UV cell stress induces oxidative cyclization of a protective reagent for DNA damage reduction in skin explants

Article abstract of DOI:10.1016/j.freeradbiomed.2018.12.037

UV irradiation is a major driver of DNA damage and ultimately skin cancer. UV exposure leads to persistent radicals that generate ROS over prolonged periods of time. Toward the goal of developing long-lasting antioxidants that can penetrate skin, we have designed a ROS-initiated protective (RIP) reagent that, upon reaction with ROS (antioxidant activity), self-cyclizes and then releases the natural product apocynin. Apocynin is a known antioxidant and inhibitor of NOX oxidase enzymes. A key phenol on the compound 1 controls ROS-initiated cyclization and makes 1 responsive to ROS with a EC50 comparable to common antioxidants in an ABTS assay. In an in vitro DNA nicking assay, the RIP reagent prevented DNA strand breaks. In cell-based assays, the reagent was not cytotoxic, apocynin was released only in cells treated with UVR, reduced UVR-induced cell death, and lowered DNA lesion formation. Finally, topical treatment of human skin explants with the RIP reagent reduced UV-induced DNA damage as monitored by quantification of cyclobutane dimer formation and DNA repair signaling via TP53. The reagent was more effective than administration of a catalase antioxidant on skin explants. This chemistry platform will expand the types of ROS-activated motifs and enable inhibitor release for potential use as a long-acting sunscreen.

Full text of DOI:10.1016/j.freeradbiomed.2018.12.037

Author’s Accepted Manuscript
UV cell stress induces oxidative cyclization of a
protective reagent for DNA damage reduction in
skin explants
Jing Liu, Haizhou Zhu, Gurdat Premnauth, Kaylin
G. Earnest, Patricia Hahn, George Gray, Jack A.
Queenan, Lisa E. Prevette, Safnas F. AbdulSalam,
Ana Luisa Kadekaro, Edward J. Merino
www.elsevier.com
PII:
S0891-5849(18)31268-1
https://doi.org/10.1016/j.freeradbiomed.2018.12.037
FRB14097
DOI:
Reference:
To appear in: Free Radical Biology and Medicine
Received date: 23 July 2018
Revised date: 10 December 2018
Accepted date: 29 December 2018
Cite this article as: Jing Liu, Haizhou Zhu, Gurdat Premnauth, Kaylin G. Earnest,
Patricia Hahn, George Gray, Jack A. Queenan, Lisa E. Prevette, Safnas F.
AbdulSalam, Ana Luisa Kadekaro and Edward J. Merino, UV cell stress induces
oxidative cyclization of a protective reagent for DNA damage reduction in skin
e x p l a n t s , Free
Radical
Biology
and
Medicine,
https://doi.org/10.1016/j.freeradbiomed.2018.12.037
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.

UV cell stress induces oxidative cyclization of a protective
reagent for DNA damage reduction in skin explants
a
a
a
a
a
Jing Liu , Haizhou Zhu , Gurdat Premnauth, Kaylin G Earnest , Patricia Hahn , George Gray ,
b
b
a
c
Jack A. Queenan , Lisa E. Prevette , Safnas F AbdulSalam , Ana Luisa Kadekaro , Edward J
a,∗
Merino
a
Department of Chemistry, McMicken College of Arts and Sciences, University of Cincinnati,
Cincinnati, OH, USA
b
Department of Chemistry, University of St. Thomas, St. Paul, MN, USA
c
Department of Dermatology, College of Medicine, University of Cincinnati, Cincinnati, OH, USA
∗Corresponding author: Email address: merinoed@ucmail.uc.edu (Edward J Merino)
Abstract:
UV irradiation is a major driver of DNA damage and ultimately skin cancer. UV exposure leads
to persistent radicals that generate ROS over prolonged periods of time. Toward the goal of
developing long-lasting antioxidants that can penetrate skin, we have designed a ROS-initiated
protective (RIP) reagent that, upon reaction with ROS (antioxidant activity), self-cyclizes and
then releases the natural product apocynin. Apocynin is a known antioxidant and inhibitor of
NOX oxidase enzymes. A key phenol on the compound 1 controls ROS-initiated cyclization and
makes 1 responsive to ROS with a EC50 comparable to common antioxidants in an ABTS assay.
In an in vitro DNA nicking assay, the RIP reagent prevented DNA strand breaks. In cell-based
assays, the reagent was not cytotoxic, apocynin was released only in cells treated with UVR,
reduced UVR-induced cell death, and lowered DNA lesion formation. Finally, topical treatment
of human skin explants with the RIP reagent reduced UV-induced DNA damage as monitored by
quantification of cyclobutane dimer formation and DNA repair signaling via TP53. The reagent
was more effective than administration of a catalase antioxidant on skin explants. This chemistry
platform will expand the types of ROS-activated motifs and enable inhibitor release for potential
use as a long-acting sunscreen.
Graphical Abstract
Keywords: ROS, antioxidant, apocynin, reactive oxygen species, melanoma, UV,
sunscreen, radical
1

1
.
Introduction
Prolonged exposure to ultraviolet radiation (UVR) generates reactive oxygen species (ROS) forms
that are highly toxic. [1] UVR-mediated excitation of melanin and ROS imbalance are important
biochemical contributors to melanoma risk. [2] Counterintuitively, skin cells, especially
melanocytes, have benign biochemistry that generates and utilizes ROS (Figure 1) such as
hydrogen
peroxide. These normal biochemical functions include melanin production and protein signaling.
[
3] Recently, it has been found that UVR not only causes direct DNA damaging events but also
results in the generation of “dark” cyclobutane pyrimidine dimers (CPDs), which derive from
radicals long after exposure. [4] Melanin-based radicals, observed using electron paramagnetic
resonance experiments, can generate the more toxic hydroxyl radical (Figure 1) long after UVR
exposure. These radicals lead to the production of exotic DNA damage products that are both lethal
and mutagenic. [5] Currently available sunscreens, which are comprised of UVR blocking and
UVR absorbing compounds, combat excessive UV exposure but do not stop these dark damaging
events. In addition, marine life toxicity issues surrounding the use of UVR-blocking compounds
like oxybenzone and the finding that repetitive sunburns are strongly associated with poor
prognosis melanoma mean that new sun-protection agents are needed. There are two design
requirements: (I) minimal activity against benign ROS biochemistry and (II) a non-stoichiometric
or catalytic cellular effect to enhance activity. In this manuscript, we detail our first molecular
design that satisfies these criteria. The ROS-initiated protective (RIP) reagent reported here
releases the natural product apocynin, an inhibitor of NADPH oxidases, upon oxidative reactions
with ROS generated in cells by UVR and later forming “dark” damage events (Figure 1).
We were among the first to report ROS-activated chemotherapeutic agents. [6, 7] The most
common ROS-responsive chemistries are boron-based. Boron-based antioxidants, [8, 9] modified
gene-targeting agents, [10] and inhibitor pro-drugs have been reported. [11, 12] Alternatives are
needed for several reasons. First, aryl boronate esters readily oxidize in the presence of hydrogen
peroxide or other ROS forms in a Chan-lam-like reaction. [13] Given that the steady-state
concentration of peroxide in blood and cells is between 200 nM and 2 µM, [14] reported aryl
boronate esters are likely activated rapidly in vivo. We have used a much different approach in this
2

work that spans several molecular designs. [15] We designed a cytotoxic molecule a few years
back. [16] This molecule had an unusual mechanism in that it underwent intramolecular
cyclization and dehydrated in the presence of hydrogen peroxide. Then we designed an antioxidant
that was not cytotoxic but this molecule had modest cell effects. [17, 18] In this work we
hypothesized that biologically-relevant effects could be achieved by taking advantage of this
oxidative reaction pathway to eject a bioactive inhibitor of ROS-producing enzymes for UVR
induced protection (1, Figure 1). We reasoned if we could release an inhibitor of cellular oxidases,
like apocynin (red, Figure 1), then selectively initiated cellular protection would be possible. This
manuscript details our first investigations into addition of a bioactive molecule to generate
catalytic antioxidants.
We sought to design a RIP reagent that would activate and release a bioactive molecule that could
globally lower ROS selectively (SI Scheme 1 for mechanism). This is important because
unselective ROS reduction can be harmful. [19] A survey of literature identified several NADPH
oxidase inhibitors that act as general ROS reducers. We focused on a natural product called
apocynin (Figure 1), a ketone version of vanilla, that has been used in traditional medicines. Its
promiscuous biological properties are ascribed to the ability to inhibit oxidases in the monomeric
and multimeric states. [20, 21] Oxidases are major generators of cell ROS, [22] and recent
literature suggests that melanocytes and keratinocytes have high levels of NOX oxidases such as
NADPH oxidase 1. [23] The NADPH oxidase 1 holoenzyme is a major producer of UV-induced
ROS. [20] The oxidase is involved in activation of the PI3K/Akt signal pathway. [24] Apocynin
may inhibit activity of other oxidases as well. Despite this poorly understood and likely complex
mechanism, its biological effects in ROS reduction are exceptional. [25] Thus, despite the complex
biological profile of apocynin it presented the best possible chance for functioning in vivo.
3

Figure 1: ROS (brown) are normally present but UV radiation enhances formation of highly toxic species
underlined) that damage DNA. RIP reagents, such as the apocynin-linked compound 1 shown, eject an
(
oxidase inhibitor in high oxidative stress environments. Apocynin (red) binds NADPH oxidase-
associated factors NCF1 and NCF2 to limit further ROS production. Structure from PDB 1K4U in
reference 21.
2. Material and methods
2.1. Synthesis and characterization
Chemicals: Boc-sarcosine, trifluoracetic acid (TFA) and dimethylformamide, chloromethyl methyl
ether were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA). Apocynin, 4-
floro-3nitro phenol, pladium on carborn, 4′-Hydroxyacetophenone, 1-Fluoro-2-nitrobenzene,
Triethylsilane, HATU were purchased from Fisher scientific international, Inc. All solutions were
prepared with water purified by a Milli-Q system (Millipore, Bedford, MA, USA). All reagents
used for buffer preparation were of analytical grade.
Full synthetic details and NMR and MS characterization data are provided in Supplementary
material.
4

2.2. Oxidation study
2
.2.1 Materials:
All stock solutions (0.1mM) of compound were dissolved in 25mM phosphate with various pH
corresponding to the different working concentrations.
HPLC condition for oxidation studies: A Beckman Coulter’s HPLC system consisting of a dual
pump Model 126 with 32 Karat Software, a System Gold 168 detector and a System Gold 508
Auto Sampler was used. A reverse phase C-18 column (Synergi™ 4μm Hydro-RP 80Å, LC
Column 100×4.6mm, Ea.) was used. The mobile phase consisted of HPLC grade Acetonitrile
(Fisher Scientific), LCMS grade formic acid (Fisher scientific) and distilled water filtered through
a Millipore Milli-Q water purification system. Solvent A for the mobile phase was 95% water, 5%
Acetonitrile for oxidation study of 1 and 95% water, 5% Acetonitrile for oxidation studies of 1a-c.
Solvent B is 95% acetonitrile and 5% water. The gradient was 0% B for 4 minute and 95% B over
1
6 minutes. Flow was 1ml/min. A detection wavelength of 250nm and 285nm was used for
oxidation of 1.
2
.2.2 Methods:
HRP oxidation: A working solution (1mg 1 was dissolved in 30mL 0.08 mM hydrogen peroxide in
5 mM phosphate buffer, pH=7.0) was made. A 1mL sample was used as the no enzyme control.
2
The reaction was 1mL of working solution and 0.1U HRP. After 60 mins, 1mL ethyl acetate was
added, shaken, and extraction performed. The organic layer was analyzed by HPLC to quantify the
relative amounts of 1 and apocynin. Moles were calculated using a standard of each compound and
yields were calculated using integration. All determinations were carried out in triplicate on
different days. On each occasion and new standards and samples were made.
Fenton reagent oxidation: The above working solution was mixed with 10μL Fe-solution (10 mM
Fe(NH ) (SO ) •6 H O, 10 mM EDTA, pH 8) by centrifugation of Fe-solution placed on the cap of
4
2
4 2
2
an Eppendorf tube. After 60mins of reaction, the reaction was processed as before.
5

KO oxidation: 1mg 1 was dissolved in 30mL 25 mM phosphate buffer, pH=7. Solid KO was
2
2
premeasured into an 15mL falcon to give a final concentration of KO . After 60mins analysis
2
proceeded as before.
UV Sensitivity of 1: 1mg 1 was dissolved in 30mL 25 mM phosphate buffer, pH=7. 1mL was
irradiated for 10 SED (10 min) using an Oriel Sol-UV-6 Solar Simulator, Oriel Instruments. After
irradiation, analysis proceeded as before.
Oxidation studies of compound 2 were performed following the same procedure as 1.
2.3. Antioxidant studies
2
.3.1 pUC19 ASSAY
Reaction mixture were prepared with 1 μg/ μL pUC19, and 0.5 mM 2’-deoxyguanosine in 20mM
pH 7.4 phosphate buffer. Reactions with 1 had a final concentration of 50 μM. Fenton conditions
were as described above. After 20 minutes reactions were stopped by addition of 10 μl quench
solution (0.5 M EDTA and 0.75M 2-mercaptoethanol in 6X agarose gel loading buffer). A 1%
agarose gel was prepared containing 0.75 μg/mL ethidium bromide. From each reaction, 12 μL of
quenched reaction was loaded per well. Gel electrophoresis was performed at 120 mV for 45 mins.
Agarose gel images were analyzed by ImageQuant 6.0 software to determine the intensity of each
band. Percentage of non-damaged DNA was calculated in each sample, by comparing the band
intensity corresponding to super-coiled pUC19 in each sample, with DNA only control.
2
.3.2 ABTS Assay
ABTS was dissolved in water to a 7 mM concentration. ABTS radical cation (ABTS•+) was
produced by reacting ABTS stock solution with 2.45 mM potassium persulfate (final
concentration) and allowing the mixture to stand in the dark at room temperature for 24 h before
use. For the study, the ABTS•+ working solution was diluted with PBS, pH 7.4, to an absorbance
of 0.70 (±0.05) at 734 nm. Then 100μL working solution was mixed with 100 μL of various
concentrations of each solution in a 96 well microplate. After 6 minutes of incubation the plate was
6

read at 743nm for absorbance. Appropriate solvent blanks were also run in each assay. All
determinations were carried out at least three times including the standards and controls. The
percentage inhibition of absorbance at 734 nm is calculated by: (Ac-At)/Ac×100%. The EC50
based on the Concentration-response curve for the absorbance at 734 nm for ABTS•+ as a function
of concentration of antioxidant solution.
2.4. Drug releasing study in cells
2
.4.1 Cell Culture
6
In these studies neonatal human keratinocytes were plated at a density of 0.60x10 cells/100 mm
dish in EpiLife® supplemented medium (Gibco) nd grown to a density of 80% confluence. The
following experimental groups: 1) Control; 2) ssUVR; 3) 50 µM 1; and 4) ssUVR+50 µM 1 were
used unless otherwise noted.
2
.4.2 Efficacy of RIP Reagent in Cellular Environment
Cells were allowed to attach for 24 h and incubated with 1 for 18 h before irradiation with 8 SED
using a Oriel Sol-UV-6 Solar Simulator set to maximum power. Immediately after irradiation,
o
fresh medium was replaced and cells were kept in 5% CO humidified incubator at 37 C for 1.5 hr.
2
After 1.5hr, cells were washed in cold PBS and detached from the dish with cell scraper in 1ml of
cold PBS. Ethyl acetate was added 1:1 (v/v) to the cell suspension. Organic layer was then
extracted and dried. Samples were resuspended in 100µL of acetonitrile and analyzed by direct
injection into the HPLC as described above. All determinations were carried out at least three
times.
2
.4.3 Quantifying Damage by LC-MS
6
Fibroblast cells (8 x 10 per sample) in PBS were treated with 1 or PBS for 18 h at 37 C. Then,
sample cells were irradiated at 10 SED using an Oriel Sol-UV-6 Solar Simulator operating at 24

C. All treatments were performed in quadruplicate. To isolate genomic DNA cells were
trypsinized, washed, and lysed by freeze/thaw for four cycles. Then, DNA was isolated using a
7

Wizard Genomic DNA Purification Kit (Promega Corp) per kit instructions. DNA yield for each
sample was quantified by A260 in triplicate. Digestion of DNA occurred using, the following
procedure. DNA was denatured, placed on ice, and incubated with 0.2 U of bovine pancreas
DNase I (Millipore Sigma) in 0.03 M ammonium acetate (pH 5.) at 45 C for 2 h. The enzyme was
then precipitated and spun down using ammonium bicarbonate (pH 8.0). Next, 5 mU of
phosphodiesterase I (Worthington Biochemical Corp.) and 0.2 U of calf intestinal alkaline
phosphatase (Thermo Fisher) were incubated at 37 C for 2 h. Finally, enzymes were again
precipitated. and samples resuspended in 200 L of LC-MS mobile phase.
LC-MS analysis was performed with a SCIEX 4000 QTRAP triple quad mass spectrometer with a
turbo spray ion source interfaced with an Agilent 1260 HPLC. Injection volume was 5 L onto an
Agela Technologies Optimix C18/amide column (2.1 x 50 mm, 5 m) flowing at 0.2 mL/min in
HILIC mode at 20 C. Solvent A was 95:5 water:acetonitrile and Solvent B was 95:5
acetonitrile:water, both with 0.1% formic acid. The following gradient was used: 0 min, 100% B;
2
.5 min, 100% B; 7.6 min, 70% B; 9.4 min, 70% B; 10.4 min, 100% B. Selected ion monitoring in
+
positive ion mode was used to detect 8-oxo-7,8-dihydro-2’-deoxyguanosine ([M+H] ; m/z =
2
2
84.1). Peak areas were normalized per g of DNA.
.4.4 DCFDA assay by flow cytometry
2
Cells were grown in 25 cm tissue culture plates to 50% confluence. Then each plate was
incubated with the listed compound for 18 hours. Cells were trypsinized, added, incubated for 30
min, and cells washed with HBSS twice. Each sample was irradiated and then 5M DCF-DA
(Sigma Aldrich) was added. Incubation for 30 min, washing, and DCF fluorescence was obtained
in using BD FACS Canto flow cytometer by exciting at 485 nm and emission at 535nm after
eliminating 7AAD positive cells. Percentage of cells with high DCF fluorescence were compared
among different treatments. Data shown are the average of three biological replicates.
3. Result and discussion
8

We envisioned that the ROS-initiated release of apocynin would occur as shown in Figure 2 and a
·
more detailed mechanism is shown in SI Scheme 1. Because the phenol of apocynin is essential to
its unique biochemical mechanism, covalent attachment to the ROS-sensitive motif through an
would block apocynin activity except under oxidative stress conditions. To examine the reaction,
two derivatives were synthesized: the RIP reagent (1, structure shown) and a control molecule (2)
where the phenol is not present to prevent oxidation and subsequent cyclization. Compound 1 was
made in five steps with a net 25% yield and 2 was made in four steps in 45% yield (see SI for
NMRs and synthetic details)
The oxidation reaction of 1 was examined (Figure 2A) under Fenton conditions, which generate
hydroxyl radical. The Fenton conditions were 0.1 mM Fe(NH ) (SO ) 6H O, and 0.08 mM H O
2
4
2
4 2
2
2
in pH 7.0 phosphate buffer and the concentration of 1 was 0.1 mM. Extraction was performed
since apocynin is not soluble in buffered solutions. After extraction with ethyl acetate the aqueous
and organic layers were analyzed by HPLC. Under our conditions, 1 eluted near 10 min, whereas
ox
apocynin eluted near 12 min (Figure 2B). The second reaction product, 1 , is observed in the
ox
aqueous layer and has a characteristic absorbance at 375 nm as does a variant of 1 identified in a
previous manuscript. [17] Exposure to Fenton conditions in phosphate buffer led to formation of
ox
products 1 and apocynin, whereas in the absence of Fenton reagent no reaction was observed
(Figure 2B). Quantitative analysis of the organic layer indicated that under these Fenton
conditions, 85% (+/- 2%) of 1 was converted to products (Figure 2C). This is close to the expected
yield based on the concentrations of 1 and H O , indicating that the first oxidative equivalent of
2
2
Fenton reagent reacts with 1 completely. Both products were collected, and MS analysis confirmed
expected product identities (Figure 2C).
We next examined the reaction of 1 with various ROS forms encountered under cellular conditions
Figure 2D. After 1 hr in Fenton conditions, 15% (+/- 2%) 1 remained. In experiments lacking
Fe(NH ) (SO ) and only containing hydrogen peroxide we find an insignificant loss of 1 after 1
4
2
4 2
hr. Importantly, the reaction is diffusional and first order in both Fenton reagents (SI Figure 1).
This indicates that self-cyclization of 1 will be strongly correlated to the total amount present in
cells. After a 1 hr exposure to 0.08 mM KO 63% (+/-8%) of 1 remained. Oxidase activation was
2
examined using 0.01 mU of HRP enzyme and 0.08 mM H O . Under these conditions 30% (+/-
2
2
5
%) remained after 1 h. We examined the control compound 2, which lacks a phenol important to
9

oxidation, under the same conditions. Compound 2 was resistant to oxidation (Figure 2D, purple)
under the same conditions, in all cases.
ox
Figure 2: (A) Proposed reaction of 1 to release apocynin and generate 1 . The control compound 2
lacks an oxidizable phenol. (B) HPLC analysis of 1incubated in phosphate buffer (gray trace) and under
Fenton conditions (black trace); the organic (left) and aqueous (right) layers were analyzed. Pure Apo
ox
was also analyzed (red trace), and peak corresponding to 1 and Fenton reagents are indicated. (C) MS
of isolated products. (D) Sensitivities of 1 (black) and 2 (purple) to various ROS forms. (E) Antioxidant
capacity as measured by ABTS assay. The obtained EC50 values are shown. (F) Fraction of 1 in solution
not subjected to irradiation and irradiated with 10 SED.
We then assessed antioxidant activity of the RIP reagent 1 (Figure 2E for EC ). To determine
5
0
antioxidant capacity a 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) assay was
performed. In the assay, green ABTS (7 mM) radicals are induced from persulfate (2 mM), and the
ability of antioxidants to reduce color formation is evaluated over a concentration range to obtain
an EC value (SI Figure 2 for data). In the presence of 10 μM 1, rapid inhibition was observed,
5
0
whereas 2 at the same concentration had limited activity. Compound 1 had an EC value of 2.4
5
0
(
+/-0.1) μM, whereas the EC of 2 is 60 (+/- 1) μM. Apocynin, catalase, glutathione, and vitamin
50
C had EC values of 11.7 (+/- 0.4), 0.7 (+/- 0.1), 3.5 (+/- 0.2), and 7 (+/- 0.1) μM, respectively.
5
0
Thus, 1 has modest antioxidant ability relative to common antioxidants but these reactions are
1
0

required to release the bioactive molecule.
It is important that 1 be stable to UVR exposure (Figure 2F). To test this, 1 (100 μM) in phosphate
buffer (pH 7) was irradiated for a total of 10 Standard Erythemal Dose (SED) units. Under these
conditions, 1 is not degraded. In summary, these data show that 1 is oxidized by various ROS, has
antioxidant capacity similar to those of common antioxidants and that the oxidation liberates
apocynin, a bioactive oxidase inhibitor.
Next, we analyzed the ability of 1 to prevent DNA damage in an in vitro DNA nicking assay
(Figure 3A). The concentration of pUC19 plasmid was 1000 ng/μL. After 1 hr incubation under
Fenton conditions, 31% (+/-3%) of the plasmids were nicked. No nicking was observed under
Fenton conditions in the presence of 50 μM 1 or 1 without Fenton reagent.
We then evaluated the cytotoxicity of 1 to primary keratinocytes in culture. After three days of
incubation with 1 no cytotoxicity was observed except at the highest dose tested. At the high dose
of 100 μM, a 15% (+/- 3%) reduction in viability was observed (Figure 3B). We also examined the
effect of 2, which is not oxidizable, and a compound that is oxidizable but does not eject apocynin
(SI Figure 3). The five step synthesis is listed in SI section 4.3. As expected 2 showed limited
cytotoxicity while the non-oxidizable derivative of 1 showed more cytotoxicity than 1. To evaluate
what was occurring inside primary keratinocytes were treated with 50 μM 1 or not and irradiated.
The organic layer was extracted and quantified (SI Figure 4 for chromatograms). In cells treated
with 1 but not subjected to UVR, little apocynin was detected (0.03 +/- 0.02 pmol). In cells that
were treated with 1 and irradiated 0.97 (+/- 0.07) pmol of apocynin were detected (Figure 3C),
thereby confirming that self-cyclization and apocynin release occurs within a cellular environment.
Based on this analysis, the concentration of apocynin within the cells is approximately 28 μM, this
is a level that is known to cause inhibitory effects. We also observed 1 inside of cells at a
concentration of 190 μM. Thus, little apocynin is released without UVR treatment and UVR leads
to ~17% release of apocynin.
Irradiation of primary keratinocytes with 10 SED reduced viability to 63% (+/-5%) relative to
DMSO-treated control cultures (Figure 3D). Cells were also treated with 50 μM 1, 50 μM
apocynin, 50 μM glutathione, or 3 U catalase without irradiation. Both apocynin and catalase
1
1

slightly but significantly enhanced growth to 106% (+/-2%) and 110% (+/- 3%) relative to DMSO-
treated controls (p< 0.05). Neither 1 nor glutathione influenced viability at this concentration
(Figure 3D). Higher concentrations of apocynin, glutathione, and catalase were cytotoxic (data not
shown). After the 10 SED irradiation, compound 1 increased viability to 88% (+/-5%) relative to
the control, while irradiated cells pretreated with apocynin showed a more modest rescue to 75%
(+/-2%) (Figure 3D). 1 is two times better at entering a cell than apocynin (SI Figure 5), but does
not affect viability without activation. Next we examined if 2, which cannot be oxidized, rescued
cells from UVR stress. SI Figure 3 shows little rescue in the presence of UVR. Similarly we
examined a compound that is oxidizable but lacks apocynin release (SI Figure 3). Limited rescue
in the presence of UVR stress is observed.
We validated if apocynin was inhibiting NOX enzymes or is just an antioxidant(SI Figure 6). Cells
were treated with apocynin and the phosphorylation status of AKT, a downstream protein
2
4
dependent on NOX1 activity, quantified . An in-cell western for phosphorylation of AKT at
Ser473, which is phosphorylated downstream of NOX activity, was performed. Apocynin reduced
the amount of phosphorylated serine by 21 (+/-9)% in the presence of UVR. The p-value was less
than 0.02. To further confirm the ability of 1 to reduce oxidative stress in cells we examined the
amount of 8-oxo-7,8-dihydro-2’-deoxyguanosine relative to non-UVR treated cells via LC-MS
(Figure 3E). Because of the large quantity of cells needed primary fibroblasts were used. In this
assay the integrated intensities at 284 m/z were compared to the total amount of DNA recovered.
For simplicity total integrated values are reported per million. We found that UVR treatment raised
the relative level of the lesion to 2.03 (+/- 0.19) from 1.41 (+/- 0.05) Int/g DNA with a p-value
less than 0.01. Treatment of 1 prior to UVR lead to a reduction of the lesion back to 1.39 (+/-
0
.21), which is not statistically different compared to untreated cells. Finally, 1 alone does not
change the amount of lesion found. Subsequent confirmatory experiments utilized the common
DCF-DA oxidative stress assay (Figure 3F). Though DCF-DA has known issues, mainly dealing
with the amount of incubation time leading to indiscriminate activation of the dye[26], the same
trend in the data is observed. Thus, of the compounds tested, 1 had the smallest effect on cell
growth in non-irradiated cells and the ability to protect cells from UVR.
1
2

Figure 3: (A) Gel electrophoretic separation of supercoiled and nicked plasmid in the presence and absence
of Fenton reagent and compound 1. (B) Viability of primary keratinocytes in a range of concentrations of 1.
(
C) Production of Apo in skin cells treated with 50 µM 1 with and without irradiation with 10 SED. (D)
Viability of primary keratinocytes without (gray bars) and with 10 SED UVR (red bars) in the presence of
the indicated antioxidants. (E) Relative LCMS quantification of 8-oxo-7,8-dihydro-2’-deoxyguanosine
under UVR and 1 treatment. (F) Dichlorodihydrofluorescein diacetate fluorescence measurements under
UVR and 1 treatment.
We then used a human skin explant model to evaluate the effect of topical treatment with 1 on UV-
induced damage (Figure 4A). In this model, topical treatments, mimicking use of sunscreen, can be
tested without compromising the dermis. Discarded and deidentified human skin was obtained
from donors undergoing elective surgical procedures. Human skin was used as rodent models do
not recapitulate human biochemistry due to the amount of hair. Subcutaneous fat tissue was
removed from the skin, and skin was cut into 3.5 cm diameter circles and placed into culture dishes
with DMEM/F12 media without phenol red, supplemented with growth factors. The dermal side
was in contact with the medium, and the epidermal was exposed to air. A solution of 50 μM 1 or
DMSO control solution was spread on the skin and allowed to dry overnight. Note that skin is
much more resistant to UVR than naked cells so a 25 SED UVR exposure was used. After
exposure skin was homogenized, protein was extracted, and DNA damage signaling quantified by
western blotting. UVR induced strong expression of the DNA repair signaling protein p53 at 24 h
post irradiation. Treatment of 1 without UVR caused no increase in expression of p53. Importantly,
p53 expression was reduced by 46% (+/-5%) in skin treated topically with 1 and then irradiated
(Figure 4B).
1
3

Figure 4: (A) Representative image of skin explant. (B) Western blot analysis of p53 production in skin
explants treated or not with 1 and subjected or not to UVR. Actin protein shown as a control. (C)
Representative immunofluorescence images of skin explants treated with DMSO or 1 and subjected to 10
SED irradiation. Red indicates the presence of CPDs; nuclei were stained with DAPI (blue). (D)
2
Quantification of CPD-positive cells per mm of non-irradiated (black bars) and irradiated (red bars) skin
treated with DMSO (-), compound 1, or catalase (CAT).
We then directly evaluated DNA damage in skin explants by staining for cyclobutane dimers
(CPDs). Skin explants were treated with RIP reagent 1 (50 μM) or catalase (3U) and dried
overnight. At 24 h after 25 SED irradiation, explants were subjected to immunofluorescence
imaging (Figure 4C). Experiments were performed on three separate skin explants per condition
and quantification was performed using ImageJ software at 20 different locations per skin explant
(Figure 4D). Cells (identified by staining with DAPI) that co-localized with CPD signal (3 fold
above background) were scored as damaged. Without irradiation, few cells were positive for CPD
2
2
(
2 +/-2 cells/mm ). Upon UVR exposure, 180 (+/- 82) CPD-positive cells per mm were detected
in DMSO-treated control explants. Treatment with 1 without UVR did not have a significant effect
+
2
+
2
(
2 +/-2 CPD cells/mm ) but catalase treatment did (13 +/- 4 CPD cells/mm ), indicating some
resting reductive stress. After UVR exposure, catalase-treated samples had 160 (+/-60) CPD-
2
positive cells per mm , not significantly different from the control UVR-treated explants. In
2
contrast, treatment with 1 significantly reduced the number of CPD-positive cells per mm to 70
(+/- 20) relative to the control explants (p<0.0005). Thus, the RIP reagent 1 greatly reduced DNA
damage and damage signaling when applied topically to human skin explants.
1
4

In conclusion, we have developed an ROS-controlled antioxidant, RIP reagent 1, that after
oxidative reaction (antioxidant function) releases a NOX inhibitor known as apocynin. The ROS-
activation is dependent on a key phenol in the molecule that leads to self-cyclization and release of
the NOX inhibitor. The molecular design of the ROS-initiated protective reagent is an alternative
to the well-known boron ester and quinone methide chemistries. Reagent 1 did not alter viability of
cultured skin cells and prevented UVR-induced cell death to a similar extent as other tested
antioxidants. In a human skin explant model, 1 applied topically reduced the formation of
cyclobutane dimers and DNA repair signaling. Thus, the described RIP reagent is a catalytic
antioxidant activated by high oxidative stress environments like UVR irradiation and these
reagents have potential for uses in sunscreens.
Acknowledgements
We acknowledge the R. Marshall Wilson Mass Spectrometric Facility of the Department of
Chemistry, University of Cincinnati. This study was supported by Department of Defense grant
W81XWH-16-1-0489to EJM and ALK.
References
[
1] A.P. Schuch, N.C. Moreno, N.J. Schuch, C.F.M. Menck, C.C.M. Garcia, Sunlight damage to cellular DNA: Focus
on oxidatively generated lesions, Free Radic Biol Med 107 (2017) 110-124.
2] N.P. Smit, F.A. Van Nieuwpoort, L. Marrot, C. Out, B. Poorthuis, H. Van Pelt, J.R. Meunier, S. Pavel, Increased
[
melanogenesis is a risk factor for oxidative DNA damage—study on cultured melanocytes and atypical nevus cells,
Photochemistry and photobiology 84(3) (2008) 550-555.
[
3] L. Denat, A.L. Kadekaro, L. Marrot, S.A. Leachman, Z.A. Abdel-Malek, Melanocytes as instigators and victims
of oxidative stress, The Journal of investigative dermatology 134(6) (2014) 1512-1518.
4] S. Premi, S. Wallisch, C.M. Mano, A.B. Weiner, A. Bacchiocchi, K. Wakamatsu, E.J. Bechara, R. Halaban, T.
[
Douki, D.E. Brash, Chemiexcitation of melanin derivatives induces DNA photoproducts long after UV exposure,
Science (New York, N.Y.) 347(6224) (2015) 842-847.
[
5] L. Zheng, M.M. Greenberg, Traceless Tandem Lesion Formation in DNA from a Nitrogen-Centered Purine
Radical, J Am Chem Soc 140 (2018) 6400-07.
6] G. Li, T. Bell, E.J. Merino, Oxidatively activated DNA-modifying agents for selective cytotoxicity,
ChemMedChem 6(5) (2011) 869-75.
7] Y. Kuang, K. Balakrishnan, V. Gandhi, X. Peng, Hydrogen peroxide inducible DNA cross-linking agents:
targeted anticancer prodrugs, J Am Chem Soc 133(48) (2011) 19278-81.
8] D. Lee, S. Park, S. Bae, D. Jeong, M. Park, C. Kang, W. Yoo, M.A. Samad, Q. Ke, G. Khang, Hydrogen
peroxide-activatable antioxidant prodrug as a targeted therapeutic agent for ischemia-reperfusion injury, Scientific
[
[
[
1
5

reports 5 (2015) 16592.
9] T.T. Hoang, T.P. Smith, R.T. Raines, A Boronic Acid Conjugate of Angiogenin that Shows ROS‐ Responsive
Neuroprotective Activity, Angewandte Chemie 129(10) (2017) 2663-2666.
10] S. Mori, K. Morihiro, T. Okuda, Y. Kasahara, S. Obika, Hydrogen peroxide-triggered gene silencing in
mammalian cells through boronated antisense oligonucleotides, Chemical science 9(5) (2018) 1112-1118.
11] D. Chen, G. Zhang, R. Li, M. Guan, X. Wang, T. Zou, Y. Zhang, C.-R. Wang, C. Shu, H. Hong, Biodegradable,
[
[
[
Hydrogen Peroxide and Glutathione Dual Responsive Nanoparticles for Potential Programmable Paclitaxel Release, J
Am Chem Soc 140 (2018) 7373-76.
[
12] S. Daum, S. Babiy, H. Konovalova, W. Hofer, A. Shtemenko, N. Shtemenko, C. Janko, C. Alexiou, A. Mokhir,
Tuning the structure of aminoferrocene-based anticancer prodrugs to prevent their aggregation in aqueous solution,
Journal of inorganic biochemistry 178 (2018) 9-17.
[
13] C. Yik-Sham Chung, G.A. Timblin, K. Saijo, C.J. Chang, Versatile Histochemical Approach to Detection of
Hydrogen Peroxide in Cells and Tissues Based on Puromycin Staining, Journal of the American Chemical Society
40(19) (2018) 6109-6121.
14] H.J. Forman, A. Bernardo, K.J. Davies, What is the concentration of hydrogen peroxide in blood and plasma?,
Archives of biochemistry and biophysics 603 (2016) 48-53.
15] F.S. Thowfeik, S.F. AbdulSalam, M. Wunderlich, M. Wyder, K.D. Greis, A.L. Kadekaro, J.C. Mulloy, E.J.
1
[
[
Merino, A ROS‐ Activatable Agent Elicits Homologous Recombination DNA Repair and Synergizes with Pathway
Compounds, Chembiochem : a European journal of chemical biology 16(17) (2015) 2513-2521.
[
16] A.K. Vadukoot, S.F. AbdulSalam, M. Wunderlich, E.D. Pullen, J. Landero-Figueroa, J.C. Mulloy, E.J. Merino,
Design of a hydrogen peroxide-activatable agent that specifically targets cancer cells, Bioorganic & medicinal
chemistry 22(24) (2014) 6885-92.
[
17] S.F. AbdulSalam, P.N. Gurjar, H. Zhu, J. Liu, E.S. Johnson, A.L. Kadekaro, J. Landero-Figueroa, E.J. Merino,
Self-Cyclizing Antioxidants to Prevent DNA Damage Caused by Hydroxyl Radical, Chembiochem : a European
journal of chemical biology 18(20) (2017) 2007-2011.
[
18] S.F. AbdulSalam, F.S. Thowfeik, E.J. Merino, Excessive Reactive Oxygen Species and Exotic DNA Lesions as
an Exploitable Liability, Biochemistry 55(38) (2016) 5341-52.
19] X. Ao, S. Bright, N. Taylor, R. Elmes, 2-Nitroimidazole based fluorescent probes for nitroreductase; monitoring
reductive stress in cellulo, Organic & biomolecular chemistry 15(29) (2017) 6104-6108.
20] L.H. Tanriverdi, H. Parlakpinar, O. Ozhan, N. Ermis, A. Polat, N. Vardi, K. Tanbek, A. Yildiz, A. Acet,
[
[
Inhibition of NADPH oxidase by apocynin promotes myocardial antioxidant response and prevents isoproterenol-
induced myocardial oxidative stress in rats, Free radical research 51(9-10) (2017) 772-786.
[
21] P. Jantaree, K. Lirdprapamongkol, W. Kaewsri, C. Thongsornkleeb, K. Choowongkomon, K. Atjanasuppat, S.
Ruchirawat, J. Svasti, Homodimers of vanillin and apocynin decrease the metastatic potential of human cancer cells
by inhibiting the FAK/PI3K/Akt signaling pathway, Journal of agricultural and food chemistry 65(11) (2017) 2299-
2
306.
22] H.J. Forman, Redox signaling: An evolution from free radicals to aging, Free Radic Biol Med 97 (2016) 398-
07.
23] A. Valencia, I.E. Kochevar, Nox1-based NADPH oxidase is the major source of UVA-induced reactive oxygen
species in human keratinocytes, Journal of Investigative Dermatology 128(1) (2008) 214-222.
24] K. Ito, A. Ota, T. Ono, T. Nakaoka, M. Wahiduzzaman, S. Karnan, H. Konishi, A. Furuhashi, T. Hayashi, Y.
[
4
[
[
Yamada, Inhibition of Nox1 induces apoptosis by attenuating the AKT signaling pathway in oral squamous cell
carcinoma cell lines, Oncology reports 36(5) (2016) 2991-2998.
[
25] M.M. Castro, E. Rizzi, G.J. Rodrigues, C.S. Ceron, L.M. Bendhack, R.F. Gerlach, J.E. Tanus-Santos,
Antioxidant treatment reduces matrix metalloproteinase-2-induced vascular changes in renovascular hypertension,
Free Radic Biol Med 46(9) (2009) 1298-1307.
[
26] B. Kalyanaraman, V. Darley-Usmar, K.J. Davies, P.A. Dennery, H.J. Forman, M.B. Grisham, G.E. Mann, K.
Moore, L.J. Roberts II, H. Ischiropoulos, Measuring reactive oxygen and nitrogen species with fluorescent probes:
challenges and limitations, Free Radic Biol Med 52(1) (2012) 1-6.
1
6

Highlights

New antioxidants are needed to prevent melanoma by reducing skin radicals and
ROS



A ROS-induced protection reagent quenches ROS leading to release of apocynin
In cells apocynin was released only with UV and reduced DNA lesions
Topical treatment of human skin explants reduced pyrimidine dimers and p53
1
7