Molecular Design
To construct a PDGF-mimicking peptide hydrogel, we envisioned the integration of a self-assembly domain and a PDGF epitope capable of activating the function of PDGFR for wound healing. Based on this rationale, we designed a peptide hydrogelator Nap-FFVLE-GG-VRKIEIVRKK (denoted as 1). 1 contains three segments. 1) FFVLE is a b-sheet forming peptide derived from b-amyloid peptide, which has been proven to promote self-assembly (21). The incorporation of the hydrophobic motif Nap enhances this self-assembly ability. 2) As a linker, Gly-Gly increases the flexibility of the binding domain, resulting in better accessibility of the binding domain to PDGFR. 3) VRKIEIVRKK, residues 153-162 on the PDGF-B chain, is the key domain that interacts with PDGFR, as shown in the crystal structures of PDGF-B in Figure 1.(28) The acetyl capping PDGF epitope Ac-VRKIEIVRKK was used as a control molecule (referred to as 2). The binding sequence is very hydrophilic in a neutral environment (five positive charges), allowing it to be exposed on the surface of the hydrophobic core of assembly 1. Therefore, we developed a PDGF-mimicking peptide hydrogel capable of interacting with PDGFR and activating its function in wound healing. The designed molecules were synthesized via standard solid-phase peptide synthesis, purified with reverse-phase high-performance liquid chromatography (RP-HPLC), and characterized by analytical HPLC and MS spectroscopy. All data are provided in Figures S1-S2.
After obtaining these peptides, we first evaluated their self-assembly behaviours. Simple mixing of an equal volume of solution 1 (2.0 wt%) and 2X PBS buffer generates a transparent hydrogel (inset of Figure 2A). Conversely, control molecule 2 hardly forms a viscous fluid at a concentration of 1.0 wt%. These results indicated that NapFFVLE is an excellent self-assembling motif to trigger the gelation of a hydrophilic molecule. As shown in the transmission electron microscope (TEM) image, 1 self-assembled to form nanofibers with widths of approximately 15 nm (Figure 2A). These long nanofibers were intertwined with each other to form a 3-dimensional network, holding a large amount of water and resulting in hydrogelation. To evaluate the self-assembling ability of 1, we measured its critical aggregation concentration (CAC) by using thioflavin T (ThT). Figure 2B shows that the CAC value of 1 was approximately 1.0 mM, further verifying its exceptional self-assembling ability. The CD spectra revealed that gel 1 has a positive peak at 211 nm and a negative peak at 238 nm (Figure 2C), suggesting the formation of a β-sheet-like structure. This is consistent with the fact that the natural PDGF-BB protein also forms a β-sheet conformation (29). Thus, assembly 1 is able to mimic PDGF’s conformation, assuming that this is the structural basis for simulating the biological activity of PDGF-BB. Additionally, the CD spectrum of Sol 1 only exhibited a positive peak at 242 nm, indicating the random coil structure of 1 before self-assembly (Figure 2C). Rheological time sweep revealed the kinetics of supramolecular hydrogelation, in which the storage modulus increased gradually and reached 1500 Pa, suggesting that 1 forms a robust hydrogel (Figure 2D). It is well known that wound dressings should maintain their integrity to protect the wound, and Gel 1 has sufficient strength to provide adequate protection to the wound.
PDGF-mimicking peptide promotes cell proliferation, migration and angiogenesis in vitro
In the injured area, PDGF-BB is an effective mitogen that stimulates the proliferation of fibroblasts and keratinocytes, primarily vascular endothelial cells. It also stimulates macrophages to activate and secrete growth factors, such as TGF-β (19). Therefore, we hypothesized that molecule 1 also has the ability to stimulate wound cell proliferation. Thus, we evaluated the capabilities of molecules 1 and 2 to promote cell proliferation. First, primary human umbilical vein endothelial cells (HUVECs) were treated with different concentrations (1-1000 nM) of self-assembling peptide 1 and peptide 2 (Ac-VRKIEIVRKK, PDGFR binding domain). As shown in Figure 3A, both 1 and 2 can stimulate cell proliferation at the measured concentration, and the optimal concentration was approximately 1.0 nM, which is comparable with PDGF protein (Figure S3). Interestingly, 1 still exhibited the capability to enhance cell proliferation even at concentrations as high as 1000 nM. Since serum protein is an essential component for cell growth in culture and might affect the activity of the designed peptide, it is more meaningful to evaluate the functions of PDGF-mimicking peptides in the absence of serum protein. Therefore, we assessed the abilities of peptide 1-2 to stimulate HUVEC proliferation in serum-free culture medium for 24 hours. At a concentration of 1.0 nM, 1 showed a significant proliferation-promoting effect compared with the blank. The effect was similar to that of PDGF-B protein and 2, indicating that both 1 and 2 had PDGF-B protein-like biological activity (Figure 3B).
The migration of skin cells is very important for wound healing. Previous reports have demonstrated that PDGF protein can increase cell migration to promote wound healing (30). We used a scratch migration assay to study the effects of 1 and 2 on HUVEC migration. HUVECs were incubated with 1, 2, and PDGF-B protein at a concentration of 1.0 nM for 24 hours. It is clear that peptide 1 increased the cell migration rate significantly; this increase was slightly higher than that of 2 and comparable to that of PDGF-B protein (Figure 3C, E). This result further proved that peptide 1 can mimic the function of PDGF-B protein to promote cell migration.
Microvessel formation plays a vital role in wound healing and tissue regeneration (31); thus, we examined the effect of PDGF-mimicking peptides on angiogenesis. Similar to the untreated group, the capillary network treated with 2 looked disorganized, finer, and discontinuous. Tissues stimulated by 1 and PDGF-B proteins formed complex, highly branched capillary-like structures, with increased branching and cross-linking of blood vessels. Quantifying the distance between the branch points along the blood vessels (called the “branch interval”), 1 and PDGF-B proteins showed reduced branching intervals, indicating a significant increase in driven vascular branches (Figure 3F). In contrast, there was no significant difference in branch interval between Group 2 and the untreated group, likely because peptide 1 self-assembles to form a b-sheet structure. Taken together, these results showed that 1 has great potential in wound healing; both 1 and 2 can promote cell proliferation and migration, but molecule 1 is superior to 2 in stimulating HUVECs to form more vascular branches.
Wound healing and angiogenesis in vivo
To investigate the efficacy of Gel 1 on wound healing, we established a full-thickness skin wound mouse model. Gel 1 was applied locally on the wound surface once during the whole experiment, and the wound area was monitored as a function of time. PBS, Sol 1, and PDGF-B were also used for comparison. It is clear that the wound area of each group decreased significantly at Day 7, and all the wounds had basically healed at Day 12 (Figure 4A). During the wound healing process, all mice were still alive, and no adverse effects were observed. Notably, no significant differences were observed between treatment groups, and all mice healed after 12 days (Figure 4B). In contrast, the new epidermis of the Gel 1 group was thicker than that of the other three groups, implying that Gel 1 increases epithelialization because it can be slowly released to promote cell proliferation.
To further assess the efficacy of Gel 1 on wound healing, we collected the tissues around the wound and stained them with H&E (Figure 4C). Compared with the PBS group, more neovascularization was observed in Gel 1 after 3 days. The regenerated epidermis was observed on Day 7, and the epidermis in all groups was completely remodelled on Day 12. Additionally, the regenerated epidermis in the Gel 1 group was thicker than that in PBS. Furthermore, a more regenerated papillary layer, sebaceous glands, and other accessory organs were observed. Therefore, Gel 1 accelerated the epithelialization of regenerated tissue and exhibited a better therapeutic effect. Collagen deposition during wound healing was examined using Masson staining (Figure 4D). The amount of collagen increased gradually in each group during the experiment. In the PBS group, only a few collagen bundles were formed and arranged loosely. The Gel 1 group had larger collagen deposition areas, indicating that Gel 1 promoted collagen deposition in the wound site. Unlike the PBS group, the collagen fibres were much clearer and more organized. Since blood vessels transport oxygen and nutrients to the wound tissue and play a key role in wound healing (32), the expression level of the vascular endothelial-specific marker CD31 was examined (33). Immunohistochemistry staining (Figure S4) on Day 3 showed that the density of neovascularization in the Gel 1 group was significantly higher than that in the PBS group and even higher than that in the Sol 1 and PDGF-B groups. These results indicated that Gel 1 provides a favourable microenvironment and stimulates angiogenesis for wound healing. Ideally, Gel 1 can slowly dissociate to release 1, prolong its action time to match the wound healing period and stimulate wound healing. However, it is worth noting that the wound healing rate of normal mice is high, and PDGF-B is unable to accelerate this process. In addition, a full-thickness skin wound mouse model with simple resection may not be the best model because of its great limitations. Unlike humans, contraction is significant in the skin healing process of rodents, which accelerates wound closure (34). This may be one of the reasons why there was no significant difference in the wound healing rate in vivo. Wang’s splint wound model can limit wound closure caused by skin contraction, thus simulating the healing process of human granulation growth and re-epithelialization and reducing differences in in vivo experiments (35). Therefore, studying the effect of Gel 1 in a mouse splint model or chronic wound models, such as diabetic wounds (36) or burn wound models, would be worthwhile (37).