Regeneration of a wound defect is a complex process that involves interaction of epithelium, trophic apparatus, fibroblasts, and inflammatory infiltration [1]. Healing includes stages of inflammation, regeneration and reorganization; at each of these stages, the mentioned interacting components have their own morphological scar features [2]. Primary intention healing requires the fastest possible transition from the stage of inflammation to the regeneration stage. Secondary intention healing, which results in a scar, typically has the inflammatory stage dominating the process [3].

The current approach to management of the oral mucosa wound defects implies use of covering materials that protect the wound surface from re-traumatization, which adds urgency to the task of developing new wound dressings that will not only protect the wound but also reduce the severity of inflammation and accelerate regeneration [4].

The polymer piezoelectric membrane tested in the context of this study was made in the Laboratory of Hybrid Materials of the National Research Tomsk Polytechnic University. In addition to piezoelectric properties, it was modified with copper ions, which have proven antimicrobial and anti-inflammatory properties [5].

The purpose of this study was to explore morphological features of regeneration of oral mucosa with an experimental wound covered with polymer piezoelectric membranes.

MATERIALS AND METHODS

The experiment involved 45 male Wistar rats bred in the vivarium of the Central Research Laboratory of the Siberian State Medical University. The animals were kept under standard vivarium conditions; their rations were limited for a day post surgery. The rats were divided into three groups: group 1, experimental (n = 15), in which the animals had the wound defect left open as prescribed by the standard oral cavity wound treatment protocol; group 2, experimental (n = 15), where the wound defect was covered with a copper-modified polymer membrane based on vinylidene fluoride with tetrafluoroethylene; group 3, control (n = 15), comprised of rats with intact mucosa. The animals were kept under standard vivarium conditions.

Before wound infliction, the animals received 0.3 mg Zoletil intramuscularly that induced narcosis. After sanitizing the surgical field with a 2% chlorhexidine solution, we excised a 7–4 mm buccal mucosa flap. Next, a polymer membrane was stitched onto the resulting wound with simple interrupted sutures along the edge.

The animals were withdrawn from the experiment through hypoxia in a CO₂ chamber on the 3^rd, 7^th, and 12^th days of the study. After withdrawal, we excised buccal mucosa at the wound defect site.

Preparation of histological specimens followed the generally accepted routine; the slides were examined in the Observer D1 light microscope (Karl Zeuss; Germany) with the AxioCam ICc5 camera (Karl Zeuss; Germany). For the purpose, after deparaffinization, the sections were stained with hematoxylineosin as per the standard procedure.

For immunohistochemistry, we deparaffinized the previously prepared serial paraffin sections 4-6 µm thick and then stained them with rabbit recombinant polyclonal VEGF antibodies and IgG isotype S-100 (Abcam; USA). The intensity of immunohistochemical staining was assessed on a four-point scale: 0 — no staining, 1 — weak staining, 2 — moderate staining, 3 — strong staining, 4 — very strong staining.  The calculation formula was as follows:

Histochemical index (H-score) = ∑ P(i) × i,

where i is the intensity of staining in points from 0 to 4, P(i) — percentage of differently stained cells (by intensity of staining).

We did the counting in three cohorts of 100 cells in different fields of view (lens ×40).

To examine the samples with an electron microscope, we put them in a 2.5% glutaraldehyde solution on the 0.2 M cacodylate buffer (1:9) for fixation and then postfixed the samples in a 1% OsO₄ solution in a refrigerator for 4 hours. The next steps involved dehydration and pouring into a mixture of epon and araldite M.

Ultrathin slides were prepared in the LKB-5 ultratome (BROMMA; Sweden), counterstained with uranyl acetate and lead citrate, and examined using a JEM-1400 CX electron microscope (JEOL; Japan).

Morphometry was performed based on the classical methods of stereometry. We established the thickness of the epithelial layer, the quantitative density of fibroblasts, the specific area of loose and dense fibrous connective tissue, that of granulation tissue and inflammatory infiltration. The software used for the purpose was the Axio Vision imaging system (Karl Zeuss; Germany) and ImageJ, version 1.52u (National Institute of Public Health; USA).

Ultrathin slides allowed us to study ultrastructure of the cells of epithelial layer, trophic apparatus, and fibroblasts.

The data were processed with Statistica 10.0 (IBM; USA). To verify distribution in the statistical hypotheses, we used the Kolmogorov–Smirnov test. The results were processed with the help of methods of descriptive and nonparametric statistics. The studied parameters were described as a median and quartiles, M (Q₁; Q₃). Comparing independent samples, we used the Kruskal-Wallis test with a median test, for paired comparisons — Wilcoxon test. The differences were considered significant at p < 0.05.

RESULTS

On the 3^rd day of the study, we registered marginal epithelialization of the wound in experimental group 1 was observed, while in group 2 the defect was completely covered by the epithelium, and there were acantholytic blisters there (fig. 1A).

At the ultrastructural level, samples from both groups had elongated width-wise basal cells, and it was not possible to identify the apical-basal polarity. In group 2, there were signs of high proliferative activity, and some cells were undergoing mitosis (fig. 1B).

We discovered a regeneration area filled with granulation tissue with numerous thin-walled vessels (fig. 1C). The specific area of granulation tissue in the experimental group 2 was 1.4 times higher than in the experimental group 1, the difference being significant (p = 0.033) (see table). There was an extensive zone of inflammatory cell infiltration separating healthy tissues from the wound defect; in group 2, is was 1.4 times smaller than in group 1, this difference being significant, too (p = 0.017). The said zone consisted of neutrophils, macrophages, plasma cells, lymphocytes, and eosinophils (see table) (fig. 1D).

Quantitatively, there were significantly more fibroblasts per 1 mm² of the section in group 2 (p = 0.035) (see table). In addition, separate bundles of connective tissue fibers were found in the animals of this group (fig. 1A).

We discovered newly formed thin-walled granulation tissue vessels in the area of the wound defect; their endotheliocytes actively expressed VEGF (fig. 1E), and the calculated H-score in group 2 was 1.4 and 4.7 times higher (significant difference) than in groups 1 and 3 (p = 0.029, p = 0.019) (see table).

Endothelial lining and basement membrane of the granulation tissue capillaries were thin, and the interendothelial spaces were enlarged. The endotheliocytes were poor in organelles. In group 2, we identified a large number of micropinocytic vesicles and numerous microvilli in the endotheliocytes near the luminal edge of the vessels.

On the 7th day of the study, epithelium completely covered the wound defect in both experimental groups. The thickness of the epithelial layer in group 1 was significantly lower than that in groups 3 and 2 (see table); there were clear active acantholytic processes (fig. 2A).

In group 2, epithelium had less pronounced pathological changes than in group 1. The surface began to assume a typical relief with papillae, yet there still were spots of acanthosis and acantholysis. The epithelium was 2.3 times less thick than in the control group (significant difference, p = 0.023) but twice as thick as in group 1, where the animals had no membranes covering the defect (significant difference, p = 0.019) (see table). Basal cells gradually assumed the typical shape elongated height-wise and apical-basal polarity, and there appeared hemidesmosomes restoring contacts with the basement membrane.

In both experimental groups, spots of loose fibrous connective tissue were identified in the lamina propria of buccal mucosa by the wound. In group 2, the specific area thereof was 4.5 times greater (significant difference, p = 0.041) than in group 1, but 2.6 times smaller than in the control group (significant difference, p = 0.034).

In the experimental group 2, the quantitative density of fibroblasts was 1.35 times higher than in the experimental group 1 (p = 0.041). Also, the prevailing type of fibroblasts in group 2 were the large differentiated branched fibroblasts with high synthetic activity. Their plasma membrane sprouted numerous outgrowths. Bundles of connective tissue fibers in different planes were found around the cells (fig. 2B). As for group 1, there still prevailed young fibroblasts with ultrastructure practically unchanged since the 3^rd day of the experiment.

The newly formed vessels in group 2 had the VEGF expression 1.7 and 3.6 times higher than in group 1 and in the control group, respectively (significant difference, p = 0.022 and p = 0.015) (see table).

At the ultrastructural level, endotheliocytes in group 1 still showed signs of transcapillary metabolism disorders, which was not registered in the experimental group 2 (fig. 2B), where the basement membrane of the vessels became continuous and uniform in thickness.

In both experimental groups, we discovered peripheral nerves near blood vessels by the wound defect's limits, the dendrons thereof having single edematous mitochondria with destruction of cristae.  As shown on fig. 2D, perineural and endoneural space were edematous (fig. 2D).

On the 12^th day of the study, the animals of group 2 had the epithelial layer as thick as it originally was, while in group 1 that layer was significantly thinner than when the mucosa was intact (see table). There were no pathological processes registered in the epithelium in group 2, and the underlying connective tissue protruded into the epithelium forming pronounced papillae with microvasculature vessels (fig. 3A). In group 1, on the contrary, we saw signs of acantholysis, which, at the ultrastructural level, caused appearance of vacuoles in the intercellular space and, with cells proliferating, acanthosis in the spinous layer (fig. 3B).

On the 12^th day of the study, granulation tissue was replaced with young loose fibrous connective tissue in both experimental groups (fig. 3A). However, the specific area of this tissues reached the control value thresholds only in the experimental group 2. In group 1, we found interstitial edema between connective tissue fibers, and there were cicatricial foci based on the dense fibrous connective tissue (fig. 3B). Table below shows the values; the maximum specific area of such connective tissue was registered in the experimental group 1.

The quantitative density of fibroblasts in group 2 equaled the respective control value; it was 6.9 times smaller than in group 1 (significant difference, p = 0.032). The predominant cells in group 2 were fibrocytes, mature and functionally inactive. There was no extracellular edema around them, but the clearly organized collagen fibers were easily identifiable. At the considered time point, the predominant cells in group 1 were still the branched cells with developed synthesis organelles and dispersed chromatin. 

In the experimental group 2 we registered mature, wellformed microvasculature vessels with no signs of sludge, stasis, and thrombosis. The vessels were surrounded with structured connective tissue fibers; there was no perivascular edema. The basement membrane was continuous, of uniform thickness. The quantity of organelles was sufficient, they had a typical structure. There was an active transcapillary exchange under way. In group 1, the capillary basement membrane was thin; we registered few microvilli and micropinocytic vesicles at endotheliocytes, with clearly seen perivascular edema.

Compared to the indicators recorded on the 7th day, expression of the VEGF decreased in both experimental groups. In the experimental group 1, the VEGF H-score was 1.45 times smaller than the control values (p = 0.026) (fig. 3C, D).

Peripheral nerves with unmyelinated nerve fibers in wound defect area were registered only in the experimental group 2 (fig. 3D). Despite this, immunohistochemistry did not reveal expression of the S-100 marker, which the indicated the nerves began to restore (fig. 3E).

DISCUSSION

With this study, we have demonstrated the key morphological aspects of the wounded oral mucosa tissue regeneration under a copper-modified polymer piezoelectric membrane.

On the 3^rd day of the experiment, in group 2, where the wound was covered with the membrane, the wound was fully covered with a thin layer of epithelium, while in group 1, the animals of which had no membrane on their wounds, only marginal epithelization was observed. The process of epithelialization translated into changes at the ultrastructural level that enabled migration of epithelial cells from the wound's edges to its center [6]. Rørth reported preservation of single desmosomal contacts in his works and explained their presence by the need for coordinated collective migration of the epithelium [7].

Basal layer cells were proliferating intensively only in the experimental group 2. In group 1, we have not detected ultrastructural signs of high proliferative activity; lack of such signs on the part of the basal layer cells may be explained by the migrating cells' inability to undergo mitosis before the wound defect is completely covered by the epithelium layer caused by the decreased content of G1/S-phase cyclins and higher activity of cyclin-dependent kinase [8].

For the first three days, at the first stage of wound healing, inflammatory reaction was the predominant one. It was aimed at fencing healthy tissue from the defect (with its necrotic tissue), microorganisms and elements of primary contamination, as well as removal of these pathological products, elimination of the consequences of damage, activation of cytokines and growth factors [9]. In this connection, the specific area of inflammatory infiltration increased in both experimental groups.

In parallel with the inflammatory response there appeared new granulation tissue pierced with many vessels. In the experimental group 2, with the wound dressed, the ratio of the specific area of granulation tissue and inflammatory infiltration shifted towards the first component thereof, while in group 1, where the animals had no cover on the defects, the ratio prevalence was the opposite, which indicated predominance of inflammatory processes over regenerative ones.

The development of granulation tissue ensured rejection of the dead substrate, created a barrier preventing the spread of microorganisms, and formed the basis for new connective tissue developing at subsequent stages of wound regeneration [10].

Neoangiogenesis was the basic mechanism of wound healing [11]. Capillaries were forming actively in the granulation tissue in both experimental groups. Neoangiogenesis and inflammatory response stimulated endothelium and the inflammatory infiltrate cells to express pro-angiogenic molecules — vascular endothelial growth factor (VEGF) and chemokines needed for capillary growth. Previously, researchers have noted that VEGF in the area of a wound defect undergoing regeneration supported the inflammatory response and increased vascular permeability, which contributed to the swelling of the surrounding tissues [12]. In our study, the expression of VEGF was increasing in both experimental groups, especially in group 2, which signaled of a more intense vascularization in the absence of exposure to aggressive factors present in the oral cavity.

On the 7^th day of the study, proliferation became the dominant process, with fibroblasts responsible for collagen synthesis and reduction of the wound area playing the key role therein [13]. Fibroblasts secreted an extracellular matrix replacing the fibrin matrix [14], which boosted the synthesis, as registered at the ultrastructural level in the experimental group 2. Their presence in the wound defect area was connected with a significant increase of the specific area of loose fibrous connective tissue in comparison with experimental group 1, where, according to the TEM data, young fibroblasts with low synthetic activity mainly appeared.

On the 7^th day of the experiment, the specific area of granulation tissue decreased, and the newly formed vessels underwent changes at the ultrastructural level, as evidenced by the results of immunohistochemistry. The expression VEGF decreased in group 1 significantly more than in group 2, which subsequently hindered blood supply in the wound defect area [15].

The characteristic features of the proliferation stage were differentiation of epithelial cells and thickening of layers of the epithelium. At the ultrastructural level, in group 2, the apicalbasal polarity of the basal layer, desmosomal contacts between cells, and hemidesmosomes with the basement membrane were restoring. We have also seen signs of acceleration of proliferation. In both experimental groups, there were registered pathological changes in the form of acanthosis and acantholysis; morphologically, they manifested as tissue detritus in the intercellular space and thickening of the spinous layer. These pathological features were more pronounced in group 2.

Cicatricial reorganization was the next stage of wound regeneration; it occurred on the 12^th day of the study. In contract with the experimental group 1, in group 2 the specific area of loose fibrous connective tissue was greater than that of scar tissue, and the predominant cells there were the synthetically inactive fibrocytes.

In both experimental groups the number of microvasculature vessels decreased on the 12^th day. Previous reports link this regression to selective apoptosis happening against the background of the increasing production of antiangiogenic factors and decreasing expression of proangiogenic factors, such as VEGF [12]. As a confirmation of these findings, in the experimental group 1 we observed that VEGF expression was significantly lower than peculiar to intact mucosa. This signaled of poor blood supply at the wound defect site, which is one of the main causes of cicatricial deformity [16]. In group 2, the expression of vascular markers was higher than the control values.

Using an electron microscope, on the 12^th day of the study we saw peripheral nerves with unmyelinated nerve fibers in the experimental group 2; while such were not found in group 1, there, along the edges of the wound defect, we discovered peripheral nerves with signs of perineural and endoneural edema the axons of which had mitochondria with destroyed cristae. The tests for expression of protein of the S-100 peripheral nerves returned negative in both groups, which indicated the nerve fibers began to grow into the wound defect area.

In the experimental group 2, where the wounds were covered with membranes, complete restoration the epithelium (both thickness and number of layers) occurred because it was protected from microorganisms. When the integrity of the epithelial layer was violated, microorganisms penetrated to the bottom of the open wound defect and excreted exotoxins that affected the wound both from the inside and the outside, causing apoptosis of epithelial cells and disrupting their proliferation and migration, such course of events associated with collective regulation of the bacterial gene expression in the biofilm, which reinforced their resistance and ability to colonize [17]. In addition, metabolic products of microorganisms boosted production of pro-inflammatory mediators in the epitheliocytes through toll-like receptors [18].  As a result, in group 1 we saw epithelium growing thinner and losing layers, while acanthosis and acantholysis, the pathological changes, persisted.

CONCLUSIONS

The polymer membrane helped form an epithelial layer completely covering the wound defect on the 3^rd day of the study and ensured absence of pathological changes on the 12th day thereof. In the group that had no such membrane, epithelization of the wound occurred only on the 7th day, and pathological changes (acanthus and acantholysis) were seen on the 12^th day.

The membrane boosted replacement of granulation tissue with loose fibrous connective tissue and decreased the severity of inflammatory infiltration and cicatricial changes. In the group without membranes, by the end of the study, the wound defect was mainly filled with dense fibrous connective tissue, and there was a pronounced inflammatory infiltration observed.

The polymer membrane boosted neoangiogenesis, which translated into a more intense expression of the VEGF (compared to the group without the membrane) registered at each time point of the study.

Overall, we have shown that use of a polymer piezoelectric membrane had a positive effect on the restoration of the oral mucosa tissues in the wound defect area, which manifested as an accelerated restoration of the epithelium, trophic apparatus and connective tissue component. Further, this helps optimize the process of treatment of wounds of this localization and improve quality of life of the patients.