A topical antioxidant solution containing

vitamins C and E stabilized by ferulic acid

provides protection for human skin against
damage caused by ultraviolet irradiation
John C. Murray, MD, James A. Burch, Robert D. Streilein, Mary Ann Iannacchione,
Russell P. Hall, MD, and Sheldon R. Pinnell, MD
Durham, North Carolina

Background: Skin cancer and photoaging changes result from ultraviolet (UV)-induced oxidative stress.
Topical antioxidants may protect skin from these effects.

Objective: We sought to determine whether a stable topical formulation of 15% L-ascorbic acid, 1% alpha-
tocopherol, and 0.5% ferulic acid (CEFer) could protect human skin in vivo from substantial amounts of
solar-simulated UV radiation.

Methods: CEFer and its vehicle were applied to separate patches of normal-appearing human skin for 4
days. Each patch was irradiated with solar-simulated UV, 2 to 10 minimal erythema doses, at 2-minimal
erythema dose intervals. One day later, skin was evaluated for erythema and sunburn cells, and
immunohistochemically for thymine dimers and p53. UV-induced cytokine formation, including interleukin
(IL)-1a, IL-6, IL-8, and IL-10, and tumor necrosis factor-a, were evaluated by real-time polymerase chain
reaction.

Results: CEFer provided significant and meaningful photoprotection for skin by all methods of evaluation.

Limitations: The number of patients evaluated was relatively small.

Conclusion: CEFer provided substantial UV photoprotection for skin. It is particularly effective for
reducing thymine dimer mutations known to be associated with skin cancer. Its mechanism of action is
different from sunscreens and would be expected to supplement the sun protection provided by
sunscreens. ( J Am Acad Dermatol 2008;59:418-25.)

U ltraviolet (UV) radiation in sunlight gener-

ates oxidative stress in skin that can result in
skin cancer and photoaging changes. The
skin protects itself using low moleculareweight
Abbreviations used:
cDNA:
CEFer:
complementary DNA
15% L-ascorbic acid, 1% alpha-tocopherol,
and 0.5% ferulic acid
antioxidants that neutralize the oxidative stress IL: interleukin
MED: minimal erythema dose
mRNA: messenger RNA
From the Division of Dermatology, Duke University Medical Center. PBS: phosphate-buffered saline
Supported by a grant from SkinCeuticals-L’Oreal, New York, NY.
PCR: polymerase chain reaction
RT: real time
Disclosure: Dr Pinnell is a consultant for SkinCeuticals-L’Oreal. Dr TNF: tumor necrosis factor
Murray, Mr Burch, Mr Streilein, Ms Iannacchione, and Dr Hall UV: ultraviolet
have no conflicts of interest to declare. C E Ferulic is a patented
formulation, US Patent No. 7,179,841 issued to SkinCeuticals.
Accepted for publication May 7, 2008.
Reprint requests: Sheldon R. Pinnell, MD, M3135 Duke University before it can cause damage. Most antioxidant pro-
Medical Center, Durham, NC 27710. E-mail: pinne002@ tection depends on dietary intake and subsequent
mc.duke.edu. delivery to skin. Because the antioxidant is destroyed
Published online July 7, 2008.
0190-9622/$34.00
or altered by oxidation during neutralization, pro-
ª 2008 by the American Academy of Dermatology, Inc. tection is often limited by the concentration of
doi:10.1016/j.jaad.2008.05.004 antioxidants remaining in skin. In an attempt to

418
 Murray et al 419

augment this protective strategy, our laboratory has METHODS

been interested in developing the topical use of Experimental protocol
antioxidants for photoprotection. To accomplish this Studies were conducted at with the approval of
goal, it has been necessary to develop formulations our institutional review board. Nine adults with
that allow for stability of these inherently unstable Fitzpatrick skin type II or III (white skin with the
substances. Of even more importance, it has been ability to tan slightly to moderately in response to UV
necessary to find formulations that allow these anti- irradiation) were entered into the protocol. Two
oxidants to enter into skin, so that they are in the right solutions were applied (2 mg/cm2) daily for 4 days to
place to provide protection. separate patches of back skin. Subjects were in-
Using porcine skin as a model for human skin, we structed not to wash the area for at least 2 hours. One
reported that 15% L-ascorbic acid, protonated to solution was an aqueous solution containing 15% L-
remove charge, was capable of entering skin and ascorbic acid, 1% dl-a tocopherol, and 0.5% trans
protecting it from UV irradiation.1 After maximizing ferulic acid (CEFer). The other solution was a vehicle
the concentration of the solution for effectiveness, only control. Solutions (C E Ferulic and its vehicle)
we demonstrated that skin concentration, once sat- were supplied by SkinCeuticals (Garland, Tex). On
urated, was somehow stabilized with a half-life day 3, each subject received solar-simulated irradia-
disappearance of about 4 days.2 This first-generation tion to an untreated patch of skin, to determine the
topical product provided protection against UVB and minimal erythema dose (MED). Irradiances were
UVA. Because the solution had no absorption against from 20 to 60 mJ/cm2 at 10-mJ/cm2 intervals of
these wavelengths, it was concluded that the pro- UVB as determined by a radiometer (IL1700,
tection was not caused by a sunscreen effect.1 In International Light, Newburyport, Miss). MED was
addition, the solution protected UV-induced immu- determined 24 hours later as the lowest dose result-
nosuppression.3 The topical L-ascorbic acid added to ing in perceptible borders of erythema (about 40
the skin’s own concentration and most probably mJ/cm2). Details of irradiation have been published.4
augmented its antioxidant effect. A 1000-W xenon solar simulator (Lightning Cure 200,
We next formulated 15% L-ascorbic acid together Hamamatsu, Hamamatsu City, Japan) fitted with a
with 1% a-tocopherol.4 These antioxidants form an dichroic mirror to filter infrared and visible light and
interactive pair in tissues. When a-tocopherol neu- a Schott 295 bandpass filter (Mainz, Germany) to
tralizes oxidative stress in lipids, its oxidation pro- remove wavelengths less than 295 nm fitted with a
duct can be regenerated by L-ascorbic acid.5-7 This liquid light guide to create a 1-cm spot of light on
interaction helps to renew antioxidant protection in skin was used at an output of about 5 mW/cm2 of
the tissues. After maximizing the solution for con- UVB and about 40 mW/cm2 of UVA. On day 4, the
centration and stability, this second-generation pro- vehicle-treated patch received 2 to 6 MED and the
duct provided doubled UV photoprotection for skin, CEFer-treated patch received 2 to 10 MED, each at
in comparison with L-ascorbic acid alone. 2 -MED intervals. One day later, skin was evaluated
More recently, we have undertaken a search for by colorimetry for erythema, and 4-mm punch
antioxidants that could be shown to increase the biopsy specimens of skin receiving 6 MED of
stability of the solution of L-ascorbic acid and a- irradiation were evaluated for sunburn cells.
tocopherol. We have demonstrated that the addition
of 0.5% ferulic acid, a common plant antioxidant, Measurement of erythema and sunburn cells
both improved stability and again doubled UV Erythema was measured by computerized color-
photoprotection for skin.8 Ferulic acid, a hydrox- imetry in the ‘‘a’’ mode of digital color skin photo-
ycinnamic acid, probably protects L-ascorbic acid graphs.9 Each spot and unirradiated adjacent skin
and a-tocopherol in solution by serving as a sacri- was measured in triplicate. The color difference
ficial substrate. We demonstrated that the antioxi- between irradiated and unirradiated skin determined
dant combination provided superior protection the erythema. Sunburn cells were determined in
when compared with the individual ingredients. formalin-fixed skin sections stained with hematoxy-
This effect is apparently unrelated to a sunscreen lin and eosin. Results are expressed as mean SD.
effect. In this article we describe the extension of The P values were determined by two-tailed Student
these observations to human skin. We demonstrate t test using unequal variance.
that this third-generation product provides effective
UV photoprotection, inhibits thymine dimer forma- Immunohistochemistry for thymine dimers
tion, and inhibits p53 activation. In addition UV and p53
activation of interleukin (IL)-1a, IL-6, IL-8, IL-10, and Thymine dimer. Formalin-fixed, paraffin-em-
tumor necrosis factor (TNF)-a is blunted. bedded tissue sections were deparaffinized with
420 Murray et al

xylene and rehydrated through a graded alcohol- Qiagen) and then stored in elution buffer at e82 C
water series. Sections underwent heat-induced epi- until quantitative RT PCR was performed.10
tope retrieval by incubating them in a solution of
citrate buffer (pH 6.0) (Zymed, San Francisco, Calif) RT quantitative PCR
for 20 minutes at 95 C, followed by equilibration in Measurement of gene expression was performed
phosphate-buffered saline (PBS). Nonspecific per- using RT PCR system (iCycler, BioRad) as previously
oxidase was blocked by 3% hydrogen peroxide, and described.11 TaqMan probes were labeled at the 5’-
nonspecific binding was blocked with 1.5% horse end with the reporter dye molecule FAM (6-carbox-
serum in PBS. Mouse monoclonal antithymine dimer yfluorescein; emission lmax = 518 nm) and at the
antibody clone KTM-53 (Kamiya Biomedical, Seattle, 3’-end with the quencher dye molecule 3’ Black Hole
Wash) was diluted in 1.5% horse serum-PBS to Quencher (Integrated DNA Technologies, Coralville,
1/1000 and incubated on sections for 1 hour at Iowa) (absorbance lmax = 534 nm). RT PCR reac-
37 C. Detection was accomplished with a horse tions of cDNA specimens and cDNA standards were
antimouse secondary antibody using an avidin-bio- conducted in a total volume of 25 L with 2 iQ
tin link (Vector Elite ABC peroxidase kit, Burlingame, SuperMix (BioRad) and primers and probes at opti-
Calif) as described by the supplier. Slides were mized concentrations. Parameters for the thermal
counterstained with hematoxylin. cycler were 3 minutes at 95 C, and 40 cycles involv-
p53. The procedure was identical to that described ing denaturation at 95 C for 30 seconds, annealing/
above except that antigen retrieval was with heat- extension at 59 C for 30 seconds. RT monitoring of
induced epitope retrieval in pH8 Tris-EDTA buffer fluorescent emission from cleavage of sequence-
(Zymed). Detection was achieved with mouse mon- specific probes by the nuclease activity of taq poly-
oclonal antihuman p53 clone DO-7 (Dako, Glostrup, merase allowed definition of the threshold cycle
Denmark) diluted in 1.5% horse serum-PBS to 1/250. during the exponential phase of amplification.11,12
For analysis of control 18S RNA, IL-1a, IL-6, IL-8,
IL-10, and TNF-a messenger RNA (mRNA) expres-
Real-time polymerase chain reaction studies
sion, a standard curve was constructed with serial
An identical study was conducted in an additional
dilutions of cDNA obtained from a single sample of
10 subjects with the exception that subjects received
human peripheral blood mononuclear cells that had
irradiation at 2 MED. The 6-mm biopsy specimens
been stimulated for 4 hours by phytohemagglutinin
of CEFer and vehicle-treated skin were bisected, half
(5 g/mL).10,13 Cells were removed, mRNA isolated,
for real-time (RT) polymerase chain reaction (PCR)
and cDNA produced as described above. Relative
studies and half for immunohistochemistry.
quantitation of cytokine mRNA expression was
determined using the relative standard curve method
Messenger RNA isolation from skin biopsy (User Bulletin No. 2, Applied Biosystems, Foster City,
specimens Calif). Individual results were normalized using
Biopsy specimens of untreated, UV vehicle, and 18S ribosomal RNA as an internal control with 18S-
UV CEFer skin of each subject were immediately specific primers and probe. All samples, analyzed by
placed in RNALater (Qiagen, Valencia, Calif) and RT PCR, were performed in triplicate. Cytokine
stored at e80 C until processing could occur. Total mRNA expression was expressed as arbitrary units.
RNA was extracted using fibrous tissue mini kits Ratios of UV-CEFeretreated skin mRNA expression
(RNeasy, Qiagen). Tissue was homogenized in RLT to UV-vehicleetreated skin 100 was determined to
buffer using a tissue homogenizer (Omni International, evaluate the percent suppression of cytokine mRNA
Bethesda, Md) and RNA isolated according to kit expression for each cytokine tested.
protocol (Qiagen). To avoid contaminating genomic
DNA, Rnase-Free Dnase (Qiagen) was used. The Design of cytokine primers and probes
concentration of RNA was measured by a spectro- Oligonucleotide primers and TaqMan probes for
photometer (ND1000, Nanodrop Inc, Wilmington, 18S, IL-1a, IL-8, and IL-10 were designed using
Del) and stored at e82 C in water. For complementary PrimerQuest (Integrated DNA Technologies). We
DNA (cDNA) synthesis, approximately 2 g of total conducted BLASTN searches to confirm the total
RNA was transcribed with cDNA transcription reagents gene specificity of the nucleotide sequences chosen
(iScript, BioRad, Hercules, Calif) using random hexa- for primers and probes and the absence of DNA
mers. Thermal cycler parameters included 5 minutes at polymorphism. To avoid amplification of contami-
25 C for incubation, 30 minutes at 42 C for RT reaction, nating genomic DNA, one of the two primers for
and 5 minutes at 85 C for enzyme inactivation. cDNA cDNA was placed either at the junction between two
was purified with PCR purification kits (QiaQuick, exons or within different exons. Cytokine primers
 Murray et al 421

Fig 1. CEFer inhibits ultraviolet (UV)-induced erythema. CEFer and vehicle were applied to
back skin (2 mg/cm2) daily for 4 days. Skin was irradiated with solar-simulated UV radiation, 2
to 10 minimal erythema doses (MED) at 2 -MED intervals to CE ferulic-treated skin and
2 - to 6 -MED intervals to vehicle-treated skin. Erythema was determined 1 day later.

Fig 2. CEFer inhibits ultraviolet (UV)-induced erythema.

CEFer and vehicle were applied to back skin (2 mg/cm2) Fig 3. CEFer inhibits ultraviolet (UV)-induced sunburn
daily for 4 days. Skin was irradiated with solar-simulated cell (SBC ) formation. CEFer and vehicle were applied to
UV radiation, 2 to 10 minimal erythema doses (MED) at back skin (2 mg/cm2) daily for 4 days. Skin was irradiated
2 -MED intervals to CE ferulic-treated skin and 2 - to 6 - with solar-simulated UV radiation, 2 to 10 minimal
MED intervals to vehicle-treated skin. Erythema was mea- erythema doses (MED) at 2 -MED intervals to CE
sured 1 day later by colorimetry of digital photographs. ferulicetreated skin and 2 - to 6 -MED intervals to
Mean SD (n = 9). *P .02 versus vehicle. **P .01 vehicle-treated skin. Skin biopsy specimens of 6 MED-
versus vehicle at 6 MED. treated skin were taken 1 day later. SBCs were counted
and are expressed as cells/mm of epidermis. Mean SD
were designed to result in amplicons less than 150 (n = 9). *P .01 vs vehicle.
base pair to enhance efficiency of PCR amplification.
Gene expression sets for IL-6 and TNF-a were treated skin (Figs 1 and 2). At 2 to 6 MED irradi-
purchased from Applied Biosystems. ance, colorimeter readings for vehicle versus CEFer
revealed significant (P .01) protection by CEFer at
Statistical analysis all irradiance levels (Fig 2). Moreover, CEFer pro-
Results were analyzed using Mann-Whitney test vided significant protection (P .01) at 8 MED and
and the Wilcoxon signed rank test using software 10 MED irradiances when compared with 6
(Analyse-it, Analyse-It Software Ltd, Leeds, United MED-irradiated vehicle-treated skin.
Kingdom). Statistical significance was assumed for a
P value less than .05. Sunburn cells
Sunburn cell enumeration (Fig 3) of 6 MED-
RESULTS irradiated skin when compared with similarly irra-
Erythema diated vehicle-treated skin revealed significant
CEFer provided substantial protection against UV- protection (P .01) by CEFer, (vehicle 31.5
induced erythema when compared with vehicle- 14.3 vs CEFer 8.4 7).