PABA/NO lead optimization: Improved targeting of cytotoxicity to glutathione S-transferase P1-overexpressing cancer cells

Article abstract of DOI:10.1016/j.bmc.2015.05.020

PABA/NO [O²-{2,4-dinitro-5-[4-(N-methylamino)benzoyloxy]phenyl} 1-(N,N-dimethylamino) diazen-1-ium-1,2-diolate] is a nitric oxide (NO)-releasing arylating agent designed to be selectively activated by reaction with glutathione (GSH) on catalysis by glutathione S-transferase P1 (GSTP1), an enzyme frequently overexpressed in cancer cells. PABA/NO has proven active in several cancer models in vitro and in vivo, but its tendency to be metabolized via a variety of pathways, some that generate inactive metabolites and hydrolysis products, limits its potential as a drug. Here we show that a simple replacement of cyano for nitro at the 4 position to give compound 4b ('p-cyano-PABA/NO') has the dual effect of slowing the undesired side reactions while enhancing the proportion of NO release and arylating activity on catalysis by GSTP1. Compound 4b showed increased resistance to hydrolysis and uncatalyzed reaction with GSH, along with a more favorable product distribution in the presence of GSTP1. It also showed significant proapoptotic activity. The data suggest p-cyano-PABA/NO to be a more promising prodrug than PABA/NO, with better selectivity toward cancer cells.

Full text of DOI:10.1016/j.bmc.2015.05.020

Bioorganic & Medicinal Chemistry 23 (2015) 4980–4988
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
PABA/NO lead optimization: Improved targeting of cytotoxicity
to glutathione S-transferase P1-overexpressing cancer cells
a
b
b
c
b,
⇑
Youseung Kim , Anna E. Maciag , Zhao Cao , Jeffrey R. Deschamps , Joseph E. Saavedra
,
a
a,
⇑
Larry K. Keefer , Ryan J. Holland
a
Chemical Biology Laboratory, National Cancer Institute at Frederick, Frederick, MD 21702, United States
Chemical Biology Laboratory, Leidos Biomedical Research, Inc., Frederick National Laboratory for Cancer Research, Frederick, MD 21702, United States
Center for Biomolecular Science and Engineering, Naval Research Laboratory, Washington, DC 20375, United States
b
c
a r t i c l e i n f o
a b s t r a c t
2
Article history:
PABA/NO [O -{2,4-dinitro-5-[4-(N-methylamino)benzoyloxy]phenyl} 1-(N,N-dimethylamino) diazen-1-
ium-1,2-diolate] is a nitric oxide (NO)-releasing arylating agent designed to be selectively activated by
reaction with glutathione (GSH) on catalysis by glutathione S-transferase P1 (GSTP1), an enzyme fre-
quently overexpressed in cancer cells. PABA/NO has proven active in several cancer models in vitro
and in vivo, but its tendency to be metabolized via a variety of pathways, some that generate inactive
metabolites and hydrolysis products, limits its potential as a drug. Here we show that a simple replace-
ment of cyano for nitro at the 4 position to give compound 4b (‘p-cyano-PABA/NO’) has the dual effect of
slowing the undesired side reactions while enhancing the proportion of NO release and arylating activity
on catalysis by GSTP1. Compound 4b showed increased resistance to hydrolysis and uncatalyzed reaction
with GSH, along with a more favorable product distribution in the presence of GSTP1. It also showed sig-
nificant proapoptotic activity. The data suggest p-cyano-PABA/NO to be a more promising prodrug than
PABA/NO, with better selectivity toward cancer cells.
Received 11 March 2015
Revised 29 April 2015
Accepted 10 May 2015
Available online 19 May 2015
Keywords:
Glutathione S-transferase P1 (GSTP1)
Glutathione
Nitric oxide (NO)
PABA/NO
Diazeniumdiolate
Arylation
Oxidative stress
Antiproliferative
Cytotoxicity
Published by Elsevier Ltd.
Apoptosis
1
. Introduction
PABA/NO, 1 (structure in Scheme 1), is an arylated diazenium-
known as glutathione S-transferases (GSTs). GSTs catalyze the con-
jugation of GSH to a broad variety of reactive electrophiles.7
Among the various GSTs, GSTP1 has attracted attention because
8
diolate prodrug designed to be activated in cancer cells overex-
pressing glutathione S-transferase P1 (GSTP1).1 The mechanism
of activation, illustrated in Scheme 1, involves an initial reaction
with a nucleophile, typically glutathione (GSH), to form the
Meisenheimer complex. Its metabolism irreversibly arylates cel-
lular GSH, exposing a diazeniumdiolate anion (3) that sponta-
neously releases up to two molar equivalents of nitric oxide
it is often overexpressed in cancers. Elevated cellular levels of
,2
7
GSTP1 are correlated with poor prognosis and resistance to
9
–11
diverse common anticancer drugs.
PABA/NO was designed to
utilize the increased detoxification ability of cancer cells and sup-
pression of apoptotic signaling through GSTP1-overexpression by
generating cytotoxic metabolites selectively in those cells.
PABA/NO has shown promising anticancer activity in preclinical
studies. In A2780 human ovarian cancer xenografts, PABA/NO
exhibited tumoristatic activity with potency comparable to that
1
,2
(
NO). The consumption of GSH through arylation and subsequent
oxidation by reactive oxygen/nitrogen species (ROS/RNS) signifi-
cantly alters the cellular redox balance, in cancer cells with a high
endogenous oxidative burden, triggering a myriad of pro-death
cellular responses.
Though this metabolism occurs spontaneously in the presence
of GSH, its rate is catalyzed by a family of detoxifying enzymes
2
of cisplatin. PABA/NO has also shown a strong growth-inhibitory
effect in U87 glioma cells in vitro as a monotherapy, and strong
synergistic toxicity when treated concomitantly with temozolo-
3
–5
1
2
mide. Evaluation of in vivo efficacy in an intracranial U87 rat
xenograft revealed intratumoral inhibition of cell proliferation
and the induction of apoptosis, unfortunately without significant
1
2
prolongation of survival.
⇑
Corresponding authors. Tel.: +1 301 846 6002; fax: +1 301 846 5946 (J.E.S.); tel.:
Furthermore, modifying the p-methylaminobenzoate (PABA)
+
1 301 846 6491; fax: +1 301 846 5946 (R.J.H.).
substituent of PABA/NO has generated analogs with lower IC50 val-
E-mail addresses: saavedjo@mail.nih.gov (J.E. Saavedra), hollandrj@mail.nih.gov
1
3
ues in leukemia cells.
(
R.J. Holland).
http://dx.doi.org/10.1016/j.bmc.2015.05.020
968-0896/Published by Elsevier Ltd.
0

Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
4981
purification with p-(methylamino)benzoate (7) to yield a mixture
of 4a and 4b (Scheme 3). These were successfully separated and
the major isomer was purified by recrystallization from hex-
ane/ethyl acetate solution.
The minor isomer from this procedure was independently pre-
pared in higher yield by changing the order of reactant addition,
that is, by reacting 5 first with 7 then with 3 (Scheme 3).
Separately, isomer mixture 6 was hydrolyzed to generate a pair
of isomeric phenols 8. These were separated by flash chromatogra-
phy and one was subjected to structure determination by X-ray
crystallography (Scheme 3, Fig. 1), confirming its structure as 8a
(a to reflect the fact that the cyano group is ortho to the diazenium-
diolate moiety, a nomenclature convention used throughout this
paper). This same isomer could also be produced by hydrolyzing
one of the isomers of 4, identifying it as being a member of the
‘a’ series, that is, 4a, cyano ortho to the diazeniumdiolate group.
These assignments specify the structures of all the other materials
in Scheme 3 including 4b.
Scheme 1. Original design principles:6 The activation of PABA/NO, a NO-releasing
arylating agent, by a S Ar reaction with GSH.
N
2
.2. Reactions of PABA/NO with GSH
Since then, other molecular modifications have been evaluated
aimed at improving both the potency and the selectivity of
PABA/NO. Most recently, Fu et al. have used oleanolic acid as a sub-
stituent to selectively target liver cancer cells and successfully hin-
der tumor growth, illustrated by a mouse H22 hepatoma cell
Previous studies1 have shown PABA/NO to readily hydrolyze
and to undergo nucleophilic attack; however, it is important to
note that in the case of the GSH reactions these prior studies were
conducted at 25 °C, and in some cases at lower pH and GSH con-
centrations than are recognized as physiological. Therefore, we
have reexamined the hydrolysis and thiolytic decomposition of
PABA/NO to determine the rates and products of each potential
metabolic pathway under conditions deemed closer to physiologi-
cal (4 mM GSH in a pH 7.4 phosphate buffer at 37 °C) using an
HPLC trapping method described in detail in the methods section.
The results are summarized in Scheme 4 and row 1 of Table 1.
PABA/NO readily hydrolyzed at 25 °C and 37 °C with half-lives
of 2 h and 20 min respectively, forming compounds 7 and 8. The
latter phenolic compound is fully ionized at physiological pH, ren-
dering it non-reactive to GSH attack and thus unable to release NO
or arylate cellular thiol groups, which are the key cytotoxic events
for arylated diazeniumdiolates. The relative instabilities at both
temperatures have detrimental implications for efficacy upon
administration and success of formulation pre-administration.
When 4 mM GSH was added to the buffer, PABA/NO underwent
nucleophilic attack at three carbon atoms, indicated by the colored
arrows in Scheme 4. Nucleophilic attack at the carbon ipso to the
diazeniumdiolate is the NO-releasing step, which is accessible
through the green and red pathways. Similarly to hydrolysis, thiol-
ysis of the ester bond (blue pathway) produced a significant
amount of the inactive phenolate 8. It is also worth noting that
both compounds 2 (product after NO release from 1 following
the red arrow) and 11 (product after aromatic substitution of
GSH following the green arrow) can undergo a second aromatic
substitution. If GSH is the second nucleophile, forming 12, then
the reaction consumes a total of two molar equivalents of GSH
per mole of drug, leading to greater stress on the cellular redox bal-
1
4
xenograft model.
The following report details a lead optimization of the PABA/NO
platform aimed at overcoming limitations we have encountered in
seeking to improve its profile as a lead anti-cancer drug. One such
limitation is its rapid hydrolysis rate, which competes with GSH
attack and produces an inactive phenolic species with the diazeni-
1
umdiolate functional group still intact, thus blocking NO release.
Another significant limitation of PABA/NO as a therapeutic agent
is the spontaneous GSH-dependent release of NO in the absence
of GSTP1. Because GSH is present in millimolar concentrations in
nearly all cell types, including red blood cells, there is an inherent
lack of control over the nonspecific release of NO. Hence, we have
sought to increase the stability of PABA/NO in aqueous media and
decrease the rates of the undesired reaction paths, thus increasing
the efficiency of PABA/NO activation in GSTP1-rich tumors.
Our specific hypothesis is that replacing nitro groups of
PABA/NO one at a time by the less electron-withdrawing cyano
group would improve the reactivity profile and accomplish at least
some of these goals. Accordingly, we chose 4a and 4b (Scheme 2)
as target molecules, and report herein on their improved reactivity.
2
2
. Results and discussion
.1. Synthesis and structure assignments
The two-step synthesis of 4b began by reacting difluoroarene 5
with diazeniumdiolate ion 3 to produce a pair of monofluoro iso-
mers whose structures are shown at the top center of
6
Scheme 3. This mixture was then reacted without further
5
ance. If a protein thiol is the second nucleophile, then a cross-link
of two peptides through the aromatic ring is achieved via a cova-
lent bond, leading to structural and hence behavioral modification
1
5
of the protein.
In addition to the deactivation of PABA/NO through the forma-
tion of compound 8, the metabolism of PABA/NO resulted in a com-
plex mixture of metabolites each requiring toxicological profiling,
thus introducing many concerns from a drug design perspective.
2
.3. Reactivity of 4a and 4b in the absence of enzymes
Scheme 2. Structures of target molecules sought in this lead optimization effort.
Nomenclature note: Compound numbers containing no letters indicate that they
refer to the PABA/NO (i.e., dinitro) series of compounds. Numbers followed by an a
or b refer to the analogous ortho and para monocyano derivatives, respectively.
To address these concerns, the reactivity of 4a and 4b were
examined in aqueous phosphate buffer, pH 7.4, with and without

4
982
Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
Scheme 3. Synthesis and structure assignments for target compounds 4b and 4a.
Figure 1. Structure of 8a as determined by X-ray crystallography. Only one of the
two molecules in the asymmetric unit is shown for clarity. Displacement ellipsoids
are at the 50% level.
GSH. Their rates of decomposition were characterized with the
same HPLC-based trapping method employed for PABA/NO. The
results for all three compounds are provided in Table 1.
The hydrolysis rates at 25 °C for both isomers of cyano-
PABA/NO (4b and 4a) indicated a remarkable increase in their sta-
bility compared to that of PABA/NO. Comparing the khydrolysis at
2
5 °C for the three compounds shows a modest 3-fold increase in
stability for 4a relative to PABA/NO, while the stability of 4b was
substantially increased close to 30-fold. Enhanced stability toward
hydrolysis was also observed at 37 °C, once again 4b demonstrat-
ing a greater stabilization than 4a when compared to the original
PABA/NO (Table 1).
An increase in stability of 4a and 4b when compared to
PABA/NO was also observed in their reaction with GSH, though
to a much lesser extent. Interestingly, 4a exhibited a greater
Scheme 4. Full product analysis of the reaction of PABA/NO (1) with 4 mM GSH in
.1 M phosphate buffer (pH 7.4).
0
stability in 4 mM GSH than 4b did, which is different from what
was observed in the hydrolysis reactions. Compound 4b showed
slightly greater than a two-fold increase in stability in 4 mM GSH
at 37 °C compared to PABA/NO, whereas 4a showed more than a

Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
4983
Table 1
Physicochemical comparison of 1, p-cyano-PABA/NO (4b), and a structural isomer of
o-cyano-PABA/NO (4a)
a
À1
b
À1
c
À1
Compound
k
hydrolysis 25 °C (s
)
k
hydrolysis 37 °C (s
)
k
GSH (s
)
À5
À4
À3
À3
À3
1
9.6 Â 10
3.3 Â 10
0.3 Â 10
5.7 Â 10
2.1 Â 10
1.1 Â 10
11.8 Â 10
2.1 Â 10
5.1 Â 10
À5
À5
À4
À4
4
4
a
b
a
b
c
First-order initial rate constants for hydrolysis at pH 7.4, 25 °C.
First-order initial rate constants for hydrolysis at pH 7.4, 37 °C.
Pseudo-first order rate constants for reaction with 4 mM GSH at pH 7.4, 37 °C.
five-fold better stability compared to PABA/NO under the same
conditions.
The products from each reaction were analyzed by LC–MS after
1
and 10 half-lives. In the case of 4a, the NO-releasing reaction was
encumbered but not eliminated (indicated by the red arrow in
Scheme 5). This leaves the thiolysis of the ester (indicated by the
blue arrow in Scheme 5) and the aromatic substitution at the p-
methylaminobenzoate (PABA) group as the dominant pathways,
neither of which resulted in the release of NO. It is significant to
note that 2a and 11a were resistant to a second aromatic substitu-
tion reaction, in contrast to 2 and 11 generated in the reaction of
PABA/NO with GSH.
In the case of 4b the aromatic substitution reactions at both the
diazeniumdiolate arrow and the PABA group were completely sup-
pressed. As a result, only the thiolysis reaction with GSH (blue
arrow) was observed, resulting in the quantitative conversion of
4
b to the inactive phenol 8b and the free PABA acid (Scheme 6).
2
.4. Reactivity in the presence of enzymes
The rates of reactivity and product distributions changed dra-
matically when glutathione S-transferase P1 was added to the
reaction mixture to a final concentration of 4 nM. Compounds 1,
Scheme 5. Products of the reaction of 1
l
M 4a with 4 mM GSH in 0.1 M phosphate
buffer (pH 7.4) at 37 °C. LC–MS trace taken after 30 s. Additional Nomenclature note:
All downstream metabolites labeled without letters refer to the PABA/NO (i.e.,
dinitro) series of compounds, while the a label refers to the o-cyano PABA/NO
metabolites and the b label to those of p-cyano PABA/NO. Accordingly, 2a is the
same as 2 except with a nitrile group ortho to the diazeniumdiolate substituent,
while 2b is the same as 2 except with a nitrile group para to the diazeniumdiolate
substituent.
4
b, and 4a were all fully decomposed in less than 5 s, which is
the fastest we could sample the reaction with the iodoacetamide
trapping HPLC method described in the Experimental section.
GSTP1 enhanced the aromatic substitution reaction at the diazeni-
umdiolate position of compound 1 (red pathway of Scheme 7).
Although this reaction is favored, the thiolysis of the ester group
of compound 1 (blue pathway of Scheme 7) was still slightly com-
petitive, resulting in the formation of a small amount of inactive
compound 8. Intermediate 2, generated from the S
underwent two competing uncatalyzed reactions with GSH, a sec-
ond S Ar reaction that displaced the PABA group, leaving com-
N
Ar reaction,
N
pounds 7 and 12, and thiolysis of the PABA ester to form
compounds 9 and 10. Thus the metabolism of compound 1 by
GSTP1 enhanced the NO-releasing mechanism but still resulted
in the formation of five distinct chemical entities, one of which is
known to be biologically inactive.
In contrast to compound 1, GSTP1 enhanced the S
at the PABA group of compound 4a, resulting in the formation of
1a (the green pathway of Scheme 8). After 5 min the distribution
N
Ar reaction
1
of reaction products remained unchanged for at least 30 min.
Compound 11a was still detected in large quantities, suggesting
stability to a second nucleophilic attack at the diazeniumdiolate
position, and thus poor nitric oxide release.
Much like the above-mentioned compounds, reaction products
of 4b and GSH were dramatically altered in the presence of GSTP1.
Unlike GSH alone, where only the inactivating thiolysis reaction
was observed (Scheme 6), in the presence of GSTP1, a quantitative
displacement of the diazeniumdiolate was observed producing 2b
Scheme 6. Products of the reaction of 1 lM 4b with 4 mM GSH in 0.1 M phosphate
buffer (pH 7.4) at 37 °C. LC–MS trace taken after 30 s.
(Scheme 9). This reaction was accompanied by a quantitative
release of NO (Fig. 2). The integration of the curve gave >1.9 mol
of NO/mol of compound. As a basis for comparison, note the poor
NO yield when GSTP1 was absent. Once the NO was released,
compound 2b decomposed through the uncatalyzed thiolysis of
the ester bond, forming 9 and 10b, which were stable for at least
30 min.

4
984
Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
Scheme 9. Product analysis of the reaction of 1
phosphate (pH 7.4) at 37 °C in the presence of 4 nM GSTP1. LC–MS trace taken after
A) 60 s or (B) 30 min of reaction.
lM 4b with 4 mM GSH in 0.1 M
(
Scheme 7. Products of the reaction of
1 lM PABA/NO (1) with 4 mM GSH
containing 4 nM GSTP1 in 0.1 M phosphate buffer (pH 7.4) at 37 °C: LC–MS trace
taken after (A) 60 s or (B) 30 min.
Figure 2. Nitric oxide release curves for the reactions of 1
lM 4b with 4 mM GSH in
0.1 M phosphate (pH 7.4) at 37 °C in the presence or absence of 4 nM GSTP1.
rate constants could not be determined. However, in all cases the
reactions with GSTA1 released less of the diazeniumdiolate than
the analogous GSTP1 reactions (Supporting information).
We evaluated which compound had the most favorable GSTA1
reaction profile by comparing the yields of 10(a,b) generated in
the red NO-releasing pathway of Scheme 10 to those of 8(a,b) pro-
duced in the concurrent blue NO-retaining pathway. As shown in
Table 2, the ratio of NO-releasing to NO-retaining reactions for
compound 1 was 3:1, whereas the corresponding ratios for both
4
a and 4b were 1:1. The data suggest that 4a and 4b will be less
apt than 1 to generate potentially toxic NO release in the liver,
where GSTA1 is particularly abundant, thus lessening possible neg-
ative effects on that organ.
Scheme 8. Product analysis of the reaction of 1
nM GSTP1 in 0.1 M phosphate (pH 7.4) at 37 °C. LC–MS trace taken after (A) 60 s
or (B) 30 min.
lM 4a with 4 mM GSH containing
4
2
.5. In vitro evaluation
The enzymatic decompositions of compounds 1, 4a, and 4b
were repeated with GSTA1 in place of GSTP1 (Scheme 10). The
enzymatic catalyses of the three prodrugs with GSTA1 occurred
very rapidly, as they did with GSTP1, such that the corresponding
Leukemia cells are known for their exceptional sensitivity to
NO. A number of diazeniumdiolate-based NO donors exhibited
potent anticancer activities against leukemia cells in vitro and

Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
4985
Scheme 10. Products of reacting 1
lM 1, 4a, or 4b with 4 mM GSH in 0.1 M phosphate (pH 7.4) at 37 °C in the presence of 4 nM GSTA1. The dashed boxes highlight the
compounds compared in Table 2.
Both 4a and 4b showed dose-dependent growth inhibitory and
Table 2
The ratio of compound 10 produced from the NO release pathway to compound 8
produced from the NO retained pathway
proapoptotic activities in leukemia cells. Compound 4b appears to
have antiproliferative activity comparable to that of PABA/NO, with
IC50 values of 9
tively. In comparison, 4a was slightly less potent, with IC50 values
of 20 M and 15 M, in HL-60 and U937, respectively.
lM and 6.5 lM, for HL-60 and U937 cells, respec-
Compound
X
Y
Ratio 10:8
1
NO
NO
CN
2
NO
CN
NO
2
2
3:1
1:1
1:1
l
l
4
4
a
b
2
Both 4a and 4b induced apoptosis in HL-60 cells, as illustrated
by activation of caspases 8 and 7 and cleavage of PARP (Fig. 3).
Twenty-four-hour incubation with 10 lM 4b resulted in a strong
apoptotic signal. Less potent 4a required slightly higher concentra-
tions to trigger apoptosis.
in vivo.16–18 In the present study, two human leukemia cell lines,
HL-60 and U937, were used to evaluate activities of 4a and 4b
in vitro. A Western blot analysis showed that both HL-60 and
U937 leukemia cell lines have an abundance of GSTP1 (Fig. S1),
but no detectable GSTA1 (not shown). This high expression level
may serve as a targeting mechanism, directing the cytotoxicity
toward cancer cells over the normal leukocytes that coexist in
the blood stream. To examine whether the enzymatic reactivity
analysis was an accurate mimetic for cellular metabolism, the
HL-60 leukemia cell line was treated with compound 4b and ana-
lyzed for metabolites as a function of time. In as early as 10 min, 4b
was absorbed and metabolized via the NO-releasing GSTP1 path-
way identified in the enzymatic reactions (Fig. S2). The quantita-
tive release of NO in the small volume of the cell should assure
that the intracellular NO reaches mid micromolar levels, which
are well within the cytotoxic range.19
4
a and 4b induced phosphorylation and activation of stress-
activated protein kinase/Jun-amino-terminal kinase (SAPK/JNK).
We have shown previously that treatment with arylated
diazeniumdiolates resulted in oxidative/nitrosative stress and
5
DNA damage. Both oxidative stress and DNA damage are widely
observed to activate a SAPK/JNK pathway, and JNK activation is
necessary for apoptosis induced by various stress stimuli in several
cell types. In non-stressed cells, GSTP1 acts as an endogenous
SAPK/JNK inhibitor, through protein-protein binding and
2
0,21
SAPK/JNK sequestration in the cytosol.
Oxidative stress condi-
tions or treatment with GSTP1 inhibitors lead to GSTP1 dimeriza-
tion and dissociation from the complex, allowing for SAPK/JNK
phosphorylation and triggering proapoptotic signaling. The role
of GSTP1 in resistance to anticancer drugs involves not only direct
Figure 3. Anti-proliferative and pro-apoptotic activities of 4a and 4b in leukemia cells. Left panel: HL-60 and U937 cells were treated with increasing concentrations of 4a
and 4b for 72 h. Right panel: PARP cleavage and caspase 8 and 7 activation as shown by Western blot in the HL-60 cells. F.L.—full length PARP or caspase-8. SAPK/JNK
phosphorylation indicates activation of this stress pathway in HL-60 cells treated with 4a and 4b.

4
986
Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
detoxification, but also blocking apoptosis through SAPK/JNK inhi-
bition. 4a and 4b are predicted to disrupt this GSTP1/JNK interac-
tion, releasing SAPK/JNK that becomes activated and signals
toward activation of apoptosis.
from Aldrich Chemical Co. (Milwaukee, WI) with the exception of
5, which was purchased from Synquest Laboratories, Inc.
(Alachua, FL). All of the proton and carbon NMR spectra were
obtained in chloroform-d, with the exception of those for 8a and
8
4
b, which were obtained in methanol-d . Chemical shifts (d) are
3
. Conclusion
reported in parts per million (ppm) downfield from tetramethylsi-
lane. All high-performance liquid chromatography (HPLC) solvents
were purchased from Fisher (Waltham, MA). Compound 1 was pre-
Here we show that a simple structural modification of PABA/NO
1
led to the development of derivatives with more favorable traits.
Our results lead us to choose 4b over 4a for further consideration
as an anticancer lead, based on the following observations. First,
it is significantly more stable toward hydrolysis (Table 1).
Second, it gives a cleaner product distribution on reaction with
pared as previously described. Sodium salt of 3 was prepared as
2
7
described earlier.
4.2. Synthesis
4
mM GSH—four products for 4a (Scheme 5) versus two for 4b
4.2.1. Synthesis of compounds 6a and 6b as a mixture
(
Scheme 6). The same ratio prevailed when catalytic amounts of
To a stirring solution of 1.4 g (7.6 mmol) of 5 dissolved in 20 mL
of tert-butanol and 2 mL of THF was added a solution of 1.1 g
(8.4 mmol) of 3 (sodium salt) dissolved in 18 mL of 5% sodium
bicarbonate. Precipitates of pale yellow were formed immediately
and this reaction was stirred for 6.5 h at room temperature. The
precipitates were isolated via vacuum filtration and dried under
high vacuum to afford 1.4 g (71% yield) of two isomers of 6 in
GSTP1 were added (four products from 4a and two from 4b,
Schemes 8 and 9, respectively). Of special note, compound 4a gives
off some NO in reactions with GSH alone while retaining most of
the NO in the form of compound 11a when GSTP1 is added.
Conversely compound 4b does not release a significant amount
of NO in the GSH alone reactions, but gives off quantitative yields
of NO in the presence of GSTP1. Compound 4b was also more
potent as an in vitro anti-leukemic.
1
3:1 ratio. Major isomer: H NMR d 3.30 (s, 6H), 7.38 (d, 1H,
1
3
J = 10.0 Hz), 8.39 (d, 1H, J = 6.8 Hz); C NMR d 42.00, 96.20,
105.69, 111.98, 132.22, 134.23, 155.48 (J = 12 Hz), 166.01
The basis for cancer cell selectivity. Our data further suggest that
b will experience a rather simple but advantageous metabolic fate
+
4
(J = 268 Hz). HRMS (ESI) m/z calculated for (M+H) (270.0633),
on intravenous administration. The compound should be stable in
serum that contains low levels of GSH. We could expect some
levels of uncatalyzed reaction with GSH in normal leukocytes or
neighboring tissues that contain abundant amounts of this peptide.
However, as shown in Scheme 6, only two products were observed
in this reaction, thioester 9 and phenol 8b, neither of which can
release NO or arylate nucleophiles, making them biologically inac-
tive and therefore harmless for these cells.
found (270.0625). Anal. Calcd for C H FO N Á0.25 H O: C, 39.50;
9
8
4
5
2
H, 3.13; F, 6.94; N, 25.59. Found: C, 39.40; H, 2.95; F, 6.98; N,
1
25.33. Minor isomer: H NMR d 3.31 (s, 6H), 7.26 (d, 1H,
1
3
J = 11.6 Hz), 8.47 (d, 1H, J = 7.6 Hz). C NMR d 41.98, 97.35,
105.42, 112.68, 132.82, 134.23, 159.52 (J = 275 Hz), 162.60
(J = 11 Hz).
4.2.2. Synthesis and isolation of compound 4a
A second anticipated route of elimination is reaction with GSH
under catalysis by GSTA1 in the liver, where that isoform is espe-
cially abundant. As shown in Scheme 10, this reaction produces
an unreactive phenol. While it also consumes GSH and generates
To a stirring solution of 1.5 g (10.0 mmol) of 7 dissolved in
10 mL of 1 M sodium hydroxide and 10 mL of distilled water was
added a solution of 1.8 g (10.0 mmol) of 5 dissolved in 25 mL of
tert-butanol and 5 mL of THF. Formation of precipitates was imme-
diately observed upon the addition. After stirring for 72 h at room
temperature, distilled water was added to crash out the precipi-
tates and they were collected by vacuum filtration, washed with
distilled water and dried under vacuum to afford 2.4 g (76% yield)
of yellow powder. A 112 mg (0.36 mmol) sample of this intermedi-
ate product (87% in purity) was then gradually added to a stirring
solution of 90 mg (0.7 mmol) of 3 (sodium salt) dissolved in 10 mL
of DMF at 0 °C. Allowing it to gradually warm to room temperature,
the reaction was stirred overnight. Saturated ammonium chloride
solution was then added and the reaction was cooled to form yel-
low precipitates which were vacuum filtered and dried to afford
41 mg (33% yield) of product: mp 148–150 °C; UV (acetonitrile)
N
NO through the S Ar reaction, it does so to a lesser extent than
its predecessor (compound 1). Interestingly, a related arylated
diazeniumdiolate (JS-K) slowed the growth of liver tumors in an
orthotopic rat model, suggesting that 4b may be promising in a
liver cancer setting as well. 22
The third predicted metabolic route is the one that occurs when
systemically delivered 4b reaches cancer cells overexpressing
GSTP1, such as in the case of the leukemia cells presented above.
Here, the metabolism is dominated by the NO-producing, GSH-
consuming pathway shown in Scheme 9. Increased expression of
GSTP1 has been observed in blasts from childhood acute lym-
2
3
phoblastic leukemia. Results obtained in relapsed non-Hodgkin
2
4
chronic lymphocytic leukemia25 or multiple
À1
À1
1
lymphoma,
kmax
(e) 314 nm (26.1 mM cm ); H NMR d 2.95 (s, 3H), 3.27
2
6
myeloma suggest an involvement of GSTP1 in drug resistance
in hematological malignancies.
(s, 6H), 4.43 (b, 1H), 6.62 (d, 2H, J = 8.2 Hz), 7.38 (s, 1H), 8.0 (d,
1
3
2H, J = 8.2 Hz), 8.46 (s, 1H);
C NMR d 30.35, 42.10, 98.02,
Taken together, our results indicate that 4b represents a sub-
stantial improvement over 1 in our lead optimization effort.
Considering especially the crucial role of selectivity for toxic attack
on GSTP1-overexpressing tumor cells while sparing their normal
counterparts, the former cells are expected to consume much more
GSH and generate much more NO on metabolism of 4b than on
metabolism of 1.
111.68, 112.46, 113.27, 115.02, 132.05, 133.34, 137.44, 150.58,
154.60, 161.56, 163.68. HRMS (ESI) m/z calculated for (M+H)+
(401.1204), found (401.1202). Anal. Calcd for C17H16N O : C,
6
6
51.00; H, 4.03; N, 20.99. Found: C, 51.13; H, 4.04; N, 20.81.
4.2.3. Synthesis and isolation of compound 4b
A solution of 0.15 g of 7 (1.0 mmol) dissolved in 1.0 mL of 1.0 M
lithium tert-butoxide (1.0 mmol) was cooled to 0 °C. To this was
added the mixture of the two isomers of 6 prepared above, still
in a 3:1 ratio (0.27 g), dissolved in DMF (4 mL). Gradually warming
to room temperature, this reaction was stirred for 48 h. Afterwards,
the reaction was cooled to 0 °C again and stirred for an additional
2 h. Then saturated ammonium chloride solution was added to
form bright yellow-orange precipitates, which were isolated by
4
4
. Experimental
.1. General
NO was purchased from Matheson Gas Products
(
Mongomeryville, PA). Other starting materials were purchased

Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
4987
vacuum filtration and washed with deionized water. Finally these
precipitates were dried overnight to afford 378 mg (95% yield) of
b and 4a in a 3:1 ratio. 4b was purified from the mixture through
recrystallization in 1:1 hexane/ethyl acetate: mp 168–169 °C; UV
the preincubated solution. At each time point 5
etamide solution was added to 95
rapidly cooled on ice, and analyzed by the protocol described
above. Pseudo-first-order rate constants were obtained from the
chromatographic data extending over 8 half-lives.
l
L of an iodoac-
l
L of the reaction mixture,
4
À1
1
(
acetonitrile) kmax (e) 318 nm (19.4 mM–1 cm ); H NMR d 2.95
(
d, 3H), 3.27 (s, 6H), 4.47 (b, 1H), 6.63 (d, 2H, J = 8.4 Hz), 7.83 (s,
1
1
1
H), 8.06 (d, 2H, J = 8.4 Hz), 8.40 (s, 1H); 13C NMR d 30.38, 42.17,
4.5. Effect of GSTP1- and GSTA1-catalysis on product
distribution for PABA/NO, 4a, and 4b
01.28, 111.71, 111.82, 113.83, 114.79, 131.47, 133.43, 154.38,
+
54.71, 157.83, 163.15. HRMS (ESI) m/z calculated for (M+H)
(
H
401.1204), found (401.1201). Anal. Calcd for C17
H
16
N
6
O
6
Á1/3
The procedure above was repeated with the addition of GSTP1
or GSTA1 (final concentration 4 nM) which was incubated at
2
O: C, 50.25; H, 4.13; N, 20.68. Found: C, 50.29; H, 4.24; N, 20.06.
3
7 °C in the GSH-containing phosphate buffered solution for at
4
.2.4. Synthesis and isolation of compounds 8a and 8b
least 30 min. The reactions occurred too rapidly to extract kinetic
To a partial solution of 1.1 g (4.0 mmol) of 6 (3:1 mixture of iso-
data but the products were confirmed by LC–MS/MS.
mers) in 60 mL of tert-butanol was added a solution of 0.4 g
10.0 mmol) of NaOH in 10 mL of water. The mixture was stirred
(
4.6. Analysis for NO
at room temperature for 3 h. Afterwards, the reaction mixture
was acidified with 3 M HCl and concentrated under vacuum.
After extracting the aqueous solution with dichloromethane, this
organic solution was further extracted with 5% sodium bicarbonate
solution to obtain the sodium salt of the desired product, which
was precipitated out of the solution by the addition of 3 M HCl.
Filtration of the precipitates gave 868 mg (81% yield) of crude
material consisting of two isomers in a 3:1 ratio. A 663 mg sample
of these isomers was then separated through flash chromatogra-
phy on silica gel using a glass column, eluted with hexane/ethyl
acetate. The major product was pooled and then dried to afford
Chemiluminescence detection and quantification of NO evolv-
ing from the reactions were conducted using a commercial nitric
oxide analyzer (NOA). Solutions of 0.1 M phosphate buffer, pH
7.4, with 50 lM DTPA, 4 mM GSH and either the presence of
absence of GSTP1 (4 nM) were sparged with inert gas until a steady
detector response was established. Compound 4a, or 4b were
added to a final concentration of 100 nM and the NO release profile
was followed over time after injection. The resulting curve was
integrated to quantify the amount of NO released/mol of
compound.
4
42 mg (89% yield) of pure 8a: mp 150–152 °C; UV (acetonitrile)
À1
À1
1
k
max
(e
) 338 nm (18.0 mM cm ); H NMR d 3.31 (s, 6H), 7.10
4.7. Biological in vitro evaluation
13
(
1
(
s, 1H), 8.56 (s, 1H); C NMR d 42.59, 93.20, 106.32, 115.27,
32.83, 134.52, 161.66, 164.04. HRMS (ESI) m/z calculated for
M+H)+ (268.0676), found (268.0677). Anal. Calcd for C
4.7.1. Cell culture and proliferation assay
9 9
H N
5
O
5
:
HL-60 and U937 cell lines were obtained from the American
Type Culture Collection (Manassas, VA) and cultured according to
the supplier’s protocol. For proliferation assays, cells were seeded
C, 40.46; H, 3.40; N, 26.21. Found: C, 40.22; H, 3.46; N, 25.94.
The minor product was pooled and then dried, affording 118 mg
4
(
71% yield) of pure 8b: mp 179–180 °C; UV (acetonitrile) kmax
80 nm (15.9 mM cm ); H NMR d 3.29 (s, 6H), 7.10 (s, 1H),
(
e
)
at 2 Â 10 per well in 96-well plates and allowed to grow for
À1
À1
1
2
8
1
24 h. Compounds were prepared as 10 mM stock solutions in
1
3
.42 (s, 1H); C NMR d 42.86, 96.71, 104.74, 116.40, 132.34,
DMSO. Increasing drug concentrations in 10
lL of PBS were added
+
34.61, 157.06, 167.52. HRMS (ESI) m/z calculated for (M+H)
to 100 L of the culture medium and incubated for 72 h. The MTT
l
(
268.0676), found (268.0675).Anal. Calcd for C
9
H
9
N
5
O
5
: C, 40.46;
assay (Promega, Madison, WI) was performed according to the
manufacturer’s protocol. Each concentration was represented in
six repeats, and the screening was performed as three independent
experiments. IC50 values were calculated by using Sigma Plot soft-
ware (Systat Software, Inc., San Jose, CA).
H, 3.40; N, 26.21. Found: C, 40.22; H, 3.46; N, 25.94.
4
.3. Hydrolysis of PABA/NO, 4a, and 4b at physiological pH
The hydrolytic decomposition of PABA/NO, 4a, and 4b (final
concentration 1
lM) in 1 mL of 0.1 M pH 7.4 phosphate buffer con-
4.7.2. Metabolism studies
taining 0.1% DMSO were analyzed by HPLC. The reactions were
studied at 25 °C and 37 °C. Because of the relative stability of 4a
and particularly 4b to hydrolysis, rate constants were determined
through an initial rate analysis. The products of decomposition
were confirmed by LC–MS/MS.
The metabolism of compound 4b was studied in the human HL-
60 leukemia cell line. Cells were plated in 75 cm flasks and incu-
2
bated overnight at 37 °C. Cells were treated with 1
4b for varying time points, followed by two cycles of freeze
L of a 3:7 ratio of 0.1% formic
lM compound
(À80 °C) and thaw (37 °C) in 500
l
HPLC separations were performed on a reverse phase C18 col-
acid and HPLC grade acetonitrile. The precipitate was removed
by centrifugation at 14,000g for 15 min, followed by syringe filtra-
tion of the supernatant and analysis by LC–MS/MS according to the
protocol above.
umn (3-
l
m, 150 Â 2.00 mm) with an acetonitrile/water mobile
phase containing 0.1% formic acid at a flow rate of 0.2 mL/min.
The gradient began at 5% acetonitrile for 2 min, followed by a linear
gradient to 85% acetonitrile over the next 12 min, whereupon it
was run isocratically for 7 min. Ions for high resolution mass spec-
trometry were generated with dual electrospray ionization and
detected with a quadrupole time-of-flight mass analyzer.
4.7.3. Protein immunoblotting
5
HL-60 cells were seeded at 5 Â 10 /mL and allowed to grow
overnight. Cells were treated with 4a and 4b for 24 h, collected
by centrifugation, rinsed with PBS, and collected again. Cells were
lysed with lysis buffer (25 mM HEPES pH 7.4, containing 1 mM
4
.4. Kinetics and product analysis of uncatalyzed reactions of
PABA/NO, 4a and 4b with GSH
2
EDTA, 10 mM MgCl , 1% Triton X-100, 150 mM NaCl, 10% glycerol,
supplemented with Complete protease inhibitors cocktail (Roche,
Indianapolis, IN). Bis–tris NuPage 4–12% polyacrylamide gels (Life
Technologies, Carlsbad, CA) were used to separate proteins. Gels
were run with both SeeBlue prestained standard and MagicMark
XP protein standards (Life Technologies). Running buffer was
A solution of GSH (final concentration 4 mM) in 1 mL of 0.1 M
phosphate buffer at pH 7.4 was incubated at 37 °C for at least
0 min. The reaction was initiated by the addition of either
3
PABA/NO, 4a, or 4b (final concentration 1 M and 0.1% DMSO) to
l

4
988
Y. Kim et al. / Bioorg. Med. Chem. 23 (2015) 4980–4988
MES NuPage buffer (Life Technologies). Gels were electroblotted on
to a PVDF membrane for 1.5–2 h at 30 V using NuPage transfer buf-
fer with 20% methanol added. Blots were blocked with 5% milk in
TBS-T (tris buffered saline with 0.1% Tween 20), then rinsed and
probed with antibodies to caspase 8 (#9746), cleaved caspase 7
publication does not necessarily reflect the views or policies of
the Department of Health and Human Services, nor does mention
of trade names, commercial products or organizations imply
endorsement by the US Government.
Crystallographic studies were supported by the National
Institute on Drug Abuse (NIDA) under Contract Y1-DA1101 and
by the Naval Research Laboratory.
(
#8438), PARP (#9542), phospho-SAPK/JNK (#4671), and
SAPK/JNK (#9258) (Cell Signaling Technology, Danvers, MA),
diluted 1:1000, and cleaved PARP p85 (#G734b, Promega,
Madison, WI), diluted 1:5000, in 5% milk/PBS-T overnight at 4 °C.
Following 6 rinses with TBS-T over 60 min, the blot was incubated
with appropriate secondary antibodies coupled to horseradish per-
oxidase (Cell Signaling) diluted 1:1000 for 2 h at room tempera-
ture. After 6 rinses as above, blots were developed with Clarity
Western ECL reagent (BioRad Laboratories, Hercules, CA). Beta-
actin was used as a loading control (antibody # ab8227-50,
Abcam, Cambridge, MA).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2015.05.020.
References and notes
1. Saavedra, J. E.; Srinivasan, A.; Buzard, G. S.; Davies, K. M.; Waterhouse, D. J.;
Inami, K.; Wilde, T. C.; Citro, M. L.; Cuellar, M.; Deschamps, J. R.; Parrish, D.;
Shami, P. J.; Findlay, V. J.; Townsend, D. M.; Tew, K. D.; Singh, S.; Jia, L.; Ji, X.;
Keefer, L. K. J. Med. Chem. 2006, 49, 1157.
4
.8. X-ray diffraction study of 8a
2.
Findlay, V. J.; Townsend, D. M.; Saavedra, J. E.; Buzard, G. S.; Citro, M. L.; Keefer,
L. K.; Ji, X.; Tew, K. D. Mol. Pharmacol. 2004, 65, 1070.
Single-crystal X-ray diffraction data on compound 8a were
3
4
.
.
Acharya, A.; Das, I.; Chandhok, D.; Saha, T. Oxid. Med. Cell Longev. 2010, 3, 23.
Maciag, A. E.; Chakrapani, H.; Saavedra, J. E.; Morris, N. L.; Holland, R. J.; Kosak,
K. M.; Shami, P. J.; Anderson, L. M.; Keefer, L. K. J. Pharmacol. Exp. Ther. 2011,
collected using MoKa radiation and a Bruker APEX-2 CCD area
detector. The crystal was prepared for data collection by coating
with high viscosity microscope oil. The oil-coated crystal was
mounted on a micro-mesh mount (Mitergen, Inc.) and transferred
3
36, 313.
5
.
Maciag, A. E.; Holland, R. J.; Robert Cheng, Y. S.; Rodriguez, L. G.; Saavedra, J. E.;
Anderson, L. M.; Keefer, L. K. Redox. Biol. 2013, 1, 115.
3
to the diffractometer. The 0.326 Â 0.324 Â 0.217 mm crystal
6. Ji, X.; Pal, A.; Kalathur, R.; Hu, X.; Gu, Y.; Saavedra, J. E.; Buzard, G. S.; Srinivasan,
A.; Keefer, L. K.; Singh, S. V. Drug Design Devel. Ther. 2008, 2, 123.
Mannervik, B.; Alin, P.; Guthenberg, C.; Jensson, H.; Tahir, M. K.; Warholm, M.;
Jomvall, H. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 7202.
was triclinic in space group P-1, with unit cell dimensions
7
.
a = 8.5304(8), b = 11.5077(13), c = 13.0487(14) Å,
b = 103.508(3)°, and = 97.873(4)°. Data were 97.3% complete to
9.22° h (ꢀ0.73 Å) with an average redundancy of 2.14. The struc-
a = 107.190(5)°,
c
8. Townsend, D. M.; Tew, K. D. Oncogene 2003, 22, 7369.
9
.
Ribrag, V.; Koscielny, S.; Carpiuc, I.; Cebotaru, C.; Vande, W. H.; Talbot, M.;
Fenaux, P.; Bosq, J. Leukemia 2003, 17, 972.
2
ture was solved by direct methods and refined by full-matrix least
squares on F2 values using the programs found in the SHELXTL
suite (Bruker, SHELXTL v6.10, 2000, Bruker AXS Inc., Madison,
WI). Corrections were applied for Lorentz, polarization, and
absorption effects. Parameters refined included atomic coordinates
and anisotropic thermal parameters for all non-hydrogen atoms.
Hydrogen atoms on carbons were included using a riding model
10. Whelan, R. D.; Waring, C. J.; Wolf, C. R.; Hayes, J. D.; Hosking, L. K.; Hill, B. T. Int.
J. Cancer 1992, 52, 241.
11. O’Brien, M. L.; Tew, K. D. Eur. J. Cancer 1996, 32A, 967.
1
1
1
1
2. Kogias, E.; Osterberg, N.; Baumer, B.; Psarras, N.; Koentges, C.; Papazoglou, A.;
Saavedra, J. E.; Keefer, L. K.; Weyerbrock, A. Int. J. Cancer 2012, 130, 1184.
3. Chakrapani, H.; Wilde, T. C.; Citro, M. L.; Goodblatt, M. M.; Keefer, L. K.;
Saavedra, J. E. Bioorg. Med. Chem. 2008, 16, 2657.
4. Fu, J.; Liu, L.; Huang, Z.; Lai, Y.; Ji, H.; Peng, S.; Tian, J.; Zhang, Y. J. Med. Chem.
2013, 56, 4641.
[
coordinate shifts of C applied to H atoms] with C–H distance set
5. Holland, R. J.; Maciag, A. E.; Kumar, V.; Shi, L.; Saavedra, J. E.; Prud’homme, R.
at 0.96 Å. The final anisotropic full matrix least-squares refinement
on F with 347 variables converged at R1 = 4.81%, for the observed
K.; Chakrapani, H.; Keefer, L. K. Chem. Res. Toxicol. 2012, 25, 2670.
16. Shami, P. J.; Sauls, D. L.; Weinberg, J. B. Leukemia 1998, 12, 1461.
7. Saavedra, J. E.; Shami, P. J.; Wang, L. Y.; Davies, K. M.; Booth, M. N.; Citro, M. L.;
2
1
data and wR2 = 15.08% for all data. Complete information on data
collection and refinement is available in the Supporting informa-
tion. Atomic coordinates have been deposited with the
Cambridge Crystallographic Data Centre (deposition number
XXXXXX). Copies of the data can be obtained, free of charge, on
application to CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK
Keefer, L. K. J. Med. Chem. 2000, 43, 261.
18. Shami, P. J.; Saavedra, J. E.; Wang, L. Y.; Bonifant, C. L.; Diwan, B. A.; Singh, S. V.;
Gu, Y.; Fox, S. D.; Buzard, G. S.; Citro, M. L.; Waterhouse, D. J.; Davies, K. M.; Ji,
X.; Keefer, L. K. Mol. Cancer Ther. 2003, 2, 409.
9. Ridnour, L. A.; Thomas, D. D.; Donzelli, S.; Espey, M. G.; Roberts, D. D.; Wink, D.
A.; Isenburg, J. S. Antioxid. Redox Signal. 2006, 8, 1329.
1
2
0. Adler, V.; Yin, Z.; Fuchs, S. Y.; Benezra, M.; Rosario, L.; Tew, K. D.; Pincus, M. R.;
Sardana, M.; Henderson, C. J.; Wolf, C. R.; Davis, R. J.; Ronai, Z. EMBO J. 1999, 18,
[
fax: +44(0) 1223 336033 or e-mail: deposit@ccdc.cam.ac.uk].
1321.
21. Wang, T.; Arifoglu, P.; Ronai, Z.; Tew, K. D. J. Biol. Chem. 2001, 276, 20999.
Acknowledgements
22. Shami, P. J.; Saavedra, J. E.; Bonifant, C. L.; Chu, J.; Udupi, V.; Malaviya, S.; Carr,
B. I.; Kar, S.; Wang, M.; Jia, L.; Ji, X.; Keefer, L. K. J. Med. Chem. 2006, 49, 4356.
2
2
3. Sauerbrey, A.; Zintl, F.; Volm, M. Br. J. Cancer 1994, 70, 1144.
4. Cheng, A. L.; Su, I. J.; Chen, Y. C.; Lee, T. C.; Wang, C. H. J. Clin. Oncol. 1993, 11,
109.
5. Schisselbauer, J. C.; Silber, R.; Papadopoulos, E.; Abrams, K.; LaCreta, F. P.; Tew,
K. D. Cancer Res. 1990, 50, 3562.
6. Stella, F.; Weich, N.; Panero, J.; Fantl, D. B.; Schutz, N.; Fundia, A. F.; Slavutsky, I.
Cancer Epidemiol. 2013, 37, 671.
7. Saavedra, J. E.; Srinivasan, A.; Bonifant, C. L.; Chu, J.; Shanklin, A. P.; Flippen-
Anderson, J. L.; Rice, W. G.; Turpin, J. A.; Davies, K. M.; Keefer, L. K. J. Org. Chem.
We thank Dr. Sergey Tarasov and Ms. Marzena A. Dyba
Biophysics Resource in the Structural Biophysics Laboratory,
Frederick National Laboratory for Cancer Research) for assistance
with HRMS.
This project has been funded with Federal funds from the
Frederick National Laboratory for Cancer Research, National
Institutes of Health, under contract HHSN261200800001E, and by
the Intramural Research Program of the NIH, National Cancer
Institute, Center for Cancer Research. The content of this
(
2
2
2
2001, 66, 3090.