**2. Results**

#### *2.1. Longitudinal SphKs and S1PRs Expression during Mouse Wound Healing*

To test our hypothesis, we first investigated S1P signaling during wound healing using the murine excisional wound splinting model [25]. SphK1 expression in the wound started increasing on Day 2 and peaked on Day 5 (88.6-fold increase compared with immediately after the wounds were generated) (Figure 1A). SphK2 expression did not change (Figure 1B). Interestingly, the expression of S1PR2 (which inhibits S1PR1 and S1PR3 signaling [12]) gradually increased towards the end of the wound healing process. S1PR1 expression did not change significantly (Figure 1C). S1PR3 expression was not detected at any time point. Thus, SphK1, but not S1PR1, is massively upregulated in the proliferative phase of wound healing.

**Figure 1.** Longitudinal sphingosine-1-phosphate (S1P) production during mouse wound healing. Splinted excisional wounds (*n* = 4–6) were generated in C57BL/6J mice, and the mRNA expression of (**A**) sphingosine kinase-1 (SphK1), (**B**) sphingosine kinase -2 (SphK2), and (**C**) sphingosine-1-phosphate reseptor-1/<sup>2</sup> (S1PR1/2) in the wound during wound healing was measured. All values shown in this figure represent the mean ± s.e.m. \* *p* < 0.05, \*\* *p* < 0.01.

#### *2.2. E*ff*ect of SphK1 Gene Knockout on Wound Healing, Vasculogenesis, and Cell Proliferation*

Compared with littermate wild-type (WT) mice, SphK1−/− mice had significantly delayed wound healing, as determined by two-factor repeated measures ANOVA (Figure 2A,B; *p* = 0.010). This reflects the tendency of the SphK1−/− mice to have larger wound sizes on Days 5, 7, and 9 after injury, as determined by Student's *t*-test (*p* = 0.056, 0.262, and 0.068, respectively). We investigated vasculogenesis on Day 5 by immunohistochemistry against CD34, which is an early marker of vasculogenesis [26]. The SphK1−/− mice exhibited significantly less angiogenesis than the WT mice (Figure 2C–E). Immunohistochemistry with Ki67 showed that SphK1 knockout also had similar suppressive effects on the proliferation of both the fibroblasts in the wound (Figure 2F) and the keratinocytes at the wound edge (Figure 2G). The Ki67<sup>+</sup> cells in the latter analyses are expressed as the percentage of Ki67<sup>+</sup> cells/total cells per field.

**Figure 2.** *Cont*.

**Figure 2.** Effect of SphK1 knockout on wound healing, vasculogenesis, and cell proliferation. (**A**) Splinted excisional wounds were generated in SphK1−/− and SphK1+/+ mice. Representative images of the closing wounds are shown. (**B**) Change in wound area over time (*n* = 6–10). ( **C**) Representative images of CD34 expression on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). ( **D**) The numbers of CD34-positive microvessels per 200-fold magnified field are graphed (*n* = 4). (**E**) The percentage of the wound area that is occupied by CD34+ cells is graphed (*n* = 4). (**F**) Representative images of Ki67 expression in fibroblasts on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). The frequencies of Ki67<sup>+</sup> fibroblasts are graphed (*n* = 4). ( **G**) Representative images of Ki67 expression in keratinocytes on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). The frequencies of Ki67<sup>+</sup> keratinocytes are graphed (*n* = 4). All values shown in this figure represent the mean ± s.e.m. \* *p* < 0.05, \*\* *p* < 0.01.

#### *2.3. E*ff*ect of SphK1 Gene Knockout on Inflammatory Cell Recruitment during Wound Healing*

Immunohistochemistry showed that the SphK1−/− mice exhibited significantly decreased macrophage numbers compared to the WT mice on Day 5 (Figure 3A–C). Flow cytometric analyses confirmed that the SphK1−/− mice had significantly lower frequencies of T cells in the wound five days after injury (Figure 3D,E).

**Figure 3.** Effect of SphK1 knockout on inflammatory cell recruitment during wound healing. (**A**) Representative images of F4/80 expression on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). (**B**) The percentage of the wound area that is occupied by F4/80<sup>+</sup> cells is graphed (*n* = 4). ( **C**) The number of F4/80<sup>+</sup> cells per field is graphed (*n* = 4). ( **D**,**E**) Representative flow cytometric plots (D) and frequency of the indicated T cell populations (E) on Day 5 after wounding, as determined by flow cytometry (*n* = 5–8). All values shown in this figure represent the mean ± s.e.m. \*\* *p* < 0.01.

#### *2.4. E*ff*ect of Nanoparticle-Mediated Topical SphK1 Gene Delivery on Wound Healing, Vasculogenesis, and Cell Proliferation*

We generated control and SphK1-expressing plasmids that were encapsulated with super carbonate apatite (sCA). sCA is a nanoparticle that is safe for in vivo gene delivery. The in vitro and in vivo safety of sCA-mediated gene delivery has been reported [27]. When mixed with sCA, the control and SphK1-expressing plasmids transfected mouse dermal fibroblast NIH3T3 cell lines in vitro with high efficiency (Figure 4A). Ointments containing the sCA-encapsulated plasmids were then generated and applied topically to the wounds of wound splinting model mice (Figure 4B). V5-tag protein expression analysis showed that the plasmids had a high transfection rate in vivo (Figure 4C). Compared with the vector, the SphK1 plasmid significantly accelerated wound closure, as determined by two-factor repeated measures ANOVA (*p* < 0.0001, Figure 4D,E). This reflected significantly greater closure on Days 7 and 9 after injury (*p* = 0.003 and 0.0002, respectively), as shown by Student's *t*-test (the vector and SphK1 plasmid did not differ significantly in terms of Day 5 closure rate; *p* = 0.122). We then investigated vasculogenesis and cell proliferation on Day 5 by immunohistochemistry. The SphK1 plasmid ointment significantly accelerated vasculogenesis (Figure 4F–H). Immunohistochemistry with Ki67 showed that the SphK1 plasmid ointment had corresponding positive effects on the proliferation of both the fibroblasts in the wound (Figure 4I) and the keratinocytes at the wound edge (Figure 4J). The Ki67<sup>+</sup> cells in the latter analyses are expressed as the percentage of positive cells/total cells per field.

**Figure 4.** *Cont*.

**Figure 4.** Effect of SphK1 overexpression on wound healing, vasculogenesis, and cell proliferation. (**A**) In vitro transfection efficiency with SphK1-expressing plasmid using super carbonate apatite (sCA) in NIH3T3 cells. (**B**) An ointment containing a SphK1-expressing plasmid encapsulated with sCA was prepared. (**C**) In vivo transfection efficiency of the sCA-encapsulated plasmid, as shown by immunoblots of V5-SphK1 expression in the wound surface tissues two days after application. (**D**) Representative images of the closing wounds are shown. (**E**) Effect of the plasmid ointment on wound closure. The change in wound area over time is graphed (*n* = 12). (**F**) Representative images of CD34 expression on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). (**G**) The numbers of CD34-positive microvessels per 200-fold magnified field are graphed (*n* = 4). (**H**) The percentage of the wound area that is occupied by CD34+ cells is graphed (*n* = 4). (**I**) Representative images of Ki67 expression in fibroblasts on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). The frequencies of Ki67<sup>+</sup> fibroblasts are graphed (*n* = 4). (**J**) Representative images of Ki67 expression in keratinocytes on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). The frequencies of Ki67<sup>+</sup> keratinocytes are graphed (*n* = 4). All values shown in this figure represent the mean ± s.e.m. \* *p* < 0.05, \*\* *p* < 0.01.

#### *2.5. E*ff*ect of Nanoparticle-Mediated Topical SphK1 Gene Delivery on Inflammatory Cell Recruitment during Wound Healing*

Immunohistochemistry showed that the SphK1 plasmid-treated wounds had significantly higher macrophage numbers on Day 5 (Figure 5A–C). Moreover, the SphK1 plasmid ointment increased the recruitment of total T cells, CD4 T cells, and CD8a T cells in the wound five days after injury (Figure 5D,E). It should be noted that uninjured SphK1−/− mice exhibit normal lymphocyte trafficking despite the fact that their blood S1P levels are about half of those in WT mice [28]. Thus, our experiments sugges<sup>t</sup> that SphK1 participates in the recruitment of inflammatory cells to the wound, and that this is needed for the normal progression of the proliferative phase of wound healing. These results are consistent with our hypothesis that after wounding, S1P generated by SphK1 promotes vasculogenesis and recruits inflammatory cells, including lymphocytes and macrophages, and that this facilitates the wound healing process during the proliferative phase. Furthermore, immunoblot analyses showed that the SphK1 plasmid ointment increased expression of the well-known wound healing-related factors VEGF, FGF-2, and IGF-1 [29–31] in the wound on Day 5 (Figure 5F,G). These findings sugges<sup>t</sup> that these wound-related factors were secreted by the recruited lymphocytes and macrophages.

**Figure 5.** Effect of SphK1 overexpression on inflammatory cell recruitment and enhanced wound-related factors. (**A**) Representative images of F4/80 expression on immunohistochemistry are shown. Arrowheads indicate positive findings (scale bars: 50 μm). (**B**) The percentage of the wound area that is occupied by F4/80<sup>+</sup> cells is graphed (*n* = 4). (**C**) The number of F4/80<sup>+</sup> cells per field is graphed (*n* = 4). (**D**,**E**) Representative flow cytometric plots (D) and frequency of the indicated T cell populations (E) on Day 5 after wounding, as determined by flow cytometry (*n* = 5–8). (**F**) Effect of the plasmid ointment on the expression of the indicated wound healing-related factors on Day 5 after wounding, as determined by immunoblot analysis. (**G**) The immunoblots were quantified and the data were graphed (*n* = 3). All values shown in this figure represent the mean ± s.e.m. \* *p* < 0.05, \*\* *p* < 0.01.

#### *2.6. E*ff*ect of SphK1 Overexpression on Granuloma Formation*

When we injected sponge granulomas in mice with the sCA-encapsulated vector or SphK1 plasmid every other day, as described previously [32], the SphK1 plasmid generated clearer collagen bundles, higher fibroblast density, and less dead cell accumulation in the center of the sponge on Day 14 (Figure 6A). Moreover, on Day 14 after injury, the SphK1 plasmid associated with significantly more granulation than the control plasmid (Figure 6B).

**Figure 6.** Topical SphK1 gene delivery promotes granulation. (**A**) Representative images of the hematoxylin and eosin (HE)-stained sponge granulomas treated with sCA-encapsulated plasmid injection on Day 14 are shown. The black boxes are shown magnified. (scale bars: Perspective: 1 mm; LPF: 400 μm; HPF: 50 μm). (**B**) Percentage of granulated area is graphed (*n* = 8). All values in this figure represent the mean ± s.e.m. \* *p* < 0.05.

#### *2.7. E*ff*ect of SphK1 and S1PR2 Gene Expression on Scar Thickness, the Interaction between Transforming Growth Factor (TGF)-*β*1 and S1P*

We treated the dermal fibroblast line NIH3T3 with SphK1 plasmid or exogenous S1P. We found that exogenous S1P, but not the SphK1 plasmid, suppressed the transcription of Collagen1a1 and Collagen3a1 in the cells (Figure 7A,B). Thus, exogenous S1P, but not endogenously produced S1P, prevents the collagen deposition of dermal fibroblasts. Notably, when the exogenous S1P-stimulated cells were treated with the S1PR1 and S1PR3 inhibitor VPC23019 or the S1PR2 inhibitor JTE013, their collagen production was restored (Figure 7C). Thus, exogenous S1P suppresses the collagen deposition of dermal fibroblasts via S1PR signaling. Transforming growth factor (TGF)-β1, which is produced during the proliferative phase of wound healing, induces fibroblasts to produce granulation tissue in vivo and extracellular matrix in vitro [33,34]. Our finding that endogenous S1P, but not exogenous S1P, also promotes granulation and participates in the proliferative phase of wound healing led us to examine the effect of TGF-β1 treatment on S1PR expression by NIH3T3 cells. We found that this treatment significantly suppressed transcription of S1PRs (Figure 7D). This suggests that TGF-β1 is a key modulator of the ability of S1P to promote the proliferative phase of wound healing. Given that S1PR2 expression in the wound increased around the end of wound closure (Figure 1C), S1PR2−/− mice had significantly smaller wounds on Day 12 after injury (Figure 7E,F). This suggests that S1PR2 signaling negatively regulates SphK1-S1PR1 signaling, thereby slowing down wound closure at the end of the proliferative phase and allowing the wound to prepare for the remodeling phase. Interestingly, we discovered that when the wounds in the mouse excisional wound splinting model were treated with SphK1-sCA ointment, the scars that formed when epithelization was completed were much thinner than the scars of the vector-sCA-treated mice. The SphK1-sCA-treated wounds also had much thinner collagen bundles, as shown by high power field images (Figure 7G,H).

**Figure 7.** Topical SphK1 gene delivery inhibits scarring. The mRNA expression of Collagen1a1 and Collagen3a1 in NIH3T3 cells (**A**) stimulated with the indicated concentration of S1P for 24 h (*n* = 4) or (**B**) transfected with SphK1-expressing plasmid or vector plasmid (*n* = 4). (**C**) The mRNA expression of Collagen1a1 and Collagen3a1 in NIH3T3 cells stimulated with 1 μM S1P for 24 h with or without 10 μM VPC23019 (inhibitor of S1PR1 and S1PR3) or JTE013 (inhibitor of S1PR2) (*n* = 3). (**D**) The mRNA expression of the indicated S1PRs in NIH3T3 cells stimulated with the indicated concentration of transforming growth factor (TGF)-β1 for 18 h (*n* = 3). (**E**) Splinted excisional wounds were generated in S1PR2−/− and S1PR2+/+ mice. Representative images of the closing wounds are shown. (**F**) The wound area over time was measured (*n* = 6). (**G**) Representative images of Masson's trichrome-stained scars at the point of epithelization after treatment with sCA-encapsulated plasmid ointment (the scale bars are LPF: 400 μm; HPF: 50 μm). "D" indicates the scar thickness. (**H**) The scar thickness was measured and graphed (*n* = 4–6). All values in this figure represent the mean ± s.e.m. \* *p* < 0.05, \*\* *p* < 0.01.
