Adrenergic Signaling in Human Oral Keratinocytes and Wound Repair
P. Steenhuis, R.E. Huntley, [...], and R.R. Isseroff
Associated Data
Abstract
Catecholamines are present in saliva, but their influence on oral epithelium is not understood. Because psychological stress increases salivary catecholamines and impairs oral mucosal wound healing, we sought to determine if epithelial adrenergic signaling could link these two findings. We found that cultured human oral keratinocytes (HOK) express the α 2B - and β 2 -adrenergic receptors (ARs). Exposure of HOK to either epinephrine or the β-AR agonist, isoproterenol, reduced migratory speed and decreased in vitro scratch wound healing. Incubation with the β-AR antagonist timolol reversed the catecholamine-induced effects, indicating that the observed response is mediated by β-AR. Epinephrine treatment decreased phosphorylation of the mitogen-activated protein kinases (MAPK) ERK1/2 and p38; these decreases were also reversed with timolol. Cultured HOK express enzymes of the epinephrine synthetic pathway, and generate epinephrine. These findings demonstrate that stress-induced elevations of salivary catecholamines signal through MAPK pathways, and result in impaired oral keratinocyte migration required for healing.
Introduction
Adrenergic receptors (ARs) are G-protein-coupled transmembrane receptors for catecholamine ligands, including the classic “stress hormones” epinephrine and norepinephrine. These receptors are expressed widely, including within the epidermis and cornea ( Walkenbach et al ., 1984 ; Steinkraus et al ., 1996 ; Schallreuter, 1997 ), where recent work has uncovered a role for AR signaling in cell migration and wound repair ( Chen et al ., 2002 ; Pullar et al ., 2006a , 2007 ; Ghoghawala et al ., 2008 ; Sivamani et al ., 2009b ). A role for ARs in the control of salivary gland function is also known (reviewed in Proctor and Carpenter, 2007 ); however, the function of ARs within the oral epithelium is not clear.
An understanding of the physiology of ARs in the oral epithelium is warranted, because this epithelium is bathed in catecholamine-containing saliva ( Okumura et al ., 1997 ). Levels of salivary catecholamines, both norepinephrine and epinephrine, increase after psychological stress ( McClelland et al ., 1985 ; Mitome et al ., 1997 ; Okumura et al ., 1997 ). Various stressors can induce differential increases in either epinephrine, norepinephrine, or both ( Vogel and Jensh, 1988 ; Goldstein, 2003 ). Therefore, one might expect that adverse effects of stress, often mediated by elevations of the “stress hormones” epinephrine and norepinephrine, could manifest in the oral cavity. Indeed, stress-impaired healing of oral mucosal wounds has been reported ( Marucha et al ., 1998 ), and the impaired healing of oral aphthous ulcers ( Natah et al ., 2004 ) has also been associated with increased psychological stress ( Albanidou-Farmaki et al ., 2008 ). Yet, the relationship of stress, catecholamine mediators, and ARs in the oral epithelium is one that has not been explored, and that was the goal of the work we report here.
Activation of β 2 -AR is one mechanism by which stress decreases migration of human skin-derived keratinocytes or corneal epithelial in vitro , delays healing of a wound in a confluent sheet of cultured cells, and impairs healing in both explanted skin and cornea ( Chen et al ., 2002 ; Pullar et al ., 2006b , 2007 ; Ghoghawala et al ., 2008 ; Sivamani et al ., 2009b ). β 2 -AR blockade with receptor-specific antagonists reverses this effect and also, even in the absence of exogenously added agonists, accelerates wound healing in vitro ( Pullar et al ., 2006b ) and in vivo in wound models ( Ghoghawala et al ., 2008 ; Sivamani et al ., 2009b ). AR antagonists, notably to the β 2 -AR, have therefore been proposed as therapeutic agents to improve healing, and are currently used in the care of burn wound patients ( Pereira et al ., 2007 ; Zhang et al ., 2009 ).
We hypothesized that the “stress hormone” epinephrine would decrease oral epithelial cell migration by activation of the β 2 -AR, if this or other ARs were present in oral epithelial cells, and that this could lead to impaired oral wound healing.
Materials & Methods
Cell Isolation
HOK were isolated from oral gingiva attached to extracted third molars of healthy individuals and cultured as described ( Krisanaprakornkit et al ., 1998 ), with the modification of epinephrine being removed from the medium. At 70% confluence, cells were frozen or passaged, and 24 hrs after passaging, the medium was replaced with growth medium (KBM 0.15 mM calcium plus keratinocyte growth medium supplements; Cambrex, Walkersville, MD, USA) to facilitate differentiation.
Microarrays
For microarray studies, fourth-passage HOKs from three donors were tested in triplicate. Total RNA was isolated with TRIzol and purified with the RNeasy mini kit (Qiagen, Valencia, CA, USA) as previously described ( Yin and Dale, 2007 ). RNA quality was assayed then processed for microarrays at the Center for Expression Array at the University of Washington for expression analysis on HG-U133A GeneChips (Affymetrix, Santa Clara, CA, USA). Array data were analyzed by GeneChip operating software (GCOS, V1.4) and further analyzed as previously described ( Yin and Dale, 2007 ). Arrays were normalized with the median intensity over the entire array. Probe sets with mean expression value < 40 were similar to negative controls and considered to be under the limits of detection.
Cell Migration
We determined the migratory speed of cultured HOK using image analysis of time-lapse images, as previously described ( Sivamani et al ., 2009b ). Cells were exposed to either growth medium alone (control) or medium containing either the β-AR agonist isoproterenol (EMD Biosciences, La Jolla, CA, USA), the β-AR agonist epinephrine, the β-AR antagonist timolol (Sigma-Aldrich, St. Louis, MO, USA), or combinations of these agonists and the antagonist, and incubated at 37°C for another 30 min before imaging.
Immunocytochemistry
immunocytochemistry was performed as previously described ( Pullar et al ., 2006b ) on HOK incubated with growth medium alone (control), or growth medium containing agonists and/or the antagonist for 1 hr prior to fixation in 10% buffered formalin. Cells were permeabilized for 5 min with 0.1% Triton-X-100 in PBS, and non-specific binding was blocked for 1 hr with 10% goat serum in PBS. Primary and secondary antibody dilutions were made in 1% goat serum in PBS. Imaging was done on a Nikon Eclipse TE2000 inverted microscope with a 60x objective (NA 1.4) with volume deconvolution and OpenLab image analysis software.
Scratch Wound Assay
The rate of healing scratch wounds made in confluent HOK cultures was examined by previously reported techniques ( Pullar et al ., 2006b ). Cells were pre-treated for 1 hr with 10 µg/mL mitomycin C (EMD Biosciences) to inhibit cell proliferation that could confound the analysis of migration ( Hayashi et al ., 2007 ). Wounded cultures were incubated with growth medium alone (control) or growth medium containing agonists and/or the antagonist. We used ImageJ (NIH.gov) to measure the scratch wound area, and performed a two-sample, unequal variance, one-tailed Student’s t test to compare the average percentage healing.
Immunoblotting
HOK treated for 60 min with growth medium alone or growth medium containing agonists and/or the antagonist underwent lysis, and lysates were subjected to immunoblotting as previously described ( Pullar et al ., 2006b ). Electrophoresis was performed with 5 µg protein per lane (P-ERK, ERK, P-p38 MAPK, and p38 MAPK) or 40 µg protein per lane (α 2B -AR and β 2 -AR). Detection was as previously described ( Pullar et al ., 2006b ).
Real-time RT-PCR
Total RNA was isolated from cells with the use of the RNeasy ® Mini Kit (Qiagen), and reverse transcription was performed on 2 ng RNA (Applied Biosystems, Foster City, CA, USA). RT-PCR was performed in duplicate on 0.056 ng cDNA in a reaction volume of 10 µL with a 7500 Fast Real-Time PCR (Applied Biosystems). Expression of tyrosine hydroxylase (TH) and phenylethanolamine-N-methyl transferase (PNMT) was determined by TaqMan ® Gene Expression Assays (Applied Biosystems) and normalized to 18S rRNA.
Epinephrine Immunoassay
We used an epinephrine enzyme immunoassay (Rocky Mountain Diagnostics, Colorado Springs, CO, USA) to measure the amount of epinephrine produced by HOK. Epinephrine levels were calculated as picogram/milligram of total protein. Epinephrine levels were also measured in HOK-conditioned medium.
Results
α 2B -AR and β 2 -AR Were Expressed by HOK
Epinephrine can activate both α- and β-AR, so we first asked which AR subtypes are expressed by HOK. Both α 2B -AR and β 2 -AR were detected by gene expression analysis ( Fig. 1A ), with the β 2 -AR being the most highly expressed AR gene in HOK. Protein expression of both α 2B -AR and β 2 -AR was also confirmed in these cells ( Fig. 1B ).
β-AR Activation Reduced HOK Cell Migration and Wound Healing in vitro
HOK (control) migrated at 1.7 µm/min, and activation of the β-AR with 1 µM isoproterenol resulted in a 43% decrease in migratory speed ( Fig. 2A , n = 7, p ≤ 0.01); 1 µM epinephrine treatment resulted in a similar 34% decrease ( Fig. 2B , n = 3, p ≤ 0.01). Dose response testing demonstrated a plateau in the inhibition of migration at the 1-µM level (Appendix Fig. 1 ). Blocking the β-AR with 20 µM timolol completely reversed the agonist-induced effect ( Fig. 2A , n = 7, p ≤ 0.01; Fig. 2B , n = 3, p ≤ 0.01). Treatment with 20 µM timolol alone resulted in a 14% increase in migratory speed ( Figs. 2A , ,2B, 2B , n = 7, p ≤ 0.01).
Similarly, in the scratch wound assay, HOK treated with 1 µM epinephrine demonstrated decreased rates of healing when compared to controls ( Fig. 2C , n = 4, p ≤ 0.01 at 6 and 24 hrs after wounding, p ≤ 0.01 at 12 hrs; Appendix Fig. 2 ). Co-incubation with 20 µM timolol reversed the epinephrine-induced effect, and timolol alone resulted in more rapid healing compared with the control (n = 4, p ≤ 0.05 at 6 and 12 hrs after wounding). Since epinephrine (unlike isoproterenol, which solely activates β-AR) is also an agonist to the α 2B -AR, that are expressed in HOK, the α 2 -AR-specific antagonist yohimbine was added in some experiments. Yohimbine did not reverse the epinephrine effect, suggesting that the epinephrine-induced impairment in healing was mediated solely by the β-AR.
The decrease in migration speed in epinephrine-treated cells was accompanied by morphologic changes characteristic of stationary cells: the presence of many large vinculin-containing focal adhesions symmetrically distributed at the cell periphery ( Fig. 2D ). Control cells, cells treated with timolol/epinephrine, or cells treated with timolol alone all displayed fewer focal adhesions, and these were concentrated in the lamellipodia at the leading edges of migrating cells. Fine branching networks of actin fibers extending into the lamellipodia were evident in migratory cells (control, epinephrine/timolol, or timolol alone), whereas broad actin fibers arranged symmetrically in the periphery characterized the stationary, epinephrine-treated cells.
Epinephrine Activation of β-AR Resulted in Decreased Phosphorylation of ERK1/2 and P38 MAPK
Activation of both ERK 1/ 2 and P38 MAP kinases are required for cell migration (reviewed in Huang et al ., 2004 ; Fitsialos et al ., 2007 ). Treatment of HOK with 1 µM epinephrine resulted in a 39-64% decrease in ERK1/2 phosphorylation ( Fig. 3A , n = 3, p ≤ 0.01), reversed when the cells were pre-treated with 20 µM timolol, indicating that the decrease in ERK1/2 phosphorylation, like the decrease in migration, was mediated by epinephrine-induced β 2 -AR activation. Phosphorylation of the pro-migratory p38 MAP kinase was also decreased (38-42%) in HOK by epinephrine activation of the β-AR ( Fig. 3B , n = 3, p ≤ 0.01), and this decrease was reversed by co-incubation with the β-AR antagonist timolol. Treatment with 20 µM timolol alone did not affect ERK phosphorylation compared with the control.
HOK Produced and Secreted Epinephrine
Addition of the β-AR antagonist timolol to cultured HOK in the absence of exogenously added β-AR agonists increased their migratory speed and ability to heal a scratch wound ( Fig. 2 ). This prompted the question of what endogenous β-AR agonist generated by the cultured HOK is blocked by the antagonist. Transcripts for 2 enzymes in the epinephrine synthetic pathway—tyrosine hydroxylase (TH, converts L-tyrosine to L-DOPA) and phenylethanolamine-N-methyltransferase (PNMT, converts norepinephrine into epinephrine)—were detected in HOK ( Figs. 4A , ,4B) 4B ) by either microarray analysis or RT-PCR. Cultured HOK were also able to generate epinephrine detectable in cell lysates ( Fig. 4C ) and in the culture medium ( Fig. 4D ).
Discussion
While the role of adrenergic signaling has been investigated in skin and corneal epithelia ( Chen et al ., 2002 ; Pullar et al ., 2006b , 2007 ; Ghoghawala et al ., 2008 ; Sivamani et al ., 2009b ), surprisingly, this receptor-mediated signaling pathway has not been explored in oral epithelium. Here, we demonstrate for the first time the existence of an adrenergic receptor signaling pathway in oral epithelial keratinocytes, report on the consequences of receptor activation and blockade, and suggest a role for endogenous catecholamine ligands in the physiology of healing wounds within the oral cavity. The β 2 -AR and α 2b -AR were found to be expressed in cultured HOK, at both the message and protein levels, with the β 2 -AR being the predominantly expressed receptor. β-AR activation by epinephrine decreased HOK migration and their ability to heal a wound within a confluent sheet of cultured cells. That these effects could reversed by a β-AR antagonist, but not an α 2 -AR-antagonist, suggests that the migratory effects are mediated by the β-AR, and since the β 2 -AR is the only β-AR expressed, it is likely that the effects are mediated by the β 2 -AR. The catecholamine agonists also decrease the activity of the pro-migratory ERK and p38 MAPK, kinases required for epithelial cell migration ( McCawley et al ., 1999 ; Fitsialos et al ., 2007 ; Storesund et al ., 2008 ; Shi et al ., 2009 ; Sivamani et al ., 2009b ) and convert HOK from a migratory to a stationary phenotype. Additionally, the HOK themselves can generate catecholamine ligands. Thus, this work establishes a regulatory autocrine pathway in oral epithelial cells wherein all the required components—the receptors, ligands, and downstream mediators—are expressed within the cell.
The presence of the β 2 -AR was previously detected by immunohistology in intact human oral mucosal epithelium, and its expression is increased in intra-oral squamous cell carcinomas (SCC) ( Shang et al ., 2009 ). Our work focused on non-transformed HOK, grown as primary cell cultures as physiologic representatives of the oral epithelium. We identified both β 2 -AR and α 2b -AR protein in HOK, providing the rationale to characterize their function in these cells.
Epinephrine in the saliva of non-stressed humans is reported to range from 0.08 to 1.0 nM ( Okumura et al ., 1997 ), derived by transfer of epinephrine from the plasma and by synthesis and secretion by salivary sympathetic nerves ( Kennedy et al ., 2001 ). Increased levels of salivary catecholamines have been reported in psychologically stressed individuals, such as those undergoing dental procedures, public speaking, or taking academic examinations ( McClelland et al ., 1985 ; Mitome et al ., 1997 ; Okumura et al ., 1997 ). These earlier methods record levels only in the nM range; however, plasma levels of epinephrine can increase to the µM range with severe stress ( Pacak et al ., 1998 ), and salivary levels may likewise rise to this range. Both acute and chronic stress have been associated with delayed wound healing ( Kiecolt-Glaser et al ., 1995 ; Marucha et al ., 1998 ; Christian et al ., 2006 ; Robles et al ., 2009 ), and both acute and chronic stress are accompanied by increases in circulating epinephrine levels ( Schmidt and Kraft, 1996 ; Pike et al ., 1997 ). Additionally, we show here that HOK themselves are a source of epinephrine and that nanomolar levels of epinephrine are secreted into the culture medium. Thus, the oral epithelium has the capacity to respond to an autocrine or paracrine signaling network that may be activated in either acute or chronic stress situations. Such a network has been suggested in the skin, where epidermal keratinocytes can generate catecholamines that then up-regulate the expression of, and/or activate, α 1 -, β 1 -, and β 2 -AR on keratinocytes or other cell types in the skin ( Schallreuter, 1997 ; Sivamani et al ., 2009a ). It is reasonable to propose that stress increases the levels of circulating catecholamine agonists and, via a paracrine signaling mechanism, impairs healing. Additionally, autocrine signaling, via epithelial-cell-generated catecholamine ligands, may also impair healing in some circumstances. Autocrine signaling may play other, more homeostatic, roles in the epithelium, and further study will be required to discover this potential role.
Because catecholamine β 2 -AR agonists can impair HOK migration and wound healing in vitro , it is logical to propose that the stress-related elevations in salivary catecholamines contribute to the well-documented impairment of wound healing in both acutely and chronically stressed individuals ( Kiecolt-Glaser et al ., 1995 ; Marucha et al ., 1998 ). Likewise, aphthous stomatitis, characterized by slowly healing ulcers in the oral epithelium, has also been associated with psychological stress ( Albanidou-Farmaki et al ., 2008 ), and here too, increased salivary catecholamines could play a role in impairing healing. Additionally, since infection of the oral epithelium with bacterial pathogens increases expression of the β 2 -AR in HOK ( Handfield et al ., 2005 ), bacterial infection may possibly impair healing in associated oral wounds, partly by β 2 -AR-mediated impairment of HOK motility and diminution of their ability to re-epithelialize a wound. Other studies have demonstrated that β 2 -AR agonists impair both endothelial angiogenesis and fibroblast-mediated gel contraction, and that β 2 -AR antagonists reverse these impairments ( Pullar and Isseroff, 2005 ; Pullar and O’Leary, 2009 ). In each of these proposed clinical scenarios, the application of topical β 2 -AR antagonists might provide a novel therapeutic approach to improve healing, by acting as a classic antagonist to prevent binding of the increased levels of epinephrine, either present in saliva or HOK-generated. Continued investigations, including correlation of salivary catecholamines and oral β 2 -AR expression with chronic vs. acute stress, as well as clinical trials to determine the efficacy of adrenergic receptor-mediated treatment in improving healing in oral ulceration, will be needed to translate these findings into the clinical arena.