β-Adrenergic Receptor Antagonists Accelerate Skin Wound Healing

The skin is our primary defense against noxious environmental agents. Upon injury, keratinocytes migrate directionally into the wound bed to initiate re-epithelialization, essential for wound repair and restoration of barrier integrity. Keratinocytes express a high level of β2-adrenergic receptors (β2-ARs) that appear to play a role in cutaneous homeostasis as aberrations in either keratinocyte β2-AR function or density are associated with various skin diseases. Here we report the novel finding that β-AR antagonists promote wound re-epithelialization in a “chronic” human skin wound-healing model. β-AR antagonists increase ERK phosphorylation, the rate of keratinocyte migration, electric field-directed migration, and ultimately accelerate human skin wound re-epithelialization. We demonstrate that keratinocytes express two key enzymes required for catecholamine (β-AR agonist) synthesis, tyrosine hydroxylase and phenylethanolamine-N-methyl transferase, both localized within keratinocyte cytoplasmic vesicles. Finally, we confirm the synthesis of epinephrine by measuring the endogenously synthesized catecholamine in keratinocyte extracts. Previously, we have demonstrated that β-AR agonists delay wound re-epithelialization. Here we report that the mechanism for the β-AR antagonist-mediated augmentation of wound repair is due to β2-AR blockade, preventing the binding of endogenously synthesized epinephrine. Our work describes an endogenous β-AR mediator network in the skin that can temporally regulate skin wound repair. Further investigation of this network will improve our understanding of both the skin repair process and the multiple modes of action of one of the most frequently prescribed class of drugs, hopefully resulting in a new treatment for chronic wounds.

␤-Adrenergic receptors (␤-ARs) 2 are expressed on a wide variety of tissues and are recognized as pivotal functional regu-lators of the cardiac, pulmonary, vascular, endocrine, and central nervous systems. Although their expression in human skin was noted over 30 years ago (1), only recently has their functional significance in this tissue been recognized. The ␤2-AR subtype is the only subtype of ␤-ARs expressed on the membranes of the major cell types in skin: keratinocytes (2)(3)(4), fibroblasts (5), and melanocytes (6). Interestingly, aberrations in either keratinocyte ␤2-AR function or density have been associated with cutaneous disease. Keratinocytes derived from patients with atopic eczema display a point mutation in the ␤2-AR gene and a low ␤2-AR density (7). In psoriasis, keratinocytes within the psoriatic lesions demonstrate a low cAMP response to ␤2-AR activation (8). These findings point to a role for the ␤2-AR in maintaining epidermal function and integrity.
Here we provide data to support a role for the ␤2-AR in regulating wound repair as well.
Cutaneous wound healing is a complex and well orchestrated biological process requiring the coordinated migration and proliferation of both keratinocytes and fibroblasts, as well as other cell types. Wounding the epidermis prompts the epidermal and dermal cells to generate cytokines, growth factors, and proteases and to synthesize extracellular matrix components, all of which can regulate the processes of keratinocyte migration and proliferation, essential for re-epithelialization (9,10). Upon injury, the cells migrate directionally toward the center of the wound bed to initiate repair and restore epithelial integrity. Many cues play a role in wound-induced keratinocyte directional migration including contact inhibition and chemotaxis (11,12). Additionally, electric fields (EFs) play an important role in epithelial cell directional migration (13)(14)(15) and wound healing (15)(16)(17)(18). Indeed, increasing or decreasing corneal wound currents pharmacologically correlates directly with an increased or decreased rate of healing in the rat cornea (19).
The first clues to a biological function for ␤2-AR in wound repair came from an early study demonstrating that ␤-AR agonists delay skin wound healing in newt limbs (20). Subsequent studies in other epithelia, however, have yielded conflicting results. For example, ␤-AR antagonists have been reported to either delay (21,22) or enhance (23) corneal epithelial wound healing. Recently, Denda et al. (24) reported that the ␤-AR could modulate epidermal barrier permeability by measuring transepidermal water loss.
Here we investigate the effects of ␤-AR antagonists on scratch wound healing, keratinocyte single cell migration, ERK phosphorylation, keratinocyte galvanotaxis, cytoskeletal orga-nization, proliferation, and, ultimately, human skin wound reepithelialization in a "chronic" wound healing model. We demonstrate that the ␤-AR antagonist is pro-motogenic, promoting human skin re-epithelialization. Ultimately, we explore the mechanism for the ␤-AR antagonist-mediated acceleration of wound repair, reporting the expression of protein for two key catecholamine (␤-AR agonist) synthesis enzymes that are localized within keratinocyte cytoplasmic granules/vesicles. Finally, we confirm that keratinocytes synthesize epinephrine. Thus, our work uncovers an endogenous ␤-AR mediator network in the skin that upon blockade accelerates wound healing. Clearly, further investigation of this network will improve our understanding of the skin wound repair process and hopefully lead to the development of new therapies to enhance wound healing.
Keratinocyte Growth-Human keratinocytes were isolated from neonatal foreskins as we have reported previously (25), under an approved exemption granted by the Internal Review Board at University of California, Davis, and cultured using a modification of the method of Rheinwald and Green (26). The cells were grown in keratinocyte growth medium (KGM) (Epilife, 0.06 mM Ca 2ϩ ), supplemented with human keratinocyte growth supplement (0.2 ng/ml epidermal growth factor, 5 g/ml insulin, 5 g/ml transferrin, 0.18 g/ml hydrocortisone, and 0.2% bovine pituitary extract) (Cascade Biologics, Inc., Portland, OR) and antibiotics (100 units/ml penicillin, 100 g/ml streptomycin, and 0.25 g/ml amphotericin) (Gemini Bio-Products, Inc., Calabasas, CA) at 37°C in a humidified atmosphere of 5% CO 2 . Cell strains isolated from at least three different foreskins were used in all of the experiments, performed with subcultured cells between passages 3 and 7.
Scratch Assay-Three different strains of keratinocytes were grown to confluence in KGM on 35-mm culture dishes (Fisher). The cells were either untreated (control) or treated with ␤-AR antagonist (ICI 118,551) at a concentration specific for the ␤2-AR (27), 10 nM, in KGM at time 0. A sterile pipette tip was used to scratch three 500-m-wide to 1-mmwide wounds around the center of the dish, and three demarcated areas of each wound were photographed on an inverted Nikon Diaphot microscope at the time of wounding (time 0) and at 16 h post-wounding (28) at 10ϫ magnification. Image J was used to measure the area of the wound at each demarcated area at time 0 and at 16 h to calculate the percentage of healing at the three positions for each wound, and all data for both the untreated and antagonist-treated wounds were statistically analyzed to determine the average percentage of healing for each condition.
Single Cell Migration Assay-Glass-bottomed 35-mm dishes (MatTek Corporation, Ashland, MA), were coated with collagen I (60 g/ml) (Cohesion Technologies, Palo Alto, CA) in PBS for 1 h at 37°C. Keratinocytes were plated at a density of 50 cells/mm 2 for 2 h at 37°C. The cells were incubated with KGM alone (control) or with KGM containing 20 M ␤-AR antagonist (timolol) at time 0. The 35-mm glass-bottomed dishes were placed in a heating chamber designed to maintain the medium between 35 and 37°C and secured to the stage of an inverted Nikon Diaphot microscope. Individual cell migration was monitored over a 1-h period at 37°C, as described previously (29). Time lapse images of the cell migratory response were digitally captured every 10 min by Q-Imaging Retiga-EX cameras (Burnaby, Canada) controlled by a custom automation written in Improvision Open Lab software (Lexington, MA) on a Macintosh G4. After the center of mass of each cell was tracked using the Open Lab software, migration speed and distance were calculated and imported to Excel (Microsoft Corporation, Redmond, WA). "Speed" is the average speed in m/min that the cells travel in a 1-h period of time. Significance was taken as p Ͻ 0.01, using Student's t test (unpaired) to compare the means of two cell populations.
Cell Treatments for Immunoblotting-1 ϫ 10 6 plated keratinocytes were incubated with either KGM alone (control and lysates for catecholamine synthesis enzyme detection) or KGM containing 20 M ␤-AR antagonist for 5-60 min. The cells were placed immediately on ice, washed twice with ice-cold PBS containing phosphatase inhibitors (50 mM NaF and 1 mM Na 3 VO 4 ) and scraped in 50 l of lysis buffer; (PBS containing 0.5% Triton X-100, 50 mM NaF, 1 mM Na 3 VO 4 , 10 g/ml leupeptin, 30 g/ml aprotinin, 200 g/ml phenylmethylsulfonyl fluoride, and 10 g/ml pepstatin A). The lysates were transferred into 1.5-ml tubes, incubated on ice for 20 min, and then centrifuged at 14,000 ϫ g for 10 min at 4°C (30). The protein concentration of the samples was determined using the Bradford Assay (Bio-Rad). The supernatants were electrophoresed immediately on 10% polyacrylamide Tris-HCl gels (Bio-Rad) or stored at Ϫ80°C.
Five g (P-ERK blots) or thirty-five g (catecholamine synthesis enzyme blots) of each protein sample was added to an equal volume of 2ϫ reducing sample loading buffer (0.0625 M Tris-HCl, pH 6.8, 3% SDS, 10% glycerol, 5% ␤-mercaptoethanol) and electrophoresed on 10% polyacrylamide Tris-HCl gels. The proteins were transferred to Immobilon membranes and immunoblotted with an anti-ERK (number 9102), phospho-ERK (number 9101), TH, or PNMT antibody at a concentration recommended by the manufacturer. The immunoblots were developed by ECL according to the manufacturer's instructions (Amersham Biosciences). Densitometry was performed on scanned images using NIH Image 1.62.
Galvanotaxis-The galvanotaxis chamber construction and EF application were performed as described previously (31). Briefly, the galvanotaxis chamber is composed of a rectangular plexiglass frame with two medium reservoirs on opposite sides to which a 45 ϫ 50-mm piece of No. 1.5 glass coverslip is attached to form the chamber bottom. This allows for continual observation of the plated cells on an inverted microscope. The keratinocytes are plated onto the collagen-coated center of the chamber between two coverslip spacers 25 ϫ 10 mm. A third 25-mm 2 coverslip is placed on top, straddling the two spacer coverslips and covering the cells plated on the collagen-coated center panel. This third coverslip rests ϳ100 -105 m above the center panel and is sealed on top of the spacer coverslips with silicone high vacuum grease (Dow Corning, Midland, MI). This small height is chosen to minimize the cross-sectional area through which current flows. A small cross-section creates a high resistance pathway, resulting in a higher voltage gradient for a fixed current. The aqueduct allows for medium and current flow across the cells. The voltage across the coverslip is measured using a voltmeter via silver-sliver chloride (Ag-AgCl) wire electrodes inserted into both medium reservoirs on either side of the center panel. Six-cm-long 2% agar/phosphate buffered saline-filled pieces of polypropylene tubing connect each end of the chamber to a medium-filled well in which the Ag-AgCl electrodes are placed to separate possible electrode byproducts from the cells themselves. The current is measured with an ammeter in series, and we only use chambers for which the current flow is kept below 0.6 mA to minimize joule heating. Furthermore, the temperature of the medium in the chambers is maintained at 36°C by placing the chamber on a metal plate heated to and maintained at 39°C. Temperature is continuously monitored during the experiment using a YSI 400 analog temperature probe (Yellow Springs Instrument Co., Inc., Yellow Springs, OH) directly attached to the metal plate and does not vary by more than 1°C over the course of the experiment.
Time Course Observation and Data Analysis-The galvanotaxis chambers rest on inverted Nikon microscopes. Time lapse images of the cell migratory response are digitally captured every 10 min over a 1-h period by a Q-Imaging Retiga-EX cameras controlled by a custom automation written in Improvision Open Lab software (Lexington, MA) on a Macintosh G4. After the center of mass of each cell is tracked using the Open Lab software, the directionality of migration is calculated and imported to Excel (Microsoft Corporation, Redmond, WA). Cosine describes the direction of migration and is a measure of the persistence of cathodal directedness, where is the angle between the field axis and the vector drawn by the cell migration path. The average cosine ϭ S i cos i /N, where S i is the summation of 70 -72 individual cells from at least three different cell strains. Angle zero ( ϭ 0 0 ) is assigned to the negative pole (cathode) and increasing angles assigned in a clockwise manner, with ϭ 180 0 aligned with the positive pole (anode). Therefore, the cosine will provide a number between Ϫ1 and ϩ1, and the average of all of the separate cell events yields an index of directional migration. For example, if a cell were to move directly to the negative pole, the angle () ϭ 0 0 , and the cosine ϭ 1. "Cosine" therefore refers to the average directional migration index of separate cell migration events at the end of a 1-h period. The results are given as average cosine Ϯ S.E. Significance is taken as p Ͻ 0.01, using Student's t test (unpaired) to compare the means of two cell populations.
Immunofluorescent Staining and Microscopy-Sterile glass coverslips were transferred into 12-well dishes and collagencoated with 60 g/ml collagen I in KGM for 1 h at 37°C as described. The coverslips were washed three times with KGM, and 3 ϫ 10 4 cells were added per well and allowed to attach overnight. The cells were left untreated or treated with 20 M ␤-AR antagonist for 15 min. The coverslips were processed at room temperature unless otherwise noted. The coverslips were washed twice in PBS and fixed for 10 min in 4% paraformaldehyde. The coverslips were washed twice in PBS between each step. The cells were permeabilized for 5 min with 0.1% Triton X-100 with PBS and blocked with 5% goat serum with PBS (vinculin) or 5% horse serum with PBS (TH and PNMT) for 20 min. Primary monoclonal anti-vinculin antibody was added dropwise in 1% goat serum with PBS (1:100) and primary anti-TH or anti-PNMT antibody were added dropwise in 1% horse serum with PBS (1:20) and incubated for 1 h at 37°C. A goat anti-mouse cy3 antibody (Jackson Laboratory, West Grove, PA) (1:100) was added in 1% goat serum with PBS for 1 h at 37°C (vinculin) or a donkey anti-rabbit cy3 antibody (Jackson Laboratory) (1:100) was added in 1% horse serum with PBS for 1 h at 37°C (TH and PNMT). Alexa 488-phalloidin (Molecular Probes, Eugene, OR) (1:40 in PBS) was added to the vinculin-stained coverslips for 20 min. Standard controls were performed. The coverslips were incubated with the primary antibody alone or the secondary antibody alone to ensure specificity. Finally, Prolong gold anti-fade reagent (Molecular Probes) was used according to the manufacturer's instructions to mount the coverslips onto glass microscope slides. The slides were viewed on an inverted fluorescent Nikon Diaphot microscope using a 40ϫ pan fluor objective. The images were captured using Q-Imaging Retiga-EX cameras and pseudo-colored green for Alexa 488 Phalloidin staining (actin) and red for Cy3 staining (vinculin) or visualized in gray scale for TH and PNMT.
Proliferation Assay-Keratinocytes were released from the tissue culture plate by treatment with 0.25% trypsin, 0.1% EDTA (Invitrogen), resuspended in KGM, and counted using a hemocytometer. 5 ϫ 10 4 cells were plated per well in a 12-well plate in triplicate and allowed to settle and attach to the plate for 2 h. The cells were then cultured in the presence or absence of 20 M ␤-AR antagonist. Triplicate wells were harvested and counted on days 2, 4, 6, and 8. The medium was changed every day. Significance was taken as p Ͻ 0.01, using Student's t test (unpaired) to compare the means of the cell populations.
Chronic Human Skin Wound Healing Assay-We have previously adapted a wound healing model developed by Kratz (32) to observe a delay in human skin re-epithelialization in the presence of ␤-AR agonist (33). Here we have adapted a chronic wound healing model also developed by Kratz (32). We have reduced the serum content of the medium from 10% to 5% to generate a chronic wound healing model, severely delaying the rate of re-epithelialization in control wounds to enable us to observe any increase in the rate of re-epithelialization in the presence 10 M ␤-AR antagonist. Normal human skin was obtained from routine breast reductions or abdominoplasties under an approved exemption granted by the Internal Review Board at University of California, Davis. Under sterile conditions, excess subcutaneous fat was trimmed from 6 ϫ 3-inch sections of skin prior to stretching and pinning onto sterile corkboard. A 3-mm punch (Sklar Instruments, West Chester, PA) was used to make wounds in the epidermis and into the superficial dermis, and the 3-mm discs of skin were excised using sterile scissors. 6-mm skin discs, with the 3-mm epidermal wound in the center, were excised using a 6 -mm biopsy punch (SMS Inc., Columbia, MD). The skin samples were immediately transferred to a 12-well dish and submerged in 2 ml of FM (Dulbecco's modified Eagle's medium (Invitrogen) containing 5% fetal bovine serum (Tissue Culture Biologicals, Tulare, CA) and antibiotics). The 12-well dishes were incubated at 37°C in a humidified atmosphere of 5% CO 2 . The medium was changed every day. Three biopsies were fixed in 4% neutral buffered formaldehyde every day for 5 days. To confirm that untreated wounds retain the capacity to heal, biopsies previously cultured for 4 days in 5% serum were cultured in the presence or absence of an additional 5% serum (i.e. 10% total) for a further 4 days prior to fixation and histological processing.
The formaldehyde-fixed biopsies were dehydrated through an ethanol-xylene series and embedded in paraffin. Cross-sections, 5 M thick, taken from the center of the wound, were stained using the hematoxylin-eosin technique. Re-epithelialization was determined using light microscopy. The slides were viewed on an inverted Nikon Diaphot microscope using a 10ϫ objective. The images were captured using Q-Imaging Retiga-EX cameras. Specimens that were damaged in the histologic process or otherwise noninterpretable were excluded from the study. The percentage of re-epithelialization was calculated by measuring the linear distance covered by new epithelium and dividing that by the linear distance between the original wound edges. The new epithelium was clearly differentiated from the epithelial wound margin by the presence of a fully stratified epithelium and fully formed stratum corneum in the latter. To eliminate observer bias, the percentage of re-epithelialization was calculated from randomly numbered pictures of the wounds by a third party. Significance was taken as p Ͻ 0.01, using Student's t test (unpaired) to compare the means of the percentage of re-epithelialization of the wounds in the presence of 5 and 10% serum for days 5-8. To observe any ␤-AR antagonist-mediated modulation in reepithelialization, the skin samples were immediately trans-ferred to a 12-well dish and submerged in 2 ml of FM in the presence or absence of 10 M ␤-AR antagonist. Three biopsies were fixed in 4% neutral buffered formaldehyde every day for 5 days. Significance was taken as p Ͻ 0.01, using Student's t test (unpaired) to compare the means of the percentage of re-epithelialization of the control and ␤-AR antagonisttreated wounds on days 3-5.
Enzyme Immunoassay for the Quantitative Determination of Epinephrine in Small Sample Volumes-1 ϫ 10 7 keratinocytes were extracted in 100 l of 0.1 N HCl and sonicated on ice for 10 min. Extracts from three strains of keratinocytes were tested yin triplicate in an epinephrine enzyme immunoassay (BIOSOURCE, Camarillo, CA) according to the manufacturer's instructions. Briefly, the assay kit provides materials for the quantitative measurement of epinephrine. Epinephrine is extracted using a cisdiol-specific affinity gel, then acylated to N-acylepinephrine, and after this converted enzymatically during the detection procedure into N-acylmetanephrine. The competitive enzyme immunoassay uses the microtiter plate format. Epinephrine is bound to the solid phase of the microtiter plate. Acylated epinephrine and solid phase bound epinephrine compete for a fixed number of antiserum-binding sites. When the system is in equilibrium, free antigen and free antigen-antiserum complexes are removed by washing. The antibody bound to the solid phase catecholamine is detected by an anti-rabbit IgG peroxidase conjugate using TMB as a substrate. The reaction is monitored at 450 nm with the amount of antibody bound to the solid phase being inversely proportional to the catecholamine concentration in the sample. A set of standards and two controls are included for determination of unknown concentrations (0, 5.6, 19, 83, 306, and 1550 pg of epinephrine/sample). The linear mean absorbance readings of the standards are plotted on the y axis versus the log of the concentrations of the standards (pg/sample) on the x axis, and a linear curve fit is applied. The concentration of epinephrine in the unknowns can then be calculated from the slope of the line. To standardize the levels of epinephrine measured in the keratinocyte extracts, a Bradford assay was performed on the extracts, as described, and the level of epinephrine detected was calculated as pg of epinephrine/mg of protein.

RESULTS
␤-AR Antagonists Accelerate the Healing of Scratch Wounds in Confluent Keratinocyte Cultures-To determine the effect of ␤-AR antagonists on keratinocyte migration, we initially measured the ability of keratinocytes to heal a "scratch wound" within a confluent sheet of cells (34). A ␤-AR antagonist, at a concentration specific for ␤2-AR (10 nM) (27), doubles the rate of scratch wound healing. Untreated wounds are only 32 Ϯ 8.5% healed within 16 h, whereas ␤-AR antagonist-treated wounds are 62 Ϯ 2% healed within the same time frame (Fig. 1). A nonspecific ␤-AR antagonist also accelerates keratinocyte scratch wound healing (results not shown).
␤-AR Antagonists Increase the Rate of Keratinocte Single Cell Migration-To more precisely measure the effect of ␤-AR antagonists on motility, we observed the effect of ␤2-AR blockade on the locomotory speed of individual keratinocytes. The ␤-AR antagonist significantly increases the keratinocyte migration rate in the 1-h period by 30% from 1.05 Ϯ 0.03 to 1.36 Ϯ 0.03 m/min.
A ␤-AR Antagonist Increases ERK Phosphorylation within Minutes in Keratinocytes-ERK plays a pivotal role in pro-migratory signaling pathways (35)(36)(37)(38) and is critical for the healing of scratch wounds in confluent monolayers of epithelial cells (39 -44). Here we demonstrate that ␤-AR antagonist treatment dramatically increases ERK phosphorylation by 5-fold within 5 min. ERK phosphorylation remains elevated for up to 60 min in the presence of ␤-AR antagonist while gradually returning toward control levels (Fig. 2).

␤-AR Antagonists Enhance Keratinocyte Electric Field-mediated Directionality of Migration-
The EF, generated immediately upon wounding (45), could be the earliest signal that cells receive to direct their migration into the wound bed. Because the center of the wound is negative with respect to the wound edges (46,47) and keratinocytes migrate toward the negative pole (cathode) in an applied EF (48), the wound-generated endogenous EF likely orients keratinocyte directional migration toward the wound center. Here we investigate the effect of ␤-AR antagonists on the ability of keratinocytes to sense and respond to an applied EF of physiological strength (100 mV/mm).
␤-AR antagonist treatment significantly increases the directionality of migration (cosine ) by 26% from 0.69 Ϯ 0.04 to 0.87 Ϯ 0.03. Additionally, as expected, the ␤-AR antagonist increases the rate of migration by 28% from 1 Ϯ 0.03 to 1.28 Ϯ 0.04 m/min as described previously (results not shown).
␤-AR Antagonists Preserve the Keratinocyte Pro-migratory Cytoskeletal Architecture-Actin remodeling plays an important role in cell polarization (49) and motility (50). Actin filaments terminate in focal adhesions, where several proteins including vinculin mediate interactions with the actin cytoskeleton (51). Focal adhesions mediate the mechanical attachment of cells to the extracellular matrix (52) and act as signaling centers, capable of regulating gene expression, cell growth, and survival (53). Additionally, small, nascent focal adhesions have been associated with actively migrating cells (54). Because we have demonstrated that ␤2-AR blockade increases the rate of keratinocyte migration, we wondered whether we could observe any alterations in the actin cytoskeleton.
Cells plated in the absence of ␤-AR antagonist are polarized and crescent-shaped with a broad lamellipodium (Fig. 3), characteristic of the migratory phenotype (49). In keratinocytes the majority of the actin fibers and focal adhesions appear to be restricted to the lamellipodia (Fig. 3). ␤-AR antagonist treat-

FIGURE 2. ␤-AR antagonists increase the phosphorylation of ERK within minutes in keratinocytes.
Keratinocytes were cultured in KGM and plated as described. The cells were either untreated (time 0) or incubated with 20 M ␤-AR antagonist for up to 60 min at 37°C. After treatment, the cell lysates were prepared as described, electrophoresed on 10% polyacrylamide gels, and transferred to membrane. Identical membranes were immunoblotted with either an anti-phospho ERK or an anti-ERK antibody (A). The data shown are representative of three independent experiments from three separate cell strains. Three blots from separate experiments were scanned for ERK and p-ERK and densitometry performed using a gel plotting macro in NIH Image 1.62. The data were normalized so that the average of the mean phosphorylated ERK signal divided by the mean ERK signal for controls was set at 100%. All of the data were averaged, statistically analyzed, and represented graphically (B). The values plotted are the means ϩ S.E. (n ϭ 3). *, p Ͻ 0.01 compared with the control. ment appears to have no effect on cytoskeletal conformation. The cell morphology, actin cytoskeleton and the number, size, and distribution of focal adhesions appear similar to untreated keratinocytes (Fig. 3).
␤-AR Antagonists Have No Effect on Keratinocyte Proliferation-Keratinocyte proliferation behind the epithelial tongue is essential for efficient human skin re-epithelialization (10). ␤-AR antagonists appear to have no effect on keratinocyte proliferation in vitro (Fig. 4).
Chronic Human Skin Wounds Retain the Capacity to Heal in the Presence of 10% Serum-Here we investigate the effect of ␤-AR antagonists on human skin re-epithelialization in an ex vivo wound healing model adapted to resemble chronic wound healing (33). Human skin punch biopsies are cultured in medium containing a reduced percentage of serum (5%) to significantly delay the re-epithelialization of untreated human skin wounds (Fig. 5A). To demonstrate that untreated wounds retained the capacity to heal, biopsies previously submerged in medium containing 5% serum for 4 days are cultured for a further 4 days in the presence or absence of an additional 5% serum prior to fixation. Wounds cultured in 10% serum for an additional 4 days are almost completely healed by day 8 (Fig. 5B). The percentage of re-epithelialization of wounds cultured for the last 4 days in 10% serum is 66% higher than wounds cultured for the entire 8-day period in the presence of just 5% serum (Fig. 5C).
␤-AR Antagonists Accelerate Skin Wound Re-epithelialization-Human skin was wounded, and the wounds were allowed to re-epithelialize in 5% serum in the presence and absence of ␤-AR antagonist. ␤-AR antagonist treatment significantly increases the rate healing of human skin wounds, and almost all ␤-AR antagonist-treated wounds are re-epithelialized by day 5 (Fig. 6). Hematoxylin and eosin-stained sections from control and ␤2-AR antagonist-treated wounds on days 3-5 are shown in Fig. 6A, highlighting the ␤-AR antagonist-mediated acceleration of skin wound repair (Fig. 6A). Because of variations in wound shape and the site within the wound from which sections were cut, leading to variation in the healing we observed on days 3-5, the percentage of re-epithelialization was calculated for each wound. ␤-AR antagonist treatment significantly increases the wound re-epithelialization by 40, 63, and 72% after 3, 4, and 5 days in culture, respectively (Fig. 6B). These results provide confirmation that ␤2-AR blockade accelerates wound re-epithelialization in human skin.
Keratinocytes Express the Enzymes Necessary to Convert L-Tyrosine to Epinephrine, Localized within Cytoplasmic Vesicles/Granules and Synthesize Epinephrine Endogenously-Catecholamines provide important biological functions, acting as both neurotransmitters and endocrine hormones. The conversion of L-tyrosine to L-dopa by TH is the rate-limiting step for catecholamine biosynthesis (55,56), and PNMT catalyzes the synthesis of epinephrine from nor epinephrine (57) (Fig. 7A). Previously, enzyme activity for TH and PNMT and mRNA for TH has been discovered in undifferentiated keratinocytes (6,58,59). To determine whether we could detect protein for catecholamine synthesis enzymes in keratinocytes, we lysed cells and immunoblotted with antibodies specific for TH and PNMT. A PC12 cell lysate was used as a positive control for TH (60) but not for PNMT because PC12 cells contain negligible PNMT (61). A dermal fibroblast lysate was used as a negative control (58). Human TH and PNMT enzymes are reported to be approximately 61-62 and 30 -32 kDa in size, respectively (62). Indeed, the TH antibody detected one major protein at approximately 61 kDa, and the PNMT antibody detected one major protein at approximately 32 kDa (Fig. 7B). To determine the localization of TH and PNMT in keratinocytes, cell cultures were immunostained with the anti-TH and anti-PNMT antibodies. Multiple brightly stained TH and PNMT-containing circular structures/granules can be observed distributed throughout the keratinocyte cytoplasm (Fig. 7C). Immunostaining was performed on dermal fibroblasts as a negative control (58) (Fig. 7C). Finally, we were able to measure 303, 468, and 888 pg of epinephrine/mg of protein in keratinocyte extracts from three different keratinocyte strains, as described. The detection of epinephrine in keratinocyte extracts provides convincing evidence that epinephrine is endogenously synthesized by keratinocytes.

DISCUSSION
Here we report the novel finding that ␤-AR antagonists promote wound re-epithelialization by blocking an autocrine ␤2-AR network within the epidermis. ␤-AR antagonists enhance the ability of keratinocytes to heal a scratch wound, increase the rate of single cell migration, increase ERK phosphorylation, enhance EF-mediated directional migration, preserve a pro-migratory cyto-architecture, maintain normal proliferation rates, and ultimately accelerate skin wound re-epithelialization. Finally, we demonstrate that keratinocytes express protein for two key enzymes in the catecholamine synthesis cascade localized to cytoplasmic granules and measure epinephrine in keratinocyte extracts. We believe that this is the first demonstration that ␤-AR antagonists can accelerate human skin wound re-epithelialization, and we hypothesize that the mechanism of action is via blockade of an endogenous autocrine ␤2-AR network that slows migration and delays wound healing (33,34).
ERK activation plays an important role in keratinocyte migration (35). Upon mechanical injury of confluent keratinocyte cultures (42) or Madin-Darby canine kidney cell cultures (43), ERK is activated by and is required for wound repair in confluent rat keratinocyte cultures (63) and in human epidermis (64). Previously, we have demonstrated that phospho-ERK is localized at the lamellipodial edge in migrating keratinocytes (33) and that ␤-AR agonists decrease ERK phosphorylation and prevent its localization to the lamellipodia via protein phosphatase 2A-dependent mechanisms (33,34). Although the function of phospho-ERK at the lamellipodial edge is unknown, it is possible that direct interactions between ERK and ␤ integrins (65) or paxillin (66) take place, suggesting an important role for ERK in integrating cell adhesion and receptor-mediated signaling in the control of cell migration. In stark contrast to the effects we observe with ␤-AR agonists, here we demonstrate that ␤-AR antagonists enhance scratch wound healing in confluent keratinocyte cultures (Fig. 1), increase the rate of keratinocyte single cell migration, and increase ERK phosphorylation (Fig. 2).
Because ERK plays such a pivotal role in cell migration (35,38,42,43,63), required for cutaneous wound repair (44), it is likely that the ␤-AR antagonist-mediated increase in ERK phosphorylation and keratinocyte migration could accelerate human skin wound re-epithelialization.
Wound currents have been measured exiting injured cornea and play a role in wound healing (16 -19) and limb regeneration in salamanders and newts (67)(68)(69). EF application initiates epithelial cell cathodal migration within minutes (13,15,47,48,70), and because an EF is generated immediately upon wounding, with the cathode at the wound center, it could be the earliest signal that epithelial cells receive to initiate directional migration into the dermal wound bed (46,71). Previously, we have demonstrated that cAMP-dependent protein kinase (31) and the ␤2-AR-mediated increase in intracellular cAMP (29) can modulate keratinocyte galvanotaxis. Here we demonstrate that a ␤-AR antagonist increases the ability of keratinocytes to sense and respond to the EF by exhibiting enhanced directional migration toward the cathode, an additional indication that ␤-AR antagonists could enhance wound healing.
Efficient cell migration, required for wound repair, is dependent on the temporally and spatially controlled reorganization of the actin cytoskeleton (50). Within hours of injury, FIGURE 5. Chronic human skin wounds retain the capacity to heal in the presence of 10% serum. Wounds 3 mm in diameter were generated in excised human skin, cultured in medium containing 5% serum. After 4 days of culture in medium containing 5% serum, biopsies were cultured for a further 4 days in the presence or absence of an additional 5% serum. Biopsies were fixed at day 8 and stained as described. Images of biopsies fixed and stained after 8 days in culture in 5% serum or 4 days in culture in 5% serum followed by 4 days in culture in 10% serum are presented in A and B at 10ϫ magnification. The arrows mark the edges of the wound, and the bars represent new epithelium. The percentage of re-epithelialization was calculated for each wound, and the data were analyzed using the Student's t test and are represented graphically (days 5-8 in 5% serum (circles with solid lines), days 5-8 in 10% serum (ϩ with dashed lines); C, *, p Ͻ 0.01). The data are combined from four independent experiments performed in triplicate on excised skin from four different individuals.
keratinocytes undergo phenotypic alterations including the formation of a fine and diffuse actin network at the advancing lammellipodium to allow for cell migration (73,74). Integrin receptors within focal adhesions stabilize the lamellipodia (75), allowing the migrating keratinocytes to interact with the variety of extracellular matrices found in the wound site, including fibronectin, vitronectin, stromal type I collagen, and fibrin (76). ␤-AR antagonists preserve the pro-migratory phenotype of migrating keratinocytes; cells are polarized and crescentshaped with a broad lamellipodium, characteristic of the migra-tory phenotype (49). Keratinocyte proliferation behind the epithelial tongue is essential for efficient human skin re-epithelialization (10) and ␤-AR antagonists also preserve normal cell proliferation in vitro.
Perhaps the strongest evidence for the role of ␤2-AR in wound repair so far is the direct demonstration that blockade of ␤2-AR receptors in excised, wounded human skin under chronic conditions (Fig. 6) significantly accelerates skin re-epithelialization. Using tissue confers the advantages of a normal extracellular matrix and the three-dimensional geometry of the FIGURE 6. ␤-AR antagonists enhance skin wound re-epithelialization. Wounds 3 mm in diameter were generated in excised human skin, cultured in medium containing 5% serum in the presence or absence of ␤-AR antagonist (10 M), fixed, and stained every day as described. Re-epithelialization was determined using light microscopy. Specimens that were damaged in the histologic process or otherwise noninterpretable were excluded from the study. Images of untreated (control) and ␤-AR antagonist-treated wounds, fixed on days 3 -5, are presented in A at 10ϫ magnification. The arrows mark the edges of the wound, and the bars represent new epithelium. The percentage of re-epithelialization was calculated for each wound, and the data were analyzed using the Student's t test and represented graphically (control, circles with solid lines; ␤-AR antagonist, squares with dashed lines; B, *, p Ͻ 0.01). The data are combined from four independent experiments performed in triplicate on excised skin from four different individuals.
healing wound not found in scratch assays or other assays using cultured cells.
Keratinocytes express a high level of ␤2-ARs (3, 4) and enzyme activity for TH and PNMT, and mRNA for TH has been discovered in undifferentiated keratinocytes (6,58,59). Inter-pretation of the enhanced keratinocyte migration and skin reepithelialization we observe in the presence of ␤-AR antagonists suggests that keratinocytes may synthesize and secrete catecholamines that are anti-motogenic and anti-mitogenic in keratinocytes (29,34) and delay skin wound re-epithelialization (33). Blockade of the ␤2-AR would negate the endogenous catecholamine negative effects on keratinocyte migration, resulting in enhanced motility and wound healing. To support our hypothesis, we lysed keratinocytes and immunoblotted with antibodies specific for two key catecholamine synthesis enzymes: tyrosine hydroxylase, an enzyme that coverts L-tyrosine to L-dopa, the rate-limiting step in catecholamine biosynthesis (55,56), and phenylethanolamine-N-methyl transferase, the enzyme that converts nor epinephrine to epinephrine (58). We demonstrate that three different strains of keratinocytes express both enzymes, confirming that mRNA for these enzymes is indeed transcribed to protein. In addition, we localized both catecholamine synthesis enzymes to granules or vesicles in the keratinocyte cytoplasm.
Finally, we measure endogenously synthesized epinephrine in keratinocyte extracts. We do observe variation in the level of catecholamine synthesis enzymes expressed and epinephrine measured in the different strains of keratinocytes, which warrants further investigation, especially because catecholamines appear to delay our ability to heal wounds. This provides convincing support for our hypothesis that blockade of ␤2-AR by antagonist prevents endogenously synthesized catecholamine from binding, negating its anti-motogenic effects and consequently accelerating wound repair.
␤-AR antagonists are widely used drugs for the treatment of cardiologic disease. Nearly 50 million Americans with hypertension are treated daily with ␤-AR antagonists (77). They are also the most frequently prescribed class of drug for the treatment of glaucoma, a disease estimated to affect 1.25% of the population over 40 years of age and the leading cause of irreversible blindness in the world (78,79). Elevated intraocular pressure is a major risk factor in glaucoma (80), and ␤-AR antagonists lower intraocular pressure, therefore minimizing damage to the optic nerve (81)(82)(83)(84). Even though ␤-AR antagonists are used prolifically, there have been no specific observations regarding the ability of patients using these agents to heal wounds. However, a number of reports, in addition to the work presented here, support the notion that exogenous ␤-AR antagonists alter wound healing. Denda et al. (24) have demonstrated that topical application of ␤-AR antagonists can accelerate the recovery of the barrier function of tape-stripped skin by measuring the transepidermal water loss. ␤-AR antagonists are also widely used in the post-burn wound recovery period, and a retrospective outcome analysis by Arbabi et al. (85) has demonstrated a shorter time for burn wound healing in a cohort of patients that received ␤-AR antagonists during their hospital stay. Thus, there are suggestions in the literature that both the endogenous ␤2-AR signaling network and exogenously supplied ␤-AR antagonists may modulate wound repair in the clinical setting. More systematic analysis of wound healing in cohorts of patients treated with ␤-AR antagonists is warranted, and such a study is currently underway at our institution. Impaired wound healing is a growing clinical problem, most FIGURE 7. Keratinocytes express the enzymes necessary to convert L-tyrosine to epinephrine, localized in cytoplasmic granules/vesicles. The catecholamine biosynthesis cascade is represented in A. Keratinocytes were lysed and separated electrophoretically, and membranes were immunoblotted with antibodies specific for TH and PNMT as described. An immunoblot of protein from three separate strains is presented in B. The data are representative of three independent experiments with at least three different keratinocyte strains. Sterile coverslips were coated with collagen, and cells were plated as described. The cells were fixed, immunostained for TH or PNMT, and photographed as described. Pictures of three cells immunostained with either TH or PNMT are presented in C. The cells presented are representative of the majority of cells immunostained for TH or PNMT from three separate cell strains. The scale bar is 20 M. evident in the remarkable numbers of chronic wounds in our aging population: 6.5 million have chronic skin ulcers caused by pressure, venous stasis, or diabetes mellitus (86), costing the United States health care system a staggering $9 billion annually (72,(87)(88)(89)(90). Defining pathways that regulate the wound healing process provides the potential for developing new therapeutic approaches. The current finding, that ␤-AR antagonists significantly accelerate wound re-epithelialization, now brings mechanistic support for the regulatory role of the ␤2-adrenergic hormonal network in the wound repair process. Clearly, further investigation of this hormonal network in skin will improve our understanding of the wound healing process and hopefully lead to the development of therapies to enhance wound repair.