Role of Cell Adhesion–Associated Protein, Pinin (DRS/memA), in Corneal Epithelial Migration

Abstract

purpose. To determine whether the cellular distribution of cell adhesion–associated protein, pinin, is altered during corneal epithelial migration in response to debridement wounding and to determine the effect of overexpression of pinin in cultured epithelial cells.

methods. Corneas from guinea pig and embryonic (day 17) chickens were excised, wounded, and placed on organ-culture rafts. At time points from 0 to 24 hours, corneas were cryosectioned and subsequently analyzed by immunofluorescence or immunoelectron microscopy for the presence and distribution of pinin. Cultured epithelial cell line MDCK (Madin Darby canine kidney) confluent monolayers were wounded by scraping and examined by immunofluorescence for pinin and desmoplakin. MDCK cells were transfected with full-length pinin cDNA. After selection in Geneticin, clones of pinin-transfected cells were isolated. Monolayers of transfected cells were scrape-wounded and assayed for their ability to migrate.

results. Within 2 hours after wounding, although morphologically identifiable desmosomes were present on migrating epithelial cells, the association of pinin to desmosomes was greatly reduced. Finally, after completion of wound closure, pinin returned to the corneal epithelial desmosome. Wounding of confluent epithelial monolayers (MDCK) in vitro demonstrated a very similar change in the distribution of pinin, whereas desmoplakin remained cell boundary–associated. Transfection of pinin into cultured epithelial cells resulted in an overexpression of pinin. Clones of cells expressing high levels of pinin exhibited marked reduction in their ability to migrate after wounding.

conclusions. Pinin is involved in corneal epithelium migration. The localization of pinin at or near the desmosome is correlated with the epithelial quiescence. The loss of pinin from the cell boundary correlates with the transition from quiescence to actively migrating. Overexpressing pinin in cultured epithelial cells affects epithelial homeostasis and, in turn, drives the epithelial cells to a hyperstable epithelial adhesive state and inhibits the transition from quiescence to migratory.

The cells of the corneal epithelium exhibit numerous specializations of their cell surfaces to ensure firm adhesion to neighboring cells and the extracellular matrix below. Both desmosomes along the lateral cell surface and hemidesmosomes of basal cell surface serve as sites of reinforcement of adhesion as well as anchorage points for the intermediate filament (IF) scaffold. ^{1 2} These junctional specializations are involved in the structural and functional integration of the corneal epithelium. 

Upon wounding, the corneal epithelium undergoes a transition from quiescent state to a state of active epithelial migration. This transition involves the alteration of numerous adhesive components of the epithelial cells. ^{3 4 5} Shortly after wounding, the basal cells of the corneal epithelium at the wound margin lose their hemidesmosome attachments, change their cell shape from columnar to an elongate flattened morphology, and project lamellapodia in the direction of migration. ^{5 6 7} Overall, the transition can be described as a loosening in cell–cell and cell–matrix attachments, with a concomitant change in the cytoskeletal organization of the epithelial cells. ^{8 9} Wound healing, however, occurs by the integrated centripetal migration of sheets of cells. The successive tiers of cells move as a continuous epithelium in a unified and coordinated manner, and therefore, cell–cell adhesion must be maintained. The desmosome is an integral component in this transition in that it serves both as adhesion reinforcement and a site of cytoskeletal linkage. 

The requisite changes in adhesion and cytoskeletal organization, which allow epithelial migration, must be reversed upon completion of wound closure. The epithelium must reestablish firm cell–matrix and cell–cell adhesion to restore the barrier function. One could envision two potential failure points in the transitions required for successful wound closure and barrier reformation: the first, the uncoupling of adhesion structures and cytoskeletal linkages at the initiation of epithelial cell migration, and the second, the restoration of firm epithelial adhesion and cytoskeletal linkages that form the competent epithelial barrier after wound closure. The cell adhesion molecule CD44 is upregulated in corneal reepithelialization, ¹⁰ and the overexpression of E-cadherin can inhibit cell migration, ^{11 12} suggesting important roles of these cell–cell adhesion molecules in epithelial migration. However, the influence of regulatory mechanisms of cell–cell adhesion molecules on cell migration and modulation of the cell–cell adhesive complexes, such as cadherin/catenin complex and desmosome during cell migration, remains largely unknown. 

Pinin was first identified and characterized as a desmosome-associated molecule. ^{13 14} It is not integral to the desmosome, but rather is found associated with only mature desmosomes. ^{13 14} Our studies of pinin suggest that its presence at the desmosome is correlated to highly organized, perpendicular bundles of keratin filaments. ¹³ We have suggested that, although it is not primarily responsible for the binding of IFs to the desmosome, the placement of pinin may stabilize the desmosome–IF association and reinforce the cell–cell adhesion of the epithelial cells. ¹⁴ Furthermore, the most recent studies from our laboratory and others have suggested that pinin/DRS/memA may function in multiple subcellular locations, including desmosomes and the nucleus. ^{15 16 17 18} Interestingly, we also have demonstrated that pinin is involved in tumorigenesis as a potential tumor suppressor, and it is absent from desmosomes in epithelial tumor cells that exhibit invasive behavior. ¹⁹ It is becoming clear that pinin/DRS/memA, like other multilocational proteins such as β-catenin and plakoglobin, ^{20 21} is a multifunctional protein involved in central cellular activity. 

We have examined the distribution of pinin within corneal epithelial cells as they undergo transitions from quiescence to migratory and from migratory to quiescence. We show that the association of pinin to desmosomes of migrating cells is greatly reduced. Thus, the regulation of pinin at the desmosome is illustrative of the dynamics involved in the transition between quiescence and migration. We next show that overexpression of pinin in an epithelial cell line (MDCK) not only dramatically enhanced epithelial cell adhesion but also depressed wound healing capability. 

Materials and Methods

Antibodies

Antibodies directed against the pinin included the mouse monoclonal antibody (mAb) 08L, ¹³ rabbit polyclonal 3A, ¹⁴ and UF215 (Shi J, Sugrue SP, unpublished results). Mouse anti-desmoplakin mAb cocktail was used to visualize desmoplakin (DP 2.15; DP 2.17; DP 2.20 IgG1; American Research Products, Belmont, MA). The anti-myc mAb 9E10 was purchased from BABCO (Richmond, CA). Fluorescein isothiocyanate–conjugated goat IgG fraction, to mouse IgG Fc and to mouse IgM (μ chain), secondary antibodies (Boehringer-Mannheim, Indianapolis, IN) and rhodamine-conjugated goat anti-mouse IgG Fc or IgM (μ chain; Cappel, Durham NC) were used in immunofluorescence staining. 

Cornea Organ Culture, Cornea Wounding, and Preparation of Frozen Tissue Sections

The experiments described in this study conform to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the NIH guidelines for use of animals. Corneas were harvested from adult Hartley guinea pigs, which were anesthetized with halothane and killed by cervical dislocation, and day 17 chicken embyros, which were killed by cervical dislocation. The corneal epithelia were wounded by outlining the wound area with 3 mm corneal trephine and then scraping the epithelium with a scalpel blade. This procedure produced superficial scrape wounds, removing the epithelium, but leaving the basement membrane and Bowman’s layer intact. After wounding of the epithelium, the endothelium was removed by scraping, and the corneas were placed in organ culture on stainless steel rafts at the air–culture medium interface. Dulbecco’s modified Eagle’s medium (DMEM) (Bio-Whittaker, Walkersville, MD), supplemented with 10% fetal calf serum (Cellgro; Mediatech, Herndon, VA) and a 1% glutamine, pen/strep cocktail (Bio-Whittaker) was used in all cultures. Both freshly isolated corneas and 4% formaldehyde–fixed corneas were embedded in OCT medium (Tissue Tek; Miles, Naperville, IL) and frozen for cryosectioning. Cryosections of 6 μm were cut perpendicular to the epithelial axis. 

Immunofluorescence Microscopy

Frozen cryostat sections of 6-μm thickness were fixed with acetone for 2 minutes at −20°C and extensively washed in phosphate-buffered saline. Prepared sections were incubated with anti-pinin mAb 08L and/or cell adhesion–related antibodies against desmoplakin. Double immunofluorescence staining was conducted as described previously, ¹³ using primary Ab mixtures of anti-pinin/anti-desmoplakin. Epifluorescence microscopy was conducted with a Zeiss Axiophot (Thornwood, NY). Images were digitally captured using a SPOT2 camera (Diagnostic Instruments, Sterling Heights, MI). 

Immuno-Electron Microscopy

Anti-pinin mAb was purified from hybridoma culture supernatant with protein-G (Pierce, Rockford, IL). Purified antibody was coupled to 7- to 10-nm-diameter colloidal gold particles as described previously. ¹³ For immunogold staining, 25-μm sections of cornea, fixed in 4% formaldehyde, were rendered permeable with buffer containing 0.5% Triton X-100 for 30 minutes at 4°C. After the pinin-gold incubation, the corneal sections were washed extensively in 20 mM Tris, pH 7.6, containing 150 mM NaCl, 1% bovine serum albumin (BSA). An indirect method was used for desmoplakin immuno-electron microscopy (EM). Briefly, corneal slices, prepared as above, were incubated in anti-desmoplakin antibody at 1:100 dilution. After extensive washes in Tris-buffered saline with 1% BSA, the sections were incubated in 50 μl of colloidal gold, conjugated to goat anti-mouse IgG (Janssen Biocemica, Beerse, Belgium) for 4 hours. The sections were then subjected to a series of washes in 20 mM Tris, pH 7.6, containing 150 mM NaCl, 1% BSA. Corneal slices incubated in colloidal gold were postfixed in 2.5% glutaraldehyde, 2% formaldehyde, and 1% picric acid buffered in 0.1 M cacodylate for 1 hour and prepared for routine EM. ¹³  

Cell Cultures, Transfections, and Cell Culture Wounding

The MDCK cell line was purchased from ATCC (Rockville, MD). Cells were cultured in DMEM containing 10% fetal bovine serum, 2 mM glutamine, and 200 U/mL each of streptomycin and penicillin G. A full-length pinin cDNA clone, derived from S₂08L, ¹⁴ was ligated into pCDNA3-myc vector (Invitrogen, San Diego, CA). Calcium phosphate transfection methodology used has been previously described. ¹⁴ Stable transfectants were created by growing cells in selection medium (DMEM, 10% fetal calf serum supplemented with 600 μg/ml Geneticin [G418]; Gibco/BRL, Grand Island, NY). After 2 weeks in selection medium, clonal colonies were isolated with cloning rings. The cloning procedure was carried out three times in succession to establish each stable clone. Clones were then analyzed for expression of pinin by Western blotting with anti-myc 9E10 and anti-pinin UF215. 

Stable pinin-transfected clones and normal MDCK cells were cultured to confluence on coverslips, and subsequently a central, 3-mm-wide linear scrape wound was placed across the coverslip. Coverslips were fixed for immunofluorescence at 0, 2, 6, 18, 24, and 48 hours. Phase micrographs of the wounded cultures were taken and the progressions of epithelial migration were documented at 4-hour intervals. The series of photographs were analyzed by measuring the distance from the wound edge of the epithelial cell sheet to the original wound site that was previously marked on the coverslip. Immunostaining for pinin and desmoplakin was then performed for cultures at 6, 18, 24, and 48 hours post-wounding. 

Immunoblotting of Whole Cell Extracts

MDCK cells and stable transfectants containing pinin cDNA were extracted as described by Ouyang and Sugrue. ¹⁴ Samples containing 30 μg of protein were loaded and run on 6% sodium dodecyl sulfate-polyacrylamide gels. The gels were transferred to nitrocellulose by the semi-dry method, and immunoblotting was carried out as described previously. ¹⁴ The primary antibodies were used at dilution 1:1000 for polyclonal serum UF215 and 1:4000 for mAb 9E10. Primary antibodies were detected by a 1:10,000 dilution of peroxidase-coupled secondary antibody (Boehringer Mannhein). The peroxidase was then visualized by ECL reagent (Amersham, Arlington Heights, IL), and blots were exposed to Kodak film (Eastman Kodak, Rochester, NY). The exposed film was then scanned in a Gel-Doc 1000 system (BIO-RAD, Richmond, CA) according to the provided procedure. The quantitative analysis of pinin was conducted using the Quantity One program in the Gel-Doc 1000 system (BIO-RAD). 

Results

Localization of Pinin in Normal Corneal Epithelium

Immunostaining of quiescent adult guinea pig corneas with both pinin and desmoplakin revealed a typical desmosome-like staining pattern (Fig. 1) . ²² The wing cells, whose surfaces are dense with desmosomes, exhibited an intense immunofluorescence. The basal cells showed less intense staining for pinin and desmoplakin, and some lateral surfaces seemed to be decorated with punctate spots, which are consistent with desmosomal staining. 

Localization of Pinin in Wounded Corneal Epithelium

After wounding of the guinea pig corneal epithelium, the distribution of pinin was remarkably changed. Shortly after wounding within 2 hours, the cells near the wound edge showed diminished cell boundary staining for pinin (Fig. 2A ). At the transition from a multilayered quiescent epithelial sheet at the periphery to a flattened, actively migrating layer of squamous cells around the wound edge, a decreased gradient of boundary-associated pinin immunoreactivity was demonstrated. However, there was clearly pinin immunoreactivity present within the cytosol of the actively migrating cells (Figs. 2A 2B) . At 6 hours post-wounding it is evident that the pinin immunoreactivity of the migrating cells appeared more diffuse within the cytosol, and pinin immunoreactivity was more limited to cell-boundary staining in the transition zone than in the peripheral epithelium, within which the immunostaining patterns for pinin and desmoplakin were very similar (Fig. 2C) . At 12 hours post-wounding, corneal epithelial cells migrating over the denuded surface of the wound bed demonstrated pinin was distributed diffusely within the cytosol of the epithelial cells participating in migration (Fig. 2D) . In contrast to pinin, desmoplakin immunostaining remained largely cell boundary–associated with some punctate accumulations along the lateral surfaces of the epithelial cells at 2 hours post-wounding (Figs. 2A′ 2B′ ). Remarkably, epithelial cells quite close to the leading edge still exhibited significant lateral cell surface–associated desmoplakin immunostaining even at 6 and 12 hours post-wounding (Figs. 2C′ 2D′ ). 

A similar change in the immunostaining pattern for pinin was observed in the wounded chicken corneal epithelium. At 4 hours post-wounding pinin immunoreactivity was significantly different from the distribution of desmoplakin (Figs. 2E 2E′ ). There was a high cytosolic signal for pinin with significantly diminished cell boundary–associated pinin immunoreactivity (Fig. 2E) . A few foci of pinin staining could be seen along the lateral borders (Fig. 2E , arrowhead), which seemed to be associated to desmoplakin, and therefore most likely represented desmosome association (Fig. 2E′ , arrowhead). 

These immunofluorescence studies revealed the redistribution of pinin in actively migrating corneal epithelia of both guinea pig and chicken, suggesting an involvement of pinin in the wound repair process of corneal epithelium. 

Immuno-EM Demonstrated Altered Subcellular Distribution of Pinin in Wounded Corneal Epithelium

Because our earlier studies have demonstrated that the chicken cornea is an appropriate model for immuno-EM localization and the penetration of antibody gold conjugates into the adult mammalian corneal epithelium was significantly restricted, we extended observations on the wounded chicken corneal epithelium to the electron microscopic level. We examined the lateral cell surfaces of the quiescent and actively migrating epithelial cells. As reported previously, ¹³ pinin immunoreactivity was localized to the IFs in the immediate vicinity of the quiescent corneal epithelial desmosome (Fig. 3A′ ). 

Although pinin was found near the desmosome shortly after wounding (2 hours) (Fig. 3A) , it appeared more loosely associated to the desmosome than it did in quiescent cells. ¹³ Six hours after wounding, pinin-gold was seen within IF tangles and aggregates some distance from the desmosome (Fig. 3B) , with greatly reduced desmosome-associated anti-pinin gold. At 12 hours post-wounding, the pinin immunoreactivity was no longer restricted to the desmosome but was associated with IFs and associated amorphous aggregates in the vicinity of the lateral membranes and extended well into the cell interior (Fig. 3C) . 

Interestingly, fewer 10-nm IFs seemed to be associated with these junctions of migrating cells although the desmosomes remained intact. Immunostaining with antibody directed against desmoplakin revealed that the plaques of desmosomes within migrating epithelial cells still contained desmoplakin (Fig. 3C′ ). There also was no non–desmosome-associated desmoplakin revealed, nor did we see evidence for internalization of desmosomes. These data suggest that, although morphologically identifiable and desmoplakin-positive desmosomes remained along the lateral borders of the actively migrating corneal epithelial cells, pinin was not present at these desmosomes but was seen associated within the cytosol associated with IF aggregates. 

Subsequent to wound closure, the desmosome–pinin association is reestablished. Immunogold localization of pinin within cells that had achieved wound closure revealed it to be both associated with the desmosome as well as with IFs scattered along the lateral surface (Fig. 3D) . Over time, pinin became more abundant at the desmosomes (Fig. 3E , arrow). In the stratifying epithelium, pinin was found to be more restricted to the immediate vicinity of the desmosome (Fig. 3F , arrow). It is noteworthy that even at 72 hours post-wounding, unlike that seen in the quiescent epithelium, ^{13 14} pinin-gold was seen in the IF tangles within the cytosol near the cell–cell boundary (Fig. 3F , arrowheads). This may indicate that the reestablishment of the cell-adhesion cytoskeletal linkage typical of the quiescent epithelial cells is a gradual process. 

Dynamics of Pinin Distribution after Wounding

We used the established kidney epithelial cell line MDCK, for in vitro epithelial wounding, cell migration, and cell adhesion studies, ^{23 24 25} to further extend our investigations. The distribution of pinin and desmoplakin in wounded MDCK cultures was very similar to that seen in wounded epithelium of the organ-cultured corneas, substantiating the correlation of change in pinin distribution and the process of pinin epithelial cell migration. Immediately after wounding, the distribution patterns of pinin and desmoplakin were nearly identical (Figs. 4 . Within 2 hours post-wounding, the immunostaining pattern of desmoplakin remained cell boundary–associated, even within the cells of the leading edge (Figs. 4B′ 4C′ 4D′ ), whereas pinin exhibited a more cytosolic distribution with weaker cell boundary staining in the cells near the wound edge (Fig. 4B) . At 6 hours post-wounding, the immunostaining of pinin was more diffuse within the cytosol with a few cytosolic accumulations of pinin immunoreactivity and desmoplakin remaining, for the most part, cell boundary–associated (Figs. 4C 4C′ ). 

The difference between the immunostaining patterns of pinin and desmoplakin remained obvious until the epithelial monolayer was reestablished. At 18 hours post-wounding the epithelial sheet had, to a great extent, closed the 3-mm wound, and obvious tension lines could be seen. These lines of tension were readily visible by immunostaining with desmoplakin and pinin (data not shown). The cells near the leading edge of the recently healed epithelia demonstrated weak pinin immunoreactivity along the cell boundary as well as high cytosolic staining (Figs. 4D 4D′ ). In cultures 24 hours post-wounding, the epithelium had again reached confluence. Although pinin immunoreactivity was again very similar to desmoplakin, there remained some cells that exhibited obvious boundary staining for desmoplakin but little if any pinin boundary staining (Figs. 4E 4E′ ). 

Very Limited Cell Spreading and Physically Constricted Hyperepithelium Resulting from Overexpressing Pinin in Normal Epithelium

Transfection of pinin cDNA into MDCK cells produced epithelial cells with greatly enhanced epithelial cell adhesion. These stable clones expressed pinin at levels roughly four times greater than MDCK endogenous levels, as determined by Western blotting (Fig. 5D , and Materials and Methods). Strikingly, the stable pinin-transfected cells exhibited very tight cell–cell adhesion with a very compact epithelium (Fig. 5C) , whereas vector-only control cells exhibited a similar phenotype as seen in normal MDCK (Figs. 5A 5B) . 

Delayed Cell Migration and Wound Closure in Pinin-Transfected Epithelium

Wounding of the stable pinin-transfected epithelium demonstrated that the pinin-overexpressing cells had limited capability to spread and migrate to cover the wound bed. The control sheet of cells actively spread and migrated into the wound bed within 6 hours (Fig. 6A ), covered the major area of the wound bed around 16 to 18 hours (Fig. 6B) , and accomplished closure in the 3-mm wound within 24 hours (Fig. 6C) . However, the epithelial sheet of the stable transfectants exhibited little cell spreading at the free edge and relatively little epithelial migration even at 18 hours (Fig. 6D) . Interestingly, the pinin-overexpressing cells do exhibit recognizable wound healing at 24 to 30 hours (Fig. 6E) , and the migration is apparent at 48 hours (Fig. 6F) . Yet, pinin-transfected cells, even at 60 hours post-wounding, had not accomplished wound closure, and the wound bed finally closed around 80 to 90 hours (data not shown). The significant delay of wound closure (2–3 days’ delay) in the wound healing of pinin-overexpressed cells suggested that pinin affects epithelial spreading and cell migration. 

The wounded pinin-transfected cells were immunostained for pinin and desmoplakin. At 6 hours post-wounding, there was no clear change for the distribution of pinin near the wounds (Fig. 7A ). These observations demonstrated a clear difference of pinin distribution between parental MDCK and pinin overexpressors after wounding. The distribution of desmoplakin in pinin overexpressors remained desmosomal as seen in parental MDCK cells after wounding (Fig. 7A′ ). Eighteen to 24 hours post-wounding only few areas within the cultures of pinin overexpressors exhibited limited reorganization and migration at the wound edge. Interestingly, it appears that pinin as well as desmoplakin was localized to the free edge of the “frontier cells.” This phenomenon had not previously been observed in the wounded parental MDCK cells (Figs. 7B 7B′ , dashed line). Rather than true free edge, this band of positive desmosome staining appeared to result from very attenuated cell processes, linked by adhesion junction, forming a restrictive belt in front of the leading edge. The cells at free edge seemed to be wrapped by long, junctionally linked processes of the neighboring cells and were thus physically restricted from migrating. The correlation between overexpression of pinin and the limited cell migration was further illustrated by immunofluorescent staining of pinin-overexpressed MDCK cells 24 hours post-wounding. It is obvious that the region in which the cells exhibited heavier pinin immunoreactivty along the lateral membrane exhibited little, if any, cell migration (Fig. 7C , bracket). These data indicate that the overexpression of pinin overrides the signals that initiate the epithelial migration, suggesting that pinin does play an important role in epithelial cell migration. 

Discussion

In this study we demonstrated the alteration in the distribution of the desmosome-associated protein, pinin, as the corneal epithelium undergoes the transitions from quiescence to active migration and from migration to quiescence, suggesting an involvement of pinin in cell migration during corneal wound healing. Using cell culture model system, we also demonstrated that overexpression of pinin enhanced cell–cell adhesion and limited the cell migration, indicating a crucial role for pinin in epithelial cell migration. 

Cellular events that allow the transition from quiescence to active migration no doubt include differential gene expression and posttranslational modification of existing proteins, such as phosphorylation and/or dephosphorylation of cell adhesion and cell adhesion–associated proteins. ^{26 27 28 29 30 31 32 33 34} Here, we show a change in the intracellular distribution of a cell–cell adhesion-related protein pinin in corneal wound healing. Although the precise molecular mechanism for the change in pinin distribution has not yet been identified, pinin is a highly phosphorylated protein and possesses multiple protein kinase recognition motifs, ¹⁴ and it is possible that changes in the phosphorylation state of pinin may regulate its association the desmosome. 

The alteration in the distribution of pinin in the migrating cells is suggestive of a weakening of the epithelial cell–cell adhesion. Remarkably, in various epithelial models young, but morphologically recognizable desmosomes, which immunostain with desmoplakin, are not decorated with pinin. ^{13 14} It is tempting to speculate that redistribution of pinin in wounded epithelial cells may destabilize the pinin–IF–desmosome complex, thereby reverting to the naïve or younger state. The alteration of pinin distribution may be one of the molecular events for modulating the epithelial cell–cell adhesion during cell migration. 

We suggest that upon transition from quiescence to migration, rather than a wholesale turnover of the epithelial desmosomes as seen in epithelial dysadhesions, there is a partial disassembly of the mature desmosomes. This partial disassembly would thereby render the desmosome more plastic and conducive to the cell rearrangements requisite for migration. However, the presence of both cadherin-based adhesions and naïve desmosomes retain the requisite cell–cell adhesion for requisite physical integration of the epithelial sheet during the process of migration. Danjo and Gipson ³⁵ have demonstrated that the E-cadherin linked to actin filament cables at the leading edge of healing circular epithelial wounds in cornea form a functional adherens junction organized in a “purse string” arrangement. One could envision that the force generated at the leading edge by the tightening purse string is transmitted along the epithelium through close epithelial cell–cell adhesions. Indeed, immunofluorescence for desmoplakin and pinin in wounded cultures of MDCK revealed obvious tension lines that were well back from the leading edge. These tension lines are typical of areas of increased cell–cell adhesion that afford increased tensile strength to the epithelial sheet. 

Overexpression of pinin in HEK 293 cells resulted in phenotypic and molecular changes culminating in increased cell–cell adhesion and promotion of epithelial phenotypic characteristics. ¹⁴ The induction of pinin expression in inducible pinin transfectants resulted in an inhibition of anchorage-independent growth of the 293 cells in soft agar. ¹⁹ Intriguingly, Degen et al. ¹⁶ have reported that the mRNA coding for pinin (DRS/memA) is upregulated in certain highly metastatic melanomas. In addition, we have reported that some renal cell carcinomas may exhibit a lack of pinin expression, whereas a small subset demonstrate a very elevated level of expression. ¹⁹ Given the fact that the overexpression of pinin in MDCK cells inhibited the migration, the upregulation of pinin in invasive melanomas and certain renal cell carcinomas may be indicative of alterations(s) in an other component of the crucial intracellular pathways within which pinin is involved. Taken together, pinin’s ability to positively influence cellular adhesive properties and its altered expression in carcinomas, which exhibit enhanced migratory phenotypes, collectively suggest that this protein may play a pivotal role in regulating epithelial adhesion and migration. 

Pinin is ubiquitously expressed in several tissues including non–desmosome-bearing tissues. ¹⁴ Recent works from our laboratory ¹⁸ and others ^{15 17} indicated that pinin is found within the nucleus as well. Although we think the claim for pinin/DRS as an exclusive nuclear protein ¹⁷ needs to be reexamined, it is now clear that pinin/DRS exhibits multiple locations in the cytosol, at the desmosome and/or within the nucleus. It is becoming evident that pinin may play important regulatory roles in different multiprotein complexes. ^{15 16 17 36 37} Because this study is mainly focusing on involvement of the cell adhesion–related function of pinin in epithelial cell wounding, the further study regarding nuclear form/function of pinin versus cytosolic and desmosomal form/function will be addressed in the future. It is, however, tempting to speculate that pinin, like β-catenin, may play a structural role in cell–cell adhesion sites and function in signaling to the nucleus. ^{38 39 40 41 42 43} These two roles can function independently or coordinately to influence cell adhesion, cell motility, and cell proliferation/differentiation. ^{20 44 45} Nevertheless, the redistribution of pinin during the process of epithelial wound healing highlights the dynamic nature of the adhesion and cytoskeletal linkages of the corneal epithelium. Our future studies will target the precise mechanism responsible for the state of assembly of pinin and the exact role for pinin playing during wound healing of corneal epithelia. 

 Supported by National Institutes of Health Grant EY07883.

 Submitted for publication July 20, 1999; revised December 1, 1999; accepted December 28, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Pinin distribution in unwounded corneal epithelium. Adult guinea pig corneal epithelium was immunostained for desmoplakin (A) and pinin (B). Note that in the intact epithelium, the staining pattern of these two desmosome-associated proteins is very similar. The wing cells exhibited abundant staining for desmoplakin and pinin, consistent with the abundance of desmosomes in this cell layer. The basal cells show pinin immunostaining that exhibited the punctate desmosome-like pattern. Bar, 10 μm.

Figure 2.

 

The distribution of pinin and desmoplakin in wounded corneal epithelium. Cryosections of scrape-wounded guinea pig and chicken corneas were double-immunostained with anti-pinin (A through E) and anti-desmoplakin (A′ through E′) mAbs. At 2 hours post-wounding (A, B), there is a clear difference in the distribution pattern of pinin (green) and desmoplakin (red) from the epithelial peripheral to the leading edge of migrating epithelium (A, from the left to the right). The redistribution of pinin is more evident in the leading edge (B), whereas desmoplakin remained largely boundary-associated (B′). The redistribution of pinin is more dramatic along the migrating zone of the guinea pig corneal epithelium at 6 hours post-wounding (C) in comparison to nearly, almost, unchanged desmoplakin distribution (C′). At 12 hours post-wounding of the guinea pig cornea, it was clearly demonstrated that pinin immunostaining was diffuse cytoplasm (D), whereas desmoplakin still remained cell boundary–associated (D′). The similar redistribution of pinin also has been observed in migrating chicken corneal epithelium after wounding. (E, E′) Pinin and desmoplakin, respectively, showed a high magnification of immunofluorescent staining profile of the leading edge of chicken cornea at 4 hours post-wounding. Bar, 15 μm.

Figure 3.

 

Immunogold localization of pinin in migrating epithelia revealed a redistribution of pinin. As early as 2 hours (A) post-wounding, anti-pinin–conjugated gold was seen associated to aggregates of IFs (arrow) in the vicinity of the desmosome but more scattered than seen in quiescent epithelia (A′). At 6 (B) and 12 (C) hours post-wounding, anti-pinin gold was seen associated to IFs (arrow) and IF aggregates (arrow) some distance from the desmosome. Desmosomes of the corneal epithelium 12 hours post-wounding still exhibited immunogold labeling for desmoplakin (arrow in C′). Shortly after wound closure, pinin was seen both associated to the desmosome as well as to IFs scattered along the lateral surface (D, arrow, 36 hours post-wounding). Over time, pinin became more restricted to the IF bundles in the immediate vicinity of the desmosome (E, 48 hours post-wounding). As the epithelium restratified, pinin became more abundant at the desmosomes (F, arrows, 72 hours post-wounding).

Figure 4.

 

Cultured MDCK cells reveal the change in pinin distribution subsequent to wounding of a confluent epithelial sheet. At 0 hours post-wounding, the distribution of pinin (A) and desmoplakin (A′) was virtually identical. Within 2 hours after wounding, however, the distribution of pinin was more diffuse within the cytosol (B). The cells near the wound edge exhibited high cytosolic pinin immunoreactivity but little cell boundary staining. Desmoplakin remained primarily cell boundary–associated and punctate, consistent with its desmosome location (B′). At 6 hours post-wounding, the loss of cell boundary staining for pinin was more apparent (C). In addition, there appeared some cytoplasmic aggregations of pinin immunoreactivity, whereas desmoplakin remained cell boundary–associated (C′). At 18 hours post-wounding, the 3-mm wound was significantly closed, and pinin started to appear more cell boundary–associated. The immunostaining for pinin and desmoplakin along a fusion-seam revealed slight cell boundary staining for pinin, with heavy cytosolic immunoreactivity (D) and largely unchanged boundary staining for desmoplakin (D′). At 24 hours post-wounding, the wound was mostly healed. Along the seam of the remaining wound site, the majority of the elongate cells demonstrated cell boundary staining for pinin as well as desmoplakin (E, E′). The nuclear staining with DAPI (blue) has been superimposed to either pinin staining or desmoplakin staining. Also note that the fluorescent tags of the secondary antibodies used for immunostaining at 6 hours post-wounding (C, C′) were intentionally switched to verify that the different staining patterns of pinin and desmoplakin were not due to secondary antibody variations or differences in the sustaining of fluorescent filter sets. Bar, 10μ m.

Figure 5.

 

Transfection of full-length pinin cDNA into MDCK cells resulted in very compact epithelial colonies. Control MDCK cells and MDCK cells transfected with C-myc–tagged full-length pinin were plated at the same density. Phase contrast photographs were taken at 24 hours post-seeding for parental MDCK and control vector–transfected cells and at 72 hours post-seeding for the pinin-transfected MDCK cells. As shown in (A), parental MDCK cells displayed typical epithelial islands. The control vector-transfected cells exhibited very similar phenotype as seen in nontransfected MDCK cells (B). However, stable pinin transfectants exhibited much more compact epithelial islands, consisting of numerous tightly packed cuboidal epithelial cells (C). The stable transfected cells seemed to exhibit enhanced cell–cell adhesion and less cell-substrate spreading. (D) Western blot analyses were conducted to estimate the levels of exogenously expressed pinin in the stable pinin transfectants on two duplicate blots. Total cellular proteins were extracted from parental MDCK (lanes 1, 3) and stable pinin transfectants (lanes 2, 4). The blot immunoreacted with anti-myc mAb 9E10 and anti-pinin polyclonal Ab UF215, respectively. The position of endogenous pinin and myc-tagged pinin on 6% PAGE gel is indicated by the arrow. Pinin was 4.2-fold overexpressed in stable pinin-transfected cells, as determined with the Quantity One densitometry program in Gel-Doc-1100. Bar, 100 μm.

Figure 6.

 

Overexpression of pinin in MDCK cells affected migration of the epithelial sheet. Confluent epithelial sheets of normal MDCK cells and the stable pinin-transfectants were scrape-wounded. Phase contrast photographs were taken at 6 hours (A), 18 hours (B), and 24 hours (C) for the parental MDCK cells and 18 hours (D), 30 hours (E), and 48 hours (F) post-wounding for transfectants. The parental epithelial cells actively migrated into the wound bed at 6 hours (A) and accomplished wound closure around 24 hours (C). The epithelial cultures of the stable transfectants, on the other hand, exhibited little cell spreading at the free edge and relatively little epithelial migration at 18 hours post-wounding (D). The stable transfectants began the process of wound healing at 24 to 30 hours (E), and epithelial spreading and migration was evident at 48 hours (F). The dashed lines in (A), (D), (E), and (F) indicated the original wound edge, whereas, the lines of wound edge could not be shown in (B) and (C) because of the extensive closure accomplished in migrating cells. Bar, 100 μm.

Figure 7.

 

Overexpression of pinin affected cell migration in the cell culture system. Double immunostaining with anti-pinin mAb (A through C) and anti-desmoplakin mAb (A′ through C′) was carried out to examine the distribution of pinin and desmoplakin in wounded pinin-transfected epithelia. At 6 hours (A), 18 hours (B), and even at 24 hours (C) post-wounding, pinin (A through C) was seen associated with the lateral cell boundaries of the cells at the wound edge and appeared in close association to desmoplakin (A′ through C′). The transfected epithelia did not spread and the cell shape remained somewhat cuboidal. Note that in some areas pinin and desmoplakin were found near the free wound edge (B, B′, dashed line). The bracket indicated the region where pinin was heavily stained and exhibited little, if any, cell migration (C, C′). Bar, 10 μm.

The authors thank Todd Barnash and Summer Carter for technical assistance.