ORIGINAL ARTICLE

Regulation of Skin Collagen Metabolism In Vitro

Using a Pulsed 660 nm LED Light Source: Clinical
Correlation with a Single-Blinded Study
Daniel Barolet1,2, Charles J. Roberge3, Franc¸ois A. Auger3,4, Annie Boucher1 and Lucie Germain3,4

It has been reported that skin aging is associated with a downregulation in collagen synthesis and an elevation
in matrix metalloproteinase (MMP) expression. This study investigated the potential of light-emitting diode
(LED) treatments with a 660 nm sequentially pulsed illumination formula in the photobiomodulation of these
molecules. Histological and biochemical changes were first evaluated in a tissue-engineered Human
Reconstructed Skin (HRS) model after 11 sham or LED light treatments. LED effects were then assessed in
aged/photoaged individuals in a split-face single-blinded study. Results yielded a mean percent difference
between LED-treated and non-LED-treated HRS of 31% in levels of type-1 procollagen and of 18% in MMP-1.
No histological changes were observed. Furthermore, profilometry quantification revealed that more than 90%
of individuals showed a reduction in rhytid depth and surface roughness, and, via a blinded clinical assessment,
that 87% experienced a reduction in the Fitzpatrick wrinkling severity score after 12 LED treatments. No adverse
events or downtime were reported. Our study showed that LED therapy reversed collagen downregulation and
MMP-1 upregulation. This could explain the improvements in skin appearance observed in LED-treated
individuals. These findings suggest that LED at 660 nm is a safe and effective collagen-enhancement strategy.
Journal of Investigative Dermatology (2009) 129, 2751–2759; doi:10.1038/jid.2009.186; published online 9 July 2009

INTRODUCTION intracellular photobiochemical reactions (Karu and Kolyakov,

Skin aging, intrinsic and extrinsic, is associated with morpho- 2005; Karu et al., 2005a, b; Hamblin and Demidova, 2006;
logical changes, including rhytids, furrows, and telangiecta- Barolet, 2008). A number of clinical studies provide evidence
sia (Kang et al., 2001; Fisher et al., 2002). It has been reported of the effectiveness of LED therapy in photorejuvenation
that collagen synthesis is reduced and interstitial matrix using a variety of LED light sources (Weiss et al., 2004; Bhat
metalloproteinases (MMP-1), the collagenase involved in et al., 2005; Russell et al., 2005; Weiss et al., 2005; Goldberg
normal turnover of skin collagen, are upregulated in aged et al., 2006; Baez and Reilly, 2007; Lee et al., 2007). An
skin (Fligiel et al., 2003; Fisher et al., 2008; Varani et al., improvement in skin appearance in aged/photoaged indivi-
2004). Hence, a possible strategy for treating and preventing duals has been documented after full-face or split-face serial
clinical manifestations of skin aging is the restoration of treatments with yellow (590 nm), red (630, 633 nm), or red in
collagen deficiency by the induction of new collagen combination with infrared (830 nm) light based on profilo-
synthesis and reduction of MMP-1. metry quantification, clinical assessment of digital photo-
It has been shown that light-emitting diode (LED) therapy, graphs, and patient reported outcomes. A correlation of
a nonthermal noninvasive treatment, can trigger natural clinical effects with further analysis for basic mechanisms
was examined in two of these studies. In Weiss et al. (2005),
1
RoseLab Skin Optics Laboratory, Montreal, Quebec, Canada; 2Department after a regimen of eight treatments delivered over 4 weeks
of Medicine, Dermatology Division, McGill University, Montreal, Quebec, with LED 590 nm, staining with anti-collagen I antibodies
Canada; 3Laboratoire d’Organogénèse Expérimentale (LOEX), Centre de
showed a 28% (range of 10–70%) average increase in density,
recherche (FRSQ) du CHA de Québec, Centre de recherche du CHA de
Québec, Hôpital du Saint-Sacrement du CHA (FRSQ), Quebec, Canada and whereas staining with anti-MMP-1 showed an average
4
Department of Surgery, Laval University, Quebec, Canada reduction of 4% (range of 2 to 40%). In the Lee et al.
Correspondence: Dr Daniel Barolet, RoseLab Skin Optics Laboratory, 3333 study, participants were randomly divided into four groups
Graham Blvd., Suite 206, Montreal, Quebec, CANADA H3R 3L5. treated with either 830 nm alone, 633 nm alone, a combina-
E-mail: [Link]@[Link]
tion of 830 and 633 nm, or a sham treatment light, two times
Abbreviations: FCS, Fitzpatrick Classification System; HRS, human
reconstructed skin; LED, light-emitting diode; MMP-1, matrix
a week for 4 weeks. In the treatment groups, a significant
metalloproteinase-1 increase in the amount of collagen was observed by
Received 7 January 2009; revised 15 April 2009; accepted 30 April 2009; histological evaluation with standard preparation with
published online 9 July 2009 hematoxylin and eosin stain. This finding was confirmed by

& 2009 The Society for Investigative Dermatology [Link] 2751

 D Barolet et al.
LED Regulation of Human Dermal Collagen

the Masson–trichrome stain for collagen. In the immunohisto- 100%

Mean percent difference between

LED-treated and Untreated HRS
chemical staining results, no significant changes in MMP-1 or
80%
MMP-2 were observed between the baseline and 2 weeks
60%
posttreatment specimens in all groups. Other models have also
been used to investigate LED-based effects. Collagen increase 40%
after LED treatment has been documented in fibroblast cultures 20%
(McDaniel et al., 2002; Huang et al., 2007), in third-degree
0%
burn wound-healing models (Meireles et al., 2008a, b), and in
–20%
human blister fluids (Barolet et al., 2005). However, con-
current MMP determinations after LED treatment were not –40%
assessed in these experiments. T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11
In this study, we opted for an innovative approach Type-1 procollagen MMP-1
involving a stable in vitro model replicating a real-life LED
Cohen’s d type-1
clinical application, and then sought to explore clinical procollagen
0.64 1.20 0.28 0.59 0.30 0.50 0.56 0.66 0.22 0.09 0.68

correlates of these effects in humans. The 660 nm LED light Cohen’s d MMP-1 –0.12 –0.42 –0.75 –1.01 –0.68 –0.45 –0.46 –0.59 –0.08 –0.56 –0.65
source delivered in a sequential pulsing mode was used as it
is a deep penetrating well-absorbed wavelength that falls into Figure 1. Increases in type I procollagen and concurrent reduction in
the absorption peaks of cytochrome c oxidase, the chromo- MMP-1 levels in HRS after LED treatment. A cyclic pattern of alternating
highs and lows was observed in response to the 11 consecutive treatments
phore thought to be responsible for LED effects (Karu and
(T1–T11) for type 1 procollagen and MMP levels. Values are percent
Kolyakov, 2005; Karu et al., 2005a, b). difference±SEM (n ¼ 9) between treated and untreated control HRS samples
A 3-D model of tissue-engineered Human Reconstructed in mean levels of type I procollagen and MMP-1 assessed in the supernatants
Skin (HRS) (Michel et al., 1999) was used to initially after each treatment. Table shows Cohen’s d for type 1 procollagen and
investigate the potential of 660 nm LED in modulating collagen MMP-1 for each time point.
and MMP-1. The HRS model offers a variety of advantages
over other in vitro and preclinical models. HRS samples are
produced exclusively from human fibroblasts and keratino- in response to each of the 11 LED treatments (T1–T11)
cytes and do not contain any synthetic material. This model conducted over a 1-month period. An increase in type I
emulates skin as a complex tissue composed of two different procollagen production with a concomitant decrease in
compartments, the continuously renewing epidermis made up MMP-1 levels was observed in the LED-treated samples
mostly of keratinocytes and the underlying dermal matrix with when compared with that in untreated samples. The latter
fibroblasts as its major cellular components. These compart- effects were found to be cyclic, with alternating high and low
ments are tightly interconnected allowing for paracrine mutual levels observed in response to consecutive LED treatments
interactions, which are essential for epidermal growth, The mean percent difference between LED-treated and
differentiation, and tissue homeostasis. This HRS model untreated HRS specimens across repeated treatments was
represents a unique opportunity to test biological changes 31% (range of 5–81%) for type I procollagen levels and 18%
after LED treatment over an extended time frame on age- and (range of 3 to 27%) for MMP-1 levels (Figure 1). Cohen’s d
gender-defined skin, much like a clinical setting. Eleven effect size was in the medium-to-large range at the various
consecutive LED treatments carried out over 4 weeks were time points for both type-1 procollagen and MMP-1
conducted on HRS and compared with the untreated control measurements, with the exception of T9 (dp0.2) (see Table
specimens for type I procollagen and MMP-1 production. The below Figure 1).
clinical study was then conducted to corroborate the in vitro
experiment. A group of chronologically aged/photoaged Histological assessments on HRS
individuals received 12 LED treatments or a sham light control Morphological changes after the last LED treatment were
over a 1-month period (split-face study). Clinical improve- assessed using Masson’s trichrome and immunofluorescent
ments were assessed with an in vivo 3-D microtopography staining. Masson staining of reconstructed skin showed no
quantitative measurement (PRIMOS readings) and a qualitative histological difference in the extracellular matrix (same
clinical assessment of digital photographs by blinded medical dermal thickness) in LED-treated compared with untreated
observers. Our results supported the hypothesis that LED control samples (data not shown). Typical results from
treatments could upregulate collagen and downregulate an LED-treated HRS are depicted in Figure 2a (HRS38).
MMP-1 in vitro, and thus lead to improvements in skin Immunofluorescent staining of HRS yielded no difference in
appearance observed in the LED-treated side in humans. the amounts of type I procollagen (same band thickness)
between the LED-treated and untreated specimens (data
not shown). A representative fluorescent immunostaining
RESULTS pattern after LED treatment for type I procollagen is displayed
LED-induced modulations in type I procollagen and MMP-1 in Figure 2b. Overall, after LED treatments, the dermal layer
production of human reconstructed skin exhibited a dense well-organized collagenous connective
Type I procollagen and MMP-1 production was assessed in the tissue material, overlaid with a well-stratified epidermis,
supernatants of HRS from individuals with aged/photoaged skin presenting an intact stratum basale.

2752 Journal of Investigative Dermatology (2009), Volume 129

 D Barolet et al.
LED Regulation of Human Dermal Collagen

Table 1. Microtopographic profilometry analysis

Percent improvement Percentage of subjects
post-treatment with improvement
LED- LED-treated Untreated
treated Untreated P-value (%) (%)

Ra 18.57±2.41 3.73±1.17 o0.0001 97 46

Rz 20.83±2.26 7.43±1.64 o0.0001 94 51
Values are mean±SEM.

–15 –15
–10
–10
–5
–5
0
0
5
5
10
10
15
15
–20 –15 –20
–15 –5 –10
0 –5 –10 10 5 0
10 5 20 15
20 15

–15
–15
–10
–10
–5
Figure 2. Histological assessments of HRS. (a) Masson stain after LED –5
0
treatment on reconstructed skin (HRS38), underlining the presence of –0
5
dermal collagen (colored in blue) in the dermal portion. Keratinocytes found 5
10
in the epidermis and dermal fibroblasts are colored in red. (b) Fluorescent 10
15
15
immunostaining of HRS42 type I collagen located in the dermal portion –20 –15 –20
–5 –10
–15 0 –5 –10
(red band pattern) after LED treatment. Nuclei are colored in blue for both 10 5 0
20 15
10 5
20 15
keratinocytes (round nuclei) and dermal fibroblasts (elongated nuclei).
No morphological changes were observed on HRS after LED treatments. Figure 3. Reduction in rhytids and skin roughness with LED therapy.
Scale bar indicates 50 mm. Photograph depicts color-coded topography images before treatment (a) and
after LED treatment (c). Each color determines a specific depth with
darker areas indicating a deeper wrinkle surface. Black and white
photographs show skin texture and pore size for the pre-LED (b) and
post-LED treatment areas (d). All photographs are for S2 aged 38 years.
Profilometry quantification of improvements in skin appearance
To assess the potential of 660 nm LED light in improving the
appearance of aged/photoaged skin, participants were treated
weekly three times for four consecutive weeks (12 treatments) treatment on Rz or Ra values. Figure 3 shows PRIMOS phase
with an LED and sham light contralaterally in a split-face shift black and white and color-coded microtopography
design. Results from a microtopographic profilometry analy- photographs before and after LED therapy of the periorbital
sis (PRIMOS reading) of rhytid depth and severity (Rz) and area for participant S2.
reduction in skin roughness (Ra) are presented in Table 1.
Analysis of the results showed that the percent improvement Skin textural enhancement in aged/photoaged participants
after treatment in Rz values was statistically different between Clinical assessment of digital photographs taken before
the LED-treated and control non-treated sides (Po0.0001). treatment and after the last treatment were analyzed by three
LED treatment produced an Rz reduction in 94% of blinded medical observers. According to the observers, an
participants, with the highest reduction in rhytids of 51% improvement in at least one subtype of the Fitzpatrick
noted on a female participant aged 46 years (S33). For Classification System (FCS), used to evaluate the degree of
the non-treated side, Rz reductions were observed in 51% wrinkling (rhytids), was obtained in 85–90% (mean of 87%)
of participants. The results from the LED-treated side of participants after LED treatments. For the untreated side,
also revealed a statistically significant reduction in surface an improvement was observed in 25–63% of participants
roughness in comparison with the untreated control side (mean of 45%). Observer assessments revealed that the
(Po0.0001), as measured by Ra. Overall, an improvement in degree of improvements in wrinkling after treatment was mild
skin appearance after LED treatment was observed in 97% to moderate. The analysis conducted on severity scores
of participants, with the highest benefit of 56% observed showed that there was a significant difference between the
on a woman aged 57 years (S29). For the untreated side, LED-treated and untreated sides for Observers 1 (Po0.0001)
improvements were observed in 46% of participants. The and 3 (P ¼ 0.001), but not according to Observer 2 (Table 2).
participant’s age was not found to influence the effect of LED Results were not found to be influenced by the participant’s

[Link] 2753
 D Barolet et al.
LED Regulation of Human Dermal Collagen

Table 2. Observers’ assessment of improvement in the severity of wrinkles

Degree of improvement Percentage of subjects X1 point
LED-treated Untreated P-value LED-treated (%) Untreated (%)

Observer 1 1.35±0.11 0.25±0.07 o0.0001 90 25

Observer 2 0.90±0.07 0.70±0.10 NS 85 63
Observer 3 1.03±0.09 0.50±0.09 0.001 85 48
Values are mean±SEM; NS: non-statistically significant.

Pre Post synthesis and MMP-1 modulation in humans. LED-induced

biochemical and histological changes in a 3-D HRS in vitro
model were first examined. A clinical study was then carried
LED out to assess in vivo the clinical correlates of this light
treatment on skin texture and appearance of individuals
with aged/photoaged skin, in whom collagen synthesis and
MMP-1 are known to be downregulated and upregulated,
respectively (Varani et al., 2000). Results from this study
provide evidence for the photobiomodulatory effects of
Control 660 nm sequentially pulsed LED treatment.
We have previously developed a tissue-engineering
approach for the production of a biological HRS from
cultured human cells (Laplante et al., 2001). In this study,
Figure 4. Improvements in skin appearance in aged/photoaged participants we took advantage of this model for our in vitro study. The
with LED treatment. Digital photograph illustrates periorbital areas at
use of this tissue-engineering method allowed the creation
pre-treatment (a) and posttreatment (c) for the LED-treated side, and at
pre-treatment (b) and posttreatment (d) for the non-treated control area.
of three differently aged HRS, which offered the unique
All photographs are for S2 aged 38 years. opportunity to test biological changes, such as type-1
collagen and MMP-1 production, after LED treatment on
age-defined skin. The stability of these constructs in culture
over time also permitted numerous consecutive LED treat-
age. The internal consistency between the observers’ findings ments to be applied. Results obtained from the three HRS
was deemed to be acceptable (Cronbach’s a ¼ 0.615). constructs tested demonstrated that 660 nm LED exposures
Clinical results for participant S2 are displayed in Figure 4. did not trigger histological changes; however, LED treatment
significantly amplified collagen production with a conco-
Skin temperature steadiness during LED treatment mitant decrease in collagenase (MMP-1) production. The
It has been reported that a mild increase in temperature (43 vs noticeable differences between our results and those of other
371C) can induce a decrease in type I procollagen secretion studies were the depth and range in collagen production
(Halper et al., 2005). Papillary dermis temperature was thus increase and MMP-1 reduction in an in vitro milieu capable
monitored in our clinical study throughout LED exposure to of dermal–epidermal communication comparable with that
ensure that skin temperature was kept normal in order not to in in vivo skin (Weiss et al., 2005; Lee et al., 2007). Interes-
hinder photobiochemical reactions associated with collagen tingly, LED treatments can induce similar changes in collagen
metabolism. Monitoring attested that physiological tempera- synthesis and MMP activity as those reported with retinoic
ture was maintained at the treatment area during the entire acid (Fisher and Voorhees 1998; Fisher et al., 1998), although
procedure. Temperature variations registered by thermo- the mechanisms at play may be different.
couple probes during LED treatment were stable or never The pattern of procollagen and MMP-1 production was
more than 0.51C (dermo-epidermal junction temperature of found to be cyclic. The observed cyclical pattern could result
33±0.51C). from the capacity of HRS to express specific endogenous cell
signaling pathways, periodically turning the factory ‘‘on’’ and
LED therapy safety profile in humans ‘‘off’’. This was seen in both treated and untreated specimens,
Adverse reactions were monitored throughout the study. LED possibly reflecting the natural variations in procollagen and
therapy was found to be well tolerated by all patients, with no MMP regulation. Hence, LED therapy did not seem to affect
adverse effects or downtime reported during or after LED the cycle per se but seemed to modulate the HRS intrinsic
treatment. ability to do so.
The HRS model used in this study emulates skin as a
DISCUSSION complex tissue composed of two different compartments: the
This study investigated the potential of 660 nm sequentially epidermis made up mostly of keratinocytes and the under-
pulsed LED treatments in the photoinduction of collagen lying dermal matrix with fibroblasts as its major cellular

2754 Journal of Investigative Dermatology (2009), Volume 129

 D Barolet et al.
LED Regulation of Human Dermal Collagen

components. The limitations of this model reside in the fact the papillary and upper reticular dermis, forming the so-
that it does not contain melanocytes, Merkel cells and called Grenz zone (Hardaway et al., 2002; Nelson et al.,
Langerhans cells of the epidermis, skin appendages (hair and 2002). Our results provide support for previous research
glands), vascularization and innervation of the dermis, as showing that light in the red spectrum, such as 660 nm,
well as its hypodermis counterpart. More complex models allows for light penetration and absorption in the dermis and
are currently under development. For example, a trilayered through the entire papillary layer, enabling the stimulation
skin substitute consisting of an epidermis, dermis, and an of collagen production (Simpson et al., 1998). The cascade
adipocyte-containing hypodermis has recently been pro- of events leading to collagen production is thought to
duced (Trottier et al., 2008). Future experiments with more be initiated by the antenna molecule mitochondrial cyto-
complex HRS models should allow extended investigations chrome c oxidase (Karu and Kolyakov, 2005; Hamblin and
into LED-related mechanisms. Demidova, 2006; Karu et al., 2005a, b). Absorbed light
Profilometry quantification confirmed a clinical improve- converted to chemical kinetic energy would cause changes in
ment pertaining to skin surface characteristics after 1 month membrane permeability, improve signaling between mito-
of LED treatment. Over 90% of participants showed a chondria, nucleus and cytosol, lead to nitric oxide formation,
reduction in rhytid depth and surface roughness measured and increase oxidative metabolism to produce more ATP
quantitatively by Rz and Ra PRIMOS profilometry after 12 (Karu, 1989; Morimoto et al., 1994; Yu et al., 1997; Zhang
LED treatments. Clinical improvement in wrinkles was et al., 2003), ultimately leading to the normalization of
observed in 87% of participants, although one observer cell activity, including increased collagen production and
reported no significant differentiation in wrinkling scores MMP regulation.
between the treated and untreated sides. Some degree of To our knowledge, these are previously unreported data
improvement was indeed noted on the untreated side in on the regulation of skin collagen metabolism in vitro with
the qualitative assessments. Although photographs were clinical correlates using a sequentially pulsed 660 nm LED
randomly presented and the observers were blinded to the source. The HRS used in this study proved to be a good
details of the experiment, the fact that the LED-treated side preclinical model to test the effects of such light treatments on
of the face was the right side in all participants may have human skin over a month, and permitted to investigate the
unblinded the study and biased observers. Given that an mechanism(s) responsible for the positive changes observed
improvement on the untreated side was also observed in the in the skin of LED-treated humans. The characteristics of the
quantitative assessments, however, suggests that the improve- participants included in this trial, the study setting, the
ment was genuine. The observed improvement on the treatment regimens tested, and the outcomes assessed are
untreated side could be due, at least in part, to the fact that similar to those encountered in day-to-day dermatology
participants were instructed on posttreatment skin care and practice. The results from this study thus support 660 nm
closely monitored by the clinical team, which might have LED as a collagen-enhancement strategy that can be used
motivated them to take better care of their skin during the trial safely in a clinical setting. Yet, additional studies are needed
period. There is also the possibility that, as the untreated side to evaluate whether 660 nm LED effects are maintained over
was not covered during treatment on the controlateral side, it time. Further studies are also warranted to ascertain the
benefited from a spill over- or a systemic effect of treatment. cellular processes involved.
Future studies with split-face designs should randomize
treatment allocation with the untreated side covered. IN VITRO STUDY MATERIALS AND METHODS
LED therapy seemed to be well tolerated, with no adverse Cell culture media
events or downtime reported, likely because of the absence Keratinocytes were grown in a complete DME-HAM medium: a
of thermal injury to the skin during treatment. From a clinical combination of Dulbecco–Vogt modification of Eagle’s medium
perspective, the nonthermal characteristics of LED treatment (DME) with Ham’s F12 in a 3:1 proportion (Invitrogen, Burlington,
may yield a significant advantage over other treatment Canada), supplemented with 5% Fetal Clone II serum (FCSII) (HyClone,
methods, given that effective improvements in the appear- Logan, UT), 10 ng ml epidermal growth factor (Austral Biologicals, San
ance of aged skin can be achieved without thermal damage Ramon, CA), 24.3 mg ml adenin (Sigma-Aldrich, Oakville, Canada),
induced to skin with associated adverse effects. Additional 5 mg ml insulin (Sigma-Aldrich), 5 mg ml transferrin (Roche Diagnostics,
studies are, however, needed to evaluate if the clinical Laval, Canada), 2  109 M 3,30 50 triiodo-L-thyronin (Sigma-Aldrich),
improvements observed in this study are maintained over 0.4 mg ml hydrocortisone (Calbiochem, La Jolla, CA), and antibiotics
time, as follow-up assessments were carried out over a 100 IU ml penicillin G (Sigma-Aldrich) and 25 mg ml gentamicin
relatively short period of follow-up time (4 weeks). Our (Schering, Pointe-Claire, Canada). Fibroblasts were cultured in DME
results are in line with previous accounts of the effectiveness containing 10% fetal calf serum (HyClone) and antibiotics.
of LED therapy in photorejuvenation with other light sources
in which effects were observed for up to 6 months follow-up Cell isolation and culture
time points (Weiss et al., 2004, 2005; Bhat et al., 2005; Skin specimens were collected from healthy women aged 38
Russell et al., 2005; Goldberg et al., 2006; Baez and Reilly, (HRS38), 42 (HRS42), and 64 (HRS64) years, during reductive breast
2007; Lee et al., 2007). surgery (HRS38 and HRS42) or face-lift (HRS64). Procedures for cell
The induction of collagen synthesis by nonablative isolation were initiated within 3 hours after surgery according to a
rejuvenation procedures has been shown to occur largely in previously published method (Germain et al., 2001; Auger et al.,

[Link] 2755
 D Barolet et al.
LED Regulation of Human Dermal Collagen

2002). Skin specimens were washed five times with a phosphate- therapy, as power density/light intensity—a key variable for optimal
buffered saline supplemented with antibiotics. Specimens were then photoinduction—is greatly influenced by the distance between the
cut into 3 mm wide strips and kept overnight at 4 1C in Hepes buffer light source and the surface of the skin (Hart and Cameron, 2005).
containing 500 mg ml thermolysin. The epidermis was mechanically Measurements of the intensity (mW) of the LED light source emitting
separated from the dermis with forceps; keratinocytes were at 660 nm (110 mA CW) were taken with a ph100-si (Gentec-eo,
dissociated from the epidermis through incubation of the epidermal Quebec, Canada) at every centimeter, up to a 10-cm distance away
fragments under agitation at 371C for 30 minutes, with 0.05% from the light source. This was important to ensure that the proper
trypsin-0.1% EDTA in PBS. After trypsin inactivation (addition of amount of photons was delivered to reach the cellular activation
culture medium containing 10% serum and centrifugation), keratino- (induction) threshold in the skin.
cytes were expanded in the presence of irradiated 3T3 fibroblasts
in T75 flasks (BD Biosciences, Mississauga, Canada) and subse- Type I collagen and MMP-1 determination
quently frozen until further use. Fibroblasts were dissociated from Human type I collagen was measured in cell culture supernatants of
the remaining dermis fragments after incubation in a colla- the selected time period with the Protype I collagen C-peptide
genase H solution at 371C under agitation. After centrifugation, the enzyme immunoassay kit purchased from Takara Mirus Bio
fibroblasts were also plated in T75 flasks for expansion and (Madison, WI), according to the manufacturer’s instructions. MMP-1
subsequently frozen until further use. levels were also measured using MMP-1 biotrak activity assay systems
according to the manufacturer’s instructions (Amersham Biosciences,
Production of tissue-engineered HRS Baie D’Urfe, Canada).
The cells isolated from the healthy women aged 38, 42, and 64 years
were used to reconstruct HRS38, HRS42, and HRS64, respectively. Data analysis
HRS were produced as previously described (Michel et al. 1999). As an indication of effect size to measure the direction and
Briefly, fibroblasts (F38, F42, or F64) were cultivated in a fibroblast magnitude of the treatment effect for each time point, Cohen’s d
culture medium containing 50 mg ml sodium ascorbate (Sigma- was used. Cohen’s d was computed as follows: d is the difference in
Aldrich) for 4 weeks until cell sheets were formed. After peeling group means divided by the pooled s.d. A standardized effect size
the fibroblast sheets from the bottom of the dishes, two sheets were of 0.2 is considered small, 0.5 is considered medium, and 0.8 is
superimposed, for a total of two sheets per HRS, and cultured for 1 considered large (Cohen, 1988).
week. After dermal equivalent maturation, keratinocytes (250,000
cells per cm2, between cellular passages 2 and 3) were seeded on Histological analysis
reconstructed stroma and cultured for 7 days in a keratinocyte After the last treatment, biopsies of untreated and LED-treated
complete medium (containing 50 mg ml sodium ascorbate) under reconstructed skin were fixed for at least 24 hours in a Bouin solution
submerged conditions. The HRS were then brought to the air–liquid (ACP, St Leonard, Canada) and embedded in paraffin.
interface and cultivated in complete DME-HAM with 5% serum and Five-mm-thick cross-sections were stained with Masson’s trichrome.
50 mg ml sodium ascorbate, without EGF, for an additional 4 weeks. Photographs were taken at the  40 objective with a digital
The culture medium was changed three times per week. Each camera (CoolSnap RS Photometrics, Roper Scientific, Munich,
experimental condition was tested in triplicate on the three HRS Germany).
tested (HRS38, HRS42, and HRS64). All plastic ware for tissue
culture was procured from BD Biosciences. Indirect immunofluorescence microscopy
At the end of the treatment series, samples of the untreated and
LED and sham light treatments LED-treated reconstructed skin were embedded and frozen in
Samples of the HRS replicates were exposed to the LED or sham light OCT compound (Somagen, Edmonton, Canada). Four-mm-thick
source under a laminar flow hood. HRS were treated 11 times (T1 to cross-sections were fixed in cold acetone and further incubated with
T11) over a 1-month period, treatments were performed three times mouse monoclonal anti-human type I collagen antibody (Chemicon,
a week for 4 consecutive weeks, with only two treatments performed Temecula, CA). The secondary antibodies (Chemicon) used were
in the fourth week. Cultures were then incubated at 371C (8% CO2). rhodamine-conjugated goat anti-mouse IgG-IgM and goat anti-rabbit
Supernatants were collected before each treatment and after the last IgG. Nuclei were stained blue with Hoechst 33258 (Sigma-Aldrich).
treatment, and stored at 201C until assessed for type I procollagen
and MMP-1. IN VIVO STUDY MATERIALS AND METHODS
The LED technique involved the application of a 660-nm Participant selection
wavelength delivered in a sequential pulsing mode at a power Forty healthy patients with aged/photoaged skin (37 women and 3
density of 50 mW cm for a total fluence of 4 J cm, for a duration of men) with a mean age of 44.9 (33–62) years were recruited from the
160 seconds (2m40 s) (LumiPhase-R, OPUSMED Inc. Montreal, Dr. Daniel Barolet Clinic in Montreal, Canada from September 2002
Canada). The pulsing patterns and time on and time off sequences to December 2002 and tested between January 03 and April 03.
were as follows: Pulse width (time on) 500 msec, pulse interval (time Inclusion criteria included patients with skin type I to III according to
off) 150 msec, four pulses per pulse train, and a pulse train interval of the FCS (Fitzpatrick, 1988). At study entry, 15% of patients were skin
1,550 msec. The non-treated HRS samples were exposed to a sham type I, 45% were skin type II, and 40% were skin type III. A total of
light for 160 seconds with a total fluence of 0 J cm. 48% were deemed to present with mild, 38% with moderate, and
During this study, a lot of care was taken to maintain a working 10% with a severe degree of wrinkles, according to the FSC for
distance of 2.5 cm (±1 mm) away from the target surface during LED degree of wrinkling (rhytids).

2756 Journal of Investigative Dermatology (2009), Volume 129

 D Barolet et al.
LED Regulation of Human Dermal Collagen

Exclusion criteria comprised patients taking cortisone (predni- experimental periorbital area. LED treatments were administered
sone), anticoagulant therapy, or any drug known to increase using the same parameters as those for the in vitro study (see In vitro
photosensitivity. In addition, during the 12 months preceding the Study Materials and Methods). To maximize LED photoinduction, a
study, patients were required not to have used isotretinoin topical moisturizer without active ingredients was applied daily, as
(Accutane) or applied topical steroids to the site to be treated. dry skin is known to enhance skin surface reflectivity (Friedman
Moreover, a previous laser or topical medication at the to-be-treated et al., 2002). To support photobiochemical reactions in the triggering
site was not permitted. Patients gave written, informed consent of gene-expression-enhanced collagen metabolism during LED
to participate in this trial in compliance with the US Code of exposure, skin temperature should be kept normal, that is,
Federal Regulations dealing with the conduct of clinical studies physiological (Halper et al., 2005). Therefore, monitoring papillary
(21CFR including parts 50 and 56 with regard to informed consent dermis temperature with a needle probe (type-T thermocouple from
and IRB regulations). The study was conducted according to Omega, Montreal, Canada) during LED treatment was carried out.
Good Clinical Practice Guidelines and the principles of the The control side was exposed to a sham light for 160 seconds with a
Declaration of Helsinki, and was approved by the Institutional total fluence of 0 J cm. No cooling method was used after exposure.
Review Board Services (Div. 1373737 Toronto, Canada). The To avoid hiding periorbital skin, a well-circumscribed external
trial was registered with [Link] ([Link] eyelid protector (Oculoplastik, Montreal, Canada) was worn to
Identifier: NCT00818246). The study flow chart is presented in protect the retina from direct illumination. Before returning home,
Figure 5. participants were instructed on posttreatment skin care, which
included applying a plain moisturizer on both sides of the face, sun
Study design avoidance, and the use of a sunscreen (SPF 30).
This was a split-face single-blinded study to assess the efficacy of
LED treatment on overall skin appearance (rhytids depth and Primary outcome measures
texture—surface roughness) of the aged/photoaged periorbital area. A topographical quantitative assessment was carried out for each
This study design allowed for within-participant assessments of study participant using PRIMOS 3D surface topography (GFM,
clinical effects. Assignment: The right periorbital area was desig- Teltow, Germany) profilometry performing phase shift rapid
nated to be the experimental side, and the left periorbital was used as in vivo measurements of skin from the experimental and control
control (sham light). Blinding: The observers who performed the periorbital areas before treatment and 4 weeks after treatment. A
clinical qualitative assessments on the basis of digital photographs PRIMOS readings (PRIMOS 4 Software, GFM, Teltow, Germany)
were blinded to the treatment regimen (LED-treated or untreated/ star analysis was carried out, from a phase shift photograph, a
control side) and to the timing of the photographs (pre- or 1.5 cm-diameter circular portion of the skin on the lateral cantus
posttreatment). area was cut into 12 equal segments, allowing depth calculation
of fine surface lines (peak and valley analysis) in that area and
Study procedure quantification of skin surface roughness. The Ra value was
Participants were treated three times weekly for 4 consecutive weeks calculated from general surface roughness characteristics, pore
(12 treatments) with the 660 nm-pulsed LED device on the size and skin texture. The Rz value obtained for each participant was

Enrollment Meeting criteria: n=40

Allocated to intervention: n=40

Spilt-face study
Allocation
Right side: LED treatment
Left side: Sham light treatment

Lost to Follow-up: n=0

Follow-up
Discontinued: n=0

Analyzed
Quantitative assessment: n=35
Analyzed
Excluded from analysis: n=5
Analysis Qualitative assessment: n=40
Given reason: Impossibility to
Excluded from analysis: n=0
match before and after
PRIMOS pictures

Figure 5. Clinical study flow chart.

[Link] 2757
 D Barolet et al.
LED Regulation of Human Dermal Collagen

estimated from average peak and valley measurement data to REFERENCES

quantify rhytid depth and severity. Pre- and posttreatment surface Auger FA, Rémy-Zolghadri M, Grenier G, Germain L (2002) A truly new
topography measurements were taken to precisely quantify clinical approach for tissue engineering: The LOEX self-assembly technique. In:
improvements. Stem cell transplantation and tissue engineering, (Haverich A, Graf H,
eds). Springer-Verlag: Berlin, 73–88
Baez F, Reilly LR (2007) The use of light-emitting diode therapy in the
Secondary outcome measures
treatment of photoaged skin. J Cosmet Dermatol 6:189–94
A clinical qualitative assessment was carried out by three blinded
Barolet D, Boucher A, Bjerring P (2005) In vivo human dermal collagen
medical observers through the evaluation of digital photographs production following LED-based therapy: the importance of treatment
(Canon Dual Flash E2, Mississauga, Canada) of the experimental and parameters. Lasers Surg Med 17:76. (abstract)
control periorbital areas taken pre-treatment and 4 weeks posttreat- Barolet D (2008) Light-emitting diodes (LEDs) in dermatology. Semin Cutan
ment for each participant. Each photograph was taken maintaining Med Surg 27:227–38
identical ambient lighting, pose, and camera angles. To avoid bias in Bhat J, Birch J, Whitehurst C, Lanigan SW (2005) A single-blinded randomised
clinical appraisal, sets of photographs for the treated and untreated controlled study to determine the efficacy of Omnilux Revive facial
treatment in skin rejuvenation. Lasers Med Sci 20:6–10
sides of the face were randomly presented to the observers who were
Cohen J (1988) Statistical power analysis for the behavioral sciences. 2nd ed,
blinded to the details of the experiment. The photographs were
Lawrence Earlbaum Associates: Hillsdale, NJ
analyzed for clinical improvement using the FCS subtype scale for
Fisher GJ, Kang S, Varani J, Bata-Csorgo Z, Wan Y, Datta S et al. (2002)
degree of wrinkling (rhytids). Their assessment was rated on a five- Mechanisms of photoaging and chronological skin aging. Arch Dermatol
point scale and scored as follows: 0 ¼ none, 1 ¼ mild, 2 ¼ moderate, 138:1462–70
3 ¼ good, 4 ¼ excellent. Adverse effects were monitored over the Fisher GJ, Talwar HS, Lin J, Lin P, McPhillips F, Wang Z et al. (1998) Retinoic
entire course of the study. acid inhibits induction of c-Jun protein by ultraviolet radiation that
occurs subsequent to activation of mitogen-activated protein kinase
pathways in human skin in vivo. J Clin Invest 101:1432–40
Data analysis Fisher GJ, Voorhees JJ (1998) Molecular mechanisms of photoaging and its
Sample sizes and power calculations were generated according to prevention by retinoic acid: ultraviolet irradiation induces MAP kinase
the primary outcome measures of the study. To have a 98% chance signal transduction cascades that induce Ap-1-regulated matrix metallo-
of detecting as significant (at the two-sided 5% level) a 10% proteinases that degrade human skin in vivo. J Investig Dermatol Symp
Proc 3:61–8
difference between the treated and untreated/control sides in the Ra
Fisher GJ, Varani J, Voorhees JJ (2008) Looking older: fibroblast collapse and
and Rz posttreatment improvement, with an assumed s.d. of 10, 33
therapeutic implications. Arch Dermatol 144:666–72
participants were required. To account for an 80% per protocol
Fitzpatrick TB (1988) The validity and practicality of sun-reactive skin types I
completion rate, the planned number of patients to be enrolled was 40. through VI. Arch Dermatol 124:869–71
The primary outcome variables were the percent improvement Fligiel SEG, Varani J, Datta SH, Kang S, Fisher GJ, Voorhees JJ (2003) Collagen
posttreatment for the Ra and Rz values measured 4 weeks degradation in aged/photoaged skin in vivo and after exposure to MMP-1
posttreatment. The primary outcome analysis was conducted on in vitro. J Invest Dermatol 120:842–8
results from 35 participants. The data from five participants were Friedman PM, Skover GR, Payonk G, Kauvar AN, Geronemus RG
removed from the quantitative assessment analysis because of the (2002) 3D in-vivo optical skin imaging for topographical quanti-
tative assessment of non-ablative laser technology. Dermatol Surg 28:
fact that it was not possible to match the before and after PRIMOS 199–204
data digital color-coded photographs. The secondary clinical Germain L, Moulin V, Berthod F, Lopez C.A, Goulet F, Auger FA (2001)
outcome variable was the medical observers’ combined gain in Multiple applications of tissue-engineered human skin. In: Cultured
the FCS score for degree of wrinkling 4 weeks after treatment. human keratinocytes and tissue engineered skin substitutes. (Horch RE,
Secondary outcome analysis was conducted on the results Munster AM, Achauer BM eds). Georg Thieme Verlag: Stuttgart,
Germany, 91–8
of the total sample of participants (n ¼ 40). For the primary and
Goldberg DJ, Amin S, Russell BA, Phelps R, Kellett N, Reilly LA (2006)
secondary outcome variables, an analysis of covariance was used to Combined 633-nm and 830-nm led treatment of photoaging skin.
assess statistical differences between the LED-treated and untreated J Drugs Dermatol 5:748–53
sides, taking into account age as the covariate. Main effects and two- Halper J, Griffin A, Hu W, Jung C, Zhang J, Pan H et al. (2005) In vitro culture
way interactions were incorporated into the statistical model. decreases the expression of TGF(beta), Hsp47 and type I procollagen and
Tukey’s HSD test was used to assess pairwise comparisons. increases the expression of CTGF in avian tendon explants. J Musculo-
skelet Neuronal Interact 5:53–63
Cronbach’s a was calculated to assess internal consistency between
Hamblin MR, Demidova TN (2006) Mechanisms of low level light therapy.
the observers’ assessments. P-values were considered significant
In: Mechanisms for Low-Light Therapy (Michael R. Hamblin, MR,
at Pp0.05. Waynant RW, Anders J eds). Proc of SPIE V 6140 1–12
Hardaway CA, Ross EV, Barnette DJ, Paithankar DY (2002) Nonablative
CONFLICT OF INTEREST cutaneous remodeling with a 1.45 microm mid-infrared diode laser:
Intellectual Property disclosure related to the LED pulsing code by the first phase I. J Cosmet Laser Ther 4:3–8
author. Hart G, Cameron R (2005) The importance of irradiance and area in neonatal
phototherapy. Arch Dis Child Fetal Neonatal Ed 90:F437–40
ACKNOWLEDGMENTS Huang PJ, Huang YC, Su MF, Yang TY, Huang JR, Jiang CP (2007) In vitro
We thank the following people for their indispensable contribution to this observations on the influence of copper peptide aids for the LED
study: Isabelle Lussier Ph.D. from MedStrategis for her assistance (article & photoirradiation of fibroblast collagen synthesis. Photomed Laser Surg
biostatistical analysis), Geneviève Chaput, Mathieu Auclair, Diane Lamalice, 25:183–90
Franc¸ois Barolet, Lyne Jodoin, Chémir Mamode, Jean Dubé, Dominique Kang S, Fisher GJ, Voorhees JJ (2001) Photoaging: pathogenesis, prevention,
Gignac, and Mariane Moisan. and treatment. Clin Geriatr Med 17:643–59

2758 Journal of Investigative Dermatology (2009), Volume 129

 D Barolet et al.
LED Regulation of Human Dermal Collagen

Karu T (1989) Photobiology of low-power laser effects. Health Phys Morimoto Y, Arai T, Kikuchi M, Nakajima S, Nakamura H (1994) Effect of low
56:691–704 intensity Argon laser irradiation on mitochondria respiration. Lasers Surg
Karu TI, Kolyakov SF (2005) Exact action spectra for cellular responses Med 15:191–9
relevant to phototherapy. Photomed Laser Surg 23:355–61 Nelson JS, Majaron B, Kelly KM (2002) What is nonablative photorejuvena-
tion of human skin? Semin Cutan Med Surg 21:238–50
Karu TI, Pyatibrat LV, Afanasyeva NI (2005a) Cellular effects of low power
laser therapy can be mediated by nitric oxide. Lasers Surg Med 36: Russell BA, Kellett N, Reilly LR (2005) A study to determine the efficacy of
307–14 combination LED light therapy (633 and 830 nm) in facial skin rejuve-
nation. J Cosmet Laser Ther 7:196–200
Karu TI, Pyatibrat LV, Kolyakov SF, Afanasyeva NI (2005b) Absorption
measurements of a cell monolayer relevant to phototherapy: reduction of Simpson CR, Kohl M, Essenpreis M, Cope M (1998) Near-infrared optical
cytochrome c oxidase under near IR radiation. J Photochem Photobiol B properties of ex vivo human skin and subcutaneous tissues measured
81:98–106 using the Monte Carlo inversion technique. Phys Med Biol 43:2465–78

Laplante AF, Germain L, Auger FA, Moulin V (2001) Mechanisms of wound Trottier V, Marceau-Fortier G, Germain L, Vincent C, Fradette J (2008) IFATS
collection: using human adipose-derived stem/stromal cells for the
reepithelialization: hints from a tissue-engineered reconstructed skin to
production of new skin substitutes. Stem Cells 26:2713–23
long-standing questions. FASEB J 15:2377–89
Varani J, Schuger L, Dame MK, Leonard C, Fligiel SE, Kang S et al. (2004)
Lee SY, Park KH, Choi JW, Kwon JK, Lee DR, Shin MS et al. (2007) A
Reduced fibroblast interaction with intact collagen as a mechanism for
prospective, randomized, placebo-controlled, double-blinded, and
depressed collagen synthesis in photodamaged skin. J Invest Dermatol
split-face clinical study on LED phototherapy for skin rejuvenation: 122:1471–9
clinical, profilometric, histologic, ultrastructural, and biochemical
evaluations and comparison of three different treatment settings. Varani J, Warner RL, Gharaee-Kermani M, Phan SH, Kang S, Chung JH et al.
(2000) Vitamin A antagonizes decreased cell growth and elevated
J Photochem Photobiol B 27:51–67
collagen-degrading matrix metalloproteinases and stimulates collagen
McDaniel DH, Weiss RA, Geronemus R, Ginn L, Newman J (2002) accumulation in naturally aged human skin. J Invest Dermatol 114:480–6
Light–tissue interactions I: Photothermolysis versus photomodulation
Weiss RA, McDaniel DH, Geronemus R, Weiss MA (2005) Clinical trial of
laboratory findings. Lasers Surg Med 14:25 (abstract)
a novel non-thermal LED array for reversal of photoaging: clinical,
Meireles GC, Santos JN, Chagas PO, Moura AP, Pinheiro AL (2008a) A histologic, and surface profilometric results. Lasers Surg Med 36:85–91
comparative study of the effects of laser photobiomodulation on the Weiss RA, Weiss MA, Geronemus RG, McDaniel DH (2004) A novel non-
healing of third-degree burns: a histological study in rats. Photomed thermal non-ablative full panel LED photomodulation device for reversal
Laser Surg 26:159–66 of photoaging: digital microscopic and clinical results in various skin
Meireles GC, Santos JN, Chagas PO, Moura AP, Pinheiro AL (2008b) types. J Drugs Dermatol 3:605–10
Effectiveness of laser photobiomodulation at 660 or 780 nanometers on Yu W, Naim JO, McGowan M, Ippolito K, Lanzafame RJ (1997) Photo-
the repair of third-degree burns in diabetic rats. Photomed Laser Surg modulation of oxidative metabolism and electron chain enzymes in rat
26:47–54 liver mitochondria. Photochem Photobiol 66:866–71
Michel M, L’Heureux N, Pouliot R, Xu W, Auger FA, Germain L (1999) Zhang Y, Song S, Fong CC, Tsang CH, Yang Z, Yang M (2003) cDNA
Characterization of a new tissue-engineered human skin equivalent with microarray analysis of gene expression profiles in human fibroblast cells
hair. In Vitro Cell Dev Biol Anim 35:318–26 irradiated with red light. J Invest Dermatol 120:849–57

[Link] 2759