PDGF-BB, TGF-Beta;1, and FGF-2 Proteins Elevate Scar Formation in a Rabbit Ear Excessive Scar Model

Excessive cutaneous scarring in the form of hypertrophic scars and keloids is an area of unmet clinical need and continues to pose significant functional, cosmetic, and psychological problems for many patients and surgeons.^1,2 Numerous surgical and pharmacological strategies have been employed to treat excessive scars as reviewed by Tsao et al.³ The molecular signals that cause an active wound healing process to turn off in the process of scar maturation are unknown, and the biology of excessive scarring remains enigmatic to scientists and physicians. The clinical treatment of scars has therefore been largely empirical.⁴ A key challenge for studying excessive scars is the lack of relevant and practical animal models for this condition because excessive scars afflict only humans.⁵ Several groups have reported their observations of hypertrophic scars and keloids in animal models. These included 1) human keloidal tissue explanted into nude mice;⁶ 2) human hyperplasic tissue developed in normal human skin that was burned 1 month after grafting to the nude mice or explanted into flaps of nude rats;^7–9 3) experimental hypertrophic scar formation formed after administrating an excisional skin wound with coal tar in guinea pigs;¹⁰ 4) hypertrophic scars developed in rabbit ears;¹¹ and 5) the female, red Duroc pig as an experimental animal model of hypertrophic scarring.¹² Most of these models have been cited with histological characteristics from that of human scars without quantitative analyses or with regressing of transplanted human excessive scars in the animal models.¹¹ Quantitative studies of excessive dermal scars in rabbit ears were first reported by Mustoe et al.¹¹ In these studies, histomorphometric analyses were used to evaluate total area of scar formation. Scar elevation indexes provide a tool to study excessive scarring. However, these quantitative and semi-quantitative analyses were employed histologically. Quantitative and kinetic evaluations of elevated scar thickness in live animals have not been conducted in these studies.
Recent technological advancements have contributed tremendously to the understanding of cellular and molecular mechanisms and differences between fetal scarless wound healing and adult scarring wound healing.¹³ These differences between fetal and adult cutaneous wound healing include intrinsic processes to fetal tissue like the unique functions of the dermal fibroblasts,^14–17 extracellular matrix components of a fetal wound rich in hyaluronic acid,^18–23 a markedly reduced inflammatory response,^24–26 relatively deficient growth factor profiles, such as transforming growth factor-b1 (TGF-β1), platelet-derived growth factor-BB (PDGF-BB), basic fibroblast growth factor (FGF-2),^27–30 and intracellular signal transduction.³¹ The most prominent differences between fetal and adult wound healing processes are the cytokine profiles in wounds and the extent of inflammatory response to cutaneous injuries.¹³ TGF-β1, PDGF-BB, and FGF families are the most extensively studied growth factor proteins or cytokines and possess profibrotic functions and promote scar formation.^28,32,33 Several studies have shown that the reduction or complete absence of TGF-β1, PDGF-BB, and FGF-2 in fetal wounds was a major factor responsible for scarless wound repair.^28,32,34
Due to previous described scar-promoting properties, TGF-β1, PDGF-BB, and FGF-2 proteins provided important information on the biological activity and mechanism of action of the elevation of scar formation. The properties of TGF-β1, PDGF-BB, and FGF-2 proteins in elevated scar formation were reviewed in an improved rabbit ear excessive scarring model.

Materials and Methods

The study was reviewed and approved by the Canji Animal Care and Use Committee. All experiments were conducted under the conditions described in the Guide for the Care and Use of Laboratory Animals of the National Institute of Health (1996). A total of 14 female New Zealand White (NZW) rabbits (Western Oregon Rabbit Company, Philomath, Ore) 4–5 months of age weighing 3.5–4.0 kg were used in 2 studies. These studies were repeated at least twice, and the results were comparable. All rabbits had free access to normal food and water throughout the studies.
Experimental design. The authors conducted 2 experiments using TGF-β1, PDGF-BB, or FGF-2 protein to produce excessive scar formation for developing the excessive scarring model. Doses of TGF-β1, FGF-2, and PDGF-BB proteins were chosen based on previous literature references in ischemic rabbit ear wounds.^35,36 All tested proteins were delivered intradermally around the margins of the rabbit ear wounds immediately after wounding. The injection volume was 100 mL per wound for all treatment groups including all control groups. There were 6 rabbit ear wounds per treatment group in the studies (n = 6). In the first experiment, a dose response of TGF-β1 and FGF-2 proteins was performed. TGF-β1 at doses of 0.25 mg, 0.5 mg, and 1.0 mg per wound and FGF-2 at doses of 0.5 mg, 1.0 mg, 3.0 mg, and 5.0 mg per wound were delivered intradermally. Wounding only, phosphate buffered saline (PBS) vehicle, and a protein stabilizer bovine serum albumin (BSA) at a dose of 200 mg per wound served as controls. As a protein diluent stabilizer, the BSA dose was the highest matching delivery dose with TGF-β1 and FGF-2 proteins. The PDGF-BB dose was evaluated in the second experiment at 1.0 mg, 2.0 mg, and 3.0 mg per wound. Scar measurements for elevated scar formation were performed in these 2 experiments. Since scar measurements with scabs intact do not represent wound elevation accurately, scar measurements were initiated on Day 14 (the first study) or 17 (the second study) post-wounding when all scabs had naturally fallen off the wounds. Scar measurements continued twice weekly thereafter. All rabbits in these 2 studies were sacrificed on Day 28 or 29 post-wounding. At this time point, all scars, regardless of treatment, were no longer significantly elevated.
Presurgical and surgical procedures. Rabbits were anesthetized with an intramuscular injection of 35 mg/kg ketamine, 5 mg/kg xylazine, and 0.1 mg/kg butorphanol. Three 8-mm wounds were created with a trephine in the ventral surface of each rabbit ear to the depth of the cartilage with approximately 8–10 mm between each wound. The cartilage and overlying skin were removed from each wound. In dermal repair investigations, all wounds were covered with a commonly used occlusive dressing (OpSite™, Smith & Nephew, Inc., Largo, Fla), which generally fell off the wound by Days 10–12 post-wounding. One rabbit ear received only 1 treatment.
Formulation of test reagents. All formulation methods were conducted according to the vendor’s product sheets. Briefly, lyophilized recombinant human TGF-β1 protein (Chemicon, Temecula, Calif) was reconstituted in sterile water to a concentration of 5 mg/100 mL. Further dilutions were made in sterile PBS containing 2 mg/mL BSA to the final working concentrations of 0.0025 mg/mL, 0.005 mg/mL, and 0.01 mg/mL. Lyophilized recombinant human FGF-2 protein (Sigma-Aldrich, St Louis, Mo) was reconstituted in sterile PBS containing 2 mg/mL BSA to the final working concentrations of 0.005 mg/mL, 0.01 mg/mL, 0.03 mg/mL, and 0.05 mg/mL. Lyophilized recombinant human PDGF-BB protein (Chemicon) was reconstituted in 10 mM acetic acid to a concentration of 10 mg/100 mL and stored at -20^oC. In the morning of the study initiating, reconstituted PDGF-BB was diluted in sterile PBS diluent to the final working concentrations of 0.01 mg/mL, 0.02 mg/mL, and 0.03 mg/mL. Vehicle controls consisted of sterile PBS and BSA (2 mg/mL in sterile PBS).
The test growth factors (0.25 mg, 0.5 mg, and 1.0 mg of TGF-β1; 0.5 mg, 1.0 mg, 3.0 mg, and 5.0 mg of FGF-2; and 1.0 mg, 2.0 mg, and 3.0 mg of PDGF-BB per wound) were injected intradermally around the wound margins in a total volume of 100 mL per wound. Injection sites on wounds were positioned 3, 6, 9, and 12 o’clock and within 2–3 mm of the wound margin. Wounding only, PBS, and BSA served as technique, vehicle, and diluent control, respectively.
Scar measurement. To evaluate wound closure differences of TGF-β1, PDGF-BB, and FGF-2 in this model macroscopically, reepithelization rates in all groups were evaluated (data not shown). There were no differences in reepithelization rates among all groups. Wounds were completely healed at approximate Days 10–12. Scars from all groups were relatively flat at the beginning of macroscopical examination at Days 10–12. Scar thickness of each wound was measured with a micrometer (series #227, Mitutoyo, Kanagawa-Ken, Japan) at Day 14 or 17 post-wounding and continued twice weekly thereafter until sacrifice at Day 28 or 29. Gross appearance of an elevated rabbit ear wound scar is shown in Figure 1A. A micrometer was used to measure the full thickness of the ear over the scar area including the dorsal side (parameter A) and the thickness of the adjacent unwounded portion of the ear (parameter B). Elevated scar thickness (parameter C) was determined by the formula A – B = C. Although elevated scar thickness has not been implemented in histological sections of rabbit ear wounds, the measurement formula should be congruent with the scars in live rabbit ears. A rabbit ear section with trichrome staining represents and better describes the measurement formula of elevated scar thickness in this model (Figure 1B).
Histological examination. At Day 28 or 29 post-wounding, animals were euthanized, and rabbit ears were amputated at the base. Full-thickness wounds were excised and bisected. Each wound sample was fixed in 4% paraformaldehyde (PFA), paraffin embedded, cross-sectioned at 6-mm thick, and processed for trichrome and hematoxylin and eosin (H&E) staining.
Statistical analysis. Data are presented as arithmetic means ± standard deviation (SD) where noted. Statistical analysis of data was conducted using the 1-way analysis of variance (ANOVA) followed by the Fisher’s procedure for least significant differences (StatView, SAS Institute Inc, Cary, NC). Results were considered significant at P ≤ 0.05.

Results

Macroscopically, there was no difference in the rate of wound closure between TGF-β1 protein, PDGF-BB protein, FGF-2 protein, wounding only without injections, PBS, and BSA groups. Visibly raised and palpable scars were observed in all groups between Days 10–12 post-wounding. Figure 1A shows an elevated scar at 29 days after 3 mg of PDGF-BB protein treatment. In all treatment groups, the majority of elevated scars had flattened and resolved by Day 28 or 29 post-wounding.
Scar measurements of TGF-β1 protein-treated wounds. Elevated scar measurements of TGF-β1 protein-treated wounds were taken at Days 14, 17, 21, 24, and 28 post-wounding. The measurement results from different doses of TGF-β1 protein-treated groups, wounding only without injections, PBS, and BSA groups are shown in Figure 2. The greatest measurements of elevated scar formations were observed at Days 14 and 17 post-wounding. Significant difference of scar formation was only observed in 0.25 mg of TGF-β1 and BSA-treated groups when compared to PBS and wounding only groups at 14 days post-wounding (Figure 2, P < 0.05). The lowest dose of TGF-β1 protein-treated wounds showed the thickest scar formation throughout the entire study when compared to the remaining TGF-β1 groups at doses of 0.5 mg and 1.0 mg per wound. The difference of the elevated scar thickness eventually resolved by the end of the study at 28 days post-wounding except for the 0.25 mg of TGF-β1 treatment group (Figure 2).
Scar measurements of PDGF-BB protein-treated wounds. Scar measurements for PDGF-BB protein-treated wounds were taken at Days 17, 20, 23, 27, and 29 post-wounding. The results of scar measurements from PDGF-BB protein-treated groups and the PBS control group are shown in Figure 3. The peak elevated scar formation was observed at Day 17 post-wounding. Significant differences of elevated scar thickness were observed between the PDGF-BB treatment group at 3 mg and the remaining groups (PDGF-BB 1 mg and 2 mg groups and PBS group) at Days 17 and 20 (P < 0.03). By the end of the study (Day 29), although no significant differences were observed between any treatment groups, the scar thickness in the highest dose of PDGF-BB treatment group was still approximately 2-fold thicker than the remaining groups (Figure 3).
Scar measurements of FGF-2 protein-treated wounds. Scar measurements for FGF-2 protein-treated wounds were performed at Days 14, 17, 21, 24, and 28 post-wounding. The scar measurement results from all FGF-2 treated groups, wounding only, BSA, and PBS groups are shown in Figure 4. The thickest elevated scar formation was observed 14 days post-wounding. Significant differences of scar thickness were observed at 3 mg of FGF-2 and BSA treatment groups when compared to PBS and wounding only groups at 14 days post-wounding (Figure 4, P < 0.05). In general, wounds in the 1 mg and 3 mg FGF-2 treatment groups showed thicker scars when compared to wounds in the 0.5 mg and 5.0 mg FGF-2 groups throughout the entire study. The differences in elevated scar thickness eventually resolved by the end of the study at 28 days post-wounding (Figure 4).
Morphological evaluation. Trichrome-stained tissue sections from rabbit ear wounds that received different doses of PDGF-BB, TGF-β1, and FGF-2 proteins were harvested at 28 or 29 days and evaluated. Microscopically, a thicker epithelial layer showing epidermal hyperplasia with regular or irregular elongation of the rete ridges was observed in all PDGF-BB protein-treated wounds, particularly in wounds treated with the highest dose of PDGF-BB (Figure 5). Additionally, there was more granulation tissue formation with higher cellularity and richer cellular collagen deposition in the wound beds when compared with the wounds treated with PBS alone (Figure 5). Morphological characteristics of TGF-β1 and FGF-2 protein-treated wounds were comparable to PDGF-BB protein-treated wounds. In addition, morphological characteristics of wounds in the wounding only and BSA groups are intermediate between the PBS and growth factor-treated wounds (data not shown). Figure 5 presents the morphological characteristics of the strongest responses observed in rabbit ear wounds treated with 3 mg of PDGF-BB, 0.25 mg of TGF-β1, or 1 mg of FGF-2.

Discussion

The reparative outcome of a cutaneous wound in a human fetus is true tissue regeneration without scarring, not just repair. In contrast, the inevitable response of cutaneous wound healing to injury in children and adult humans is a fibroproliferative response that eventually leads to a fibrotic scar. Normal wound healing in adults involves elevation of inflammatory responses and enhancements of production and activity of growth factors with receptors. It has been reported that the repair process of scarless fetal wound healing involves mild inflammation accompanied by significantly reduced levels of growth factors.^13,33 The precise mechanism of scarless fetal wound repair remains unknown. Several publications report that growth factors, such as the TGF-β1, PDGF-BB, and FGF-2 families, are profibrotic cytokines and all have profibrotic functions and promote scar formation.^28,32,33 Additionally, Liu et al characterized the differences between fetal and adult wound healing and concluded that the reduction or complete absence of TGF-β1 and PDGF-BB in fetal wounds was a major factor responsible for scarless wound repair.¹³ In this report, the authors have explored the development of a new model for studying human excessive scar formation. By applying growth factors to rabbit ear excisional wounds, the model mimics the period of time in human scarring where there is a known elevation of growth factors. Results from this study have shown that significantly elevated scar formation occurs after TGF-β1 (low dose only), PDGF-BB, or FGF-2 application, suggesting that these growth factors are important contributors to the formation of excessive scars.
The biological effects of TGF-β1 include chemotaxis, angiogenesis, and production and remodeling of extracellular matrix. The profibrotic function and scar-promoting properties of TGF-β1 have been widely studied. TGF-β1 overexpression has been observed in patient tissues with systemic sclerosis, hypertrophic scars, and keloids.^37–39 Moreover, the scar formation of cutaneous wounds is reduced by neutralizing antibodies to TGF-β1.⁴⁰ PDGF-BB has also been shown to play an important role in wound healing, which includes chemotaxis stimulation, fibroproliferation, angiogenesis, and more. Studies have demonstrated that overactivity of PDGF-BB in scarless fetal wounds induces fetal wound fibrosis, and elevated levels of PDGF-BB have been involved in liver cirrhosis.^28,41 In addition to TGF-β1 and PDGF-BB, FGF-2 is also involved in the process of dermal wound healing. A range of studies has indicated that FGF-2 is a mitogen for a wide variety of cell types including vascular endothelial cells, vascular smooth muscle cells, fibroblasts, and keratinocytes.⁴² Recent observations have demonstrated that FGF-2 has greater expression with increasing gestational age in fetal skin and during adult wounding.³²
To the best of the authors’ knowledge, the use of TGF-β1, PDGF-BB, and FGF-2 as stimulators of scar formation for the development of an excessive scarring model in rabbit ears has not been reported. The authors hypothesize that TGF-β1, PDGF-BB, and FGF-2 play important roles in elevating excessive scars because of their well established roles in the wound healing process. TGF-β1, PDGF-BB, and FGF-2 doses used to enhance scar formation in the present studies were chosen on the basis of other investigators’ studies that showed these growth factors are potent in stimulating granulation tissue formation in animal wounds.^35,36 The results from current studies demonstrate that TGF-β1, PDGF-BB, and FGF-2 contribute to excessive scar formation in the rabbit ear model. Although a dose response of TGF-β1 was not generated in the study, all doses of TGF-β1 elevated scar thickness. It is not clear why the lowest dose of TGF-β1 (0.25 mg per wound) showed relatively thicker scar formation than the other 2 higher doses of TGF-β1. PDGF-BB protein demonstrated a dose-dependent effect in stimulating scar formation in this model. Wounds treated with the highest dose of PDGF-BB displayed the thickest elevated scar formation. This thickest elevated scar formation was observed not only in all doses of PDGF-BB protein administrative groups but also between the TGF-β1, PDGF-BB, and FGF-2 protein-treated wounds in this model. Given the thickest elevated scar formation was observed in PDGF-BB protein-treated wounds, it seems likely that PDGF-BB protein is a more potent profibrotic stimulator than TGF-β1 and FGF-2 proteins in this model. In FGF-2 protein-treated wounds, 2 middle doses of FGF-2 (1 mg and 3 mg per wound) stimulated thicker scar formation than the lowest (0.5 mg per wound) and the highest (5.0 mg per wound) doses of FGF-2, suggesting FGF-2 activity in this model follows a bell-shaped curve. Mustoe et al demonstrated that FGF-2 at a dose of 5.0 mg per wound did not show an increase in granulation tissue formation in a similar rabbit ear wound healing model.³⁵ In contrast, Pierce et al showed that FGF-2 at a dose of 2 mg per wound stimulated granulation tissue formation in the same model.³⁶ Although it is not clear why the high dose FGF-2 did not show the highest scar elevation in this model, these results are comparable with the evaluations reported by other investigators. Impressively, scar elevations were significant at the early measurement time points (Days 14–20 post-wounding) only after a single intradermal delivery of TGF-β1, PDGF-BB, and FGF-2 proteins, suggesting that the peak response of elevated scarring to these growth factors in this model occurs during a 1-week time period, approximately. The small sample size might also be a limitation in the study, and a relatively larger sample size in future studies would be helpful to observe significant differences in these growth factors at all time points post-treatment. Interestingly, scar measurement results suggest that the BSA diluent may contribute to stimulate scar formation in the early stage of wound healing in this model, raising the possibility that a foreign protein immune response may also contribute to the scar thickness in TGF-β1 and FGF-2 protein-treated wounds. However, the data of scar measurement from TGF-β1 and FGF-2 treatment groups did not support this standpoint.
In recent years, significant work has shown that epithelial-mesenchymal interactions play an important role in regulating skin homeostasis between epidermis and dermis.^43,44 These interactions may be involved in governing excessive scar formation, such as in keloid pathogenesis. Lim et al reported that keloid-derived human keratinocytes cocultured with normal dermal human fibroblasts resulted in increased proliferation.⁴⁵ Interestingly, an even higher proliferation level was observed after keloid-derived human keratinocytes were cocultured with keloid-derived human fibroblasts.⁴⁶ These results indicate that both keratinocytes and fibroblasts possess an intrinsic abnormality and suggest that both types of cells are involved in keloid pathogenesis. The present studies showed that an elevated scar with a thick epithelial layer that manifests epidermal hyperplasia in conjunction with an enhanced level of granulation tissue existed in all TGF-β1, PDGF-BB, and FGF-2 protein-treated wounds, particularly wounds treated with PDGF-BB protein. These specific morphological manifestations support the notion that TGF-β1, PDGF-BB, and FGF-2 may play important roles in excessive scarring in adult wound healing.
An animal model that exactly simulates the biological, pathophysiological, and morphological parameters of a human excessive scar does not exist. In this brief report, the authors have presented evidence that the exogenously administrated growth factors TGF-β1, PDGF-BB, and FGF-2 increase elevated scar thickness in epidermal and dermal layers, which are associated with a human excessive scar. The mechanism of actions of these growth factors in the rabbit ear excessive scar model is unknown, and additional factors may also contribute to human excessive scar formation. This improved model might still provide a useful tool for evaluation of newly produced anti-scarring agents.

Conclusion

PDGF-BB, TGF-β1, and FGF-2 proteins significantly elevate scar formation in epithelial and dermal layers in a rabbit ear excessive scar model after a single intradermal administration.

Acknowledgments

The authors thank Wendy Hancock for helpful discussions and review. The authors are grateful to Suganto Sutjipto and Doug Cornell in Canji Process Science for the production of recombinant adenoviruses, Jeff Coleman in the Canji Histology Core for technical assistance, and Debbie Muilwijk-Cornell in the Canji Animal Facility for hosting the animals.