Differential Evaluation of Excisional Non-occluded Wound Healing in db/db Mice

Abstract

The full-thickness wound in the genetically diabetic (db/db) mouse is a commonly used model of impaired wound healing. We investigated delayed healing of non-occluded, excisional, full-thickness, dermal wounds in db/db mice in comparison to their normal littermate controls and refined methods for monitoring skin wound re-epithelialization, contraction, granulation tissue formation, and inflammation. We have confirmed with a computer-assisted planimetry method the results of previous studies showing that healing of non-occluded full excision wounds in db/db mice does not occur by contraction as much as in healthy mice. In addition, we have developed separate histological methods for the assessment of re-epithelialization, contraction, granulation tissue (mature, immature, fibrosis), and inflammation (lipogranulomas, secondary, nonspecific). Using a new approach to histological assessment, we have shown that wound closure in db/db mice is delayed owing to: (1) delayed granulation tissue maturation; (2) ‘‘laced,’’ widely distributed granulation tissue around fat lobules; and (3) obstruction by lipogranulomas, whereas the rate of re-epithelialization seems to be the same as in C57Bl/6 mice. This methodology should permit a more precise differentiation of effects of novel therapeutic agents on the wound healing process in db/db mice.

Introduction

Chronic open foot ulcers are a common problem in diabetes and represent a major concern, from both a quality of life and an economic standpoint (Brem et al. 2004; Gallucci et al. 2000; Livant et al. 2000; Pierce 2001; Quatresooz et al. 2003). In these patients, delayed wound healing is characterized by inhibition of the initial inflammatory response, delayed macrophage infiltration, angiogenesis, fibroplasia, and reparative collagen accumulation (Cohen et al. 2001; Pierce 2001).

The wound healing process is a complex cascade of inflammation, granulation tissue formation, re-epithelialization and angiogenesis initiated by the release of various growth factors, cytokines, and low-molecular-weight compounds from the serum and degranulating platelets. Inflammatory cells, namely neutrophils, monocytes, and lymphocytes, produce a wide variety of proteinases and reactive oxygen species as a defense against contaminating microorganisms. They are also an important source of growth factors and cytokines, which initiate the proliferative phase of wound repair that starts with the migration and proliferation of keratinocytes, forming a new epithelial layer, and dermal fibroblasts. Wound fibroblasts transform into myofibroblasts, which play a major role in wound contraction. At the same time, massive angiogenesis leads to the formation of new blood vessels. The resulting wound connective tissue is known as granulation, which finally transforms into mature scar by continued collagen synthesis (Falanga 2005; Werner and Grose 2003).

Genetically diabetic mice (db/db mice) are often used as an animal model for wound-healing studies, since wound healing in these animals is markedly delayed when compared with non-diabetic littermates (DiPietro and Burns 2003; Greenhalgh 2005; LeGrand 1998; Michaels et al. 2007; Senter et al. 1995; Sullivan et al. 2004). Although the subject of numerous investigations, the mechanism underlying the healing impairment in diabetic mice is still not known. Several studies reported in the literature have elucidated the function of various genes in the healing process, revealing the role of growth factors, cytokines and their downstream effectors in wound repair. The expression of platelet-derived growth factor (PDGFs) and its receptors is reduced in db/db mice (Beer et al. 1997), as well as that of fibroblast growth factors (FGF) 1 and 2 (Werner et al. 1994) and vascular endothelial growth factor (VEGF) (Lerman et al. 2003). Also, Wetzler et al. (2000) showed that prolonged persistence of neutrophils and macrophages in the wounds of db/db mice correlates with a large and sustained induction of the CC chemokine, macrophage chemoattractant protein (MCP-1/CCL2), and macrophage inflammatory protein 2 (MIP-2). In the same study, the expression of the pro-inflammatory cytokines interleukin (IL)-1α , IL-1β , IL-6 and tumor necrosis factor (TNF)-α was prolonged in genetically diabetic db/db mice. As a result, healing impairment in db/db mice is characterized by delayed cellular infiltration and granulation tissue formation, reduced angiogenesis, and decreased collagen and its organization (DiPietro and Burns 2003; LeGrand 1998; Senter et al. 1995). In contrast to normal mice, contraction is a minor mechanism of wound healing in db/db mice. Instead, re-epithelialization plays a major role in wound closure (DiPietro and Burns 2003; Greenhalgh et al. 1990; Michaels et al. 2007; Senter et al. 1995).

In previous methodological studies of excision wound healing in db/db mice, wound area, re-epithelialization, and granulation were all evaluated by planimetry (Galiano et al. 2004; Michaels et al. 2007; Obara et al. 2005; Senter et al. 1995), and the quantity of granulation tissue was also assessed by scoring (Senter et al. 1995). The present study sought to confirm the use of planimetry for wound area and to extend histological analysis to provide methods for a better differentiation of the types of tissue formed in wounds in db/db mice in comparison to normal mice. For planimetry we adapted existing computerized technologies, as already described in previously performed experiments on alloxan-hyperglycemic rats (Ševeljević-Jaran et al. 2006). Histological evaluation involved morphometric evaluation of re-epithelialization, wound contraction, maturity of granulation tissue, and inflammation.

Methods

Animals

Genetically diabetic female C57BL/KsJ db+ /db+ mice (38–45 g) and female C57Bl/6 control mice, all six weeks old, were obtained from Charles River Laboratories, Belgium. The diabetic mice were markedly hyperglycemic (mean blood glucose = 527 ± 9 mg/dL) compared with nondiabetic animals (mean blood glucose = 205 ± 9 mg/dL). The hyperglycemia was clinically manifested as polydypsia, polyuria, and glycosuria.

Animals were allowed to acclimatize for ten days, marked, and individually housed for the experiments. Mice were kept on wire mesh floors with a 3 cm thick layer of dust-free, irradiated corn cob grit (Scobis Due, Mucedola, Italy) bedding and maintained under standard laboratory conditions (temperature 23° C–24° C, relative humidity 60% ± 5%, about 15 air changes per hour, controlled photoperiod with white light on at 7:00 AM, and red light on at 7:00 PM). Food (food pellets, Mucedola, Italy) and water were provided ad libitum.

Experimental Plan and Procedure

To investigate wound healing dynamics, twenty diabetic db/db mice and twenty nondiabetic C57Bl/6 mice were each divided into four groups of five animals designated for future wound tissue sampling at four different time points (one group each per time point). On day 1, all mice were anesthetized by inhalation of isoflurane (Forane, isoflurane, inhalation anaesthetic, Abbott Laboratories, England), delivered in an anesthesia induction chamber (Stoelting Co., Wood Dale, IL, USA). Gas scavenging was provided using the Fluovac 240V system (International Market Supply, England). Subsequently, anesthesia was maintained using a mask (Stoelting Co., Wood Dale, IL, USA), and the pedal reflex response was checked at intervals. In the fully anesthetized animal, the interscapular region was surgically close shaved (Contura cordless clipper, International Market Supply, England), and excess loose hair was brushed off. Twenty-four hours later (Day 0), animals were anesthetized as on day 1 and the shaved region disinfected (Pursept-A, Merz Hygiene, Germany). Using strictly aseptic procedures, a single full-thickness excisional wound 8 mm in diameter was made midline in the shaved region of each animal, with a sterile, disposable biopsy punch (Stiefel Laboratories Ltd., Ireland), exposing the underlying fascia muscularis. The clearly visible wound was next photographed. Animals were returned to their cages with some form of distractive enrichment, like food pellet, and allowed to recover from the anesthesia. On days 3, 6, 9, and 13, one group of each strain was anesthetized and photographed as before and exsanguinated by puncturing the v. jugularis and a. carotis communis. The wounds with surrounding healthy tissue were excised post mortem, clamped onto a strip of hard paper, and stored in 10% formalin for future histopathological assessment.

All procedures on animals were performed in accordance with (1) the EEC Council Directive 86/609 of November 24, 1986 on the approximation of laws, regulations, and administrative provisions of the Member States regarding the protection of animals used for experimental and other scientific purposes; and (2) regional legislative equivalent, animal protection law, revised edition published December 13, 2006, in Official Gazette, Issue No. 135.

Processing and Analysis of Wound Images

Serial standard 2D photographs of each wound were made immediately after wounding (day 0) and on days 1, 3, 6, 9, and 13 while animals were under isoflurane anesthesia.

A digital camera (Olympus C-2040 Zoom, Olympus Optical Co., Ltd., Japan) was used, facing down, statically mounted on a tripod (Gruppo Manfrotto, Italy), 20 cm above the mouse, permitting direct comparison of wounds between individual mice. The mouse was placed flat on its belly, directly under the camera, front legs and neck slightly stretched out, on white paper, with the head and shoulders (i.e., wound-bearing area) visible on the display (not the viewfinder). A paper tag was used to indicate the number of the photographic session and the number of the mouse. The camera was operated with a 128 MB Smart Media Card (256 shots available) using a mains adaptor, set to the macro mode (range 0.2–0.8 m), suitable for taking still pictures and employing the high quality (HQ, 1200 x 1600 pixels) recording mode. The zoom was set to ensure the highest possible magnification of the wound, and the focus was locked. A USB Port Digital Film Reader (Lexar Media Inc., Fremont, CA, USA) was used to transfer the data from the camera to a computer.

Digital images were processed using the LEICA QWin Image Processing and Analysis System (Leica Microsystems Ltd., UK); the Manual tool bar allowed manual tracing of the wound margins (Figure 1).

The Interactive Measurements window provides a Draw Area option, which, when chosen, displays the manual results for the measured object area as 3.22 μm pixels. Each result was copied into an Excel file, previously created for this purpose, that calculated the average of the area measured for each object. Wound closure was expressed as the relative change in wound area in comparison to day 0 and was calculated by dividing the wound areas measured on days 1, 3, 6, 9, and 13 by the wound area measured on day 0. In this way, wound areas on day 0 in all animals were standardized and equal to 1.

Tissue Processing and Histological Evaluation

Tissue was fixed in formalin, cut through the center of the wound, and embedded in paraffin, ‘‘sideways-on.’’ Slides were stained with hematoxylineosin, Mallory, or van Gieson. The edge of the wound was defined by the acellular connective tissue of the dermis. The wound bed was defined by adipose tissue in db/db mice and by large arteries, nerves, and/or adipose cells in C57Bl/6 mice.

Morphometrical evaluation of re-epithelialization and contraction was performed on a Leica DMRXA microscope using LEICA QWin Image Processing and Analysis System (Leica Microsystems Ltd., UK). The magnification used was 50X. For each wound the percentage of re-epithelialization was calculated as follows:

$$(\text{distance\hspace{0.17em}covered\hspace{0.17em}by}\hspace{0.17em}\text{epithelium}/\text{width\hspace{0.17em}of\hspace{0.17em}wound\hspace{0.17em}bed})\times 100$$

Wound contraction on each day was assessed as histological wound diameter in 3.22 μm pixels.

Granulation tissue was described as:

• immature granulation tissue: loose granulation tissue (macrophages, fibroblasts) with emerging vessels (Figure 2A)

• mature granulation tissue: fibroblasts and sparse extracellular matrix proteins forming layers, vessels running perpendicular (Figure 2B)

• fibrosis: extracellular matrix proteins (mainly collagen) dominating the granulation tissue, fewer fibro-blasts and vessels (Figure 2C)

The score for each wound was given in a semiquantitative manner with a score of 1 to 3 for each of the three granulation tissue categories (Figure 2D):

1. Wound bed partially covered with granulation tissue

2. Thin granulation tissue over the whole wound bed

3. Thick granulation tissue over the whole wound bed

Inflammation was scored as the presence (1) or absence (0) of each of the following types of inflammatory response:

• lipogranuloma: granulomatous inflammation around the characteristically shaped cholesterol crystals or fat cells (Figure 3)

• secondary inflammation: wound-unrelated inflammation characterized by accumulation of polymorphonuclear leukocytes

• nonspecific inflammation: neutrophils and sometimes eosinophils diffusely scattered throughout granulation tissue, polymorphonuclear leukocytes forming crust were excluded

Data Analysis and Statistical Evaluation

Data calculations for all parameters were performed using Microsoft Excel software. GraphPad InStat v. 3.06 software (GraphPad Software Inc., San Diego, CA, USA) was used for statistical analysis. Wound closure in nondiabetic C57Bl/6 and db/db mice, assessed by planimetry, was tabulated as mean ± standard error of mean. All histological parameters, except wound diameter, in nondiabetic C57Bl/6 and db/db mice were tabulated as medians. However, since the rates of re-epithelialization and granulation tissue formation were presented as scores, standard deviations were determined only for the values of histological wound diameter measured as 3.22 μm pixels and tabulated as average. The Mann-Whitney U test was used for all comparisons between groups, at each time point.

In all cases, the level of significance was p < .05. The values were presented graphically to demonstrate the healing dynamics with time.

Results

Closure of Wound Area

The computer-assisted planimetry method revealed that the excision of skin resulted in retraction of wounds within the first day in all animals, and retraction was significantly greater in C57Bl/6 mice (Figure 4). However, in db/db mice wound retraction continued until day 3, whereas in C57Bl/6 mice at this time, wound closure had already started. On day 6, relative wound size in C57Bl/6 was the same as in db/db mice, whereas on day 9 in C57Bl/6 mice it was significantly smaller than in db/db mice, indicating significantly faster wound closure in nondiabetic C57Bl/6 mice.

Contraction and Re-epithelialization

Histological wound diameter in C57Bl/6 mice was significantly smaller than in db/db mice on days 6 and 13, indicating a significantly faster rate of wound contraction in normoglycemic mice in comparison to db/db mice (Figure 5A). The mean re-epithelialization rate in terms of proportion of wound area covered in db/db and C57Bl/6 mice was not significantly different between the two groups at any time point (Figures 5B and 7). In C57Bl/6 mice, re-epithelialization was a continuous process that ended thirteen days after wounding. In contrast to normoglycemic mice, the re-epithelialization rate in db/db mice was slightly slower in the first six days after wounding, but full wound re-epithelialization was achieved nine days after wounding.

Granulation Tissue

Immature granulation tissue, consisting of macrophages, lymphocytes, and plump fibroblasts accompanied by early vascular sprouting within edematous fluid, was well formed and covered almost the whole wound bed in the majority of the C57Bl/6 mice three days after the wounding, whereas in db/db mice it was just emerging from the wound edges (Figures 6A and 7). This finding is in accordance with literature data (Senter et al. 1995). By six days after wounding, the whole wound bed was covered with mature granulation tissue in C57Bl/6 mice (Figures 6B and 7). Transformation of ‘‘immature’’ granulation tissue into ‘‘mature’’ granulation tissue was characterized by elongation of fibroblasts, early deposition of collagen fibers, and angiogenesis. Mature granulation tissue appeared in db/db mice with a three-day delay. The wound healing process was complete thirteen days after wounding in C57Bl/6 mice; the whole wound bed was filled with collagen and cells were scarce, indicating the appearance of fibrosis (Figures 6C and 7). In contrast, in db/db mice thirteen days after wounding, the mature granulation tissue dominated in the wound bed, whereas deposition of collagen in the extracellular matrix had just begun at the wound edges. There was a difference between the granulation tissue growth pattern in C57Bl/6 mice and that in db/db mice (Figure 8). In C57Bl/6 mice, granulation tissue emerged from the wound bed and from the wound edges, filling the tissue defect. In db/db mice, the wound was filled with abundant fat tissue, whereas granulation tissue formed a lace growing through the interlobular septa of adipose tissue.

Inflammation

It has been reported that in obesity, cytokines produced by macrophages cause an increase in white adipose tissue–resident macrophages, some of which fuse to generate giant multi-nucleated cells (Fantuzzi 2005). Histological evaluation of wounds in our experiment showed that lipogranulomas were present in all db/db mice. Lipogranulomas developed around single fat cells and/or characteristically shaped cholesterol crystals. Crystals were mostly engulfed into the cytoplasm of a giant multinucleated cell of a foreign-body type (Figure 3). On the other hand, secondary or nonspecific inflammation was not found in any of animals in either group.

Discussion

In an attempt to further standardize the model of full excisional wound healing in diabetic mice, we have established differential methods for monitoring skin wound closure, re-epithelialization, inflammation, and granulation tissue formation.

Using a method developed for wounds in rats (Ševeljević-Jaran et al. 2006), we have shown that the computerized digital camera technique provides an objective, precise, and reproducible method for planimetric measurement of wound area in the mouse, confirming other studies with similar techniques (Galiano et al. 2004; Obara et al. 2005; Senter et al. 1995; Thawer et al. 2002). A delay in the closure of the excisional wounds in db/db mice in comparison to normal C57Bl/6 mice was clearly seen at most time points. This finding confirms previous studies showing that wound contraction does not play a major role in wound closure in the db/db mouse and is not affected by wound occlusion with dressings (LeGrand 1998; Senter et al. 1995; Sullivan et al. 2004). Also, prolonged initial wound retraction in db/db mice was observed in our experiment. It is well known that the modulation of fibroblasts toward the myofibroblast phenotype is essential for tissue remodeling during normal and pathological wound healing (Conrad et al. 1993; Tomasek et al. 2002). Prolonged wound retraction in db/db mice might be a result of delayed myosin appearance.

To make wound healing evaluation more precise, we have developed separate histological methods for the assessment of re-epithelialization, contraction, granulation tissue, and inflammation.

Re-epithelialization was assessed, according to the literature, as a percentage of the width of the sectioned wound covered with newly formed epithelial layer (Low et al. 2001). Wound contraction was determined as the change in diameter of the whole wound, to differentiate it from re-epithelialization. Although the rate of contraction was significantly higher in C57Bl/6 mice, there was no statistically significant difference in the rate of re-epithelialization between C57Bl/6 and db/db mice, which is in accordance with literature data (Greenhalgh et al. 1990; LeGrand 1998; Michaels et al. 2007; Senter et al. 1995). Our observations support the conclusion that re-epithelialization plays a major role in wound healing in db/db mice and that impaired contraction is the main reason for delayed wound closure.

Granulation tissue was initially assessed using a scoring method based on literature data (Greenhalgh et al. 1990) by which cell infiltration, granulation tissue thickness, amount of acellular matrix, and the number of newly formed capillaries was assessed. The scoring turned out to be time consuming, and it was hard to gain an overall picture of wound healing dynamics. After careful and critical evaluation of a number of wounds at different time points, we designed a new scoring system for granulation tissue. The main principle of the scoring system was the observation that elements of granulation tissue, namely cellular infiltrate, vessels, and extracellular matrix proteins, lead to different granulation tissue architecture as a function of time. Fibroblasts, for example, changed their shape from plump into elongated, and their environment changed from edematous to fibrous granulation tissue with changing architecture or ‘‘granulation tissue maturation.’’ Fibroblasts play a central role in wound healing and are important for the wound contraction, secretion of growth factors, and deposition of collagen fibers (Lerman et al. 2003; Obara et al. 2005; Thorey et al. 2004). In vitro studies conducted on fibroblasts from db/db mice showed that they exhibit selective impairment in cellular migration, VEGF production, and response to hypoxia compared to fibroblasts isolated from their wild-type littermates (Lerman et al. 2003). This finding was not a result of absence of the leptin-receptor in db/db mice. Werner et al. (1994) demonstrated reduced mRNA levels of FGF1 and FGF2 during wound healing in healing-impaired, genetically diabetic mice. FGF 1 and 2 were shown to stimulate angiogenesis (Risau 1990) and are mitogenic for several cell types present at the wound site, including fibroblasts and keratinocytes (Abraham and Klagsbrun 1996). Therefore, the impaired function of fibroblasts in adult db/db mice could explain the delay in granulation tissue maturation and abundance observed in our study. Furthermore, it has been observed that expression of homeobox D3 (HoxD3), a collagen-inducing transcription factor, and expression of collagen type I were reduced in an animal model of diabetic wound repair in db/db mice (Hansen et al. 2003). In our experiment, a remarkable delay in collagen deposition (described as ‘‘fibrosis’’) was observed, reflecting inefficient collagen synthesis in diabetic wounds.

In addition to granulation tissue architecture, a semiquantitative assessment of granulation tissue volume was incorporated into the scoring system. By comparing the two elements of architecture and abundance of granulation tissue, the dynamics of wound healing could be analyzed. The scoring system was easily reproducible and suitable for following the time-dependent changes in the wound healing process. Thus, although granulation tissue was deposited earlier in wounds of C57Bl/6 mice and ‘‘fibrosis’’ with collagen deposition occurred rapidly between nine and thirteen days after wounding, maturation of granulation tissue in wounds in db/db mice was slower and collagen deposition markedly delayed. These results are in accordance with previously conducted studies of wound healing in diabetes (Goodson and Hung 1977; Silhi 1998). We have demonstrated a difference between the granulation tissue growth pattern in C57Bl/6 mice and db/db mice resulting from the abundance of fat tissue in db/db mice. Histological analysis showed that granulation tissue in db/db mice forms a ‘‘lace’’ growing through the interlobular septa of adipose tissue which therefore represents a mechanical barrier for collagen deposition and might be one of the reasons for its delay and, consequently, for reduced wound closure.

White adipose tissue has recently emerged as an active participant in regulating physiologic and pathologic processes, including immunity and inflammation (Fantuzzi 2005). It is composed of adipocytes and the stromovascular fraction, which includes macrophages. It has been reported that many inflammation- and macrophage-specific genes are dramatically up-regulated in white adipose tissue in mouse models of genetic induced obesity (Xu et al. 2003), suggesting that macrophage-related inflammatory activities in adipose tissue may contribute to the pathogenesis of obesity. Also, Fantuzzi et al. (2005) reported that in obesity, cytokines produced by adipocytes and macrophages up-regulate adhesion molecules on endothelial cells leading to transmigration of bone marrow–derived monocytes and thus an increase in white adipose tissue-resident macrophages, some of which fuse to generate giant multinucleated cells. In our experiment, lipogranuloma formation (inflammation) was observed in all db/db mice, and, most probably, was a consequence of initial fat tissue injury and liberation of fatty acids into the cellular surroundings as already described in the literature (Lin et al. 1979).

In conclusion, we have confirmed with a computer-assisted planimetry method the results of previous studies showing that healing of non-occluded full excision wounds in db/db mice does not occur by contraction as much as in healthy mice. In addition, using a new approach to histological assessment, we have been able to distinguish differential effects of genetic diabetes in db/db mice on the deposition and maturation of granulation tissue, significantly shortening and simplifying the histological scoring, by obviating the need to count capillaries and neutrophils. We have shown that wound closure in db/db mice is delayed because of: (1) delayed granulation tissue maturation; (2) ‘‘laced,’’ widely distributed granulation tissue around fat lobules; and (3) obstruction by lipogranulomas, whereas the rate of re-epithelialization seems to be the same as in C57Bl/6 mice. This methodology should permit a more precise differentiation of effects of novel therapeutic agents on the wound healing process in db/db mice.