0.42 mmol) was added, thereby resulting in
a clear mixture, and the reaction mixture was stirred at RT for
90 min. The solvent and unreacted Ac2O were evaporated under
vacuum, and the crude was crystallized from hexane/chloroform
Experimental Section
1
(3:1) to afford purified crystals of the product in 65% yield. H NMR
General procedures: Solvents were of high purity. Anhydrous sol-
vents were used as received. Commercial compounds were used
without further purification. All synthetic compounds have been
reported earlier.[19–22] 1H NMR and 13C NMR spectra were obtained
on Bruker DPX-300, Avance-400, and DMX-600 spectrometers.
125Te NMR spectra were obtained on a Bruker DMX-600; chemical
shifts are reported in ppm relative to diphenyl telluride as an inter-
nal reference. Mass spectra were recorded in CI, ESI, and MALDI-
TOF modes with methane as the reagent gas. The purities of all
synthetic compounds were >95% as determined by elemental
analyses on a FlashEA 1112 analyzer (Thermo Fisher Scientific). ICP
analyses were made in an ULTIMA 2 spectrometer (Jobin Yvon
Horiba); classical calibration with standard solutions was used to
analyze tellurium.
(CDCl3,): d=7.76 (d, J=8.8 Hz, 4H; H-3), 7.00 (d, J=8.8 Hz, 4H; H-
2), 3.84 (s, 6H; CH3O), 1.96 ppm (s, 6H; CH3CO); 13C NMR (CDCl3):
d=177.5 (q, CO2), 161.9 (q, C1), 134.9 (CH, C3), 125.7 (q, C4), 115.3
(CH, C2), 55.4 (CH3O), 22.4 ppm (CH3CO); 125Te NMR ([D6]DMSO):
d=983.8 ppm; elemental analysis calcd (%) for C18H20O6Te : C 47.00,
H 4.38; found: C 47.04, H 4.31; HRMS (MALDI): m/z calcd: 403.0189;
found: 403.023 ([M+- CH3COOÀ]).
Bis(4-methoxyphenyl)tellurane (4): Organotellurane 1 (0.2 g,
0.49 mmol) was dissolved in acetone (10 mL), and the mixture was
added to a suspension of sodium ascorbate (0.2 g, 1 mmol) in
MeOH/H2O (8:2, 10 mL), which was already pre-stirred for 30 min.
The resulting clear mixture was stirred for 27 h. Then, the product
was extracted with CH2Cl2. The organic phase was dried with CaCl2,
vacuum filtered, and evaporated to provide the highly pure prod-
Cells: Murine melanoma B16F10 cells were cultured at 378C in
DMEM containing glucose (4.5 gLÀ1) and FCS (10%) under CO2
(5%). The cells were used for up to 20 passages.
1
uct in 93% yield. H NMR (CDCl3): d=7.61 (d, J=8.7 Hz, 4H; H-3),
6.74 (d, J=9.0 Hz, 4H; H-2), 3.74 ppm (s, 6H; CH3); 13C NMR
(CDCl3): d=159.8 (q, C1), 139.8 (CH, C3), 115.5 (CH, C2), 104.4 (q,
C4), 55.2 ppm (CH3); 125Te NMR ([D6]DMSO): d=639.7 ppm; elemen-
tal analysis calcd (%) for C14H14O2Te : C 49.19, H 4.13; found: C
50.38, H 4.51; HRMS (CI+): m/z calcd: 342.0038; found: 342.0037
([M+]).
Mice: Male C57Bl/6 mice (8–12 weeks) were purchased from
Harlan Laboratories (Jerusalem, Israel). Mice were kept in a specific
pathogen-free environment and were fed a standard pellet diet
and tap water. Mice were allowed to acclimate for seven days
before the experiments. Animal experiments were performed in ac-
cordance with confirmed institutional protocol and were approved
by the Institutional Animal Care and Use Committee (researcher li-
cense number: BS 20A2000).
NMR studies
General procedure for the reaction of organotelluranes with l-cys-
teine and N-acetylcysteine: Each organotellurane (0.025 mmol) was
dissolved in [D6]DMSO (1.2 mL) and transferred into a 10 mm NMR
tube for recording its 125Te NMR spectrum. Next, l-cysteine HCl
monohydrate (0.05 mmol) or N-acetylcysteine (0.05 mmol) was
added to the test tube, which was shaken for a few seconds. The
reaction products were followed by 125Te and 13C NMR.
Chemistry
Synthesis of organotellurium compounds: Dichlorobis(4-methoxy-
phenyl)-l4-tellane (1): TeCl4 (5.28 g, 19.6 mmol) was placed in
a 50 mL three-neck flame-dried flask. Anisole (12.74 g, 117.6 mmol)
was added, and the mixture was stirred for 6 h at reflux (1608C)
under argon. The mixture was cooled to RT, and the solvent was
evaporated to dryness under vacuum. Next, boiled acetonitrile
(45 mL) was added, and the mixture was filtered. The filtrate was
Biology
In vitro assays
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
925
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Adhesion to fibronectin, VCAM-1, and MadCAM-1: A 96-well plate
was coated with FN, VCAM-1, or MadCAM-1 (80 mL, 5 mgmLÀ1) or
with BSA (2%) and left overnight. Cells (105 in 200 mL serum-free
DMEM) were incubated with various concentrations of organotellu-
rium for 1 h. Next, the wells were washed twice in PBS to remove
unattached cells, then serum-free DMEM (200 mL) and XTT mixture
(50 mL, Biological Industries, Israel) was added to each well and in-
cubated for approximately 4–6 h. Plates were read at 450 nm for
the quantification of live attached cells.
group). Freshly prepared 1 was made each week, and the concen-
tration of the tellurium was determined by ICP analysis. The mice
were injected IP with 1 for three consecutive days and then every
other day. After 18 days, the mice were sacrificed and subjected to
liver analysis.
Statistical analysis: Data are expressed as meanÆSD. Unless other-
wise stated, all experiments were performed at least three times
on different days. Differences between groups in the adhesion
assays, relative to control, were analyzed by a Student’s t-test. Dif-
ferences between groups in the migration assay were analyzed by
one-way ANOVA. Differences in average weight between groups
and control in the acute toxicity model were analyzed with a Stu-
dent’s t-test. Differences between groups for in vivo metastases
were analyzed by one-way ANOVA. p<0.05 was considered statisti-
cally significant.
XTT cytotoxicity assay: Cells (105 in serum-free DMEM (200 mL)) were
incubated in a 96-well plate with 1 (6 mm), DMSO (1%), or cyclo-
heximide (5 mgmLÀ1). XTT (50 mL) was added to each well. Plates
were read after 6 h at 450 nm for the quantification of live cells.
FACS analysis: B16F10 cells were cultured to 60% confluence. The
medium was replaced, and cells were incubated with either 1 (2.4
or 6 mm) or DMSO (1%) for 1 h, or overnight with etoposide
(40 mm). Floating and trypsinized cells were collected, centrifuged
at 394g and suspended in PBS (1 mL) with PI (1 mL). Cell death was
measured with an LSRII flow cytometer (BD Biosciences).
Acknowledgement
This study was partially supported by the Raoul Wallenberg Chair
for Immunological Chemistry.
Cell migration. Cells (2”105) were incubated for 1 h in the presence
of either mobilized FN (13.2 mgmLÀ1) or BSA (2%), with MnCl2
(0.25 mm) and 1 (1.2–6 mm) or pCMBS (300 mm). The cells were
washed twice with PBS, and equal amounts of cells were loaded
onto 8 mm polycarbonate membrane inserts. The bottom chambers
were filled with DMEM (800 mL) containing FBS (20%) and FN
(13.2 mgmLÀ1) serving as a chemo attractant. Migrated cells were
quantified after 24 h by Trypan Blue exclusion test using a hemacy-
tometer.
The authors declare no conflict of interest.
Keywords: cancer · inhibitors · integrins · labile ligands ·
organotelluranes · VLA-4
[1] E. C. McGary, D. C. Lev, M. Bar-Eli, Cancer Biol. Ther. 2002, 1, 459–465.
[2] C. Chothia, E. Y. Jones, Annu. Rev. Biochem. 1997, 66, 823–862.
[3] S. Johansson, G. Svineng, K. Wennerberg, A. Armulik, L. Lohikangas,
Front. Biosci. 1997, 2, 126–146.
[4] M. A. Arnaout, B. Mahalingam, J. P. Xiong, Annu. Rev. Cell Dev. Biol. 2005,
21, 381–410.
[5] A. Chigaev, G. J. Zwartz, T. Buranda, B. S. Edwards, E. R. Prossnitz, L. A.
Sklar, J. Biol. Chem. 2004, 279, 32435–32443.
[6] B. Sredni, Semin. Cancer Biol. 2012, 22, 60–69.
[7] B. Sredni, R. R. Caspi, A. Klein, Y. Kalechman, Y. Danziger, M. Ben Ya’akov,
T. Tamari, F. Shalit, M. Albeck, Nature 1987, 330, 173–176.
[8] A. Layani-Bazar, I. Skornick, A. Berrebi, M. H. Pauker, E. Noy, A. Silberman,
M. Albeck, D. L. Longo, Y. Kalechman, B. Sredni, Cancer Res. 2014, 74,
3092–3103.
[9] M. Valcµrcel, V. Carrascal, O. Crende, F. Vidal-Vanaclocha, J. Invest. Der-
matol. 2014, 134, 470–480.
[10] M. Klemke, T. Weschenfelder, M. H. Konstandin, Y. Samstag, J. Cell. Physi-
ol. 2007, 212, 368–374.
Zimography: A 24-well plate was coated with FN. Cells (2.5”105)
were incubated for 72 h in DMEM containing BSA (0.1%) on top of
the FN-coated wells. Next, the upper fluid was collected, concen-
trated in Centricon tubes (>30 kDa; Merck Millipore), and tested
for protein concentration by a Bradford assay. For every sample,
the same amount of protein was loaded on a zimography gel (Bio-
Rad, Israel) containing gelatin (10%). After electrophoresis (30 mA,
1.5 h), the gels were washed and incubated for 30 min with Triton
X-100 (2.5%), washed, and incubated in a development buffer to
initiate enzyme activity for 30 min. The latter was replaced with a
fresh development buffer, and the gels were incubated overnight.
Then, the gels were stained with Coomassie Blue for 45 min and
washed in a destaining solution (acetic acid/ methanol/water;
5:10:85) until the appearance of transparent white bands. All re-
agents were purchased from Bio-Rad.
[11] S. Yosef, M. Brodsky, B. Sredni, A. Albeck, M. Albeck, ChemMedChem
2007, 2, 1601–1606.
In vivo activity
[12] A. Albeck, H. Weitman, B. Sredni, M. Albeck, Inorg. Chem. 1998, 37,
1704–1712.
[13] F. S. Pessoto, P. A. Faria, R. L. O. R. Cunha, J. V. Comasseto, T. Rodrigues,
I. L. Nantes, Chem. Res. Toxicol. 2007, 20, 1453–1461.
[14] C. B. C. Lima, W. W. Arrais-Silva, R. L. O. R. Cunha, S. Giorgio, Korean J.
Parasitol. 2009, 47, 213–218.
[15] D. S. Persike, R. L. O. R. Cunha, L. Juliano, I. R. Silva, F. E. Rosim, T. Vignoli,
F. Dona, E. A. Cavalheiro, M. J. da Silva Fernandes, Neurobiol. Dis. 2008,
31, 120–126.
[16] R. L. O R. Cunha, M. E. Urano, J. R. Chagas, P. C. Almeida, C. Bincoletto,
I. L. S. Tersariol, J. V. Comasseto, Bioorg. Med. Chem. Lett. 2005, 15, 755–
760.
[17] R. L. O. R. Cunha, I. E. GouvÞa, G. P. V. Feitosa, M. F. M. Alves, D. Brçmme,
J. V. Comasseto, I. L. S. Tersariol, L. Juliano, Biol. Chem. 2009, 390, 1205–
1212.
[18] T. S. Abondanza, C. R. Oliveira, C. M. V. Barbosa, F. E. G. Pereira, R. L. O. R.
Cunha, A. C. F. Caires, J. V. Comasseto, M. L. S. Queiroz, M. C. Valadares,
C. Bincoletto, Food Chem. Toxicol. 2008, 46, 2540–2545.
[19] J. Bergman, Tetrahedron 1972, 28, 3323–3331.
Acute toxicity studies: Organotellurane 1 (0.9–5.4 mgkgÀ1 in 200 mL
PBS containing DMSO (1%)) was injected IP into C57Bl/6 mice (8–
12 weeks; N=10) every other day, for a total of three weeks. PBS
containing DMSO (1%) served as the vehicle and as a control. The
mice were monitored for abnormal behavior or death, and each
group was measured for average weight once a week. After 3
weeks, the dose of the compound was raised and injected into
a new group of C57Bl/6. For each dose, freshly prepared solutions
of 1 (dissolved in PBS containing DMSO (1%)) were made every
week, and the concentration of 1 was calculated based on ICP
analysis.
Experimental liver metastases model: B16F10 cells (5.5”105) were
injected through the tail vein of C57Bl/6 mice (8–12 weeks old),
which were then randomly divided into four groups (N=10–14).
Each group was treated with a different dose of 1 (0.25, 0.9, or
1.8 mgkgÀ1) or with DMSO (1%) in PBS (vehicle and control
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
926
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[20] J. Beckmann, D. Dakternieks, A. Duthie, F. Ribot, M. Schürmann, N. A.
Lewcenko, Organometallics 2003, 22, 3257–3261.
[29] R. L. O. R. Cunha, I. E. Gouvea, L. Juliano, An. Acad. Bras. Cienc. 2009, 81,
393–407.
[21] L. Engman, J. Persson, Synth. Commun. 1993, 23, 445–448.
[22] S. Tamagaki, I. Hatanaka, S. Kozuka, Bull. Chem. Soc. Japan 1977, 50,
2501–2502.
[23] S. Schmidt, P. Friedl, Cell Tissue Res. 2010, 339, 83–92.
[24] N. Manickam, S. S. Ahmad, D. W. Essex, J. Thromb. Haemostasis 2011, 9,
1207–1215.
[30] S. Liang, C. Dong, Am. J. Physiol. Cell Physiol. 2008, 295, C701–C707.
[31] S. R. Elshaw, K. Sisley, N. Cross, A. K. Murray, S. M. MacNeil, M. Wagner,
C. E. Nichols, I. G. Rennie, Br. J. Ophthalmol. 2001, 85, 732–738.
[32] B. Bauvois, Biochim. Biophys. Acta Rev. Cancer 2012, 1825, 29–36.
[33] T. Turpeenniemi-Hujanen, Biochimie 2005, 87, 287–297.
[34] A. Barth, L. A. Wanek, D. L. Morton, J. Am. Coll. Surg. 1995, 181, 193–
201.
[25] J. S. Desgrosellier, D. A. Cheresh, Nat. Rev. Cancer 2010, 10, 9–22.
[26] S. Mohanty, L. Xu, J. Vis. Exp. 2010, 42, 1942.
[27] R. Xu, Y. Kalechman, M. Albeck, H. Kung, B. Sredni, Int. J. Oncol. 1996, 9,
319–325.
[28] Y. Kalechman, U. Gafter, T. Weinstein, A. Chagnac, I. Freidkin, A. Tobar,
M. Albeck, B. Sredni, J. Biol. Chem. 2004, 279, 24724–24732.
Manuscript received: November 16, 2015
Accepted article published: March 15, 2016
Final article published: March 25, 2016
ChemBioChem 2016, 17, 918 – 927
www.chembiochem.org
927
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim