Secretion of basic fibroblast
growth factor (bFGF) by normal dermal fibroblasts 24 hours after application
of tretinoin (1 × 10 −5 mol/L). In each cell line,
treated cells secreted more bFGF than did controls (mean comparison, P<.05).
Comparison of normal and keloid-producing
dermal fibroblast mean secretion of transforming growth factor-β1 (TGF-β1)
120 hours after application of tretinoin (1 × 10 −5
mol/L). Keloid fibroblasts show sensitivity to tretinoin treatment (P<.05), whereas normal dermal fibroblasts do not.
Secretion of transforming growth
factor-β1 (TGF-β1) by keloid-producing fibroblasts 120 hours after
application of tretinoin (1 × 10 −5 mol/L). In each
cell line, treated cells secreted more TGF-β1 than did controls (mean
Comparison of normal and keloid-producing
dermal fibroblast mean secretion of transforming growth factor-β1 (TGF-β1)
24 hours after application of glycyl-L-histidyl-L-lysine-Cu2+ (GHK-Cu)
(1 × 10 −9 mol/L). Although both cell types show sensitivity
to GHK-Cu treatment, only keloid fibroblasts demonstrated statistically significant
Secretion of transforming growth
factor-β1 (TGF-β1) by keloid-producing fibroblasts 24 hours after
application of glycyl-L-histidyl-L-lysine-Cu2+ (GHK-Cu) (1 ×
10 −9 mol/L). In each cell line, treated cells secreted less
TGF-β1 than did controls (mean comparison, P<.05).
McCormack MC, Nowak KC, Koch RJ. The Effect of Copper Tripeptide and Tretinoin on Growth Factor Production in a Serum-Free Fibroblast Model. Arch Facial Plast Surg. 2001;3(1):28-32. doi:
From the Wound Healing and Tissue Engineering Laboratory, Division
of Otolaryngology–Head and Neck Surgery, Stanford University Medical
Center, Stanford, Calif.
Copyright 2001 American Medical Association. All Rights Reserved.
Applicable FARS/DFARS Restrictions Apply to Government Use.2001
Objective To evaluate the effect of copper tripeptide and tretinoin on normal
and keloid-producing dermal fibroblasts in a serum-free in vitro model. The
cellular response was described in terms of viability and secretion of basic
fibroblast growth factor (bFGF) and transforming growth factor-β1 (TGF-β1).
Methods Primary cell lines were established from patient facial skin samples
obtained during surgery and plated in serum-free media. At 0 hour, copper
tripeptide (1 × 10 −9 mol/L), tretinoin (1 ×
10 −5 mol/L), or appropriate control vehicle was added. Cell
counts and viability were established at 24, 72, and 120 hours. Supernatants
were collected at the same intervals and were assessed for bFGF and TGF-β1
concentrations using the enzyme-linked immunosorbent assay technique.
Results Cell lines showed viability between 86% and 96% (mean, 92%) throughout
the experiment. Tretinoin-treated normal fibroblasts secreted more bFGF than
did controls at 24 hours (P<.05). Tretinoin-treated
keloid-producing fibroblasts secreted more TGF-β1 than did controls at
120 hours (P<.05). Keloid-producing fibroblasts
treated with copper tripeptide secreted less TGF-β1 than did controls
at 24 hours (P<.05); a similar trend was observed
in normal fibroblasts.
Conclusions Normal fibroblasts treated with tretinoin produced more bFGF than did
controls, and this might partially explain the clinically observed tightening
effects of tretinoin. Normal and keloid-producing dermal fibroblasts treated
with copper tripeptide secreted less TGF-β1 than did controls, suggesting
a possible clinical use for decreasing excessive scar formation.
ABERRANT WOUND healing is a significant problem for many surgical patients.
Inadequate healing is often due to an underlying medical condition such as
diabetes, previous radiation therapy, poor nutritional status, or malignancy.
Other patients, at the opposite end of the spectrum, may form hypertrophic
scars or keloid tissue. Keloids do not represent a more severe form of hypertrophic
scars. There is no simple continuum from normal skin scar to hypertrophic
scar to keloid tissue. Collagen bundles in hypertrophic scars remain parallel
in orientation (as in normal skin), whereas keloids have randomly organized
sheets of collagen. Also, whereas contractile myofibroblasts are common in
hypertrophic scars, they are relatively absent in keloids.
Multiple treatment modalities have attempted to reduce such excess scarring,
yet none have established long-standing results.1
Keloid-producing dermal fibroblasts (KFs) have been shown2-3
to produce substantially more procollagen and fibronectin than do normal dermal
fibroblasts (NFs) in culture. The current research indicates that at the core
of keloid etiology are wound-healing cytokines—growth factors that regulate
production of extracellular matrix (ECM) components.
An imbalance in the levels of various cytokines generated in the wound-healing
process may lead to keloid formation. In addition, it may be the proper balance
of cytokines in the wound environment that allows for normal wound healing.
Such growth factors, which promote cell growth, division, and migration in
wounded tissue, are secreted by dermal fibroblasts. Two of these key growth
factors were considered in the present study.
Transforming growth factor-β1 (TGF-β1) is a key cytokine in
the initiation and termination of tissue repair. In relation to other known
tissue repair cytokines, it strongly stimulates synthesis of the major ECM
proteins, namely, collagen, proteoglycan, and fibronectin.4-5
Younai et al6 investigated the in vitro effects
of TGF-β1 on fibroblasts and found that KFs produce many times more collagen
than do NFs when stimulated by TGF-β1. Cultured KFs also secrete more
TGF-β1 than do NFs—further evidence that excessive amounts of this
cytokine in the wound environment may be central to keloid formation and growth.7
The second growth factor assayed in our study, basic fibroblast growth
factor (bFGF), has been shown to inhibit hydroxyproline biosynthesis, an index
of collagen production, in cultured KFs.8 Methods
that increase or stabilize bFGF secretion might therefore decrease aberrant
scar or keloid formation by reducing the amount of collagen deposited during
The modulators selected for evaluation were the copper tripeptide complex
glycyl-L-histidyl-L-lysine-Cu2+ (GHK-Cu) and all-trans retinoic acid (tretinoin). A naturally occurring tripeptide,
GHK-Cu has been shown to have significant clinical application in the field
of wound healing and tissue repair. It has been demonstrated that GHK-Cu stimulates
cultured NFs to synthesize collagen and induces a dose-dependent increase
in the synthesis of glycosaminoglycans.9-11
It also has been shown that ECM accumulation increases in the rat wound model
as a result of GHK-Cu application.12 Clinical
studies13-14 have demonstrated
the efficacy of copper tripeptide preparations in facilitating an increased
rate of wound healing in diabetic ulcers and in patients who have undergone
Mohs surgery. Additional cosmetic uses of copper tripeptide complex are currently
Tretinoin is frequently prescribed for its collagen-tightening effects
and is useful in the treatment of a variety of skin conditions, including
wrinkling, acne vulgaris, photoaging, early stretch marks, and hyperkeratosis.15-16 Several studies17-19
have also demonstrated the possible antitumor effects of tretinoin in the
treatment of basal cell carcinoma, dysplastic nevi, and cutaneous malignant
melanoma. The dermal tightening effect of tretinoin in the treatment of photodamaged
skin has been well studied and its histologic effects described.20
Central to the clinical effects of reduced wrinkles and skin roughness is
the partial restoration of the facility of normal skin cells to produce collagen.
Retinoids, however, also have been shown to decrease the amount of collagen
produced in KF cultures, and results of clinical trials21-22
indicate that topical treatment with retinoic acid (0.05%) reduces keloid
size in most cases.
Cell culture–based research is an effective means of studying
wound healing at the cellular level because it offers a controlled environment.
The presence of serum in the culture medium has long been a necessary component
of this model because it allows for sustained cell growth. The presence of
serum components, however, hinders any experiment seeking to accurately measure
growth factor production by the cells themselves. Use of a serum-free cell
culture model addresses this shortcoming because it allows for a controlled
environment in which the growth factor–secreting effects of potential
wound-healing modulators can be evaluated without confounding effects from
serum. The Wound Healing and Tissue Engineering Laboratory of Stanford University
Medical Center, Stanford, Calif, has developed a fibroblast in vitro model
that uses serum-free growth medium.23
The purpose of this study was to grow NFs and KFs in a serum-free model,
treat them with GHK-Cu or tretinoin, and assess the cellular response in terms
of cell viability and autocrine growth factor production. An attempt is made
to characterize the growth factor profiles of KFs and NFs based on these 2
Primary cultures of dermal fibroblasts were established from excisional
biopsies of 3 different keloid and 3 different normal facial skin specimens
using a standard explant technique. All cell lines were directly established
from operative specimens. Keloid specimens were from the lobule and normal
skin samples were from the preauricular and mental region from 5 different
patients. The described method was approved by the institutional review board
at Stanford University Medical Center.
The dermis was isolated from the specimens and minced. Antimicrobial
treatment consisted of washing the specimens in Dulbecco phosphate-buffered
saline solution (PBS) with 5% penicillin, streptomycin, and amphotericin (GIBCO,
Grand Island, NY). The minced specimens were placed in scored 25-cm2 tissue (T25) flasks (Falcon; Becton-Dickinson, Franklin Lakes, NJ)
with a 2.5-mL solution of primary culture media (20% fetal bovine serum in
Dulbecco modified Eagle medium; 1% penicillin, streptomycin, and amphotericin;
and 1% L-glutamine) (GIBCO). The dermal specimens were stored and maintained
at 37°C in a humidified 5% carbon dioxide atmosphere.
After 24 hours, the media were changed with 5.0 mL of primary culture
media. The media were then changed every 2 days until fibroblasts were visualized
under light microscope to be growing outward from the explanted tissue. At
this time the tissue was removed. With sufficient outgrowth of fibroblasts,
cells were passed into 75-cm2 tissue (T75) flasks using 0.05% trypsin
(GIBCO) in PBS. Primary culture media were changed every third day, and successive
cultures were passed at confluence. Cells from passages 4 and 5 were used
for experimentation. All work was performed under a laminar flow hood using
a sterile technique.
A concentrate was prepared by dissolving the GHK-Cu (Procyte Corp, Kirkland,
Wash) in PBS just before experimentation, which was then diluted in serum-free
media to a concentration of 1.0 × 10 −9 mol/L. Previous
studies9-10 have shown this concentration
to be effective in stimulating collagen and glycosaminoglycan production by
Tretinoin (Sigma-Aldrich Corp, St Louis, Mo) was dissolved in ethanol
immediately before use, then diluted in commercially available serum-free
media (UltraCulture; Biowhittaker, Walkersville, Md) to achieve a final concentration
of 1.0 × 10 −5 mol/L. This concentration has proven
to be most effective for in vitro studies of procollagen inhibition.21, 24 Total concentration of solvent was
less than 0.1%. This concentration has been shown to be nontoxic to fibroblasts
and has no effect on collagen metabolism by these cells.21
The corresponding amount of ethanol was added to the serum-free media in control
At the time of experimentation, fibroblasts were released from flask
walls using 0.05% trypsin solution. The trypsin was inactivated using trypsin
soybean inhibitor (GIBCO) in a 1:1 ratio. Cells were suspended in UltraCulture
and then transferred into 24-well culture plates (Falcon; Becton-Dickinson)
at a concentration of 6 × 104 cells/well. UltraCulture was
selected for its ability to sustain dermal fibroblast growth to at least 7
days with greater than 90% viability.25 Cells
were counted in duplicate using phase-contrast microscopy and a hemocytometer.
Viable cells were determined using trypan blue exclusion. The plates were
incubated for 48 hours to allow for adequate settling. After this time, cells
were washed in PBS and fresh UltraCulture was added, this time with the appropriate
modulator or vehicle included (0 hour).
The supernatant was drawn from the culture wells at 24, 72, and 120
hours and stored at –70°C until the time of assay. Each sample was
assayed using the enzyme-linked immunosorbent assay technique. Growth factors
bFGF and TGF-β1 were assayed using Quantikine High Sensitivity and Quantikine
assay kits (R&D Systems, Minneapolis, Minn), respectively. Assays were
read using an automated plate reader (Elx800; Bio-Tek Instruments Inc, Winooski,
Vt). Optical densities were analyzed with KC4 software (Bio-Tek Instruments
Inc). Assays were read with the specified filter for each assay with application
of a reference filter to correct for optical imperfections in the plate.
Each data point represents duplicate cell counts with assays performed
in duplicate. Statistical differences were assessed using 2-sample and paired t tests. Differences at the 5% level were considered statistically
All cell lines grew in the modulated, serum-free environment, with cell
viability ranging from 86% to 96% (mean, 92%). Differential growth factor
secretion patterns were observed and are described in the following 2 subsections.
Levels of bFGF measured in supernatant samples peaked at 24 hours and
progressively declined throughout the experiment for KFs and NFs, with no
significant difference between the 2 cell types. Greater concentrations of
bFGF were detected in samples of NFs treated with tretinoin than in controls
at 24 hours (mean, 19.6 and 8.0 pg/mL, respectively; P<.05)
(Figure 1). No differential secretion
pattern was observed in NFs treated with GHK-Cu. Similarly, KFs treated with
either modulator did not demonstrate any trend in bFGF secretion compared
Samples obtained from KFs had higher levels of TGF-β1 than did
those from NFs in control and tretinoin-treated groups at 120 hours (not statistically
significant) (Figure 2). Tretinoin-treated
KFs also secreted more TGF-β1 than did controls at 120 hours (mean, 58.5
and 24.6 pg/mL, respectively; P<.05), whereas
NFs showed no such sensitivity to tretinoin in terms of TGF-β1 concentrations
(Figure 2). This pattern was consistent
for each keloid specimen (Figure 3).
Both fibroblast types treated with GHK-Cu secreted less TGF-β1 than did
controls, although this was significant only for KFs at 24 hours (mean, 29.9
and 17.7 pg/mL, respectively; P<.05) (Figure 4). This pattern was consistent for
each keloid specimen (Figure 5).
Cell culture has long been a primary means toward understanding the
activity of human dermal fibroblasts—normal or keloid. It has often
proved an especially useful way of differentiating the behavior of these 2
cell types as they relate to the wound-healing environment. Keloid fibroblasts,
eg, have been shown in vitro to produce more collagen than their normal dermal
counterparts. Although many studies have demonstrated differences by measuring
specific cellular proteins such as collagen, rarely has a model specifically
focused on the growth factors that trigger their production. This is largely
because of the difficulty in controlling for growth factor levels in a serum-containing
cell culture model, traditionally used to maintain cell growth. Our study
demonstrated that NFs treated with tretinoin produce more bFGF than do controls,
whereas NFs and KFs treated with copper tripeptide secrete less TGF-β1
than do controls. Each result suggests a correlation between growth factor
production and known clinical effects.
Only recently have studies23 established
KF and NF cell lines in serum-free media. The present study demonstrates the
viability of such cells in a serum-free model and assayed for growth factors
known to figure prominently in the wound-healing process. Such methods have
been adopted previously, although serum-free media have not been used past
the incubation phase of cell culture.7 A potential
disadvantage of serum-free media is that fibroblast proliferative characteristics
and viability are generally not as good as with serum-based models. In short-term
culture, however, the medium used in our experiments supported similar growth
characteristics and comparable cell viability to that of serum-based models
of similar experiment duration.25
Altering the wound environment through chemical modulators, as demonstrated
in the present study, might provide insight as to the link between proven
clinical applications and the induced cellular response. The mechanism by
which tretinoin exerts its cellular effects is linked to the retinoinc acid
receptors, discovered in 1987.26 Tretinoin
binds these intracellular receptors—similar in makeup and function to
steroid/thyroid hormone receptors—which in turn bind regulatory regions
of cellular DNA, causing activation of gene transcription. Several of these
target sequences are contained by genes that have been shown to be markers
of tretinoin stimulation, including cellular retinoic acid binding protein.
By directly stimulating the transcriptional machinery of the cell, tretinoin
is able to modulate the production of proteins central to cell growth and
differentiation. These proteins might then initiate a cascade effect whereby
other DNA segments are transcribed, including those coding for ECM proteins,
and thus account for the ultimate clinical effect of tretinoin.
In the present study, retinoic acid seems to stimulate secretion of
"collagen tightening" growth factor (bFGF) by NFs. This may partially explain
its known clinical utility. That maximal levels of bFGF were observed at 24
hours in all cell lines is consistent with the half-life of bFGF (25 hours).27 The application of tretinoin initially stimulated
fibroblasts to produce bFGF. Levels then gradually declined over the course
of the experiment as the bFGF degraded. Although some clinical studies have
shown topical tretinoin use to reduce the size of keloid scars, our data suggest
a mechanism other than modulation of TGF-β1, as keloid cell lines secreted
more of this growth factor than did controls.
Treatment with GHK-Cu stimulates glycosaminoglycan and collagen production
in human fibroblasts, critical to the postinflammatory phase of wound healing.
More recently it has been postulated that GHK-Cu stimulates specific matrix
metalloproteinases.28 Other studies29-30 have demonstrated that in addition
to its direct wound-healing effects, GHK-Cu enables angiogenesis and leukocyte
chemoattraction. The exact mechanism whereby copper tripeptides alter cellular
activities has yet to be worked out, despite such observed phenomenon. Modulation
of local growth factor production at the site of active wound healing might
be involved in these processes.
As already described,7 KFs produce more
TGF-β1 than do NFs in culture, and our data reinforce this property.
In the present study, copper tripeptide therapy seems to suppress secretion
of "fibrogenic" growth factor (TGF-β1) in NFs and especially in KFs,
and it may have application in decreasing excess scar formation.
Because of the many antagonisms of growth factor activities, it may
be possible to correct a deficiency or overabundance with local application
of another factor that modulates the wound cells' growth factor production
profile. Once a modulator's (or combination thereof) autocrine growth factor
stimulatory properties are known, it could be placed into a wound to achieve
the desired healing response. Routine wound application of recombinant-produced
or autologous-derived growth factors would be expensive. Using obtainable
modulators such as tretinoin and copper tripeptide as cytokine stimulators
would circumvent this problem.
In the larger scheme, using cytokine manipulations to vary the makeup
of ECM components (such as collagen) might have a great impact in precisely
controlling the wound-healing process. For example, if a person with diabetes
has a nonhealing ulcer, the wound could be treated with a modulator that stimulates
production of a fibrosis-producing growth factor. The appearance of the wound
in this case is not as important as closure by scar tissue. Also, if an irradiated
patient has an open wound because of poor tissue blood supply, the wound could
be treated with a modulator that stimulates production of an angiogenic growth
factor, which will cause local development of blood vessels. Finally, anyone
undergoing surgery may benefit from wound treatment with a modulator causing
production of a growth factor that causes an increase of collagen with tighter
bundles, thus forming a smaller yet stronger scar.
In summary, the results of our study demonstrate that NFs treated with
retinoic acid produce more bFGF than do controls, and this might partially
explain the clinically observed tightening effects of tretinoin. Both NFs
and KFs treated with copper tripeptide secreted less TGF-β1 than did
controls, and this suggests possible clinical use for decreasing excessive
scar and keloid formation.
Accepted for publication February 23, 2000.
Presented in part at the American Academy of Facial Plastic and Reconstructive
Surgery 1999 Spring Meeting as part of the Combined Otolaryngological Spring
Meetings, Palm Desert, Calif, April 28, 1999.
Corresponding author: R. James Koch, MD, MS, Facial Plastic and Reconstructive
Surgery, Division of Otolaryngology–Head and Neck Surgery, Stanford
University Medical Center, Stanford, CA 94305-5328 (e-mail: RJK@stanford.edu).