Interleukin-33 encourages scar formation in murine fetal skin wounds

Abstract

The magnitude of the inflammatory response after skin injury is important for determining whether wounds in developing fetal skin will heal scarlessly (minimal inflammation) or with prominent scars (robust inflammation). One class of inflammatory mediators gaining attention for their role in wound inflammation is alarmins. In the current study, the alarmin IL-33 was examined in a mouse model of fetal wound healing. IL-33 expression was elevated in scar-forming embryonic day 18 wounds compared to scarless embryonic day 15 wounds. Furthermore, injection of IL-33 into embryonic day 15 wounds caused scarring when wounds were analyzed at 7 days post-wounding. The introduction of IL-33 into embryonic day 15 wounds did not induce statistically significant changes in the number of neutrophils, mast cells, or macrophages in vivo. However, IL-33 treatment enhanced collagen expression in cultured fibroblasts derived from adult and fetal murine skin, suggesting that IL-33 may directly stimulate fibroblasts. In vitro studies suggested that the stimulation of collagen production by IL-33 in fibroblasts was partially dependent on NF-κB activation. Overall, the data suggest an association between IL-33 and scar formation in fetal wounds.

INTRODUCTION

Cutaneous repair typically includes periods of inflammation, proliferation, and scar formation.^{1, 2} Scarring can lead to serious functional impairments as well as psychosocial problems. In contrast to developed skin, fetal skin heals with minimal inflammation and no scarring at early stages of development.^{3, 4} This unique healing ability is lost as the skin becomes more developed, with wounds in late-gestation fetal skin inducing a vigorous inflammatory response and subsequent scarring. The absence of significant inflammation in early-gestation wounds is known to be a key feature in enabling scarless healing.^{5, 6} The importance of this minimal inflammatory response is highlighted by the fact that stimulating inflammation in early-gestation wounds that would otherwise heal without a scar induces a fibrotic response.^6–9

Alarmins are a group of endogenous inflammatory mediators that stimulate inflammation in response to tissue damage, even in the absence of microbes.^{10, 11} Alarmins are typically intracellular host proteins that are not in contact with inflammatory cells under normal conditions. However, upon cell damage or stimulation, alarmins can be released extracellularly where they are recognized by inflammatory cells and trigger inflammation.¹⁰ Alarmins include proteins such as HMGB-1 (high-mobility group box-1), S100 proteins, heat shock proteins, and several IL-1 cytokine family members, including IL-33 (interleukin-33).^10–13

IL-33 normally resides in the nucleus where it regulates gene transcription, but it functions as a pro-inflammatory alarmin when released outside of the cell.^{11, 13} Extracellular IL-33 stimulates inflammation by binding to ST2 (suppression of tumorigenicity 2) and signaling through IL-1RAcP (IL-1 receptor accessory protein), resulting in MAPK signaling and NF-κB activation.^{11, 13} The goal of this study was to assess the role of IL-33 in fetal wound healing. The rationale for examining IL-33 was based on two previous observations. First, studies have shown that the alarmin HMGB-1 plays a role in the transition from scarless to fibrotic healing in fetal wounds,⁷ suggesting that other alarmins such as IL-33 could also be important. Second, IL-33 has been shown to stimulate mast cells,^14–16 which are known to promote scar formation in late-gestation fetal wounds.¹⁷ In the current study, IL-33 expression was compared in scarless and fibrotic fetal wounds and the effects of IL-33 on scar formation were determined.

MATERIALS AND METHODS

Animal experiments

A murine model was used to examine the role of IL-33 in fetal wound healing. Surgical procedures were approved by The Ohio State University Institutional Animal Care and Use Committee. Full-thickness dorsal skin wounds were created in utero on fetuses at E15 or E18 in FVB mice (Taconic, Germantown, NY) as described previously.¹⁸ Some experiments were also performed using Col3.6-GFPtpz mice, which express green fluorescent protein (GFP) under the control of the collagen 1a1 promoter, as indicated in the results.^{19, 20} At these ages, skin wounds undergo scarless (E15) or fibrotic (E18) healing.⁶ 10% India ink in sterile saline was injected subcutaneously to mark the wound. Some wounds received 800 ng recombinant murine IL-33 (R&D Systems, Minneapolis, MN) diluted in 10% India ink. This IL-33 dose was within the range of the effective injection doses used for other models in published studies.^{14, 15} Incisional wounds were also generated on adult FVB mice as described previously.¹⁸ Skin samples from uninjured animals (control tissue) and wounds were harvested for analysis at various time points post-wounding. Samples were fixed in 10% buffered formalin overnight and embedded in paraffin or frozen in TBS tissue freezing media (Triangle Biomedical Sciences, Durham, NC) for histology and immunohistochemistry. Sample numbers vary for each parameter examined and are reported in the figure legends. Each sample represents tissue isolated from a separate animal.

Immunohistochemistry and histology

Cryosections (10 μm) were used for the detection of IL-33 and F4/80 (macrophages) by immunohistochemistry following procedures outlined previously,⁷ except that the following primary antibodies were used: IL-33 (R&D Systems) and F4/80 (Abcam, Cambridge, MA). Paraffin sections were used for the detection of mast cells by toluidine blue staining and Ly-6G (neutrophils; BD Biosciences, San Jose, CA) by immunohistochemical staining as described.^{7, 17} For quantification of cell density, cells were counted in digital images taken in high-power fields immediately adjacent to either side of the wound and area was determined using Axiovision software (Carl Zeiss Imaging Solutions, Thornwood, NY). Cell densities (cell number per mm²) were then calculated. Cell densities for the right and left sides of each wound were averaged to obtain the mean cell density for each sample.

For immunofluorescence staining of IL-33 and GFP and immunostaining for ST2, paraformaldehyde-fixed frozen sections were used in conjunction with primary antibodies specific for IL-33 and ST2 (R&D Systems) or GFP (Abcam) and appropriate Alexa Fluor-conjugated (Thermo Fisher) or biotinylated (Vector Laboratories) secondary antibodies, using standard methods. Staining was visualized using a Nikon A1Rsi resonant scanning confocal microscope.

Evaluation of scar tissue

Masson’s trichrome staining, which differentially stains collagen blue, was performed in paraffin sections as described¹⁸ to evaluate healing outcomes. The presence of scar tissue was determined based on histological features of scar tissue, including a lack of skin appendages and the replacement of normal dermal tissue with disorganized collagen. Axiovision software was used to measure the width of the scars as described previously.^{7, 17, 18}

Fibroblast studies

Fibroblasts were cultured using skin explants harvested from adult or E15 Col3.6-GFPtpz mice, which express green fluorescent protein (GFP) under the control of the collagen 1a1 promoter.^{19, 20} Explants were cultured in Dulbecco’s Modified Eagles Medium (high glucose) supplemented with non-essential amino acids, sodium pyruvate, L-glutamine, penicillin/streptomycin, and 10% fetal bovine serum (all reagents from Life Technologies, Grand Island, NY) using procedures outlined before.²¹ Cells were expanded and used at passage 3 to assess NF-κB activation, ST2 protein expression, or GFP fluorescence as a read out for collagen expression. Some cells were grown on Lab-Tek chamber slides (Thermo Fisher), fixed in acetone, and stained for ST2 using the immunohistochemistry methods described above.

To assess NF-κB activation, cells were serum starved overnight, then treated with fresh serum-free media containing equal volumes of either PBS as a negative control, 100 ng/ml recombinant murine IL-33 (R&D Systems), or 20 ng/ml murine TNF-α (Peprotech, Rocky Hill, NJ) as a positive control. After 2 hours, nuclear and cytoplasmic proteins were isolated using the NE-PER nuclear and cytoplasmic extraction kit (Thermo Scientific, Rockford, IL) per manufacturer’s instructions. Western blotting was used to detect the presence of the p65 subunit of NF-κB in cytoplasmic and nuclear fractions. Briefly, protein was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (Bio-Rad Laboratories, Hercules, CA). Membranes were blocked and probed using anti-p65 antibodies (Cell Signaling, Danvers, MA). Appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies (Cell Signaling) were used in conjunction with Immun-Star WesternC chemiluminescence reagents (Bio-Rad) for protein detection. Digital images were captured using the Chemidoc XRS imaging system (Bio-Rad). Blots were stripped using Restore PLUS Western blot stripping buffer (Thermo Scientific) and re-probed with antibodies against the cytoplasmic marker α-tubulin (Cell Signaling) to demonstrate effective separation of cytoplasmic and nuclear proteins. Similar methods were used to evaluate ST2 expression by Western blotting, except that total protein lysates were probed with ST2 (R&D Systems) or beta-actin (Cell Signaling) antibodies.

To analyze GFP fluorescence as a measure of collagen 1a1 expression, cells were serum starved overnight then treated with media containing 1% serum plus equal volumes of either PBS as a negative control, 100 ng/ml recombinant murine IL-33 (R&D Systems) with or without the addition of the NF-κB inhibitor JSH-23 (30 mM; Sigma, St. Louis, MO), or 10 ng/ml human TGF-β1 (Peprotech) as a positive control. After 48 hours, cells were trypsinized and fluorescence intensity was determined by flow cytometry using a FACS Calibur (BD Biosciences, San Jose, CA). The mean fluorescence intensity per cell is reported.

Statistical analysis

Data were analyzed using GraphPad Prism (GraphPad Software, Inc., San Diego, CA). Statistical differences were determined by unpaired t-test or ANOVA as appropriate, with p-values < 0.05 considered statistically significant.

RESULTS

IL-33 levels correlate with scar formation in fetal wounds

Immunohistochemistry was used to examine IL-33 expression patterns in fetal skin and wounds. Minimal staining was observed in unwounded fetal skin harvested at E15 or E18 (Fig. 1 A-B). This differed from unwounded adult skin, which showed positive staining for IL-33 in epidermal keratinocytes prior to injury (Fig. 2 A). In fetal wounds, strong nuclear staining was observed in keratinocytes at the wound margin, with more keratinocytes expressing IL-33 for a longer period of time in E18 wounds compared to E15 wounds (Fig. 1 C-G). E18 wounds displayed a similar staining pattern to adult wounds (Fig. 2 B-D), which also had prominent nuclear staining in keratinocytes at the wound edge. The results suggest that higher epidermal levels of IL-33 are present in scar-forming wounds.

Figure 1. IL-33 localization in fetal skin and wounds.

Immunohistochemistry was used to examine IL-33 expression in frozen sections of E15 and E18 unwounded skin and wounds. Representative photomicrographs of E15 (left panels; A, C, E) and E18 (right panels; B, D, F) normal skin (A, B) and wounds harvested at 24 (C, D) or 48 (E, F) hours post-wounding are shown. Wound margins are marked with arrows and scale bars = 100 μm. IL-33-positive keratinocytes were counted at the wound margin and the data are represented graphically in G. Bars represent average number of IL-33-positive keratinocytes per high power field +/− S.E.M. (***p<0.001 by two-way ANOVA with Bonferonni post-hoc testing; n = 3–4 per group).

Figure 2. IL-33 localization in adult skin and wounds.

Immunohistochemistry was used to examine IL-33 expression adult murine skin and wounds. Representative photomicrographs are shown for unwounded skin (A) and wounds at 24 hours (B), 3 days (C), and 7 days (D) post-injury (scale bars = 100 μm; n = 3 per group). Wound margins are marked with arrows for time points at which wounds were not reepithelialized (B, C).

Positive staining was also observed in dermal cells at early time points, particularly in E18 wounds (Fig. 1 D and F). The cells appeared to be inflammatory cells based on morphology and the typical timing of inflammation in this model. However, we could not rule out the possibility that dermal fibroblasts were expressing IL-33 using standard immunohistochemical staining. Therefore, we generated additional wounds in collagen 1a1-GFP transgenic mice, in which fibroblasts express GFP, and performed immunofluorescence staining for IL-33 and GFP. Double positive cells were not observed at early time points (Fig. 3 A-C). However, double positive cells were detected in E18 wounds at 7 days post-wounding (Fig. 3 D-E), suggesting that fibroblasts express IL-33 at later stages of healing in scar-forming fetal wounds.

Figure 3. Immunofluorescence analysis of IL-33 in fetal wounds.

Immunofluorescence was used to localize IL-33 (red) and GFP (green) in paraformaldehyde-fixed frozen sections of E15 and E18 wounds. Samples were counterstained with DAPI (blue). Representative photomicrographs of E15 (A) and E18 (B-C) wounds at 24 hours post-wounding are shown. The area marked with a * in B is shown at higher magnification in panel C. Wound margins are to the right. Representative photomicrographs are also shown for E15 (D) and E18 (E) wounds at 7 days post-wounding, with the wound bed/scar in the middle of the image. Scale bars = 100 μm in panels A, B, D, and E. Scale bars = 50 μm in panel C. Images are representative of n = 3–5 samples per group.

Expression of ST2, the receptor that binds IL-33, was also examined by immunohistochemistry. In both E15 and E18 skin, positive staining was observed in basal keratinocytes, with intense staining near the dermo-epidermal junction in normal skin adjacent to the wound site (Fig. 4). The staining diminished in keratinocytes near the wound margin (Fig. 4 A-B, D) but reappeared once the wounds were reepithelialized (Fig. 4 C). Positive cells were also observed in the dermis in both E15 and E18 wounds, which could be immune cells or fibroblasts. In addition, large granular cells were positive in E18 wounds, which are likely to be dermal mast cells based on previous studies showing that they are prominent in E18 wounds¹⁷ and that they express ST2.²²

Figure 4. Immunohistochemical staining for ST2 in fetal wounds.

Immunohistochemistry was used to analyze ST2 expression in paraformaldehyde-fixed frozen sections of E15 and E18 wounds. Representative photomicrographs of E15 (left panels; A, C) and E18 (right panels; B, D) wounds at 24 hours (top panels; A-B) and 72 hours (bottom panels; C-D) post-wounding are shown. Wound margins are to the right. Insets show magnified images of a positive cell with morphology consistent with a mast cell (B) and positively stained dermal cells (C-D). Scale bars = 50 μm. Images are representative of n = 4–8 samples per group.

IL-33 promotes scar formation in E15 wounds

To determine whether IL-33 stimulates scar formation in fetal wounds, E15 wounds were injected with recombinant IL-33. Trichrome staining was performed to assess scar formation. PBS-injected E15 control wounds healed scarlessly as expected (Fig. 5 A), but nearly 60% of IL-33-injected wounds healed with discernable scars (Fig. 5 B). All scars in IL-33-injected fetuses were over 100 µm in width, and the average width of the scars was 180.6 +/− 27.8 µm.

Figure 5. Changes in scar formation with IL-33 exposure.

E15 wounds were injected with India ink combined with PBS as a control or 800 ng recombinant murine IL-33. Wounds were harvested at 7 days and the presence of scar tissue was assessed in Masson’s trichrome-stained sections. Representative photomicrographs from wounds injected with PBS (A) and IL-33 (B) are shown. Wound beds/scars are marked with arrows. The number of wounds that healed with scars is also reported for each group. Representative images for E15 (C) and E18 (D) wounds harvested at 3 days post-wounding are also shown for samples from collagen 1a1-GFP transgenic mice. Wound margins are to the right. Samples were stained for GFP (green) and counterstained with DAPI (blue), with n = 3–5 samples examined per group. Scale bars for all panels = 100 μm.

To examine the activity level of the fibroblasts in response to IL-33 injection, wounds were created in collagen 1a1-GFP transgenic mice and GFP was assessed by immunofluorescence as an indicator of collagen expression. Compared to control wounds (Fig. 5 C), more prominent GFP staining was observed in IL-33 injected wounds (Fig. 5 D), suggesting that IL-33 was either directly or indirectly stimulating fibroblast activity in vivo.

IL-33 injection has minimal effects on inflammation in E15 wounds

To identify potential mechanisms underlying IL-33-induced scar formation, various inflammatory cell types were compared in E15 control and IL-33-injected wounds, including neutrophils, mast cells, and macrophages. Neutrophils were absent in the dermis (Fig. 6 A-B; adult wounds were used as a positive control – data not shown) and there was no evidence of significant mast cell degranulation 24 hours post-wounding, even in IL-33-injected wounds (Fig. 6 C-D). Additionally, no statistically significant differences in the density of mast cells (Fig. 6 C-D, G) or macrophages were detected (Fig. 6 E-F, H).

Figure 6. Effects of IL-33 on inflammatory cells.

Ly-6G immunostaining was used to identify neutrophils in E15 wound injected with PBS (A) or IL-33 (B) at 24 hours in paraffin sections (samples lacked positive cells). Toluidine blue staining was used to detect the presence of mast cells in paraffin sections in E15 wounds injected with PBS (C) or IL-33 (D) at 24 hours. F4/80 immunostaining in cryosections was used to identify macrophages in E15 wounds injected with PBS (E) or IL-33 (F) in 48 hour wounds. Cell numbers at the wound margins were counted at multiple time points for mast cells (G) and macrophages (H) and area was determined using image analysis. The bars on the graphs represent average cell density (cell number per mm²) +/− S.E.M. (n = 4–8 per group); differences were not significant by unpaired t-test.

IL-33 stimulates cultured fibroblasts

Because there were no substantial differences in inflammation in IL-33-treated wounds, the potential for IL-33 to directly affect fibroblasts was assessed in vitro. IL-33 signaling through ST2/IL-1RAcP is known to result in NF-κB activation, so NF-κB nuclear translocation was examined in IL-33-treated and control cells. Murine fibroblasts from adult skin treated with IL-33 showed an increase in the presence of the p65 subunit of NF-κB in the nucleus compared to untreated cells (Fig. 7 A), suggesting that IL-33 can directly activate IL-33 receptor signaling in dermal fibroblasts. Since IL-33 was able to induce signaling in fibroblasts, the effect of IL-33 on collagen expression in adult fibroblasts was examined. The fibroblasts were derived from collagen 1a1-GFP transgenic mice, so GFP was used as a measure of collagen expression. IL-33 treatment significantly increased the mean fluorescence intensity of GFP (fluorescence per cell) as measured by flow cytometry compared to control cells (Fig. 7 B), suggesting that IL-33 is capable of stimulating collagen expression by fibroblasts.

Figure 7. Response of cultured fibroblasts to IL-33.

Dermal fibroblasts were cultured from skin explants harvested from collagen 1a1-GFP transgenic mice. To determine the effects of IL-33 on NF-κB activation, adult cells were treated with PBS, 100 ng/ml IL-33 or 20 ng/ml TNF-α as a positive control (A). Cells were collected at 2 hours post-treatment, and nuclear and cytoplasmic proteins were isolated. Western blotting was used to detect the p65 subunit of NF-κB in cytoplasmic (cyt) and nuclear (nuc) fractions. Blots were stripped and re-probed for the cytoplasmic protein α-tubulin to demonstrate effective separation of nuclear and cytoplasmic proteins. The blot shown is representative of three separate samples. To determine the effects of IL-33 on collagen 1a1 gene expression, adult cells were treated with PBS, 100 ng/ml IL-33 or 10 ng/ml TGF-β1 as a positive control (B). Cells were collected at 48 hours post-treatment and green fluorescence was assessed using flow cytometry. E15 fibroblasts were grown on chamber slides (C-D) or lysed (E) for protein analysis of ST2 by immunohistochemistry or Western blotting, respectively. Samples stained with anti-ST2 antibodies (C) and IgG as a negative control (D) are shown. Scale bar = 50 µm. In panel E, two E15 fibroblast protein samples are shown in Western blots probed for ST2 (top) or beta-actin as a loading control (bottom). To confirm the effects of IL-33 on collagen 1a1 gene expression in E15 fibroblasts and determine the importance of NF-κB signaling in this response, cells were treated with PBS, 100 ng/ml IL-33 or 100 ng/ml IL-33 with 30 mM JSH-23 (F). Cells were collected at 48 hours post-treatment and green fluorescence was assessed using flow cytometry. Bars on graphs represent the geometric mean of fluorescence intensity ± S.E.M. (n = 3–6 per treatment; ***p<0.001 or *p<0.05 by one-way ANOVA with Tukey post-hoc testing).

To determine whether fetal fibroblasts respond to IL-33 and whether the induction of collagen expression is dependent on NF-κB, cultured fibroblasts from E15 skin were examined. Immunohistochemical staining and Western blotting for ST2 demonstrated the presence of the binding receptor for IL-33 on cultured E15 fibroblasts (Fig. 7 C-E). In addition, flow analysis showed that treatment with recombinant IL-33 increased the mean fluorescence intensity of GFP (collagen expression) and that the addition of the NF-κB inhibitor JSH-23 significantly reduced GFP fluorescence in cells stimulated with IL-33 (Fig. 7 F), suggesting that NF-κB signaling plays a role in the stimulation of collagen expression by IL-33.

Discussion

Inflammation is a known regulator of scar formation in fetal wounds and alarmins are becoming increasingly recognized as mediators of wound inflammation. Previous work by our lab has shown that the alarmin HMGB-1 is released more extensively in scar-forming wounds and that HMGB-1 promotes scar formation in fetal wounds in vivo.⁷ In the current study, IL-33, which has also been described as an alarmin,^{23, 24} was examined. Recently, several studies have demonstrated a role for IL-33 in the healing of adult wounds,^25–28 but detailed analyses of scar formation were not performed. Here, we found that IL-33 is more highly expressed in scar-forming wounds and that the addition of IL-33 to E15 fetal wounds promotes scar formation.

Previous studies examining adult skin have shown that total levels of IL-33 are elevated after wounding.^{27, 28} Based on immunohistochemical analysis, prominent upregulation of IL-33 in keratinocytes was observed at the wound edge in fetal wounds, which was exaggerated in scar-forming E18 wounds compared to scarless E15 wounds. Strong keratinocyte staining was also present in scar-forming adult wounds, suggesting that high IL-33 expression is associated with scar formation. Previous studies have suggested that IL-33 acts as an alarmin in adult skin wounds.²³ Interestingly, the staining patterns observed here for IL-33 differed from what has been described for traditional alarmins such as HMGB-1. In scar-forming fetal and adult wounds, as well as other injury types in the skin and damaged tissue in other organs, HMGB-1 is present in the nucleus of epithelial cells in normal tissue and injury stimulates cellular release of HMGB-1 as indicated by a loss of nuclear staining.^{7, 29–32} In contrast, the present studies showed very little IL-33 in unwounded E15 or E18 fetal skin, but there was upregulation of IL-33 in keratinocytes at the wound edge after injury. In addition, nuclear localization of IL-33 remained strong in keratinocytes after injury. It is possible that IL-33 does not act as a traditional alarmin in wound keratinocytes and/or is not released to the extent that a loss of staining can be detected. It is also possible that IL-33 may be acting as more of a traditional cytokine in wounds and is being released gradually by keratinocytes, which would not necessarily be detectable as a reduction in staining by immunohistochemistry. More work will need to be done to understand whether IL-33 is being released by wound keratinocytes in vivo and if so, how it is being released. Regardless, the levels of IL-33 were clearly elevated in scar-forming wounds compared to scarless wounds.

Based on the high levels of IL-33 expression observed in scar-forming wounds, functional studies were performed to assess the role for IL-33 in scar formation. Injection of recombinant IL-33 into E15 wounds, which normally heal scarlessly, caused the wounds to heal with scars. These results are in line with studies demonstrating a role for IL-33 in cutaneous fibrosis.³³ While thorough studies on IL-33 and scar formation in cutaneous wounds have not been performed to our knowledge, previously published studies in IL-33 knockout mice have reported no significant changes in granulation tissue area in adult wounds.²⁷ There are several possible explanations for the discrepancies between adult wound studies, in which strong evidence for the involvement of IL-33 in scar formation is lacking, and the fetal studies described here which indicate a role for IL-33 in scar formation. First, it is possible that a comparison of excisional wounds in adult wild-type and IL-33 knockout mice would have revealed differences in scar formation if wounds were analyzed at later time points. Second, the approach of injecting IL-33 into scarless fetal wounds differs from adult wound healing studies utilizing IL-33 knockout mice. Adding exogenous IL-33, as done in the present work, will provide information about the extracellular function of IL-33. In contrast, both nuclear and extracellular activities of IL-33 are abrogated in IL-33 knockout mice. This could be an important distinction, as nuclear and extracellular IL-33 reportedly have opposing actions that could impact aspects of healing.^{26, 34} Finally, it is possible that the fetal wound environment, which contains less pro-inflammatory and pro-fibrotic mediators, is more sensitive to alterations in IL-33 levels than adult wounds, which contain a larger array of mediators with potentially redundant functions. More studies will have to be done to completely understand the role of IL-33 in scar formation in mature skin, including determining whether IL-33 knockout mice, neutralization of IL-33, or inhibition of IL-33 signaling can reduce scarring in adult skin wounds and investigating the mechanisms by which IL-33 may promote scarring in these wounds.

Given the function of IL-33 as a pro-inflammatory cytokine and the importance of inflammation in the transition from scarless to fibrotic healing in fetal wounds, we examined inflammation as a potential mechanism to explain the ability of IL-33 to induce scarring in E15 wounds. While it has been suggested that IL-33 increases inflammation in other skin models,^{35, 36} we did not find statistically significant changes in inflammation in wounds injected with IL-33, at least at the time points examined in this study. Some inflammatory cell types display a less mature phenotype in fetal skin, so it is possible that the inflammatory cells were not responding to IL-33 in E15 wounds to the same extent that adult cells would. It is also possible that IL-33 was affecting the inflammatory cells in some way (i.e., inducing the production of cytokines or growth factors that could be stimulating fibroblast activity), but not to a degree to which statistically significant changes in overall inflammatory cell density could be detected. Interestingly, IL-33 has been shown to stimulate expression of the antimicrobial protein REG3A in keratinocytes, which reduces TLR3-induced inflammation in skin wounds through induction of the phosphatase SHP-1.³⁷ It is possible that the lack of significant changes in inflammation observed after injection of IL-33 into E15 wounds is due in part to IL-33 increasing the levels of REG3A, which is helping to limit inflammation. However, further work will have to be done to determine whether this pathway is regulating inflammation in IL-33-injected E15 wounds.

Because dermal fibroblasts have been reported to have the ST2 receptor,^{38, 39} another possible explanation for the increase in scarring in IL-33-injected wounds is that IL-33 was directly stimulating fibroblasts. IL-33-ST2 binding and receptor signaling events mediated by IL-1RAcP are known to result in NF-κB activation. Experiments using cultured dermal fibroblasts showed that IL-33 treatment induced NF-κB activation. IL-33 treatment caused an increase in collagen 1a1 expression, and this was significantly reduced in the presence of an NF-κB inhibitor. In addition, the binding receptor, ST2, was detected in cultured fibroblasts by immunohistochemistry and Western blot. Taken together, these results suggest that IL-33 is capable of activating downstream signaling and enhancing collagen expression in dermal fibroblasts. Additional studies will be required to determine whether IL-33 is able to directly promote collagen production and scar formation by fibroblasts in vivo.

In summary, these studies suggest that higher levels of IL-33 are associated with scarring in murine skin wounds and that IL-33 promotes scar formation in early-gestation fetal wounds. Further studies are required to determine whether IL-33 regulates scar formation in adult skin wounds and to assess its involvement in human scarring.