The Effects of P-selectin Receptor Antagonist rPSGL-Ig on Wound Healing in a Rodent Model

Introduction P-selectin glycoprotein ligand-Ig (rPSGL-Ig) is a novel P-selectin glycoprotein ligand-Ig receptor antagonist that decreases leukocyte rolling and tethering onto stimulated endothelial cells and platelets, thus decreasing the inflammatory response.1 Recent research in venous thrombosis using rPSGL-Ig showed that pretreatment, prior to thrombus induction, significantly inhibited vein wall inflammation. However, the monocyte population, which is necessary for the normal progression of wound healing, was not significantly altered.2 As rPSGL-Ig will likely be used in clinical situations, such as ischemia/reperfusion, transplantation, and even deep venous thrombosis (DVT), the purpose of this investigation is to determine that such usage will not result in compromised wound healing. Materials and Methods Phase 1. Male Sprague Dawley rats (n = 30), weighing approximately 325g, were utilized in this experiment. Full-thickness skin incisions were created on the dorsum of the rat and left open to heal by second intention. Half (n = 15) were pre-treated with 1mg/kg rPSGL-Ig (reconstituted in 10mM NaCl, 320mOsm/L, 6.94mg/mL) via tail vein injection, and half (n = 15) served as saline controls. The rPSGL-Ig dose of 1mg/kg was chosen based on pharmacokinetic studies performed at Wyeth, Inc.3 The ideal control for rPSGL-Ig is a nonactive mutated human Fc-rPSGL-Ig chimera. However, it is very expensive and has been difficult to produce in sufficient quantities of this control for experimental use. Studies using a low activity form of rPSGL-Ig have shown saline to serve as an adequate control.4 Under anesthesia, the dorsum of each rat was surgically prepped and three 1cm full-thickness skin wounds were surgically created. Skin samples were harvested and evaluated at selective time points (6 hours, day 2, and day 7). Morphometrics. Full-thickness biopsies (6mm) of rat skin were stained with hematoxylin and eosin from paraffin embedded tissues. Five representative high-power fields (oil immersion 1000X) were examined around the wound interface and the cell count of the skin analyzed. Cells were identified as neutrophils, monocytes, or lymphocytes based on standard histologic criteria including nuclear size, cytoplasmic content, and total cell size. Results from five high-power fields were added together and mean ± SE calculated for each skin sample studied. In each experimental group, skin samples were added together and then the mean ± SE calculated for the cellularity of the group as a whole. Skin wound sandwich ELISA. This ELISA technique quantitated protein levels from wound homogenates for the chemokine MCP-1 and cytokines PDGF-bb, TGF-beta-1, and IL-1b. Full-thickness skin samples were placed in 1mL “Complete” protein inhibitor buffer (Roche Diagnostics, Mannheim, Germany) and homogenized at high speed for 45 seconds using a TH homogenizer (Omni International, Warrenton, Virginia). Skin samples were then sonicated for 30 seconds. The samples were spun down using a benchtop microcentrifuge for two minutes at 14,000 RPM. The supernatant from each sample was then used for protein quantification. A 96-well Nunc-Immuno Plate Maxisorp microplate (Nunc, Denmark) was coated the night before with capture antibody (Ab) diluted in coat buffer (120mM NaCl, 50mM H3BO3, 0.16 N NaOH) and incubated at 4 degrees C. Antibodies used were pAb 1:100 IL-1b (Peprotech, New Jersey), mAb 1:100 TGF- beta -1 (R&D Systems, Minnesota), mAb 1:250 PDGF-bb (R&D Systems, Minnesota), and pAb1:250 MCP-1 (Peprotech, New Jersey). Plates were washed three times in wash buffer (1xPBS, 0.05% Tween), then blocked in blocking buffer (1xPBS and 2% BSA) for one hour at 37 degrees C. The wash step was then repeated. Standards were prepared from 100ng/mL down to 0ng/mL in dilution buffer (wash buffer and two-percent fetal bovine serum). 50µL of standards and samples were added in duplicate to plates. Plates were incubated at 37 degrees C for one hour, and the wash step was repeated. Detection antibodies were diluted in dilution buffer; these included pAb biotinylated 1:100 IL-1b (Peprotech, New Jersey), pAb biotinylated 1:250 TGF-beta-1 (R&D Systems, Minnesota), pAb biotinylated 1:250 PDGF-bb (R&D Systems, Minnesota), and pAb biotinylated 1:250 MCP-1 (Peprotech, New Jersey). The detection Abs were incubated at 37 degrees C for one hour. Avidin D-horseradish peroxidase (Vector Laboratories, California), in a 1:10,000 ratio with dilution buffer, was then added to all plates, incubated at 37 degrees C for 30 minutes, and then washed. TMB substrate was added to the plates and allowed to incubate at room temperature (Kirkegard and Perry Laboratories, Maryland). Color development was stopped with 1M phosphoric acid, and the samples were read on an Elx808 plate reader (Biotek, Vermont) at 450nm wavelength. Total protein was measured using the same supernatant as ELISA samples and a bicinchoninic acid protein assay (BCA). Samples were read at 590nm (Pierce, Rockford, Illinois). All samples were run in duplicate. Bovine serum albumin standards ranged from 1mg/mL to 0mg/mL. Hydroxyproline analysis. Collagen content in healing wounds was determined by acid hydrolysis of rat skin homogenates to quantitate free hydroxyproline in hydrolyzates. The addition of Ehrlich’s reagent results in the formation of a chromophore that can be measured by an ELISA plate reader at 550nm. Full-thickness rat skin samples were placed in 1mL “Complete” protein inhibitor buffer (Roche Diagnostics, Mannheim, Germany) and homogenized using high speed for 45 seconds. Skin samples were sonicated for 30 seconds on high and 200µL of full homogenate (not spun down) was removed and placed in a glass vial. 400µL of 6 N HCl was added to sample and baked for eight hours to overnight at 120 degrees C. 10µL of standard (100µg/mL–0µg/mL) or sample (in duplicate) was added to a sterile 96-well plate. 10µL of citrate/acetate buffer (5% citric acid, 1.2% glacial acetic acid, 7.24% sodium acetate, 3.4% sodium hydroxide, 10mL sterile water) and 100µL of chloramine T solution (0.282g chloramine T, 2mL N-propanol, 2mL sterile H20, 16mL citrate/acetate buffer) was then added to the samples and incubated for 20 minutes at room temperature. To each well, 100µL of Ehrlich’s solution (2.5g 4-DMAB Ehrlich’s reagent, 9.3mL of N-propanol, 3.9mL of 70% perchloric acid) was added and incubated for 15 minutes at 65?C in a water bath and read at 550nm wavelength. Sample numbers were then normalized to total protein levels. This was determined using the BCA assay mentioned above. Brillmeyer’s trichrome staining. This technique was used to visualize collagen in 4µm paraffin histology sections. A plasma stain (acid fuchsin) 0.2g diluted in 100µL was utilized followed by aniline blue 0.5g, orange G 2.0g, and phosphomolybdic acid 1.0g added to 100µL of water to increase selectivity for collagen. Brillmeyer’s trichrome stains collagen blue, nuclei black or blue, cytoplasm red, and erythrocytes yellow. The trichrome stain qualitatively differentiates between collagen and smooth muscle cells in healing wounds over time. Phase 2. Male Sprague Dawley rats (n = 30) were utilized in this experiment. Half (n = 15) were pretreated with 1mg/kg rPSGL-Ig via tail vein injection, and the other half (n = 15) served as saline controls. Each rat had four 1cm full-thickness dorsal skin wounds surgically created (2 open and 2 closed with suture) and was evaluated on day 0, day 7, day 14, and day 21 for peripheral blood perfusion using laser Doppler perfusion imaging to document angiogenesis. On day 21, open and closed skin wounds were harvested and prepped for biomechanical tensiometry to evaluate skin tensile strength. Laser Doppler perfusion imaging (LDPI). The laser Doppler perfusion imager (LDPI, Perimed Inc., North Royalton, Ohio) with computer software was used to evaluate skin perfusion in surgically created wounds. Anesthetized rats were placed in sternal recumbency and had the LDPI 670nm helium-neon laser beam placed 15cm above the skin to sequentially scan the surface of healing tissue and detect moving blood cells. At each measurement site the beam illuminates the tissue volume to a depth of a few hundred micrometers. In the presence of moving blood cells, partially Doppler-broadened back-scattered light is detected by a photo detector and processed to determine a percentage perfusion value.5,6 The rats had surgically created open and closed wounds scanned with a helium-neon laser at a wavelength of 670nm on day 0, day 7, day 14, and day 21. In phase II, it was imperative to evaluate the skin wounds of the rat while the animal was anesthetized to prevent distorted readings caused by movement. Maximum perfusion of each open and closed wound was calculated using the computer-generated color scale of increasing intensity (gray, dark blue, light blue, green, yellow, rose, and red). Percent perfusion was normalized by evaluating the green, yellow, rose, and red areas of pixel intensity divided by the total scan area of the wounds (3589 pixels) using a computer program (NIH Image, Bethesda, Maryland). Biomechanical tensiometry. Tensiometry was performed on skin biopsies to determine ultimate tensile stress. One-hundred and twenty wound skin samples (60 open and 60 closed) from rats on day 21 were removed from the rats and frozen at –20 degrees C until testing. Prior to material testing, each specimen was thawed in a warm water bath for approximately 30 minutes. Tissue was cut perpendicular to the wound to form a 3mm x 15mm dog-bone shape using a custom-made instrument containing stainless steel skin graft knife blades. The wound was centered in the narrow region, and the thickness of the wound was measured in the center using a standardized soft-tissue micrometer. After the tissue was cut and measured, the dog-bone ends were placed between sandpaper in sinusoidal clamps and attached to an 858 Mini Bionix II system (MTS Systems Corporation, Minneapolis, Minnesota). The tissue was aligned so that tensile load would occur perpendicular to the wound. A tensile pre-load of 0.1N was applied. The skin was then preconditioned 10 times with a sine wave at a rate of 0.6N/s with a maximum load of 0.5N. After the tissue was examined for irregularities due to the pre-load or pre-conditioning, the specimen was loaded in tension to a failure rate of 0.25mm/s, which is approximately 100 percent/min strain rate based on the cut-out region.7 Data were collected by a microcomputer for analysis. Ultimate stress for each specimen was computed by dividing the ultimate load by the product of the tissue thickness times 3mm (load divided by cross sectional area). Statistical evaluation and animal use. Statistical analysis included mean ± standard error of mean, analysis of variance (ANOVA), and unpaired student t-tests for parametric data. A minimum number of 14 animals per group were needed for this experiment as determined by a statistical power/sample size analysis calculated for an alpha value of 0.05 for a desired power of 0.90. Significance was defined as p Phase 1. Morphometric (cells in 5HPFs) evaluation showed a statistically significant increase in PMNs in the rPSGL-Ig group versus controls on day 2 (19 ± 2.4 vs. 11 ± 3.0, p Phase 2. Laser Doppler perfusion imaging showed increased percent perfusion in wounds on control rats with open wounds versus rPSGL-Ig treated rats (0.4 ± 0.04 vs. 0.2 ± 0.03, p 8 Both neutrophils and monocytes share similar cellular receptors and migrate into a wound at the same time with neutrophils in greater proportions than that are found in the general circulation. However, the presence of neutrophils is short lived in a noninfectious state, and macrophages become the predominate cell type within three to five days.9,10 Platelet activation that initiates homeostasis is also responsible for chemically active substances that modulate inflammation by attracting fibrocytes and monocytes to promote wound healing.9,10 Approximately 24 hours after the wounding event, macrophages take over the regulatory function from platelets. Substances released from macrophages stimulate angiogenesis, collagen production, and fibroplasia that are needed for normal wound healing.11–16 Recombinant PSGL-Ig (rPSGL-Ig) is a novel P-selectin glycoprotein ligand-Ig receptor antagonist that inhibits leukocyte rolling and tethering onto stimulated vascular endothelial cells and platelets. The receptor antagonist includes the amino terminal 47 amino acids of PSGL-1 bound to the Fc portion of IgG (disabling Fc receptor binding and complement activation) and is called rPSGL-Ig. This study set out to determine whether the previously documented anti-inflammatory effects of rPSGL-Ig would have an adverse effect on the wound healing of surgically created skin wounds.1,2 Phase I of this experiment investigated 6mm full-thickness skin biopsies from surgical skin wounds at selected time points (6 hours, 2 days, and 7 days). No statistically significant differences were noted in total inflammatory cell populations between the control and rPSGL-Ig pretreated animals at six hours. However, at days 2 and 7, the rPSGL-Ig treated animals had significant increases in PMNs. This rebound response in PMNs has been documented in our laboratory in evaluating rodent models of venous thrombosis.17 The recruitment of PMNs into the wound of rPSGL-Ig treated animals is likely due to other inflammatory and chemical mediators, such as C5a, IL-8, and leukotriene B4, which attract inflammatory cells to the wound site18 or a bypassing of the selectin-dependent mechanism for inflammatory cell extravasation using other adhesion mechanisms, such as CD11b/CD18. Of interest, seven-day–treated animals showed a monocyte population that was significantly lower as compared to controls. Despite minor differences in subtypes of inflammatory cells noted between the groups, the total inflammatory cell populations in the skin wounds were not significantly different. Therefore, rPSGL-Ig treated animals showed no interference with the normal inflammatory cell migration into the skin wound margins. This finding supports that the normal wound healing process can progress in rPSGL-Ig treated animals. The chemokine MCP-1 was significantly up-regulated in the rPSGL-Ig group versus the controls at all time points. This response indicated that rPSGL-Ig did not interfere with the role of this chemokine in the wound repair process and that other mechanisms in the leukocyte recruitment pathway exist that facilitate MCP-1 up-regulation. Macrophage-derived growth factor expression (PDGF-bb, TGF-beta-1) were not adversely affected by rPSGL-Ig. Additionally, a significant up-regulation of PDGF-bb was observed in the rPSGL-Ig group versus controls at six hours (0.09 ± 0.01ng/mg vs. 0.05 ± 0.007ng/mg, p = 0.03). It has been previously documented that rPSGL-Ig reduces inflammation by limiting the activation of proinflammatory substances in the vein wall in animal models of venous thrombosis.2,17,19,20 Therefore, we feel rPSGL-Ig has properties that promote early down-regulation of the proinflammatory cytokine IL-1b that showed significant decrease expression in the rPSGL-Ig group versus controls at six hours. The evaluation of collagen content by hydroxyproline analysis and by trichrome staining of skin samples showed no quantitative or qualitative differences between the experimental groups. In phase II of this study, laser Doppler perfusion imaging showed increased percent perfusion in open wounds on control rats versus 1mg/kg rPSGL-Ig treated rats at day 0 (p 21 The location of the open wound on the dorsal lateral surface of the thorax and abdomen of the rat had loose skin that did not impair wound contraction. Alternating their placement on the dorsal surface of the rat’s back to account for differences in wound healing randomized the wounds evaluated for this study. Tensile strength more closely describes an intrinsic property of the tissue and should be used when comparing different tissues from different animals or from different areas on the same animal.22 The rPSGL-Ig treated rats with open skin wounds had decreased wound tensile strength. Of interest, rPSGL-Ig treated wounds that were closed with suture and healing by primary intention showed no statistical differences in ultimate tensile strength at 21 days as compared to controls. All surgically created skin wounds, regardless of the assigned experimental group, healed uneventfully. Planimetry evaluation of open wound margins over time would allow the calculation of wound contraction rates. This may be the best method to study significant differences in healing open wounds in future investigations. Conclusions The pre-treatment of rPSGL-Ig before surgically created skin wounds had no adverse effects on acute wound healing in a rodent model. rPSGL-Ig promoted increased early chemokine and growth factor expression (MCP-1, PDGF-bb) that would promote the normal progression of wound healing. It is extremely promising that rPSGL-Ig, which has been studied as a therapeutic treatment for venous thrombosis, did not compromise wound peripheral perfusions or tensile strength in closed surgical wounds. This study has shown that using a selectin antagonist approach to modulate vascular inflammation does not interfere with the normal collagen deposition and tensile strength of healing full-thickness skin wounds.