Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Dystroglycan (DG) is part of a complex that anchors the sarcolemna of skeletal muscle to the extracellular matrix protein laminin (1, 2). Disruption of this dystrophin glycoprotein complex (DGC) underlies several forms of muscular dystrophy (3). The complex consists of several glycoproteins within the sarcolemna, including -DG and -DG (4). Both DGs are encoded by a single gene and are cleaved by post-translational processing (7, 8). -DG is a 156-kD glycosylated extracellular protein on a 50- to 65-kD peptide backbone that noncovalently binds -DG at the C-terminus and binds to the E3-like G domains of both laminin-2 (3, 6, 9) and laminin-1 (7, 10) at the N-terminus of the  chain. The glycosylation of -DG varies widely in different species (11) and in different tissues within the same species (12); these differences lead to dramatic differences in how -DG binds to laminin-2 (7, 12). By its recognition of specific glycosyl residues on laminin, -DG functions as a lectin, and this function can be blocked by competition with glycosaminoglycan sugars, such as dextran sulfate and heparin (13, 14), or by lectins competing for the same binding site, such as wheat-germ agglutinin (WGA) (15). -DG is a 43-kD glycosylated transmembrane protein that binds to the C-terminal domain of dystrophin (16, 17). The DGC may colocalize with focal adhesions and integrins in some tissues (18, 19).

Recent studies demonstrate that the DGC can be found outside of skeletal muscle, in peripheral nerves (20, 21), brain (22), and blood vessels (13, 23). -DG is localized to the basal side of developing rat kidney epithelial cells and is expressed early in kidney morphogenesis (24). Monoclonal antibodies (mAbs) that blocked -DG binding to laminin-1 perturbed development of these epithelial cells. DGs can be found in both embryonic and adult human lung (25). However, the localization and function of lung DGs is not clear. DG has been demonstrated in rabbit airway smooth muscle, where it is associated with sarcoglycans (26). In contrast, neither rabbit epithelial cell nor Madin Darby canine kidney (MDCK) cell DG is associated with sarcoglycans. The functional differences between the DGs in these cell types is unclear.

The role of DGs in binding laminin makes them a potentially useful matrix receptor in central airways. Airway epithelium rests on a basement membrane comprised of several extracellular matrix proteins, including collagens and laminin-1 (27). Laminin-2 is expressed in the basement membrane of asthmatic airways (28). Alterations in either the composition of the basement membrane or the ability of the airway epithelium to bind these proteins may, in particular disease states, impair the repair process required during and after inflammatory injury to the mucosa. Airway epithelial cells use integrin receptors to adhere and migrate over collagen, laminin-1, and laminin-2 after acute mechanical injury (29). Although blocking specific -integrin subunits impaired migration over collagen-IV, similar treatment to cells grown on either laminin substrate did not. This suggests that airway epithelial cells may use other matrix receptors to facilitate spreading and migration during repair.

We present here the first report demonstrating expression of both -DG and -DG in human central airway epithelium. We also show that -DG is a functional laminin receptor in airway epithelium, and that its lectin domain can be inhibited by competition for its binding site to laminin. The presence of DGs in airway epithelium suggests a potential role for these proteins in epithelial cell migration and repair after injury.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

Penicillin, streptomycin, dextran, dextran sulfate, sodium heparin, N-acetylneuraminic acid (NAN), hydrocortisone, triiodothyrinine, transferrin, human epidermal growth factor (hEGF), insulin, bovine pituitary extract, type 25 protease from Bacillus polymyxia, and L-glutamine were obtained from Sigma (St. Louis, MO). Fetal calf serum (FCS) was obtained from Hyclone (Logan, UT) and was heat-denatured before use. The mouse anti--DG mAb 43DAG and the anti--DG mAb 8B4 were obtained from Novocastra (Burlingame, CA). The mouse antidystrophin mAb XIXC2 was obtained from Upstate Biotechnologies (Lake Placid, NY). The mouse anti--DG mAb IIH6 was a gift from Kevin Campbell, University of Iowa (Iowa City, IA). The mouse anti--DG mAb 6Cl was a gift of Neil Smalheiser, University of Illinois at Chicago. Mouse laminin-1 was obtained from ICN (Costa Mesa, CA). Fluorescein isothiocyanate (FITC)-conjugated and Texas Red-conjugated goat antimouse immunoglobulin (Ig) G and IgM were purchased from Molecular Probes (Eugene, OR). Succinyl-WGA (Triticum vulgare lectin), biotinylated WGA, and Texas Red-conjugated WGA were purchased from EY Laboratories (San Mateo, CA).

Culture of Human Airway Epithelial Cells

Primary normal human bronchial epithelial (NHBE) cells were purchased from Clonetics, Inc. (Walkersville, MD). These cells were derived from a single donor and were supplied as first-passage cells. Cells were plated into defined medium (Clonetics medium: 5 µg/ml insulin, 0.5 µg/ml hEGF, 10 mg/ml transferrin, 6.5 µg/ml triiodothyrinine, 0.5 mg/ml hydrocortisone, 0.5 mg/ml epinephrine, and 2 ml/liter bovine pituitary extract) on plastic plates. Cells were subcultured and used between passages 3 and 5 when ~ 40% confluent.

We also used cells collected from surgical specimens via a method described previously (30, 31). Briefly, airway segments collected from surgical specimens were cleaned of extraneous tissue and then incubated in phosphate-buffered saline (PBS) containing 0.1% protease and penicillin/streptomycin solution for 2 h. Bronchi then were filleted and epithelial cells were collected with a rubber spatula. Cells were plated in F12 supplemented with 10 µg/ml insulin, 5 ng/ml EGF, 0.5 µg/ml transferrin, 0.4 µg/ml hydrocortisone, 2 µg/ml triiodothyrinine, 100 µg/ml penicillin, and 5% FCS into collagen-coated six-well plates or two-chamber covered slides and incubated at 37°C in 5% CO₂ until confluent.

We also used the 1HAEo cell line (32), simian virus 40 transformed cells that have cell-surface markers similar to primary airway basal epithelial cells (33). Cells were grown on flasks, six-well plates, or chamber slides previously coated with laminin-1 (10 µg/ml) in Dulbecco's modified essential medium containing 10% FCS, 2 mM L-glutamine, 100 µg/ml streptomycin, and 100 U/ml penicillin G. Cells were used when ~ 90% confluent.

Human Tissue

Human tissue for all analyses was collected in accordance with the Declaration of Helsinki. Cross sections of fifth-generation airways were obtained from a lung that had been collected for transplantation but subsequently rejected for clinical use, and from a lung resected from a patient with non-small cell lung cancer. A block of human skeletal muscle was obtained from a human lower leg that had been freshly amputated. Tissue sections from lung or muscle were embedded in OCT resin (Sakura FineChemicals, Torrance, CA) and snap-frozen in liquid isoethane. Sections, 1 µm, were cut, placed onto polylysine-coated glass slides, and kept at 70°C until use.

Flow Cytometry

Cells were collected with a rubber spatula, washed in cold PBS containing 0.5% bovine serum albumin (BSA) and 0.2% NaN₃, and blocked in 5% human serum for 10 min at 4°C. Cells were stained with either IIH6 (neat) or 43DAG (1:20) for 45 min. IgG₁ or IgM isotype antibodies served as controls. Cells were then washed and incubated with a FITC-conjugated goat antimouse IgG or IgM for 30 min. Cells were washed in fluorescence-activated cell sorter (FACS) buffer, fixed in 2% paraformaldehyde, and analyzed on a Becton-Dickinson FACScan cytometer.

Immunofluorescent Staining of Cells

Cells in chamber slides were blocked with 3% BSA in PBS for 20 min at room temperature (rt). Tissue sections were stained without blocking. Slides were washed and incubated with 43DAG (1:100 dilution), 6C1 (1:50), or 8B4 mAb (1:10) for 1 h at rt. Slides were washed three times with PBS, after which 1:300 of Texas Red-conjugated 2°C goat antimouse IgG or IgM, as required, was added for 1 h at rt. Slides were washed three times in PBS and then counterstained with 1 mM Hoechst 33258 in water for 5 min. Slides then were washed five times in distilled, deionized water and visualized immediately.

Western Blot Analysis

Protein from 1HAEo cells was collected after lysing in protein lysis buffer (1% Nonidet P-40, 0.25% sodium deoxycholic acid (Na-DOC), 150 mM NaCl, 1 mM ethyleneglycol-bis-[-aminoethyl ether]-N,N'-tetraacetic acid, 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin, 1 µg/ml pepstatin, 1 µg/ml leupeptin, 1 mM Na₃VO₄, and 1 mM NaF) for 15 min at 4°C. Protein from primary epithelial cells was collected after dissecting the epithelial layer from a previously collected airway, digesting in 0.1% protease in PBS, and incubating pelleted protein in lysis buffer on ice for 5 min. All samples were then homogenized, shaken at 4°C for 10 min, centrifuged at 4,000 rpm for 10 min at 4°C, and stored at 70°C until analyzed. Equal loading of protein was assured by prior quantitation using the Bradford assay. Protein aliquots were separated on a 3 to 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) minigel and transferred onto nitrocellulose using a semidry technique. Immunoblotting was performed and developed using an ECL protocol (Amersham, Buckinghamshire, UK).

Isolation of RNA and Generation of Primers

Total RNA was isolated using a RNeasy mini kit (Qiagen, Inc., Valencia, CA). Kit directions were followed. Primers were generated to create full-length complementary DNA (cDNA) sequences of both polymerase chain reaction (PCR) products. Primers for the -DG portion of the DG gene were determined from the cDNA sequence (GenBank accession number L19711). The upstream external primer was 5'GACTTGAGCAAACTTGGACCTGGG-3', and the downstream external primer was 5'-GTTGGTCCATTCCACCACGATGGA-3'. The upstream nested primer was 5'-ATGAGGATGTCTGTGGGCTCA-3', and the downstream nested primer was 5'-GCCCCGGGTGATATTCTGCAG-3'. The expected fragment size after nested PCR was 1,959 base pairs (bp), starting at bp 395. Primers for the -DG portion of the DG gene were determined from the same GenBank cDNA sequence. The upstream primer was 5'-TCCATCGTGGTGGAATGGACC-3', and the downstream primer was 5'-TTAAGGTGGGACATAGGGAGG-3'. The expected fragment size was 729 bp, starting at bp 2,354.

Demonstration of Messenger RNA Presence by Reverse Transcriptase-PCR

Postive controls for all PCR reaction were performed using primers for glyceraldehyde-3-phosphate dehydrogenase. Total RNA samples (1 µg) had first-strand synthesis using MuLV reverse transcriptase (RT) in a Perkin-Elmer Gene Amp RT-PCR kit, following kit directions. Negative controls were done omitting the RT. Reverse transcription was carried out at 42°C for 60 min. Subsequent amplification of DNA was done with the specific gene primers using the following protocol: denaturation for 5 min at 94°C, followed by 30 cycles of 30 s at 94°C, 30 s at 58°C, and 30 s at 72°C, with a final extension of 7 min at 72°C. Samples were analyzed on a 1.5% agarose/TAE gel. For -DG, PCR was first done using the external primers. A second PCR reaction using the nested primers and the first PCR product then was done.

Cloning and Sequencing Amplified DNA Fragments

The recovered PCR product generated from the -DG primers was cloned using an Invitrogen TOPO TA Cloning kit, following kit directions. Clones were sequenced bidirectionally using a Sequenase-based ³⁵S dideoxy-sequencing kit (USB-Amersham) and sequencing primers located in the multiple cloning region of the vector.

Monolayer Wound Repair Assay

We have previously published details of this method (29). 1HAEo monolayers were washed and placed in serum-free medium. Wounds were made in the confluent monolayer with a rubber stylet to remove cells without disturbing the underlying protein matrix. Monolayers were treated with sham diluent, with 15 ng/ml EGF (an accelerant for epithelial wound closure [29]), or with 15 ng/ml EGF plus either the lectin WGA (30 µg/ml), NAN, dextran, dextran sulfate, or heparin (2 mg/ml each). Wound closure was photographed serially for 24 h using a Sony Iris CCD camera on a Nikon Diaphot inverted-stage microscope. Images were digitized using a Macintosh computer and Apple Video Player software (Apple Computer, Cupertino, CA). The area of the remaining wound in each image was measured using NIH Image software. Repair data are expressed as mean ± standard error of the mean of the percent remaining area of the original wound. Differences were examined using factorial analysis of variance; when significant differences were found, post hoc testing was done using Fisher's protected least significant difference test.

Additional monolayer wounds treated with 15 ng/ml EGF were fixed in neutral-buffered formalin 2 to 24 h after creation. Slides were blocked in lectin buffer containing 0.1% BSA for 20 min, then washed and incubated with 50 µl of Texas Red-conjugated WGA (20 µg/ml) overnight at 4°C. Slides were rinsed three times with PBS and stained for 45 s with 1 mM Hoechst 33258, after which they were rinsed in tap water and visualized immediately. Controls were processed omitting the lectin.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of DGs in Airway Epithelial Cells

We first screened for the presence of the DG gene in both 1HAEo and primary airway epithelial cells. Both -DG- and -DG-specific sequences of the common gene could be amplified from total RNA from each set of cells and were demonstrated at expected sizes (Figure 1). Subsequent sequencing of the amplified -DG and -DG fragments matched to the reported sequence (accession #LY19711) for human DG at > 99% homology.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Presence of DG transcripts in human airway epithelial cells. RT-PCR was performed on RNA isolated from both 1HAEo and primary airway epithelial cells. Negative lanes omitted RT from the initial reaction. Positive controls were done using the control plasmid pAW109 (not shown). Weights of bp are as marked.

Because it is possible that transcripts are present in cells without being translated, we next determined whether the DG protein products were expressed in human airway epithelial cells. We grew 1HAEo cells to confluence on laminin-1, then lifted and stained cells for the presence of -DG by flow cytometry. -DG was identified at the surface of 1HAEo cells using the 43DAG antibody (Figure 2). We then examined, by immunoblotting, whether -DG was present both in 1HAEo cells grown in similar fashion and in protein lysates collected from freshly harvested human central airway epithelium. The 43-kD -DG was clearly detected in both cell samples at the expected molecular weight (Figure 3). We then asked whether translation was ongoing, indicative of turnover of this protein, or whether -DG was a long-lived protein. 1HAEo cells grown to confluence on laminin-1 were treated with 25 µg/ml cycloheximide, which blocks messenger RNA (mRNA) translation, for 3 to 24 h, after which cell protein lysates were immunoblotted. The abundance of -DG protein decreased substantially in 1HAEo cells within 3 h of treatment (Figure 3).

View larger version (24K): [in this window] [in a new window]   Figure 2.   Flow cytometry analysis for the expression of - and -DG in 1HAEo cells. Cells were lifted using a spatula, blocked, and stained using either the 43DAG (-DG) or the IIH6 (-DG) mAb. Control isotype IgG1 and IgM isotype antibodies, respectively, are shown. Representative of four experiments each.

View larger version (69K): [in this window] [in a new window]   Figure 3.   Protein abundance of -DG in whole-cell protein lysates of primary human airway epithelium and in 1HAEo cells. In each set of experiments, lysates were resolved on 10% SDS-PAGE and probed for the presence of -DG using the 43DAG antibody. (A) Detection of -DG in primary airway epithelium. Cells were harvested from three different samples of human lung, from which lysates were collected and immunoblotted. (B) Detection of -DG in 1HAEo cells. Cells were treated with 25 µg/ ml cycloheximide for the time intervals shown immediately before protein collection. (C ) 1HAEo cells were grown to confluence on either collagen-IV or laminin-1, and were lifted from substrate either by using EDTA or by scraping from the culture flask. Lysates then were resolved and probed with 43DAG. CE, cells grown on collagen-IV and lifted with EDTA; CS, cells grown on collagen-IV and lifted by scraping; LE, cells grown on laminin-1 and lifted with EDTA; LS, cells grown on laminin-1 and lifted by scraping.

Because airway epithelium can be grown on a variety of extracellular matrix proteins, we then asked whether the underlying matrix protein influenced expression of -DG. In these experiments, 1HAEo cells were grown to confluence on either laminin-1 or collagen-IV, after which cell protein lysates were probed for the presence of -DG. -DG was found in 1HAEo cells grown either on collagen-IV or laminin-1 matrix, and after lifting cells from culture plates either with ethylenediaminetetraacetic acid (EDTA) or by scraping (Figure 3).

In contrast to the preceding experiments, the 156-kD -DG could not be detected in either set of cell protein lysates on immunoblotting using either the IIH6 or the 6C1 mAb. -DG also could not be localized to the cell surface by flow cytometry using the IIH6 antibody (Figure 2).

Localization of DGs in 1HAEo Cells and Airway Tissue

-DG was identified in 1HAEo cells grown on laminin-1 by immunofluorescent labeling using the 43DAG mAb (Figure 4). Protein was identified at the cell rim with a characteristic punctate nodularity. -DG staining was most intense in cells clustered in small groups, but was nearly absent in cells that were in the interior of large groups of cells. -DG was also identified in 1HAEo cell monolayers using the 6C1 mAb (Figure 4). Staining for -DG was similar to that noted for -DG. Localization was to the cell membrane in sparsely populated monolayers (< 30% confluent) and was never found in confluent 1HAEo cells. Staining with the IIH6 mAb did not identify -DG.

View larger version (35K): [in this window] [in a new window]   Figure 4.   Immunofluorescent localization of DGs in 1HAEo cells grown on laminin-1. (A) 1HAEo cells were fixed and labeled by indirect immunofluorescence using either an irrelevant IgG1 (a) or the 43DAG mAb for -DG (c), and counterstained with Hoechst 33258 (b and d, same fields as a and c, respectively). Arrow indicates localization of -DG. Bar: 20 µm. (B) 1HAEo cells were fixed and labeled by indirect immunofluorescence using either an irrelevant IgM (a), the XIXC2 mAb for dystrophin (c), or the 6C1 mAb for -DG (e), and counterstained with Hoechst 33258 (b, d, and f, same fields as a, c, and e, respectively). Bar: 20 µm.

-DG was identified in NHBE cells using both the 6C1 and the 8B4 mAb. Staining for -DG was seen in both the perinuclear region and at the cell membrane (Figure 5). As with 1HAEo cells, localization was greatest in single cells and cells in small clumps, and was not seen in more confluent regions. -DG was identified less well in NHBE cells using the 43DAG mAb (Figure 5). When present, it localized to the perinuclear region, unlike that seen in the 1HAEo cell line. For both DGs, staining was most intense in single cells and cells clustered in small groups, and nearly absent in cells that were in the interior of large groups of cells.

View larger version (25K): [in this window] [in a new window]   Figure 5.   Immunofluorescent localization of DGs in primary human airway epithelium. (A) Primary airway epithelial cells were fixed and labeled by indirect immunofluorescence using either an irrelevant IgM (a), or the 6C1 (c) or the 8B4 (e) mAb for -DG, and counterstained with Hoechst 33258 (b, d, and f, same fields as a, c, and e, respectively). Bar: 20 µm. (B) Primary airway epithelial cells were fixed and labeled by indirect immunofluorescence using either an irrelevant IgG1 (a) or the 43DAG mAb for -DG (c), and counterstained with Hoechst 33258 (b and d, same fields as a and c, respectively). Arrow indicates localization of -DG. Bar: 20 µm.

Neither - nor -DG was identified in the airway epithelium in cross sections of airway tissue using the respective 6C1, 8B4, and 43DAG antibodies. As a positive control, human skeletal muscle labeled appropriately for both DGs using these antibodies (not shown).

Effect of Blocking -DG on Epithelial Monolayer Wound Closure

DG is known to bind laminin, and kidney fetal epithelial cells are known to require DG/laminin binding during development (24, 34). However, although DG has been identified in rabbit airway epithelium (24), its role has not been formally established. To determine whether DG expression in airway epithelial cells was functional, we examined the migration and spreading of 1HAEo cell monolayers grown on laminin-1 matrix after mechanical injury. The initial area for 79 wounds was 2.15 ± 0.08 mm². Wounds treated with 15 ng/ml EGF closed substantially (29 ± 3% of time 0 area) compared with control wounds (48 ± 3% of time 0 area) (P = 0.0001). We then tested whether addition of a competing sugar for the -DG/laminin interaction would prevent cell spreading and migration over laminin matrix after wound creation. Concurrent treatment with EGF plus either dextran sulfate or heparin slowed repair to that of control wounds (Figure 6). Treatment with the nonsulfated dextran did not attenuate wound repair. Treatment with the monosaccharide NAN also attenuated wound repair stimulated by EGF with equal potency to dextran sulfate and heparin (Figure 6).

View larger version (32K): [in this window] [in a new window]   Figure 6.   Effect of competing sugars on wound closure in 1HAEo cells. Cells were grown to confluence on laminin-1. A small wound was made in each monolayer using a rubber stylet. Cells were then treated with 15 ng/ml EGF alone, or EGF plus 2 mg/ml of dextran (A), dextran sulfate (DS) (B), heparin (C ), or NAN (D). Wound closure was followed for 24 h by video microscopy, and remaining wound area was calculated. The control and EGF-alone curves in each panel are the same. n = 6 to 10 experiments in each group. *P  0.02, §P  0.005, and P  0.001, versus EGF alone at 24 h.

WGA competes with laminin for the binding site on -DG (15), and thus may interfere with cell migration and spreading over this matrix protein. To test this, we added WGA, 30 µg/ml, plus EGF at the time of wound creation. Wound closure was attenuated in these monolayers compared with monolayer wounds treated with EGF alone (Figure 7). Attenuation in repair was noted more after the first 6 h.

View larger version (21K): [in this window] [in a new window]   Figure 7.   Effect of the lectin WGA on wound closure in 1HAEo cells. Cells were grown to confluence on laminin-1. A small wound was made in each monolayer using a rubber stylet. Cells were then treated with 15 ng/ml EGF alone, or EGF plus 30 µg/ ml WGA. Wound closure was followed for 24 h by video microscopy, and remaining wound area was calculated. Control and EGF-treated monolayers are the same as in Figure 7. n = 8 to 10 experiments in each group. *P  0.001 versus EGF alone at 24 h.

On the basis of previous staining experiments in 1HAEo cells, the expression of -DG occurred only in single epithelial cells and cells at the edge of a cluster. We hypothesized that this expression would also occur in cells at the edge of a wound. In separate monolayers, we examined -DG abundance in wound margins 2 to 24 h after injury, using WGA lectin staining. WGA labeled cells at the time of injury near the wound edge only in a nuclear pattern (Figure 8). By 6 h after wound creation, cells at the wound edge labeled with a granular appearance. By 12 and 24 h, cells at the wound edge and migrated cells in the interior of the original wound had WGA labeling at the cell membrane (Figure 8). At these time points, cells away from the wound edge continued to label inconsistently and only with a nuclear pattern.

View larger version (80K): [in this window] [in a new window]   Figure 8.   Localization of -DG on 1HAEo cells during wound repair. Cells were grown to confluence on laminin-1. A small wound was made in each monolayer using a rubber stylet. Cells then were treated with 15 ng/ml EGF. Wound closure was followed for 0 (top row), 6 (second row), 12 (third row), and 24 (bottom row) h by video microscopy, at which time monolayers were fixed using 10% neutral-buffered formalin. Monolayers then were labeled by direct fluorescence using 20 µg/ml Texas Red- conjugated WGA and with Hoechst 33258. Matched low-magnification images showing Hoechst (left column) and WGA (middle column) labeling, and high-magnification images showing WGA (right column) are shown. Bar in each panel: 60 µm; L, wound lumen; e, wound edge. Arrows (right column) indicate localization of -DG by WGA lectin within cells.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The purpose of this study was to evaluate the expression and functional role of DGs in human airway epithelial cells. The presence of proteins and their common gene were demonstrated by several methods to avoid the limitations of any one method. Both - and -DG are clearly identified in airway epithelial cell lines and in human primary airway epithelial cells.

The expression of -DG at the edges of cell clusters, and that of -DG in single, isolated cells, strongly suggests that both proteins are involved in cell-matrix interactions and not involved in cell-cell junctions. The rim and nodular staining suggests that both colocalize at the cell membrane. The disappearance of -DG after treatment with cycloheximide suggests that the expression of -DG on the cell surface is dependent upon constitutive, active transcription, and is not a holdover in expression from development or cell plating. The absence of -DG in the interior of a cluster suggests that the expression and translation of this protein may be dynamically regulated.

The absence of labeling using antibodies for either - or -DG in airway tissue sections suggests some binding-site differences between airway epithelium and muscle, which may in turn be related to differences in glycosylation. These differences are in addition to the differences in the submembrane sarcoglycans seen between rabbit airway epithelium and smooth muscle (24), and suggest multiple differences in the DGC between airway epithelium and other cell types. Further, the lack of binding with IIH6 in human epithelial cells is contrasted to the efficacy of this antibody in identifying -DG in rabbit and canine epithelium (24). This suggests species differences in the binding site for this antibody. A third possibility is that the lack of staining may be related to the low abundance of the DGs in normal, nonrepairing, noninjured epithelium. Our data should be interpreted with caution in comparing the expression of the DGs in cultured airway epithelium with that in the in vivo situation.

Blocking the interaction between -DG and laminin with a sulfated glycosaminoglycan can block adhesion to this matrix protein in RT4 Schwannoma cells, bovine aortic endothelial cells, and neural cells (10, 13, 24, 35). Sulfate-containing glycosaminoglycans bind to the carboxyl-terminal domain of the laminin  chain (36). This same region in the E3 fragment of laminin  chain is implicated in the binding of -DG to laminin, and competitive binding between -DG and heparin sulfate at this site has recently been demonstrated (37). Addition of dextran sulfate or heparin attenuated wound repair over laminin-1 matrix in 1HAEo cells. In contrast, treatment with nonsulfated dextran did not block wound closure. Our data suggest strongly that sulfated glycosaminoglycans attenuate airway epithelial cell wound repair over laminin matrix by interfering with the binding of -DG to laminin.

In contrast to previous studies using endothelial cells and Schwannoma cells, in which monosaccharides such as NAN did not block cell adhesion (10, 13), addition of NAN did attenuate wound repair in 1HAEo cells. This suggests differences between these cell types in the interactions between -DG and laminin. The reasons are not clear but may relate to differences in the glycosylation of -DG in different cell types and species, which in turn may affect specificity and local charge at binding sites.

The lectin succinyl-WGA also attenuated EGF-stimulated wound repair. This lectin binds to N-acetylglucosamine (38) and can be used to isolate -DG (15). WGA binding is independent of the competition of -DG for the laminin binding site that is blocked by a glycosaminoglycan. Pretreatment of laminin-coated surfaces with WGA blocks subsequent adhesion and migration of fibroblasts via binding of N-acetylglucosamine residues on laminin (39). After addition of WGA, attenuation of wound repair was noted within 6 h, corresponding to the time when there was a change in WGA labeling of wound edge cells seen on lectin fluorescent histochemistry. These data suggest that WGA competes effectively with -DG for the binding site on laminin. Thus, two independent methods of interference with DG/laminin binding strongly suggest a role for DG in repair over laminin matrix.

Regulation of -DG expression was demonstrated after mechanical injury as shown by WGA labeling. Little WGA staining could be appreciated immediately after wounding. Granular staining was present within 6 h, consistent with new synthesis of -DG. There was considerable rim staining of cells within 12 h, consistent with accumulation of new -DG on the cell membrane. The nuclear staining in sections and monolayers may represent binding to nuclear core complex proteins, which have O-linked N-acetylglucosaminyl residues that bind WGA (40). These data suggest that injury, or an event early in the repair after injury, stimulates -DG production and mobilization to the cell membrane so as to facilitate repair.

Our present data suggest that -DG is a functional nonintegrin laminin receptor in airway epithelium with a role in cell migration and repair after injury. This result extends our previous finding that blocking ₂- or -₆-integrin subunits in human airway epithelial cells attenuates repair over laminin matrix only modestly, even though these subunits regulate repair over collagen-IV matrix (29) and bind laminin matrix in adhesion assays (41). These data suggest that there is a redundancy of matrix receptor systems in airway epithelium, and that the relative importance of each may depend on the presented matrix and other external factors. Abnormalities or changes in regulation of any of these may mediate needed repair after injury. Conversely, the existence of redundant systems may partially ameliorate an absence or dysregulation of another matrix receptor system.

The demonstration of functional DG receptors for laminin in airway epithelium adds to what is known about this receptor complex in other epithelial cell types. -DG has an apparent role in renal epithelial cell morphogenesis, such that blocking the binding of -DG to laminin inhibits development (24). DG is required for the formation of a basement membrane in embryoid bodies during development, and DG/laminin interactions are a prerequisite for the deposition of other basement membrane proteins (42). DGs have been identified recently in several epithelial types in adult mice, including salivary gland, kidney, trachea, and digestive tract (43). The role of the DGC in non-airway, adult epithelial tissues is not clear. One recent report demonstrates the presence of DG (24) in cultured MDCK cells. These were restricted to the basolateral membrane and may serve to help anchor cells to laminin.

There are limitations to these experiments. A major limitation to the immunolocalization of -DG in the human epithelial cell line is the extensive glycosylation of this molecule, which is both species- and tissue-specific (8, 12). The negative staining for -DG on flow cytometry may relate to changes in or damage to this extracellular protein on lifting cells from culture plates for analysis. A limitation in the wound-repair experiments is the concern that over time, wound closure may result from cell proliferation and not migration and spreading into the injury region. We have previously demonstrated that significant cell proliferation is not seen in human airway epithelial cell lines in the first 24 h after injury in this model, even after treatment with EGF (29).

A third limitation is the specificity of the WGA lectin, which can bind not only -DG (15, 24) but also other proteins (40). Blocking cell migration and spreading with WGA therefore may not be completely specific for the interaction between -DG and laminin. Similarly, the use of glycosaminoglycan sugars to block the lectin function of -DG is not completely specific in that other adhesion molecules with similar lectin function and utility in mediating cell migration and spreading could be blocked in a similar way. To date, no report has identified such a lectin function for integrins; however, other receptors on the cell surface may have a role in this process.

In summary, we show here the mRNA and protein expression of - and -DG in human airway epithelium. This cell surface receptor complex interacts with laminin matrix to modulate repair after injury. The role of these proteins in airway epithelial function under homeostatic conditions and after injury in situ remains to be defined.