Genetic- or transforming growth factor-beta 1-induced changes in epidermal peroxisome proliferator-activated receptor beta/delta expression dictate wound repair kinetics.

Advances in wound care are of great importance in clinical injury management. In this respect, the nuclear receptor peroxisome proliferator-activated receptor (PPAR)beta/delta occupies a unique position at the intersection of diverse inflammatory or anti-inflammatory signals that influence wound repair. This study shows how changes in PPARbeta/delta expression have a profound effect on wound healing. Using two different in vivo models based on topical application of recombinant transforming growth factor (TGF)-beta1 and ablation of the Smad3 gene, we show that prolonged expression and activity of PPARbeta/delta accelerate wound closure. The results reveal a dual role of TGF-beta1 as a chemoattractant of inflammatory cells and repressor of inflammation-induced PPARbeta/delta expression. Also, they provide insight into the so far reported paradoxical effects of the application of exogenous TGF-beta1 at wound sites.

Of the numerous cytokines produced at the wound site, TGF-␤1 1 has the broadest effects, influencing nearly every aspect of tissue repair (1). TGF-␤1 signals through a membrane receptor complex that recruits and phosphorylates Smad2 and 3 proteins. Once phosphorylated, they associate with Smad4 and translocate into the nucleus to control gene transcription in several ways (2,3). Smad3-null mice, whose TGF-␤1 signaling pathway is impaired, show accelerated wound healing characterized by enhanced reepithelialization, increased keratinocyte proliferation, and reduced immune cell infiltration (4). Interestingly, TGF-␤1-null mice display either delayed or accelerated wound healing depending on the model analyzed (5,6). Equally complex is the effect of exogenous TGF-␤1 on a wound because the healing rate was found to be dependent on the delivery vehicle, dose of the growth factor, its application schedule, and wound model (7)(8)(9). In a rabbit ear dermal ulcer model, a single application of TGF-␤1 at the time of wounding had a beneficial effect equal to that of multiple doses, whereas application 24 h after wounding did not improve healing (10).
Doses of TGF-␤1 used to treat wounds topically have varied over a 1,000-fold concentration range from a few ng to several g, depending on delivery vehicle and wound models (8,9,11,12). All studies have shown a bell-shaped dose-dependent response to TGF-␤1, with either no or decreasing effects at lower or higher than optimal concentrations. The reason for such diverse responses to exogenous TGF-␤1 is part of the present investigation.
The nuclear receptor PPAR␤/␦ plays critical roles in the keratinocyte response to inflammation signals produced immediately after a skin injury. Its inflammation-induced increase in activity at the wound edge maintains a sufficient number of viable keratinocytes for reepithelialization (13)(14)(15). Using exogenous TGF-␤1 application as well as Smad3 and PPAR␤/␦ single-and double-knock-out mice, we provide evidence for an in vivo cross-talk between TGF-␤1 and PPAR␤/␦ signaling.
Wounding and healing analyses were performed on 6-week-old females as described previously (6,14). Briefly, the hair follicle cycle of each mouse was synchronized by shaving the back of the animal 2 weeks before starting the experiment. Mice were then anesthetized, shaved, and a full thickness middorsal wound (0.5 cm 2 , square-shaped) was created by excising the skin and the underlying panniculus carnosus. Wound closure was measured daily in a double-blinded fashion until complete wound closure. A subset of wild type animals was treated topically with either vehicle (3% methylcellulose in phosphate-buffered saline and 4 mM HCl) or 100 ng of TGF-␤1. Treatments were rotated to avoid site bias. At the indicated days postwounding, the entire wound including a 5-mm margin was excised. Wounds were dissected for immunohistochemistry, RNA, and protein analyses, as described previously (16). Immunohistochemistry was performed on 8-m cryosections of wound biopsies with the following antibody dilutions: K6, involucrin, 1:1,000; filaggrin, Akt1-phosphoserine 473, 1:500; ␣-SMAfluorescein isothiocyanate, 1:400; Ki67, Akt1-phosphothreonine 308, 1:200; PPAR␤/␦, Mac3, 1:100. The signal was detected either by immunofluorescence or amplified using the ABC-peroxidase method (Vector Laboratories) and revealed using 3,3Ј-diaminobenzidine with (dark blue coloration) or without (brown coloration) metal enhancer. Apoptotic cells were detected by TUNEL assay (Roche Applied Science).
RNase Protection Assay-Direct RNase protection assay was performed as described previously (15), for PPAR␤/␦, L27, ILK, and/or PDK1, using RNA isolated from wound biopsies of various mutant mice excised at the indicated postinjury time points. The measured expression levels were normalized against the mRNA levels of the ribosomal protein L27. The normalized value of unwounded (day 0) skin biopsies was assigned a value of 1.
TGF-␤1(Day 0) Treatment-The TGF-␤1(day 0) treatment accelerated wound repair with a complete closure at days 13-14 compared with the vehicle-treated wounds, which closed at day 17 (Figs. 1A and S1A). Wound biopsy analyses showed that the expression of PPAR␤/␦ was already elevated 24 h after application, an accelerated response compared with vehicletreated wounds (Figs. 1B and S2A). Although PPAR␤/␦ levels did not reach the peak observed in vehicle-treated wounds, they remained elevated for up to 5 days after injury. As expected, the expression patterns of two PPAR␤/␦ target genes, ILK and PDK1, and hence Akt1 activity, were changed according to the altered PPAR␤/␦ expression profile (Figs. 1B and S2A) (13).
Wound contraction also contributes to wound closure and takes place when the granulation tissue is populated with myofibroblasts. Because TGF-␤1 is a major cytokine of myofibroblast differentiation (17), we next examined the presence of these cells in wound biopsies using ␣-SMA as a specific marker. Consistent with other reports, the expression of ␣-SMA peaked between days 7 and 10 after injury (Figs. 1B and S2A) (18,19). In TGF-␤1(day 0)-treated wounds, a higher expression of ␣-SMA was already detected at day 5. Subse- The expression of ␣-SMA and PPAR␤/␦ is triggered by TGF-␤1 and inflammatory cytokines, respectively. Thus, their expression profiles are good indicators of active TGF-␤1 and the inflammatory response at wound site. TGF-␤1-mediated chemotaxis follows a bell-shaped dose dependence. In vehicle-treated wounds, the optimal dose occurs between days 3 and 4, corresponding to high PPAR␤/␦ and low ␣-SMA expression (open box). In TGF-␤1(day 0) treatment, this dose was shifted toward the left (gray arrow), i.e. between days 1 and 2 (yellow box) and was accompanied by an early infiltration of macrophages. A consequential increased production of inflammatory cytokines overcomes TGF-␤1 inhibition of PPAR␤/␦ expression, leading to an early and prolonged expression (red line).
We propose that these early recruited macrophages produce enough inflammatory cytokines to overcome the inhibitory action of TGF-␤1 on PPAR␤/␦ (16) and to up-regulate its expression. Indeed, the chemotactic dose response to TGF-␤1 follows a bell-shaped curve (2). Consistent with other reports showing that the infiltration of macrophages begins 48 h postinjury (22,23), the optimal response occurs around days 2-4 in the vehicle-treated mice, when the expression of PPAR␤/␦ is high and that of ␣-SMA low (Fig. 1D). In TGF-␤1(day 0)-treated wound, the infiltration of macrophages occurs earlier, between days 1 and 2 (Fig. 1D), producing a sufficient amount of cytokines to overcome the inhibitory action of TGF-␤1 on PPAR␤/␦. Accordingly, a different outcome with a later application of the TGF-␤1 treatment is expected.
TGF-␤1(Day 2) Treatment-A later exposure to TGF-␤1, 48 h after injury (TGF-␤1(day 2) treatment), resulted in a significant, but transient delay in reepithelialization, with a wound closure time similar to that of the vehicle-treated wound (Figs. 2A and S1B). Interestingly, the peak of increased PPAR␤/␦ expression typically observed in vehicle-treated wound biopsies was strongly repressed by the TGF-␤1(day 2) treatment (Figs. 2B and S2B). The diminished PPAR␤/␦ expression also resulted in reduced ILK and PDK expression as well as reduced Akt1 activity (Figs. 2B and S2B). In contrast, an early increase in ␣-SMA expression was detected already at day 3 and peaked between days 5 and 10 (Figs. 2B, S2B, and S3).
To gain further insights into the dual effect of exogenous TGF-␤1, similar experiments were performed on Smad3-null mice. Regardless of the application schedule, TGF-␤1(day 0) or TGF-␤1(day 2), no significant increase in the number of Mac3positive macrophages was detected in 24-h post-treatment biopsies compared with vehicle-treated wounds (Fig. S5). This suggests that Smad3 is critical for TGF-␤1-mediated chemotaxis, which is consistent with earlier reports showing that topical applications of TGF-␤1 immediately before wounding did not influence inflammatory cell recruitment in Smad3deficient mice (4,24).
Taken together, these results underscore a finely tuned temporal balance between inflammatory signals and TGF-␤1 in the regulation of PPAR␤/␦ expression and hence of wound repair. They also suggest that prolonged PPAR␤/␦ expression promotes rapid wound closure.
To evaluate the effect of the prolonged PPAR␤/␦ expression on reepithelialization, we performed immunohistochemistry on wound biopsies at day 5, when the expression levels of PPAR␤/␦, PDK1, ILK, and the activity of Akt1 were dramatically higher in Smad3 Ϫ/Ϫ wounds. Compared with WT mice, Smad3 Ϫ/Ϫ mice exhibited a thicker hyperproliferative and more extended epithelial tongue at the wound site, as revealed by K6 staining (Fig. 4A). Consistent with the sustained PPAR␤/␦ mRNA and Akt1 activity levels, a higher number of PPAR␤/␦-positive keratinocytes was detected in Smad3 Ϫ/Ϫ biopsies (WT versus Smad3 Ϫ/Ϫ , 32.4 Ϯ 3.1 versus 68.4 Ϯ 2.1/ section; Fig. S4), and more phosphorylated Akt1 was detected at the leading wound edges and in the basal wound keratinocytes (Fig. 4B).
Together, these results provide evidence for prolonged epidermal PPAR␤/␦ expression in Smad3 Ϫ/Ϫ mice, which produces a prolonged Akt1 activity, higher keratinocyte proliferation, and lower apoptosis rates in the regenerating epithelium. Most importantly, these events are associated with much faster healing in Smad3 Ϫ/Ϫ mice.
These results reveal that in vivo PPAR␤/␦ and Smad3 coordinate the balance between keratinocyte apoptosis and proliferation. Importantly again, a prolonged elevated PPAR␤/␦ expression profile favors a rapid wound closure.

DISCUSSION
Efficient wound repair after an injury is crucial to the survival of any organism. The results herein demonstrate that, either via external conditioning of the wound site by topical application of TGF-␤1 or genetic ablation of the Smad3 gene, a prolonged elevated expression and activity of PPAR␤/␦ dictates an accelerated wound closure. This prolonged PPAR␤/␦ activity allows for extended Akt1 activity that reduces apoptosis and increases proliferation of keratinocytes at the wound site, and hence promotes faster healing.
The differential effects of exogenous TGF-␤1 on wound closure underscore the complexity of the wound repair mechanism. The TGF-␤1(day 0) treatment accelerates, whereas the TGF-␤1(day 2) application transiently delays wound healing. The balance between promotion and slowdown of healing depends on doses and the application schedule of TGF-␤1. Analyses of these parameters suggest that the described paradoxical actions of exogenous TGF-␤1 on wound healing most likely reside in its dual role. As a chemoattractant of immune cells, it stimulates PPAR␤/␦ expression via the production of inflammatory cytokines (28). As a repressor, it reduces the inflammation-induced PPAR␤/␦ expression via Smad3 signaling. Early TGF-␤1-mediated recruitment of macrophages, achieved via a timely appropriate level of TGF-␤1 as illustrated by the TGF-␤1(day 0) treatment, can overcome the TGF-␤1-repressive action on inflammation-induced PPAR␤/␦ and promotes rapid wound closure. Conversely, delayed TGF-␤1 application, as in the TGF-␤1(day 2) treatment, most likely resulting in a higher TGF-␤1 level, fails to recruit increased amounts of macrophages and to overcome its repression effect on PPAR␤/␦ expression. In this model, an early very high dose of TGF-␤1 which exceeds the concentration range for TGF-␤1-mediated chemotaxis should also have a negative effect on wound healing. In agreement with this proposition, Beck et al. (10) showed positive effects of TGF-␤1 on wound healing, at doses of 5-100 ng, which was lost at higher doses. Further investigation of the dual role of TGF-␤1 would benefit from the availability of conditional epidermal Smad3-deficient mice.
At first sight, it may appear difficult to reconcile the observations between those obtained from Smad3 Ϫ/Ϫ mice and TGF-␤1(day 0)-treated animals. Notably, the TGF-␤1(day 0) treat-ment influenced the PPAR␤/␦ expression profile such that it was almost identical to that of Smad3 mutant mice, which explains the similarities of the healing kinetics. Despite a similar outcome, the means by which it is reached in the two different experimental models are different. In the TGF-␤1(day 0) protocol, the chemoattractant effect of exogenous TGF-␤1 triggers an early infiltration of immune cells, which produce inflammatory cytokines that up-regulate PPAR␤/␦ expression (15). In the Smad3 Ϫ/Ϫ mice, despite an impaired local inflammatory response (4), the enhanced and sustained expression of PPAR␤/␦ results from a loss of the suppressive effect of Smad3 on the expression of PPAR␤/␦ (17).
In the TGF-␤1(day 2) protocol, despite the strong suppression of inflammation-induced PPAR␤/␦ expression and a transient healing delay, a similar time point of complete wound closure was observed compared with vehicle-treated mice. This apparent discrepancy with our earlier reports describing a delayed wound closure in PPAR␤/␦-deficient mice (14, 28) is most likely for two reasons. First, the TGF-␤1(day 2) treatment resulted in only a local suppression of PPAR␤/␦ expression compared with the absence of the receptor in the entire PPAR␤/ ␦-deficient animal. Second, an enhanced wound contraction can also contribute to wound closure. Indeed, the TGF-␤1(day 2) protocol produced an earlier and robust up-regulation of ␣-SMA, a marker of myofibroblasts, than vehicle and, surprisingly, TGF-␤1(day 0)-treated biopsies. The different effects seen in the two TGF-␤1 treatments suggest the involvement of additional cytokines at different time points. Indeed, it was Values were subjected to the Mann-Whitney test compared with WT (*, p Ͻ 0.05 and **, p Ͻ 0.01).
shown recently that ␣-SMA expression is antagonistically regulated by TGF-␤1 and interleukin-1. The latter is present abundantly during the early inflammation phase of wound repair (29). Of interest is the inverse expression pattern between PPAR␤/␦ and Akt1 activity on the one hand and ␣-SMA on the other hand. This is consistent with earlier studies (3,16,30,31), which showed that elevated phosphorylated Akt may inhibit the transcriptional activity of Smad3, which is crucial to the up-regulation of ␣-SMA expression (17,32).
For PPAR␤/␦ ligands to be considered as potential wound repair drugs, several factors including application schedule, bioavailability, and dose have to be taken into account. The narrow peak of increase in PPAR␤/␦ expression after injury corresponds to a small time window for effective ligand activation of the receptor. A timely biochemically controlled attenu-ation of TGF-␤1/Smad3 signaling, which leads to a prolonged PPAR␤/␦ expression and also reduces fibrotic tissue (4,33), coupled with an increase in PPAR␤/␦ activity by an agonist, should result in an accelerated rate of wound repair with minimal scarring.
Our results suggest that monitoring changes in the PPAR␤/␦ expression profile may serve as a very useful indicator of the differential effects of exogenous TGF-␤1 or alternative treatments on wound closure. Such insights into how incoming extracellular signals regulate PPAR␤/␦, which in turn coordinates their action on wound healing, may aid in the development of better treatments for ulcers and chronic wound disorders.