Critical roles of PPARβ/δ in keratinocyte response to inflammation

Reactivation of PPARβ in keratinocytes is mediated by an AP-1 site in the PPARβ promoter.

Treatment of cultured primary keratinocytes with TNF-α resulted in an expression profile of PPARβ, which was closest to that observed after injury in an in vivo model (Michalik et al. 2001). Therefore, the molecular events by which TNF-α regulates the promoter of the PPARβ gene were studied next. The mouse PPARβ promoter was isolated (accession no. {"type":"entrez-nucleotide","attrs":{"text":"AF187850","term_id":"6166440"}}AF187850) and a 1.9-kb fragment containing the transcription initiation site was subcloned into a CAT reporter gene. The effect of TNF-α on the activity of a series of truncations of this 1.9-kb PPARβ promoter region were analyzed by use of transfection into primary keratinocytes. The effects of ceramide (Cer), a downstream effector of TNF-α (Hannun and Luberto 2000), and TPA were also studied. The addition of each of the three agents increased the promoter activity of pPPARβ(−1880), as well as that of the shorter promoter constructs pPPARβ(−864), pPPARβ(−587), and pPPARβ(−445), by two to threefold (Fig. ​(Fig.2A).2A). In contrast, pPPARβ(−277), pPPARβ(−223), and pPPARβ(−11) were not responsive to TNF-α, TPA, or Cer, suggesting that the response elements are located between −445 and −277.

Open in a separate window

Figure 2

Identification of the region in the PPARβ promoter that mediates the TNF-α, TPA, and ceramide responses. (A) PPARβ promoter 5′-deletion mutants. Plasmids are named according to the length of the promoter region they contain upstream from the transcription start site. Keratinocytes in culture were cotransfected with a CAT reporter driven by the various PPARβ promoter constructs and pCMV-β-galactosidase as a control of transfection efficiency. Keratinocyte cultures were treated for 24 h with the inducers indicated. Results were normalized to β-galactosidase activity and are presented as relative CAT activity compared with that of the promoterless vector pBLCAT alone. (B) TNFα-, TPA-, and Cer-signaling target an AP-1(−414) of the PPARβ promoter. The hatched bar depicts site-directed mutations of the transcription factor binding sites. The mutations introduced into the site are indicated as underlined nucleotides above the respective hatched bar. The numbers associated with the sites indicate their position relative to the transcription start site. Data are means of six independent experiments. Vehicles for TPA, TNF-α, and Cer were ethanol, PBS, and DMSO, respectively.

This proximal part of the PPARβ promoter contains sequences related to AP-1, Ets, and C/EBP transcription factor binding sites (Fig. ​(Fig.2A),2A), as identified by using MatInspector (Wingender et al. 2000). Mutations were introduced in each of these sites and their relative importance was assessed via transactivation studies (Fig. ​(Fig.2B).2B). Whereas mutation of the Ets or C/EBP sites did not affect regulation by the three agents, mutation of the AP-1 site, located at −414, completely abolished responsiveness to all three inducers. Hence, the signaling pathways that are activated by TPA, TNF-α, or Cer terminate at the AP-1(−414) site in the PPARβ promoter.

TNF-α-induced expression of PPARβ is mediated by the factor-associated-with-neutral-sphingomyelinase and the stress-associated kinase cascade

Next, experiments were designed to identify the individual steps relaying the signal initiated from TNF-α to the AP-1(−414) site in the PPARβ promoter. TNFα-induced trimerization of its receptor (p55TNF-R1) leads to the stimulation of several signaling cascades. One of these pathways involves the production of bioactive Cer, which, in turn, acts as second messengers to initiate downstream effects (Kronke 1999). Cer is of particular interest in this study as it is a component of the epidermal stratum corneum. Adam-Klages et al. (1996) reported that FAN (factor associated to neutral sphingomyelinase) couples the p55TNF-R1 to neutral sphingomyelinase (N-Smase), which is the major enzyme responsible for Cer generation via sphingomyelin hydrolysis (Fig ​(Fig3A).3A). Thus, the involvement of FAN in transducing the effect of TNF-α to the PPARβ promoter was explored. Keratinocytes were transfected with wild-type (wt) or a dominant-negative (dn) FAN and treated with TNF-α. Overexpression of wtFAN potentiated the ability of TNF-α to induce PPARβ expression (Fig. ​(Fig.3B,3B, second panel) and triggered an early onset of keratinocyte differentiation as indicated by the induction of INV expression. In contrast, overexpression of dnFAN abolished the effect of TNF-α on PPARβ expression and resulted in delayed cell differentiation. The repressive effect of dnFAN could be rescued by addition of exogenous Cer (Fig. ​(Fig.3B,3B, third and fourth panels), in accordance with the role of Cer as a downstream effector of FAN. Treatment of untransfected keratinocytes with Cer alone resulted in an elevated PPARβ expression and an early expression of INV.

Open in a separate window

Figure 3

Signaling cascades that mediate the effects of TNF-α, TPA, and Cer on the PPARβ gene in keratinocytes. (A) (Overall scheme) The solid arrows represent the major TNFα-activated, FAN-mediated, pathway regulating PPARβ expression. The cascade invoked by TPA is depicted by the dashed arrows. (B) TNF-α induces PPARβ gene expression through FAN and Cer. Keratinocytes were cotransfected with expression vectors for wild-type FAN (wtFAN, second panel) or a dominant-negative form of FAN (dnFAN, third and fourth panels) and pCMV-β-galactosidase as internal control. Transfected keratinocytes were treated with TNF-α and RPA were performed at the indicated times. (fourth panel) Suppression of PPARβ expression by dnFAN was rescued by addition of exogenous ceramide at 12 h post-transfection. (first panel) Control experiment with TNF-α alone; (fifth panel) effect of ceramide in absence of TNF-α treatment. (C) Regulation of PPARβ via the stress-associated kinase pathway. Keratinocytes were transfected with a CAT reporter driven by the pPPARβ(−445) promoter region together with expression vectors encoding dominant negative (dn) or constitutively active (ca) kinases. The empty vector pCDNA 3.1 was used as a control (vector). Transfected keratinocytes were treated with an inducer (TNF-α, ceramide, or TPA) and CAT activity was measured 24 h post-treatment. SB203580 (10 μM) is a specific inhibitor of p38 and was added 1 h prior to treatment with the inducers. Values are means of six independent experiments.

Transmission of the TNF-α signal downstream of Cer involves the mitogen-activated protein (MAP) kinase pathways. These pathways lead to the phosphorylation of transcription factors, such as those belonging to the AP-1 family, and consequently to a modulation of gene expression (Kyriakis and Avruch 1996). TNF-α and Cer are known to activate MEKK1, which is a key player of the MAPK cascade. MEKK1, in turn, can phosphorylate several downstream kinases including SEK1 (Davis 2000; Fig. ​Fig.3A).3A). To evaluate the role of these kinases in regulating PPARβ expression, the pPPARβ(−445) construct, which harbors the AP-1 site that is responsive to TNF-α, was cotransfected with expression vectors encoding either constitutively active (ca) or dominant-negative forms of several kinases. Alternatively, specific known kinase inhibitors were used. Transfection with dnMEKK1 abolished TNFα-, TPA-, and Cer-dependent promoter activity (Fig. ​(Fig.3C).3C). In contrast, caMEKK1 increased the basal activity by two to threefold and a further twofold upon addition of the cytokines. The cytokine-mediated induction of the PPARβ promoter is also suppressed by dnSEK1.

Downstream of SEK1, both JNK/SAPK and p38 have been shown to modulate AP-1 expression and activity (Davis 2000). The dnJNK/SAPK completely abolished the responsiveness of the PPARβ promoter to the three inducers. Interestingly, SB20358, a specific inhibitor of p38, only partially attenuated the effect of TPA stimulation and had no effect on stimulation by TNF-α or Cer (Fig. ​(Fig.3C).3C). Finally, additional experiments ruled out the possible involvement of S6K, PKA, and MEK1 signaling in regulating PPARβ expression (Fig. ​(Fig.3C).3C). Hence, activation of the PPARβ promoter by TNF-α, TPA, and Cer is mediated through MEKK1, SEK1, and SAPK/JNK (Fig. ​(Fig.33C).

Taken together, these observations show that TNF-α, TPA, and Cer utilize overlapping signal transduction cascades that converge to JNK/SAPK and AP-1. Incidentally, the cascades that are activated by the three agents are stress-associated pathways and are invoked in response to inflammation (Kyriakis and Avruch 1996). Hence, these data attest to an important role for PPARβ in the stress- and injury-associated responses of keratinocytes.

Expression and activation of PPARβ lead to keratinocyte differentiation

The results presented so far show that expression of PPARβ, induced by inflammatory cytokines, correlates with keratinocyte cell cycle arrest and differentiation. Furthermore, experiments with PPARβ^−/− keratinocytes have shown that PPARβ is required for these processes to take place (Fig. ​(Fig.1).1). Therefore, we hypothesized that PPARβ overexpression would be sufficient to induce keratinocyte differentiation in our primary cell culture assay. To test this possibility, primary keratinocytes overexpressing PPARβ were treated or not with the synthetic PPARβ ligand L165041 (LD). A high level of PPARβ alone was not sufficient to trigger keratinocyte differentiation, whereas in conjunction with a 6-h LD treatment, these cells expressed the differentiation marker, INV (Fig. ​(Fig.4A,4A, first and second panels). Repeated LD exposure of the PPARβ overexpressing cells led both to cell cycle arrest and to keratinocyte differentiation (Fig. ​(Fig.4A,4A, third panel). However, the same treatment applied to untransfected keratinocytes resulted only in a mild expression of INV (Fig. ​(Fig.4A,4A, fourth panel). Thus, the induction of keratinocyte differentiation requires both an elevated expression of PPARβ and the presence of a PPARβ ligand that activates the receptor.

Open in a separate window

Figure 4

PPARβ accelerates keratinocyte differentiation. (A) Overexpression and activation of PPARβ stimulates keratinocyte differentiation. Keratinocytes were transfected with a wtPPARβ expression vector and lysed for RPA at various time points post-transfection, as indicated. Overexpression of PPARβ alone is not sufficient to trigger keratinocyte differentiation (first panel), which requires a 6-h exposure to a PPARβ-specific ligand L165041 (LD) (second panel). Better results were obtained with a repeated exposure to LD [6 h, followed by an additional 18 h with fresh medium containing the second dose (third panel)]. This latter dose regime was also able to slightly induce the expression of differentiation markers in absence of ectopical expression of PPARβ (fourth panel). (B) Activation of the PPARβ LBD by cytokines. Keratinocytes were cotransfected with an expression vector for a GAL4 DBD–PPARβ LBD, the Gal4 responsive reporter vector pG5CAT (Clontech), and the pCMV-β-galactosidase vector. CAT and β-galactosidase activities were measured after 24-h exposure to two different concentrations (open and solid bars) of PPARβ selective ligand L165041 (LD; 1, 5 μM), IFN-γ (10, 25 ng/ml), TNF-α (10, 25 ng/ml), TGF-β (10, 25 ng/ml), TPA (10, 20 ng/ml), Ceramide (Cer; 10, 20 μM), or arachidonic acid (AA; 10, 20 μM). Exposure to IFN-γ, TNF-α, TPA, and Cer resulted in a dose-dependent increase of the CAT reporter activity. (C) Total lipid extract of cytokine-activated keratinocytes activates PPARβ. Total lipid extract were prepared from keratinocytes untreated (control), or exposed to vehicle or to various pro-inflammatory mediators, as indicated. (Left) Transfections were performed as in A. Cells were then exposed to 1 or 2 μL of a total lipid extract for 24 h and CAT activity was measured. (Right) Keratinocytes were transfected with the wtPPARβ expression vector, pGL-3xPPRE-tk-luc reporter gene, and pCMV-β-galactosidase. Normalized reporter activity is shown as fold increase as compared with control.

TNF-α, TPA, and ceramide can generate endogenous ligands of PPARβ

Together, our results suggest that signaling by inflammatory molecules result not only in stimulation of the PPARβ gene but also leads to the production of an endogenous ligand. To test this hypothesis, an expression vector encoding a fusion protein comprising the GAL4–DNA-binding domain (DBD) and the PPARβ ligand-binding domain (LBD) was constructed. This fusion protein allows for monitoring the activation of the PPARβ LBD independently of the expression of the endogenous PPARβ gene. Keratinocytes were cotransfected with this fusion construct and a CAT reporter gene controlled by GAL4 response elements. The responsiveness of this system was validated with the synthetic PPARβ ligand L165041, which activated the fusion protein in a dose-dependent manner (Fig. ​(Fig.4B).4B). Treatment of the transfected cells with either IFN-γ, TNF-α, TPA, or Cer that were shown above to up-regulate PPARβ gene expression, also resulted in a dose-dependent increase in the transcriptional activity of the fusion protein. Interestingly, both the bioactive lipid Cer and arachidonic acid (AA) led to a dramatic increase in CAT activity (Fig. ​(Fig.4B).4B). In contrast, TGF-β had no effect on CAT activity.

To verify that the activation of the PPARβ LBD was not due to its phosphorylation but to the production of a ligand, we tested the ability of an organic extract from cytokine-activated keratinocytes to transactivate PPARβ. This experiment was performed in keratinocytes transfected either with the GAL4 DBD/PPARβ LBD fusion protein and its corresponding reporter CAT gene (Fig ​(Fig4C,4C, left), or with wild-type PPARβ and a PPRE-containing reporter luciferase gene (Fig ​(Fig4C,4C, right). In both systems, a dose-dependent increase in reporter activity was observed, showing the presence of a ligand in the organic extracts. Thus, in addition to elevating PPARβ gene expression, inflammatory signals also trigger the production of endogenous ligand(s) necessary for PPARβ activation.

PPARβ regulates genes associated with apoptosis

We have established that inflammation stimulates PPARβ expression and produces ligand(s) that activate the receptor, which together induce keratinocyte differentiation. Interestingly, amidst the TNF-α-induced keratinocyte differentiation experiments, we observed that PPARβ^−/− keratinocytes, upon a prolonged exposure to TNF-α, died at a higher rate than wild-type cells. This differential response to this cytokine, which is an inducer of apoptosis (Natoli et al. 1998), led us to test whether PPARβ controls genes involved in apoptosis. The expression of such genes was analyzed by RPA in PPARβ^−/− versus wild-type keratinocytes. Expression of the proapoptotic genes CIDE-A and CIDE-B (cell death-inducing DFF45-like effector A and B), involved in nuclear condensation and DNA fragmentation (Inohara et al. 1998), was 1.8–2.5-fold higher in the PPARβ^−/− cells (Fig. ​(Fig.5).5). In contrast, the expression of the anti-apoptotic ICAD-S and ICAD-L (inhibitor of caspase-activated deoxyribonuclease short and long forms), which are highly homologous to the human DFF45, were reduced in the PPARβ^−/− keratinocytes (Fig. ​(Fig.5).5). This is consistent with the known role of DFF45 (DNA fragmentation factor 45) in the inhibition of the DNase/apoptotic activity of CIDEs (Inohara et al. 1998). Thus, PPARβ appears to up-regulate antiapoptotic and down-regulate proapoptotic genes in keratinocytes. Whether this regulation is direct or indirect, that is, does not or does require de novo protein synthesis, was analyzed using keratinocytes treated for 12 and 24 h with the PPARβ ligand LD, in the presence or absence of cycloheximide (CH), a protein synthesis inhibitor. Consistent with the results presented above, all the genes tested responded to the PPARβ agonist in the absence of the protein synthesis inhibitor (Fig. ​(Fig.5).5). However, in the presence of CH, the down-regulation of CIDE-A and CIDE-B was lost, whereas the up-regulation of ICAD-S and ICAD-L was maintained. This result indicates that de novo protein synthesis is required for PPARβ-dependent down-regulation of CIDE-A and CIDE-B, whereas up-regulation of ICAD-S and ICAD-L is a direct effect of PPARβ. A further analysis of the molecular mechanisms by which PPARβ modulates apoptosis is underway. In addition to the genes involved in apoptosis, the expression of β-actin in the PPARβ^−/− cells was ∼2.3-fold higher than in wild-type cells, a fact worth mentioning because of the frequent use of β-actin mRNA as a internal control in expression studies. Furthermore, this result is consistent with our previous observation that PPARβ^−/− keratinocytes exhibited altered cell morphology (Michalik et al. 2001).

Open in a separate window

Figure 5

PPARβ target genes are associated with apoptosis. RPA was performed using total RNA (2.5 μg) isolated from keratinocytes derived from PPARβ^+/+ and PPARβ^−/− mice (left); or from wild-type keratinocytes not treated (−) or treated for 12 and 24 h with 1μM of the PPARβ specific ligand L165041 (LD; middle); and with 5 or 10 μg/ml of cycloheximide (CH; right). Two representative samples are shown per condition. Numbers below each band represent the fold differences in mRNA expression levels with respect to the wild-type or untreated keratinocytes.

Elevated PPARβ confers increased resistance to TNF-α-induced apoptosis

The above results implicate a role of PPARβ in the repression and stimulation of pro-apoptotic and anti-apoptotic genes, respectively. Regardless of the origin of the apoptotic stimuli, the commitment to apoptosis occurs through the activation of caspases and the fragmentation of nuclear DNA. In accordance with this process, TNF-α receptors can trigger apoptosis via activation of caspase 8, followed by DNA fragmentation. This prompted us to measure this enzymatic activity and to monitor DNA fragmentation in wild-type and PPARβ^−/− keratinocytes.

Primary keratinocytes displayed minimal caspase 8 activity (Fig. ​(Fig.6A).6A). Upon treatment with TNF-α, the caspase activity increased by 12- and 14-fold after 6 and 24 h of treatment, respectively. Interestingly, keratinocytes expressing PPARβ ectopically showed a reduced TNFα-induced caspase activity (Fig. ​(Fig.6A).6A). Incubation of these cells with a PPARβ selective ligand LD further reduced the caspase activity by 2.5-fold, when compared with untransfected wild-type keratinocytes. The importance of PPARβ in controlling caspase activity levels was further examined in PPARβ^+/− and PPARβ^−/− keratinocytes. In unchallenged keratinocytes, only PPARβ^−/− cells displayed higher basal levels of caspase 8 activity. However, both mutant keratinocytes exhibited significantly enhanced TNFα-induced caspase activity when compared with wild-type cells (Fig. ​(Fig.6A).6A). Finally, the reduced response to TNF-α seen in wild-type cells could be attributed to PPARβ expression, as the response of PPARβ^−/− keratinocytes to TNF-α could be normalized by transfection of a PPARβ-expression vector.

Open in a separate window

Figure 6

Inflammation-induced PPARβ expression protects keratinocytes from apoptosis. (A) Elevated PPARβ expression inhibits TNFα-induced apoptosis. Keratinocytes prepared from PPARβ^+/+,PPARβ^+/−, and PPARβ^−/− mice were either transfected with expression vectors harboring the wtPPARβ and/or cultured in the absence or presence of TNF-α and L165041 (LD, 5 μM) as indicated. The level of TNFα-induced apoptosis was monitored by caspase 8 activity. Keratinocytes expressing wtPPARβ were more resistant to apoptosis signals, whereas PPARβ^+/− cells showed an increase susceptibility. PPARβ^−/− keratinocytes exhibited higher basal caspase 8 activity and were more sensitive to TNFα-induced apoptosis, but could be rescued by transfection with the vector expressing wtPPARβ. Values are means of three independent experiments. (B) PPARβ^−/− keratinocytes are more susceptible to TNFα-induced apoptosis. Culture of primary keratinocytes from wild-type and from PPARβ^−/− mice were exposed to TNF-α. Apoptosis was measured by radioactive DNA fragmentation assay over indicated periods of time. Values are means of three independent experiments.

The fragmentation of genomic DNA is also a hallmark of apoptosis (Duke 1983) and, hence, was also assessed in PPARβ^−/− and PPARβ^+/+ keratinocytes upon TNFα-induced apoptosis. As expected from the caspase 8 activity measurements presented above, PPARβ^−/− keratinocytes showed higher levels of TNF-α-induced DNA fragmentation than wild-type cells over the time course monitored (Fig. ​(Fig.66B).

The observed higher caspase 8 activity, higher DNA fragmentation rate, and resulting reduced resistance to apoptosis of PPARβ^−/− cells, provide evidence for an important role of PPARβ in the control of keratinocyte death during inflammation.

PPARβ promotes terminal differentiation of keratinocytes

As mentioned above, IFN-γ and TNF-α-induced activation of PPARβ accelerates keratinocyte differentiation in addition to the antiapoptotic effect of the receptor. The inflammatory cytokines trigger both elevated expression and ligand-induced activation of PPARβ, which is required to achieve their effects. We were able to recapitulate the cytokine-dependent differentiation simply by both high-ectopic expression of PPARβ in undifferentiated keratinocytes and its activation by a specific synthetic ligand. With respect to keratinocyte differentiation, it is unclear as yet why Peters et al. (2000) did not observe any difference in INV expression upon TPA treatment between PPARβ^+/+ and PPARβ^−/− mice. Possible explanations are that their expression data were normalized to the β-actin expression level that, as shown here, is increased in PPARβ^−/− cells. An increase in INV expression, if paralleled by an elevation of β-actin synthesis would obviously have remained undetected using this normalization procedure. Alternatively, the pleiotropic effects of TPA treatment might activate both PPARβ-dependent and independent pathways of keratinocyte differentiation (Dotto 1999). Finally, topical application of TPA on skin may also have effects on the dermal fibroblasts that might modulate the keratinocyte response (Szabowski et al. 2000).

Whether the PPARβ-dependent cell differentiation noted here is keratinocyte specific or if activation of PPARβ in other tissues can accelerate cell differentiation as well, remains uncertain. In support of the latter, ectopic expression of PPARβ in NIH-3T3 fibroblasts was reported to promote early adipocyte differentiation that was dependent on the addition of PPARβ-selective ligands (Amri et al. 1995; Hansen et al. 2000). Also, PPARβ agonists were shown to stimulate oligodendrocyte differentiation and regulate the size of membrane sheets (Saluja et al. 2001).

PPARβ expression is regulated by pro-inflammatory cytokines

The epidermis is the first line of defense of the organism against aggressions from the environment, and thus must often respond to various types of injuries. As such, epidermal injury is also typically associated with inflammation, particularly during wound repair, which involves modulation of cell growth and differentiation of keratinocytes. Although PPARβ is undetectable by in situ hybridization in adult mouse interfollicular epidermis, its reactivation after a topical application of TPA, during wound repair and the healing delay observed in PPARβ^+/− mice (Michalik et al. 2001), suggested a link between this receptor and keratinocyte biology during inflammation. The findings reported herein reveal that pro-inflammatory cytokines, such as TNF-α and IFN-γ produced during skin injury, can increase expression of PPARβ and produce its ligands. The inflammatory response is one of the major differences between normal skin renewal and wound repair (Savill and Fadok 2000). The response of keratinocytes to inflammatory signals was tested by use of necrotic cells in the MLR assay, and also by treating them with the noncytokine inflammatory agent, TPA. The stimulation of the PPARβ gene by the inflammatory molecules, as studied herein using TNF-α, comprises the recruitment of FAN by p55TNF-R1, which, via activation of N-Smase, induces the production of ceramide. In turn, ceramide activates a stress-associated kinase cascade terminating at an AP-1 site in the PPARβ promoter. It is worth noting that FAN-deficient mice, which show an impaired N-Smase activation, have a delayed cutaneous barrier homeostasis and repair phenotype (Kreder et al. 1999). Hence, PPARβ up-regulation and activation is clearly a stress-associated event, such as it occurs after injury.

Inflammation-induced PPARβ expression and increased resistance to apoptosis

Certainly, the up-regulation of PPARβ in any biological system can lead to several outcomes. Here, these effects will be considered in the context of keratinocytes exposed to TNF-α. This cytokine induces oligomerization of the receptor p55TNF-R1, which leads to the formation of a multimolecular signal-transducing complex, from which several pathways originate rapidly. Paradoxically, activation of p55TNF-R1 initiates two different types of signals, the first leading to cell death, and the second protects the cell from it. p55TNF-R1 interacts with TRADD, which recruits the downstream transducers FADD and TRAF2. FADD interacts with apoptotic proteases triggering cell death (Natoli et al. 1998). In contrast, TRAF2 can activate the transcription factors NFκB, c-Jun, ATF, TCF/Elk, or SAPK/JNK, which are necessary, but not sufficient to confer protection against TNFα-induced apoptosis (Natoli et al. 1998). This points to a critical role for additional p55TNF-R1 complex components that are involved in cytoprotection against TNFα-induced apoptosis. We showed herein that such an event is the stimulation of PPARβ expression and its ligand-dependent activation, which confer increased resistance to TNFα-induced apoptosis, as measured by caspase activity and DNA fragmentation. PPARβ up-regulates antiapoptotic and down-regulates proapoptotic genes, in response to inflammatory signals in keratinocytes. Thus, up-regulation of PPARβ by the FAN-mediated pathway, which is independent of the TRAF2-associated signaling (Adam-Klages et al. 1996), is an additional way of directing TNF-α signaling away from apoptosis and toward cell survival and differentiation. This novel pathway should contribute to avoid excessive cell death during inflammation (see below). Further support to this anti-apoptotic role of PPARβ comes from the observation that increased PPARβ levels in colorectal carcinoma confers resistance to NSAID-induced apoptosis (He et al. 1999).

Cell cultures

Dendritic cells (DCs) were cultured as described by Gallucci et al. (1999). Splenic T lymphocytes were isolated as described by Peterson et al. (2000). Mouse keratinocytes were obtained from 3-day-old pups from PPARβ^+/+, PPARβ^+/−, and PPARβ^−/− (Michalik et al. 2001) and isolated from epidermis as reported by Hager et al. (1999) with some modifications; epidermis was separated from the dermis following overnight incubation at 4°C in 2.5 U/mL of Dispase. Epidermis was placed in a 50-mL centrifuge tube with 10 mL of KSFM, and the tube was given 50 firm shakes. Keratinocytes were resuspended in KSFM containing 0.05 mM Ca²⁺ and 0.1 ng/ml epidermal growth factor, and seeded at 5 × 10⁴ cells/cm². Keratinocytes were used after two to three passages. Fibroblasts were similarly isolated from the dermal pieces. Fibroblasts were plated in IMDM containing 5% FCS and were used after three to five passages.

Mixed leukocyte reaction

Induction of apoptosis in fibroblasts was performed by UV irradiation as reported (Fadok et al. 1998). Apoptotic efficiency was monitored by caspase 8 activity. Necrosis was achieved by repeated freezing at −80°C and thawing at 37°C, or by mincing newborn mouse skin (1g/mL in PBS). Large cellular debris and any remaining viable cells were removed by centrifugation at 2000g. Necrosis was assessed by trypan blue uptake.

MLR were carried out as reported (Gallucci et al. 1999). Briefly, DCs were exposed for 24 h to supernatant of necrotic cells or medium from apoptotic cells in culture, after which the medium was discarded and the activated DCs were cocultured with T-cells for 48 h. Conditioned medium obtained from this MLR (CM) was clarified of cells by centrifugation and desalted via ultrafiltration with 3000 MW cutoff YM membrane (Ambion), prior to addition to keratinocytes. Ca²⁺ concentration of ultrafiltered medium was below 0.02 mM.

For TNF-α neutralization experiment, necrosis-derived CM was preincubated at 37°C for 1 h with 20 μg/mL of anti-TNF-α antibody. After treatment, the CM was added onto the keratinocytes. The 50% neutralization dose of anti-TNF-α was 0.6μg/mL in the presence of 0.25 ng/mL of TNF-α.

RNase protection assay

The cDNA corresponding to the mouse PPARα and PPARβ A/B domain, as well as gene-specific probes corresponding to Cyc A, INV, TgaseI, K6, K10, K17, and L27 were subcloned into pGEM3Zf(+) (Promega). Gene-specific antisense riboprobes were synthesized by in vitro transcription with either T7 or Sp6 RNA polymerase (Ambion). For all riboprobes, except L27, a ratio of 1:1 of α³²P-UTP to cold UTP was used, whereas a ratio 1:20 was used for L27 probe.

RPA was carried out as described by the manufacturer (Ambion) with the following modifications: 1–2 × 10⁶ keratinocytes were lysed in 200 μL of Lysis/Denaturation buffer and clarified by centrifugation through a Qiashredder (QIAGEN). A total of 45 μL of lysate was incubated with 1 ng of gene-specific riboprobes (1 × 10⁹ cpm/μg) and 10 ng of L27 probe (1 × 10⁷ cpm/μg). RPA products were resolved in a 6% electrolyte-gradient denaturing polyacrylamide gel. Gels were dried and exposed to X-ray film or to Phosphor screen.

Transfection of primary keratinocytes

Keratinocytes were trypsinized and replated into 12-well culture plates, and incubated for 24 h to reach 40%–50% confluence. Cells were transfected using Superfect (QIAGEN). At the indicated time, cells were rinsed with phosphate buffer and lysed in Reporter Lysis buffer (Promega) for β-galactosidase and CAT assay or in Lysis/Denaturation Solution for RPA. Activity or mRNA of cotransfected pCMV-β-galactosidase reporter was used to monitor transfection efficiency. The various MKK7 and SEK1 expression vectors were from E. Nishida (Kyoto University, Japan) and the various MEKK1 and SAPK/JNK expression constructs were from R.J. Davis (University of Massachusetts Medical School, Worcester) and D.J. Templeton (University of Virginia Medical School, Charlottesville).

The proximal 1.9-kb mouse PPARβ promoter region was subcloned upstream of the CAT reporter gene in pBLCAT. The 5′ deletion promoter mutants were generated using available restriction enzyme sites. Site-directed mutagenesis of transcription factor binding sites was achieved using the QuikChange Site-Directed mutagenesis kit (Stratagene).

The LBD of PPARβ, corresponding to amino acids 136–440, was PCR amplified, and subcloned into EcoRI/HindIII sites of pMGAL4 (Clontech). The PPRE-TK-Luc reporter was from R. Evans (Salk Institute for Biological Sciences, La Jolla, CA). Lipids of cytokine-induced keratinocytes were extracted as reported (Buck et al. 1991). Briefly, 5 × 10⁶ keratinocytes were treated with indicated agents for 12 h. The organic extract was vacuum dried and the pellet dissolved in 10 μL of ethanol.

Apoptosis assays

TNF-α-induced apoptosis was carried out as reported (Reinartz et al. 1996). The apoptotic keratinocytes were measured using ApoAlert Caspase 8 Colorimetric assays (Clontech). Apoptosis was also monitored using radioactive DNA fragmentation assay (Duke 1983). Briefly, the cells, at a density of 2 × 10⁴ cells/cm², were prelabeled via incubation with 5 μCi/mL of [³H]thymidine for 24 h. TNFα-induced apoptosis were performed on these labeled cells. As background control, labeled cells were unchallenged. At indicated time points, the cells were lysed in 150 μL of Cell Lysis buffer (Boehringer Mannheim), and subsequently, 100 μL of supernatant were counted.

We thank E. Nishida, R.J. Davis, D.J. Templeton, and R. Evans for the gift of plasmids; J. Tschopp for critical reading of the manuscript; and N. Deriaz, T. Favez, and M. Narce for their assistance. This work was supported by grants from the SNSF, the Etat de Vaud, and the HFSP to B.D. and W.W. and by the Novartis Foundation to N.N.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.