Burn wound healing involves a complex set of overlapping processes in an environment conducive to ischaemia, inflammation and infection costing $7.5 billion/year in the U.S.A. alone, in addition to the morbidity and mortality that occur when the burns are extensive. We previously showed that insulin, when topically applied to skin excision wounds, accelerates re-epithelialization and stimulates angiogenesis. More recently, we developed an alginate sponge dressing (ASD) containing insulin encapsulated in PLGA [poly(D,L-lactic-co-glycolic acid)] microparticles that provides a sustained release of bioactive insulin for >20 days in a moist and protective environment. We hypothesized that insulin-containing ASD accelerates burn healing and stimulates a more regenerative, less scarring healing. Using heat-induced burn injury in rats, we show that burns treated with dressings containing 0.04 mg insulin/cm2 every 3 days for 9 days have faster closure, a higher rate of disintegration of dead tissue and decreased oxidative stress. In addition, in insulin-treated wounds, the pattern of neutrophil inflammatory response suggests faster clearing of the burned dead tissue. We also observe faster resolution of the pro-inflammatory macrophages. We also found that insulin stimulates collagen deposition and maturation with the fibres organized more like a basket weave (normal skin) than aligned and cross-linked (scar tissue). In summary, application of ASD-containing insulin-loaded PLGA particles on burns every 3 days stimulates faster and more regenerative healing. These results suggest insulin as a potential therapeutic agent in burn healing and, because of its long history of safe use in humans, insulin could become one of the treatments of choice when repair and regeneration are critical for proper tissue function.
The study was undertaken to investigate the efficacy of sustained release of insulin from PLGA microparticles embedded in alginate gels in burn wound healing and to understand the mechanism of the accelerated healing observed in insulin-treated burns.
Our findings show that sustained release of insulin from PLGA microparticles led to accelerated healing via decrease in oxidative stress and tissue damage, early recruitment of neutrophils, management of inflammatory cells, enhanced angiogenesis and proper collagen deposition and maturation.
Use of sustained release of insulin in a moist environment in burn wounds could result in faster and improved healing outcomes.
The American Burn Association has reported that approximately 486000 people/year received medical treatment for burn injuries over the last 10 years . Moreover, inadequate wound treatment leads to increased morbidity and impaired quality of life while consuming substantial health resources in developed countries . Burns are one of the most devastating forms of soft tissue injury and are most commonly caused by exposure to heat, electricity, radiation, chemicals and/or friction . The damage may only affect the epidermal layer of the skin (superficial burns), penetrate a part of the dermis (deep-partial thickness burns) or completely extend through both epidermal and dermal layers (full-thickness burns) .
During the first 48 h post-injury, cardiac output, oxygen consumption, decrease in metabolic rate and hyperglycaemia associated with impaired glucose tolerance occur [5,6–8]. These parameters then gradually rise, developing a hyper-metabolic response, until a plateau is reached at 5 days after the burn. This response is most probably initiated by the secretion of several regulatory proteins, including, but not limited to, hormones such as catecholamine, glucagon and dopamine, cytokines such as interleukin (IL)-1, IL-6 and tumour necrosis factor (TNF)-α, formation of neutrophil adherence complexes and reactive oxygen species (ROS), ultimately leading to protein and lipid catabolism [8,9]. The rise in oxygen consumption results from the increased energy requirements by major organs and tissues, such as liver and muscle . During this early period, increases in inflammation, accompanied by endothelial cell dysfunction and microvascular permeability, result in oedema . Furthermore, intense metabolic changes associated with hyperglycaemia and insulin resistance occur, which are common complications of severe burn .
The successful use of insulin on problematic wounds began in the 1930’s with improvement in the healing of non-diabetic bed sores upon injections of insulin . The topical use of insulin was later proven to effectively heal small uncomplicated decubitus ulcers . Insulin was then subcutaneously injected on scalded rats and shown to promote faster healing, with a better and thicker epidermal layer, abundant retia and higher number of cells in S-phase . Moreover, local injection of long-acting insulin zinc suspension accelerated skin wound healing without major side effects . Insulin administration to hospitalized burn patients also has been shown to enhance protein synthesis in the muscle and accelerate the healing process . Although a variety of therapeutic agents, including growth factors such as epidermal growth factor (EGF)  or transforming growth factor-β (TGF-β) , hormones , simvastatin  and propranolol  have been successfully used in treating patients with burn wounds, the elevated costs associated with the production of these molecules discourages their application .
Tissue-engineered wound dressings have gained increasing interest as an approach towards more rapid wound healing, while reducing pain and discomfort to the patient. This results in faster rehabilitation with less extended hospital stays. Wound dressings should prevent bacterial infection and be non-toxic, biodegradable, capable of removing wound exudates and painless on both application and removal, while providing a moist environment to the wound area [23,24]. Among the different types of wound dressings, alginates have been used because of their haemostatic properties during burn healing [25,26]. Alginate is a low-cost, linear, unbranched polysaccharide obtained from brown seaweed, comprising different amounts of (1→4)β-D-mannuronic and α-L-guluronic acid units. It is biocompatible, non-toxic, non-immunogenic, biodegradable and can undergo gelation upon addition of divalent ions, such as calcium [24,26].
We have previously published a procedure describing encapsulation of human recombinant insulin in poly(D,L-lactic-co-glycolic acid) (PLGA) microspheres . Using this system, we obtained sustained release of insulin in aqueous solution for up to 25 days. In order to achieve an effective insulin stabilization and release in the wound tissue, insulin-loaded PLGA microspheres were further incorporated into alginate sponge dressings (ASDs) . This resulted in the sustained release of insulin, while providing a moist environment and protection to the wounds. In the present study we show that slow release of insulin from PLGA particles in alginate dressings stimulates healing of burn wounds and that the healing occurs with less scarring.
MATERIALS AND METHODS
Human recombinant crystalline insulin (hRcI), PEG (average molecular mass 1450 Da and 10 kDa), poly(vinyl alcohol) [PVA; average molecular mass 13–23 kDa], CaCl2, NaCl, NaOH and SDS were supplied by Sigma–Aldrich. PLGA (D,L, 50:50, average molecular mass 17 kDa) was obtained from Purac Biomaterials. Dichloromethane was supplied by Caledon Laboratories. Protanal LF10/60 LS (35–45% guluronic/55–65% mannuronic) was obtained from FMC Biopolymer. A Micro BCA protein assay kit was obtained through Thermo Fisher. Other materials were obtained as follows: Vectashield from Vector Laboratories; trichrome stain kit from ScyTek Laboratories; ketamine, xylazine, buprenorphine from Henry Schein.
Microparticle and alginate sponge dressing preparation
Burn wound model
Animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of California, Riverside. Female adult Sprague–Dawley rats weighing 200–240 g were obtained from the University of California Riverside vivarium and housed under a 12-h dark/12-h light cycle in a room with controlled temperature and humidity. Animals were supplied with standard rat chow and water ad libitum. Prior to burning, the rats were anaesthetized with an intraperitoneal dose of 80 mg/kg ketamine and 16 mg/kg xylazine. The analgesic buprenorphine (0.05–0.1 mg/kg) was administered right after the burn to prevent post-burn pain. Photographs of the wounds were taken using a digital camera at a distance of 4 cm and wound size was then determined using ImageJ software (NIH) over time during healing until re-epithelialization was achieved. Wound tissues were collected 1 h after the burn was made and at different time points post-burn for tissue analysis. Normal skin was also collected from an area distant from the burn. Over the period of the experiment, levels of blood glucose were monitored with an electronic glucose meter (Bayer) and animal weight was determined using a digital electronic scale (Mettler Toledo). Insulin levels in the tissues were measured using a commercially available kit (Abcam). Insulin in the tissue samples was bound to an immobilized antibody and then conjugated with biotinylated insulin antibody that was further bound to horseradish peroxidase (HRP)-conjugated streptavidin. A chromogenic substrate solution was then added to develop colour that was read at 450 nm and expressed as microlitre-unit per millilitre.
At different time points, tissue was collected for histology by excising the wound tissues, including an area of 5 mm around the wound edge. Tissues were fixed in 4% paraformaldehyde for 2 h, followed by 1 h incubation with 0.1 M glycine/PBS and two incubations with 15% and 30% sucrose, for ∼6 h and overnight respectively. Tissues were then embedded in optimal cutting temperature compound (OCT) (Tissue-Tek), frozen in a mixture of ethanol/dry ice slush and stored at–80°C. Sections were stained with haematoxylin and eosin and with a modified Masson's trichrome stain, as previously described . Sections were visualized using a Nikon Microphot-FXA microscope with a Nikon DS-Fi1 digital camera. Collagen stained blue, whereas keratin showed a red colour, cytoplasm stained pink and cell nuclei were black.
Second harmonic generation imaging
Collagen's intrinsic second harmonic generation (SHG) signal were detected by using an inverted Zeiss LSM 510 non-linear optics (NLO)META laser-scanning microscope (Carl Zeiss Microscopy) based on the Axiovert 200 M inverted microscope equipped with standard illumination systems for transmitted light and epifluorescence detection. This microscope was also equipped with an NLO interface for a femtosecond titanium–sapphire laser excitation source (Chameleon-Ultra) for multi-photon excitation. The Chameleon laser provided femtosecond pulses at a repetition rate of ∼80 MHz, with the centre frequency tunable from 690 to 1040 nm. Images were acquired with a long working distance objective (Zeiss, ×40 water, numerical aperture 0.8). The two-photon signals from the sample were epi-collected and discriminated by the short-pass 650 nm dichroic beam splitter. The SHG images were collected using a META detection module with signals sampled in a 394–405 nm detection range (λex=800 nm). Each image presented is 12 bit, 512 pixels×512 pixels representing 225 μm × 225 μm field of view.
Frozen sections were fixed in 4% paraformaldehyde for 20 min, rinsed with PBS, incubated in a solution of 0.1 M glycine in PBS for 20 min and blocked in 10% non-immune serum of the secondary antibody species in PBS containing 0.1% Triton X-100 for 30 min. Slides were then incubated for 2 h at room temperature with the primary antibodies in 1% BSA in PBS. The primary antibodies used were the following: 1:200 diluted FITC-labelled mouse anti-rat α-smooth muscle actin (α-SMA, Sigma–Aldrich) and 1:100 rabbit anti-rat myeloperoxidase (MPO, eBioscience). Sections were then washed three times with 0.1% BSA in PBS and incubated with 1:25 diluted goat anti-rabbit or 1:100 diluted goat anti-mouse dilutions of FITC-labelled or 1:100 diluted goat anti-rabbit Alexa Fluor 594-labelled secondary antibodies (Life Technologies), for 1 h at room temperature. After washing, sections were mounted in Vectashield containing DAPI. Immunofluorescence was visualized and imaged using a Nikon Microphot-FXA fluorescence microscope with a Nikon DS-Fi1 digital camera and Nikon NIS-Elements software (Nikon Instruments Inc.).
Flow cytometry assay
Tissue disaggregation and digestion were performed in 4 ml of RPMI 1640 medium in the presence of collagenase type 1 (Worthington Biochemical Corp.) at 37°C for 45 min. Dissociated tissues were passed two or three times through an 18 gauge needle and then a 20 gauge needle. Harvested cells were washed with RPMI 1640 medium. Single cell suspension was obtained by straining the homogenate via cell strainers. Single cells were then washed with RPMI 1640 medium and stained with cluster of differentiation (CD)80, CD86 (eBioscience) and CD163 (AbD Serotec) for 30 min. Cells for isotype controls were obtained from the single cell populations and stained with isotype control allophycocyanin (APC), isotype control FITC and isotype control phycoerythrin (PE) (eBioscience). Stained cells were washed with RPMI 1640 medium, suspended in FACS buffer and read using a FACSAria Cell Sorting System (BD Biosciences). Analysis of data was performed using FLOWJO software.
Burn wound tissues were homogenized as described above. ProcartaPlex™ cytokine immunoassay was performed according to the manufacturer's protocol (eBioscience) using equilibrated protein tissue homogenate. The level of IL-10 was quantified using a Luminex™ 200 instrument (Millipore) by monitoring the fluorescence associated with the bead set.
In vivo angiogenesis assay
The hair on the backs of C57BL/6 mice was removed using Nair. The next day, they were injected subcutaneously with either insulin (Humulin; 1 μg/15 μl of saline) or 15 μl of saline using an insulin syringe. Both insulin and PBS were injected in different sites of the back of each mouse every 24 h for 4 days. The areas surrounding the injection sites were labelled using a permanent marker to ensure that the injections were always done at the same site. A possible position-dependent effect was discarded by changing the site of insulin and PBS administration from mouse to mouse. Skin samples from the injected areas were collected and observed at day 5. Blood vessels were highlighted using ImageJ software (NIH). Briefly, the scale for all the images was set using a ruler with known distance and ImageJ function Analyze–>Scale. The free-hand selection tool was used to trace and highlight the blood vessels and the image was saved. The total vessel length was then evaluated using the analyse function in ImageJ.
Vascular endothelial (VE)-cadherin expression was quantified by Western blotting. Burn wound tissues were homogenized in radioimmunoprecipitation assay (RIPA) buffer (1% NP-40, 0.1% SDS, 50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 0.5% sodium deoxycholate and 1 mM EDTA) as previously described and protein was quantified using the DC protein assay kit (Bio-Rad Laboratories). Equal amounts of protein for each sample were mixed with sample buffer, boiled for 5 min and analysed using a SDS/PAGE (10% gel). Following 1 h of blocking in 5% BSA, the immunoblotting was performed by incubating the membrane with 1:1000 rabbit anti-mouse VE-cadherin (phospho-Tyr731) primary antibody (Abcam) overnight at 4°C. Blots were then washed and incubated with the appropriate HRP-conjugated goat anti-rabbit secondary antibody, followed by a 5-min incubation with West Dura extended duration substrate (Pierce Biotechnology, Thermo Fisher Scientific Inc.). Blots were also probed for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (1:1000 dilution, Cell Signaling Technology), a housekeeping protein to evaluate equal loading. Band intensities were quantified with Image Lab 5.0 (Bio-Rad Laboratories) and normalized against the GAPDH control. Results were then expressed as fold change relative to the control.
Determination of the antioxidant status in wounded tissues
Antioxidant enzymes activity was detected in the tissue extracts as previously described . Briefly, tissues (40–50 mg) were homogenized in a bullet blender for 5 min at 4°C using zirconium oxide beads and 10 μl of RIPA buffer per mg of tissue. The extracts were then centrifuged at 16904 X g for 15 min at 4°C. The supernatants were either immediately used or divided into aliquots and stored at–80°C for later use.
Superoxide dismutase (SOD) activity in the tissues was determined using a commercially available kit (Cayman Chemical), according to the manufacturer's protocol. Results were expressed as units per ml of tissue extract, one unit of SOD being defined as the amount of enzyme needed to achieve 50% dismutation of the superoxide radical.
Hydrogen peroxide (H2O2) levels in the tissues were measured using a commercially available kit (Cell Technology Inc.). The assay is based on the peroxidase-catalysed oxidation by H2O2 of the non-fluorescent substrate 10-acetyl-3,7-dihydroxyphenoxazine to a fluorescent resorufin read fluorometrically at 530 nm (excitation)/590 nm (emission) using a Victor 2 microplate reader. The amount of H2O2 in the supernatants was calculated from a standard curve generated with known concentrations of H2O2.
Catalase activity in the tissues was measured using a commercial kit (Cayman Chemical), according to the manufacturer's procedure. This assay is based on the peroxidase ability of catalase together with methanol to produce formaldehyde in the presence of an optimal concentration of H2O2. This formaldehyde was then quantified spectrophotometrically at 540 nm using 4-amino-3-hydrazino-5-mercapto-1,2,4-triazole (Purpald) as the chromogen. Catalase activity was expressed as nmol/min per ml.
Glutathione peroxidase (GPx) activity was measured indirectly by a coupled reaction with glutathione reductase (GR), using a commercially available kit (Cayman Chemical). GPx reduces H2O2 to water, while producing GSSG in the process, which is then recycled to its reduced state (GSH) by GR and NADPH. Oxidation of NADPH to NADP+ is accompanied by a decrease in absorbance at 340 nm. Under rate-limiting conditions of GPx activity, the rate of decrease in the absorbance measured at 340 nm, read at 1-min intervals for a total of 5 min in a Victor 2 microplate reader, will be directly proportional to the GPx activity of the sample. GPx activity was expressed as nmol/min per ml of tissue extract.
DNA damage assay
Wound tissue (5–6 mg) was collected and homogenized in RIPA buffer, as described above. DNA was then extracted using the DNeasy Blood and Tissue kit (Qiagen) according to the manufacturer's protocol. DNA concentration was measured and purity confirmed by determining the A260/A280 ratio. The levels of 8-hydroxy-2′-deoxyguanosine (8-OH-dG), which correlate with the amount of oxidative DNA damage, were measured spectrophotometrically at 405 nm using a DNA/RNA Oxidative Damage EIA kit (Cayman Chemical) as per the manufacturer's instructions.
Lipid peroxidation assay
Thiobarbituric acid reactive substances (TBARS), an indicator of lipid peroxidation in the tissues, were measured by using a commercially available kit (Cell Biolabs Inc.). By-products of lipid peroxidation, such as malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE), form adducts with TBARS that can be detected fluorometrically at 540 nm/590 nm excitation/emission. TBARS levels were then calculated from a pre-determined MDA standard curve.
Protein nitration assay
Nitrotyrosine levels in the tissue were measured by an enzyme immunoassay, using a commercially available kit (Cell Biolabs Inc.). Briefly, tissue samples were incubated with an anti-nitrotyrosine primary antibody, followed by an HRP-conjugated secondary antibody and enzyme substrate, in an ELISA plate. Nitrotyrosine content was then measured spectrophotometrically at 412 nm against a standard curve prepared from pre-determined nitrated BSA standards.
Statistical analysis was performed using Prism 5 (GraphPad Software). A one-way ANOVA, followed by a pairwise comparison post-hoc test, was conducted wherever appropriate. A 95% confidence interval was considered statistically significant.
Partial thickness burn wounds were made on the dorsum of anaesthetized rats with a 1.5 cm diameter brass cylinder heated in a water bath at 80°C for 2 min and pressed against the shaved rat skin for 6 s (Figures 1A–1C). Rats were then treated with ASDs, containing either PLGA microparticles loaded with 125 μg of insulin or ASD with PLGA microparticles without insulin. These 3-cm-diameter dressings were placed on top of the burn wounds and covered with Tegaderm (3M) to keep the dressing in place over the burn (Figures 1D and 1E). Dressings were replaced every 3 days and wounds were cleaned with sterile saline.
Partial thickness burn wound model system
Body weight, measured every 2 days starting from the day insulin dressings were applied and lasting until the end of the healing process, did not show any significant difference between placebo and insulin-treated rats (Figure 1F). We also observed a similar physiological rhythm of changes in blood glucose in both insulin and placebo-treated rats; the levels of blood glucose decreased from day time to night time, which is normal as it depends on the circadian rhythm. However, there was a significant decrease in blood glucose at 4 and 6 h after insulin application, suggesting the gradual releasing of insulin from the dressing. Because in both groups and all times the blood glucose level was higher than 60 mg/dl, no hypoglycaemia occurred (Figure 1G).
Insulin levels in the tissue were significantly elevated in the insulin-treated burns as seen at 12 h and remained significantly elevated at 3 and 6 days post-burn (each cycle of treatment), suggesting the release of insulin from the ASDs. The insulin level in placebo-treated burns stayed low until day 9 when they became elevated (Figure 1H), suggesting the influx of insulin to aid in the later stages of burn healing.
Insulin improves burn wound healing
Insulin treatment accelerated the wound healing process only moderately as shown by a decrease in wound area when compared with placebo-treated wounds, in particular after day 9; this improvement was consistent in all of the rats we tested (Figures 2A and 2B). However, histological evaluation revealed improved quality of the healing tissue. At 3 days post-burn, insulin treatment resulted in early disjunction of the epidermal layer when compared with the placebo-treated burn (Figures 2C–2F), suggesting a more rapid clearance of the dead burned tissue. Simultaneously, the granulation tissue in the insulin-treated burns was sparser than the same tissue in the placebo, indicating faster disintegration of the collagen fibres (Figures 2G and 2H). To examine the condition of the collagen in the burn wounds, we performed second harmonic generated imaging and show that insulin treatment resulted in faster dismantling of collagen fibres; the fibres in the placebo-treated burns appear thicker (Figures 2I and 2J). Furthermore, insulin-treated burns showed an earlier increase in inflammatory cells (Figures 2K and 2L), primarily neutrophils and M1 macrophages, cells important in clearance of bacterial load and debris respectively (Figures 3A and 3C).
Insulin-treated burns show improved healing
Insulin-treated burns show early and improved inflammatory response
Effects of insulin treatment on inflammation
Inflammation levels in the burn wounds were assessed by quantification of inflammatory cells, specifically neutrophils and macrophages, at different time points post-burn. A significant increase in the number of neutrophils occurred during the first 12 h in the insulin-treated wounds, whereas neutrophils were not detected at that time in the placebo-treated wounds (Figure 3A). The early recruitment of neutrophils to the site of insulin-treated burns suggests a faster immune response when compared with placebo-treated burns that, in conjunction with an increase in the M1 macrophages (Figure 3C), could be responsible for the more rapid clearance of the dead burned tissue.
The number of macrophages present in the burn wounds was determined by flow cytometry. Results of the analysis of macrophages with various markers are shown in Figure 3(B). Singlet cell population was defined by plotting forward scattered light height (FSC-H) compared with FSC area (FSC-A) [1,2]. The cell population was then defined with markers for M1 (CD80+, CD86+ and CD163 low) and M2 (CD163+, CD80 low and CD86 low) macrophages. The number of M1 macrophages steadily increased over time in the placebo-treated burns, and at 12 days post-burn, the number of M1 (pro-inflammatory) macrophages was approximately 3-fold higher than at day 3, suggesting continued pro-inflammatory activity in the tissue. In contrast, the number of M1 macrophages initially rose faster in insulin-treated animals, becoming significantly higher by day 6, perhaps to clear the higher number of neutrophils. However, by day 9, levels of M1 macrophages in insulin-treated burns returned to that of the placebo and by day 12, insulin treatment had significantly decreased the number of these pro-inflammatory cells (Figures 3B and 3C). The pattern of M2 macrophages was distinctly different (Figures 3B and 3D). At day 3 post-burn, the placebo was high and then fell off progressively through day 12. In contrast, insulin-treated burns started lower at day 3, decreased through day 9 as the placebo, but then increased significantly at day 12 post-burn, suggesting that insulin induces a more anti-inflammatory environment later in healing that potentially contributes to less scarring. Because insulin stimulates an anti-inflammatory environment, we investigated the levels of IL-10, an anti-inflammatory cytokine, which is known to be produced by M2 macrophages , as well as fibroblasts . We found that, with the exception of day 6, IL-10 was significantly elevated throughout healing in insulin-treated burns than was the case with placebo-treated burns (Figure 4). Taken together, these findings suggest that insulin treatment induces a more anti-inflammatory environment at a time at which the granulation tissue is developing (9–12 days post-burn; full wound closure occurs 22–24 days post-burn), decreasing the late inflammatory response and leading to less scarring.
IL-10 levels are elevated in insulin-treated burns
Insulin stimulates angiogenesis in healing of burn wounds
The presence of new blood vessels was determined by immuno-labelling the tissue sections for α-SMA, a protein present in the smooth muscle cells of the blood vessel walls. We observed an approximately 3-fold increase in the number of newly formed blood vessels in insulin-treated wounds (Figures 5A and 5B). However, this was not accompanied by changes in the average blood vessel width or area (Figures 5C and 5D). The pro-angiogenic properties of insulin were further confirmed in a C57BL/6 mouse model by subcutaneously injecting insulin for 4 days and collecting the skin tissue to measure angiogenesis and the integrity of the blood vessels formed. Results show that insulin induced a 1.9-fold increase in the total blood vessel length, compared with placebo (Figures 5E–5G). Similarly, VE-cadherin expression was increased ∼1.7-fold in the presence of insulin (Figures 5H and 5I).
Insulin-treated burns show increased angiogenesis and improved permeability of blood vessels
Insulin improves collagen deposition and maturation
Granulation tissue has collagen as the major component of the extracellular matrix (ECM). Collagen scaffold plays an intricate and important role in maintaining the biological and structural integrity of the wound tissue where both proper collagen deposition and architectural maturation contribute to the quality of healing. We measured hydroxyproline levels in the burn tissue to evaluate the collagen content in the granulation tissue during the course of healing and found that, when compared with placebo treatment, insulin treatment stimulates a steady increase in hydroxyproline levels after day 3 with a brief decrease at day 12 (Figure 6A). This suggests that insulin stimulates higher deposition of collagen in the wound tissue. Indeed, we found more collagen fibres by Masson's trichrome staining (Figure 6B) and, when using second harmonic imaging microscopy, we see that the fibres are thicker and the organization is close to the ‘basket-weave’ pattern found in normal skin (Figure 6C). Therefore, increases in hydroxyproline levels during the granulation tissue formation and proper maturation of collagen fibres show that insulin is able to improve the quality of the newly formed healed tissue.
Collagen deposition is increased and more mature in insulin-treated burns
Insulin treatment decreases the levels of reactive oxygen species
It is known that low levels of oxidative stress are beneficial to wound healing, whereas excessive levels are problematic for the healing process. We hypothesized that one of the mechanisms by which insulin improves burn healing is by decreasing oxidative stress early post-burn. Upon injury, NADPH oxidase (NOX) acts as the primary source of superoxide anion (O2−), a highly reactive species that, in the presence of SOD, is dismutated to H2O2. H2O2, although less reactive, causes tissue damage upon accumulation and hence it is important that it is broken down to H2O and O2 or only H2O by the antioxidant enzymes catalase and GPx respectively (Figure 7A). Also, H2O2 in the presence of iron can enter the Fenton reaction to give rise to OH•– radicals and HO− anions. Furthermore, O2−, in the presence of nitric oxide (NO), produced by nitric oxide synthase (NOS), generates peroxinitrite (ONOO−; Figure 7A), a very reactive NO species that also causes molecular damage. We found that insulin treatment causes a decrease in SOD activity and consequently H2O2 levels, during the first 3 days of healing (Figures 7B and 7C), suggesting that insulin reduces the levels of O2− generated in response to injury. Simultaneously, the levels of catalase and GPx were increased during this same period of time, contributing to the decrease in H2O2 levels by breaking it down to H2O and O2 or H2O respectively (Figures 7D and 7E). In addition, it is well known that ROS can result in deleterious effects on lipids, proteins and DNA (Figure 7A). Lipid peroxidation, measured by the concentration of MDA present in the burn tissues, was similar in both insulin- and placebo-treated wounds during the first 2 days post-burn, but decreased suddenly at day 3 in insulin-treated burns and then increased gradually over time following the curve for placebo but at a lower level (Figure 8A). Damage to proteins, as indicated by nitrotyrosine (Figure 8B), displays an abrupt dip at day 3 as was seen for MDA, but then slowly climbs back up to the value of placebo at day 9. DNA as measured by 8-OH-dG levels (Figure 8C), on the other hand, showed a small, but significant, dip at days 2–3 in insulin-treated burns but then followed the big decrease at day 6 seen in the placebo, all the while staying slightly below the placebo. The insulin-induced decrease in oxidative stress this early during wound healing has powerful consequences on the subsequent healing processes in terms of inflammation, angiogenesis and scar formation.
Regulation of redox state is enhanced in insulin-treated burns
Damage to lipids, proteins and DNA is reduced in insulin-treated burns
Burn wounds are one of the leading causes of traumatic injury and death. Survivors are frequently faced with physical deformities post-healing and often suffer from psychological problems. Insulin treatment resulted in an early and significantly elevated influx of neutrophils to the burn wound site during the first 12 h. This early increase can help to contain microbial infection in the burn wounds and help in the removal of necrotic debris [32–34]. The early influx of the neutrophils may also suggest a role in the faster epidermal disjunction noted in the insulin-treated burns (Figures 2A–2D). The early disintegration of the necrotic tissue may lead to enhanced and better wound healing. The cellular and molecular mechanism(s) involved in insulin-induced early influx of neutrophils to the site of burn injury warrants further investigation. We are currently undertaking such studies and expect to report on them in the very near future.
An increase in M1 macrophages, the pro-inflammatory phenotype, in insulin-treated burns suggests escalation in the host defence mechanism(s). The initial presence of M1 macrophages has been suggested to increase phagocytic activity and protect the tissue from bacterial infection and promote the removal of damaged tissue . Furthermore, this increase in M1 macrophages may explain the observed early clearance of the damaged granulation tissue in insulin-treated burns. M2 macrophages have been shown to possess anti-inflammatory properties and promote angiogenesis [36,37]. Although the percentage of macrophages in the placebo burns at day 3 was higher in comparison with insulin treatment, the M2 macrophages in insulin-treated burns significantly increased after day 6 as compared with the placebo and was significantly higher by day 12, suggesting a potential role in the less scarring burn healing observed when the burns are treated with insulin .
Burn wounds cause excessive inflammation and ROS generation, hence the presence of anti-inflammatory cytokines such as IL-10 may help to reduce inflammation and consequently scar formation. The IL-10 expression profile in both the placebo- and the insulin-treated burns was biphasic as previously shown by others in burn healing [39,40]. Our data show that insulin treatment significantly stimulates IL-10 production in both the early and the late phases. This suggests that one way insulin improves burn healing is by stimulating IL-10. Although the number of M2 macrophages, known to produce IL-10 , was lower in insulin-treated wounds, the source of elevated IL-10 stimulated by insulin could be due to fibroblast production  at both day 3 and after day 9. Fibroblast stimulation and proliferation has been shown to occur in the presence of insulin . This anti-inflammatory cytokine has been shown to inhibit the activation of macrophages and the presence of reactive oxygen and nitrogen species [42–45]. The significant increase in IL-10 levels beyond day 6 in the insulin-treated burns may indicate its effects on collagen deposition and reduced scarring observed in the insulin-treated burn wounds [46–49].
An effective angiogenesis process brings nutrients and oxygen to the cells in the wound tissue, increasing their proliferation and migration . We have shown that insulin has a pro-angiogenic effect in rats; it increases the density of newly formed microvessels in the healing tissue. For better visualization and evaluation of newly formed blood vessels, we used an in vivo mouse model with thinner skin. In agreement with our findings in the rat model, we observed an increased number of blood vessels after treatment with insulin subcutaneously. An increase in development of microvessels can help the tissue regain the state of normoxia from a state of hypoxia. Indeed, burn wound tissues have increased hypoxia and diminished oxygenation for a prolonged period of time, contributing to impaired healing. Therefore, the increased supply of oxygen and nutrients to the burn tissue stimulates better healing.
Proper granulation tissue formation and remodelling are important to acquire the desired structural integrity post-wounding. The very significant levels of hydroxyproline, a major component of collagen, in insulin-treated burns past day 6 (Figure 6A) correlates with the increase in collagen production and deposition and basket-weave-like collagen fibre structure during the granulation tissue formation observed at day 21 (Figure 6B), suggesting better tissue integrity.
Burn tissues are characterized by initially having high levels of oxidative stress as a result of increased formation of ROS that surpasses the cells' antioxidant enzymes activity . ROS such as O2− and H2O2 can further react with other molecules and/or ions, forming other ROS, such as peroxynitrite, when reacting with NO or OH− radicals via the Fenton reaction, both of which lead to oxidative damage to most biomolecules in particular to lipids, proteins and DNA .
Insulin has been characterized as regulating several biological processes [53–55]. In the present study, we show that the levels of SOD activity are significantly increased in insulin-treated burn wounds shortly after burning. This suggests elevated dismutation of O2− resulting in generation of H2O2 which is effectively removed by increased activity of both catalase and GPx in insulin-treated burns, resulting in significantly lower levels of H2O2, thereby indicating a decrease in oxidative stress.
In conclusion, sustained release of insulin from PLGA microparticles incorporated into alginate dressings accelerates wound healing of burns and improves the restoration of tissue integrity. This was associated with decreased initial oxidative damage, early recruitment of neutrophils, reduced damage to macromolecules and well-controlled inflammation, as well as increased angiogenesis and proper collagen deposition and maturation. The result is a completely healed tissue 21 days post-burn, with proper tissue organization, indicating that insulin, particularly when incorporated in the alginate–PLGA delivery system, may be an improved therapy for burn wounds.
Sandeep Dhall, João Silva, Yan Liu, Michael Hrynyk, Ronald Neufeld and Manuela Martins-Green conceived and designed the experiments. Sandeep Dhall, João Silva, Yan Liu, Michael Hrynyk, Monika Garcia, Julia Lyubovitsky and Manuela Martins-Green performed the experiments. Sandeep Dhall, João Silva, Yan Liu, Alex Chan and Manuela Martins-Green analysed the data. Manuela Martins-Green, Ronald Neufeld, Michael Hrynyk and Julia Lyubovitsky contributed reagents/materials/analysis tools. Sandeep Dhall, João Silva and Manuela Martins-Green wrote the paper.
We thank the Genomics core facility and the Stem Cell Center core facility at the University of California, Riverside for the use of FACS Aria and Luminex 200 respectively. We thank Dr Julia Lyubovitsky for her help with second harmonic generated imaging.
This work was supported by the National Natural Science Fund of China [grant numbers 81170761 and 81270909 (to Y.L.)]; the Natural Sciences and Engineering Research Council of Canada [grant numbers 204794–2011 (to M.H.) and private donor (to M.M.-G.)].
α-smooth muscle actin
alginate sponge dressing
Cluster of differentiation
forward scattered light area
forward scattered light height
reactive oxygen species
second harmonic generation
thiobarbituric acid reactive substance(s)