Keratinocyte-targeted overexpression of the glucocorticoid receptor delays cutaneous wound healing

Abstract

Delayed wound healing is one of the most common secondary adverse effects associated to the therapeutic use of glucocorticoid (GC) analogs, which act through the ligand-dependent transcription factor GC-receptor (GR). GR function is exerted through DNA-binding-dependent and -independent mechanisms, classically referred to as transactivation (TA) and transrepression (TR). Currently both TA and TR are thought to contribute to the therapeutical effects mediated by GR; however their relative contribution to unwanted side effects such as delayed wound healing is unknown. We evaluated skin wound healing in transgenic mice with keratinocyte-restricted expression of either wild type GR or a mutant GR that is TA-defective but efficient in TR (K5-GR and K5-GR-TR mice, respectively). Our data show that at days (d) 4 and 8 following wounding, healing in K5-GR mice was delayed relative to WT, with reduced recruitment of granulocytes and macrophages and diminished TNF-α and IL-1β expression. TGF-β1 and Kgf expression was repressed in K5-GR skin whereas TGF-β3 was up-regulated. The re-epithelialization rate was reduced in K5-GR relative to WT, as was formation of granulation tissue. In contrast, K5-GR-TR mice showed delays in healing at d4 but re-established the skin breach at d8 concomitant with decreased repression of pro-inflammatory cytokines and growth factors relative to K5-GR mice. Keratinocytes from both transgenic mice closed in vitro wounds slower relative to WT, consistent with the in vivo defects in cell migration. Overall, the delay in the early stages of wound healing in both transgenic models is similar to that elicited by systemic treatment with dexamethasone. Wound responses in the transgenic keratinocytes correlated with reduced ERK activity both in vivo and in vitro. We conclude that the TR function of GR is sufficient for negatively regulating early stages of wound closure, while TA by GR is required for delaying later stages of healing.

Figure 1. In vivo overexpression of GR and GR-TR in keratinocytes delays wound healing with different kinetics of repair.

A) Similar transgene expression in K5-GR and K5-GR-TR mice. Relative protein levels of the transgene vs endogenous GR were determined by immunoblotting using an antibody that recognizes both mouse (endogenous) and rat (transgene) GR. Actin was used as a loading control. A representative panel of three independent experiments and quantitation of three independent experiments is shown. Statistical significance was assessed by ANOVA (**p<0.01). B) Macroscopic changes in full-thickness cutaneous wounds. The wound sites were photographed at the time indicated after wounding. Representative results from 12 mice of each genotype are shown. C) The percentage of wound closure area at each time point was estimated relative to the original wound area. Values represent mean ± SD (n = 12 for all groups). Statistical significance was determined using ANOVA (*p<0.05, **p<0.01). D) Histological evaluation of skin wounds in WT, K5-GR and K5-GR-TR mice at d4 and 8 after injury. Arrowheads and arrows indicate original wound edges and re-epithelialized edges, respectively. Trichrome staining shows distinct collagen deposition in the dermis at d8 in the indicated genotypes. Immunostaining using an anti-GR specific antibody shows GR expression and localization in WT vs transgenic mice (IH: GR). Note the nuclear localization of GR in both transgenic skin samples. Hematoxylin/eosin and Trichrome staining, Bars: 100 µm; IH: GR, Bar: 50 µm. E) The percentage of re-epithelialization was evaluated in WT, K5-GR and K5-GR-TR mice at d4 and d8 after wounding. All values represent the mean ±SD (n = 12 for all groups). Statistical significance was estimated using ANOVA (*p<0.05, **p<0.01).

Figure 2. Reduced migration and increased differentiation of keratinocytes in K5-GR wounds relative to K5-GR-TR and WT mice.

Skin wounds were collected from WT, K5-GR or K5-GR-TR mice at d4 and d8 to evaluate the expression of K6 and K10 by immunostaining using specific antibodies. Representative results from 5 mice of each genotype are shown. The arrow indicates the migrating tip of the WT wound and the asterisks indicate the impaired keratinocyte migration in the transgenics. Boxed areas are enlarged. Higher magnification of boxed areas is at the lower right of each image. Bar: 100 µm.

Figure 3. Delayed wound healing response in K5-GR mice is equivalent to that of Dex-treated WT mice.

Histological analysis of re-epithelialization of skin wounds in WT mice untreated or treated with Dex (1 mg/kg/day) for 7 d before wounding, at d9 after injury. Arrowheads indicate original wound edges. Immunostaining using anti-K6 and anti-K10 specific antibodies shows the delayed healing in Dex-treated vs untreated WT mice (n = 6). Representative images of three independent experiments are shown. Bar: 100 µm.

Figure 4. K5-GR and K5-GR-TR mice exhibit reduced proliferation and inflammatory response following injury.

A) Keratinocyte proliferation in WT and transgenic wounds. The proliferation rate was quantified by counting BrdU-labeled nuclei in the basal layer of interfollicular epidermal keratinocytes at d4 and d8. Positive nuclei were counted from the BrdU-labeled keratinocyte most proximal to the wound margin to the area at which the BrdU labeling index fell below 5%. The graph illustrates the percentage of positive BrdU keratinocytes vs total nuclei. Mean values ± SD are shown (n = 36). Asterisks denote statistically significant differences among genotypes within a given timepoint, as determined by ANOVA (*, p<0.05; **, p<0.01). B) Cutaneous inflammation in WT and transgenic wounds. Immunostaining of WT, K5-GR and K5-GR-TR wounds at d4 and d8 showing granulocytes (Ly6G, arrows), macrophages (F4/80) (inset), and TNF-α (inset). Representative images of three independent experiments are shown. Bar: 50 µm. C) Altered gene expression in transgenic skin wounds. Transcript levels of Tnf- α, Il-1 β, Tgf-β1, Tgf-β3, and Kgf were determined in the wounds of the indicated genotypes at d4 and d8 after injury by quantitative RT-PCR. Experiments were performed using at least three individuals of each genotype per timepoint. The mRNA levels of WT mice before wounding for each gene were set to 100, and fold-changes related to this basal level. Mean values ± SD are shown. Asterisks denote statistically significant differences among genotypes for each gene at a given timepoint, as determined by ANOVA (*, p<0.05; **, p<0.01).

Figure 5. K5-GR and K5-GR-TR wounds show reduced ERK activity.

A) Changes in protein expression after wounding. Total skin protein was extracted from d4 and d8 wounds of WT, K5-GR and K5-GR-TR mice and subjected to immunoblotting using the indicated specific antibodies (p-ERK, ERK, c-Fos, and cyclin D1). Actin was used as a loading control. Panels are representative of the relative expression levels of the indicated proteins for each condition. B) Quantitation of the experiments performed using at least three individuals of each genotype per timepoint. Protein levels were normalized to actin. The graph shows the p-ERK/ERK ratio as well as c-Fos and cyclin D1 protein levels in K5-GR, K5-GR-TR and WT wounds. The protein levels of WT mice before wounding were set to 100, and fold-induction related to this basal level. Mean values ±SD are shown. Asterisks denote statistically significant differences at each timepoint among all genotypes, as determined by ANOVA (*, p<0.05; **, p<0.01). C) Immunostaining of sections of WT, K5-GR and K5-GR-TR wounds at d4 and d8 showing the expression pattern of β-catenin. Arrows denote the nuclear localization of β-catenin in transgenic keratinocytes. Representative images of each condition are shown. Bar: 50 µm.

Figure 6. GR and GR-TR delay keratinocyte migration in in vitro wound scratch assays.

A) In vitro migration of K5-GR and K5-GR-TR primary keratinocytes isolated from WT, K5-GR and K5-GR-TR were evaluated for in vitro wound closure at 8 h and 24 h following scratching. Phase contrast photographs are representative of four independent experiments. Bar: 100 µm. B) The graph shows the average of wound closure ± SD. Statistically significance was determined by ANOVA (*, p<0.05; **, p<0.01). The percentage of cell migration was estimated as described in Materials and Methods. C) Quantitation of the p-ERK/ERK and p-AKT/AKT ratios in WT, K5-GR and K5-GR-TR keratinocytes before (t0) and after (t1) scratch wounding. Protein levels were determined by immunoblotting using specific antibodies against the total and phosphorylated proteins, and then normalized to actin. Mean values ± SD are shown. Asterisks denote statistically significant differences, as determined by ANOVA (*, p<0.05; **, p<0.005).