Esterase-sensitive nitric oxide donors of the diazeniumdiolate family: In vitro antileukemic activity

Article abstract of DOI:10.1021/jm9903850

We have designed a novel prodrug class that is stable in neutral aqueous media but releases bioactive nitric oxide (NO) on metabolism by esterase. Diazeniumdiolates of structure R₂NN(O)=N-OR', in which R' = Na, were reacted with BrCH₂OAc to convert the spontaneously NO-releasing salts 1a (R₂N = diethylamino) and 1b (R₂N = pyrrolidino) to prodrugs 2a (AcOM-DEA/NO) and 2b (AcOM-PYRRO/NO), respectively, where R' = CH₂OAc. In contrast to anions 1a and 1b (half-lives in pH 7.4 phosphate at 37 °C of 2 min and 3 s, respectively), 2a and 2b showed only minimal decomposition after 16 h under these conditions. Very rapid hydrolysis occurred in the presence of porcine liver esterase, however, with free anion 1a being observed as an intermediate in the esterase-induced generation of NO from 2a. The potential utility of this prodrug class is illustrated with a comparison of 1 and 2 as antiproliferative agents in NO-sensitive human leukemia cell lines HL-60 and U937. While the 72-h IC₅₀'S for 1a and 1b (which generate NO throughout the medium) in HL-60 cell cultures were >600 μM, those of 2a and 2b were 8.3 and 6.4 μM, respectively. This result is consistent with our hypothesis that 2 is selectively hydrolyzed to 1 and thence to NO intracellularly. For U937 cells, the 72-h IC₅₀ for both 2a and 2b was 53 μM. By contrast, relatively high antiproliferative IC₅₀'s (≥ 100 μM in U937 cells) were observed for analogues in which R' = CH₂CH₂SC(O)Me, from which acetyl and 2- mercaptoethyl groups must be successively cleaved to free the NO-releasing diazeniumdiolate function. Within 24 h at initial concentrations of 50 μM, 2a and 2b induced apoptosis in 50% and 57% of the HL-60 cells, respectively (35% and 40% of the U937 cells, respectively). The data reveal significant in vitro antileukemic activity on the part of these novel compounds. Moreover, their substantial ease-of-handling advantages over the anionic diazeniumdiolates from which they are derived suggest their use as convenient agents for probing the biological roles of NO.

Full text of DOI:10.1021/jm9903850

J . Med. Chem. 2000, 43, 261-269
261
Ester a se-Sen sitive Nitr ic Oxid e Don or s of th e Dia zen iu m d iola te F a m ily: In
Vitr o An tileu k em ic Activity
J oseph E. Saavedra,†,‡ Paul J . Shami,†,§ Lai Yi Wang,§ Keith M. Davies,| Melissa N. Booth,
Michael L. Citro,‡ and Larry K. Keefer*,
Intramural Research Support Program, SAIC Frederick, National Cancer Institute, Frederick Cancer Research and
Development Center, Frederick, Maryland 21702, Salt Lake City Veterans Administration Medical Center, Salt Lake City,
Utah 84148, Department of Chemistry, George Mason University, Fairfax, Virginia 22030, and Chemistry Section,
Laboratory of Comparative Carcinogenesis, National Cancer Institute, Frederick Cancer Research and Development Center,
Frederick, Maryland 21702
Received J uly 27, 1999
We have designed a novel prodrug class that is stable in neutral aqueous media but releases
bioactive nitric oxide (NO) on metabolism by esterase. Diazeniumdiolates of structure R2N-
N(O)dN-OR′, in which R′ ) Na, were reacted with BrCH2OAc to convert the spontaneously
NO-releasing salts 1a (R2N ) diethylamino) and 1b (R2N ) pyrrolidino) to prodrugs 2a (AcOM-
DEA/NO) and 2b (AcOM-PYRRO/NO), respectively, where R′ ) CH2OAc. In contrast to anions
1a and 1b (half-lives in pH 7.4 phosphate at 37 °C of 2 min and 3 s, respectively), 2a and 2b
showed only minimal decomposition after 16 h under these conditions. Very rapid hydrolysis
occurred in the presence of porcine liver esterase, however, with free anion 1a being observed
as an intermediate in the esterase-induced generation of NO from 2a . The potential utility of
this prodrug class is illustrated with a comparison of 1 and 2 as antiproliferative agents in
NO-sensitive human leukemia cell lines HL-60 and U937. While the 72-h IC50’s for 1a and 1b
(which generate NO throughout the medium) in HL-60 cell cultures were >600 µM, those of
2a and 2b were 8.3 and 6.4 µM, respectively. This result is consistent with our hypothesis
that 2 is selectively hydrolyzed to 1 and thence to NO intracellularly. For U937 cells, the 72-h
IC50 for both 2a and 2b was 53 µM. By contrast, relatively high antiproliferative IC50’s (>100
µM in U937 cells) were observed for analogues in which R′ ) CH2CH2SC(O)Me, from which
acetyl and 2-mercaptoethyl groups must be successively cleaved to free the NO-releasing
diazeniumdiolate function. Within 24 h at initial concentrations of 50 µM, 2a and 2b induced
apoptosis in 50% and 57% of the HL-60 cells, respectively (35% and 40% of the U937 cells,
respectively). The data reveal significant in vitro antileukemic activity on the part of these
novel compounds. Moreover, their substantial ease-of-handling advantages over the anionic
diazeniumdiolates from which they are derived suggest their use as convenient agents for
probing the biological roles of NO.
Sch em e 1. Synthesis of O2-Acetoxymethylated
Diazeniumdiolates 2
In tr od u ction
A major advantage of the anionic diazeniumdiolates
1 (Scheme 1) as tools for investigating the bioregulatory
roles of nitric oxide (NO) is their ability to generate NO
spontaneously in aqueous media at rates that depend
only upon pH, temperature, and the identity of the R2N
moiety.1 Half-lives for commercially available diazeni-
umdiolates in pH 7.4 phosphate buffer range from 2 s
to 20 h at 37 °C, with their potencies and durations of
action in a variety of in vivo applications correlating well
with predictions based on their solution chemistry.2
But this inexorability of NO release can be a major
disadvantage in many biomedical applications, particu-
larly in the design of drugs for remediating disorders
resulting from local deficiencies of endogenous NO.
Accordingly, the ultimate goal of our drug discovery
effort is to devise strategies for selectively delivering NO
* To whom correspondence should be addressed. Tel: 301-846-1467.
Fax: 301-846-5946. E-mail: keefer@mail.ncifcrf.gov.
† These authors contributed equally to this work.
‡ SAIC Frederick.
to the sites of need while avoiding exposing other NO-
sensitive parts of the body to its multipotent effects.2
One means of doing this is to convert these spontane-
ous NO generators to stable prodrug form by attaching
§ Veterans Administration.
| George Mason University.
National Cancer Institute.
10.1021/jm9903850 CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/30/1999

262 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2
Saavedra et al.
a protecting group (R′) to the terminal oxygen to produce
agents of structure R2N-N(O)dNOR′.3 If one could
choose a protecting group that would be selectively
removed in the target organ or cell type of interest,
dosage with NO could be concentrated at that site even
if the prodrug were systemically distributed.
The purpose of this report is to introduce a prodrug
class that can be activated to NO-generating form by
the action of esterases.
Resu lts
Syn th esis a n d Hyd r olysis of O2-Acetoxym eth yl-
a ted Dia zen iu m d iola tes. The nucleophilic character
of the terminal oxygen3 in anions of structure 1 allowed
for their straightforward conversion to acetals 2 by
reacting them with bromomethyl acetate in tetrahydro-
furan, as shown in Scheme 1. Attachment of the
acetoxymethyl group to the terminal oxygen of the
anionic diazeniumdiolate resulted in great stabilization.
For example, unlike sodium salt 1a (DEA/NO), whose
half-life at 37 °C in pH 7.4 phosphate buffer is 2 min,1
acetoxymethylated DEA/NO 2a (AcOM-DEA/NO) was
found to be stable in neutral aqueous media. Its
characteristic ultraviolet absorption (λmax ) 230 nm)
showed less than a 2% decrease after 16 h in phosphate-
buffered saline (pH 7.4) at 37 °C. Comparable stability
was also noted in phosphate buffers at pH 5.5, 6.5, and
8.0. The stability of 2a decreased markedly, however,
at both higher and lower pH. In 1 mM sodium hydrox-
ide, the 230-nm maximum of 2a decayed exponentially
with a half-life of 1.4 min, producing the free diazen-
iumdiolate 1a (λmax ) 250 nm). In 1 mM hydrochloric
acid, the measured half-life of 2a was 19 min.
Exposure of 2a as a stable aqueous solution to
esterase resulted in its rapid disappearance. As il-
lustrated in Figure 1A, when a 0.19 mM solution of 2a
at pH 8 was treated at 37 °C with porcine liver esterase,
its characteristic chromophore at 230 nm completely
disappeared within 1 min and the spectral character-
istics of 1a were the only ones detectable. In accord with
its known first-order dissociation to diethylamine and
NO, the spectrum of 1a had also disappeared after the
mixture was allowed to stand overnight (not shown).
Comparable results were seen at pH 7.4 (Figure 1B),
except that the isosbestic behavior of Figure 1A was
compromised because subsequent dissociation of the
initially formed 1a to diethylamine and NO at this lower
pH is more rapid than at pH 8. The overall course of
esterase-catalyzed hydrolysis of 2a and 2b is shown in
eq 1.
F igu r e 1. Ultraviolet spectral changes observed after addition
of a 0.19 mM solution of 2a in phosphate buffer at 37 °C to
porcine liver esterase at a pH of 8.0 (A) or 7.4 (B). Panel A
shows spectra recorded 5, 10, 15, and 30 s after the addition,
while panel 1B shows corresponding changes after 10, 25, 35,
60, and 90 s. The peak at 230 nm is characteristic of the
starting ester, while that at 250 nm is due to anionic diazeni-
umdiolate 1a . The loss of the isosbestic point in panel B is
due to the more rapid decomposition of 1a at pH 7.4 (t1/2 ) 2
min).
after addition of the esterase (chemiluminescence traces
shown as Figure A in Supporting Information). Both NO
release profiles integrated for the near-theoretical yield
of this gas (1.83 and 1.85 mol of NO/mol of 2, respec-
tively). The time course of NO release from 2a was
significantly more prolonged than that obtained for 2b.
This is consistent with the known half-lives (eq 1) of
anions 1a and 1b, which are the intermediates in the
mechanism of NO release.
To determine whether nonhepatic esterases could also
effect the hydrolysis, mixtures of 2a with rat blood were
studied. We found that 6.7 mM 2a in whole blood
rapidly decomposed even at 4 °C. The apparent half-
life under these conditions as measured by gas chro-
matography was 32 min. Correspondingly faster disap-
pearance was noted as temperature was increased, half-
life estimates of 8 and 2 min being recorded at 25 and
37 °C, respectively. Both plasma and the cell fraction
were catalytically active; a half-life of 8 min was
observed when 12.3 mM 2a was dissolved at 25 °C in
either rat plasma or a suspension of blood cells in
phosphate-buffered saline, pH 7.4.
Syn th esis a n d Hyd r olysis of Th ioa ceta te Ester s
4. Capitalizing on an alternate strategy of esterase-
sensitive prodrug design employed previously by Lefe-
bvre et al.,4 we also performed the stepwise synthesis
of thioacetates 4 as outlined in Scheme 2. The 2-bromo-
ethyl derivatives (3) of 1 were prepared as previously
described3,5 and then reacted with MeC(O)S- to give
stable products (4). According to this strategy, the
thioester bond should hydrolyze rapidly under the
influence of esterase, freeing the mercaptoethyl deriva-
tive as a discrete intermediate. This in turn could
hypothetically suffer intramolecular attack by sulfur on
To confirm that molecular NO is produced in the
esterase-mediated hydrolysis of compounds of structure
2, the reaction was conducted in a flask being continu-
ously swept with inert gas to purge NO as it formed in
the dissociation into an NO-specific chemiluminescence
analyzer. Both 2a and 2b began to generate copious NO

Esterase-Activated Nitric Oxide Prodrugs
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2 263
Sch em e 2. Synthesis of, and Proposed Mechanism of
Esterase-Induced NO Release from, Thioacetate Esters 4
its â-carbon,4 displacing diazeniumdiolate anion 1 with
formation of thiirane as a byproduct.
That this was actually the course followed was
indicated by studying the ultraviolet spectra of DEA/
NO derivative 4a at 37 °C in the absence and presence
of porcine liver esterase. The characteristic chromophore
of 4a at 236 nm did not change in intensity in pH 10
buffer over 30 min but shifted to 250 nm within 5 min
when the esterase was added, consistent with the rapid
esterase-induced regeneration of free anion 1a from 4a .
This process was considerably slower at pH 7.4,
presumably because the initial hydrolysis produced the
thiol rather than the thiolate anion at this pH,6 decreas-
ing the nucleophilicity of the sulfur atom and slowing
the rate of thiirane departure. Nevertheless, formation
and dissociation of 1a did occur in physiological buffer,
a total of 1.6 mol of NO/mol of starting material
eventually being detected by chemiluminescence when
an otherwise stable solution of 4a was treated with
esterase.
Hyd r olysis of 2 a n d 4 by Leu k em ia Cells. To
determine whether NO-sensitive leukemia cell lines HL-
60 and U9377 are capable of converting diazeniumdi-
olates 2 and 4 to NO intracellularly, we conducted a
series of experiments in serum-free phosphate-buffered
saline (PBS) with and without the cells. Nitrate and
nitrite, the oxidative end products of NO release, were
measured in the supernatants. Figure 2 shows the
quantities of these indicators of total NO release de-
tected under various circumstances. In cell-free PBS,
there was minimal spontaneous release of NO by either
acetoxymethyl derivative (2a or 2b) when added at
concentrations as high as 100 µM. By contrast, when
added at the same concentrations in cell-containing
PBS, there was significant and concentration-dependent
NO release from both compounds. HL-60 cells appeared
to induce the release of slightly more NO than U937
cells. When the NO donors were added to HL-60 cells
at a concentration of 100 µM, the concentrations of
nitrate/nitrite at 24 h were 83 ( 6.1 and 102 ( 0.2 µM
for 2a and 2b, respectively. The same measurements
with U937 cells were 74 ( 12.5 and 95 ( 0.3 µM for 2a
and 2b, respectively. By comparison, when 4a was
F igu r e 2. NO release by 2a and 2b. HL-60 or U937 cells were
cultured at a density of 150 000/mL in serum-free PBS. At the
time of culture initiation, we added 2a or 2b to the cells or to
cell-free PBS. NO release was determined by measuring
nitrate/nitrite levels in the supernatants at 24 h. When added
to leukemia cells, 2a and 2b released a significant amount of
NO in a concentration-dependent fashion. There was very little
spontaneous release by either compound. Results are the
means ( SEM of at least three separate experiments for each
data point.
added to HL-60 or U937 cells at 100 µM, the measured
nitrate/nitrite concentrations at 24 h were elevated over
background by 12.8 ( 0.75 and 18.4 ( 4.8 µM, respec-
tively (not shown).
To determine the time course of intracellular activa-
tion of the compounds, 2a , 2b, and 4a were added to
HL-60 or U937 cells at a concentration of 100 µM and
nitrate/nitrite levels in the supernatants were measured
at time points of 2, 4, 6, 8, and 24 h. Acetoxymethyl
compounds 2a and 2b achieved maximal NO release
6-8 h after addition (Figure 3). No significant NO
release was observed from thioacetate 4a until 24 h
after addition. In a separate set of experiments, 2a and
2b were added at concentrations of 100 µM to HL-60 or
U937 cells cultured at different densities in serum-free
PBS. In these experiments, the amount of NO released
after 24 h increased with increasing numbers of leuke-
mia cells in the cultures (Figure B in Supporting
Information).
To determine whether activation of compounds 2a
and 2b occurred intracellularly, as hypothesized, and
not by release of esterases into the medium, HL-60 and
U937 cells were incubated in serum-free PBS for a
period of 8 h before 2a or 2b was added at a concentra-
tion of 100 µM. Supernatants were then collected at 2,
4, 6, 8, and 24 h for measurement of nitrate + nitrite
levels using the Griess assay. The results, summarized
in Figure 4, reveal an accumulation of NO oxidation
products over the first 8 h after adding the drugs; this

264 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2
Saavedra et al.
F igu r e 4. Time course of NO release by 2a and 2b after
delayed addition to leukemia cells. HL-60 or U937 cells were
incubated in serum-free PBS at a density of 150 000/mL for 8
h whereupon 2a or 2b was added at a concentration of 100
µM. Nitrate + nitrite levels were measured by the Griess
reaction at the indicated time points after the addition of 2a
and 2b. The kinetics of NO release was similar to those in
experiments in which 2a and 2b were added at the time of
culture initiation (Figure 3). The higher levels of NO metabo-
lites at the 24-h time point (32 h after culture initiation) likely
reflect spontaneous cell death rather than secretion of es-
terases into the medium by viable leukemia cells. Results are
the means ( SEM of two separate experiments at each time
point.
F igu r e 3. Time course of NO release by 2a , 2b, and 4a . HL-
60 or U937 cells were cultured at a density of 150 000/mL in
serum-free PBS. At the time of culture initiation, 2a , 2b, or
4a was added to the cells at a concentration of 100 µM. NO
release was determined by measuring nitrate/nitrite levels in
the supernatants at the indicated time points. NO release from
2a and 2b peaked at 6-8 h. NO release from 4a was very slow
and minimal at 24 h. Results are the means ( SEM of at least
three separate experiments for each data point.
result was kinetically reminiscent of that seen in Figure
3, except that buildup of metabolites was slower and
did not plateau at 8 h but rather was highest at the
24-h time point. The differences can be explained by the
fact that in these serum-free media (i.e., containing no
growth factors and inadequate nutrients) cellular well-
being declines with time, with death occurring roughly
1 day after initiating the culture. Thus at the 24-h time
point in Figure 4, which is actually 32 h after culture
initiation, a significant portion of the cells can be
expected to have lysed, releasing their intracellular
esterases into the medium for continued activity.
3). These data provide further evidence for the intrac-
ellular activation of compounds 2.
An tip r olifer a tive Effects of 2 a n d 4 on Leu k em ia
Cells in Cu ltu r e. We next determined the effect of
esterase-sensitive diazeniumdiolates 2 and 4 on leuke-
mia cell growth in serum-containing medium using
Trypan blue exclusion. When measured at 24, 48, and
72 h, there was a concentration-dependent growth
inhibition by both 2a and 2b, but even after 3 days of
exposure and at a concentration as high as 100 µM, 4b
inhibited cellular growth minimally (Figure 6A). While
not very effective at lower concentrations, 4a at 100 µM
inhibited HL-60 and U937 cell growth at 72 h by 90%
and 48%, respectively. The most potent growth inhibi-
tors were 2a and 2b, with significant growth inhibition
in the low micromolar range. For HL-60 cells, the
calculated 50% inhibitory concentrations (IC50) at 72 h
were 8.3 and 6.4 µM for 2a and 2b, respectively. For
U937 cells, the 72-h IC50 for both 2a and 2b was 53 µM.
To determine whether growth inhibitory potency is
influenced by the presence of the protecting groups on
the terminal oxygen, HL-60 and U937 cells were treated
To confirm that 2a and 2b are hydrolyzed intracel-
lularly in these leukemia cell cultures, rather than
through the activity of esterases secreted into the
extracellular medium, HL-60 or U937 cells were incu-
bated in serum-free PBS for 16 h and centrifuged to
remove the cells. Compounds 2a and 2b were added to
the resulting cell-free supernatants at a concentration
of 100 µM, and nitrite/nitrate levels were measured by
the Griess assay 24 h later. As shown in Figure 5,
concentrations of NO metabolites in the range of 4-5
µM were observed. This is far less than the 70-100 µM
concentrations seen at the same time point when the
compounds were added to cell-containing media (Figure

Esterase-Activated Nitric Oxide Prodrugs
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2 265
produce formaldehyde, acetate, and an amine but not
NO. When assayed 24, 48, and 72 h after culture
initiation, 5 did not induce significant apoptosis in either
HL-60 or U937 cells (Figure 7). By contrast, when added
at the same concentrations, 2a and 2b induced extensive
apoptosis in both cell lines (Figure 7). These flow
cytometric apoptosis assays were confirmed using DNA
laddering assays (Figure 8).
Discu ssion
The family of NO donors introduced in this work
constitutes a novel class of compounds with a unique
mechanism of activation requiring the enzymatic cleav-
age of an ester bond. In enzyme-free aqueous media,
they showed little tendency to release NO spontane-
ously. When added to leukemia cells, however, they
released large amounts of NO as evidenced by the high
levels of NO oxidation products in the supernatants of
these cultures (Figures 2 and 3). The suggestion in these
results that the NO donors were activated intracellu-
larly is supported by the fact that the amount of NO
released by the individual compounds in a given time
interval was proportional to the number of cells in the
cultures. The conclusion that these esters are activated
for NO release intracellularly was further supported by
the similarity of nitrite/nitrate accumulation rates when
substrate was added to the medium at 8 versus 0 h after
initiating the cultures (Figure 4). Additional confirma-
tion was found in the fact that cell-free supernatants
prepared from cell cultures incubated for 16 h were
essentially devoid of esterase activity (Figure 5).
The compounds we screened showed differences in
potency as far as leukemia cell growth inhibition is
concerned (Figure 6). Indeed, only acetoxymethyl de-
rivatives 2a and 2b were potent growth inhibitors in
the concentration range studied. Furthermore, this
growth inhibition seemed to be induced by triggering
apoptosis (Figures 7 and 8). Since control compound 5
did not induce leukemia cell apoptosis, we conclude that
growth inhibition was due to NO release rather than
to toxicity of formaldehyde, acetate, or the amine carrier
moiety.
Since compounds 1, 2, and 4 are all expected to
release the same amount of NO per molecule on hy-
drolysis, their differences in growth inhibitory potency
are likely due to differences in their rates and location
of NO release. Indeed, when added to leukemia cells,
2a and 2b (the two most potent growth inhibitors)
generated more NO at faster rates over a period of 24 h
than 4a (which was much less potent). HL-60 cells
seemed to be somewhat more sensitive than U937 cells
to 2a and 2b even though both compounds released
roughly equal amounts of NO when added to each cell
line. Consequently, the observed difference in suscep-
tibility is more likely due to greater inherent resistance
to NO by U937 cells than to differing rates of activation
of the compounds by the two cell lines.
F igu r e 5. Activation of 2a and 2b in cell supernatants. HL-
60 or U937 cells were incubated at a density of 150 000/mL in
serum-free PBS for 16 h whereupon supernatants were col-
lected. Compounds 2a and 2b were added to the cell-free
supernatants at a concentration of 100 µM and incubated for
24 h, at which point nitrate + nitrite levels were measured
using the Griess assay. There was minimal activation of 2a
and 2b in cell-free supernatants, further supporting the
intracellular mechanism for hydrolysis of these compounds.
Results are the means ( SEM of three separate experiments.
with 1a and 1b. Even after 72 h of exposure at initial
concentrations up to 100 µM, underivatized diazenium-
diolate ions 1a and 1b did not substantially affect
leukemia cell growth (Figure 6B). We did not observe
any significant growth inhibition by either salt until
concentrations of 1000 µM were reached. The observed
IC50 values for 1a were 620 and 500 µM for HL-60 and
U937 cells, respectively. Even at a concentration of 1000
µM, the IC50 for 1b was not reached.
In d u ction of Leu k em ia Cell Ap op tosis by 2a a n d
2b. We then determined the effect of the two most
potent growth inhibitors (2a and 2b) on HL-60 and
U937 cell apoptosis. To ascertain that the observed
effects were due to NO release and not to generation of
formaldehyde, acetate, and/or the amine byproducts of
dissociation of 2, we used compound 5 [prepared by
reacting 2-(N,N-diethylamino)ethanol first with sodium
hydride then with bromomethyl acetate] as a negative
Once cleaved by esterases, 2a and 2b are converted
to 1a and 1b, respectively. While these anions have very
short half-lives of NO release (2 min and 3 s, respec-
tively), they proved to be far less potent growth inhibi-
tors than either 2a or 2b at identical concentrations.
This discrepancy can be explained by the fact that 1a
control since its esterase-induced hydrolysis would

266 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2
Saavedra et al.
F igu r e 6. Effect of different NO donors on leukemia cell growth. HL-60 or U937 cells were cultured with the NO donors at the
indicated concentrations. Cell growth was assayed by Trypan blue exclusion 72 h after culture initiation. There was a concentration-
dependent growth inhibition by the NO donors, with 2a and 2b being the most potent agents (A). Similar trends were seen at 24
and 48 h (not shown). By contrast, anions 1a and 1b were about 2 orders of magnitude less potent when introduced directly into
the culture medium (B). Results are the means ( SEM of at least three separate experiments for each data point.
and 1b release NO spontaneously and probably do not
penetrate leukemia cells significantly before releasing
NO. Consequently, they generate NO extracellularly
such that a substantial proportion is oxidized before it
can enter the cell, wasting its potential antileukemic
activity. Additionally, we have previously shown that
in order to be effective growth inhibitors, compounds of
this class must have a long half-life for NO release (up
to several hours).7 On the other hand, 2a and 2b do not
release NO spontaneously but rather require activation
through esterase cleavage. Since the esterase activity
in these experiments was found to be concentrated in
the cytoplasm, the amount of NO continuously delivered
to its molecular targets was maximized with 2a and 2b.
considerable potential utility for intracellular targeting
of NO in a variety of in vitro applications. Rapid
hydrolysis in rat blood of the acetyl esters described here
obviously limits the suitability of these specific com-
pounds for in vivo applications. Nevertheless, analogues
that are less susceptible to the action of such nonspecific
esterases can be conceived. An important goal of this
work will be to design agents that are tailored for
selective activation by esterases unique to a given target
cell type. Noteworthy too are the demonstrable ease-
of-handling advantages of the esters reported herein
relative to the hydrolytically unstable parent anions.
Further work should considerably extend the biomedical
utility of this novel class of NO donor.
Sign ifica n ce
Exp er im en ta l Section
One of the main obstacles to the use of NO in the
treatment of neoplastic diseases is its effects on other
biological targets, particularly in the cardiovascular
system.8,9 Accordingly, we are seeking ways of maximiz-
ing NO exposure to the tumor cells while minimizing
undesired collateral effects elsewhere in the body. The
present work describes a selection of esterase-activated
prodrugs that appear to offer substantial cell-targeting
advantages over the anions from which they were
derived. The observed increases in leukemia cell anti-
proliferative potency (up to 2 orders of magnitude or
more relative to the corresponding anions) suggest
NO was purchased from Matheson Gas Products (Mont-
gomeryville, PA). Ultraviolet (UV) spectra were recorded on a
Hewlett-Packard model 8451A diode array spectrophotometer.
Nuclear magnetic resonance spectra were collected with a 300-
MHz Varian Unity Plus or a Varian XL-200 NMR spectrom-
eter; spectra were obtained in deuteriochloroform. Chemical
shifts are reported in parts per million (ppm) downfield from
tetramethylsilane. Starting materials 1 and 3 were prepared
as previously described.3,5,10 Elemental analyses were done by
Atlantic Microlab, Inc. (Norcross, GA).
O2-(Acet oxym et h yl) 1-(N,N-Diet h yla m in o)d ia zen -1-
iu m -1,2-d iola te (2a ). A slurry of 1.1 g (7.1 mmol) of sodium
1-(N,N-diethylamino)diazen-1-ium-1,2-diolate (1a )10 and 200

Esterase-Activated Nitric Oxide Prodrugs
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2 267
allowed to warm to room temperature and stirred for 72 h.
The mixture was evaporated in vacuo and the residue was
extracted with dichloromethane, dried over sodium sulfate,
filtered, and evaporated to give 611 mg of an amber oil. The
crude mixture was chromatographed on silica gel and eluted
with dichloromethane to give 321 mg (22%) of product, 2a :
NMR δ 1.12 (t, 6 H), 2.11 (s, 3 H), 3.21 (q, 4 H), 5.81 (s, 2 H);
UV (H2O) λmax (ꢀ) 230 nm (7.6 mM-1 cm-1). Anal. (C7H15N3O4)
C, H, N.
Decom p osition of 2a in Ra t Blood . Samples of 2a (20
mM in pH 7.4 phosphate buffer) were mixed with whole blood
at various temperatures to provide a final concentration in the
blood of 6.7 mM. Aliquots of 200 µL each were quenched at 0
°C with 100 µL of 10% trichloroacetic acid, filtered through
glass wool at various times after mixing, and injected into a
Shimadzu model 4BM gas chromatograph equipped with a
Hewlett-Packard 18652A A/D converter coupled to the recorder
of a flame ionization detector. A 3% OV-17 on 80/100 Supel-
coport glass column (Supelco, Bellefonte, PA) was used. Oven
temperature was kept constant at 150 °C, with a helium flow
rate of 40 mL/min. The retention time for 2a was 5.0 min
under these conditions. The decrease in the integral of the peak
appearing at the retention time of 2a was monitored and used
to estimate the half-lives under the indicated conditions. Later
experiments tracking disappearance of 2a in plasma and the
blood cell fraction were similarly conducted, except that the
initial substrate concentration was 12.3 mM, aliquots were
quenched with ethanol, and chromatography was performed
on a Varian 3700 gas chromatograph equipped with a 30-foot
× 0.3-mm DB-1 capillary column (J &W Scientific, Folsom, CA)
with a helium flow rate of 19 mL/min; oven temperature was
kept at 75 °C for 5 min, then programmed at 10 °C/min to
250 °C. The retention time of 2a was 14.5 min.
O2-(Acetoxym eth yl) 1-(P yr r olid in -1-yl)d ia zen -1-iu m -
1,2-d iola te (2b). A slurry of 3.07 g (0.02 mol) of sodium
1-(pyrrolidin-1-yl)diazen-1-ium-1,2-diolate (1b),5 1 g of anhy-
drous sodium carbonate, and 20 mg (0.12 mmol) of silver
acetate (needed because 1b seemed to react with bromomethyl
acetate more sluggishly than 1a , requiring silver ion as-
sistance) in 25 mL of dimethyl sulfoxide was stirred with 1.4
mL (0.014 mol) of bromomethyl acetate for 72 h at room
temperature. The resulting solution was poured over 100 g of
ice-water, extracted with ether, and dried over anhydrous
sodium sulfate. The solution was filtered through magnesium
sulfate and evaporated to give 1.67 g of an oil. Preliminary
purification was carried out on a 4-cm × 15-cm KP-Sil column
connected to a Biotage Flash 40 system and eluted with pure
F igu r e 7. Induction of leukemia cell apoptosis by 2a and 2b.
HL-60 or U937 cells were cultured with 2a , 2b, or 5 at the
indicated concentrations. At 24, 48, and 72 h, the percent of
apoptotic cells was determined by flow cytometry. Control
compound 5, whose esterase-induced hydrolysis would gener-
ate formaldehyde but not NO, did not induce significant
leukemia cell apoptosis. At the same concentrations, 2a and
2b induced a substantial degree of apoptosis in both cell lines.
Results are the means ( SEM of at least three separate
experiments for each data point.
dichloromethane at
a rate of 25 mL/min. The fractions
containing the pure product were combined and evaporated
in vacuo to give 1.33 g of a solid that was recrystallized from
petroleum ether to give white plates: mp 62-63 °C; NMR δ
1.97 (m, 4 H), 2.12 (s, 3 H), 3.60 (m, 4 H), 5.76 (s, 2 H); UV
(H2O) λmax (ꢀ) 218 nm (5.2 mM-1 cm-1) and 256 nm (7.8 mM-1
cm-1). Anal. (C7H13N3O4) C, H, N.
Hyd r olysis of Ester s w ith P or cin e Liver Ester a se. To
1 mL of a 189 µM solution of 2a in pH 8 buffer was added 1
µL (1 unit) of porcine liver esterase (suspension in 3.2 M
ammonium sulfate; Sigma) (1.05 mg of protein in a 0.5-mL
solution). The course of the reaction was monitored by ultra-
violet spectrophotometry at 37 °C by observing the disappear-
ance of the chromophore at 230 nm and the ensuing formation
of 1a (250 nm). To obtain the early data points for Figure 1,
the reaction was also initiated by adding 1.0 mL of a 189 µM
2a solution in pH 8 buffer at 37 °C to 1 µL of the esterase
contained in a 1-mL cuvette prepositioned in the spectropho-
tometer. Esterase-induced hydrolysis of esters 2b and 4 was
effected under similar conditions.
Gen er a l Met h od for t h e P r ep a r a t ion of O2-(2-S-
Acetylm er ca p toeth yl) 1-Su bstitu ted Dia zen -1-iu m -1,2-
d iola t es. To a stirring solution of the appropriate O2-
bromoethyl 1-substituted diazen-1-ium-1,2-diolate 3 in dichloro-
methane at 0 °C under N2 was added first triethylamine and
then thiolacetic acid (Aldrich). The reaction mixture was
allowed to warm gradually to 25 °C over the course of several
F igu r e 8. Induction of apoptosis by NO donors in HL-60 cells.
Cells were treated with compound 2a for a period of 24 h at a
concentration of 100 µM. Cellular DNA was then isolated and
electrophoresed in a 1.2% agarose gel: lane 1 ) molecular
weight marker; lane 2 ) untreated control; lane 3 ) 100 µM
2a . The NO donor induced apoptosis (as evidenced by DNA
laddering). The figure shows the results of one experiment that
is representative of three performed. Similar results were
obtained with U937 cells and with compound 2b.
mg of anhydrous sodium carbonate in 20 mL of tetrahydro-
furan was cooled to 0 °C under a stream of dry nitrogen. To
this was added a solution of 1 g (6.5 mmol) of bromomethyl
acetate in 5 mL of tetrahydrofuran. The reaction mixture was

268 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2
Saavedra et al.
hours. The solution resulting from filtration of the reaction
mixture under vacuum was washed first with water and then
with cold dilute hydrochloric acid. The washed dichloro-
methane solution was dried with sodium sulfate and filtered
under vacuum through magnesium sulfate. Removal of solvent
yielded the crude product, which was purified by flash chro-
matography on a Flash 40 system using a 90-g KP-Sil column
eluted with 10:1 dichloromethane:ethyl acetate to give the final
product.
O2-(2-S-Acetylm er ca p toeth yl) 1-(N,N-Dieth yla m in o)-
d ia zen -1-iu m -1,2-d iola te (4a ). O2-(2-Bromoethyl) 1-(N,N-
diethylamino)diazen-1-ium-1,2-diolate (3a )3 (2.3 g, 9.6 mmol),
dichloromethane (35 mL), triethylamine (4.0 mL, 2.9 g, 29
mmol), and thiolacetic acid (2.0 mL, 2.1 g, 28 mmol) were
mixed and stirred for 3 h as described above. The crude product
was flash chromatographed to yield a pale yellow oil (775 mg,
3.3 mmol, 34%): NMR δ 1.10 (t, 6 H), 2.35 (s, 3 H), 3.12 (q, 4
H), 3.24 (t, 2 H), 4.36 (t, 2 H); UV (H2O) λmax (ꢀ) 236 nm (12
mM-1 cm-1). Anal. (C8H17N3SO3) C, H, N, S.
O2-(2-S-Acetylm er ca p toeth yl) 1-(P yr r olid in -1-yl)d ia -
zen -1-iu m -1,2-d iola te (4b). O2-(2-Bromoethyl) 1-(pyrrolidin-
l-yl)diazen-1-ium-1,2-diolate (3b)5 (1.7 g, 7.0 mmol) in dichloro-
methane (15 mL), triethylamine (3.2 mL, 2.3 g, 23 mmol), and
thiolacetic acid (1.4 mL, 1.49 g, 19.6 mmol) were used as
described in the general procedure. The reaction mixture was
allowed to stir at 25 °C for 18 h. The crude product, a yellow
oil, was chromatographed to yield a pale yellow oil, which
crystallized on standing to yield 690 mg (43%) of pale yellow
crystals: mp 38-40 °C; NMR δ 1.95 (m, 4 H), 2.35 (s, 3 H),
3.23 (t, 2 H), 3.56 (m, 4 H), 4.26 (t, 2 H); UV (H2O) λmax (ꢀ) 236
nm (10.9 mM-1 cm-1) and 250 nm (9.75 mM-1 cm-1). Anal.
(C8H15N3SO3) C, H, N, S.
2-(N,N-Dieth yla m in o)eth yl Acetoxym eth yl Eth er (5).
A solution of 711 mL (5.4 mmol) of N,N-diethylethanolamine
in 10 mL of tetrahydrofuran was flushed with dry nitrogen.
To this solution was added 198 mg (8.25 mmol) of sodium
hydride and the resulting reaction mixture was heated at
reflux for 1 h. The mixture was cooled to 0 °C under a stream
of nitrogen, whereupon 833 mg (5.4 mmol) of bromomethyl
acetate was introduced slowly through a syringe and stirred
overnight. Methanol (500 mL) was added and the mixture was
stirred for 15 min. The reaction mixture was concentrated on
a rotary evaporator; the residue was extracted with dichloro-
methane and the solution was dried over sodium sulfate,
filtered through a layer of magnesium sulfate, and evaporated
under vacuum to give 633 mg of a colorless oil. The crude
mixture containing roughly 44% of the desired product and
56% of 2-(N,N-diethylamino)ethyl acetate was fractionally
distilled under vacuum to give 120 mg of 5: bp 80 °C at 5
mmHg; NMR δ 1.05 (t, 6 H), 2.10 (s, 3 H), 2.63 (q, 4 H), 2.73
(t, 2 H), 3.75 (t, 2 H), 5.29 (s, 2 H). Anal. (C9H19NO3) C, H, N.
cells were harvested and washed twice in PBS, whereupon
assays for growth and apoptosis were done.
Cell Gr ow th a n d Ap op tosis Assa ys. The number of viable
cells was determined by Trypan blue exclusion. Apoptosis was
assayed by flow cytometry and by determining DNA laddering
using gel electrophoresis as previously described.7 For the flow
cytometry assay, we used the method of Nicoletti et al. with
slight modification.12 In brief, after two PBS washes, cells were
resuspended in 1 mL of lysis buffer consisting of 0.2% NP40
solution containing 5% bovine serum albumin in PBS. Ten
microliters of a 4 mg/mL propidium iodide solution and 10 µL
of a 10 mg/mL RNAse A solution were then added. The fraction
of cells with low DNA content (which is consistent with the
apoptotic fraction)12,13 was determined using a Becton-Dick-
enson flow cytometer.
For the DNA laddering assay, cells were cultured in suspen-
sion medium with the NO donors as outlined above. They were
then centrifuged and washed in PBS. DNA was isolated with
the DNA Blood Kit from QIAGEN (Chatsworth, CA) using the
manufacturer’s protocol. Five micrograms of DNA was elec-
trophoresed in a 1.2% agarose gel and stained with ethidium
bromide. NP40, RNAse A, and propidium iodide were from
Sigma (St. Louis, MO). Results are expressed as averages of
multiple experiments ( the standard error of the mean (SEM).
SEM was calculated as the standard deviation of different
measurements divided by the square root of the number of
measurements.
Ack n ow led gm en t. We thank Mary McGuire and
J ohn Klose for determining the NMR spectra. P.J .S. is
a Translational Research Awardee from the Leukemia
Society of America. This project was funded in part by
the National Cancer Institute under Contract No. NO1-
CO-56000.
Su p p or tin g In for m a tion Ava ila ble: Time course of NO
generation after adding esterase to aqueous 2 solutions (Figure
A) and dependence of NO generation rate on cell density in
cultures of HL-60 or U937 cells (Figure B). This material is
available free of charge via the Internet at http://pubs.acs.org.
Refer en ces
(1) Keefer, L. K.; Nims, R. W.; Davies, K. M.; Wink, D. A. “NO-
NOates” (1-Substituted Diazen-1-ium-1,2-diolates) as Nitric
Oxide Donors: Convenient Nitric Oxide Dosage Forms. Methods
Enzymol. 1996, 268, 281-293.
(2) Keefer, L. K. Nitric Oxide-Releasing Compounds: From Basic
Research to Promising Drugs. CHEMTECH 1998, 28, 30-35.
(3) Saavedra, J . E.; Dunams, T. M.; Flippen-Anderson, J . L.; Keefer,
L. K. Secondary Amine/Nitric Oxide Complex Ions, R2N-
[N(O)NO]-. O-Functionalization Chemistry. J . Org. Chem. 1992,
57, 6134-6138.
(4) Lefebvre, I.; Pe´rigaud, C.; Pompon, A.; Aubertin, A.-M.; Girardet,
J .-L.; Kirn, A.; Gosselin, G.; Imbach, J .-L. Mononucleoside
Phosphotriester Derivatives with S-Acyl-2-thioethyl Biorevers-
ible Phosphate-Protecting Groups: Intracellular Delivery of 3′-
Azido-2′,3′-dideoxythymidine 5′-monophosphate. J . Med. Chem.
1995, 38, 3941-3950.
(5) Saavedra, J . E.; Billiar, T. R.; Williams, D. L.; Kim, Y.-M.;
Watkins, S. C.; Keefer, L. K. Targeting Nitric Oxide (NO)
Delivery in Vivo. Design of a Liver-Selective NO Donor Prodrug
That Blocks Tumor Necrosis Factor-R-Induced Apoptosis and
Toxicity in the Liver. J . Med. Chem. 1997, 40, 1947-1954.
(6) The pKa of the thiol group in cysteine is 8.3. See: Zubay, G. L.
Biochemistry, 4th ed.; Wm. C. Brown: Dubuque, IA, 1998; p 64.
(7) Shami, P. J .; Sauls, D. L.; Weinberg, J . B. Schedule and
Concentration-Dependent Induction of Apoptosis in Leukemia
Cells by Nitric Oxide. Leukemia 1998, 12, 1461-1466.
(8) Griffith, O. W.; Kilbourn, R. G. Design of Nitric Oxide Synthase
Inhibitors and Their Use to Reverse Hypotension Associated
with Cancer Immunotherapy. Adv. Enzyme Regul. 1997, 37,
171-194.
(9) Shahidi, H.; Kilbourn, R. G. The Role of Nitric Oxide in
Interleukin-2 Therapy Induced Hypotension. Cancer Metastasis
Rev. 1998, 17, 119-126.
(10) Maragos, C. M.; Morley, D.; Wink, D. A.; Dunams, T. M.;
Saavedra, J . E.; Hoffman, A.; Bove, A. A.; Isaac, L.; Hrabie, J .
A.; Keefer, L. K. Complexes of ‚NO with Nucleophiles as Agents
for the Controlled Biological Release of Nitric Oxide. Vasore-
laxant Effects. J . Med. Chem. 1991, 34, 3242-3247.
Mea su r em en t of NO Gen er a tion Ra tes. Generation of
molecular NO was quantified as previously described1 to obtain
the solution chemistry data of Figure A (Supporting Informa-
tion) and associated text.
For the cell culture experiments, to avoid the confounding
effect of nitrate in the medium or serum, cells were cultured
at 37 °C in a 5% CO2 humidified atmosphere in serum-free
PBS and at a density of 25 000-150 000 cells/mL as indicated
in the individual experiments. The NO donors were added at
the indicated concentrations. In parallel, NO donors were
added at the same concentrations in cell-free PBS and incu-
bated under the same conditions. At the indicated time points,
100-µL aliquots of the supernatant were removed and used
for measurement of nitrite levels after treatment with nitrate
reductase using the Griess reaction as previously described.11
Cell Lin es a n d Cu ltu r e Con d ition s. HL-60 and U937
cells were from ATCC (Rockville, MD). For the cell growth and
apoptosis experiments, cells were cultured at a density of
150 000 cells/mL in RPMI-1640 with 10% fetal bovine serum
at 37 °C in a 5% CO2 humidified atmosphere. Cells were
cultured in a total volume of 1 mL with one well per variable.
The NO donors were added at the indicated concentrations at
the time of culture initiation. At the indicated time intervals,

Esterase-Activated Nitric Oxide Prodrugs
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 2 269
(11) Granger, D. L.; Anstey, N. M.; Miller, W. C.; Weinberg, J . B.
Measuring Nitric Oxide Production in Human Clinical Studies.
Methods Enzymol. 1999, 301, 49-61.
(13) Darzynkiewicz, Z.; Bruno, S.; Del Bino, G.; Gorczyca, W.; Hotz,
M. A.; Lassota, P.; Traganos, F. Features of Apoptotic Cells
Measured by Flow Cytometry. Cytometry 1992, 13, 795-808.
(12) Nicoletti, I.; Migliorati, G.; Pagliacci, M. C.; Grignani, F.;
Riccardi, C. A Rapid and Simple Method for Measuring Thy-
mocyte Apoptosis by Propidium Iodide Staining and Flow
Cytometry. J . Immunol. Methods 1991, 139, 271-279.
J M9903850