The purpose of the present study was to examine whether exogenous liposomal cDNA gene transfer is recognized by the cell and causes endogenous cellular and physiological responses. When administered as a protein, IGF-I is known to cause adverse side effects due to lack of cellular responses. Therefore, we used IGF-I cDNA as a vector to study cellular and physiological effects after liposomal administration to wounded skin. Sprague-Dawley rats were given a scald burn to inflict an acute wound and were divided into two groups to receive weekly subcutaneous injections of liposomes plus the Lac-Z gene (0.2 μg vehicle) or liposomes plus the IGF-I cDNA (2.2 μg) and Lac Z gene (0.22 μg). Transfection was confirmed by histochemical assays for β-galactosidase. Planimetry, immunological assays, and histological and immunohistochemical techniques were used to determine molecular mechanisms after gene transfer, protein expression, and dermal and epidermal regeneration. IGF-I cDNA transfer increased IGF-I protein expression and caused concomitant cellular responses by increasing IGF binding protein (IGFBP)-3 and decreasing IGFBP-1. IGF-I cDNA gene transfer increased keratinocyte growth factor expression and exerted promitogenic antiapoptotic effects on basal keratinocytes, thus improving epidermal regeneration. IGF-I cDNA improved dermal regeneration by an increased collagen deposition and morphology. IGF-I cDNA increased VEGF concentrations and thus neovascularization. Exogenous-administered IGF-I cDNA is recognized by the cell and leads to similar intracellular responses as the endogenous gene. Liposomal IGF-I gene transfer further leads to improved dermal and epidermal regeneration by interacting with other growth factors.
- wound healing
- growth factors
- gene therapy
nonviral gene therapy has several advantages over viral gene transfer methods, such as being easy to apply, simple, direct, and inexpensive, and also not requiring ex vivo manipulation (1). Finally, the liposomal constructs can be administered repeatedly without causing an immune response or tachyphylaxia (1, 4, 22). Hence, liposomal cDNA gene transfer represents a unique therapeutic approach for several pathophysiolgical states (1, 4, 22). There are still some unanswered questions regarding cellular, molecular, and physiological mechanisms, e.g., whether exogenous nonviral liposomal cDNA gene transfer is recognized by the cell and causes concomitant endogenous cellular and physiological responses. To answer some of these questions, we constructed a vector with a cytomegalovirus-driven promoter containing the sequence for IGF-I and defined cellular and physiological responses in vivo.
When administered as a protein, IGF-I is known to cause adverse side effects due to lack of cellular responses, in particular, the synthesis and expression of the IGF binding proteins (IGFBPs) (2, 3, 8). In the circulatory system, 95-99% of IGF-I is bound and transported by one of its six binding proteins (IGFBP-1 to -6), whereas IGFBP-3 represents the most important binding protein (2). The synthesis and expression of the binding proteins are regulated intracellularly via feedback signals by the same cell that synthesizes the IGF-I protein, which means that the production of IGF-I causes a concomitant increase of IGFBP-3 and a decrease of IGFBP-1 (2, 3, 8). This intracellular signal cascade would explain why exogenous administration of the IGF-I protein does not stimulate the binding protein synthesis, leading to supraphysiological doses of free IGF-I and thus to the adverse side effects (3). The purpose of the present study was 1) to determine whether IGF-I cDNA gene transfer results in an increased IGF-I protein concentration, with a concomitant change of IGFBP-1 and IGFBP-3, showing that the transferred gene is recognized and incorporated into the cellular transcription and translation process; and 2) to determine physiological interactions of the transcribed and translated IGF-I protein with other cell signals and growth factors. To study cellular and physiological effects after gene transfer, we chose the skin and induced an acute wound, because the skin is easily accessible and physiological responses can be well monitored (1, 9, 10, 22). In addition, experimental and clinical studies showed the great potential of IGF-I to improve the wound healing process (9, 10, 12, 18). Fibroblasts and keratinocytes have IGF-I receptors, which probably mediate IGF-I stimulation of mitogenicity and proliferative activity (12). Therefore, in addition to the cellular effects of IGF-I, the physiological effect of IGF-I on the process of dermal and epidermal regeneration could be defined.
MATERIALS AND METHODS
Twenty-two adult male Sprague-Dawley rats (350-375 g) were placed in wire-bottom cages housed in a temperature-controlled room with a 12:12-h light-dark cycle. Rats were acclimatized to their environment for 7 days before the blinded study. All animals received similar amounts of a liquid diet of Fresubin (Fresenius Medical Care) and water ad libitum throughout the study. Each rat received a 30% total body surface area (TBSA), full-thickness scald burn under general anesthesia (50 mg/kg body wt pentobarbital) and analgesia (1 mg/kg body wt buprenorphin) following a modified procedure as previously described (7). Rats were anesthetized, shaved, and received a 30% TBSA scald burn (99°C hot water contact 10 s to the back). After the thermal injury, rats were immediately resuscitated by intraperitoneal injection of Ringer lactate (50 ml/kg body wt). Thermally injured rats were then randomly divided into two groups to receive either weekly subcutaneous injections of liposomes (10 μl liposomes in 180 μl saline) containing 0.2 μg of the reporter gene for β-galactosidase Lac Z cDNA construct, vehicle (n = 11) or weekly subcutaneous injections of liposomes (10 μl liposomes in 180 μl saline) containing 2.2 μg of a IGF-I cDNA construct, plus 0.2 μg of the reporter gene for β-galactosidase Lac Z cDNA construct (n = 11).
These studies were reviewed and approved by the Animal Care and Use Committee of the Regierung der Oberpfalz, Bayern, Germany and University Texas Medical Branch (UTMB), Galveston, Texas, assuring that all animals received humane care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” [DHEW Publication No. (NIH) 85-23, Revised 1985, Office of Science and Health Reports, DRR/NIH, Bethesda, MD 20205]. The animals were visited twice daily by the group and daily by the Animal Care and Use Committee to ensure that animals are not suffering or were not in pain. Animals were treated humanely, and pain medication, special nutrition, and fluid substitution according to human burn treatment were given.
In previous experiments, we found that animals receiving liposomes containing the Lac-Z gene represented a better control compared with animals receiving saline in having better wound healing rates and an attenuated acute phase response. Therefore, we chose, in the present study, to use liposomes with Lac-Z as the control group. The rat IGF-I cDNA construct consisted of a cytomegalovirus-driven IGF-I cDNA plasmid prepared at the UTMB Sealy Center for Molecular Science Recombinant DNA Core Facility (kind gift from Dr. J. R. Perez-Polo). The liposomes were formulated from 1:1 (M/M) 1,2-dimyristyloxypropyl-3-dimethyl-hydroxyl ethyl ammonium bromide and cholesterol suspended in membrane-filtered water (Life Technologies, Rockville, MD). This reagent interacts spontaneously with IGF-I cDNA to form the lipid cDNA complex. The dose of 2.2 μg for the IGF-I cDNA was defined in a dose-response study (11).
Immediately after the thermal injury, each rat received 0.2 ml of the lipoplexes subcutaneously injected at two sites opposite each other. This was repeated each week for 4 wk. Mixtures were prepared fresh before injections. Animals were humanely killed by decapitation 5 days after the last injection. Skin samples from the back were harvested, fixated in 4% paraformaldehyde and 1% glutaraldehyde, or snap-frozen in liquid nitrogen, and stored at -73°C for analysis.
Transfection was determined in skin samples taken 33 days after burn and 5 days after the last injection by measuring the presence of β-galactosidase. The presence of the β-galactosidase protein was detected by histochemical staining with halogenated indolyl-β-d-galactoside (Life Technologies, Gaithersburg, MD) for β-galactosidase in the skin. Linear skin biopsies ∼4-mm thick in width and extended from the center of the burn wound well into surrounding normal skin were taken at the end of the experiment and processed as previously published (9, 10).
IGF-I Protein Concentration
IGF-I protein concentrations in dermal and epidermal tissues were measured by using two methods, radioimmunoassay (RIA) and standard immunohistochemical techniques for antibodies against IGF-I. IGF-I protein concentrations were measured by rat RIA (Diagnostic System Laboratories, Webster, TX) in skin biopsies. Proteins were extracted, recovered, centrifuged, and then measured as previously published (10).
IGF-I protein concentrations in dermal and epidermal tissues were measured by using antibodies against IGF-I with standard immunohistochemical techniques. Paraffin-embedded samples were cut 4-μm thick, placed in increasing alcohol concentrations, and placed in PBS. Protease K (100 μg/ml) was applied at 37°C for 30 min. After samples were washed with PBS, endogen peroxidases were blocked by using methanol-H2O2 for 15 min. The samples were then again washed. Samples were incubated with pig serum (1:5; DAKO X901) for 20 min. After the samples were washed, the primary antibody rabbit anti-IGF-I (1:10; Peprotech, Tebu, Germany) was applied to the samples and then samples were incubated at 4°C twice overnight. After another washing, the samples were incubated for 1 h at 37°C with the secondary antibody DAKO E-431 (biotinylated swine anti-rabbit, 1:300) and followed by another washing. The samples were then incubated with streptavidin (1:300; DAKO P-0397) for 1 h at 37°C. The samples were thoroughly washed and diaminobenzidine-hydrogen peroxidase was applied for color development. Counter-staining was performed by using hematoxylin. After the samples were placed in increasing alcohol concentrations, they were mounted. IGF-I concentration was determined by the number of positive cells per high-power field (HPF) and per micrometer square. Three observers blinded for treatment counted each sample at three different sites for IGF-I positive cells.
IGFBP-1 and IGFBP-3 Protein Concentration
IGFBP-1 and IGFBP-3 concentrations were determined by using antibodies against IGFBP-1 or IGFBP-3 with standard immunohistochemical techniques as described in the paragraph above. The antibody used was a goat anti-IGFBP-1 or -IGFBP-3, respectively (Santa Cruz-Biotechnology, 1:10).
Keratinocyte Growth Factor and VEGF Protein Concentration
Keratinocyte growth factor (KGF) protein concentrations in the skin were measured by antibodies against KGF using standard immunohistochemical techniques. Paraffin-embedded samples were cut 4-μm thick, placed in an increasing alcohol concentration, and finally placed in PBS. Protease K (100 μg/ml) was applied at 37°C for 30 min. After samples were washed with PBS, endogen peroxidases were blocked by using methanol-H2O2 for 15 min. The samples were again washed. The primary antibody goat anti-KGF (Santa Cruz Biotechnology) was applied to the samples and then incubated at 4°C overnight. After another washing, the samples were incubated for 1 h at 37°C with the secondary antibody DAKO E-0466 (biotinylated rabbit anti-goat, 1:400) and followed by another washing. The samples were then incubated with streptavidin (1:300; DAKO P-0397) for 1 h at 37°C. The samples were thoroughly washed and diaminobenzidine-hydrogen peroxidase was applied for color development. The counterstaining was performed by using hematoxylin. After placing the samples in an increasing alcohol concentration they were mounted. KGF concentration was determined by the number of positive cells per HPF and per micrometer square. Three observers blinded for treatment counted each sample at three different sites for KGF-positive cells.
VEGF protein concentrations in the skin were measured by antibodies against VEGF using standard immunohistochemical techniques. Paraffin-embedded samples were cut 4-μm thick, placed in an increasing alcohol concentration, and finally placed in PBS. Pretreatment was in a microwave for 30 min at 240 W. After samples were washed with PBS, endogen peroxidases were blocked by using methanol-H2O2 for 15 min. The samples were then again washed. The primary antibody rabbit anti-VEGF (1:50, PC 315; Oncogene Research Products) was applied to the samples and then incubated at 4°C overnight. After another washing, the samples were incubated for 1 h at 37°C with the secondary antibody (DAKO E-0431 biotinylated anti-rabbit, 1:300) and followed by another washing. The samples were then incubated with streptavidin (1:300; DAKO P-0397) for 1 h at 37°C and again washed. Counterstaining was performed with hematoxylin. After placing the samples in an increasing alcohol row they were embedded. Positive controls used the blood vessels of the small bowel. VEGF concentration was determined by the number of positive cells per blood vessel. Three observers blinded for treatment counted each sample at three different sites for VEGF positive cells.
Reepithelization. Reepithelization was determined by planimetry and by histological examinations. Planimetry was performed as follows. The wound eschar was left intact for the first 28 days and then removed by gentle traction, with caution being taken to not disturb or destroy the healing edge along the periphery. After the eschar was removed, animals were placed on a standard surface and the wound area was traced onto acetate sheets along the well-demarcated reepithelized and nonburned interface and the leading edge of the neoepithelium. The areas of these tracings were calculated by computerized planimetry (Sigma Scan and Sigma Plot Software, San Rafael, CA). The area of reepithelization was calculated by the following formula. Values are expressed as percent reepithelization from the original burn wound. with outer area (o) 4 or 5 wk postburn, inner area (i) 4 or 5 wk postburn, and original area (b) at the time of burn.
Histologically, reepithelization was determined in skin samples taken 40 days after burn and 5 days after the last injection. Linear skin biopsies were taken and were ∼4-mm thick in width and extended from the center of the burn wound well into surrounding normal skin. Skin specimens were fixed overnight at 4°C in 4% paraformaldehyde in a HEPES-buffered Hanks' balanced salt solution at pH 7.6. Skin was then stained for hematoxylin and eosin (2a).
Length of reepithelization was determined by measuring the distance from the normal, uninjured skin (identified by skin appendix structures, such as hair follicle and glands) to the end of the neoepithelium covering the wound and granulation tissue. Measurements were repeated twice by observers blinded to the treatment. Length was calculated by forming the average of the two measurements.
Epithelial cell layer. Hematoxylin and eosin-stained skin samples were examined to determine the highest neoepidermis. Epithelial cell layers from the basal membrane to the surface were counted by three observers blinded to treatment. The epithelial cell layer was calculated by forming the average of the three measurements.
Skin cell proliferation and apoptosis. The balance between proliferation and apoptosis is an important indicator for organ homeostasis, which in this case was the skin. Skin cell proliferation was determined by using antibodies against Ki-67. Ki-67 stains cells that underwent mitosis over the last 24 h. Paraffin sections 3- to 5-μm thick were mounted on Superfrost Plus slides, heated for 20 min at 72°C, deparaffinized, and rehydrated. Sections were placed in a microwave for 30 min at 240 W in a citrate buffer at pH = 7.3 and then cooled to room temperature. After antigen retrieval, the slides were rinsed, endogenous peroxidase activity was blocked by using methanolic peroxide, the slides were rinsed again, and primary monoclonal antibody against Ki-67 (Cat. No. 36521A; BD PharMingen) in a dilution of 1:100 was applied. All slides were stained by using a Ventana machine, and each antibody incubation was performed at 37°C and labeled. The streptavidin-biotin-peroxidase method at 37°C was used to visualize positive reaction (Ventana Medical Systems basic DAB detection kit, Ventana Medical Systems). A dark-brown nuclear precipitate was evaluated as positive reaction. Crypts of the large bowel were used for positive control, and sections incubated without primary antibody did not yield positive immunoreactivity.
All skin cells in the pannus and along the basal layer were counted. Proliferation was then determined by positive-stained skin cells per 100 counted skin cells. Three observers blinded to the treatment groups were recruited to count the cells.
Apoptosis was measured by using the terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) immunohistochemical method (Apoptag; Oncogene, Baltimore, MD) to allow histological identification of apoptotic cells in the skin. Formalin-fixed tissues were processed and embedded into paraffin. Sections of 4 μm obtained at 40- to 50-μm intervals were deparaffinized, rehydrated in graded alcohol, and washed in deionized water. Protein was digested by using proteinase K (20 μl/ml in PBS) to decrease background contamination. The sections were then incubated with freshly prepared terminal deoxyribonucleotidyl transferase enzyme at 37°C for 2 h. After enzyme incubation, slides were incubated with antidigoxigenin peroxidase at room temperature for 30 min. Sections were thoroughly washed and diaminobenzidine-hydrogen peroxidase was applied for color development. Last, the sections were counterstained with Mayer's hematoxylin and mounted.
Two sections (4 μm) of each block of tissue were obtained at 40- to 50-μm intervals. In each section, three observers, blinded to treatment, counted TUNEL-positive cells. Apoptotic cells were identified as those with a brown staining of the nucleus, or as apoptotic bodies, which are fragments of apoptotic cells engulfed by neighboring epithelial cells. All epithelial cells within the basal skin cells and dermal cells were counted and apoptosis was expressed as a percentage of apoptotic cells per 100 epithelial cells. Values for all sections were averaged to calculate apoptosis for the apoptotic rate in the skin.
To determine skin cell net balance, whether skin cells survive or undergo apoptosis, we divided proliferation/apoptosis. A large ratio indicates cell survival, and a small ratio indicates that cell survival is decreased.
Collagen morphology and deposition. Quality and quantity of collagen was determined by the Masson Goldner Trichrome staining (2a). Qualitative measurements of the wound collagen were performed by examining the pannus for collagen boundless and the staining intensity of collagen in the area at the wound edge. The collagen morphology and staining intensity were determined by three observers blinded to the treatment. A score from 1 to 4 was given to each sample, whereby 1 = collagen bundles are linear and 4 = collagen bundles appear to have normal histological structures (angel curl shape). Similarly a score from 1 to 4 was given for staining intensity for the collagen pannus, 1 = weak blue-green staining, few collagen concentration, 4 = strong blue-green staining, high density of collagen in the pannus.
The length of the collagen pannus being deposited in the dermal wound structures was determined by measuring the distance from the normal, uninjured skin to the end of the collagen pannus in the granulation tissue. Measurements were repeated twice by observers blinded to the treatment. Length was calculated by forming the average of the two measurements. Depth of the collagen pannus was determined in the same fashion. The area of the collagen pannus was calculated by multiplying the length times the depth of the pannus.
Neoangiogenesis. Neoangiogenesis was determined by quantifying blood vessels in the granulation tissue. Three observers blinded to treatment counted the blood vessels in three different sections of the granulation tissue using a magnification of ×40 (HPF). Values were averaged and calculated as blood vessels per micrometer square.
Statistical comparisons were made by ANOVA and Student's t-test with the Bonferroni correction. Data are expressed as means ± SD or means ± SE, where appropriate. Significance was accepted at P < 0.05.
IGF-I, IGFBP-1, and IGFBP-3 protein concentration
IGF-I protein concentration in the skin, determined by RIA, showed that animals receiving the IGF-I cDNA construct had significantly increased IGF-I protein concentrations when compared with controls. IGF-I cDNA-treated animals had an IGF-I concentration of 175 ng/ml, whereas control animals had an IGF-I concentration of 96 ng/ml (P < 0.001) (Fig. 1A). Immunohistochemistry confirmed the RIA findings. By using antibodies against IGF-I, we detected IGF-I expression in the same area in which we found a positive reaction for transfection, indicated by the histological examinations for β-galactosidase. Rats receiving 2.2 μg of the IGF-I cDNA construct had an increased number of positive cells per HPF and thus per micrometer square, using antibodies against IGF-I when compared with controls, P < 0.05 (Figs. 1B and 2, C and D).
IGFBP-1 concentration was significantly decreased in rats with the IGF-I cDNA treatment. Rats receiving the IGF-I cDNA construct demonstrated 3 ± 1 positive cells/HPF (13 positive cells/μm2), whereas rats receiving the liposomes plus the Lac Z gene had 7 ± 1 positive cells (30 positive cells/μm2) (P < 0.05) (Fig. 1C). In contrast to IGFBP-1, IGFBP-3 protein concentration was significantly increased in rats receiving the IGF-I cDNA compared with control rats. There were 27 ± 2 IGFBP-3 positive cells/HPF (116 positive cells/μm2) in the IGF-I cDNA group, whereas there were only 12 ± 2 positive cells (51 positive cells/μm2) in the control group (P < 0.001) (Figs. 1D and 2, E and F).
KGF and VEGF protein concentration
IGF-I cDNA gene transfer increased KGF protein concentration in the skin. Whereas 14 ± 1 KGF positive cells/HPF (60 positive cells/μm2) could be identified in the IGF-I cDNA group, only 9 ± 0.7 KGF-positive cells (38 positive cells/μm2) were found to be positive in the control group (P < 0.05) (Fig. 1E). Similarly, IGF-I cDNA gene transfer increased VEGF protein expression. In the IGF-I gene group, 14 ± 2 positive cells/HPF were found (60 positive cells/μm2), whereas 7 ± 1 positive cells (30 positive cells/μm2) could be identified in the control group (P < 0.05) (Fig. 1F).
Confirming previous experiments in both groups, the fine granular blue-green reaction product of the β-galactosidase reaction was predominantly present in the granulation tissue underlying the wound. Cells consistently staining for β-galactosidase were myofibroblasts, endothelial cells, and macro-phages located in the areas of inflammation and multinucleate giant cells (Fig. 2, A and B). The highest concentration of transfected cells was next to the injection site, but transfection is almost equally detectable all around the wound area; thus the size of the skin area affected by the local injection is almost equal all around the wound.
Reepithelization Reepithelization, determined by planimetry, was significantly increased in rats receiving the liposomal IGF-I cDNA complexes when compared with rats receiving the liposomes containing the cDNA for β-galactosidase (P < 0.05) (Fig. 3A). Histological examinations of reepithelization revealed similar results as planimetric measurements. Rats treated with the IGF-I cDNA gene complex at a dose of 2.2 μg showed an accelerated reepithelization of almost 160%. Five weeks after the wound infliction in the IGF-I gene group, linear reepithelization was 380 ± 30 μm. In the control group, reepithelization was significantly less with 240 ± 19 μm, P < 0.001 (Fig. 3B).
Epithelial cell layers. No differences in epithelial cell layers could be found between the treatment (13 ± 1 layers) and the control group (13 ± 1 layers).
Skin cell proliferation and apoptosis. IGF-I cDNA gene transfer in a dose of 2.2 μg significantly increased basal cell proliferation by 45% when compared with controls (P < 0.01). Animals receiving the IGF-I cDNA gene demonstrated a cell proliferation of 23 ± 3 positive cells/HPF, whereas 14 ± 3 positive cells were found in animals receiving liposomes (P < 0.05) (Fig. 4, A and B). Furthermore, skin cell apoptosis was significantly decreased by the biological factor 2 in the IGF-I cDNA 2.2 μg group when compared with the liposomes plus Lac Z gene group (P < 0.05). Apoptosis was in the IGF-I group 1.3 ± 0.3/HPF, compared with 3.3 ± 0.4 apoptotic cells in the liposomes group (P < 0.05) (Fig. 4, C and D). Thus the IGF-I cDNA group had a significantly improved net balance calculated as proliferation divided by apoptosis (15 ± 1) when compared with liposomes alone (4 ± 1) (P < 0.05).
Collagen morphology and deposition. IGF-I cDNA gene transfer improved quality and quantity of deposited collagen (Fig. 4, E and F). In rats receiving the IGF-I cDNA at a dose of 2.2 μg, collagen formation was significantly higher than the collagen formation in rats receiving vehicle. Bundle morphology was 1.7 ± 0.1 in the control group compared with 2.4 ± 0.2 in the IGF-I gene group (P < 0.05). Furthermore, the staining intensity of collagen deposition was significantly higher in rats treated with IGF-I cDNA when compared with vehicle (liposomes 2.1 ± 0.1 vs. IGF-I 2.7 ± 0.1, P < 0.05). We also found that IGF-I cDNA at 2.2 μg significantly increased the length of the collagen pannus (liposomes 200 ± 10 μm vs. IGF-I 300 ± 20 μm) and the area of the collagen pannus compared with vehicle (liposomes 20,000 ± 2,000 vs. 36,000 ± 6,000 μm2) (P < 0.05) (Fig. 4, E and F).
Neoangiogenesis. IGF-I cDNA treatment increased the amount of blood vessels per HPF and thus neoangiogenesis. The IGF-I cDNA-treated group had 12 ± 0.7 blood vessels/HPF (50 vessels/μm2) compared with 7 ± 0.4 blood vessels/HPF in the vehicle group (30 vessels/μm2) (P < 0.05).
In the present study, we showed that IGF-I cDNA gene transfer results in an increased IGF-I protein concentration with a concomitant change of IGFBP-1 and IGFBP-3, indicating that the transferred gene is recognized and incorporated into the cellular transcription and translation process. Furthermore, we showed interactions of the transcribed and translated IGF-I protein with other cell signals and growth factors and the benefit of IGF-I gene transfer on dermal and epidermal regeneration. Physiologically, IGF-I is bound to one of six binding proteins IGFBP 1-6, whereas IGFBP-1 and IGFBP-3 are the most important factors to regulate IGF-I concentration and efficacy (2). An increase of IGFBP-1 stimulates IGF-I synthesis and expression. Increased synthesis of IGF-I leads to decreased concentration of IGFBP-1. IGFBP-3 has the opposite effect. Intracellular synthesis of IGF-I causes a concomitant increase of IGFBP-3, which binds IGF-I before it is systemically released and transported (2). A possible reason for exogenous IGF-I protein administration to cause adverse effects is that it does not affect IGFBP synthesis and thus leads to a supraphysiological concentration of free IGF-I (2, 3). We hypothesized that IGF-I cDNA gene transfer leads to IGF-I protein synthesis but furthermore, and more importantly, to a concomitant increase of IGFBP-3 and decrease of IGFBP-1, similar to the physiological cellular function. We showed that in fact IGF-I gene transfer affected IGFBP-1 and IGFBP-3 protein expression. IGF-I cDNA gene transfer increased IGF-I mRNA and IGF-I protein concentration in the area of transfection (10). The exogenous-transferred IGF-I gene caused an endogenous increase of IGFBP-3 and decrease of IGFBP-1. These data indicate that the transferred gene exerts the same intracellular responses as the endogenous gene, which is probably the reason why IGF-I cDNA gene transfer is safe and efficacious.
A new finding was that IGF-I, a mesenchymal growth factor, interacts and stimulates other growth factors, such as KGF and VEGF. KGF is synthesized by several mesenchymal cells, such as fibroblasts, microvascular endothelial cells, and smooth muscle cells (23, 24). For KGF to exert its effects, which is mainly in a paracrine fashion, KGF needs to bind to its receptor fibroblast growth factor receptor 2-IIIb, which is expressed in keratinocytes and hair follicles (6, 14, 17, 25). Activating the KGF-receptor leads through phosphorylation to proliferation of epithelial cells and is responsible for the morphogenesis of the skin and thus plays an important role during wound healing (24). After wounding, KGF protein and mRNA concentration increases and they are mainly present in the dermal fibroblasts below the wound and at the wound edges, whereas KGF receptor transcripts and also the corresponding protein were exclusively detected in keratinocytes of the epidermis and the hair follicles (15). These cells migrate over the wound and subsequently proliferate, which leads to complete wound reepithelization.
Development of a vascular supply is essential for wound healing and is reliant on neovascularization. During wound healing, angiogenic capillary sprouts invade the fibrin/fi-bronectin-rich wound and within a few days organize into a microvascular network in the granulation tissue (21). Neovascularization depends on a dynamic interaction among endothelial cells, cytokines, and the extracellular matrix (5). VEGF stimulates and increases neovascularization (5). Chronic or impaired wounds have decreased concentrations of VEGF and fewer blood vessels in the granulation tissue when compared with physiological wounds, which is probably due to increased proteolysis of VEGF (20). We found that IGF-I cDNA transfer increased the number of newly formed blood vessels in the granulation tissue. Supporting this finding is the increased concentration of VEGF, which we showed by immunohistochemical staining with antibodies against VEGF. IGF-I increased VEGF concentration in the wound, which in turn leads to increased neoangiogenesis.
The most important function of the skin is to serve as a protective barrier against external influences, such as the environment (16). The loss of dermal integrity may lead to loss of fluid, energy, and increased permeability accompanied by an increased risk of bacterial translocation and thus the development of infection or even sepsis (16). Wound healing is critical for morbidity and mortality in acute and chronic wounds. In previous studies (9, 10), we determined the effects of IGF-I on wound healing by injecting IGF-I cDNA constructs into the skin. We showed that after the injection of liposomal cDNA complexes, reepithelization was significantly improved. However, we did not determine the mechanisms by which accelerated wound healing occurred, which was one aim of the present study. We examined the effect of IGF-I on epithelial and mesenchymal structures. Confirming previous experiments, we found that IGF-I cDNA accelerated reepithelization. A possible mechanism of how IGF-I improved epithelial growth onto the wound bed is that IGF-I significantly increased epithelial cell proliferation and decreased apoptosis, resulting in a better epithelial cell balance. Furthermore, IGF-I cDNA gene transfer increased the length and area of the collagen pannus deposited from the wound edge into the newly formed skin. IGF-I did not affect only the length of the collagen pannus, but also the morphology of the collagen deposition. IGF-I cDNA increased the amount of collagen at the wound site, and the structure of the newly formed collagen was more similar to the physiological structure of collagen. Increased and improved collagen formation may be due to a stimulation of fibroblasts to increase their synthesis capacity. The increase of collagen at the wound edge can explain the increased wound breaking strength as found in many different studies.
We and others (1, 9, 10, 19) have shown that liposomal gene transfer is useful for the treatment of the skin. Liposomes can be applied by either topical administration or by direct injection of liposomal gene constructs. Alexander and Akhurst (1) demonstrated that topical application of liposomal constructs containing the Lac Z-gene in shaved 4-wk-old mice resulted in a transfection and expression in the epidermis, dermis, and hair follicle. Topical administration of liposomal complexes, however, has been speculated not to be as effective as injection in cell transfection and expression (1, 22). This is due to the stratum corneum, which is the protective layer of the epidermis. Thus injecting the liposomal gene complexes improves transfection of dermal and epidermal cells. After normal skin transfection, which is transient, expression is predominantly in keratinocytes with lower levels of expression in dermal cells (1, 9, 10, 22). The skin is a good model to study mechanisms mediating nonviral gene transfer because of its accessibility and the ease of monitoring the modified area, and it is likely that the mechanisms are found in other organs and modes of application as well (1). We used a control group with 0.2 μg DNA, whereas IGF-I-treated rats received 2.2 μg IGF-I cDNA and 0.2 μg Lac Z cDNA, making 2.4 μg cDNA. Although it may be possible that the observed effects result from the difference in cDNA, we previously used 2.4 μg cDNA coding for Lac Z and did not observe any of the effects seen with IGF-I cDNA transfer. We conclude that the effects are due to the transferred IGF-I cDNA and not to the amount of cDNA.
In the present study, we demonstrated some new mechanistic insights after nonviral gene transfer. Exogenous IGF-I cDNA gene transfer caused cellular and physiological responses similar to endogenous IGF-I, which indicates that the exogenous-transferred gene is recognized by the cell and initiates the same molecular cascade. This is probably the reason why IGF-I cDNA gene transfer is safe and efficacious in improving dermal and epidermal regeneration. We therefore suggest that liposomal IGF-I cDNA gene transfer has great potential to affect multiple pathophysiological states.
This study was supported by Deutsche Forschungsgemeinschaft Grant DFG-Je-233/1-2.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society