Direct measurement of actual levels of nitric oxide (NO) in cell culture conditions using soluble NO donors

Article abstract of DOI:10.1016/j.redox.2016.05.002

Applying soluble nitric oxide (NO) donors is the most widely used method to expose cells of interest to exogenous NO. Because of the complex equilibria that exist between components in culture media, the donor compound and NO itself, it is very challenging to predict the dose and duration of NO cells actually experience. To determine the actual level of NO experienced by cells exposed to soluble NO donors, we developed the CellNO Trap, a device that allows continuous, real-time monitoring of the level of NO adherent cells produce and/or experience in culture without the need to alter cell culturing procedures. Herein, we directly measured the level of NO that cells grown in the CellNO Trap experienced when soluble NO donors were added to solutions in culture wells and we characterized environmental conditions that effected the level of NO in in vitro culture conditions. Specifically, the dose and duration of NO generated by the soluble donors S-nitroso-N-acetylpenicillamine (SNAP), S-nitrosoglutathione (GSNO), S-nitrosocysteine (CysNO) and the diazeniumdiolate diethyltriamine (DETA/NO) were investigated in both phosphate buffered saline (PBS) and cell culture media. Other factors that were studied that potentially affect the ultimate NO level achieved with these donors included pH, presence of transition metals (ion species), redox level, presence of free thiol and relative volume of media. Then murine smooth muscle cell (MOVAS) with different NO donors but with the same effective concentration of available NO were examined and it was demonstrated that the cell proliferation ratio observed does not correlate with the half-lives of NO donors characterized in PBS, but does correlate well with the real-time NO profiles measured under the actual culture conditions. This data demonstrates the dynamic characteristic of the NO and NO donor in different biological systems and clearly illustrates the importance of tracking individual NO profiles under the actual biological conditions.

Full text of DOI:10.1016/j.redox.2016.05.002

Redox Biology 9 (2016) 1–14
Contents lists available at ScienceDirect
Redox Biology
journal homepage: www.elsevier.com/locate/redox
Research Paper
Direct measurement of actual levels of nitric oxide (NO) in cell culture
conditions using soluble NO donors
n
Weilue He, Megan C. Frost
Department of Biomedical Engineering, Michigan Technological University, 309 Minerals and Materials Building, 1400 Townsend Dr., Houghton,
MI 49931-1295, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 27 April 2016
Received in revised form
Applying soluble nitric oxide (NO) donors is the most widely used method to expose cells of interest to
exogenous NO. Because of the complex equilibria that exist between components in culture media, the
donor compound and NO itself, it is very challenging to predict the dose and duration of NO cells actually
experience. To determine the actual level of NO experienced by cells exposed to soluble NO donors, we
developed the CellNO Trap, a device that allows continuous, real-time monitoring of the level of NO
adherent cells produce and/or experience in culture without the need to alter cell culturing procedures.
Herein, we directly measured the level of NO that cells grown in the CellNO Trap experienced when
soluble NO donors were added to solutions in culture wells and we characterized environmental con-
ditions that effected the level of NO in in vitro culture conditions. Specifically, the dose and duration of
NO generated by the soluble donors S-nitroso-N-acetylpenicillamine (SNAP), S-nitrosoglutathione
10 May 2016
Accepted 12 May 2016
Available online 16 May 2016
Keywords:
Direct NO measurement
In vitro cell culture
Complete media
Soluble NO donors
(
GSNO), S-nitrosocysteine (CysNO) and the diazeniumdiolate diethyltriamine (DETA/NO) were in-
vestigated in both phosphate buffered saline (PBS) and cell culture media. Other factors that were stu-
died that potentially affect the ultimate NO level achieved with these donors included pH, presence of
transition metals (ion species), redox level, presence of free thiol and relative volume of media. Then
murine smooth muscle cell (MOVAS) with different NO donors but with the same effective concentration
of available NO were examined and it was demonstrated that the cell proliferation ratio observed does
not correlate with the half-lives of NO donors characterized in PBS, but does correlate well with the real-
time NO profiles measured under the actual culture conditions. This data demonstrates the dynamic
characteristic of the NO and NO donor in different biological systems and clearly illustrates the im-
portance of tracking individual NO profiles under the actual biological conditions.
&
2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1
. Introduction
Nitric oxide (NO) is a very common immune regulator [1], cri-
lives of NO donors are commonly used to predict the level of NO
achieved and the time frame over which cells experience NO while
using these soluble donors, the potencies of NO releasing chemi-
cals are difficult to predict. The underlying issue with inaccurate
predictability of the potency of NO donors lies in the poorly
characterized NO levels actually achieved in the complex and
changing environments of biological conditions. It is of great im-
portance to directly understand how much actual NO cells ex-
perience rather than making assumptions regarding approximated
NO levels based only on the concentration and chemical properties
of the parent NO donors when trying to understand the role NO
plays in cellular behavior.
Some critical limitations associated with using soluble NO do-
nors that need to be recognized include: (1) the NO donor's ana-
lytical concentration does not equal to NO level that cells or tissues
experience; (2) NO level that cells actually experience is influ-
enced by species present within the buffer solutions and culture
media and by cells and tissues; (3) the NO donor treatment time
may not represent the true NO exposure time; (4) NO donor's
tical neurotransmitter [2] and potent vasodilator [3]. NO deficiency
is closely related to chronic cardiovascular diseases such as hy-
pertension, coronary heart diseases, and arterial thrombotic dis-
orders [4]. To restore physiologically useful levels of NO, NO re-
leasing drugs (NO donors) have been widely used for more than one
hundred years. Currently GSNO has been clinically used to prevent
thrombosis [5]. Additionally SNAP derivatives have also shown
considerable potential to maintain vascular tone [6,7]. However,
caution needs to be exercised when using NO as a drug, since NO
potentially brings about fatal side-effect such as shock [8].
Many experiments have shown that NO donors exhibited dif-
ferent potencies in different circumstances [9–12]. Although half-
n
Corresponding author.
E-mail address: mcfrost@mtu.edu (M.C. Frost).
http://dx.doi.org/10.1016/j.redox.2016.05.002
2213-2317/& 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

2
W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
biological effects should not be considered as equal to the effect of
NO.
(St. Louis, MO). Tert-butyl nitrite was purchased from Acros Or-
ganics, (Pittsburgh, PA). Gelatin was obtained from Bio-rad (Her-
cules, CA). Pyridine was purchased from EMD Chemical Inc.
Considering the high diffusivity and reactivity of NO, Lancaster
s
[13] proposed the following expression (Eq. (1)) to represent the
(Darmstadt, Germany). Ethidium bromide, Click-iT EdU assay kit
concentration of NO at any given site and given time point under
physiological conditions:
and Hoechst dye were purchased from Invitrogen (Grand Island,
NY). Smooth muscle cell line MOVAS, Dulbecco's modified eagle
medium (DMEM), fetal bovine serum (FBS), were all purchased
from ATCC (Manassas, VA).
2
⎡
⎤
∂
⎣NO⎦
∂
[NO]x,t
x,t
2
x,t
=
v − D
− k [O ][NO]
2
1
2
∂t
∂x
(1)
2.2. CellNO Trap device manufacturing
where v represents the rate of NO generation at a specific site and
2
⎡⎣ NO ⎤⎦
∂
x,t
time point, −D
is the diffusion of NO according to Fick's law,
Detailed fabrication and characterization of the CellNO Trap can
2
∂x
be found elsewhere [16]. Briefly, the semipermeable membrane for
the CellNO Trap was manufactured through top treating glass fiber
2
x,t
and−k [O ][NO] is the consumption of NO due to auto-oxidation.
1
2
Eq. (1) demonstrates that the ultimate NO level is determined by
the rates of NO generation, consumption and the distance between
the NO source and site of interest when the system is non-
homogeneous (i.e., layer of cells in a culture dish). The con-
centration of the NO donor can affect the generation rate and
greatly affects the ultimate NO level but is not equal to the ana-
lytical concentration of the NO donor. Biological systems are
complicated, where different species within the biological system
can react with NO and complicate the consumption rate of NO,
greatly influencing the ultimate NO level. NO itself has a very short
half-life. The half-life of NO varies from less than a second to
minutes depending on its concentration and specific environment
s
filter paper by manually casting 5% Sylgard hexanes solution
2
(v/v) for 3 times (72
μ
l/cm /cast) and air-drying. Then membrane
was placed into 60 °C oven for overnight heat to stiffen the poly-
mer. The surface and cross-section of the polymer treated foam-
like micro structure of the membrane are shown in Fig. 1A and B
by SEM. The membrane was cut according to chamber frames and
attached in between the two-chamber system by dropping to-
luene to temporally dissolve the plastic frame to adhere the
membrane. The final device is shown in Fig. 1 C. For cell culture
experiments, the device was sterilized by ethylene oxide and then
top treated with 20 g/ml collagen I solution for 2 h, or top coated
μ
with 1 mg/ml dopamine solution (10 mM Tris buffer, pH¼8.5) for
overnight, sterilized by ethylene oxide and treated by 2 mg/ml
gelatin solution (dissolve in PBS) for 1 h. Air-dried device will be
ready for cell culture.
[14,15]. This means that the duration of NO is primarily de-
termined by the generation process. Once the generation process
is over, there will be no further NO exposure (assuming other in-
termediate NO sources such as S-nitrosothiols are not formed). NO
donors consist of parent compounds that contain various func-
tional groups which remain after the release of NO and further
complicate the environment under study. Additionally some re-
actions that affects cellular response may not be directly through
NO such as direct S-NO transfer [9], which can easily be confused
with the effect of NO.
To fully understand the roles of NO and the factors that may
affect NO levels when using soluble NO donors, there is a great
need to understand how much NO is actually present at the cell
layer. Previously, our laboratory developed the CellNO Trap, a
chemiluminescence-based device for directly monitoring NO le-
vels within solutions and NO levels that cells produce during the
entire duration of experiments in real-time [16]. Herein, we used
this same device to directly measure and analyze the NO genera-
tion profiles of different commonly used NO donor compounds
2.3. SNAP synthesis and quantification
S-nitroso-N-acetylpenicillamine crystals were synthesized by
dissolving 200 mg of N-acetylpenicillamine (NAP) in 5 ml metha-
nol along with sonication. Two to three milliliter of HCl and 100
of concentrated H SO were slowly added in order to acidify the
solution. The acidified NAP solution was combined with 144.9 mg
NaNO and vortex-mixed avoiding light till all the sodium nitrite
was dissolved. After dark green/red color was gradually developed
SNAP solution), the solution was ice-cooled for 45 min. SNAP
crystals were collected by rotary evaporation and vacuum filtra-
tion, and washed by diH O repeatedly for 3 times and air-dried.
μl
2
4
2
(
2
SNAP content was tested by injecting known mass percent of SNAP
solution into excessive triiodide solution (both with and without
acidified sulfanilamide) according to Yang et al.[17].
(
SNAP, GSNO, CysNO, and DETA/NO) during in vitro culturing.
Different factors that influence these profiles including cell culture
conditions, pH/CO , free thiol levels, oxidative stress and solution
2.4. GSNO synthesis and quantification
2
volume were examined in detail with the CellNO Trap. A demon-
stration of using this device to illustrate the potencies of different
NO donors in inhibiting MOVAS cell proliferation to different de-
grees was introduced. This data indicates that the NO generation
profiles of all NO donors investigated are very dynamic. Real-time
NO data is a powerful tool to help explain observed biological data
regarding potencies of NO donors used in experiments.
GSNO synthesis was adapted from Hart et al. [18], and all the
synthetic procedures were shielded from light. Briefly, 1.54 g of
GSH was dissolved within 6.52 ml 1 M HCl and stirred on ice;
0
2
.345 g of NaNO dissolved in 1 ml water was dropped gently into
the GSH solution and to allow for reaction for 10 min; then 710
μl
newly prepared 10 N NaOH was carefully added to neutralize the
reaction system to pH between 3 and 4; the end product was
aliquoted and stored in ꢀ80 °C freezer until use. The produced
GSNO concentration should be close to 500 mM. Before each use,
one vial was taken out to measure the concentration by UV–vis
spectrometry at 335 nm, and the extinction coefficient used was
2
. Materials and methods
ꢀ
1
ꢀ1
2.1. Chemicals and cells
0
.92 mM cm
.
s
Silicone elastomer base & curing agent (Sylgard 184) were
2.5. CysNO synthesis and quantification
ordered from Dow Corning Co. (Midland, MI). Penicillin-strepto-
mycin (pen/strep), G418 disulfate, collagen I, N-acetyl-D,L-peni-
cillamine, acetic anhydride, calcein-AM, hydrogen peroxide, di-
tert-butyl peroxide, and cyclam were obtained from Sigma-Aldrich
CysNO synthesis was accomplished through nitrosating acid-
ified cysteine (Cys). Cysteine was dissolved in 1 M HCl (final con-
centration 0.2 M) and stirred on ice on the magnetic plate,

W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
3
Fig. 1. Illustration of the CellNo Trap experimental set-up. (A) SEM image of the surface the semi-permeable membrane; (B) SEM of a cross-section of the semi-permeable
membrane showing the foam-like structure which allows NO to pass through and provides a surface upon which cells can grow; (C) Illustration of the two-chamber NO
measurement system; (D) real-time measurement of NO that cells experience using chemiluminescence detection.
protected from light. Equimolar amounts of NaNO
2
was carefully
according to different needs) and calibration gas calibrated by
using 45 PPM NO standard (Air Liquid Healthcare America Corp.
Plumsteadville, PA) using 200 ml/min flow rate. Then the NOA
pipetted into the Cys solution and allowed to react for 10 min.
Freshly prepared 10 N NaOH was carefully titrated into the CysNO
solution until solution pH was between 3 and 4.The end product
was aliquoted and stored in ꢀ80 °C until used. Before each use,
one vial was taken out to measure the concentration with UV–vis
calibration constant was determined by injecting 100
μ
l of 500 M
μ
NaNO2 standard to stirring acidified Lugol solution. NO flux
through the CellNO Trap was measured according to our previous
published method [16]. In Brief, the CellNO Trap was connected to
the NOA. Data recording was initiated before adding the donor
solutions into the device to let the background signal stabilize.
Then 37 °C pre-warmed bathing solution with specified con-
centrations of the NO donor of interest, with or without other
chemicals such as free thiols or extra acid, were freshly prepared
and quickly added to CellNO Trap for measurement (as shown in
Fig. 1D).
ꢀ
1
ꢀ1
spectrometry at 543 nm (
ε
¼16.8 M cm ).
2.6. Cell culture and proliferation assays
Mouse smooth muscle cells (MOVAS) were cultured in DMEM
media, with 10% FBS, 1% pen/strep, and 0.2 mg/ml G418 in 37 °C 5%
CO incubator. Cells were seeded onto the cover-slip at an initial
2
2
density of 10,000 cell/cm then placed into 6-well plate for culture,
or cells were seeded directly into the surface of the CellNO Trap.
After overnight culturing to allow cell recovery, the specific NO
donor was added by freshly dissolving donor stock solutions into
the culture media to 2X specific concentrations and gently adding
equal volume of this 2X donor media solution to the volume of
culturing media containing cells. Immediately before using all the
NO donors, the concentration of the different NO donor stocks
were determined by UV–vis spectrometry. Then culturing of cells
continued for different durations, with a 24 h maximum time
frame. Immediately after the designated treatment time ended,
proliferating cells were labeled by using EdU imaging kit. In brief,
DNA synthesis was labeled by EdU (5-ethynyl-2′-deoxyuridine)
2.8. Statistical analysis
All NO measurement experiments were independently run
3
times unless specified otherwise. Cell proliferation and cell
number experiments were independently repeated in triplicate
with 3 samples for each treatment. Error bars represent sample
standard deviations among each repeat. The data was analyzed by
either student t-student test or one-way analysis of variance
(ANOVA). All statistical assays were achieved through R pro-
gramming unless specific notification.
incorporation, and this incorporation is visualized by alkyne-azide
s
(
EdU provides alkyne; Alexa Fluor dye provides azide) reaction,
3
. Results and discussion
which was performed after 3.7% formaldehyde fixation and 0.5%
Triton X-100 permeabilization. Nuclei were then stained by
Hoechst. Proliferating cells were imaged under Olympus BX51
microscope and cell number was quantified by ImageJ.
3.1. Definition and quantification of cellular NO level
Schmidt et al. [19] presented a straightforward method to cal-
2
.7. Real-time NO measurement through chemiluminescence NOA
culate NO concentration from soluble NO donors by considering
the NO generation rate of donors and the auto-oxidation of NO by
oxygen. The primary limitation of this method is the calculation is
only applicable to solutions with simple composition (where NO
consumption routes are limited), making its application to real
Before measurement, Sievers 280i Nitric Oxide Analyzer (GE
Instruments, Boulder, CO) was zero calibrated by using sampling
gas (either N gas, atmospheric air, or air within the incubator
2

4
W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
biological systems largely inaccurate. Herein, we introduce a
1 2
cells tightly attach to the membrane, J and J were considered to
be equal, such that:
ꢀ
2
method
to
measure
NO
flux
(in pmol cm
s
or
ꢀ
10
ꢀ2
ꢀ1
1
0
mol cm min ), which can be empirically measured by
J1 = J
2
(3)
using the CellNO Trap without making assumptions regarding NO
generation and consumption pathways or rates.
Also since the sampling process by a sweep gas in the lower
chamber is a continuous and very fast process, it is considered that
immediately after the NO travels through the membrane, the
signal can be detected by the NOA, regardless of whether N2 or
ambient air is used as the sweep gas, such that:
As illustrated in Fig. 1, the upper chamber serves as a conven-
tional cell culture environment to which drugs (including NO
donors) can be directly applied to cells using established cell cul-
ture protocols. While NO is present in the cell environment, NO
diffuses in all directions, and most importantly, through the
semipermeable membrane into the lower chamber. NO in the
lower chamber can be detected via chemiluminescence without
influencing activities in the upper chamber. Once a solution with a
specific NO donor is applied in the upper chamber, the actual NO
level (dose and duration) at the surface of the cell layer of this
particular donor can be tracked in real-time.
The key question in describing the NO levels cells experience is
what this measured flux (or NO level) means to cells cultured in
the device. To illustrate this, two models were shown below. First,
assume a tightly attached intact mono layer of cell is cultured onto
the device membrane to obtain the cell sheet model (see Fig. 2A):
J3 = Jt
(4)
We designed an experiment to investigate how much NO
consumption occurs between J and J . The highly controllable NO
2
3
releasing polymer SNAP-PDMS was cast on a cover-slip and top
coated with RTV-3140 PDMS layer. This SNAP-PDMS film releases
NO at a certain constant rate at specific temperature (due to
thermal degradation) [20]. By measuring the NO flux when the
polymer layer faces up from the upper chamber, the original NO
flux coming out of the polymer (J ) was obtained and facing down
p
and sampling from the bottom chamber, the NO diffusing across
the membrane (JTr) was detected (Fig. 3A). A membrane specific
NO permeability
η was defined as
Suppose J
0
is NO flux entering the cell;
Jp
J
J
J
1
2
3
is the NO flux exiting the cell;
is the NO flux entering the semipermeable membrane;
is the NO flux exiting the semipermeable membrane;
η = J ×100%
(5)
Tr
Result (Fig. 3B) shows that direct NO measurement and cross-
membrane measurement produced almost identical NO releasing
profiles. After 5 replicate measurements, the steady state NO flux
values were averaged. Values were applied to paired t-test
and J
t
is the NO signal measured by the NOA.
NO will be consumed through each interface such as J
This decay is represented as J , so:
c
0
1
to J .
(
Fig. 3C). Data showed that no detectable difference exists be-
tween the two sets of measurements using these two methods.
This implies that the relative NO permeability can be considered
00%, meaning there is negligible signal loss between J and JTr, i.e.
J0 = J + Jc
(2)
1
η
Predicting measured NO flux (J
will allow the actual NO level around cells to be known. Because
t
) at time (t) relative to J
0
or J
1
1
p
from J
2 3
to J , indicating:
J2 = J3
(6)
In this model, from Eqs. (3), (4) and (6), it is reasonable to use J
to represent J . So the NO level that was measured by the NOA is
equivalent to the NO flux coming from the cell sheet.
t
1
In the second model, if cells are not confluent (assume k% is the
fractional area covered in Fig. 2B), since Eqs. (4) and (6) are still
valid here, it is easy to obtain the following relation to represent J
t
:
J = J × k%+J ×(1−k)%
(7)
t
1
0
In addition, while the cell coverage ratio is very low and the
reaction occurred in the cells is negligible (i.e. a very small J ), such
that J and J might be considered as equal in the very low cell
c
t
0
t
density model. This NO flux J is used to reflect the dose and
duration of NO that cells actually experience when NO donors are
introduced in solution to the culture media.
3.2. NO release levels
The NO release levels (the level of NO and the time duration
over which this level is experienced) of four different commonly
used NO donors were examined by using the CellNO Trap device.
In order to elucidate the NO level in real biological conditions, all
the experiments were run at 37 °C in PBS buffer or complete cell
culture DMEM media. The same effective concentration of NO
donors (i.e. such that the same moles of NO were in theory gen-
erated) was used for direct comparison. Biologically irrelevant
concentrations were avoided, so 50 M of donor was used in most
μ
of the following experiments. Considering the reported NO donors’
Fig. 2. NO quantification in the confluent mono cell layer model (A) and the
nonconfluent cell culture model (B).
half-lives [13,21], it is estimated that these soluble NO donor

W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
5
SNAP-PDMS
Direct
test
Sweep gas
Sweep gas
To NOAs
Water seal
PDMS
membrane
Cover-slip
Through
PDMS
To NOAs
Water seal
Fig. 3. Evaluation of the cross-membrane signal dampening of NO. Panel (A) shows the experimental design of the evaluation of the NO cross-membrane signal dampening.
Panel (B) illustrates real-time NO flux data of both J and JTr; while panel (C) shows the comparison of direct measurement and cross-membrane measurement of NO. The
p
average NO flux (n¼5) at steady-state (defined as the NO flux from time 60–70 min after polymer was applied for measurement, where the NO signal became almost stable)
was calculated; the NO flux at steady-state measured by direct measurement was normalized to itself and the corresponding cross-membrane NO flux was normalized to
this value; P 40.9 according to paired t-test.
systems should not have an NO level that is likely to last days.
Therefore we only reported NO release profiles for the initial 24 h
after adding donors, though GSNO and DETA/NO both continued
releasing detectable amounts of NO after 24 h under some
conditions.
were dissolved in 10 ml freshly prepared pre-warmed complete
DMEM (with 10% FBS and 1% pen/strep) and the CellNO Trap was
placed in 37 °C incubator with ambient air for real-time NO
monitoring. It should be noted that CO
cubator for this initial work, but because the use of 5% CO
common condition for cell culture and it might potentially affect
the solution pH and carbonate concentration, the use of 5% CO
2
2
was not used in the in-
2
is a
Ten milliliter of PBS with 50
μ
M of CysNO, SNAP and GSNO and
2
5
μM of DETA/NO were applied to the NO measurement device in
separate experiments. Fig. 4A to D (blue traces) show their NO
release levels. By integrating the area under the release curves, the
total NO released in mole can be determined. The total NO re-
leased from these donors is recorded in Table 1. The reported half-
lives of each NO donor are also summarized in Table 1 [9,22–24].
CysNO showed the most rapid release, while GSNO and DETA/NO
generated a gradual and long-term release pattern. However, both
the reported SNAP half-life and the NO generation duration mea-
sured by the device showed large variations. Fig. 4E shows each
individual repeat for SNAP, which differs greatly from other do-
nors’ consistent and repeatable release profiles.
was considered a separate variable for investigation and will be
discussed below. Fig. 4 shows that NO profiles of each NO donor in
DMEM without CO
(50 M), the NO release profiles of all NO donors differed greatly
in PBS from the corresponding release in DMEM (without CO ). In
2
(red traces). At this specific concentration
μ
2
DMEM, the total NO that was captured by the sweep gas was
5.4 times and 11.0 times lower for CysNO and DETA/NO, respec-
tively, while SNAP and GSNO both showed that more total NO (1.3
and 1.2 times, respectively) compared with in PBS group within
24 h (Table 1). The change of NO flux also diverged; reduced in
CysNO, SNAP and DETA/NO but significantly increased in the GSNO
samples (Fig. 4A–D blue traces vs. red traces and quantified in
Table 1). Those results clearly demonstrated that different buffers
do significantly affect NO generation of all four NO donors
examined.
The inconsistencies of the NO release from SNAP also casts
doubt on how rational it is to use half-life of each NO donor
measured in PBS as the reference to predict the NO levels achieved
under real biological conditions. To further investigate this in a
more biologically relevant scenario, NO release profiles of different
NO donors dissolved in DMEM were directly measured. NO donors
2
When 5% CO was applied to the incubator atmosphere, making
the conditions more representative of typical cell culture

6
W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
Fig. 4. The NO release profiles of three different RSNOs and DETA/NO in different buffer conditions are shown. Ten milliliter of 50 μM CysNO, SNAP, GSNO and 25 μM DETA/
NO prepared in PBS or DMEM were applied to the CellNO Trap for real-time monitoring, where panels A through D are CysNO, SNAP, GSNO, and DETA/NO, respectively.
Under each condition, triplicate experiments were run independently. Data were presented as average (n¼3) with error bar representing the standard deviation of the three
trials. Panel (E) shows the NO release profile of 50 μM SNAP in 10 ml PBS. Since the release profiles of SNAP showed large variation among replicates, all repeats are presented
for readers’ reference. Red marks indicate the highest flux values, red arrows show sharp decrease of the NO flux, green arrow, sharp increase. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)
conditions, the NO generation levels were further affected. All the
RSNOs showed longer NO release duration in the DMEM with 5%
CO compared to samples run in DMEM without CO (CysNO and
2 2
SNAP released NO for approximately 2 and 20 h longer, Fig. 4B, red
and green traces, respectively), while DETA/NO released NO in a
significant faster fashion (Fig. 4D, red and green traces). The results
why the reported GSNO’s half-life is much longer than SNAP’s but
in some studies its potency did not follow this trend [9,24]. This
data clearly shows that simply using an NO donor's half-life
measured in a simple buffer solution to estimate the level and
duration of NO that cells experience is grossly inaccurate.
might be explained by the lower media pH brought about by CO
2
.
3.3. Factors that may affect NO release profiles of NO donors
GSNO and DETA/NO no longer constantly released low level NO for
over 24 h but released faster in the initial stage, and after 24 h, the
NO level can barely be detected. This might be one of the reasons
Other factors were also investigated to understand how the
level of NO that is present at the cell surface changes when soluble

W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
7
Table 1.
Quantitative analysis of the NO release profiles of different NO donors.
CysNO (50 μM )
0.023
7.37E-08 (4.76E-09)
2.49E-12 (1.61E-13)
2.12E-11 (1.89E-12)
0.62 (0.024)
SNAP (50 μM )
GSNO (50 μM )
DETA NONOate (25 μM )
Reported t1/2 (in PBS or PSS, h)
1.15, 4.6, Up to 6
8.87E-08 (4.87E-08)
4.89E-13 (2.92E-13)
1.11E-12 (6.72E-13)
4.04 (2.76)
159
20
PBS
w/out CO
2
Total NO (mol)
1.09E-07 (6.30E-08)
9.06E-14 (5.21E-14)
4.36E-13 (4.80E-13)
424 h*
9.30E-08 (1.30E-08)
8.11E-14 (1.14E-14)
1.19E-13 (1.32E-14)
424 h*
2
Average flux (mol/cm /s)
Max flux (mol/cm /s)
Duration (h)
2
DMEM
w/out CO
2
Total NO (mol)
Average flux (mol/cm /s)
Max flu (mol/cm /s)
Duration (h)
1.38E-08 (8.88E-10)
5.76E-13 (3.70E-14)
2.65E-12 (1.08E-12)
0.45 (0.056)
1.18E-07 (2.96E-09)
6.96E-13 (2.86E-14)
3.05E-13 (7.67E-15)
6.42 (0.12)
1.36E-07 (2.70E-08)
5.68E-13 (1.25E-13)
1.24E-12 (2.22E-13)
22.01 (0.83)
8.44E-09 (3.94E-09)
7.22E-15 (3.30E-15)
2.55E-14 (1.14E-14)
424 h*
2
2
with CO
2
Total NO (mol)
Average flux (mol/cm /s)
Max flux (mol/cm /s)
Duration (h)
7.43E-08 (1.14E-09)
5.07E-13 (1.49E-14)
1.73E-12 (6.81E-14)
2.44 (0.46)
1.10E-07 (2.78E-08)
7.11E-14 (8.50E-15)
1.25E-13 (1.41E-14)
424 h*
1.64E-07 (4.94E-09)
1.44E-13 (6.42E-15)
4.31E-13 (1.59E-14)
4.48 (0.80)
1.55E-07 (2.30E-08)
1.32E-13 (1.96E-14)
2.33E-13 (3.39E-14)
424 h*
2
2
The total moles of NO were calculated by integrating the area under the NO release curves; maximum NO flux was directly read from the NO releasing profile; average NO
flux refers to the entire releasing duration; t1/2 is the half-life of donors; * duration is longer than 24 h, however we only reported the initial 24 h. Data is presented as the
average and (Std. dev.).
NO donors were used. As previously described, the NO level cre-
ated by NO donors is determined by both the NO generation and
consumption reactions. In first order kinetics model, NO genera-
tion can be expressed as follows:
d[NO]
dt
⎡
⎤
=
k ⎣D⎦eNO.
1
(8)
And in NO auto-oxidation model, the ultimate NO level can be
influenced by the consumption described as:
d[NO]
dt
2
⎡
⎤
=
k ⎣D⎦e − k [O ][NO]
1
NO
2
2
(9)
where:
[
NO] concentration of NO at given time t,
t time,
is the rate constant for the decomposition of the donor,
D] is the concentration of the donor at a given time t,
k
[
1
eNO is the factor representing moles of NO released per mole of
donor,
] is the concentration of oxygen,
is the rate constant for the oxidation of NO by oxygen,
[O
2
k
2
Any factor that can change generation, consumption or in-
troduce other consumption mechanisms will change the actual NO
level obtained. Several known mechanisms can initiate the release
of NO from RSNO, including transition metal ions mediated NO
releasing, photocleavage, and ascorbate initiated NO release [25].
In addition to direct decomposition, RSNO can undergo transni-
trosation reactions with nucleophiles and other thiols without
generating any free NO, producing new RSNOs which may affect
the NO release profiles as well [25]. Specific factors (presence of
2
transition metals, free thiols, change in pH/CO , redox status, and
solution volume) were systematically investigated to understand
how they affect the ultimate NO level adherent cells experience
when these soluble NO donors are used.
Fig. 5. The NO release profiles of 50 μM of three different RSNOs in 10 ml PBS with
0 mM EDTA. Panel (A) shows the real-time NO release three independent ex-
1
3
.4. The effect of transition metal ions
periments for each RSNO were run and data was presented as the average of the
three, error bar representing standard deviation. Panel (B) shows the real-time
measured from time 0–1 h of 50 μM CysNO in 10 mM EDTA PBS (blue arrow shows
the signal peak); at 1 h, 1 mM ascorbic acid was applied (red arrow), initiating rapid
NO release. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)
PBS with 10 mM EDTA was used as the NO donor buffer to
eliminate the effect of the transition metal ions which are ubiquitous
contaminants in buffer salts. Using the same experimental set-up

8
W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
Table 2.
Quantitative analysis of transient ion effect on RSNO's NO generation.
1
50
20
CysNO
50 μM )
SNAP
(50 μM )
GSNO
(50 μM )
1
(
9
0
0
PBS
First 10 min NO re- 6.64E-08
4.69E-09
(2.32E-09)
1.53E-09
(5.97E-10)
8.88E-15
(2.06E-15)
1.46E-14
(3.22E-15)
8.24E-11
(2.49E-11)
56.9
2.13E-09
(2.07E-09)
3.53E-11
(7.80E-12)
5.43E-14
(2.84E-14)
8.99E-15
(1.71E-15)
2.05E-10
(1.47E-10)
10.4
lease(mol)
(4.12E-09)
1.12E-09
6
PBS with Total NO (mol)
EDTA
(2.41E-10)
2.97E-14
(6.75E-15)
1.64E-13
Average flux (mol/
cm /s)
30
0
2
Max flux (mol/
2
cm /s)
(2.42E-14)
0
1
2
3
4
5
6
7
8
9
10 11 12 13
First 10 min NO re- 4.74E-10
Time (h)
lease (mol)
(1.07E-10)
140.1
Ratio (with TIs/without TIs)
NO moles were calculated by integrating the area under NO releasing curve be-
tween 2 specific time points; first 10 min was chosen because in PBS solutions
CysNO's NO releasing in PBS and all RSNOs' releasing in 10 mM EDTA PBS became
almost undetectable. Data is presented as the average and (Std. dev.).
and conditions, SNAP’s NO generation profile under metal ion-free
conditions appeared to be significantly slower and more
repeatable (red traces in Fig. 5A). This result suggested that transi-
tion metal ions mediated NO releasing is the main release me-
chanism for SNAP decomposition in PBS, and without transition
metal ions, SNAP may be preserved for a much longer time. (After
SNAP was placed in 37 °C incubator for 48 h in 10 mM EDTA PBS,
1
mM ascorbic acid was applied to initiate NO release, which rapidly
Fig. 6. The NO level was changed by introducing free thiols, in panel (A), 10 ml of
00 μM DETA/NO solution was applied to the device for NO flux measurement until
the signal reached steady state. The NO flux increased immediately after introdu-
cing PBS and mixing as a control (red arrow), then 200 μM GSH was added (blue
arrow), causing the level of NO to sharply decrease but it returned to the initial
level of NO after approximately 4 h; Panel (B) shows the effect when 10 ml of
produced a huge amount of NO release, indicating the high stability
of SNAP in solutions without transition metal ions (data not
shown)). When transition metal ions were removed in CysNO and
GSNO solutions, similarly, significant decrease in NO generation
was observed in both groups (blue and green traces respectively
in Fig. 5A).
1
5
0 μM SNAP dissolved in PBS was applied to a solution of 500 μM Cys (n¼3). The
presences of the free thiol (Cys) made the level of NO produced by SNAP decom-
position more reproducible when compared to NO release without the free thiol
After 1 h, the NO generation from both CysNO and GSNO be-
came almost undetectable, but after adding ascorbic acid, NO
generation recovered (example shown in Fig. 5B). The amount of
NO generated in the initial 10 min by these three RSNOs in both
normal PBS and PBS with 10 mM EDTA are summarized in Table 2.
With transition metal ions RSNOs’ NO generation increased to
different degrees, where CysNO generated over 140 times more
total NO in the initial 10 min in normal PBS compared with PBS
with EDTA while the total NO generation only increased by ap-
proximately 50 times and 10 times respectively in SNAP and GSNO.
This demonstrates that CysNO is more sensitive to transition metal
ions than SNAP and GSNO. Removal of the transition metal ions
from RSNOs environment (even CysNO, which is normally con-
sidered to be highly reactive), can make RSNOs much more stable,
which is consistent with Singh et al.‘s [23] statement that SNAP’s
stability can be significantly longer in transition metal ion free
solutions. The NO generation from RSNOs in untreated buffers and
growth media may be mainly through the transition metal ion
mediated mechanism at 37 °C.
(see Fig. 4, Panel E). (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
Table 3.
Quantitative analysis of free thiols’ effect on RSNO's NO generation.
SNAP (50 μM )
PBS
Normal PBS
With Cys
Total NO (mol)
Average flux (mol/cm /s)
Max flux (mol/cm /s)
Duration (h)
Total NO (mol)
Average flux (mol/cm /s)
Max flux (mol/cm /s)
Duration (h)
8.87E-08 (4.87E-08)
4.89E-13 (2.92E-13)
1.11E-12 (6.72E-13)
4.04 (2.76)
1.26E-07 (2.30E-9)
3.39E-13 (4.56E-14)
5.38E-13 (9.02E-14)
7.54 (0.97)
2
2
2
2
NO moles were calculated by integrating the area under NO release curves. Data
presented as average (Std.dev.).
inhibition of NO release was reversible. To further investigate the
effect of thiol content on other donor’s NO generation profiles, the
same amount of GSH was added to the SNAP PBS solution, but no
significant change in the NO release profile was observed (data not
shown). In contrast, when the same amount of Cys was added to
SNAP, there was a significant reduction in total NO and average
flux (see Table 3).
Table 3 summarized the NO release parameters observed,
clearly showing that instead of greatly changing those numbers,
the addition of thiols significantly reduced the variations, making
the result more repeatable. Interestingly, in DMEM, the NO gen-
eration from SNAP is observed to be more repeatable and the
profile increases slowly and decreases slowly, indicating that thiols
3
.5. The effect of free thiols
The free thiols of GSH and Cys were added to the NO generating
solutions to achieve a final concentration of 500
μM to examine
their effect on the overall NO levels cells experience. It is noted
that high concentration of thiol (mM range) should be avoided
because of the potential to change the pH of the buffer. When GSH
was directly added to a solution of DETA/NO in PBS, which releases
NO in a continuous and near constant rate fashion at 37 °C, a
significant NO drop was observed (Fig. 6A), suggesting the ex-
istence of NO-GSH or DETA/NO-GSH interaction. However, after
approximately 4 h, the NO level started to return, suggesting this

W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
9
Table 4.
Quantitative analysis of pH's effect on DETA/NO NO generation.
DETA/NO (25 μM )
PBS pH¼7.4 PBS
pH¼6.0 PBS
Total NO (mol)
Average flux (mol/cm /s) 8.11E-14 (1.14E-14)
Max flux (mol/cm /s)
Duration (h)
Total NO (mol)
Average flux (mol/cm /s) 1.69E-13 (2.63E-14)
Max flux (mol/cm /s)
Duration (h)
9.30E-08 (1.30E-08)
2
2
1.19E-13 (1.32E-14)
424 h*
1.31E-07 (1.81E-08)
2
2
5.48E-13 (2.59E-14)
15.86 (1.19)
pH¼7.4 PBS with CO
2
Total NO (mol)
Average flux (mol/cm /s) 1.47E-13 (2.96E-14)
Max flux (mol/cm /s)
1.86E-07 (3.76E-08)
2
2
3.62E-13 (7.37E-14)
Duration (h)
424 h*
NO moles were calculated by integrating the area under the NO release curves. *
means that duration is longer than 24 h, however we only examined the initial
24 h. Data is presented as the average and (Std. dev.).
Fig. 7. The NO release profiles of 25 μM DETA/NO in PBS (pH¼7.4 and 6.0) and PBS
). Three independent experiments under each different con-
dition were run and data was presented as average of the three, error bars re-
presenting standard deviation.
(
pH¼7.4 with 5% CO
2
3.7. The effect of redox environment
may be a factor that account for this slowly increasing NO release.
Free thiols are important NO reactants in the body. They can
elongate the effective NO release and help with NO transportation
Oxidative condition of the cellular environment may also in-
fluence the ultimate NO levels obtained. It has been reported by
many groups that elevated ROS levels suppressed biological po-
tency of NO. The underlying mechanisms is not clearly understood,
but some ROSs directly react with NO, decreasing the NO that cells
can access and some are through other mediators [34,35]. The
influence of ROS (such as peroxide and superoxide) to NO level
was investigated by real-time monitoring NO from DETA/NO so-
[
[
26]. They also serve as important regulators of protein nitrosation
27]. Understanding the relationship between NO and thiols can
help us understand the underlying mechanisms affecting NO le-
vels present in biological systems.
lutions. Different peroxide sources were used (H
butyl peroxide(DTBP)) to complete this examination. There was no
significant decrease in NO level observed when H was applied
data not shown) and only very mild NO decrease was observed
2 2
O and di-tert-
3
.7. The effect of pH and CO
2
2 2
O
Different tissues have different pHs, which may result in different
NO generation profiles from the same NO donor [28,29]. Fig. 7
showed the NO release profile of 25 M DETA/NO in PBS solutions
(
when up to 1 mM DTBP was applied into 100 M DETA/NO solu-
μ
μ
tion (Fig. 8). And NO level returned to normal after approximately
with different pH's. At pH 6 although total NO that cells might ex-
perience was only around 1.4 times larger (within 24 h), the max-
imum NO flux increased around 4.6 times and NO release duration
was shortened significantly to around 16 h. Considering the biologi-
cal effect of NO is determined by the level and duration of NO de-
livery that specific cells experience, this data indicates specific NO
donors may have different potencies in different tissues or cell
compartments based on the pH of the target cellular environment.
0.5–1 h, indicating only a mild interaction of peroxide and NO.
Using the same principle, superoxide was applied to DETA/NO
solutions by using hypoxanthine-xanthine oxidase (HX/XO) sys-
tem [36]. Superoxide generation relies on the availability of the
substrate HX. Different concentration (100 M and 1 mM) of
μ
substrate was applied to the system to modulate the amount and
duration of superoxide. Data shows that immediately after ap-
plying superoxide, the NO signal quickly decreased, indicating that
superoxide greatly suppressed the resulting NO level (Fig. 9). The
NO level recovered gradually to the original level, and the time
needed for this recover is correlated to the amount of hypox-
anthine added.
Because blood is a pH buffered system that relies on the balance
-
between CO
2
and HCO
3
, it is hypothesized that 5% CO
2
within the
incubator may exert similar effects on the NO level by changing pH
of the solution. To test this, PBS buffer was applied to the CellNO
Trap and placed into the cell culture incubator with 5% of CO . Di-
2
rect measurement of the pH of the PBS shows that the solution pH
decreased to around 6.6 within 90 min (data not shown). Using
3.8. The effect of solution volume
pH¼7.4 PBS buffer to prepare 25
monstrates that CO has a significant effect on the NO generation
profiles (see Fig. 7). One fold increase of total NO and almost 3 times
greater maximum flux compared with without CO were observed
see Table 4). After 24 h, the NO signal was barely detectable. The
decomposition of diazeniumdiolates to generate NO has been
shown to be pH dependent [30]. Factors that affect the dissociation
rates of diazeniumdiolates include pH, concentration and the pre-
sence of metal ions [31,32]. Protonation is necessary for the de-
composition reaction and thus dependent on the pH [33]. The no-
tion that DETA/NO releases NO in a long-lasting fashion might be
true only in simple buffer conditions, but not in more complex
solutions such as culture medium. The micro-environments of
many inflammation situation and tumor masses can achieve rela-
tively low pH value in vivo (even lower than pH¼6 [28]), which is
likely to greatly change the effective NO delivery pattern.
μ
2
M DETA/NO with 5% CO , de-
In in vitro experiments, the quantitative study of NO is normally
2
accomplished by applying media with specific concentration of an NO
donor. However, most studies do not precisely specify the final volume
of solution in culture wells. Different solution volume may lead to
different final NO levels experienced by cell layer at the bottom of the
culture vessel, even when the amount of donor added was carefully
controlled due to the diffusivity and reactivity of NO through the vo-
2
(
lume of the liquid medium. To investigate this, 5, 10, 15 ml of 50 M
μ
CysNO DMEM solution was applied to the CellNO Trap and NO levels
sustained at the bottom of the culture vessel (which importantly
corresponds to the level experienced by an adherent layer of
cells) are shown in Fig. 10A. Keeping all other conditions constant and
varying only volume, 5 ml solution generated a peak flux of
ꢀ
2
ꢀ1
2.0 pmol cm min , the highest NO flux (P¼0.023, 0.022 compared
to 10 ml and 15 ml group respectively shown by * and , P40.05
between 10 and 15 ml by Tukey's-test), compared to approximately
Δ

10
W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
Fig. 9. The NO level decreased upon the introduction of superoxide. Hypoxanthine
1 mM in blue and 100 μM in red) was added into 10 ml 100 μM DETA/NO PBS
(
solution pH¼7.4 (with 10 mU/ml xanthine oxidase, time points indicated by black
arrows). All experiments were run at 37 °C. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
Fig. 8. The NO level at the surface where cells grow in the CellNO Trap decreased by
introducing peroxide. Panel (A) shows that the NO level changed in accordance with
the addition of peroxide (blue arrow) compared to the PBS (red arrow); Panels
(
B) through (D) illustrate three repeats of the experiment (blue curves representing
with peroxide, red the control). Ten milliliter of 100 μM DETA/NO solution were
applied to the device for NO flux measurement. Once the NO flux reached the steady
state, t-butyl peroxide was added into the solution (to the final concentration of
Fig. 10. Different media volumes illustrate an effect on the NO level experienced by
the cultured cells at the bottom of the CellNO Trap. Panel (A), shows the NO release
profiles of different volumes of 50 μM CysNO in DMEM with 5% CO ; three in-
2
dependent experiments of each volume were run and data was presented as
average of the three, error bar representing standard deviation. Panel (B) shows the
change of the total NO and NO flux experienced by cells along with the change of
the media volume, * and Δ indicate a significant difference of the total NO in be-
tween 5 ml and 10 ml, and between 5 ml and 15 ml respectively, Po0.05 by AN-
OVA and Tukey's test; star indicates a significant difference of the maximum NO
flux between 5 ml and 15 ml, Po0.05 by ANOVA and Tukey's test.
1
mM). All experiments were run at 37 °C. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
ꢀ
2
ꢀ1
1
.8 and 1.7 pmol cm min for 10 and 15 ml, respectively. There is a
decrease in the NO flux level with increasing solution volume, while
the total accumulative NO that cell experienced is greater for the larger
volumes. The total NO delivered was 73,000, 76,000 and 78,000 pmol
for 5, 10 and 15 ml, respectively (see Fig. 10B (P¼0.0370 between
5
and 15 ml groups shown by ⧋, P40.05 in other groups in Tukey's-
3.9. Different NO donors have different ultimate effect on MOVAS cells
test)). Although it is expected that the larger volume of solution would
deliver a larger total dose of NO, there is a striking difference in the
real-time flux resulting from different volumes of solution.
In the previous section, NO release profiles of the four NO do-
nors in real culture conditions (DMEM at 37 °C and 5% CO ) were
2

W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
11
Fig. 12. The real-time measurement of the NO levels that cells experienced when in
panel (A) they were cultured in the collagen surface-treated CellNO Trap and
treated with 10 ml of media with 50 μM CysNO, SNAP, GSNO or 25 μM DETA/NO;
and panel(B) shows cells cultured on the polydopamine surface-treated CellNO
Trap when treated with 10 ml of media containing 50 μM CysNO, SNAP, GSNO or
25 μM DETA/NO. Under each condition, triplicate experiments were run in-
dependently. Data was presented as the average value of the triplicate; error bar
represents the standard deviation of the three values.
(
P¼0.0565 by ANOVA and Po0.05 by t-test), while cell pro-
liferation ratios of all other groups were slightly smaller but close
to the control group (P40.2 by ANOVA).
Fig. 11. Different NO donors exert different effects on MOVAS cell proliferation. Cell
2
seeding density was 10,000 cell/cm and before applying the NO donors, cells were
The CellNO Trap was used to obtain real-time NO measure-
ments during the actual cell culture to allow correlation the level
of NO experienced by the cells with the cell proliferation. Result
and quantification analysis were summarized in Fig. 12A and Ta-
ble 5. This clearly shows that, with cells (under actual cell culture
conditions) GSNO did not release NO in low and long-lasting
fashion as people normal think, but DETA/NO did. Instead GSNO’s
release was more similar to CysNO, which released most of its NO
in the initial 4 h of the experiment. Compared with the media only
condition, the average and maximum flux of all RSNOs increased
significantly (around 2, 4, and 9 times more for CysNO, SNAP and
GSNO respectively), though the total accumulated NO experienced
by cells did not change much (a little over 1 fold for CysNO and
only 50% more and 20% less for SNAP and GSNO). And both NO flux
and total NO diffusing to and through the cell layer greatly de-
creased in the DETA/NO treated group (only 30% and 15% com-
pared with media only conditions).
cultured overnight to allow recovery; 24 h after treatment, cells were screened for
live-dead assay and proliferation assay (A) Relative living cell number. ** represents
Po0.01 by ANOVA and Tukey's-test compared with the control (CTRL). (B) Cell
proliferation ratio. * represents P¼0.0565 by ANOVA and Tukey's-test but Po0.05
by t-test compared with the CTRL. Data represents the average of the 3 in-
dependent repeats and error bar represents standard deviation among.
measured. With this information regarding predicted NO fluxes,
the inhibition of MOVAS cell proliferation was studied [37–39].
The same four NO donors with the same effective NO con-
centration (50
μ
M for RSNOs, and 25 M for DETA/NO) were used
μ
to treat MOVAS cells. After 24 h both cell number and cell pro-
liferation ratio were examined as shown in Fig. 11. Result shows
that DETA/NO treatment had the least potent effect on restricting
cell numbers after 24 h. All three RSNOs showed significantly
greater inhibitory effect on cell growth and proliferation as in-
dicated by ANOVA and Tukey's test, however, Tukey's test did not
suggest any statistic difference existed among the three RSNO
groups, which is difficult to explain by the NO real-time profiles
This data clearly demonstrates that during actual cell culture
experiment, the NO donors may release NO in a very different
manner than predicted from the release profile obtained from the
controlled experimental conditions. Simply knowing that different
NO donors have different NO releasing profiles is not enough to
2
obtained using DMEM media with CO . Meanwhile cell prolifera-
tion assays implicated that after 24 h only SNAP treated cell pro-
liferation ratio differed significantly from the control group

12
W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
Table 5.
Quantitative analysis of cell culture NO profiles treated by soluble NO donors.
CysNO (50 μM )
SNAP (50 μM )
GSNO (50 μM )
DETA/NO (25 μM )
2.54E-08 (4.21E-09)
3.21E-14 (1.05E-14)
7.38E-14 (3.40E-15)
Complete media with
cells
Collagen I coating with
cells
Total NO (mol)
1.67E-07 (5.17E-
08)
1.71E-07 (2.39E-
08)
3.77E-13 (8.33E-
14)
5.92E-13 (7.42E-
14)
1.32E-07 (2.15E-
08)
4.70E-13 (3.35E-
14)
1.80E-12 (3.58E-
13)
Average flux (mol/cm2/ 1.07E-12 (2.56E-
s)
13)
2
Max flux (mol/cm /s)
4.96E-12 (1.18E-
12)
Duration (h)
Total NO (mol)
3.15 (0.98)
1.30E-07 (3.15E-
08)
9.46 (2.55)
1.87E-07 (4.92E-
08)
4.61E-13 (9.00E-
14)
8.40E-13 (2.94E-
14)
5.33 (2.20)
4.44 (0.43)
1.31E-07 (2.60E-
08)
5.95E-13 (1.05E-
13)
1.24E-12 (3.85E-
13)
4.06 (0.64)
14.60 (3.12)
9.30E-08 (3.20E-08)
Dopamine coating with
cells
Average flux (mol/cm2/ 6.86E-13 (4.68E-
s)
Max flux (mol/cm /s)
1.38E-13 (7.42E-14)
2.48E-13 (9.03E-14)
13.15 (1.06)
13)
2
2.67E-12 (5.46E-
1
3)
Duration (h)
2.76 (0.19)
Cells were cultured in either a collagen I top-coated or a dopamine-gelatin treated CellNO Trap device, respectively. NO donors were dissolved in media to twice the desired
final concentration in equal volume media to original cell culture media and applied to the cultured cells. Data is presented as the average and (Std. dev.).
explain some of their biological outcomes. It is also important to
remember that a donor’s NO delivery behavior is highly change-
able, and may result in a totally different profile when experi-
mental conditions are changed slightly. For example, CysNO and
GSNO are normally considered to have very different stability, but
actually both of them released NO rapidly in the initial stages of
value was at 12 h. DETA/NO showed a consistently high cell pro-
liferation ratios, but the resultant cell numbers were still smaller
than the controls (Fig. 13A). By using the CellNO Trap, it is de-
2
monstrated that an NO flux of around 0.4–1 pmol/cm /s on aver-
age showed good inhibition of MOVAS cell proliferation. A con-
stantly high flux (around 41.5 pmol lasting for over 2 h) may kill
the cells (data not shown). It has now been clearly illustrated that
the actual NO status experienced during the culture experiment is
critical, not the empirical values determined or measured under
other conditions (for example, it is not recommended to use half-
lives of NO donors’ measured in PBS or complete media even with
2
the in vitro experiment with a very high flux (41 pmol/cm /s,
Fig. 12A), and no NO was detected by the end of the 24 h treat-
ment. This result explains the data shown in Fig. 11 as to why the
cell number and cell proliferation ratio in CysNO and GSNO group
were very similar, and DETA/NO's least potency may be resulting
from the very low NO level generated during the actual cell culture
experiment. SNAP has an intermediate NO level in real cell culture
2
5% CO to explain observed biological data). The results presented
here directly show the significance of tracking the actual NO level
generated by NO donors in biological experiments throughout the
duration of the culturing experiments.
2
conditions (around 0.4 pmol/cm /s) and lasted a significant longer
time compared with other RSNO (persisted up to 12 h duration),
this might be one of the reasons why after 24 h the smallest cell
proliferation ratio was observed in SNAP group. But overall, the
SNAP treated group did not show significantly smaller cell number
after 24 h compared with the CysNO and GSNO groups. This ob-
served inconsistency between cell number and cell proliferation
ratio may result from the fluctuation of cell proliferation ratios
within the 24 h experiment time because of the changing of the
NO level. To better understand NO’s effect on MOVAS, instead of
doing a single end-point examination on cells, cells need to be
examined at multiple intermediate time points during the course
of the 24 h experiment.
Though NO’s biological roles became much clearer during the
past several decades, many contradictions and enigmas still exist.
One problem is that NO’s effect on cells in cell culture is hard to be
separated from the behavior of soluble NO donors. One systematic
study on biological effect of different RSNO donors was reported
by Mathews et al.[9]. Considering the general idea that RSNO’s
effect was through NO released from the cleavage of S-NO bond,
they designed gas-purging experiment which showed that re-
moving NO had negligible influence on smooth muscle relaxation,
suggesting that at least other activation pathways to soluble gua-
nylate cyclase (sGC) exist. They also showed that the correlation
between RSNO decomposition first-order half-life and the degree
of the smooth muscle relaxation to be low. However, the half-life
values they used were measured within 0.5 mM physiologic saline
solution, while the biological assays were in more complicated
systems. Wink et al. [10] showed NO's cyto-protective effect on
hydrogen peroxide stressed V79 cells and no protective effect on
superoxide stress, even after superoxide dismutase (SOD) was
added. However, in the same paper they also showed different NO
donors of the same working concentration had very different po-
tencies in counteracting peroxide mediated cytotoxicity, showing
that choosing the proper NO donor is critical. Using a system with
both HX/XO and RSNO showed more complicated behavior ac-
cording to Trujillo et al. [40], where peroxynitrite might form and
stress cells when simultaneously using these two chemicals in the
same system. Sodium nitroprusside (SNP) was shown to be cyto-
toxic because of the synergetic effect of peroxide and the release of
3.10. NO profile and cell proliferation at different time points
Cell number and cell proliferation ratios were examined at
different time points to assess the effect of the selection of time
points in end-point testing. According to the NO profile data Fig. 12
(
1
A), CysNO and GSNO’s NO release lasted for around 3–4 h, and at
1–12 h SNAP finished releasing NO, this lead to the examination
of cells at 4 h and 12 h and 24 h.
Cell number data clearly showed that even though at 24 h cell
numbers in three RSNO groups were similar, the change of cell
number/proliferation ratio was very different at the intermediate
time points (Fig. 13A). The increase of cell number in CysNO and
GSNO groups was mainly after 12 h and between 4 and 8 h the
increase is low, however, SNAP showed the inverse pattern, where
the increase was mainly in the first 12 h. DETA/NO has relatively
stable with a slow increase of cell number. And these results are
also consistent with the cell proliferation ratio data (Fig. 13B),
where in CysNO and GSNO treatments the smallest proliferation
ratios were observed at 4 h, and in the SNAP treatment, the lowest
ꢀ
other ions such as CN and iron complex [41], and 3-morpholi-
2 2
nosydnonimine (SIN-1) may facilitate the accumulation of H O
for cell cancer cytotoxicity [42,43]. So tracking the real-time NO
status is critical in helping researchers evaluate the net effect of

W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
13
NO generation [54]. And these processes are very likely to be pH
dependent. For RSNOs, lower pH potentially generates more HNO
2
þ
and NO , which are strong nitrosating agents, preventing NO
generation [55], while for diazeniumdiolates NO releasing is
through the protonation, which favors lower pH [33]. In biological
system, cases can be more complicated because of different en-
vironments such as ion strength [31], making final NO level diffi-
cult to predict. To investigate the net biological effect of NO, the
CellNO Trap device directly measures NO that cells experience,
helping us understand the biological function of NO.
4. Conclusions
The NO generation profiles of different NO donors (CysNO,
SNAP, GSNO and DETA/NO) at physiological relevant levels in dif-
ferent buffer conditions were examined, proving that NO genera-
tion profiles of all NO donors can be very dynamically affected by
their specific environments by using the CellNO Trap, a two-
chamber real-time NO measurement device developed by our lab
previously. This information allows the direct correlation of NO
level (not NO donor concentration or NOS level) with the observed
biological response. MOVAS cells was directly cultured in the de-
vice and treated with different NO donors, while at the same time
the actual NO experienced by cells during the whole culture was
tracked. The result showed that NO experienced by cells cannot be
simply predict or estimated by the NO donor half-life or from a
measured value obtained in PBS. The different effects of NO shown
by different NO donors on inhibiting MOVAS proliferation was
studies by using CellNO Trap, highlighting the importance of
choosing correct examination time points and tracking NO dura-
tion of release. To precisely study NO’s biological effect, it is highly
recommended to monitor and report real-time NO level of each
specific experiment to eliminate the reporting of complicating
circumstances that lead to inaccurate results.
References
[
[
1] C. Bogdan, Nitric oxide and the immune response, Nat. Immunol. 2 (10) (2001)
07–916.
2] E.M. Schuman, D.V. Madison, Nitric oxide and synaptic function, Annu. Rev.
9
Fig. 13. Different NO donors inhibited MOVAS cell proliferation to different degrees
at different time points. Panel (A) shows the relative cell numbers normalized to
cell number of control groups at 24 h and panel (B) shows the MOVAS’s cell pro-
liferation ratio under different NO donor treatments at different time points. Cell
seeding density was 10,000 cell/cm and before applying NO donors cells were
cultured overnight to allow recovery; 4, 12 and 24 h after treatment, cells were
stained for life-death assay and proliferation assay. Data represents the average of
Neurosci. 17 (1994) 153–183.
[3] L.J. Ignarro, Nitric oxide: a unique endogenous signaling molecule in vascular
biology (Nobel lecture), Angew. Chem. Int. Ed. 38 (13–14) (1999) 1882–1892.
4] M. Cooke, P. John, M. Dzau, J. Victor, Nitric oxide synthase: role in the genesis
of vascular disease, Annu. Rev. Med. 48 (1) (1997) 489–509.
5] E. Langford, R. Wainwright, J. Martin, Platelet activation in acute myocardial
infarction and unstable angina is inhibited by nitric oxide donors, Arter.
Thromb. Vasc. Biol. 16 (1) (1996) 51–55.
6] I. Megson, et al., Prolonged effect of a novel S-nitrosated glyco-amino acid in
endothelium-denuded rat femoral arteries: potential as a slow release nitric
oxide donor drug, Br. J. Pharmacol. 122 (8) (1997) 1617–1624.
[
[
2
3
independent repeats and error bar represents sample standard deviation.
[
[
7] I. Megson, et al., N-substituted analogues of S-nitroso-N-acetyl-D, L-peni-
cillamine: chemical stability and prolonged nitric oxide mediated vasodilata-
tion in isolated rat femoral arteries, Br. J. Pharmacol. 126 (3) (1999) 639–648.
NO in complicated systems. Instead of using new NO donors, some
groups developed NO delivery systems that can directly deliver
only NO into the biological systems [44–46]. Though problems still
exist such as the controllability, consumption of NO through the
biological system, biocompatibility and cytotoxic effect of NO’s
oxidative products, the use of NO releasing polymer and device is
still promising approaches to understand the effects NO exerts on
cellular systems.
A huge number of potential biological species can react with
NO, including ROS [34,35], thiols [47], lipid [48], DNA [49,50], and
proteins [51–53], which makes the actual NO level present in
cellular environments hard to predict. For example, the effect of
free thiol to NO is complicated. On one hand, thiols may react with
NO forming new RSNOs, which may increase the effective half-life
of NO but decrease the effective NO level. On the other hand, free
thiol can facilitate RSNO decomposition resulting in a more rapid
[8] A.B. Levine, D. Punihaole, T.B. Levine, Characterization of the role of nitric
oxide and its clinical applications, Cardiology 122 (1) (2012) 55–68.
[9] W.R. Mathews, S.W. Kerr, Biological activity of S-nitrosothiols: the role of nitric
oxide, J. Pharmacol. Exp. Ther. 267 (3) (1993) 1529–1537.
[10] D.A. Wink, et al., The effect of various nitric oxide-donor agents on hydrogen
peroxide-mediated toxicity: a direct correlation between nitric oxide forma-
tion and protection, Arch. Biochem. Biophys. 331 (2) (1996) 241–248.
11] K.B. Sandau, J. Fandrey, B. Brüne, Accumulation of HIF-1α under the influence
[
of nitric oxide, Blood 97 (4) (2001) 1009–1015.
[12] D.A. Wink, et al., Nitric oxide and some nitric oxide donor compounds en-
hance the cytotoxicity of cisplatin, Nitric Oxide 1 (1) (1997) 88–94.
[
13] J.R. Lancaster, A tutorial on the diffusibility and reactivity of free nitric oxide,
Nitric Oxide 1 (1) (1997) 18–30.
[14] B.G. Hill, et al., What part of NO don’t you understand? Some answers to the
cardinal questions in nitric oxide biology, J. Biol. Chem. 285 (26) (2010)
19699–19704.
[
15] B.Y. Owusu, R. Stapley, R.P. Patel, Nitric oxide formation versus scavenging: the
red blood cell balancing act, J. Physiol. 590 (20) (2012) 4993–5000.

14
W. He, M.C. Frost / Redox Biology 9 (2016) 1–14
[
[
[
[
[
16] W. He, M.C. Frost, CellNO Trap: novel device for quantitative, real-time, direct
oxygen species, Toxicol. Lett. 82 (1995) 221–226.
measurement of nitric oxide from cultured RAW 267.4 macrophages, Redox
Biol., 8 (2016) 383–397.
17] B.K. Yang, et al., Methodologies for the sensitive and specific measurement of
S-nitrosothiols, iron-nitrosyls, and nitrite in biological samples, Free Radic.
Res. 37 (1) (2003) 1–10.
[36] T. Sawa, T. Akaike, H. Maeda, Tyrosine nitration by peroxynitrite formed from
nitric oxide and superoxide generated by xanthine oxidase, J. Biol. Chem. 275
(42) (2000) 32467–32474.
[37] F.C. Tanner, et al., Nitric oxide modulates expression of cell cycle regulatory
proteins A cytostatic strategy for inhibition of human vascular smooth muscle
cell proliferation, Circulation 101 (16) (2000) 1982–1989.
[38] U.C. Garg, A. Hassid, Nitric oxide-generating vasodilators and 8-bromo-cyclic
guanosine monophosphate inhibit mitogenesis and proliferation of cultured
rat vascular smooth muscle cells, J. Clin. Investig. 83 (5) (1989) 1774.
[39] K.-I. Kariya, et al., Antiproliferative action of cyclic GMP-elevating vasodilators
in cultured rabbit aortic smooth muscle cells, Atherosclerosis 80 (2) (1989)
143–147.
18] T.W. Hart, Some observations concerning the S-nitroso and S-phenylsulphonyl
derivatives of L-cysteine and glutathione, Tetrahedron Lett. 26 (16) (1985)
2013–2016.
19] K. Schmidt, et al., Release of nitric oxide from donors with known half-life: a
mathematical model for calculating nitric oxide concentrations in aerobic
solutions, Naunyn-Schmiedeberg's Arch. Pharmacol. 355 (4) (1997) 457–462.
20] G.E. Gierke, M. Nielsen, M.C. Frost, S-Nitroso-N-acetyl-D-penicillamine cova-
lently linked to polydimethylsiloxane (SNAP–PDMS) for use as a controlled
photoinitiated nitric oxide release polymer, Sci. Technol. Adv. Mater. 12 (5)
[40] M. Trujillo, et al., Xanthine Oxidase-mediated decomposition of S-Ni-
trosothiols, J. Biol. Chem. 273 (14) (1998) 7828–7834.
(
2011) 055007.
[41] J.M. Campbell, et al., The interaction of sodium nitroprusside with peripheral
white blood cells in vitro: a rationale for cyanide release in vivo, Biochim.
Biophys. Acta – Gen. Subj. 1156 (3) (1993) 327–333.
[42] D. Gergeľ, et al., Increased cytotoxicity of 3-morpholinosydnonimine to HepG2
cells in the presence of superoxide dismutase role of hydrogen peroxide and
iron, J. Biol. Chem. 270 (36) (1995) 20922–20929.
[
21] J.S. Beckman, W.H. Koppenol, Nitric oxide, superoxide, and peroxynitrite: the
good, the bad, and ugly, Am. J. Physiol. – Cell Physiol. 271 (5) (1996)
C1424–C1437.
[
22] C.M. Maragos, et al., Complexes of ꢁ NO with nucleophiles as agents for the
controlled biological release of nitric oxide. Vasorelaxant effects, J. Med. Chem.
3
4 (11) (1991) 3242–3247.
[43] R. Farias-Eisner, et al., The chemistry and tumoricidal activity of nitric oxide/
hydrogen peroxide and the implications to cell resistance/susceptibility, J. Biol.
Chem. 271 (11) (1996) 6144–6151.
[
[
23] R.J. Singh, et al., Mechanism of nitric oxide release from S-nitrosothiols, J. Biol.
Chem. 271 (31) (1996) 18596–18603.
24] D.L. Mooradian, T.C. Hutsell, L.K. Keefer, Nitric oxide (NO) donor molecules:
effect of NO release rate on vascular smooth muscle cell proliferation in vitro, J.
Cardiovasc. Pharmacol. 25 (4) (1995) 674–678.
[44] C. Wang, W.M. Deen, Nitric oxide delivery system for cell culture studies, Ann.
Biomed. Eng. 31 (1) (2003) 65–79.
[45] M.A. Starrett, et al., Wireless platform for controlled nitric oxide releasing
optical fibers for mediating biological response to implanted devices, Nitric
Oxide 27 (4) (2012) 228–234.
[46] G.E. Romanowicz, et al., Novel device for continuous spatial control and
temporal delivery of nitric oxide for in vitro cell culture, Redox Biol. 1 (1)
(2013) 332–339.
[
[
25] D.L.H. Williams, The chemistry of S-nitrosothiols, Acc. Chem. Res. 32 (10)
(
1999) 869–876.
26] J.S. Stamler, et al., S-nitrosylation of proteins with nitric oxide: synthesis and
characterization of biologically active compounds, Proc. Natl. Acad. Sci. USA 89
(
1) (1992) 444–448.
[
27] V.G. Kharitonov, A.R. Sundquist, V.S. Sharma, Kinetics of nitrosation of thiols
[47] R. Radi, et al., Peroxynitrite oxidation of sulfhydryls. The cytotoxic potential of
superoxide and nitric oxide, J. Biol. Chem. 266 (7) (1991) 4244–4250.
[48] R. Radi, et al., Peroxynitrite-induced membrane lipid peroxidation: the cyto-
toxic potential of superoxide and nitric oxide, Arch. Biochem. Biophys. 288 (2)
(1991) 481–487.
[49] D.A. Wink, K.S. Kasprzak, DNA deaminating ability and genotoxicity of nitric
oxide and its progenitors, Science 254 (5034) (1991) 1001.
[50] T. Nguyen, et al., DNA damage and mutation in human cells exposed to nitric
oxide in vitro, Proc. Natl. Acad. Sci. USA 89 (7) (1992) 3030–3034.
[51] J.S. Stamler, Redox signaling: nitrosylation and related target interactions of
nitric oxide, Cell 78 (6) (1994) 931–936.
[52] F.J. Schopfer, P.R. Baker, B.A. Freeman, NO-dependent protein nitration: a cell
signaling event or an oxidative inflammatory response? Trends Biochem. Sci.
28 (12) (2003) 646–654.
[53] R. Radi, Nitric oxide, oxidants, and protein tyrosine nitration, Proc. Natl. Acad.
Sci. USA 101 (12) (2004) 4003–4008.
by nitric oxide in the presence of oxygen, J. Biol. Chem. 270 (47) (1995)
28158–28164.
[
[
[
[
28] I.F. Tannock, D. Rotin, Acid pH in tumors and its potential for therapeutic
exploitation, Cancer Res. 49 (16) (1989) 4373–4384.
29] J. Wike-Hooley, J. Haveman, H. Reinhold, The relevance of tumour pH to the
treatment of malignant disease, Radiother. Oncol. 2 (4) (1984) 343–366.
30] T. Hansen, A. Croisy, L. Keefer, N-nitrosation of secondary amines by nitric
oxide via the 'Drago complex', IARC Sci. Publ. 41 (1981) 21–29.
31] J.A. Hrabie, L.K. Keefer, Chemistry of the nitric oxide-releasing diaze-
niumdiolate (“nitrosohydroxylamine”) functional group and its oxygen-sub-
stituted derivatives, Chem. Rev. 102 (4) (2002) 1135–1154.
[
32] L.K. Keefer, Progress toward clinical application of the nitric oxide-releasing
diazeniumdiolates 1, Annu. Rev. Pharmacol. Toxicol. 43 (1) (2003) 585–607.
33] K.M. Davies, et al., Chemistry of the diazeniumdiolates. 2. Kinetics and me-
chanism of dissociation to nitric oxide in aqueous solution, J. Am. Chem. Soc.
[
1
23 (23) (2001) 5473–5481.
[54] T.-M. Hu, T.-C. Chou, The kinetics of thiol-mediated decomposition of S-ni-
trosothiols, AAPS J. 8 (3) (2006) E485–E492.
[55] M.K., Morakinyo, S-Nitrosothiols: Formation, Decomposition, Reactivity and
Possible Physiological Effects. 2010.
[
34] D.A. Wink, et al., Nitric oxide protects against cellular damage and cytotoxicity
from reactive oxygen species, Proc. Natl. Acad. Sci. 90 (21) (1993) 9813–9817.
35] D.A. Wink, et al., Nitric oxide (NO) protects against cellular damage by reactive
[