4-Aminopyridine (4-AP) expedites wound closure and enhances skin regeneration. The beneficial effects of 4-AP on healing of skin wounds were first noticed in studies monitoring wound closure (Fig. 1A-C). We created 5-mm full thickness dorsal excisional wounds in 10-wk-old male C57BL/6 mice34; mice were then randomized and treated with either saline or systemic 4-AP20,23,29,35,36 daily for 14 days. Wounds were splinted with silicone rings to prevent wound contraction and were monitored by digital imaging for morphometry, percentage of wound healing and tissue regeneration on days 3, 5, 7, 9, 12 and 14 post wound (PWD).
We found that the extent of wound closure in 4-AP-treated mice was more than twice that of saline treated mice at PWD3 (day 3: 22.35 ± 0.30% vs. 9.88 ± 0.28%; P < 0.0199). Significant differences in wound closure were found at every time point examined, including PWD14 (P-values < 0.001). At this point, 4-AP-treated mice had complete wound closure, while saline-treated mice still showed incomplete healing (day 14: 98.24 ± 0.10% vs. 75.28 ± 0.32%; P < 0.0001) (Fig. 1B and C).
Analysis of tissue at PWD14 revealed that 4-AP treatment also increased epidermal thickness to that of uninjured tissue (Fig. 1D and E). Wound tissue was analyzed by morphometric analysis on hematoxylin and eosin (H&E) stained sections2,4,34. The newly formed epidermis in 4-AP-treated mice was slightly thicker than uninjured tissue (29.74 ± 1.30 µm in 4AP-treated mice vs. 25.41 ± 0.12 µm in uninjured mice; P = 0.0033) and was significantly thicker than in saline-treated mice (19.38 ± 0.96 µm; P = 0.0067).
4-AP treated mice also showed a significant increase in wound-induced hair neogenesis (WIHN) when compared with saline-treated mice (Fig. 1D and F), a feature known to correlate with successful skin regeneration. 4-AP-treated mice showed a 1.8-fold increase in hair follicles over saline-treated mice (45 ± 3 vs. 25 ± 3 follicles/400 µm2; P = 0.0067) (Fig. 1F)37,38.
4-AP increases keratinocyte number and epithelial stem-cell markers in healed wounds. We observed 2.3-fold increases in keratin-14 positive (K14+) keratinocytes in the epidermis and in de novo hair follicles of 4-AP treated mice (Fig. 2A) compared with saline-treated mice (cells: 16.65 ± 0.97% cells/mm2 vs. 7.23 ± 0.73% cells/mm2; P < 0.0001) and (protein: 186666667 ± 25271955 integrated density/mm2 vs. 70298114 ± 10356742 integrated density/mm2, P = 0.0063) (Fig. 2B and C). In contrast, expression of K10, a marker of epidermal differentiation, was not impacted by 4-AP treatment at this time point (Supplementary Fig. S1A and B).
In agreement with the increase in hair follicles seen with H&E staining (Fig. 1D and F), we also saw increases in K17+ and K15+ cells and protein expression, which are markers of hair follicles. We saw a 1.5-fold increase in the percentage of K17+ cells as a percentage of total DAPI+ cells (cells: 33.95 ± 2.80% cells/mm2 vs. 22.94 ± 1.49% cells/mm2; P = 0.0014), and in K17 protein (protein: 132329400 ± 12238815 integrated density/mm2 vs. 50989532 ± 3610166 integrated density/mm2, P < 0.0001) (Fig. 2D, E and F). Similarly, there were >115% increases in K15+ cells (cells: 16.05 ± 1.38% cells/mm2 vs. 7.44 ± 0.75% cells/mm2, P < 0.0001), a marker of hair follicle bulge stem cells37–39. There were also increases in K15 protein (protein: 71374407 ± 5506935 integrated density/mm2 vs. 23022240 ± 2208956 integrated density/mm2; P < 0.0001) (Fig. 2G, H and I).
K14 and K17 expression also increased in the overlying epidermis in both saline and 4-AP-treated mice, which is consistent with their known expression following wounding (Fig. 2A and D)39.
4-AP treatment promotes increases in fibroblasts, myofibroblasts and transforming growth factor-β (TGF-β). Fibroblast migration and maturation contribute to contraction, granulation, and proliferation phases of wound healing. We therefore next examined effects of 4-AP treatment on fibroblasts and on a known regulator of fibroblast differentiation, TGF-β.
To test whether 4-AP treatment altered fibroblast maturation during wound healing, we first performed Masson’s Trichrome staining to examine collagen deposition in the healing wound (Fig. 3). This staining revealed elevated collagen deposition in 4-AP treated mice compared to saline treated mice (P = 0.0025; Fig. 3A and B), with collagen levels like those seen in normal tissue. This staining also revealed a tissue structure and collagen deposition pattern very much like that seen in normal tissue.
4-AP treatment also increased the expression of fibroblast proteins, vimentin and α-smooth muscle actin (α-SMA). Immunofluorescence analysis revealed more vimentin+ fibroblasts and elevated vimentin levels in wound tissue from 4-AP treated mice than saline treated mice (Fig. 3C and D). We also observed increases in α-SMA, which signifies fibroblast differentiation into collagen-producing myofibroblasts2–4. Increases were seen in the number of α-SMA+ cells and in α-SMA protein (P = 0.0063; Fig. 3C and E).
TGF-β plays an important role in promoting myofibroblast differentiation2–4, and we found significant increases in TGF-β protein expression with 4-AP treatment compared to saline treatment (P = 0.0001; Fig. 3F and G).
4-AP promotes reinnervation and neuropeptide expression. Normal skin wound healing is also associated with increases in cell division and increases in non-dividing neurons. 4-AP treatment caused increases in both of these measures.
Expression of the proliferation marker, Ki-67, was significantly increased in mice treated with 4-AP compared with saline treated controls. The proportion of Ki-67+ cells within hair follicles and epidermis was increased 2.8 fold (cells: 21.87 ± 2.763 cells/mm2 vs. 7.754 ± 1.664% cells/mm2; P = 0.0001) (Fig. 4A and B).
The number of neurons in the skin of 4-AP-treated mice also was increased over that seen in saline-treated animals. Neuronal number was determined by staining with antibodies against high molecular weight neurofilament protein (NF-H)2,8,24,40. NF-H axonal counts were increased 2.5 fold (cells: 194 ± 50.51 count/mm2 vs. 74.19 ± 15.85 count/mm2; P = 0.007) in 4-AP treated mice compared with saline treated controls (Fig. 4A and C).
We also found that NF-H stained axons in the 4-AP treated mice were more often encountered in direct association with Ki-67+ hair follicles (Fig. 4A) than in saline treated controls – an important qualitative finding given that hair follicles are known to be associated with sympathetically innervated arrector pili muscles39.
Another example of the ability of 4-AP to restore aspects of skin structure like that seen in uninjured tissue was revealed by staining for protein gene product 9.5 (PGP9.5), a neuronal peptide that promotes wound healing13,41,42. Fourteen days post wounding, PGP9.5+ nerve fibres in the healed wounds were twice as abundant in 4-AP treated mice, as reflected by increased amounts of PGP 9.5 (protein: 4096938 ± 713297 integrated density/mm2), compared with saline treated mice (protein: 2107970 ± 325039 integrated density/mm2) (P = 0.0012; Fig. 4D and E). The levels of PGP 9.5 in 4-AP-treated mice were not significantly different from seen in uninjured tissue (Supplementary Fig. S1C and D).
4-AP increases numbers of Schwann cells (SC) and expression of markers of an early differentiation state. Schwann cells (SC) are critical players in wound healing and are associated with axons around hair-follicles in the wound bed. We found that the number of SCs was significantly increased in the wounds of 4-AP-treated mice. Analysis of expression of S100, a pan-SC marker2,28, identified SCs within both the hypodermis and dermis of the healed wounds (Fig. 5A).
The number of SCs was 3-fold greater in 4-AP treated mice than in saline treated controls (cells: 6.483 ± 1.163% cells/mm2 vs. 2.242 ± 0.3159% cells/mm2; P = 0.0007) (Fig. 5A-B). SCs were preferentially located around nerve bundles, as predicted by the known affiliation of SCs with nerve cells. 4-AP treatment also increased the expression of p75-NTR (Fig. 5A, C-F), which is thought to be expressed in SC-100+ cells as a marker of de-differentiation2,43,44.
We also found elevated expression of SOX10, substance-P, and NGF in 4-AP treated mice. SOX10 is required for myelin production in SCs45 and elevated SOX10 expression promotes conversion of mesenchymal cells into p75-NTR expressing neural crest stem cells (NCSC)45,46. Conversely, depletion of SOX10 expression significantly delays wound healing and tissue regeneration2. NGF plays a significant role in the wound healing process9,40,43,44,46−48 by inducing nerve sprouting from injured nerve endings. NGF also promotes keratinocyte proliferation49, and migration of dermal fibroblasts7. NGF also acts on non-neuronal cells to sensitize them to substance-P, which in turn further stimulates more NGF secretion48,50 ensuring that keratinocytes, for example, can elaborate and respond to neuronal factors along with neurons.
We found significantly increased SOX10 in both immunofluorescence and western blot analysis (immunofluorescence: 3726901 ± 151884 vs. 3125946 ± 62780 integrated density/mm2; P = 0.0018 and western blot analysis: 6312394 ± 114415 vs. 3780312 ± 437118 normalized integrated density; P = 0.0304) (Fig. 6A, B, E and F), substance-P (immunofluorescence: 3916297 ± 247329 vs. 3120195 ± 107296 integrated density/mm2; P = 0.0063) (Fig. 6A, C), and NGF (4-fold increase in immunofluorescence: 2429872 ± 375280 vs. 626967 ± 101856 integrated density/mm2; P = 0.0002 and in western blot analysis: 4337935 ± 103543 vs. 2645874 ± 129171 normalized integrated density; P = 0.0005) (Fig. 6A, D, E and G) expression in healed wounds from 4-AP treated mice compared to saline treated mice (Fig. 6). In all cases, 4-AP increased these neuronally active molecules to levels found in uninjured skin.
4-AP promotes neo-angiogenesis in granulation tissue. Neo-angiogenesis is necessary to provide nutrients and oxygen to healing wounds. To assess whether 4-AP treatment enhanced wound neo-vascularization, we performed immunofluorescence staining for the endothelial specific marker, CD314.
We observed larger and more abundant blood-vessel networks in dermal tissue from 4-AP treated mice than saline treated mice (Supplementary Fig. S1E). Although the number of blood vessels in 4-AP-treated mice was still less than in uninjured tissue, vessels in the newly healed tissue were notably larger. Quantification of CD31 intensity revealed statistically significant increases in blood vessels with 4-AP treatment compared to saline treatment (Supplementary Fig. S1F).
4-AP effectively stimulates proliferation and migration in primary cultures of human skin derived primary cells in-vitro. We next found that effects of 4-AP on keratinocytes, fibroblasts and SCs in-vitro were similar to outcomes observed in-vivo, suggesting 4AP may have direct effects on these cell types. In these experiments, we cultured primary, normal human epidermal keratinocytes (NHEKs)51,52, fibroblasts51 and dermal SCs53 in the presence or absence of 4-AP.
We first confirmed the purity and identity of each cell type with immunohistochemistry for characteristic markers (Supplementary Fig. S2A-C). There was a modest decrease in MTT staining at higher 4-AP concentrations, but no obvious decrease in cell health up to 2 mM concentrations of 4-AP used in previous work on cells in-vitro24(Supplementary Fig. S2D-F). Automated wound scratch assays were performed19,54,55 on confluent monolayers of keratinocytes, fibroblasts, and SCs, with and without 1 mM 4-AP24,30,56, a standard 4-AP dose used in studies in-vitro. 4-AP exposure accelerated scratch closure and keratinocyte migration (Fig. 7A and B; Supplementary Movie 1 and 2) within 3 hours, with complete scratch closure occurring at 18 hours. In contrast, control cultures not exposed to 4-AP closed at 32 hours (Fig. 7A and B).
SOX10 and NGF expression in 4-AP treated keratinocytes were both increased (Fig. 7C-F). Expression of K14 and K17, which are associated with basal, proliferating keratinocytes, were also increased (Supplementary Fig. S3). These data suggest that 4-AP promotes a more proliferative phenotype in keratinocytes, as is necessary for accelerated scratch closure.
In contrast with the effects on keratinocytes, fibroblast migration was unaffected by 4-AP (Supplementary Fig. S4A and B). 4-AP did, however, induce a conversion of fibroblasts to myofibroblasts with increased vimentin and α-SMA protein expression (Supplementary Fig. S4C; and Supplementary Movie 3 and Supplementary Movie 4).
In SCs, 4-AP accelerated scratch closure (80% by 11 hours with 4-AP treatment compared with 23 hours without treatment) (Supplementary Fig. S5A-B and Supplementary Movie 5 and Supplementary Movie 6). 4-AP treatment also increased SOX10, p75-NTR and NGF expression compared to controls (Supplementary Fig. S5C-E).
We next found that co-culturing cells accelerated scratch culture closure rates and that 4-AP exposure further increased these effects. Given that keratinocytes, fibroblasts and SCs cells all interact during wound healing in-vivo, we conducted co-culture experiments with pairs of cell types55 with and without 4-AP treatment to determine if 4-AP promotes cellular interactions. Keratinocytes co-cultured with SCs in a ratio mimicking that of epidermal skin (10:1 ratios of keratinocytes:SCs) closed a scratch within 15 hours with 4-AP treatment, which was significantly faster than scratch closure without treatment, which took 20 hours. At 15 hours, the percentage of wound closure was at 98% in 4-AP treated cultures but only 80% in non-treated scratch cultures (Fig. 8A and B; and Supplementary Movie 7 and Supplementary Movie 8).
The ability of 4-AP to accelerate scratch closure in cultures of keratinocytes and SCs was not unique to this cell combination. We also found 4-AP treatment led to faster scratch closure in keratinocytes co-cultured with fibroblasts (ratio, 10:1) (measured at 15 hours: 93% in 4-AP treated vs. 72% in non-treated) (Fig. 8C and D; and Supplementary Movie 9 and Supplementary Movie 10).