Introduction: The treatment of diabetic wounds remains one of the most difficult challenges for clinicians due to the inherent poor healing of these wounds. Previous studies have demonstrated that diabetic C57Bl/ks-db/db mice exhibit significantly impaired wound healing and that topical application of PDGF improves healing of these wounds. The present study investigates whether transfection of the regenerating dermis in diabetic mice using lentiviral gene therapy with the PDGF gene will also improve diabetic wound healing.

Methods: Platelet-derived growth factor (PDGF) cDNA was cloned from HUVEC RNA using RT-PCR. Lentiviral vector plasmids were constructed with the PDGF and blastosidin resistance gene. Lentiviral vector plasmids were also constructed with the green fluorescent protein (GFP) gene. A 2x2 cm full thickness dermal wound was made on the dorsum of each db/db mouse. The animals were divided into three groups (7 animals per group) as follows: Group I: empty wound, Group II: lentiviral PDGF, Group III: lentiviral GFP. The lentiviral vectors were injected into the margin and base of the wounds (total volume of 1cc of 1x106 lentiviral vector particle/cc). The mice were sacrificed at 21 days and the wounds were harvested. Samples were processed for H&E in order to measure the epithelial gap and area of granulation tissue within the wounds. Samples were also processed for immunohistochemistry for PDGF, immunohistochemistry for CD31 as a marker for angiogenesis, and Picrosirius Red for quality of collagen formation.

Results: Successful construction of PDGF lentiviral vectors was confirmed by demonstrating growth of transfected dermal fibroblasts in blastosidin supplemented media. These transfected cells also expressed PDGF mRNA as measured by RT-PCR. Successful construction of GFP lentiviral vectors was confirmed by transfecting dermal fibroblasts, exposing the cells to blastosidin, and analyzing the cells for green fluorescence. There was successful in vivo transfection of the PDGF gene into the regenerating dermis as confirmed by immunohistochemical staining for PDGF. In vivo transfection of the GFP gene was confirmed by fluorescent microscopy. At 21 days, there was no statistical difference in the epithelialization or granulation tissue area between the control groups (empty wound and lentiviral GFP groups) and the lentiviral PDGF group. However, CD31 immunohistochemistry revealed a statistically significant increase in neovascularization with the lentiviral PDGF group versus controls [mean vascular density 6.94 ± 1.22 for PDGF group vs 3.85 ± 1.04 for the empty control and 3.37 ± 1.23 for the GFP control (p<0.05)]. Furthermore, Picrosirius Red staining revealed substantially thicker, more coherently aligned collagen fibers in the PDGF group versus controls.

Conclusions: PDGF lentiviral vectors were successfully constructed and used to transfect the regenerating dermis in diabetic wounds. The animals treated with lentiviral PDGF produced PDGF while the control and GFP animals did not. While the epithelial gap and area of granulation tissue showed no difference between the groups, there was significantly increased neovascularization and improved quality of collagen formation in the PDGF group versus controls. Studies investigating longer time interval of healing and increased titers of lentiviral vector are currently in progress.