| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratoire d’organogénèse expérimentale (LOEX), Hôpital Saint-Sacrement du CHA de Québec and Surgery Department, Laval University, Québec, Canada
1Correspondence: LOEX, Surgery Department, Laval University, LOEX, Hôpital Saint-Sacrement du CHA de Québec, 1050 Ch. Ste-Foy, Québec, PQ, Canada, G1S 4L8. E-mail: veronique.moulin{at}chg.ulaval.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: wound healing • keratinocyte • epithelia • drug screening
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
, and on the terminal differentiation of keratinocytes into a protective cornified layer (4
, 5)
.
The excision of full-thickness skin creates a defect that is wider and longer to heal than an incision (6)
. In excisional wounds that are left open, the reepithelialization progresses from the surrounding wound margins toward the center, creating a continuum in the regeneration of a differentiated epidermis over a DEJ in reorganization (1
, 5
, 7)
. The chronology of events occurring during reepithelialization can therefore be studied with the full spectrum of changes that develop spatially (8)
.
The mechanisms of wound reepithelialization have been debated for a long time but remain unclear. Different hypotheses have been envisioned for the process of keratinocyte migration over the wound bed (9
, 10)
. A sliding mechanism as a coherent sheet of cells was proposed based on tight intercellular cohesion (11
12
13
14)
. In this conception, basal cells are actively motile whereas the above cluster of superficial cells is passively dragged along. A different hypothesis involving the crawling of individual cells over each other (leap-frog model) at the tip of the regenerating epidermis was suggested (7
, 15
16
17)
. Another concern is the implication of proliferation during reepithelialization (9)
. Some studies, most of which were conducted on linear incisional wounds, have emphasized hyperproliferation at the wound margin (7
, 18
19
20)
. Fragmentary information about proliferation in the regenerating epidermis has also been reported in these and other studies (21
22
23)
. Its contribution to wound closure remains controversial: negligible (18
, 22
, 23)
or important (7
, 19
, 20)
, depending on the report.
In this study, we present a tissue-engineered wound-healing model we developed in order to study in vitro the mechanisms of wound reepithelialization in a controlled manner. Reconstructed human skin (24)
(RHS) was wounded with a 6 mm biopsy punch and the reepithelialization from the surrounding margins was studied. We first verified that the model behaved as in vivo by observing features of human wound healing such as fibroblast and keratinocyte migration, acceleration of wound closure in the presence of exogenous factors, and reorganization of a DEJ over a regenerating epidermis. We then took advantage of the spectrum of changes developing along the neoepidermis (NE) to divide it into regions for investigating the mechanisms underlying wound reepithelialization. Based on our observations, we propose that keratinocytes are displaced toward the wound center by two independent mechanisms that occur in distinct regions of the NE and that both contribute to wound reepithelialization.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
were used between the fifth to ninth passage. Human keratinocytes were isolated from newborn foreskin, cultured as described previously (26)
, and used at third or fourth passage.
Wound-healing model
A completely biological RHS was produced as described previously (24)
. Human fibroblasts grown for 28 days in the presence of 50 µg/ml of ascorbate (Sigma-Aldrich Canada, Oakville, Ontario, Canada) secreted an abundant extracellular matrix and formed manipulatable sheets (27)
(Fig. 1
a). Two fibroblast sheets were superimposed and left to adhere together for 7 days. Human keratinocytes were seeded on the stack of fibroblast sheets and cultured in submerged conditions for 7 days. They were induced to differentiate by elevating the tissue construct to the air-liquid interface for 14 days (Fig. 1b
).
|
The mature RHS was wounded (Fig. 1c
) with a 6 mm diameter biopsy punch (Laboratoire Stiefel, Nanterre, France) and placed over a third fibroblast sheet (obtained as stated above) to allow reepithelialization by keratinocytes onto a natural matrix. The wounded RHS was cultured at the air-liquid interface for 3, 7, or 14 days. The reepithelialized surface was photographed at the time of biopsy (Fig. 1d
). The biopsies were removed to produce the wounds and the wounded RHS with adjacent uninjured skin and human foreskin were processed for histology, electron microscopy, and immunofluorescence staining.
Wound treatments and calculation of the reepithelialization rate
Treatments consisted of drying the surface of the wounded RHS with Whatman paper #4 (Fisher Scientific, Nepean, Ontario, Canada) before applying 20 µl of various solutions: 0.9% NaCl saline solution, culture medium, platelet-rich plasma prepared by the hematology service, or fibrin. The fibrin clot was prepared by combining 3 mg/ml of fibrinogen (Sigma) with 0.01 UI/µl of thrombin (Sigma) in a 5:1 ratio. Treatments were twice a day during 3 days for dried wounds (controls), saline, or culture medium-treated wounds. Clots generated from platelet-rich plasma or a combination of fibrinogen with thrombin were deposited only once, i.e., after wounding.
The rate of reepithelialization was calculated for each treatment (n=4, i.e., two independent experiments done in duplicate) using the surfaces photographed (Fig. 1d
) and measured with NIH Image software. Although contraction was weak, we eliminated its contribution to wound closure by measuring the un-reepithelialized and total wound surfaces on the third day after wounding. The radius of these circular surfaces was subtracted and the result divided by 3 days to obtain the reepithelialization rate. The Wilcoxon rank sum statistical test (P<0.05) was conducted to compare the treated wounds with the control (dried wounds). Calculations based on the initial wound diameter of 6 mm gave identical statistical conclusions.
Histology
Tissues were fixed in 4% paraformaldehyde and paraffin embedded. Microtome sections were stained with hematoxylin, phloxine and saffron, toluidine blue, Masson’s trichrome, or periodic acid Schiff (PAS).
Immunohistochemistry
Indirect immunofluorescence was performed on acetone- or methanol-fixed cryosections as previously reported (8)
. The primary antibodies used were mouse monoclonal anti-human Ki67 (BD PharMingen, Missisauga, Ontario, Canada), anti-human filaggrin (BTI, Stoughton, MA), anti-human laminin 5 (chain 
2) (Chemicon International, Temecula, CA), and anti-bovine keratin 1 and 10 (K1/K10) (clone K8.60, Sigma), rabbit polyclonal anti-human collagen IV (Chemicon), and anti-mouse loricrin (BAbCo, Richmond, CA). The secondary antibodies (Chemicon) were rhodamine conjugated goat anti-rabbit IgG or goat anti-mouse IgG-IgM. Negligible background was observed for controls (primary antibodies omitted).
Transmission electron microscopy
Samples were fixed in 2.5% glutaraldehyde and processed as described previously (28)
.
Calculation of the proliferative index
We studied the expression of Ki67 to determine the proliferation rate in different regions of the 3 day NE and of the unwounded epidermis as used by others (29)
. The proliferative index was calculated by dividing the number of labeled basal cells by the total number of basal cells in the same 40x microscope field. The regions of 12 samples were used for the calculations. The Wilcoxon signed rank statistical test (P<0.05) was conducted to compare the indexes obtained for each region with their paired indexes in the unaffected epidermis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). At the air-liquid interface, its macroscopic aspect became more opaque and whitish (Fig. 1b
). A 6 mm circular wound was produced (Fig. 1c
). Keratinocytes reepithelialized it by migrating onto the natural matrix secreted by fibroblasts. This model allowed us to follow the wound closure macroscopically by observing the ring of reepithelialization that progressed toward the wound center (Fig. 1d
). Thus, the effect of diverse factors on healing was quantified by measuring the reepithelialization rate (Fig. 2
). Fibrin and platelet-rich plasma significantly accelerated the reepithelialization rate (
0.75 mm/day) compared with the control (0.4 mm/day). This shows the sensitivity of the model for the evaluation of the effect of exogenous factors on wound reepithelialization.
|
Histological and electron microscopical analysis of the wound-healing model
The unwounded RHS presented the histological features of a differentiated epidermis (Fig. 3
a, b). Basal keratinocytes were cuboidal or prismatic. Numerous spinous processes characterized the 6- to 10-cell thick spinous layer. Cells of the granular layer (2–4 cells thick) presented abundant granules and progressively more contrasted plasma membranes. Flattened corneocytes with a thickened plasma membrane formed a thick cornified layer. The dermal sheets supporting the keratinocytes were composed of fibroblasts embedded in their endogenous matrix. The wound margin could be easily recognized by locating the cut made by the biopsy punch into the two initial fibroblast sheets (Fig. 3a
). The third fibroblast sheet, placed underneath after wounding, served as a natural support for keratinocyte migration from the surrounding epidermis. The NE progressed toward the wound center. Partially reepithelialized after 3 days, the wounds were closed after 7 days.
|
We separated the 3-day NE into 4 regions based on histological criteria corresponding to the progressive differentiation of the NE from the wound center toward the margin (Fig. 3a
). 1) In the wound, the region called ‘near the wound margin’ had a thick cornified layer that ended abruptly due to the cut made by the biopsy punch. 2) From this easily recognizable reference point and toward the wound center, the ‘intermediate region’ presented a progressively discontinuous granular layer. 3) ‘Behind the migrating epidermal tongue’ (bMET) no granular cells were present, but the NE contained a significant number of organized layers composed of differentiating keratinocytes (flattening with spinous processes). 4) The ‘migrating epidermal tongue’ (MET) was constituted of keratinocytes forming the tip of the NE without showing significant features of organization and differentiation.
Electron microscopy of the unwounded RHS showed many desmosomes throughout the epidermis (Fig. 4
a, d). They were particularly abundant in the spinous layer and their remnants were still present in the stratum corneum. The flattened corneocytes possessed cornified envelopes. In the granular layer, we observed lamellar bodies (Fig. 4b
) and keratohyalin granules (Fig. 4c
). These ultrastructural features of differentiation were also present near the wound margin of 3, 7, and 14 day NE.
|
Reformation of the dermoepidermal junction in the wound
By electron microscopy, we observed an organized DEJ in the RHS (Fig. 4f
). Hemidesmosomes attached basal keratinocytes to the basal lamina (Fig. 4e
, 4f
). Cytoplasmic tonofilaments were inserted into the hemidesmosomal dense plate from which anchoring filaments traversed the sub-basal dense plate in the lamina lucida and reached the lamina densa. Anchoring fibrils extended from the latter into the connective tissue, where anchoring plaques were observed. The DEJ was mostly continuous. Reduplication foci of lamina densa and plasmalemmal vesicles in basal keratinocytes were noted, suggesting DEJ turnover (30)
. ‘Near the margin’ 7 days after wounding, we observed that the DEJ was discontinuous. However, near the margin of 14 day wounds, the DEJ was comparable to adjacent RHS in terms of continuity and organization. The DEJ was progressively more discontinuous and hemidesmosomes were less frequent toward the wound center. Foci of DEJ in reorganization were separated by regions without lamina densa, which were often seen under cytoplasmic projections (Fig. 4g
). These foci of reorganizing DEJ contained hemidesmosomes in reformation. Accordingly, the deposition of the DEJ, which was also studied with PAS, progressed over time but remained weak and discontinuous after 14 days compared with the adjacent RHS (data not shown).
The deposition of laminin 5 and collagen IV along the DEJ of unwounded RHS was comparable to that of human skin (summarized in Table 1
). Near the margin of 3 day wounds (Fig. 5
a, e), their deposition at the DEJ was almost continuous. Toward the wound center, their labeling at the DEJ became progressively less intense and more discontinuous. Under the MET, the light and discontinuous laminin 5 labeling (Fig. 5a-c
) along with the absence of collagen IV (Fig. 5e
) indicates that the deposition of laminin 5 was initiated rapidly after keratinocytes had settled down and that it occurred earlier than collagen IV secretion at the DEJ (Fig. 5a
vs. e). The keratinocyte cytoplasmic expression of laminin 5 and collagen IV (Fig. 5a
, 5e
), mainly present in the basal layer near the wound margin, increased and involved more suprabasal cells toward the MET, where most of the keratinocytes were labeled. In the center of 7 day wounds (data not shown), the deposition of both proteins was still discontinuous and the presence of collagen IV at the DEJ remained weak whereas that of laminin 5 had recovered the same intensity as in the adjacent RHS. Their deposition after 14 days (still weak for collagen IV) was almost continuous (data not shown). Collagen IV was also focally deposited into the fibroblast sheets of the RHS. Under the NE, the collagen IV deposition into the dermal matrix increased and extended farther toward the center with time (3 to 14 days). Collagen IV was present in the cytoplasm of fibroblasts with the same intensity, whatever their location in the wound or the RHS. Taken with the electron microscopic study, these immunofluorescence observations of the DEJ indicate that the deposition of individual proteins occurs earlier than its ultrastructural reorganization.
|
|
Fibrin and platelet-rich plasma delayed DEJ reformation
When fibrin and platelet-rich plasma were added onto the wounds for 3 days, the reepithelialization by keratinocytes was significantly accelerated (Fig. 2)
. Fibroblasts had begun to invade the fibrin matrix 3 days after wounding (Fig. 3c
, 3e
). However, the deposition of laminin 5 and collagen IV into the DEJ was delayed as shown by the faint labeling of collagen IV over fibrin (Fig. 5d
). Near the wound margin, fibrin was absent probably due to its degradation; laminin 5 and collagen IV were in the course of deposition at the DEJ, as in untreated wounds, suggesting that their presence resulted from the proximity between keratinocytes and fibroblasts. Laminin 5 and collagen IV began to be deposited into areas where fibrin or platelet-rich plasma matrices were thin and colonized by fibroblasts.
Progressive regeneration of a differentiated epidermis in the wound
K1/K10 (Fig. 6
a; summarized in Table 1
) in wounded or unwounded RHS, as in human skin, were expressed by the suprabasal keratinocytes, which had a differentiated morphology as seen by phase-contrast microscopy. From the margin of 3 day wounds up to the bMET, the K1/K10 labeling became progressively restricted to the most superficial keratinocytes, discontinuous, and less intense. In the 3 day NE, the unlabeled keratinocytes had an undifferentiated morphology and were located in the deep layers. The number of negative layers increased toward the MET, which was completely unlabeled. In the center of the reepithelialized 7 day wounds, the K1/K10 labeling was still thin, light, and discontinuous but was normalized after 14 days (data not shown).
|
Filaggrin (Fig. 6b
) and loricrin (Fig. 6c
) were expressed in RHS, as in human skin, by keratinocytes of the upper layers (summarized in Table 1
). Expressed with a granular pattern in the granular layer, filaggrin spread in the cornified layer to produce a diffuse labeling in the corneocytes’ cytoplasm. Loricrin was abundant and finely granular in cells of the granular layer, almost giving the impression of a uniform labeling. It was also deposited along the cytoplasmic membrane. Both cytoplasmic and membranous components of the labeling were maximally present in the upper granular cells and suddenly decreased in the cornified layer, where a brick wall-like staining resulted. Therefore, filaggrin and loricrin both showed a different pattern of expression in the granular compared with the cornified layer. They conserved their specific pattern of expression near the margin of 3 day wounds. In the intermediate region, filaggrin labeling decreased toward the wound center as fewer granular cells were present and as fewer granules were found per cell, although they had the same intensity. In the same region, loricrin was lightly expressed in the cytoplasm of isolated granular cells without membranous labeling. At day 7 and 14 after wounding, filaggrin and loricrin were expressed farther toward the wound center (data not shown).
The wounding procedure created a cut in the fibroblast sheets that served to locate the wound margin and also produced a sharp edge in the stratum corneum that was still recognizable 3 days later, indicating conservation of the cornified layer’s architecture (Figs. 3
, 5
, 6)
. Initially located at the wound margin (data not shown), the sharp edge in the cornified layer was displaced into the wound after 3 days of reepithelialization. Both the cornified and granular layers, identified by their typical filaggrin and loricrin expression pattern, had slid into the wound (Fig. 6b
, 6c
). In most samples, the edge of the granular layer was located closer to the wound margin than the cut in the cornified layer. Therefore, this observation indicates that the cornified layer had slid farther toward the wound center than the granular layer.
Keratinocyte migration in the migrating epidermal tongue
The MET was much thinner than the unwounded epidermis, being only a few cell layers thick (Fig. 3a
, d
, e
). At the tip of the NE, the keratinocyte immediately behind and above the foremost basal one was elongated. Its cytoplasm was partially in the suprabasal layer, but its forefront (toward the wound center) extended over the latter to make contact with the third fibroblast sheet (Fig. 3d
) or the fibrin matrix (Fig. 3e
). These elongated suprabasal keratinocytes migrated over their neighbor basal cells already in contact with the extracellular matrix.
Toward the wound center, the increasing expression of laminin 5 (Fig. 5a
) and collagen IV (Fig. 5e
) by keratinocytes, which adopted a characteristic spatial distribution in the MET, gave hints about their mode of migration. Easily observable in thick MET and particularly evident for laminin 5, the labeling was characteristically present in deep and superficial cells (Fig. 5a
, 5b
). This observation is consistent with cells crawling over each other at the tip of the MET. Between the deep and superficial labeling in thick MET, we observed columns of keratinocytes expressing laminin 5 (Fig. 5c
), which suggests that the positive superficial cells may originate from the underlying labeled basal keratinocytes.
Increased proliferation in the recent regions of the neoepidermis
We used the proliferation marker Ki67 (Fig. 7
) to evaluate the proliferative index in the reepithelializing basal keratinocytes. The proliferative index was calculated for each region of the 3 day NE and compared with that of the unaffected epidermis from the same sample. The proliferative indexes of the unwounded epidermis, close to or far from the wound margin, and of the NE near the wound margin were all similar. In contrast, the proliferative index was significantly increased in the intermediate region, bMET, and MET. Moreover, many low suprabasal cells were Ki67 positive in these regions, further showing their high proliferative activity. Many K1/K10 unlabeled suprabasal cells were also observed in these regions (Fig. 6)
. In summary, the newest NE regions had increased proliferative indexes.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In this report, we have established in vitro a new human wound-healing model that provides a unique tool to reexamine long-standing questions concerning wound reepithelialization. It presents the following advantages: a 3-dimensional and completely biological cellular environment; the major morphological and ultrastructural features of human skin; controlled and reproducible in vitro conditions; no obstacle for tissue processing; and obvious landmarks for locating the wound margins. In addition to its interest for fundamental studies of the wound-healing process, this model offers an accurate tool to screen for factors accelerating wound reepithelialization.
Acceleration of wound closure and potential use of the model for drug screening
By following macroscopically the wound closure, this human wound-healing model offers a tool to compare, under standardized conditions, the effect of various factors on the rate of reepithelialization. The reepithelialization of the wound-healing model occurred slightly faster than in vivo—i.e., 0.4 instead of 0.15–0.3 mm/day for human and pig wounds (7
, 31
, 32)
—likely because of increased humidity, which is known to have an accelerating effect (7)
. Indeed, the humidity in the incubator was high (95%), and probably also explains why wounds treated with saline or culture medium reepithelialized at a rate comparable to wounds dried regularly. After treatment with fibrin or platelet-rich plasma, the reepithelialization was significantly accelerated although the deposition of the DEJ proteins was retarded, as discussed below. Therefore, modulation of reepithelialization by the addition of exogenous factors can be studied and easily quantified in vitro with this wound-healing model, indicating its potential use for drug screening in the pharmaceutical industry.
Differentiation of the regenerating epidermis and reformation of the DEJ
The differentiated RHS presented features of human skin (33)
. The epidermis was ultrastructurally well organized with desmosomes, keratohyaline granules, lamellar bodies, and cornified envelopes. We took advantage of the differential expression of various markers during keratinocyte differentiation (34
35
36
37
38)
to follow the epidermal regeneration during healing. Before wounding, K1/K10 were expressed in all suprabasal layers. Filaggrin and loricrin had a granular appearance in the granular layer, but their change to a diffuse (filaggrin) and a membranous (loricrin) labeling in the cornified layer allowed us to easily distinguish these layers in the healing wound. As the NE advanced from the surrounding margins, it reproduced the centripetally progressive differentiation that occurs during reepithelialization (5
, 20)
. ‘Near the margin’ 3 days after wounding, the NE had recovered a differentiated phenotype comparable to that of the adjacent RHS. In contrast, keratinocytes of the MET, the newest region of the NE, were undifferentiated. In the NE, the progressive reexpression of K1/K10, followed by that of filaggrin and loricrin, is consistent with their sequential expression during keratinocyte differentiation in normal skin (33
, 35
, 39)
.
As in vivo, the DEJ of the unwounded RHS was ultrastructurally organized (33
, 40)
and composed of laminin 5 and collagen IV (41
, 42)
. Both keratinocytes and fibroblasts express collagen IV (43
, 44)
and probably contribute to its continual turnover at the DEJ of the unwounded RHS considering their light cytoplasmic labeling. In contrast, laminin 5 turnover appears to be assumed only by keratinocytes, as suggested by the observations of others (45)
. Contrary to fibroblasts, wound keratinocytes had an increased level of collagen IV cytoplasmic labeling, which suggests that, during DEJ regeneration, fibroblasts kept the same baseline level of expression whereas keratinocytes temporarily increased their synthesis. Collagen IV began to be deposited only under the bMET as in pig (1
, 31)
and human (46)
wounds. The deposition of laminin 5, which was initiated under the MET, therefore occurred before that of collagen IV as reported by others (44
45
46)
. The ultrastructural reorganization of the DEJ during healing, as after grafting (44)
, was delayed compared with the deposition of laminin 5 and collagen IV. The reorganization occurred by foci where we and others have observed precursors of hemidesmosomes (28
, 40
, 44)
. From these foci, an almost continuous and organized DEJ was regenerated.
The kinetics of DEJ reformation was modified in wounds treated with fibrin or platelet-rich plasma. Even though fibrin and platelet-rich plasma accelerated wound reepithelialization and migrating fibroblasts had begun to colonize these matrices after 3 days, the deposition of laminin 5 and collagen IV was retarded. A delay in the deposition of DEJ proteins was also reported when epidermal sheets were grafted with fibrin glue on athymic mice although the quality of the graft take (adhesion) was improved (47
, 48)
. The migration of keratinocytes occurs in vivo over a fibrin matrix composed of fibronectin (31)
, which is probably present in wounds treated with fibrin and platelet-rich plasma considering the nature of these blood derived matrices and the use of culture serum. Fibronectin might be responsible for the accelerated reepithelialization as it supports keratinocyte migration (49)
; consequently, it could be involved in the delay of stationary matrix deposition. Alternatively, the observation that laminin 5 and collagen IV were deposited when migrating fibroblasts came into close contact with keratinocytes supports the hypothesis that effective epithelial–mesenchymal interactions are crucial for deposition of the DEJ (50)
. After the disappearance of these matrices, laminin 5 and collagen IV were then in the course of deposition at the DEJ similarly to untreated wounds.
The increased proliferation would supply cells for migration and stratification
To better define the implication of proliferation during reepithelialization, we have investigated the distribution of proliferating keratinocytes along the NE. The newest regions of the 3 day NE were in hyperproliferation compared with the adjacent unaffected RHS and possessed many K1/K10 unlabeled suprabasal cells. In the epidermis immediately adjacent to 3 day wounds and in the NE near the margin, the proliferative indexes were at baseline level, indicating that the well-known transient burst of proliferation reported at the margin of recent wounds had already passed (7
, 18
19
20)
. Considering that the proliferation had normalized at the margin of 3 day wounds and that the latter were still reepithelializing, a new equilibrium must had been created in the MET, at the cellular scale, to ensure an effective reepithelialization based on the proliferation and migration of keratinocytes present locally in the wound. We think that the hyperproliferation measured in the MET should be responsible for the maintenance of the migrating cell mass, as others have suggested: mitosis within 24 h after cells had settled down on the wound surface (7)
, tritiated thymidine labeled cells at the tip of the NE (19)
, clusters of genetically marked keratinocytes just behind the leading edge of the NE (20)
, and slow reepithelialization of colchicine-treated wounds (19)
. In the bMET and intermediate region, the increased proliferation should generate cells for stratification and differentiation as suggested from observations made during the healing of pig wounds (7)
, after wound closure (19
20
21)
, and after removing suprabasal cells by tape stripping (51
, 52)
. Therefore, our detailed observations along the NE, which conciliate and complement previous data, show that keratinocytes hyperproliferate in the central regions to supply migrating and differentiating cells, after which they normalize their proliferation rate in the peripheral area that had reepithelialized earlier.
Two complementary mechanisms are simultaneously involved in wound reepithelialization
The features observed near the margin and in the MET suggest that two independent and complementary mechanisms are involved in wound reepithelialization.
The first mechanism involves the passive displacement of the superficial layers near the margin. It is based on the observation of precise landmarks: 1) the cut in the fibroblast sheets indicating the wound margin and 2) the sharp edge in the cornified layer, produced at the time of wounding, which is located within the wound instead of being at the wound margin. This indicates that the superficial layers had advanced over the wound. Since corneocytes are inert and tightly joined for effective barrier function (53)
, this displacement, which we have also recognized elsewhere (21)
, was likely caused by their passive sliding as a coherent sheet over the living layers of keratinocytes. Our observation that the cornified layer had advanced farther toward the wound center than the granular layer (distinguished by filaggrin and loricrin expression pattern) indicates that the passive sliding of these superficial layers results from a pushing rather than from a dragging force. This is consistent with observations made during mouse corneal healing (13)
. It is likely that the pushing force, coming from the adjacent unwounded epidermis, originates from the mitotic pressure (proliferating keratinocytes pushing suprabasal cells toward the surface). The absence of resistance from the wound side due to the absence of epidermis favors a lateral displacement toward the free space over the wound. This lateral migration force could also displace keratinocytes of the deep layers (20)
. The contribution of this passive displacement of the superficial layers to wound closure is particularly significant for small or incisional wounds; after 3 days, it represented 
30% of the 6 mm wounds studied. Such a phenomenon would be difficult to demonstrate in vivo because of the presence of scabs in the wound (7
, 52)
that limit this lateral movement from the margins and because of wound contraction (6)
, which is particularly important in rodents.
The second mechanism takes place in the MET where keratinocytes migrate individually over each other. We showed that suprabasal keratinocytes at the tip of the MET were elongated over and beyond the basal foremost cell to reach the connective tissue and thus the basal layer, in agreement with the leap-frog model of migration (7
, 15
16
17)
. Consistent with their increased expression of laminin 5 during migration (45
, 46
, 54
, 55)
, keratinocytes highly expressed it in the superficial layers at the tip of the MET. Laminin 5 was expressed by keratinocytes in both the deep and superficial layers. We have recognized a similar pattern of expression in studies conducted in vivo by others for laminin 5 (45)
, integrins 
3, 5, ß1, and possibly v (32)
, and the urokinase-type plasminogen activator (56)
. Their characteristic and common spatial distribution in the MET is probably of functional significance. We showed that suprabasal keratinocytes at the tip of the MET had a migratory phenotype and expressed laminin 5. Once in contact with the connective tissue, they soon become anchored (15)
, and we have observed they begin to deposit laminin 5 under the MET. The newly synthesized laminin 5 may play a crucial role in migration since it supports keratinocyte motility, whereas its proteolytic cleavage by plasmin induces the establishment of hemidesmosomes and cell anchorage to the extracellular matrix in an integrin-dependent manner (54
, 55
, 57
, 58)
. Therefore, the expression of laminin 5 by keratinocytes plays two successive functions: stimulating migration and then cell anchorage. The induction of the migratory phenotype and laminin 5 expression in the superficial layer of the MET may come from 1) the altered keratinocyte polarity, since these cells are not completely surrounded by other cells or 2) the columns of labeled cells observed in the midlayers, which may be generated by the proliferation of newly anchored keratinocytes. The series of events occurring in the MET, now better understood, ensures an effective reepithelialization in the wound center based on the proliferation and migration of keratinocytes present locally once the wound margin has become distant (7)
.
The completely biological human wound-healing model we developed has allowed us to bring a new perspective to long-standing questions concerning the mechanisms of wound reepithelialization. It constitutes another example of the various applications of human tissues reconstructed in vitro by tissue engineering (24
25
26
27
28
, 50
, 59
60
61
62
63)
. The conclusions we have elaborated can be enlarged to other pluristratified epithelia and may have practical considerations. Although the superficial layers slide passively over a limited distance, this may represent a significant proportion of the surface area, especially in small wounds, due to the circumferential effect from the wound periphery. As these layers are endowed with protective features, their displacement over the wound surface contributes to the rapid regeneration of a functional barrier. Scab formation, which occurs in a dry environment, represents a probable limitation for the sliding of the superficial layers and justifies keeping the wounds moist with appropriate dressings. To sustain an effective reepithelialization from de novo, an equilibrium should be reached between the proliferation, migration, stratification, and differentiation of keratinocytes present locally in the wound. After proliferation, the fate of keratinocytes can be conceived in the following way. In the MET, keratinocytes stratify to produce a mass of cells in the superficial compartment, ready to migrate over the wound bed. In the bMET and toward the wound margin, keratinocytes stratify to regenerate progressively a differentiated epidermis endowed with barrier properties. Therefore, the two mechanisms of wound reepithelialization act in a complementary fashion to efficiently regenerate a functional epidermal barrier over the shortest possible time period.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 29, 2001.
Revision received July 24, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3 laminin subunit. 6ß4 and 3ß1 integrin coordinately regulate wound healing in cultured epithelial cells and in the skin. J. Cell Sci. 112,2615-2629[Abstract]
This article has been cited by other articles:
|
|
 |
|
  D. Larouche, X. Tong, J. Fradette, P. A. Coulombe, and L. Germain Vibrissa hair bulge houses two populations of skin epithelial stem cells distinct by their keratin profile FASEB J, May 1, 2008; 22(5): 1404 - 1415. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Carrier, A. Deschambeault, M. Talbot, C. J. Giasson, F. A. Auger, S. L. Guerin, and L. Germain Characterization of Wound Reepithelialization Using a New Human Tissue-Engineered Corneal Wound Healing Model Invest. Ophthalmol. Vis. Sci., April 1, 2008; 49(4): 1376 - 1385. |
