| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
School of Medicine, Division of Respiratory Cell and Molecular Biology, Southampton General Hospital, Southampton, SO16 6YD, U.K.
1Correspondence: Medical Specialties Level D, Centre Block (810), Southampton General Hospital, Tremona Rd., Southampton SO16 6YD, U.K. E-mail donnad{at}soton.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: phosphorylation • tyrphostin • remodeling • TGF-ß • c-erbB1
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
. This damage is associated with selective
loss of columnar epithelial cells, which appear to be shed from the
epithelial surface leading to detection of large clusters of cells
known as creola bodies in bronchial alveolar lavage fluid
(2)
. Evidence of epithelial shedding is also apparent in
bronchial biopsies from asthmatic subjects that show extensive loss of
columnar epithelial cells while the basal cells usually remain attached
to the basement membrane (4)
. Disruption of the bronchial
epithelium has important structural and functional consequences. It
leads to disruption of mucociliary clearance and loss of barrier
function, enabling tissue-damaging molecules to pass unimpeded from the
lumen into the airway wall. It also results in an altered epithelial
phenotype, with the epithelium becoming a significant source of
bioactive molecules (reviewed in refs 5
, 6
) that have the
capacity to maintain the ongoing inflammatory response, as well as
influence airway remodeling.
Airways remodeling is typically present in patients with asthma and
characteristically includes thickening of the subepithelial basement
membrane (SBM) collagen layer (7)
. Recent studies indicate
that this abnormality is apparent in young children who later develop
asthma up to 2 years before the onset of clinical symptoms
(8)
. SBM-collagen thickness has been shown to correlate
with disease severity, chronicity, and bronchial hyperresponsiveness
(BHR) (9
10
11)
. SBM thickening is due to the deposition of
interstitial collagens types I, III, and V and fibronectin in the
lamina reticularis (7)
accompanied by laminin
2 and ß2 chains
(12)
and tenascin (13)
. Collagen gene
expression originates from myofibroblasts whose numbers and activity
are increased in asthma (14)
and further enhanced by
allergen exposure (15)
. Along with eosinophils, in asthma
the epithelium is a major source of TGF-ß and other profibrogenic
growth factors (11
, 16)
whose levels correlate well with
SBM-collagen deposition, myofibroblast numbers (16)
, and
BHR (11)
. Higher levels of TGF-ß have been found in BAL
fluid in asthmatic compared with normal individuals, increasing
significantly after allergen challenge (17)
. Direct
evidence for a relationship between epithelial injury and enhanced
remodeling responses has come from in vitro studies using
cocultures of bronchial epithelial cells and myofibroblasts. In these
studies, polyarginine (as a surrogate for eosinophil basic proteins) or
mechanical damage to confluent monolayers of bronchial epithelial cells
grown on a collagen gel seeded with human myofibroblasts resulted in
enhanced proliferation and increased collagen gene expression due to
the combined effects of basic fibroblast growth factor (bFGF) (FGF-2),
insulin-like growth factor 1, platelet-derived growth factor-BB,
TGF-ß, and ET-1 (18)
.
In response to injury to the bronchial epithelium, there is an urgent
requirement to initiate tissue repair and to restore barrier function.
The immediate response involves migration of epithelial cells adjacent
to the area of damage into the wound to form a temporary squamous
barrier consisting of poorly differentiated and highly spread cells
often associated with inflammatory cells (19)
. This
interim repair is likely to provide polarity and some barrier function.
However, as these cells are unlikely to perform the normal
differentiated functions of the epithelium (seromucous secretion,
cilial motility, etc.), there follows a period of cell division and
redifferentiation leading to complete restoration of normal epithelial
barrier function.
Members of the epidermal growth factor (EGF) family [i.e., EGF,
transforming growth factor alpha (TGF-), heparin-binding EGF-like
growth factor (HB-EGF), amphiregulin, betacellulin, and epiregulin]
are likely to be important regulators of epithelial restitution by
virtue of their ability to stimulate cell migration, proliferation,
differentiation, and survival (20)
. Indeed, a direct role
for EGF, HB-EGF, and TGF- in cutaneous wound healing is already well
established (21
, 22)
and EGF enhances repair of sheep
tracheal epithelium after cotton smoke injury (23)
.
Furthermore, cell culture experiments using guinea pig tracheal
epithelial cells (24)
or rat type II alveolar cells
(25)
suggest that EGF and related factors assist in the
early chemotactic and migratory responses associated with wound
closure.
The EGF family of growth factors exert their biological effects by
binding to and activating the EGF receptor (c-erbB1, HER1, EGFR), a 170
kDa receptor tyrosine kinase (26)
. Although the presence
of EGF and the EGFR in human lung tissues has been demonstrated by
radioimmunoassay and immunohistochemistry, much of the work on the EGFR
has been performed in the context of cancer, where elevated EGFR
expression is a frequent observation (27)
. In normal adult
human lung, EGFR immunoreactivity has been found on the basal cells of
the bronchial epithelium and is limited to the intercellular lateral
cell membrane, whereas the basal surface attached to the basement
membrane is negative (28
, 29)
. Although elevated EGFR
expression has been reported in bronchial mucosa taken from elderly
subjects at autopsy from fatal asthma or from asthmatics undergoing
surgical resection for lung cancer (30)
, its expression
has not been examined in bronchial biopsies taken from young,
clinically characterized asthmatic subjects with no other respiratory
disease. We have therefore quantified bronchial epithelial expression
of immunoreactive EGFR in biopsies from subjects with mild or severe
asthma and control subjects without asthma and examined its
relationship with SBM thickness. To assess the relationship between
EGFR-mediated repair and tissue remodeling, we studied the involvement
of the EGFR in the repair responses of monolayers of bronchial
epithelial cells subjected to mechanical injury in vitro.
The effect of stimulating or inhibiting EGFR function on the release of
the profibrogenic mediator, TGF-ß2, during repair of the monolayer
was also measured.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Definition of disease severity was based on treatment and was
according to the GINA guidelines (i.e., mild asthmatics were stage 1 or
2 and severe were stage 3 or 4). The study was approved by the
Southampton Joint University and Hospitals Ethics Committee, and all
subjects gave their written consent after being fully informed about
the nature and purpose of the study.
|
Spirometry, bronchial reactivity, and atopic status
FEV1 and bronchial reactivity to inhaled
histamine were measured 1 to 4 wk before bronchoscopy in all subjects.
On this occasion, subjects with asthma were asked to withhold inhaled
short-acting ß2-agonists for at least 6 h. The baseline
FEV1 was recorded with a wedge bellows spirometer
(Vitalograph Ltd., Buckinghamshire, U.K.).
The assessment of bronchial reactivity has been described in detail
(31)
. In brief, subjects first inhaled five breaths of
normal saline solution from functional residual capacity to total lung
capacity from an Inspiron Mininebulizer (C. R. Bard International,
Sunderland, U.K.) through a mouthpiece, and further measurements of
FEV1 were made after 1 and 3 min. Subjects then
inhaled increasing doubling concentrations (0.03 to 16 mg/ml saline
solution) of histamine (BDH, Lutterworth, Leicestershire, U.K.) and
FEV1 measured at 1 and 3 min after each
inhalation. These stepwise inhalations were discontinued when
FEV1 had fallen by > 20% of the postsaline
value or when the maximum concentration had been reached. The bronchial
responses to the inhaled agonist were expressed as the provocation
concentration of agonist causing a 20% fall in
FEV1 (PC20). This was
derived by linear interpolation from the concentration-response curve
constructed on a logarithmic scale by plotting the percentage change in
FEV1 from the postsaline value against the
cumulative concentration of histamine administered.
To determine atopic status, skin prick testing was performed with the common aeroallergens Dermatophagoides pteronyssinus, Dermatophagoides farinae, mixed grass pollen, dog allergen, mixed tree pollen, and cat allergen (Miles Inc., Hollister Stier, Elkhart, Ind.). Development of a wheal >3 mm in diameter 15 min after exposure to one or more of these allergens in the presence of negative (normal saline solution) and positive (histamine acid phosphate) controls was considered a positive response.
Fiberoptic bronchoscopy and bronchial biopsy
Fiberoptic bronchoscopy and bronchial biopsy were performed as
described (32)
and in accordance with current NIH
guidelines (33)
. Subjects received premedication
consisting of nebulized salbutamol (5 mg) and ipratropium bromide (0.5
mg). Atropine (0.6 mg) was used intravenously as an antisecretory agent
and midazolam (4 to 10 mg) was administered to achieve mild sedation.
Lignocaine (10% and 4% spray) was applied to the upper airways, and
an Olympus BF IT20 fiberoptic bronchoscope (Olympus Company, Tokyo,
Japan) was inserted through the nose or mouth. Additional aliquots of
lignocaine (1%) were applied to the larynx and subsequently to the
lower airways. Oxygen (28%) was administered via nasal cannulas
throughout the procedure, and oxygen saturation was monitored
continuously with a digital pulse oximeter (Pulsox-7, Minolta, Tokyo,
Japan). Bronchial biopsy specimens were obtained from second- or
third-generation airway carinas by using FB15C biopsy forceps (Olympus
Company). Close clinical supervision was maintained throughout the
procedure, and subjects were kept under observation for a minimum of
4 h afterward.
Sample processing and immunohistochemistry
Biopsy specimens from each subject were placed immediately into
ice-cooled acetone solution containing the protease inhibitors
iodoacetamide (20 mM) and phenylmethylsulfonyl fluoride (2 µM). After
overnight fixation at -20°C, specimens were embedded in the
water-soluble resin glycolmethacrylate (GMA) (Park Scientific,
Northampton, U.K.). Processing of the tissue into GMA and the
immunohistochemical method applied to this material have been described
in detail (34)
. Briefly, 2 µm sections were cut with an
ultramicrotome (OM U3, Reichert, Vienna, Austria), floated onto 0.2%
(v/v) ammonia solution for 1 min, transferred to poly-L-lysine-coated
slides, and air dried at room temperature for 1 to 4 h. Endogenous
peroxidase activity was inhibited by incubating sections in 0.1%
sodium azide and 0.3% hydrogen peroxide in Tris-buffered saline (TBS)
pH 7.6 for 30 min. After washing three times with TBS, blocking medium
consisting of Dulbecco’s modified Eagles medium (DMEM) containing 10%
(v/v) fetal bovine serum (FBS) and 1% (w/v) bovine serum albumin, was
applied for a further 30 min. The sections were then incubated with a
sheep anti-EGFR polyclonal antibody at 25 µg/ml (an IgG fraction
obtained from immune serum raised against EGFRs that had been purified
from A431 cell plasma membranes by EGF-affinity chromatography) for
1 h at room temperature (29)
. After washing,
anti-goat biotinylated IgG Fab fragment at 1:100 (Dako Ltd, Wycombe,
U.K.) was applied to the sections for 1 h. This was followed by a
1 h incubation with streptavidin-biotin horseradish peroxidase
complex at 1:200 (Dako Ltd.) before visualization with 0.5 mg/ml
diaminobenzidine in Tris/HCl buffer containing 0.01%
H2O2. Sections were
counterstained with Mayer’s hematoxylin and mounted in DPX.
Quantitative analysis
Airway epithelial expression of immunoreactive EGFR in GMA
sections was quantified by computer-assisted image analysis
(Colorvision 1.7.6, Improvision, Coventry, U.K.). For each biopsy
specimen, the entire intact epithelium in two nonserial sections was
systematically assessed based on red, blue, green (RGB) color balance.
At the beginning of each session, the image analysis system was
standardized using the same section of bronchial mucosa stained for the
EGFR to ensure reproducibility of analysis. The digitized image of the
standard section was used to interactively sample an example of the
positive staining and the system was then allowed to select all the
pixels of the same RGB color balance (i.e., positive staining) within
the image. The area of the epithelium was then delineated interactively
and the percentage of positive staining within the epithelium was
determined; the color balance and percent staining value was recorded
for future sessions. At the beginning of each subsequent session, the
image analyser was calibrated using this section and adjusted to
within ± 5% of the original pixel reading. Once the system had
been set up using the‘standard’slide, the test sections were
analyzed, using the same parameters. Measurements of EGFR expression
were performed by an observer who was unaware of the clinical group
from which the biopsy specimen was derived. Each tissue section was
analyzed on two separate occasions by the same observer; the
intraobserver coefficient of variation ranged from 6% to 16%.
Quantitation of SBM thickening
The thickness of SBM was measured from the base of the bronchial
epithelium to the outer limit of the reticular lamina of the basement
membrane at regular intervals of 200 µm along the length of the
section with a digital micrometer. The final figure was the mean of all
the measurements obtained for each biopsy specimen.
Statistical analyses
Data for age and percent predicted FEV1 were expressed as
means ± SE; data for PC20 were
expressed as geometric means (ranges); data for EGFR immunoreactivity
and SBM thickening were expressed as medians (ranges). Comparisons
between clinical groups were made by using one-way ANOVA with post hoc
testing, with least-squared difference. Correlations were sought by
using Pearson’s and Spearman’s tests for normally and abnormally
distributed data, respectively. P<0.05 was regarded as
statistically significant.
Scrape wounding of 16HBE 14o- monolayers
16HBE 14o- cells (a gift from Professor Gruenert; ref
35
) were grown in 10% FBS/MEM containing 50 IU/ml
penicillin, 50 µg/ml streptomycin, and 2 mM L-glutamine in 4-chamber
glass slides (Life Technologies). At 90% confluence the cells were
placed in Ultraculture serum-free medium (SFM) (BioWhittaker,
Walkersville, Md.) for 24 h, and the cell monolayer was washed in
fresh SFM before scrape wounding. A series of 12 parallel wounds was
made in each monolayer using a sterile, comb and any damaged cells
washed away before the wounded cells were allowed to repair in fresh
SFM with the additions as indicated in Results. Slides were fixed in
ice-cold methanol at 0, 3, 6, and 9 h and EGFRs or TGF-ß were
detected by immunoperoxidase staining using the polyclonal sheep
anti-EGFR antibody described above or mouse anti-TGF-ß (R&D Systems,
Oxford, U.K.), with diaminobenzidine as a chromagen and hematoxylin as
counterstain. The slides were scanned using a Hewlett Packard Desk Scan
at a resolution of 720 dpi; the wound areas (mm2)
were quantitated using Scion Image analysis software and analyzed using
Student’s unpaired t test.
Immunoprecipitation and Western blotting
16HBE 14o- cells were grown to 90% confluence in 57
cm2 petri dishes and serum starved in SFM for
24 h prior to wounding. The cells were scrape wounded using a 4 cm
sterile comb. Semicircular wounds were made across each side of the
monolayer and then the procedure was repeated serially after turning
the tray through 45° until rotation through 180° had been achieved.
The monolayer was then washed to remove cell debris and the medium was
replaced with fresh SFM medium in the absence or presence of EGF (1.7
nM) for 0, 1, 5, 15, 30 min, 1 h, and 4 h. Unwounded
monolayers were grown and processed as above, with the exception that
cells were not scrape wounded. At the end of each incubation, the cells
were washed in phosphate-buffered saline containing protease and
phosphatase inhibitors (1 mM sodium orthovanadate, 50 mM sodium
fluoride, 1 mM PMSF, 1 µM leupeptin, and 1.54 µM aprotinin) and
lysed into 0.5 ml boiling lysis buffer [1% sodium dodecyl sulfate
(SDS), 10 mM Tris-HCl pH 7.4, 5 mM EDTA, 5 mM EGTA containing protease
and phosphatase inhibitors, as above]. Equal amounts of lysates (900
µg of protein) were diluted 1:1 with 2x Triton buffer (20 mM
Tris-HCl pH 7.4, containing 2% Triton X-100, 2% Nonidet P-40, 300 mM
NaCl, 10 mM EDTA, 10 mM EGTA, protease, and phosphatase inhibitors) and
incubated with 30 µl of PY-agarose conjugate (Transduction Labs,
Lexington, Ky.) for 1.5 h at 4°C. Phosphotyrosine containing
immune complexes were then pelleted, washed three times with 1x Triton
buffer, and solubilized into 50 µl of sample buffer (62.5 mM Tris-HCl
pH 6.8, 10% glycerol, 5% ß-mercaptoethanol, 2% SDS, and 0.02%
bromphenol blue, 5 mM EDTA, 5 mM EGTA containing protease and
phosphatase inhibitors) at 95°C for 5 min. Immunoprecipitates or cell
lysates were analyzed using SDS-polyacrylamide gel electrophoresis and
Western blotting with a polyclonal sheep anti-EGFR antibody or
antiphosphotyrosine (PY20, Transduction Labs) with enhanced
chemiluminescence detection following the manufacturers instructions.
The levels of immunoreactive EGFR were then quantitated by densitometry
using Kodak 1D software.
Measurement of TGF-ß2 release from repairing bronchial
epithelial monolayers
16HBE 14o- cells were grown to 90% confluence in 57
cm2 petri dishes, serum starved, and scrape
wounded as described for the Western blot experiments. After washing,
the monolayers were washed to remove cell debris and the medium was
replaced with fresh SFM containing EGF and/or tyrphostin, as stated in
Results. The cells were incubated for 24, 48, and 72 h before the
medium was harvested, centrifuged to remove cell debris, and frozen
prior to analysis. Cell medium was acidified with 1M HCl to enable
measurement of total TGF-ß2 activity in the conditioned media. After
neutralization, samples were analyzed using a TGF-ß2
Emax ELISA kit (Promega, Madison, Wis.),
according to manufacturer’s instructions. Data were analyzed using
Student’s unpaired t test.
Mitogenesis assay
The ability of EGF or acidic FGF to induce DNA synthesis in
confluent and quiescent H292 bronchial epithelial cells or NR6/HER
fibroblasts was measured in a modification of a standard mitogenesis
assay (36)
. In brief, cells were grown to confluence in
96-well opaque cell culture trays in RPMI/10% FBS and rendered
quiescent by serum reduction. Growth factors and/or tyrphostin AG1478
(stored as a 20 mM stock dissolved in DMSO) were added to the cells in
mitogenesis assay buffer and DNA synthesis was determined 18 h
later by incorporation of [125I]UdR over a
2 h pulse period. The cells were fixed and washed with 5%
trichloroacetic acid, followed by methanol. After drying,
acid-insoluble material was dissolved in 40 µl/well of 0.2M NaOH and
radioactivity was determined on a Topcount Scintillation counter
(Canberra Packard, Pangbourne, Berks, Australia) after addition of 150
µl of Microscint-40 (Canberra Packard) to each well.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. As we have recently documented (29)
, normal
bronchial epithelium was found to exhibit a uniform pattern of positive
immunoreactivity for the EGFR, with the majority of the staining being
confined to the basal cells and their junctions with columnar
epithelial cells; no staining was evident on the luminal surface of the
epithelium (Fig. 1A
). In asthmatic biopsies, EGFR immunostaining was visibly
increased, particularly in the samples from severe asthmatics, and
occurred uniformly throughout the epithelial layer (Fig. 1B
); image analysis of the tissue sections confirmed that
the percentage of epithelial EGFR immunostaining in asthmatic
epithelium increased in relation to disease severity (Fig. 2
). As EGFR expression remained elevated in patients receiving
corticosteroids, we further evaluated five subjects before and after 12
wk treatment with inhaled budesonide (800 µg per day) and found no
significant difference in bronchial epithelial EGFR expression [median
value (range) = 15.6 (9.8–22.3) before treatment vs. 16.9
(10.4–18.1) after treatment].
|
|
As previously reported (7)
, the subepithelial
basement membrane was significantly thicker [median (range) in the
biopsies taken from either severe (11.4, range 7.1–18.5 µm) or
mildly asthmatic subjects (10.2, range 4.7–17.2 µm] when compared
with the healthy controls [5.7 (3.5–11.6) µm) (P<0.01
asthma vs. normal)]. A positive correlation (r=0.62
P<0.001) was observed when EGFR expression was analyzed in
relation to SBM thickening using Spearman’s test.
Examination of the biopsies further revealed that a significant
proportion of the bronchial epithelium from the asthmatic subjects was
damaged and lacked columnar epithelial cells, as found previously
(4)
. In these areas of epithelium, membrane staining for
the EGFR was particularly intense and was evident on all surfaces
including the apical surface (Fig. 1D, F
), which is
reported to be bathed in secretory fluid containing EGF-like growth
factors (37)
. In contrast, epithelial immunostaining in
areas of biopsy-induced damage from normal subjects was not affected
(Fig. 1C
), suggesting that the intense staining seen in the
damaged asthmatic epithelium was unlikely to be an artifact due to
altered antibody accessibility.
EGF promotes bronchial epithelial repair in vitro
To determine whether the EGFR is involved in human bronchial
epithelial repair, the ability of EGF to stimulate wound closure was
investigated in vitro. A series of single, parallel scrape
wounds were made in confluent monolayers of 16HBE 14o- bronchial
epithelial cells so that well-defined wound margins were delineated
(Fig. 3A
), enabling wound closure to be measured simply by image
analysis (Fig. 3B
). Under these conditions, wound closure
occurred in serum-free medium lacking exogenous growth factors, with
closure almost complete by 9 h after wounding. Addition of 1.7 nM,
EGF was found to enhance the rate of repair, with closure occurring
6 h after wounding (Fig. 3B
), and comparable results
were obtained when the EGF concentration was reduced to 420 pM (data
not shown). In contrast, another epithelial mitogen, keratinocyte
growth factor (KGF) (Fig. 3B
, inset), failed to mimic the
effect of EGF on wound closure, but it was not determined whether 16HBE
14o- cells express KGF receptors (KGFR, FGFR-2iiib splice variant). The
synthetic corticosteroid dexamethasone (1 µM) was without effect on
basal or EGF-stimulated wound closure (Fig. 3C
); similarly,
no significant effect was observed when dexamethasone was tested at
0.1, 1.0, or 10.0 µM (wound area at 6 h, expressed as percent of
untreated control = 127±11%, 99±10%, and 137±21%,
respectively), but significant inhibition was observed at 100 µM
(165±24%; P<0.005). The latter result probably reflects a
nonspecific effect of dexamethasone, given the very high level of drug
required to elicit any effect on wound closure.
|
The involvement of the EGFR in epithelial repair was further examined
using the tyrosine kinase inhibitor tyrphostin AG1478
(38)
. As 16HBE 14o- cells failed to show significant
mitogenic responses to growth factors unrelated to EGF, the specificity
of tyrphostin AG1478 was tested using H292 bronchial epithelial cells
or NR6/HER fibroblasts. In mitogenesis assays, the drug blocked
EGF-induced DNA synthesis in either cell line but was without effect on
that induced by aFGF, KGF, or bFGF (Fig. 4
).
|
As expected, AG1478 blocked the stimulatory effect of EGF on
wound closure in 16HBE 14o- cells; however, it also reduced the rate of
basal wound repair (Fig. 3
and Fig. 5A
). As our control experiments indicated that AG1478 was a
selective inhibitor of EGFR signal transduction (Fig. 4)
, blockade of
basal repair suggested that activation of the EGFR was an intrinsic
component of the repair process and implied release of an autocrine
ligand. Although we examined the effect of a neutralizing anti-EGFR
antibody on wound closure, results were inconclusive as the antibody
appeared to activate the EGFRs by direct cross-linking (data not
shown). Instead, as HB-EGF has been implicated in wound repair
processes in other systems (22)
, we determined the
effect of a neutralizing anti-HB-EGF antibody or the diphtheria
toxin analog CRM197 that binds to and neutralizes HB-EGF activity
(39)
. However, both reagents failed to significantly
affect basal wound repair, even though they blocked the ability of
exogenous HB-EGF to promote wound closure similarly to EGF (Fig. 5B
).
|
Scrape wounding stimulates phosphorylation of the EGFR
To confirm that activation of the EGFR occurred in response to
wounding, we examined the pattern of tyrosine phosphorylation in cell
lysates. To facilitate detection of phosphorylated proteins, the cell
monolayers were subjected to multiple scrape wounds so that a much
larger proportion of the total cell population had experienced the
physical trauma. As shown in Fig. 6A
, there was a rapid increase in tyrosine phosphorylated
proteins in the cells after wounding and the pattern of phosphorylation
was similar whether or not the cells were damaged in the presence of
EGF. Of particular note was the 170 kDa band whose phosphorylation
increased after wounding; this band was identified as the EGFR by
Western blotting with an anti-EGFR antibody (Fig. 6B
).
Consistent with the selective nature of AG1478, Fig. 6C
shows that phosphorylation of the EGFR was blocked by the tyrphostin.
|
The effect of scrape wounding on EGFR expression
We also examined the effect of mechanical damage on EGFR levels in
scrape-wounded monolayers of 16HBE 14o- cells. Immunocytochemical
staining of the wounded monolayer showed that EGFR immunoreactivity was
locally increased at the edges of the damaged monolayers (Fig. 7B
vs. panel A), consistent with that observed in
bronchial biopsies. This increase was still evident 6 and 9 h
after damage and was unaffected by the presence of dexamethasone in the
medium during repair of the monolayers (data not shown). The staining
observed did not appear to be artifactual (e.g., due to folding of the
monolayer or to altered cell number), as staining for the structurally
related c-erbB2 and c-erbB3 receptors was not altered
(40)
.
|
The effect of disruption of EGFR-mediated epithelial repair on
profibrogenic growth factor production
As damage causes bronchial epithelial cells to release
profibrogenic growth factors (18)
, we examined the effect
of wounding on TGF-ß expression in 16HBE 14o- cells. Using
scrape-wounded cells as described above, immunocytochemical staining
for TGF-ß was found throughout the monolayer. However, when the
monolayer was damaged and allowed to repair in either the absence or
presence of EGF, immunostaining was locally increased around the wound
edge (Fig. 7
, panels D–F vs. panel C; data not
shown) in a pattern similar to that observed for the EGFR. To directly
quantify the effect of wounding, EGF, and tyrphostin AG1478 on TGF-ß2
production, we used extensively wounded monolayers of 16HBE 14o- cells,
as in the phosphorylation experiments. In this case, repair of the
monolayers took more than 72 h, even when SFM was supplemented
with EGF. Under these conditions, epithelial damage increased TGF-ß2
production (Fig. 8A
), as reported previously (18)
. Whereas EGF
failed to significantly affect TGF-ß2 release in either wounded or
unwounded cells (Fig. 8B
), the presence of AG1478 markedly
augmented TGF-ß2 release only from the repairing epithelial cells
(Fig. 8A, B
). This stimulatory effect of tyrphostin could
not be attributed to any nonspecific damaging effect as AG1478 had no
effect on TGF-ß2 production by control nonwounded cultures.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. It is known to be involved
in epithelial repair in skin (41)
, the gastrointestinal
tract (42)
, and animal models of acute lung injury
(23
, 43)
. In the present in vitro studies, we
have established a role for EGF and HB-EGF in promoting repair of human
bronchial epithelial cells consistent with previous studies using
guinea pig tracheal epithelial cells (24)
. Using the
in vitro model, we demonstrated that injury resulted in a
localized increase in EGFR immunostaining at the wound edge in a manner
similar to that observed in areas of damage in normal bronchial
epithelium (44)
. We also showed that activation of the
EGFR occurred rapidly after damage and was independent of exogenous
ligand. This suggested that EGFR signaling was an intrinsic component
of the response to injury. Furthermore, the speed of the
phosphorylation response (Fig. 6)
suggested that induction of gene
expression was not required for EGFR activation and that it most likely
resulted from release of a membrane-associated ligand by a mechanism
involving regulated proteolysis (45)
. Although we
postulated that HB-EGF was the most likely candidate, we failed to
confirm our hypothesis using a neutralizing anti-HB-EGF antibody. As
16HBE 14o- cells express TGF- and amphiregulin in addition to HB-EGF
(29)
, it seems likely that one of these other ligands is
responsible for EGFR activation. However, at this stage we cannot
exclude the possibility of other ligand-independent mechanisms
(46)
that are known to lead to EGFR activation.
Studies of airways of patients dying of status asthmaticus have
described extensive epithelial damage with complete sloughing of large
sections of columnar epithelium into the lumen (47
, 48)
.
Even though epithelial shedding has also been observed in bronchial
biopsies and lavage fluid obtained fiberoptic bronchoscopy of patients
with mild asthma, the possibility of artifactual epithelial cell loss
during the collection and preparation of the biopsy specimens has to
date made unequivocal interpretation of these observations difficult.
Our observation that EGFR expression is markedly increased in areas of
epithelial damage in bronchial biopsies from asthmatic airways, but not
in similar areas of control subjects, provides for the first time
convincing evidence that epithelial injury and repair is an ongoing
process, even in mild asthma.
Unlike the bronchial epithelium in normal individuals, where elevated
EGFR expression is observed only in areas of structural damage
(44)
, in the present studies we found a striking
disease-related increase in EGFR expression in morphologically intact
asthmatic epithelium from mild-moderate asthmatics. Furthermore,
although we were only able to obtain biopsies from five severe
asthmatics, there was a further increase in immunostaining suggesting
that EGFR expression increases in proportion with disease severity.
Further contrasting with normal epithelium, EGFR immunostaining in
asthma was found to occur throughout the epithelial layer indicative of
widespread functional changes. This confirms a recent Japanese study
using asthmatic bronchial tissue postmortem or from surgical resections
(30)
. In gingival tissue, increased epithelial and stromal
cell EGFR immunoreactivity is caused by inflammation (49)
,
and a similar mechanism may operate in asthma. A possible mediator in
this effect is TNF-, which can induce EGFR expression in rat trachea
(50)
, neonatal skin explants (51)
, pancreatic
cells (52)
, and psoriatic skin (53)
. However,
in the present study it is noteworthy that anti-inflammatory
corticosteroid treatment failed to down-regulate EGFR expression in
intact bronchial epithelium, even though corticosteroids are reported
to enhance epithelial repair in vivo (54)
. As
steroids had no effect on epithelial EGFR expression or repair in
vitro, we postulate that their effects in vivo are
indirect and are due to their ability to reduce inflammation. Our
observations that EGFR expression is not modulated by corticosteroid
treatment is suggestive of a steroid-insensitive component
that may contribute to severe and chronic disease and is thus of
therapeutic significance. For example, one consequence of persistently
high levels of EGFR expression is that this may cause structurally
intact epithelium to be ‘locked’ into a repair phenotype in which it
sustains airways inflammation and remodeling by provision of
proinflammatory mediators and profibrogenic growth factors.
It has been reported that exposure to cigarette smoke (55)
or ozone (56)
results in increased EGFR in airways
epithelium. Thus, high EGFR expression in asthma may be a ‘reporter’
of damage indicating that the extent of epithelial injury is more
widespread than previously appreciated. In a normal wound healing
response, as well as in several hyperproliferative diseases including
carcinomas (27)
and psoriasis (53)
, the usual
outcome of increased EGFR expression is increased proliferation.
Indeed, it has recently been reported that induction of EGFR expression
in rat trachea results in goblet cell hyperplasia (50)
,
which is observed in both asthma and COPD. However, while there is
evidence of epithelial shedding in asthma, the level of basal cell
expression of proliferating cell nuclear antigen expression does not
increase, even though this proliferation marker is elevated in other
inflammatory diseases such as chronic bronchitis (57)
.
Failure to mount an adequate proliferative response after injury may be
a key feature of the asthmatic bronchial epithelium, explaining the
extent of epithelial damage as compared with other inflammatory
diseases such as COPD.
The lack of correlation between EGFR expression and proliferative
activity, together with the observation that EGFR overexpression
correlated with SBM thickening, led us to consider the effect of injury
on the production of profibrogenic growth factors. As shown in Fig. 7
,
immunoreactive TGF-ß was detectable in bronchial epithelial cell
monolayers and, after injury, this was increased at the wound edge in a
pattern similar to that seen for the EGFR. This suggests that the cells
that responded to the physical trauma were those local to the area of
damage, although the possibility of a more widespread or graded
response passing back from the wound edge could not be completely
excluded. When we then quantified TGF-ß2 release from wounded
monolayers, highest levels were found in cultures in which epithelial
repair was impaired with tyrphostin AG1478, i.e., TGF-ß2 production
was independent of EGFR activation. In view of the immunocytochemical
localization of both TGF-ß and EGFR in cells at the wound edge, we
propose that there are parallel pathways operating within these
repairing epithelial cells. Some of these pathways direct efficient
repair and are regulated by the EGFR, whereas others control
profibrogenic growth factor production (and possibly proinflammatory
cytokine production) and are independent of the EGFR. Given that EGFRs
are capable of regulating a number of different stages of epithelial
repair (survival, migration, proliferation, and differentiation), any
inhibitory mechanisms (e.g., ligand damage caused by proteolysis or
high levels of antiproliferative growth factors such as TGF-ß) that
act on these EGFR-mediated processes may cooperate to prolong
epithelial repair. As a consequence, the duration of activation of
parallel, EGFR-independent pathways that are associated with remodeling
responses will increase (Fig. 9
). These observations provide evidence for a reciprocal relationship
between the rate of epithelial repair and the production of
profibrogenic growth factors and suggest functional mechanisms that may
underlie the remodeling responses that are characteristic of chronic
asthma.
|
Epithelial–mesenchymal interactions are known to play important roles
during lung development, repair, and inflammation (58)
.
These interactions appear to be locally regulated by a resident layer
of myofibroblasts and fibroblasts in close proximity to the epithelium
and the term epithelial-mesenchymal trophic unit has been used to
describe their close anatomical and functional relationship
(59)
. Overall, the observations reported in the present
study suggest that epithelial damage causes a functional disturbance of
the epithelial-mesenchymal trophic unit and that this may underlie the
remodeling responses characteristic of chronic asthma.
|
ACKNOWLEDGMENTS |
|---|
Received for publication June 23, 1999.
Revision received February 9, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, R., Wilson, J., Holgate, S. T. (1990) Myofibroblasts and subepithelial fibrosis in bronchial asthma. Am. J. Resp. Cell Mol. Biol. 3,507-511
, R., Wilson, J. W., Lai, C. K. W., Holgate, S. T., Howarth, P. H. (1991) The safety aspects of fibreoptic bronchoscopy, bronchoalveolar lavage, and endobronchial biopsy in asthma. Am. Rev. Resp. Dis. 143,772-777[Medline]
and epidermal growth factor receptor is increased following bleomycin-induced lung injury in rats. Am. J. Resp. Cell Mol. Biol. 11,540-551[Abstract]
induces the expression of transforming growth factor and the epidermal growth factor receptor in human pancreatic cancer cells. Proc. Natl. Acad. Sci. USA 90,863-867This article has been cited by other articles:
|
|
 |
|
  H. H. Kariyawasam, G. Xanthou, J. Barkans, M. Aizen, A. B. Kay, and D. S. Robinson Basal Expression of Bone Morphogenetic Protein Receptor Is Reduced in Mild Asthma Am. J. Respir. Crit. Care Med., May 15, 2008; 177(10): 1074 - 1081. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Liu, R. C. Gualano, M. L. Hibbs, G. P. Anderson, and S. Bozinovski Epidermal Growth Factor Receptor Signaling to Erk1/2 and STATs Control the Intensity of the Epithelial Inflammatory Responses to Rhinovirus Infection J. Biol. Chem., April 11, 2008; 283(15): 9977 - 9985. [Abstract] [Full Text] [PDF]
|
|
) in the presence of a range of AG1478 concentrations. In the
absence of AG1478, the mean incorporation of 125IUdR was
4 x 103 cpm. Induction of DNA synthesis was measured
by incorporation of the thymidine analog 125IUdR as
described in Materials and Methods. Data are mean ±
SD, n = 3. B) The effect
of 1 µM tyrphostin (black bars) on induction of DNA synthesis in H292
or NR6/HER cells by the mitogens (gray bars) as indicated. Data are
mean of duplicate determinations and are representative of three
independent experiments.
, •) or EGF + 10 µM AG1478 (
,
). Data are mean of
duplicate observations and are representative of three independent
experiments. B) Release of TGF-ß2 by unwounded (open
bars) or wounded cells (hatched bars) cultured for 72 h in
serum-free medium (SFM) alone or with EGF (1.7 nM), 1 µM AG1478 (T1),
or 10 µM AG1478 (T10) without or with EGF (ET1, ET10). Data are
mean ± SD, n = 8. Significance
values were determined using Student’s t test for
unpaired data for treatment conditions vs. control unwounded cultures
(SFM), **P<0.001 and *P<0.05.