Optimization of the antitumor efficacy of a synthetic mitochondrial toxin by increasing the residence time in the cytosol

Article abstract of DOI:10.1021/jm9008339

Plasma membrane drug efflux pumps of the multidrug resistance associated protein (MRP) family blunt the effectiveness of anticancer drugs and are often associated with drug resistance. GSAO, a tripeptide trivalent arsenical that targets a key mitochondria

Full text of DOI:10.1021/jm9008339

J. Med. Chem. 2009, 52, 6209–6216 6209
DOI: 10.1021/jm9008339
Optimization of the Antitumor Efficacy of a Synthetic Mitochondrial Toxin by Increasing the Residence
Time in the Cytosol
†
†,
Pierre J. Dilda, Stephanie Decollogne, Lakmini Weerakoon,†, Murray D. Norris,‡ Michelle Haber,‡ John D. Allen,§ and
ꢀ
Philip J. Hogg*,†,‡
†UNSW Cancer Research Centre, University of New South Wales, Sydney 2052, Australia, ‡Children’s Cancer Institute Australia for Medical
Research, Randwick NSW 2031, Australia, and The Centenary Institute of Cancer Medicine and Cell Biology, Newtown NSW 2042, Australia.
Both authors contributed equally to this work.
§
Received June 9, 2009
Plasma membrane drug efflux pumps of the multidrug resistance associated protein (MRP) family blunt
the effectiveness of anticancer drugs and are often associated with drug resistance. GSAO, a tripeptide
trivalent arsenical that targets a key mitochondrial transporter in angiogenic endothelial cells, is an
example of a compound whose efficacy is limited by tumor cell expression of MRP isoforms 1 and 2.
A cysteine mimetic analogue of GSAO was made, PENAO, which accumulates in cells 85 times faster
than GSAO due to increased rate of entry and decreased rate of export via MRP1/2. The faster rate
of accumulation of PENAO corresponds to a 44-fold increase in antiproliferative activity in vitro and
∼20-fold better antitumor efficacy in vivo. This information could be used to improve the efficacy of
other small molecule cancer therapeutics.
Introduction
ANT is the most abundant protein of the inner mitochon-
drial membrane.5 It is a 30 kDa protein that spans the
membrane six times and has two primary functions. The
transporter plays an important role in oxidative phosphoryla-
tion by exchanging newly formed ATP in the matrix with
spent ADP located in the inner membrane space across the
inner membrane. It is also a component of the mitochondrial
permeability transition pore.6
Mitochondria are an attractive cancer drug target
because their proper functioning is fundamental to cell survi-
val.1 4-(N-(S-Glutathionylacetyl)amino)phenylarsonous acid
(GSAOa) represents a novel class of organoarsenical com-
pound that selectively targets a key mitochondrial transporter
in tumor endothelial cells.2-4 It is currently being tested in a
phase I/IIa clinical trial in cancer patients.
The transition pore allows passage of molecules smaller
than1500 Da acrossthe inner mitochondrial membrane and is
induced by matrix Ca2þ overload, particularly when accom-
panied by oxidative stress, elevated phosphate concentration,
and depletionof adenine nucleotides. Transition pore opening
is thought to be triggered by Ca2þ-dependent binding of
cyclophilin-D to ANT.6 The trivalent arsenical of GCAO or
CAO reacts with two cysteine thiols located on the peptide
loops of ANT that protrude into the mitochondrial matrix,
forming a stable cyclic dithioarsinite complex in which both
sulfur atoms are bound to the arsenic atom.2,7 Calcium
content within the mitochondrial matrix increases several-
fold in proliferating cells,8 which sensitizes proliferating cells
to CAO-induced pore formation.2 Formation of the pore
coupled with the higher respiration rate in proliferating cells
and associated elevated levels of O2- are likely responsible for
the proliferation arrest observed in endothelial cells.2
The cell-surface cleavage of GSAO by γ-glutamyl trans-
peptidase is the rate-limiting step in its antimitochondrial
action. Our aim was to improve the efficacy of GSAO by
making analogues that have enhanced rate of accumulation in
the cytosol. A cysteine mimetic analogue of CAO, 4-(N-(S-
penicillaminylacetyl)amino) phenylarsonous acid (PENAO),
was made that bypasses the extracellular and cytosolic proces-
sing of GSAO. PENAO accumulates in cells much more
rapidly than GSAO, which translates to more potent effects
GSAO is processed at the cell surface and in the cytosol
beforethe trivalent arsenical moietyreactswithmitochondria.
The γ-glutamyl residue of GSAO is cleaved at the cell surface
by γ-glutamyl transpeptidase and the product of this reaction,
4-(N-(S-cysteinylglycylacetyl)amino)phenylarsonous acid
(GCAO), is transported across the plasma membrane by an
organic ion transporter where it may be further processed by
dipeptidases to 4-(N-(S-cysteinylacetyl)amino)phenylarson-
ous acid (CAO).4 Cytosolic concentrations of GCAO and
CAO are balanced by export from the cell by the multidrug
resistance associated protein isoforms 1 and 2 (MRP1/2)3
and rate of reaction with the mitochondrial inner mem-
brane transporter, adenine nucleotide translocase (ANT)
(Figure 1A).
*To whom correspondence should be addressed. Phone: 61-2-9385-
1004. Fax: 61-2-9385-1389. E-mail: p.hogg@unsw.edu.au.
a Abbreviations: ANT, adenine nucleotide translocase; BRAA, 4-(2-
bromoacetylamino)benzenearsonic acid; BRAO, 4-(2-bromoacetyl-
amino)benzenearsonous acid; CAO, 4-(N-(S-cysteinylacetyl)amino)-
phenylarsonous acid; DIDS, 4,40-diisothiocyanostilbene-2,20-disulfonic
acid; GCAO, 4-(N-(S-cysteinylglycylacetyl)amino)phenylarsonous
acid; GSAO, 4-(N-(S-glutathionylacetyl)amino)phenylarsonous acid;
MRP, multidrug resistance associated protein; OATP, organic anion
transporting polypeptide; PENAA, 4-(N-(S-penicillaminylacetyl)-
amino)phenylarsonic acid; PENAO, 4-(N-(S-penicillaminylacetyl)-
amino)phenylarsonous acid.
r
2009 American Chemical Society
Published on Web 09/29/2009
pubs.acs.org/jmc

6210 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 20
Dilda et al.
Figure 1. Mechanism of action of GSAO. (A) GSAO is processed by γ-GT at the cell surface, and the resulting GCAO is transported across the
plasma membrane by an OATP. GCAO is likely further processed to CAO in the cytosol before it reacts with ANT of the innermitochondrial
membrane. The cytosolic levels of GCAO and CAO are controlled by export from the cell by MRP isoforms 1 and 2. (B) Summary of the
synthesis of PENAO. (C) C18 reverse phase HPLC analysis of PENAO. Purity of PENAO by peak area is 97%.
on endothelial and tumor cells in culture and markedly better
antitumor efficacy in mice.
than GCAO, and an 85-fold faster rate than GSAO
(Figure 2A). Cellular concentration of the arsenicals was
measured from the arsenic content by inductively coupled
plasma spectrometry.
Results
Synthesis and Characterization of PENAO. The synthesis
of PENAO is summarized in Figure 1B. Briefly, 4-(2-bro-
moacetylamino)benzenearsonic acid (BRAA) was synthe-
sized from p-arsanilic and bromoacetyl bromide and reduced
to 4-(2-bromoacetylamino)benzenearsonous acid (BRAO)
using sodium sulphite with sodium iodide as catalyst.9
PENAO was produced by coupling BRAO and D-penicilla-
mine and purified by C18 silica gel chromatography. Purity
by HPLC was 97% (Figure 1C).
The organic anion transporting polypeptide (OATP) fa-
mily of transporters mediate uptake of glutathione-S-con-
jugates into cells,10 including GCAO.4 4,40-Diisothiocyano-
stilbene-2,20-disulfonic acid (DIDS),10 an inhibitor of
OATP’s, reduced PENAO uptake (Figure 2B), which implies
that this transporter is involved in PENAO translocation
across the plasma membrane.
The plasma membrane MRP isoforms 1 and 2 mediate
export of GCAO and CAO from endothelial cells.3,4 To
determine if PENAO is also a substrate for MRP’s, mam-
malian cells overexpressing MRP1, 2, 3, 4, 5, or 6, or MDR1
or BCRP were tested for resistance to PENAO. MRP1,
PENAO Accumulates More Rapidly in Endothelial Cells
than GSAO or Its Metabolites. PENAO accumulated in BAE
cells at a 10-fold faster rate than CAO, a 9-fold faster rate

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 20 6211
Figure 2. PENAO accumulates more rapidly in endothelial cells than GSAO or its metabolites. (A) BAE cells were incubated with 50 μM
GSAO, GCAO, CAO, or PENAO for up to 4 h, and cytosolic arsenic was measured by inductively coupled plasma spectrometry. The rates of
accumulation of GSAO, GCAO, CAO, and PENAO are 0.4 ( 0.1, 3.8 ( 0.1, 3.3 ( 0.1, and 34 ( 0.6 pmol As atoms per 106 cells per min. Data
points are the mean ( SD of three determinations, and the solid line is the linear least-squares fit of the data. The results are representative
of two experiments. (B) BAE cells were pretreated or not with the OATP inhibitor DIDS (500 μM) for 30 min and then incubated for 2 h with
20 μM PENAO. Cytosolic arsenic was measured by inductively coupled plasma spectrometry. (C) BAE cells were incubated for 2 h with 50 μM
GCAO, CAO, or PENAO in the absence or presence of 5 μM of the MRP1/2 inhibitor Reversan and cytosolic arsenic was measured by
inductively coupled plasma spectrometry. The data is expressed as the % increase in accumulation in the presence of Reversan. The values are
the mean ( SD of triplicate determinations and are representative of two experiments. ** is p < 0.01 and * is p < 0.05.
Table 1. Resistance of Mammalian Cells Overexpressing Different
Drug Transporters to PENAOa
MRP2 or MRP3 was overexpressed in the canine kid-
ney epithelial MDCKII cell line,11,12 MRP4 or MRP5 in
the human embryonic kidney HEK293 cell line,13 while
MRP6, MDR1, or BCRP was overexpressed in the murine
embryo fibroblast MEF3.8 line.3 The IC50 (concentra-
tion of compound that inhibits proliferation by 50%) for
PENAO was determined for each transfected cell line and
the value divided by the IC50 for proliferation arrest of
nontransfected parental cells to give the resistance factor
(Table 1). The resistance factor for MRP1 was 3.7, while
for MRP2 it was 4.6. The resistance factors for MRP3,
MRP4, MRP5, MRP6, MDR1, and BCRP were in the
range 0.4-1.3. These results indicate that PENAO, like
GCAO and CAO, is exported from cells by MRP isoforms
1 and 2.
cell type/drug transporter
resistance factor
MDCKII/MRP1
MDCKII/MRP2
MDCKII/MRP3
HEK293/MRP4
HEK293/MRP5
MEF3.8/MRP6
MEF3.8/MDR1
MEF3.8/BCRP
3.7
4.6
0.8
0.5
0.4
1.3
1.2
1.1
a Resistance factor is the ratio of the PENAO IC50 value for prolif-
eration arrest of transfected cells and the IC50 value for non-transfected
parental cells.
the cytosol via MRP1/2 using the inhibitors Reversan or
(E)-3-[[[3-[2-(7-chloro-2-quinolinyl)ethenyl]phenyl]-[[3-dime-
thylamino)-3-oxopropyl]thio]methyl]thio]-propanoic acid
(MK571)15 (Figure 3C) correlated with decreased or in-
creased antiproliferative action, respectively. The inhibitors
alone at the concentrations used had no effect on BAE cell
proliferation (data not shown). There is a strong correlation
between the antiproliferative activity of GSAO, GCAO,
CAO, and PENAO and the rate of cellular accumulation
of the arsenicals (Figure 3D). These results indicate that the
antiproliferative efficacy of the arsenicals is proportional to
their cytosolic levels.
Blocking de novo synthesis of glutathione in BAE cells
with buthionine sulfoximine (BSO), an inhibitor of γ-gluta-
myl cysteine synthase, enhanced GSAO’s antiproliferative
activity by almost 100-fold.3 Similarly, treating BAE cells
with BSO enhanced the PENAO IC50 for proliferation arrest
by 28-fold (Figure 3E). IC50 values for proliferation arrest
decreased from 440 ( 130 to 16 ( 10 nM in the presence of
100 μM BSO. BSO alone at this concentration had no effect
on cell growth (data not shown).
To confirm the role of MRP1/2 in regulating cytosolic
levels of PENAO, endothelial cells were incubated for 2 h
with 50 μM GCAO, CAO, or PENAO in the absence or
presence of 5 μM of the MRP1/2 inhibitor, 5,7-diphenyl-
pyrazolo[1,5-a]pyrimidine-3-carboxylic acid (3-morpholin-
4-ylpropyl)amide (Reversan),14 and cellular arsenic was
measured by inductively coupled plasma spectrometry.
MRP1/2 inhibition increased the cellular levels of GCAO
by 197 ( 31%, CAO levels by 114 ( 18%, and PENAO levels
by only 52 ( 5% (Figure 2C). This result indicates that
export of PENAO from the cell by MRP1/2 is less efficient
than for GCAO or CAO.
The Rate of Accumulation of PENAO in Endothelial Cells
Correlates with Antiproliferative Activity. The faster rate of
accumulation of PENAO in endothelial cells corresponded
to a 44-fold increased antiproliferative activity compared to
GSAO, an 8-fold increased antiproliferative activity com-
pared to GCAO, and a 6-fold increased antiproliferative
activity compared to CAO (Figure 3A, 24 h assay). The
marked time-dependence of the antiproliferative effect of
GSAO is due to the requirement for processing of GSAO at
the cell surface before it can react with mitochondrial ANT.
There was a small decrease in the GCAO, CAO, and
PENAO IC50 with time of incubation but not nearly to the
same extent as for GSAO.
Inhibiting MRP1/2 activity and blocking de novo synth-
esis of glutathione in endothelial cells resulted in additive
enhancement of PENAO’s antiproliferative activity. The
IC50 value for proliferation arrest decreased from 1100 to
130 nM in the presence of 3 μM Reversan, to 70 nM in the
presence of 20 μM BSO, and to 20 nM in the presence of
both 3 μM Reversan and 20 μM BSO (Figure 3F). Reversan
Blocking entry of PENAO into endothelial cells with the
OATP inhibitor DIDS (Figure 3B) or blocking export from

6212 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 20
Dilda et al.
Figure 3. Rate of accumulation of PENAO in endothelial cells correlates with antiproliferative activity. (A) Comparison of the GSAO,
GCAO, CAO, and PENAO IC50 (concentration of compound that inhibits proliferation by 50%) values for proliferation arrest of endothelial
cells. BAE cells were incubated with 0.8-250 μM of the compounds for 24, 48, or 72 h and cell viability determined using MTT. Values are mean
( SD of triplicate determinations. The results are representative of three experiments. (B) BAE cells were untreated or pretreated with 500 μM
DIDS for 30 min and then incubated with 3 μM PENAO for 24 h. Cell viability was determined using MTT and results expressed as percentage
of control. Values are mean ( SD of triplicate determinations. Results are representative of two experiments. (C) BAE cells were incubated for
30 min with 15 μM MK571 or 2 μM Reversan prior to incubation with 0.3 μM PENAO for 72 h. Cell viability was determined using MTT and
results expressed as percentage of control. The values are the mean ( SD of triplicate determinations. (D) Correlation between the IC50 for
proliferation arrest of endothelial cells and rate of cellular accumulation of GSAO, GCAO, CAO, and PENAO. (E) BAE cells were treated with
0.001-3 μM PENAO and the indicated concentrations of BSO for 72 h and the IC50 for proliferation arrest was calculated. The data points and
error bars are the mean ( SD from two experiments performed in triplicate. (F) BAE cells were untreated or pretreated with 20 μM BSO and/or
3 μM Reversan for 30 min and then incubated with 0-3 μM PENAO for 24 h. Cell viability was determined using MTT and results expressed as
percentage of control. Values are mean ( SD of triplicate determinations. *** is p < 0.001 and ** is p < 0.01.
and/or BSO alone at these concentrations had no effect on
cell growth (data not shown).
was human pancreatic carcinoma BxPC-3 cells. The IC50 for
proliferation arrest of BxPC-3 cells by PENAO is 19-fold
higher than for endothelial cells and 27-fold higher for
GSAO.
The Antimitochondrial Action of PENAO is Comparable to
that of the GSAO Metabolite, CAO. Like GSAO, GCAO,
and CAO,4 PENAO triggered swelling of isolated rat liver
mitochondria in a time- and concentration-dependent man-
ner (Figure 4A). Comparison of the time for 25% maximal
swelling as a function CAO or PENAO concentration
indicates that the two compounds have comparable effects
on mitochondrial integrity (Figure 4B).
The trivalent arsenical moiety of PENAO was responsible
for its antiproliferative activity as the corresponding penta-
valent arsenical (4-(N-(S-penicillaminylacetyl)amino)phenyl-
arsonic acid, PENAA) had no effect on BAE cell prolifera-
tion (see Supporting Information).
Prolonged incubation of endothelial cells with GSAO
results in apoptosis of the cells.2 Incubation of BAE cells
with PENAO also resulted in loss of viability, with an IC50 of
∼3.5 μM in a 48 h assay (see Supporting Information). This
IC50 is 21-fold lower than for GSAO.2
PENAO is a More Effective Inhibitor of Proliferating Cells
than GSAO. PENAO is a 5- to 17-fold better inhibitor
than GSAO of proliferation of both primary and trans-
formed cells in 72 h assays (Table 2). The antiproliferative
activity for bovine, human, and murine endothelial
cells, canine epithelial cells, and seven different tumor
cell lines is within a factor of 19 for PENAO and a factor
of 27 for GSAO. The most resistant tumor cell line tested
PENAO is ∼20-Fold More Efficacious than GSAO at
Inhibiting Growth of Human Pancreatic Carcinoma Tumors
in Mice. BalbC nude mice bearing subcutaneous human
BxPC-3 pancreatic carcinoma tumors in the proximal mid-
line were implanted with micro-osmotic alzet pumps sub-
cutaneously in the flank. The pumps delivered vehicle
control or 0.25 or 0.5 mg/kg/day PENAO. The tumor doubl-
ing times were 9.4 ( 1.4, 15.6 ( 2.1, and 13.8 ( 2.1 days for

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 20 6213
Table 2. Comparison of PENAO vs GSAO IC50 Values for Proliferation Arrest of Endothelial and Tumor Cells. a
cell type
cell line
PENAO IC50, μM
GSAO IC50, μM
bovine aortic endothelial
BAE
1.1 ( 0.4
2.0 ( 0.3
2.5 ( 0.2
2.4 ( 0.2
1.8 ( 0.4
4.1 ( 0.1
5.3 ( 0.1
5.9 ( 0.5
6.4 ( 0.1
9.4 ( 0.5
21 ( 0.7
10 ( 1.3
23 ( 1.0
42 ( 2.1
13 ( 2.7
14 ( 0.8
25 ( 4.8
41 ( 3.5
43 ( 5.3
43 ( 6.2
51 ( 1.5
270 ( 54
human microvascular endothelial
murine cerebral cortex endothelial
canine kidney epithelial
HMEC-1
bEnd.3
MDCKII
K562
human chronic myelogenous leukemia
human fibrosarcoma
mouse lung carcinoma
HT1080
LLC
human colorectal carcinoma
human pancreatic carcinoma
human mammary carcinoma
human pancreatic carcinoma
HCT1116
PANC-1
MCF-7
BxPC-3
a IC50 values were determined from a 72 h proliferation assay.
Figure 4. Antimitochondrial action of PENAO is comparable to that of the GSAO metabolite, CAO. (A) Mitochondrial swelling induced by
0-300 μM PENAO as measured by decrease in light scattering at 520 nm over 60 min. The traces are representative of two experiments
performed in duplicate on two different mitochondrial preparations. (B) Time for 25% maximum swelling as a function of CAO or PENAO
concentration. The results are representative of two experiments.
groups treated with vehicle or 0.25 or 0.5 mg/kg/day
PENAO, respectively (Figure 5A).
mitochondrial matrix where it reacts with ANT, triggering
proliferation arrest. We reasoned that CAO would be a more
effective mitochondrial toxin than GSAO because it would
enter the cell more rapidly and would bypass the dipeptidase
processing in the cytoplasm.
There were no obvious signs or symptoms of toxicity of the
compounds at the doses tested. There was a small accumula-
tion of connective tissue at the delivery site of the osmotic
pump in 7 of the 10 mice at the higher dose of PENAO. This
is likely the result of a local inflammatory response triggered
by the high concentrations of PENAO at the delivery site.
Immunohistochemical analysis of the vehicle and 0.25 mg/
kg/day PENAO tumors following 18 days of treatment
indicated a significant reduction in both blood vessel density
in the treated tumors (p < 0.01, Figure 5B) and the prolif-
erative index of the tumor cells (p < 0.05, Figure 5C). These
findings indicate that PENAO has both antiangiogenic and
antitumor cell activity in vivo.
Dose-dependence and comparison of the antitumor effi-
cacy of GSAO versus PENAO is shown in Figure 5D.
PENAO is ∼20-fold more efficacious that GSAO at inhibit-
ing growth of human pancreatic carcinoma tumors in im-
munodeficient mice. Comparable antitumor activity was
observed with administration of 5 mg/kg/day GSAO and
0.25 mg/kg/day PENAO.
In our hands, CAO was difficult to produce synthetically,
mostly because of incompatibility with the solvents used for
GSAO production. We also observed that CAO was some-
what unstable in aqueous solutions at 4 or 37 °C when
compared to GSAO. We instead chose to make an analogue
of CAO, PENAO, in which the cysteine residue has been
replaced with the cysteine mimetic, D-penicillamine. Penicilla-
mine contains two methyl groups in place of the hydrogens
at the β-carbon atom of cysteine. PENAO was produced
to a purity of 97% by HPLC. The compound was stable in
phosphate-buffered saline for at least 2 weeks at 4 or 37 °C.
PENAO accumulated in endothelial cells at an 85-fold
faster rate than GSAO and a ∼10-fold faster rate than GCAO
or CAO. Like for GCAO and CAO,4 PENAO uptake was
mediated, at least in part, by the OATP and export was
via MRP1 and MRP2. The faster rate of accumulation of
PENAO compared to CAO is mostly due to decreased rate of
export by MRP1/2. It is not known why the two additional
methyl groups in PENAO would influence its handling by
MRP1/2. Presumably, the cysteine pendant of CAO binds
with higher affinity to MRP1/2 than the cysteine mimetic of
PENAO. PENAO and CAO had comparable effects on the
integrity of isolated mitochondria, which implies that rate of
entry into the mitochondrial matrix and reactivity with ANT
is similar for the two compounds.
Discussion and Conclusions
The rate-limiting step in GSAO’s antimitochondrial action
is cleavage at the cell surface by γ-glutamyl transpeptidase.
The resulting GCAO then enters the cell via OATP and is
converted by dipeptidases in the cytoplasm to CAO. CAO is
either exported from the cytosol by MRP1/2 or enters the

6214 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 20
Dilda et al.
Figure 5. PENAO is ∼20-fold more efficacious that GSAO at inhibiting growth of human pancreatic carcinoma tumors in mice. (A) BxPC-3
tumors were established in the proximal midline of female 7-9 week old BalbC nude mice. Mice bearing ∼50 mm3 tumors were randomized
into three groups (n = 10) and implanted with 28 day alzet micro-osmotic pumps subcutaneously in the flank. The pumps delivered vehicle
alone or 0.25 or 0.5 mg/kg/day PENAO. The data points are the mean ( SE of the tumor volumes. (B) and (C) Tumors were excised at day 18 in
mice treated with 0.25 mg/kg/day PENAO (see part A), fixed, sectioned, and analyzed for vascularity (CD31 hotspots, (B)) and tumor cell
proliferation (PCNA, (C)). ** is p < 0.01 and * is p < 0.05. (D) Comparison of the antitumor efficacy of PENAO and GSAO. BalbC nude mice
bearing ∼50 mm3 subcutaneous BxPC-3 tumors in the proximal midline were treated for 18 days with the indicated doses of either GSAO (white
bars) or PENAO (black bars). The compounds were administered subcutaneously via micro-osmotic pumps implanted in the flank. There were
5 and 7 mice, respectively, for the 0.25 and 5 mg/kg/day GSAO groups and 2 and 7 mice, respectively, for the 0.025 and 0.25 mg/kg/day PENAO
groups. The data points are the mean ( SE of the treatment/control values expressed as %. The dotted line represents no change in tumor
growth rate.
The 85-fold faster rate of accumulation of PENAO in
endothelial cells compared to GSAO corresponded to a
44-fold increased antiproliferative activity, while the ∼10-fold
faster rate of accumulation compared to GCAO or CAO
equated to ∼7-fold increased antiproliferative activity. There
is a marked time-dependence of the antiproliferative effect of
GSAO, which is due to the requirement for cell surface
processing of GSAO by γ-glutamyl transpeptidase before it
can exert its cellular effect.4 There was little difference in
the PENAO IC50 value for proliferation arrest in 24 or 72 h
assays, however. This finding is consistent with the direct
mitochondrial action of PENAO. It is neither processed at the
cell surface nor in the cytoplasm before it reacts with mito-
chondrial ANT.
The antiproliferative activity of PENAO is dependent on
cellular glutathione levels as it is for GSAO and GCAO.3
Depletion of endothelial cell glutathione reduced the PENAO
IC50 for proliferation arrest from 440 to 16 nM. This strong
correlation between glutathione level and antiproliferative
activity likely relates to the mechanism of export of PENAO.
MRP1 transports trivalent arsenic as a complex with three
molecules of glutathione.16 The arsenic atom of PENAO,
therefore, is probably glutathionylated before the conjugate
is exported by MRP1/2. Low cellular glutathione would
translate to impaired export and, therefore, increased anti-
proliferative activity.
The relative antiproliferative activity of PENAO and
GSAO for 11 different primary or transformed cell types in
culture is within a factor of 19 for PENAO and a factor of 27
for GSAO. This finding indicates that PENAO has similar
cellular selectivity as GSAO.
PENAO is a ∼20-fold more effective inhibitor of tumor
growth in mice than GSAO. Comparable antitumor activity
was observed with administration of 0.25 mg/kg/day PENAO
and 5 mg/kg/day GSAO. No systemic toxicity of either
PENAO or GSAO was apparent at these doses. GSAO-
treated tumors are characterized by a significant reduction
in blood vessel density and no effect on the tumor cell
proliferative index,2 which is consistent with an antiangiogen-
esis mechanism of action. Analysis of PENAO treated tumors
showed a marked reduction in vascular hot spots (p < 0.01) as
well as significant inhibition of BxPC-3 tumor cell prolifera-
tion (p < 0.05). This can be explained by the difference in the
effects of the two compounds on proliferation of BxPC-3cells.
The IC50 for proliferation arrest of BxPC-3 cells is 270 μM for
GSAO2 compared to 21 μM for PENAO.
In summary, the cytotoxic and antitumor efficacy of
GSAO has been markedly improved by constructing an

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 20 6215
analogue of a GSAO metabolite that enters cells faster and is
exported slower. The increased residence time of PENAO in
the cytosol correlates with its increased mitochondrial toxi-
city. PENAO is one of the most potent antitumor agents in
mice that has been reported and is well tolerated at efficacious
doses.
cultured in RPMI medium. HMEC-1 cells were cultured in
MCDB131 medium (Gibco, Gaithersburg, MD) supplemented
with 10 ng per mL EGF and 1 μg per mL hydrocortisone. The
cells were supplemented with 10% fetal calf serum (FBS), 2 mM
L-glutamine, and 10 units per mL penicillin and 10 μg per mL
streptomycin. Cell culture plasticware was from Techno Plastic
Products (Trasadingen, Switzerland). All other cell culture
reagents were from Gibco.
Accumulation of the Arsenicals in BAE Cells. BAE cells were
seeded in 6-well plates and allowed to attach overnight. The
medium was replaced, and the cells were incubated for 30 min in
the absence or presence of transport inhibitors and then with
50 μM GSAO, GCAO, CAO, or PENAO for up to 4 h.
Cytosolic arsenic measured by inductively coupled plasma
spectrometry.4
Drug Transporter Transfectants. Canine kidney epithe-
lial MDCKII cells stably overexpressing MRP1, MRP2, or
MRP3,11,12 human embryonic kidney HEK293 cells stably
overexpressing MRP4 and MRP5,13 and murine embryo fibro-
blast MEF3.8 cells stably overexpressing MRP6, MDR1, or
BCRP3 were used as described. Cells were grown and main-
tained as adherent monolayers in DMEM containing 10% calf
serum (Cosmic, Hyclone, Tauranga, New Zealand), 10 units per
mL penicillin, and 10 μg per mL streptomycin. Cytotoxicity
assays were performed as described previously.17
Cell Proliferation Assays. BAE, NB4, K562, MDCKII,
HT1080, LLC, HCT116, PANC-1, MCF-7, and BxPC-3 cells
were seeded at a density of 1.5 ꢀ 103, 3 ꢀ 103, 4 ꢀ 103, 5 ꢀ 102,
2 ꢀ 103, 5 ꢀ 102, 5 ꢀ 102, 6 ꢀ 103, 6 ꢀ 103, and 1 ꢀ 104 cells per
well, respectively, into 96-well plates. Cells were allowed to
adhere overnight at 37 °C in a 5% CO2, 95% air atmosphere.
Cells were untreated or pretreated with transporter inhibitors
and then cultured for 24, 48, or 72 h in medium containing 10%
fetal calf serum and PENAO. See figure legends for reagent
concentrations and times of incubation. Viable cells were de-
termined by incubating cells with the tetrazolium salt MTT
(Sigma, St. Louis, MO), which is metabolized by viable cells to
form insoluble purple formazan crystals. DMSO was added to
lyse cells, the contents of the wells were homogenized, and the
absorbance at 550 nm measured. Cell number in the untreated
control was normalized as 100%, and viable cell number for all
treatments was expressed as percentage of control.
Experimental Section
Synthesis and Purification of PENAO. An aqueous solution
of sodium carbonate (36.6 g, 345 mmol, 250 mL) was added to
a suspension of p-arsanilic acid (50.0 g, 230 mmol) in water
(250 mL) and cooled to 0 °C. Bromoacetyl bromide (92.6 g,
460 mmol) was added over a period of 10 min to the cooled
solution. The reaction was maintained for 1 h below 20 °C. The
precipitated product, BRAA, was collected by filtration and
washed with water. The wet cake was added to a mixture of
methanol (375 mL), 47% aqueous hydrobromic acid (375 mL),
and sodium iodide (5.14 g, 34.5 mmol), followed by dropwise
addition of a solution of sodium sulphite (86.9 g, 690 mmol) in
water (325 mL) over 15 min. The heterogeneous mixture was
maintained at room temperature for 2 h and filtered. The filter
cake was slurried in a mixture of methanol, 47% aqueous
hydrobromic acid, and water (375:150:225 mL), followed by
water (375 mL) and 5% aqueous sodium bicarbonate solution
(375 mL). The product was filtered and dried under vacuum at
1
45 °C for 12 h giving 54 g of BRAO (72.8% yield). H NMR
(CF3COOD): δ 4.14 (s, 2H), 7.80 (d, J = 7.8 Hz, 2H), 7.95 (d,
J = 7.6 Hz, 2H), 11.5 (s, 1H). 13C NMR (CF3COOD): δ 28.47,
123.69, 133.26, 134.07, 171.4.
To a freshly prepared solution of sodium (10.7 g 465 mmol)
in methanol (500 mL) was added D-penicillamine (25.53 g,
171-mmol) under argon atmosphere. The reaction mixture
stirred for 5 min. BRAO (50 g, 155 mmol) was added, stirred
for 10 min, and pH was adjusted to 3-5 using 5% sulfuric acid in
methanol and filtered. The clear filtrate was taken and dry
distilled under vacuum to give crude PENAO (190 g, 125%
yield). The crude PENAO was further purified by C18 silica gel
chromatography using water as eluent. The pure fractions were
collected and distilled completely to give 30 g of purified
material. The obtained solid was dissolved in methanol
(90 mL) and added to a solution of acetone and water mixture
(600 mL:45 mL) and stirred for 14 h. The precipitated white
product was filtered and dried at 40 °C for 4 h to give 18.1 g
(30%) of PENAO. 1H NMR (D2O): δ1.40 (s, 3H), 1.61 (s, 3H),
3.61 (d, J = 6.34 Hz, 2H), 3.72 (s, 1H), 7.58 (d, J = 7.81 Hz, 2H),
7.74 (d, j = 7.81 Hz, 2H). 13C NMR (D2O): δ 25.93, 29.68, 35.91,
49.58, 64.10, 124.02, 132.77, 141.55, 146.86, 173.44 ppm.
PENAO purity was characterized by HPLC (1200 Series;
Agilent Technologies, Santa Clara, CA) on a Zorbax Eclipse
XDB-C18 column (4.6 mm ꢀ 150 mm, 5 μm; Agilent Tech-
nologies) using a mobile phase of acetonitrile-water (25:75 vol/
vol), flow rate of 0.5 mL per min, and detection by absorbance at
256 nm. Purity of PENAO by peak area was 97%.
Mitochondrial Swelling Assay. Mitochondrial permeability
transition induction was assessed by light scattering using
isolated rat liver mitochondria.4
Primary Tumor Growth Assays. Female 7-9 week old Balb C
nude were used (UNSW Biological Resource Centre). Mice were
held in groups of 3-5 at a 12 h day and night cycle and were
given animal chow and water ad libitum. A suspension of 3 ꢀ 106
BxPC-3 cells in 0.1 mL of PBS was injected subcutaneously in
the proximal midline. Tumors were allowed to establish and
grow to a size of ∼50 mm3, after which they were randomized
into groups of 2-10. Animals were implanted with 28 day alzet
model 1004 micro-osmotic pumps (ALZA Corporation, Palo
Alto, CA) subcutaneously in the flank. The pumps delivered
vehicle, 0.025, 0.25, or 0.5 mg/kg/day PENAO or 0.25 or 5 mg/
kg/day GSAO. Tumor volume and animal weight was measured
every 2 or 3 days. Tumor volume was calculated using the
relationship length ꢀ height ꢀ width ꢀ 0.523. Tumor doubling
time (TD) was calculated from the tumor growth rate curve
during exponential growth using the formula TD = 0.693/
ln(VF/VI)/ID, where VF is final tumor volume, VI is initial tumor
volume, and ID the interval in days.18
GSAO, GCAO, and CAO were prepared as previously
described.2,4 All arsenical compounds were dissolved in phos-
phate-buffered saline containing 100 mM glycine, and concen-
trations were determined by titrating with dimercaptopropanol
and calculating the remaining free thiols with 5,50-dithiobis-
(2-nitrobenzoic acid). The solutions were stored at 4 °C in the
dark until use. There was no significant loss in the active
concentration of stock solutions of the arsenicals for at least a
week when stored under these conditions.
Cell Culture. Bovine aortic endothelial (BAE) cells were from
Cell Applications, San Diego, CA, and BxPC-3, HT1080, LLC,
PANC-1, MCF-7, HCT116, bEnd.3, HMEC-1, and K562 cells
were from ATCC, Bethesda, MD. MDCKII cells were from
P. Borst (The Netherlands Cancer Institute, Amsterdam). BAE,
HT1080, PANC-1, MCF-7, HCT116, MDCKII, bEnd.3, and
LLC cells were cultured in DMEM. K562 and BxPC-3 cells were
Immunohistochemistry. Blood vessel density was estimated by
immunostaining sectioned tumors for CD31 (rat antimouse
CD31; BD Pharmingen, Australia). Tumor cell proliferation
was estimated by immunostaining for PCNA (mouse antihu-
man PC10; Dako, Carpinteria, CA). Excised tumors were fixed
in formalin, embedded in paraffin, and 4 μm sections were cut
and mounted on Superfrost slides. Sections were deparaffinized

6216 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 20
Dilda et al.
in xylene, rehydrated in PBS using an ethanol gradient, and
microwaved for 5 min to increase antigen retrieval. Sections
were incubated with anti-CD31 at 1 in 50 and with anti-PCNA
at 1 in 100 dilutions for 60 min at 37 °C and secondary
peroxidase-conjugated antibodies for 30 min at 37 °C. Micro-
vessels were counted in each of the five most vascularized,
separately located areas or hot spots.19 Tumor cell proliferative
index was determined by counting PCNA-positive cells in 100
tumor cells in three random fields in each tumor tissue. Magni-
fication was 400ꢀ.
(5) Halestrap, A. P.; McStay, G. P.; Clarke, S. J. The permeability
transition pore complex: another view. Biochimie 2002, 84, 153–
166.
(6) Halestrap, A. P. What is the mitochondrial permeability transition
pore? J. Mol. Cell. Cardiol. 2009, 46, 821–831.
(7) McStay, G. P.; Clarke, S. J.; Halestrap, A. P. Role of critical thiol
groups on the matrix surface of the adenine nucleotide translocase
in the mechanism of the mitochondrial permeability transition
pore. Biochem. J. 2002, 367, 541–548.
(8) Murgia, M.; Giorgi, C.; Pinton, P.; Rizzuto, R. Controlling
metabolism and cell death: at the heart of mitochondrial calcium
signalling. J. Mol. Cell. Cardiol. 2009, 46, 781–788.
(9) Adams, E.; Jeter, D.; Cordes, A. W.; Kolis, J. W. Chemistry of
organometalloid complexes with potential antidotes: structure of
an organoarsenic(III) dithiolate ring. Inorg. Chem. 1990, 29, 1500–
1503.
(10) Kobayashi, D.; Nozawa, T.; Imai, K.; Nezu, J.; Tsuji, A.; Tamai, I.
Involvement of human organic anion transporting polypep-
tide OATP-B (SLC21A9) in pH-dependent transport across
intestinal apical membrane. J. Pharmacol. Exp. Ther. 2003, 306,
703–708.
Statistical Analyses. Results are presented as means ( SD,
except for the results shown in Figure 5 where SE are used. All
variables were examined for normality and homogeneity of
variance. Normality of the distributions was assessed using the
Shapiro-Wilk test. Results of blood vessel density and tumor
cell proliferative index were analyzed using one-way ANOVA.
Posthoc tests, parametric, or nonparametric as appropriate
were used to compare mean values. All analyses were performed
using GraphPad Prism (GraphPad, San Diego, CA). All tests of
statistical significance were two-sided, and p values <0.05 were
considered statistically significant.
(11) Kool, M.; van der Linden, M.; de Haas, M.; Scheffer, G. L.;
de Vree, J. M.; Smith, A. J.; Jansen, G.; Peters, G. J.; Ponne, N.;
Scheper, R. J.; Elferink, R. P.; Baas, F.; Borst, P. MRP3, an organic
anion transporter able to transport anti-cancer drugs. Proc. Natl.
Acad. Sci. U.S.A 1999, 96, 6914–6919.
(12) Evers, R.; Kool, M.; Smith, A. J.; van Deemter, L.; de Haas, M.;
Borst, P. Inhibitory effect of the reversal agents V-104, GF120918
and Pluronic L61 on MDR1 Pgp-, MRP1- and MRP2-mediated
transport. Br. J. Cancer 2000, 83, 366–374.
(13) Wielinga, P. R.; Reid, G.; Challa, E. E.; van der Heijden, I.; van
Deemter, L.; de Haas, M.; Mol, C.; Kuil, A. J.; Groeneveld, E.;
Schuetz, J. D.; Brouwer, C.; De Abreu, R. A.; Wijnholds, J.;
Beijnen, J. H.; Borst, P. Thiopurine metabolism and identification
of the thiopurine metabolites transported by MRP4 and MRP5
overexpressed in human embryonic kidney cells. Mol. Pharmacol.
2002, 62, 1321–1331.
Acknowledgment. We are grateful to Rabeya Akter from
the UNSW Analytical Centre for inductively coupled plasma
spectrometry. We also thank Shane Thomas for his assistance
with HPLC. The technical assistance of Diana Lau is appre-
ciated. This study was supported by grants from the National
Health and Medical Research Council of Australia, the
Cancer Council NSW, and the Cancer Institute NSW. Pierre
J. Dilda is a Research Fellow of the Cancer Council NSW.
(14) Burkhart, C. A.; Watt, F.; Murray, J.; Pajic, M.; Prokvolit, A.;
Xue, C.; Flemming, C.; Smith, J.; Purmal, A.; Isachenko, N.;
Komarov, P. G.; Gurova, K. V.; Sartorelli, A. C.; Marshall,
G. M.; Norris, M. D.; Gudkov, A. V.; Haber, M. Small-molecule
multidrug resistance-associated protein 1 inhibitor reversan in-
creases the therapeutic index of chemotherapy in mouse models
of neuroblastoma. Cancer Res. 2009, 69, 6573–6580.
(15) Jedlitschky, G.; Leier, I.; Buchholz, U.; Barnouin, K.; Kurz, G.;
Keppler, D. Transport of glutathione, glucuronate, and sulfate
conjugates by the MRP gene-encoded conjugate export pump.
Cancer Res. 1996, 56, 988–994.
(16) Leslie, E. M.; Haimeur, A.; Waalkes, M. P. Arsenic transport by
the human multidrug resistance protein 1 (MRP1/ABCC1). Evi-
dence that a tri-glutathione conjugate is required. J. Biol. Chem.
2004, 279, 32700–32708.
(17) Allen, J. D.; Brinkhuis, R. F.; Wijnholds, J.; Schinkel, A. H. The
mouse Bcrp1/Mxr/Abcp gene: amplification and overexpression in
cell lines selected for resistance to topotecan, mitoxantrone, or
doxorubicin. Cancer Res. 1999, 59, 4237–4241.
Supporting Information Available: Results are presented that
demonstrate that the trivalent arsenical moiety of PENAO is
responsible for its antiproliferative activity and that PENAO
also reduces the viability of endothelial cells, albeit at concen-
trations that are higher than those that trigger proliferation
arrest. This material is available free of charge via the Internet at
http://pubs.acs.org.
References
(1) Don, A. S.; Hogg, P. J. Mitochondria as cancer drug targets.
Trends Mol. Med. 2004, 10, 372–378.
(2) Don, A. S.; Kisker, O.; Dilda, P.; Donoghue, N.; Zhao, X.;
Decollogne, S.; Creighton, B.; Flynn, E.; Folkman, J.; Hogg,
P. J. A peptide trivalent arsenical inhibits tumor angiogenesis by
perturbing mitochondrial function in angiogenic endothelial cells.
Cancer Cell 2003, 3, 497–509.
(3) Dilda, P. J.; Don, A. S.; Tanabe, K. M.; Higgins, V. J.; Allen, J. D.;
Dawes, I. W.; Hogg, P. J. Mechanism of selectivity of an angiogene-
sis inhibitor from screening a genome-wide set of Saccharomyces
cerevisiae deletion strains. J. Natl. Cancer Inst. 2005, 97, 1539–1547.
(4) Dilda, P. J.; Ramsay, E. E.; Corti, A.; Pompella, A.; Hogg, P. J.
Metabolism of the tumor angiogenesis inhibitor 4-(N-(S-glutathio-
nylacetyl)amino)phenylarsonous acid. J. Biol. Chem. 2008, 283,
35428–35434.
(18) Wolff, N. C.; Randle, D. E.; Egorin, M. J.; Minna, J. D.; Ilaria,
R. L., Jr. Imatinib mesylate efficiently achieves therapeutic intra-
tumor concentrations in vivo but has limited activity in a xenograft
model of small cell lung cancer. Clin. Cancer Res. 2004, 10, 3528–
3534.
(19) Weidner, N. Current pathologic methods for measuring intratu-
moral microvessel density within breast carcinoma and other solid
tumors. Breast Cancer Res. Treat. 1995, 36, 169–180.