Research Article
Skin Pharmacol Physiol
DOI: 10.1159/000501731
A Novel Butylated Caffeic Acid Derivative
Protects HaCaT Keratinocytes from
Squalene Peroxidation-Induced Stress
Dan Zhou Xinchu Weng
School of Life Sciences, Shanghai University, Shanghai, PR China
Keywords
Peroxidized squalene · HaCaT keratinocyte · Caffeic acid ·
Butylated caffeic acid · Antioxidant · Inflammation
Abstract
Squalene is a major sebum lipid which is easily peroxidized
by ultraviolet A (UVA) irradiation, generating products that
can stress keratinocytes. Prevention of squalene photooxi-
dation with antioxidants could potentially help defend skin
from environmental stress. Previously, we have systemati-
cally characterized butylated caffeic acid (BCA) as a more li-
pophilic alternative of caffeic acid (CA). The current study
aimed to test the hypothesis that the lipophilic property of
BCA makes it a good candidate as a skin antioxidant. The
UVA-induced peroxidation of squalene and the antioxida-
tive activity of CA and BCA were measured for peroxide value
and further characterized with ultraperformance liquid chro-
matography coupled with ultraviolet and mass spectrome-
try analysis. Both CA and BCA showed strong antioxidant ca-
pacity during squalene peroxidation under direct UVA chal-
lenge. In HaCaT keratinocyte culture, BCA’s and CA’s impact
on the damage induced by peroxidized squalene (P-SQ) was
investigated through quantification of reactive oxygen spe-
cies (ROS) generation and expression of interleukin-1 (IL-1)
cytokines. Both BCA and CA decreased P-SQ-induced IL-1β
secretion in HaCaT cells. However, only BCA could reduce
© 2019 S. Karger AG, Basel
E-Mail karger@karger.com
www.karger.com/spp
Received: January 15, 2019
Accepted after revision: June 25, 2019
Published online: September 24, 2019
P-SQ-induced ROS generation as well as expression of in-
flammatory cytokines in the IL-1 family. This advantage in
biological efficacy could have been attributed to BCA’s supe-
rior intracellular bioavailability due to its higher lipophilicity
compared with CA, as indicated by higher intracellular con-
centration of BCA. Based on this observation, we propose
using BCA as a functional antioxidant to prevent sebum lipid
from peroxidation and its detrimental effect to skin.
© 2019 S. Karger AG, Basel
Introduction
The skin, the first and major interface between the
body and the external environment, protects the body
from various chemical agents and exogenous physical
carcinogens, including ultraviolet (UV) radiation. UV-
induced oxidative stress is associated with acute and
chronic detrimental cutaneous effects, which may result
in the development of skin malignancies [1]. Oxidative
stress can be described as a reductant-oxidant imbalance
resulting in the generation of excess oxidants or mole-
cules which accept electrons. Reactive oxygen species
(ROS) are free radical derivatives of oxygen (O2)-contain-
ing molecules, including superoxide anion (O2–), hydro-
gen peroxide (H2O2), hydroxyl radicals (OH), singlet ox-
ygen (O21), and peroxyl radicals (ROO) [2]. ROS oxidize
130.238.7.40 - 10/7/2019 3:30:23 PM
Xinchu Weng
School of Life Sciences, Shanghai University
Uppsala Universitetsbibl.
333 Nanchen Road
Shanghai 200444 (PR China)
Downloaded by:
E-Mail wxch @ staff.shu.edu.cn

O O
HO HO
OH OH
HO HO
a b
Fig. 1. Chemical structures of CA (a) and BCA (b). BCA, butyl-
ated caffeic acid; CA, caffeic acid.
the unsaturated bonds of lipids in the sebum to yield lip-
id hydroperoxides. ROS also generate toxic byproducts in
cellular metabolism, leading to the oxidation of cellular
macromolecules, such as membrane lipid, proteins, DNA,
and other components of the cell [3]. Numerous studies
have demonstrated that environmental and human-gen-
erated stressors, such as UV irradiation and hydrogen
peroxide, could enhance ROS generation in skin cells and
create an assortment of adverse impacts [4].
Squalene, triglycerides, and wax esters are the major
components of sebum lipids [3]. Sunlight exposure in-
duced the photooxidation of sebum lipids. Squalene is sus-
ceptibly peroxidized by ultraviolet A (UVA) irradiation
due to the presence of its six double bonds [5]. Generation
of squalene peroxides on the skin disrupts the skin barrier
function as an acute response and induces skin roughness
as a chronic response [1]. Squalene accounts for approxi-
mately 13% of total sebum lipids [3]. Therefore, many
studies have been carried out on the major components of
squalene peroxides and their negative effects in skin cells.
Squalene monohydroperoxide is the primary peroxidation
product of squalene, together with several byproducts,
when squalene is subjected to photooxidation [6]. It has
been reported that the concentration of squalene monohy-
droperoxide in forehead skin increased by about three
times after 3 h of sunlight exposure [7]. Peroxidized squa-
lene (P-SQ) has also been demonstrated to be involved in
the pathogenesis of certain skin conditions such as skin
aging, cutaneous autoimmune disease, and skin cancer [8].
Natural and synthetic antioxidants have been studied
to prevent the photooxidation process of squalene [7].
Polyphenols exhibit a wide variety of biological effects,
such as antibacterial, anti-inflammatory, hepatoprotec-
tive, and so on because of their great capability of scav-
enging free radicals and their antioxidative activity [9].
Polyphenols exert antioxidant capacity by scavenging ac-
tive radicals to restrain the initiation or stop the propaga-
tion of free radicals chain reactions [10]. Among polyphe-
2 Skin Pharmacol Physiol
DOI: 10.1159/000501731
nols, caffeic acid (CA; 3,4-dihydroxycinnamic acid), the
natural antioxidant in many plants and fungi, has a wide
range of pharmacological activities including anti-in-
flammatory, antioxidative, immune-modulatory, and an-
ticarcinogenic effects [11]. CA has shown significant anti-
inflammatory and antioxidative activity in skin cells in
previous reports [12]. CA is allowed to be added in skin
care products and cosmetics as an active ingredient, ac-
cording to International Nomenclature Cosmetic Ingre-
dient. Publications also studied the skin care effects of
various plant extracts containing CA and related cosmet-
ic applications [13]. However, CA’s carboxyl group in its
molecular structure contributes CA’s poor lipophilicity,
which constrains CA’s biological efficacy and application.
Thus, a novel derivate of CA with considerable fat solubil-
ity was developed as a new candidate substitute. Addition
of a tert-butyl group at the 3-position in the phenyl ring
could improve CA’s lipophilicity according to structural
chemistry [14]. The new compound is butylated caffeic
acid (BCA; (E)-3-(3-(tert-butyl)-4,5-dihydroxyphenyl)
acrylic acid, BCA). BCA’s significant antioxidative capac-
ity was validated with Rancimat experiment, deep frying
test, and 1,1-diphenyl-2-picrylhydrazyl radical-scaveng-
ing assay in a previous report [14]. The chemical struc-
tures of CA and BCA are shown in Figure 1. We therefore
propose that BCA is a more potent antioxidant to be ap-
plied on skin to prevent squalene peroxidation.
In this study, the protective effects of CA and BCA
were first evaluated during the photooxidation process of
squalene under UVA irradiation, and then actives were
applied in the HaCaT cells challenged with P-SQ. The two
approaches provide a comprehensive view on their com-
parative benefit against skin surface oxidative stress.
Materials and Methods
Materials
BCA was synthesized in our laboratory as described by Shi et
al. [14]. A spontaneously immortalized human keratinocyte cell
line (HaCaT) was purchased from American Type Culture Col-
lection (Manassas, VA, USA). Dulbecco’s modified Eagle’s me-
dium (DMEM) and Hank’s balanced salt solution (HBSS) were
purchased from Gibco (Grand island, NY, USA). Fetal bovine se-
rum (FBS) was acquired from Life Technologies (Carlsbad, CA,
USA). One percent penicillin (100 U/mL)/streptomycin (100 μg/
mL) were acquired from GE Healthcare Life Sciences (Boston,
MA, USA). A cell counting kit (CCK-8) was purchased from
Dojindo Laboratories (Kumamoto, Japan), and 2’,7’-dichloro­
fluorescein diacetate (DCF-DA) was purchased from Sigma-
Aldrich Chemical (St. Louis, MO, USA). A human interleukin-
1β (IL-1β) ELISA kit was purchased from BD Biosciences
(San Diego, CA, USA). A human IL-1α and human IL-1 receptor
130.238.7.40 - 10/7/2019 3:30:23 PM
Zhou/Weng
Uppsala Universitetsbibl.
Downloaded by:
antagonist (IL-1ra) ELISA kit was purchase from Luminex (Aus-
tin, TX, USA). The RNeasy plus Mini kit was purchased from
Qiagen (Valencia, CA, USA). The SYBR® Premix Ex TaqTM kit
was purchased from Takara (Dalian, China). All reagents and sol-
vents were of analytical grade.
Preparation of P-SQ
A mixture of 2 mL of squalene and isopropanol (1:5; v/v) was
irradiated with UVA at 2.09 J/cm2 using a UV crosslinker which
fitted with UVA-specific light source (UVP, Upland, CA, USA). In
order to measure the antioxidant capacity of CA and BCA, these
two actives were added to the above squalene-isopropanol mixture
at 0.02% (w/w), respectively. Afterwards, squalene samples were
irradiated for 0, 3, and 6 h, with or without antioxidant, before be-
ing collected for subsequent characterization.
Peroxide Value Determination
The peroxide value (POV) was measured as an index of oxida-
tion degree of irradiated squalene samples. Analysis of POV was
performed according to a method adopted and slightly modified
from Wang et al. [15]. In brief, P-SQ was dissolved in 30 mL acetic
acid-chloroform (3:2; v/v). Then, 1 mL saturated potassium iodide
solution was added and thoroughly mixed before 5 min incubation
protected from light. After dilution with 100 mL deionized water,
the mixture was titrated against sodium thiosulphate. The POV
was calculated as follows:
X = [(V – V0) × N × 0.129] / m,
where X is the POV (%), V and V0 represent the volumes of sodium
thiosulphate exhausted by the samples and the blank, respectively
(mL), and m is the mass of squalene.
Ultraperformance Liquid Chromatography Coupled with
Ultraviolet and Mass Spectrometry Analysis of P-SQ
A Waters ACQUITY ultraperformance liquid chromatogra-
phy coupled to Waters Quattro Micro mass spectrometry or a pho-
todiode array detector (UPLC-UV/MS) was used for qualitative
analysis of P-SQ (Waters, Milford, MA, USA). This method pro-
vides high sensitivity to analyze the diversity between different
samples. The separation was carried out by a reverse-phase LC
column (Waters ACQUITY UPLC BEH C18, 2.1 × 50 mm, 1.7 μm).
The mobile phase was methanol at a flow rate of 0.3 mL/min under
ambient temperature. A total of 2 μL samples were injected. The
MS analysis was operated using atmospheric pressure chemical
ionization in positive mode. The atmospheric pressure chemical
ionization probe temperature and ionization source temperature
were set at 300 ° C. The m/z detection was set at full mass scan mode
from m/z 300 to 900. The total analysis time was 4 min.
Cell Culture
HaCaT cells were cultured in DMEM supplemented with 10%
FBS and 1% penicillin (100 U/mL)/streptomycin (100 μg/mL) at
37 ° C in a humidified atmosphere of 95% air and 5% CO2. Cells
were passaged every 2 days and kept at 90% confluency.
Cell Viability
Cell viability was determined using a CCK-8 kit. HaCaT cells
in 0.1% FBS/DMEM were seeded into 96-well plates at a density of
1.5 × 104 cells/well. The medium was changed to treatment me-
dium containing P-SQ and actives 16 h post plating. After another
CA and BCA Reduce Direct Peroxidation
of Squalene under UVA Irradiation
24 h incubation, cell viability was quantified according to the man-
ufacturer’s protocol [16].
Evaluation of ROS Generation
The DCF-DA method was used to evaluate the generation of
intracellular ROS [17]. HaCaT cells at a density of 1.5 × 104 cells/
well in 96-well plates were challenged with P-SQ applied together
with or without actives. Following 24 h of incubation, the cells were
incubated with DCF-DA (10 μM, in HBSS) for 30 min. Then cells
were washed three times with HBSS. Fluorescence intensity was
measured at Ex: 488/Em: 525 nm using a multimode microplate
reader (Tecan Inc., Männedorf, Switzerland).
ELISA
HaCaT cells in 6-well plates were treated with P-SQ together
with or without actives in HBSS for 24 h. After the treatment, su-
pernatants were collected and subjected to ELISA quantification
of IL-1α, IL-1β, and IL-1ra, according to the manufacturer’s in-
structions.
RNA Isolation and Quantitative Real-Time PCR
Total RNA was isolated from HaCaT cells after the abovemen-
tioned treatment using the RNeasy plus Mini kit according to the
manufacturer’s protocols. The amount and quality of total RNA
were analyzed using a Nano Drop spectrophotometer (Thermo
Fisher, Waltham, MA, USA). Then mRNA expression was deter-
mined by quantitative real-time PCR using the SYBR® Premix Ex
TaqTM kit according to the manufacturer’s guidelines. PCR was
conducted at 95 ° C for 30 s, followed by 40 cycles of 95 ° C for 5 s
and 60 ° C for 34 s in an ABI 7500 real-time PCR system (Applied
Biosystems, Foster City, CA, USA). Reactions containing qRT-
PCR primer sets specific for human IL-1β (F: 5′-CTG TCC TGC
GTG TTG AAA GA-3′; R: 5′-TTC TGC TTG AGA GGT GCT GA-
3′), β-actin (F: 5′-GGA CTT CGA GCA AGA TGG-3′; R: 5′-GAA
GGA AGG CTG GAA GAG-3′), and YWHAZ (F: 5′-CCT GCA
TGA AGT CTG TAA CTG AG-3′; R: 5′-GAC CTA CGG GCT
CCT ACA ACA-3′) were used as internal control for normaliza-
tion.
Analysis of Intracellular Active Concentration
The uptakes of CA and BCA by HaCaT cells were analyzed with
UPLC-MS according to previous reports [18]. Briefly, HaCaT cells
in 6-well plates were incubated with 200 μM CA or BCA for 8 or
24 h. After the treatment, supernatant was collected and cells were
scraped into 1 mL of 75% aqueous ethanol. Considering CA’s
property in UPLC-MS, cells treated with CA were redissolved in
deionized water after nitrogen blowing for UPLC-MS detection.
The UPLC-MS analysis was performed with a Waters ACQUI-
TY UPLC BEH C18 column (2.1 × 50 mm, 1.7 μm) coupled to an
AB Sciex 6500 QTRAP MS (Framingham, MA, USA) equipped
with an electrospray ionization (ESI) source operating in negative
mode at 550 ° C. Mobile phase A was methanol and mobile phase
B was 0.1% formic acid in water. The following gradient elution
program at a flow of 0.3 mL/min was applied: 0–2 min, 5% B; 2–
2.5 min, 95% B; 2.5–3 min, 5% B. The injection volume was 10 μL.
The source parameters were as follows: capillary voltage, 5,500 V;
capillary temperature, 550 ° C; curtain gas, 30 psi; nebulizer gas,
50 psi; turbo gas, 60 psi; MRM transition for CA, 179.1/135.2
(CE = –25.0 V); MRM transition for BCA, 235.2/191.3 (CE =
–35.0 V). The total analysis time was 3 min.
130.238.7.40 - 10/7/2019 3:30:23 PM
Skin Pharmacol Physiol 3
DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
Table 1. POV of squalene samples with or without actives during Table 2. Comparison of peak areas of squalene samples under 6 h
UVA irradiation UVA irradiation in UPLC

Irradiation Control With CA With BCA Content Blank P-SQ (6 h)

time
control with CA with BCA
0h 0.058±0.0028 0.058±0.0028 0.058±0.0028
3h 0.144±0.0037 0.095±0.0031** Peak 1 (Rt = 0.46) 161,050 308,223 239,256 182,653
0.0873±0.0014**
Peak 2 (Rt = 0.66) 14,170 52,001 13,007 17,234
6h 0.167±0.0004 0.1412±0.0021## 0.1466±0.0011##
Peak 3 (Rt = 0.80) 34,001 38,005 15,130 18,081
Values are expressed as mean ± SEM (n = 2 replicates). BCA, Peak 4 (Rt = 0.84) 17,230 39,492 9,317 1,441
butylated caffeic acid; CA, caffeic acid; POV, peroxide value; UVA, Peak 5 (Rt = 0.97) 138,351 183,922 304,604 213,924
ultraviolet A. ** p < 0.01 vs. 3-h control group; ## p < 0.01 vs. 6-h Peak 6 (Rt = 1.20) 67,082 242,324 47,785 71,480
control group.
The blank group was pure squalene, the control group irradiated
squalene. Squalene samples containing CA and BCA (0.02%, w/w)
were irradiated identical to the control group. BCA, butylated
caffeic acid; CA, caffeic acid; P-SQ, peroxidized squalene; UPLC,
Statistical Analysis ultraperformance liquid chromatography; UVA, ultraviolet A.
Results were expressed as mean ± SEM. The data were analyzed
by one-way analysis of variance followed by Dunnett’s post hoc
comparisons to determine the significant difference between treat-
ment groups and the control group. Differences with a p value
<0.05 were considered statistically significant. Figure 3a, 40 μg/mL irradiated squalene was taken as an
example to analyze the effects of UVA duration. Within
4.5 h UVA irradiation, P-SQ products induced a consis-
Results tent increase in ROS level along with irradiation time.
Also, P-SQ from 4 and 4.5 h UVA irradiation showed a
CA and BCA Alleviated UV-Induced Squalene significant effect on ROS induction compared to the
Peroxidation blank group (p < 0.05). After 4.5 h, it was found that fur-
First, POV was measured to show the degree of squa- ther UVA irradiation dramatically decreased ROS gen-
lene peroxidation. As indicated in Table 1, UVA irradia- eration, probably due to the generation of secondary or
tion induced escalating levels of squalene peroxidation even tertiary metabolites. Therefore, when all factors
product with increasing irradiation time towards 6 h. The were considered, 4 h was selected as the preferred irra-
antioxidative potentials of CA and BCA were confirmed diation time. To study the effect of concentrations of
by their inhibitory effects on the peroxidation process of P-SQ, P-SQ with 4 h UVA irradiation was illustrated as
squalene. Next, UPLC-UV/MS was employed to further an example in HaCaT cells and subsequently used in fol-
verify the antioxidant effects of CA and BCA in the UVA lowing experiments. P-SQ from 4 h UVA irradiation was
irradiation period. Chromatography for each sample is added at different concentrations in the range of 4–80 μg/
shown in Figure 2, peak area value in the chromatography mL into the culture medium (Fig. 3b). It was found that
is listed in Table 2. There were significant differences be- ROS level was dose-dependently increased by P-SQ con-
tween the control group (squalene with 6 h UVA irradia- centration. Thus, 40 μg/mL P-SQ with 4 h UVA irradia-
tion) and the blank group (pure squalene). The antioxi- tion was used in follow-up cellular experiments as the
dant capacity of CA and BCA was also confirmed during standardized inducer to assess the moderating effects of
the squalene peroxidation process. CA and BCA effec- actives.
tively reduced squalene peroxidized products (peaks 1, 2,
3, 4, and 6). Effects of CA and BCA on ROS Generation in HaCaT
Cells
Model Establishment of P-SQ-Induced ROS in HaCaT The cytotoxic effects of CA and BCA on HaCaT cells
Keratinocytes were measured first. The cells were exposed to various
To understand P-SQ’s impact on HaCaT cells, various concentrations (1–500 μM) of CA and BCA for 24 h (data
UVA irradiation durations and concentrations of P-SQ not shown). Then, the nontoxic concentrations of CA
were characterized, with the generation of intracellular and BCA were used in HaCaT cells stressed by P-SQ to
ROS as the key endpoint for optimization. As shown in evaluate the cellular antioxidant capacity of actives. As
130.238.7.40 - 10/7/2019 3:30:23 PM

4 Skin Pharmacol Physiol Zhou/Weng

DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
 4.00E+07

3.50E+07
3.00E+07
2.50E+07
2.00E+07

AU
1.50E+07
1.00E+07

5.00E+06
0
1.5 2 2.5 3 3.5 4
Minutes

Blank Control CA BCA

ed
rg
7.00E+06

la
En
1
5 6

600E+06

5.00E+06
2

4.00E+06
3
4
AU

3.00E+06

2.00E+06

1.00E+06

0
0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5
Minutes

Fig. 2. UPLC of squalene samples with or without actives. The blank group was pure squalene, the control group
was irradiated squalene. Squalene samples containing CA and BCA (0.02%, w/w) were irradiated identical to the
control group. AU, arbitrary units; BCA, butylated caffeic acid; CA, caffeic acid; UPLC, ultraperformance liquid
chromatography.

shown in Figure 4, actives in the dosage of 10–200 μM had
no significant effects on the cell viability of P-SQ-chal-
lenged HaCaT cells. Hence, actives within this nontoxic
dose range were applied in subsequent cellular experi-
ments. The antioxidative capacity of these actives against
P-SQ-induced ROS escalation is show in Figure 5. We
observed that P-SQ caused a significant increase in ROS
level in the HaCaT compared with the blank group. BCA
dose-dependently ameliorated ROS level and completely
abrogated ROS at 50 μM, while CA demonstrated no sig-
nificant effect on ROS generation (Fig. 5).
Effects of CA and BCA on IL-1 Family Cytokines in
HaCaT Cells
The proinflammatory effects of P-SQ and the anti-in-
flammatory potentials of the active compounds are shown
in Figure 6. Protein secretion of IL-1α, IL-1β, and IL-1ra
as well as IL-1β mRNA expression were assessed. As
shown in Figure 6, there were significant increases in the
secretion of IL-1α and IL-1ra as well as IL-1β secretion
and IL-1β mRNA expression in HaCaT cells under P-SQ
stimulation. For the anti-inflammatory effects of actives,
only BCA, at 50 and 200 μM, was able to decrease all IL-1
130.238.7.40 - 10/7/2019 3:30:23 PM

CA and BCA Reduce Direct Peroxidation Skin Pharmacol Physiol 5

of Squalene under UVA Irradiation DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
 200 200

* * *** ***
normalized to cell viability

normalized to cell viability

150 150 * **
ROS generation, %

ROS generation, %
100 100

50 50

0 0
Blank Control 3 3.5 4 4.5 5 7 Blank 4 80 4 20 30 40 80
(0 h)
Irradiation time, h Squalene, P-SQ with 4 h UVA irradiation,
µg/mL µg/mL
a 40 µg/mL squalene b

Fig. 3. Effects of UVA irradiation time of P-SQ (a) and concentration of P-SQ (b) on ROS generation in HaCaT
cells. Fluorescence intensities normalized to cell viability indicate ROS generation. Data are all expressed as per-
centage of the blank group. Values are means ± SEM (n = 6 replicates). P-SQ, peroxidized squalene; ROS, reactive
oxygen species; UVA, ultraviolet A. * p < 0.05, ** p < 0.01, *** p < 0.001 versus blank.

150 150

100 100
Cell viability, %

Cell viability, %

50 50

0 0
Blank Control 10 50 100 200 Blank Control 10 50 100 200
CA, µM BCA, µM
a P-SQ b P-SQ

Fig. 4. Effects of different concentrations of CA (a) and BCA (b) on cell viability in HaCaT cells challenged with
P-SQ. Data are all expressed as percentage of the blank group. Values are means ± SEM (n = 6 replicates). BCA,
butylated caffeic acid; CA, caffeic acid; P-SQ, peroxidized squalene.

family secretion and IL-1β mRNA expression induced by
P-SQ treatment in HaCaT cells. Under the same treat-
ment, CA decreased IL-1β secretion dose-dependently,
but failed to show significant inhibitory effects on IL-1β
mRNA expression and secretion of IL-1α and IL-1ra
stimulated by P-SQ treatment.
Analysis of Intracellular Active Concentrations in
HaCaT Cells
Intracellular concentrations of CA and BCA were
quantified in order to test the hypothesis that the solubil-
ity difference between CA and BCA affected their relative
efficacies in the cellular microenvironment. The concen-
trations of intra- and extracellular actives and recovery
rates were detected when 0.4 μmol actives were used in
cells for 8 or 24 h. As shown in Table 3, the results of the
130.238.7.40 - 10/7/2019 3:30:23 PM

6 Skin Pharmacol Physiol Zhou/Weng

DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
 normalized to cell viability 200 200

normalized to cell viability

150 150
ROS generation, %

ROS generation, %
* * *
100 100 **
***
50 50

0 0
Blank Control 10 50 100 200 Blank Control 10 50 100 200
CA, µM BCA, µM
a P-SQ b P-SQ

Fig. 5. Effects of different concentrations of CA (a) and BCA (b) against P-SQ-induced ROS generation in HaCaT
cells. Fluorescence intensities normalized to cell viability indicate ROS generation. Data are all expressed as per-
centage of the blank group. Values are means ± SEM (n = 6 replicates). BCA, butylated caffeic acid; CA, caffeic
acid; P-SQ, peroxidized squalene; ROS, reactive oxygen species. * p < 0.05, ** p < 0.01, *** p < 0.001 versus control
(cells only treated with P-SQ). The difference of ROS level between the control and the blank group was analyzed
using the Student t test.

Table 3. Intracellular content of CA and BCA at incubation times of 8 and 24 h detected by UPLC-MS

Content 8 h treatment 24 h treatment

CA BCA CA BCA

Cells, μmol 1.65±0.028×10–4 7.16±0.081×10–4** 4.30±0.036×10–4 1.30±0.137×10–3##

Supernatant, μmol 0.51±0.039 0.50±0.019 0.50±0.046 0.27±0.009##
Recovery, % 127.13±9.823 124.53±4.758 125.13±11.468 68.02±2.257##

Values are expressed as mean ± SEM (n = 3 replicates). BCA, butylated caffeic acid; CA, caffeic acid; MS, mass
spectrometry; UPLC, ultraperformance liquid chromatography. ** p < 0.01 vs. CA 8-h treatment; ## p < 0.01 vs.
CA 24-h treatment.

samples revealed that the intracellular concentration of
BCA was superior to and nearly triple that of CA for both
timepoints. The recovery rate of the BCA group within
24 h was only 68.02 ± 2.257%, which was significantly
lower than that of the other groups (about 125%).
Discussion
CA, a plant-based phenolic antioxidant, has shown its
chemical and biological activity as a promising antioxi-
dant in many publications [19, 20]. However, CA exhibits
poor lipophilicity due to its high polarity increased by the
bisphenol group [21]. This structure of CA is not condu-
cive to the absorption and metabolism of cells in the or-
ganism. A tert-butyl group was introduced at the 3-posi-
tion in the ortho-aromatic ring, thus BCA was synthe-
sized as a lipophilic CA derivative [14]. Considering the
higher ability to release hydroxyl radicals and higher sta-
bility during the hydrogen transferring process, the tert-
butyl group was added at the 3-position instead of the 2-
or 6-position in the phenyl ring [22]. The modification of
structure has the following beneficial effects: (1) the tert-
butyl group with a large steric hindrance speeds up the
release of ortho-hydrogen; (2) it increases the compound’s
lipophilicity; (3) it enables high yield and convenience for
130.238.7.40 - 10/7/2019 3:30:23 PM

CA and BCA Reduce Direct Peroxidation Skin Pharmacol Physiol 7

of Squalene under UVA Irradiation DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
 500

1,000
400

300

IL-1ra, %
IL-1α, %

*
50 200
*
100 *
***
0 0
Blank Control 50 200 50 200 Blank Control 50 200 50 200
CA, µM BCA, µM CA, µM BCA, µM
a P-SQ b P-SQ

200 2.0

Relative mRNA expression for IL-1β

150 1.5

(fold change)
*
IL-1β, %

100 * 1.0
*
*

50 0.5
***
0 0 ***
Blank Control 50 200 50 200 Blank Control 50 200 50 200
CA, µM BCA, µM CA, µM BCA, µM
c P-SQ d P-SQ

Fig. 6. Effects of actives against P-SQ-induced cytokine secretion of IL-1α (a), IL-1ra (b), IL-1β (c), and IL-1β
mRNA expression (d) in HaCaT cells. Data are all expressed as percentage of the blank group. Values are means
± SEM (n = 3 replicates). BCA, butylated caffeic acid; CA, caffeic acid; IL, interleukin; IL-1ra, interleukin-1 re-
ceptor antagonist; P-SQ, peroxidized squalene. * p < 0.05, *** p < 0.001 versus control (cells only treated with
P-SQ). The difference of IL-1α and IL-1β secretion between each group was analyzed using the Student t test
separately.

synthesis [23]. BCA’s significant antioxidative capacity
has been validated in the lipid oxidation process in a pre-
vious report [14]. The efficacy of BCA against skin surface
oxidative stress was further investigated and compared
with CA in this report.
Squalene, a main component of skin surface lipids, is
subjected to the peroxidation process under chemical and
physical conditions due to abundant double bonds in its
chemical structure [24]. Photooxidation products of
squalene have been proved to cause cell damage in HaCaT
cells [5]. In this study, we investigated the antioxidant ca-
pacity of CA and BCA in squalene peroxidation under
UVA irradiation and the potential roles of CA and BCA
against P-SQ-induced damage and stress in HaCaT cells.
Peroxidation of squalene was confirmed and analyzed
by POV measurement and the UPLC-UV/MS method;
CA and BCA were able to inhibit the peroxidation pro-
cess and the production of P-SQ as expected. CA and
BCA, as the free radical scavenging antioxidants, exhib-
ited their antioxidant capacity by catching free radicals
and delaying propagation of free radicals during the in-
duction period of the free radical chain reaction [9]. These
results are consistent with previous reports which dem-
onstrated the effect of polyphenol on the stability of squa-
lene and the performance of novel polyphenolic antioxi-
dants under storage and frying conditions [25]. Hence,
CA and BCA can both be considered as effective antioxi-
dants against squalene peroxidation.
130.238.7.40 - 10/7/2019 3:30:23 PM

8 Skin Pharmacol Physiol Zhou/Weng

DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
 Squalene peroxidation induced by UVA irradiation
is a dynamic process and can result in different perox-
ides products. ROS is not only associated with oxidative
damage to proteins, lipids, and DNA which leading to
death and aging of cells, but also perturbs the redox bal-
ance in cells. Thus, cellular ROS level is known to play
important roles in certain pathological and physiologi-
cal events. The establishment of an HaCaT model about
P-SQ-induced oxidative stress in the cellular system
was studied with ROS level as the key endpoint. The re-
sults indicated that progressive UV-induced squalene
peroxidation led to increasing stress when cells were
exposed to P-SQ. Interestingly, after 4.5 h, a decrease in
ROS level was observed, which might be due to the pro-
duction of malondialdehyde and other aldehyde prod-
ucts under excessive UVA irradiation [3, 7]. The reduc-
ing nature of malondialdehyde might have interfered
with the inducing effect of primary products of peroxi-
dation of squalene. Dennis and Shibamoto [8] reported
that a maximum yield of malondialdehyde, as a decom-
position product of oxidized lipids, was obtained after
6–8 h irradiation, which was in accordance with the ob-
served decrease in ROS level with P-SQ irradiated for
> 5 h. P-SQ increased ROS level in a dose-dependent
manner in the range of 4–80 μg/mL in HaCaT cells.
Given the solubility of P-SQ in the HBSS culture sys-
tem, we expect that the damage can plateau once reach-
ing the highest concentration, suggesting more physi-
ologically relevant 40 μg/mL dosage for ensuing active
evaluation.
While the two compounds showed similar direct squa-
lene protection effects in the cell-free environment, only
BCA decreased ROS generation dose-dependently when
CA and BCA were applied to P-SQ-stressed HaCaT cells
at nontoxic dosages. Based on chemical nature of the
compounds, we hypothesize that this discrepancy could
be attributed to polarity, which rendered higher lipophil-
ic potential for BCA due to its tert-butyl group [26]. To
validate this hypothesis, intracellular contents of these ac-
tives were studied with UPLC-MS after incubating the
cells with equal amounts of CA and BCA for 8 or 24 h.
The intracellular level of BCA was significantly higher
than that of CA, regardless of incubation time. However,
the recovery of the BCA group for 24 h was significantly
lower than that of the CA group, which is worth discuss-
ing but should not change the conclusion on its better
cellular bioavailability. The partial tert-butyl group in
BCA was easily detached or fragmented from the mole-
cule during the ESI-MS ionization if the compound was
left in cell culture. It was not only in accordance with ESI-
CA and BCA Reduce Direct Peroxidation
of Squalene under UVA Irradiation
MS fragmentation regularity, but in line with the spec-
trum of similar compounds containing a tert-butyl group
in LC-MS [27]. Hence, this could have been the main rea-
son behind the lowered recovery for BCA at 24 h. For
other groups, the volatilization of HBSS during incuba-
tion may be responsible for >100% recovery. The differ-
ence of intracellular levels between CA and BCA is con-
sistent with the feature of the tert-butyl group, and there
is also accordance with the principle that substances with
high fat solubility readily diffuse through cell membranes,
which are phospholipid bilayers [28]. Therefore, we argue
that the lipid solubility of BCA allowed it to enter cells
through the lipid bilayer membrane, making it readier to
access lipid peroxides than CA.
Another common cellular stress response when ex-
posed to P-SQ is inflammatory reaction. Secretion of the
IL-1 family, a highly active and pleiotropic proinflamma-
tory cytokine, is directly involved in the pathogenesis of
acne and plays a central role in mediating the link be-
tween sebum peroxide and inflammation [29, 30]. Here-
in, we demonstrated that P-SQ elevated the secretion of
IL-1α, IL-1β, and IL-1ra, which is consistent with the
publication by Nakagawa et al. [7]. Besides, IL-1β mRNA
expression was also quantified because the resulting IL-
1β protein level was determined by both mRNA tran-
scription, translation, and posttranslational activation by
caspase-1 [31]. For actives, BCA decreased and even in-
hibited IL-1β mRNA expression and IL-1β secretion as
well as the secretion of IL-1α and IL-1ra dose-dependent-
ly. However, CA only showed its dose-dependent inhibi-
tion on IL-1β secretion and had no significant effect on
other cytokine expressions. IL-1α could regulate the in-
tracellular environment because it is localized in the cy-
tosol or cell membrane. In contrast, IL-1β is secreted and
works extracellularly. The commonality of IL-1α and IL-
1β is that they release a variety of cytokine mediators of
inflammation and that they have competitive inhibition
against binding of IL-1ra to IL-1 receptors [32]. Varia-
tions of the IL-1 family were uniform in the experiment,
which is in accordance with their changing tendency in
most treatments in previous reports [33]. The activity of
caspase-1, which controls the activation and secretion of
IL-1β posttranscriptionally, is a likely target for CA given
the different results between protein and mRNA level of
IL-1β [34, 35]. CA depressed IL-1β secretion, but not IL-
1α and IL-1ra secretion, and BCA had a better inhibitory
effect on cytokine expression. Whether it is relevant to the
different localization of IL-1 and intracellular distribu-
tion of CA and BCA influenced by lipophilicity remains
to be further studied.
130.238.7.40 - 10/7/2019 3:30:23 PM
Skin Pharmacol Physiol 9
DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
 In summary, we propose both CA and BCA as good Statement of Ethics
scavenging agents that can reduce direct peroxidation of
The authors have no ethical conflicts to declare.
squalene under UVA irradiation. However, the latter
demonstrated additional superior protective benefit in
HaCaT cells challenged with P-SQ. Such a difference in Disclosure Statement
biological efficacy could be a manifestation of the critical
role of lipophilicity that contributed to superior cellular The authors declare no conflict of interest.
bioavailability. Based on this observation, we conclude
that BCA is an attractive novel active for protection
against P-SQ-induced stress on skin. Future work will be Funding Sources
needed to further evaluate function, safety, and formula-
This research received no external funding.
tion compatibility before BCA can be applied in skin care
products.
Author Contributions

Acknowledgment D. Zhou: formal analysis, investigation, methodology, and

writing of the original draft. X. Weng: conceptualization, data cu-
Many thanks to Gaosheng Shi for synthesis of pure BCA and ration, project administration, supervision, and manuscript re-
execution of the relevant characterization assay. view and editing.

References
1 Masaki H. Role of antioxidants in the skin:
anti-aging effects. J Dermatol Sci. 2010 May;
58(2):85–90.
2 Segal LM, Wilson RA. Reactive oxygen spe-
cies metabolism and plant-fungal interac-
tions. Fungal Genet Biol. 2018 Jan;110:1–9.
3 Ohsawa K, Watanabe T, Matsukawa R, Yo-
shimura Y, Imaeda K. The possible role of
squalene and its peroxide of the sebum in the
occurrence of sunburn and protection from
the damage caused by U.V. irradiation. J Tox-
icol Sci. 1984 May;9(2):151–9.
4 Phull AR, Nasir B, Haq IU, Kim SJ. Oxidative
stress, consequences and ROS mediated cel-
lular signaling in rheumatoid arthritis. Chem
Biol Interact. 2018 Feb;281:121–36.
5 Ottaviani M, Alestas T, Flori E, Mastrofran-
cesco A, Zouboulis CC, Picardo M. Peroxi-
dated squalene induces the production of in-
flammatory mediators in HaCaT keratino-
cytes: a possible role in acne vulgaris. J Invest
Dermatol. 2006 Nov;126(11):2430–7.
6 Nakagawa K, Ibusuki D, Suzuki Y, Yamashita
S, Higuchi O, Oikawa S, et al. Ion-trap tan-
dem mass spectrometric analysis of squalene
monohydroperoxide isomers in sunlight-ex-
posed human skin. J Lipid Res. 2007 Dec;
48(12):2779–87.
7 Nakagawa K, Shibata A, Maruko T, Sookwong
P, Tsuduki T, Kawakami K, et al. γ-Tocotrienol
reduces squalene hydroperoxide-induced in-
flammatory responses in HaCaT keratino-
cytes. Lipids. 2010 Sep;45(9):833–41.
8 Dennis KJ, Shibamoto T. Production of malo-
naldehyde from squalene, a major skin sur-
face lipid, during UV-irradiation. Photochem
Photobiol. 1989 May;49(5):711–6.
9 Soobrattee MA, Neergheen VS, Luximon-
Ramma A, Aruoma OI, Bahorun T. Phenolics
as potential antioxidant therapeutic agents:
mechanism and actions. Mutat Res. 2005
Nov;579(1–2):200–13.
10 Ighodaro OM, Akinloye OA. First line de-
fence antioxidants-superoxide dismutase
(SOD), catalase (CAT) and glutathione per-
oxidase (GPX): their fundamental role in the
entire antioxidant defence grid. Alex J Med.
2018;54:287–93.
11 Prasad NR, Jeyanthimala K, Ramachandran
S. Caffeic acid modulates ultraviolet radia-
tion-B induced oxidative damage in human
blood lymphocytes. J Photochem Photobiol
B. 2009 Jun;95(3):196–203.
12 Agilan B, Rajendra Prasad N, Kanimozhi G,
Karthikeyan R, Ganesan M, Mohana S, et al.
Caffeic Acid Inhibits Chronic UVB-Induced
Cellular Proliferation Through JAK-STAT3
Signaling in Mouse Skin. Photochem Photo-
biol. 2016 May;92(3):467–74.
13 Sahin SC. The potential of Arthrospira pla-
tensis extract as a tyrosinase inhibitor for
pharmaceutical or cosmetic applications. S
Afr J Bot. 2018;119:236–43.
14 Shi G, Liao X, Olajide TM, Liu J, Jiang X,
Weng X. Butylated caffeic acid: an efficient
novel antioxidant. Grasas Aceites. 2017;68(3):
201.
15 Wang H, Wang H, Yang L, Zu YG, Liu F, Liu
TT. Comparative effect of carnosic acid, BHT
and α-tocopherol on the stability of squalene
under heating and UV irradiation. Food Res
Int. 2011;44(9):2730–4.
16 Tominaga H, Ishiyama M, Ohseto F, Sasamo-
to K, Hamamoto T, Suzuki K, et al. A water-
soluble tetrazolium salt useful for colorimet-
ric cell viability assay. Anal Commun. 1999;
36(2):47–50.
17 Takei K, Hashimoto-Hachiya A, Takahara M,
Tsuji G, Nakahara T, Furue M. Cynaropicrin
attenuates UVB-induced oxidative stress via
the AhR-Nrf2-Nqo1 pathway. Toxicol Lett.
2015 Apr;234(2):74–80.
18 Qian BJ, Liu JH, Zhao SJ, Cai JX, Jing P. The
effects of gallic/ferulic/caffeic acids on col-
our intensification and anthocyanin stability.
Food Chem. 2017 Aug;228:526–32.
19 Pluemsamran T, Onkoksoong T, Panich U.
Caffeic acid and ferulic acid inhibit UVA-in-
duced matrix metalloproteinase-1 through
regulation of antioxidant defense system in
keratinocyte HaCaT cells. Photochem Photo-
biol. 2012 Jul–Aug;88(4):961–8.
20 Sato Y, Itagaki S, Kurokawa T, Ogura J, Ko-
bayashi M, Hirano T, et al. In vitro and in vivo
antioxidant properties of chlorogenic acid
and caffeic acid. Int J Pharm. 2011 Jan;403(1–
2):136–8.
21 Roleira FM, Siquet C, Orrù E, Garrido EM,
Garrido J, Milhazes N, et al. Lipophilic phe-
nolic antioxidants: correlation between anti-
oxidant profile, partition coefficients and re-
dox properties. Bioorg Med Chem. 2010 Aug;
18(16):5816–25.
22 Huang Y, Jiang ZW, Liao XY, Hou JP, Weng
XC. Antioxidant activities of two novel syn-
thetic methylbenzenediol derivatives. Czech J
Food Sci. 2014;32(4):348–53.
130.238.7.40 - 10/7/2019 3:30:23 PM

10 Skin Pharmacol Physiol Zhou/Weng

DOI: 10.1159/000501731
Uppsala Universitetsbibl.
Downloaded by:
23 Zhang CX, Wu H, Weng XC. Two novel syn-
thetic antioxidants for deep frying oils. Food
Chem. 2004;84(2):219–22.
24 Reddy LH, Couvreur P. Squalene: A natural
triterpene for use in disease management and
therapy. Adv Drug Deliv Rev. 2009 Dec;
61(15):1412–26.
25 Aladedunye F, Catel Y, Przybylski R. Novel
caffeic acid amide antioxidants: Synthesis,
radical scavenging activity and performance
under storage and frying conditions. Food
Chem. 2012;130(4):945–52.
26 Howie L. Essential Guide to Food Additives.
Int J Dairy Technol. 2006;59(1):49–50.
27 Hao PP, Ni JR, Sun WL, Huang W. Determi-
nation of tertiary butylhydroquinone in edi-
ble vegetable oil by liquid chromatography/
ion trap mass spectrometry. Food Chem.
2007;105(4):1732–7.
CA and BCA Reduce Direct Peroxidation
of Squalene under UVA Irradiation
28 Chillistone S, Hardman JG. Factors affecting
drug absorption and distribution. Anaesth
Intensive Care Med. 2017;18(7):335–9.
29 Qin M, Pirouz A, Kim MH, Krutzik SR, Gar-
bán HJ, Kim J. Propionibacterium acnes In-
duces IL-1β secretion via the NLRP3 inflam-
masome in human monocytes. J Invest Der-
matol. 2014 Feb;134(2):381–8.
30 Tanaka Y, Uchi H, Hashimoto-Hachiya A,
Furue M. Tryptophan Photoproduct FICZ
Upregulates IL1A, IL1B, and IL6 Expression
via Oxidative Stress in Keratinocytes. Oxid
Med Cell Longev. 2018 Nov;2018:9298052.
31 Yoshizumi M, Nakamura T, Kato M, Ishioka
T, Kozawa K, Wakamatsu K, et al. Release of
cytokines/chemokines and cell death in UVB-
irradiated human keratinocytes, HaCaT. Cell
Biol Int. 2008 Nov;32(11):1405–11.
Skin Pharmacol Physiol
DOI: 10.1159/000501731
32 Lewis AM, Varghese S, Xu H, Alexander HR.
Interleukin-1 and cancer progression: the
emerging role of interleukin-1 receptor an-
tagonist as a novel therapeutic agent in cancer
treatment. J Transl Med. 2006 Nov;4(1):48.
33 Lee H, Cheong KA, Kim JY, Kim NH, Noh M,
Lee AY. IL-1 Receptor Antagonist Reduced
Chemical-Induced Keratinocyte Apoptosis
through Antagonism to IL-1α/IL-1β. Biomol
Ther (Seoul). 2018 Jul;26(4):417–23.
34 Choi J, Kim SH, Rah YC, Chae SW, Lee JD,
Md BD, et al. Effects of caffeic acid on cispla-
tin-induced hair cell damage in HEI-OC1 au-
ditory cells. Int J Pediatr Otorhinolaryngol.
2014 Dec;78(12):2198–204.
35 Shen J, Yin Y, Mai J, Xiong X, Pansuria M, Liu
J, et al. Caspase-1 recognizes extended cleav-
age sites in its natural substrates. Atheroscle-
rosis. 2010 Jun;210(2):422–9.
130.238.7.40 - 10/7/2019 3:30:23 PM
11
Uppsala Universitetsbibl.
Downloaded by: