NFκB-dependent Down-regulation of Tumor Necrosis Factor Receptor-associated Proteins Contributes to Interleukin-1-mediated Enhancement of Ultraviolet B-induced Apoptosis*

Activation of the transcription factor nuclear factor-κB (NFκB) by inflammatory cytokines like tumor necrosis (TNF) factor and interleukin-1 (IL-1) is generally associated with the induction of antiapoptotic pathways. Therefore, NFκB inhibits both intrinsically and extrinsically induced apoptosis and thus is regarded to act universally in an antiapoptotic fashion. Accordingly, activation of NFκB by IL-1 was shown to result in reduction of death ligand-induced apoptosis via up-regulation of antiapoptotic inhibitor of apoptosis proteins (IAPs). In contrast, apoptosis induced by ultraviolet-B radiation (UVB) was shown to be enhanced in an NFκB-dependent manner, indicating that NFκB can also act in a proapoptotic fashion. This study investigates the molecular mechanisms underlying IL-1-mediated enhancement of UVB-induced apoptosis. We show that NFκB activation in costimulation with UVB treatment results in repression of antiapoptotic genes and consequently in down-regulation of the respective proteins, like c-IAP, FLICE-inhibitory protein (FLIP), and some members of the TNF receptor-associated (TRAF)2 protein family. In parallel, TNFα is released, leading to activation of signaling pathways mediated by TNF receptor-1 (TNF-R1). Although TNF is well known to induce both proapoptotic and antiapoptotic effects, the down-regulated levels of TRAF-1, -2, and -6 proteins by IL-1 plus UVB action leads to a shift toward promotion of the proapoptotic pathway. In concert with the down-regulation of IAPs and FLIP, TNF-R1 activation as an additional proapoptotic stimulus now results in significant enhancement of UVB-induced apoptosis. Taken together, elucidation of the molecular mechanisms underlying IL-1-mediated enhancement of UVB-induced apoptosis revealed that NFκB does not exclusively act in an antiapoptotic fashion but may also mediate proapoptotic effects.

necrosis factor receptor 1 (TNF-R1), 1 CD95 (Fas/APO-1), and tumor necrosis factor-related apoptosis-inducing ligand receptors (TRAIL-R) (1). Upon binding of the specific ligands TNF␣, CD95L, or TRAIL clustering of the receptor leads to activation of the intracellular death domain resulting in apoptotic signal transduction (2,3). In contrast to CD95L, TRAIL induces apoptosis preferentially in transformed cells, whereas normal cells remain resistant (4). This selectivity suggested TRAIL to be a promising candidate for an anticancer drug (5).
The TRAIL receptor family consists of four members, of which TRAIL-R1 and TRAIL-R2 are potent inducers of apoptosis (6,7). Although evidence exists showing that TRAIL-R1 and TRAIL-R2 activation leads to nuclear factor B (NFB) activation (8,9), the major role is induction of apoptosis, starting with recruitment of Fas-associated protein with death domain (FADD) followed by activation of initiator procaspase-8, which triggers activation of effector caspases (3,10). Activation of procaspase-8 can be antagonized by competitive binding of FLICE-inhibitory protein (FLIP) to FADD or by heterodimerization of FLIP with procaspase-8 (11,12). Another way to prevent apoptosis is expression of inhibitors of apoptosis proteins (c-IAP), which interfere with activation of effector caspases (13,14).
In contrast to TRAIL, TNF␣ is a weaker inducer of apoptosis but plays an important role in mediation of immune and inflammatory responses (15). TNF-R1 triggers two different pathways upon TNF␣ binding, either a pro-or an antiapoptotic one (16). Both pathways coexist in a certain balance, and each can be promoted depending on the physiological conditions. Either recruitment of FADD leads to induction of apoptosis, or recruitment of TNF receptor-associated factor (TRAF) proteins results in activation of the transcription factor NFB, which mediates cell survival pathways (17)(18)(19). Besides its important role in controlling immune function, cell proliferation, and differentiation, NFB so far has mainly been described to induce expression of antiapoptotic genes (14,20). Therefore, NFB has been designated a tumor-promoting molecule. Accordingly, NFB was found to be constitutively up-regulated in a variety of tumor cells (21)(22)(23). Consequently, inhibition of the NFBactivating pathway is a major target for alternative anticancer drugs.
Regarding the tumor selectivity of TRAIL, it has to be taken into account that tumor cells are generally surrounded by immune cells that constantly release proinflammatory cytokines like IL-1. Therefore, IL-1-mediated activation of NFB may represent a pathway by which tumor cells can escape the cytotoxic effects of selective anticancer agents, like TRAIL. Accordingly, earlier work from our laboratory revealed that activation of NFB by interleukin-1 (IL-1) resulted in reduction of TRAIL-induced apoptosis in transformed keratinocytes, which coincided with up-regulation of c-IAP-1 and c-IAP-2 proteins (24). The same was observed for CD95-induced apoptosis, supporting the assumption that NFB protects from apoptosis universally. In contrast, apoptosis induced by UVB was shown to be significantly enhanced upon IL-1-mediated NFB activation (25). UVB-induced apoptosis is a complex process involving several pathways like induction of genomic DNA damage (26), ligand-independent clustering of death receptors (27,28), and generation of reactive oxygen species (29). All three pathways contribute independently to the complete apoptotic response to UVB. Analyzing the molecular mechanisms underlying enhancement of UVB-induced apoptosis by IL-1 we observed pronounced down-regulation of c-IAP proteins (25). Even more importantly, costimulation with IL-1 resulted in strong release of TNF␣, which additively induced apoptosis (25), suggesting that, upon UVB exposure NFB suddenly mediates effects that promote apoptosis. The present study was performed to investigate the molecular mechanisms underlying IL-1-mediated enhancement of UVB-induced apoptosis. Additionally, we wanted to determine whether NFB has the potential to mediate proapoptotic effects, because this may have important implications for photocarcinogenesis and for the potential use of NFB inhibitors in the prevention and therapy of skin cancer.

MATERIALS AND METHODS
Cells and Reagents-The human epithelial carcinoma cell line KB (American Tissue Culture Collection, Rockville, MD) was cultured in RPMI 1640 with 10% fetal calf serum. Stimulation of cells was carried out in colorless medium containing 2.5% fetal calf serum. For UVB irradiation a bank of six TL12 fluorescent bulbs (Philips, Eindhoven, The Netherlands) was used that emit most of their energy within the UVB range (290 -320 nm) with an emission peak at 313 nm. Throughout this study a dose of 350 J/m 2 was used. Control cells were subjected to the identical procedure without being exposed to UVB. To induce TRAIL-mediated apoptosis, 80 ng/ml of recombinant human TRAIL was added to the cells. This recombinant protein was N-terminally fused to a leucine zipper motif (lzTRAIL) to constitutively build the trimerized active form (49) and was obtained from Immunex Corp. (Seattle, WA). Recombinant human IL-1␤ (R&D Systems Inc., Minneapolis, MN) was applied in a final concentration of 10 ng/ml. The amount of TNF␣ secreted from cells was measured with an ultrasensitive TNF␣ ELISA (BioSource, Nivelles, Belgium). Activation of caspases was blocked by preincubating cells with 20 M zVAD (Enzyme Systems Products, Livermore, CA) for 1 h. Plasmids constructed to overexpress TRAF-1, -2, -6, FLIP, and c-IAP, respectively, and the corresponding empty vector were kindly provided by Jü rg Tschopp, Department of Biochemistry, University of Lausanne, Switzerland.
Determination of Cell Death-16 h after stimulation cells were detached from culture dishes, and apoptosis was analyzed by a cell death detection ELISA (Roche Diagnostics GmbH, Mannheim, Federal Republic of Germany). The enrichment of mono-and oligonucleosomes released into the cytoplasm of cell lysates is detected by biotinylated antihistone-and peroxidase-coupled anti-DNA-antibodies and is calculated according to the formula: absorbance of sample cells/absorbance of control cells. Unless otherwise stated, this factor was used as a parameter of apoptosis and is given as the mean Ϯ S.D. of triplicates, representing one of three independently performed experiments. Throughout the experimental settings an enrichment of 1 corresponds to 5-7% apoptotic cells as determined by Annexin V staining followed by FACS analysis (data not shown).
Analysis of Gene Expression Rates by GeneChip® Arrays-To screen KB cells for NFB-dependent differential gene expression, GeneChip® array analysis was performed utilizing Human Genome U95Av2 chips (Affymetrix®, Santa Clara, CA). 1.5 h after stimulation total RNA was extracted by the phenol/chloroform extraction method utilizing phase lock tubes (Eppendorf, Hamburg, Federal Republic of Germany). 25 g total RNA was subjected to first and second strand cDNA synthesis using T7-(dT24)-primer (MWG-Biotech, Ebersberg, Federal Republic of Germany) and enzymes from Invitrogen according to the protocol provided by Affymetrix®. Biotinylated cRNA was in vitro transcribed from 10 g of cDNA utilizing the ENZO-kit from Affymetrix® and subsequently fragmented according to the manufacturer's recommendation (see technical manual: GeneChip® Expression Analysis, by Affy-metrix®). Samples were analyzed using an Affymetrix® fluidic station 400 and a Affymetrix® 3000 Scanner with high density upgrade. The data were scanned and generated by the software MicroArray Suite 5.0 (MAS 5.0) from Affymetrix®. The signals were log transformed and calculated according to the statistical algorithms of the MAS 5.0, which calculates an intensity value and a detection call corresponding to "absent," "present," or "marginal" for each transcript (Affymetrix® User Manual). Furthermore, we used the software package Expressionist Suite (GeneData, Basel, Switzerland). The software package includes the Expressionist Refiner version 3.0.4. This allows an additional global quality control of the GeneChip® with detection and masking of defect areas on the chip and outliers, gradient correction of fluorescence signal, and regulation of variance.
The statistical analysis was performed with the Expressionist Analyst version 4.0.5. Data were first normalized on a logarithmic mean of 210 in all experiments. We built groups with four experiments per group for the IL-1 stimulation and five experiments per group for the other stimulations and the control. Data were filtered for genes with valid values (expression over background respective a detection call "present" or "on" (see Affymetrix® User Manual) in 50% or more per group. For sample comparison, we calculated each stimulation group against the control group with n-fold regulation by comparing the mean and the p value by the Student's t test for statistical significance of the respective groups.
Validation of Gene Expression Rates by Real-time PCR-3 h after stimulation, total RNA was extracted from KB cells by the phenol/ chloroform extraction method utilizing phase lock tubes (Eppendorf). 3 g of total RNA was reverse transcribed using a SuperScript TM II kit (Invitrogen). Assays-on-demand® were used to quantify gene expression of TRAF-1 (Hs00194638_m1), TRAF-2 (Hs00184192_m1), and TRAF-6 (Hs00371508_m1) (Applied Biosystems, Foster City, CA). Expression rates were normalized to the expression of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (Hs99999905_m1) using the TaqMan® Universal kit in an ABI Prism 7900HT Real-time PCR System supplied with SDS 2.1 software (Applied Biosystems). Expression rates were quantified using the 2 Ϫ⌬⌬CT method as described by Livak and Schmittgen (30).
Staining of Intracellular Proteins and FACS Analysis-6 h after stimulation cells (1 ϫ 10 6 ) were removed from culture dishes, washed with PBS and fixed with 0.8% paraformaldehyde in PBS for 5 min on ice. After washing, cells were permeated with 0.3% saponin (Sigma) in PBS and incubated with the first antibody (20 g/ml for TRAFs and 10 g/ml for c-IAP and FLIP) overnight at 4°C. Monoclonal mouse-anti c-IAP and anti-TRAF antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA): anti-c-IAP (N19, sc-1867), anti-TRAF-1 (H3, sc-6253), anti-TRAF-2 (H249, sc-7187), and anti-TRAF-6 antibody (H274, sc-7221). Polyclonal rabbit-anti-FLIP (556567, BD Pharmingen) was used. After washing, the respective secondary antibody (10 g/ml donkey-anti-goat IgG, phycoerythrin (PE) labeled for c-IAP; 10 g/ml goat-anti-mouse, PE labeled for TRAFs and bcl-2 and 20 g/ml goatanti-rabbit IgG, PE labeled for FLIP) was incubated in 0.3% saponin for 30 min on ice. Subsequently, cells were washed in PBS and subjected to flow cytometry analysis in PBS with 0.03% saponin. FACS analysis was performed using an EPICS® XL-MCL flow cytometer (Coulter, Miami, FL). 20,000 cells were analyzed for each sample. Data analysis was performed by using the WIN/MDI 2.8 software program (Flow Cytometry Core Facilities, The Scripps Research Institute, La Jolla, CA).

RESULTS
Stimulation of cells from the epithelial cell line KB with TRAIL or irradiation with UVB induces apoptosis. IL-1 causes activation of the transcription factor NFB, which is known to be responsible for the expression of a number of antiapoptotic genes. Accordingly, costimulation of cells with IL-1 resulted in reduction of TRAIL-induced apoptosis, as demonstrated previously (24). In contrast, costimulation with IL-1 caused a significant enhancement of UVB-induced apoptosis (Fig. 1A). This enhancement coincided with a strong release of the proapoptotic cytokine TNF␣ (ϳ100 pg/ml), whereas no TNF␣ was released upon stimulation with IL-1 alone or in combination with TRAIL ( Fig. 1B). TNF␣ at 100 pg/ml does not induce apoptosis in KB cells itself. Addition of TNF␣ and of supernatants of IL-1 plus UVB-stimulated cells to UVB-irradiated cells only slightly enhanced UVBinduced apoptosis in comparison to an addition of IL-1 (Fig. 1C). This indicates that other intracellular mechanisms are responsible for intensifying the proapoptotic effect of TNF␣ that results in enhancement of UVB-induced apoptosis.
IL-1-mediated activation of NFB essentially requires phosphorylation of two specific Ser residues (Ser-32/36) of its cytoplasmic inhibitor IB followed by ubiquitination and proteasomal degradation. In the super-suppressor variant utilized in this experiment both Ser-32/36 had been substituted by Ala thereby preventing phosphorylation and degradation of IB (IB-DN). Cell death analysis revealed that in cells transiently transfected with IB-DN the enhancing effect of IL-1 on UVBinduced apoptosis as well as its inhibiting effect of TRAILinduced apoptosis was abolished, indicating NFB to be critically involved in the underlying mechanisms (Fig. 2).
We have previously demonstrated that NFB mediates differential regulation of c-IAPs at the protein level depending on the apoptotic costimulus (25). To expand on these findings and to further study transcription rates under these conditions, GeneChip® arrays were performed using cells stimulated with IL-1 alone or in combination with either TRAIL or UVB in comparison to untreated cells. Stimulation with IL-1 alone or IL-1 plus TRAIL resulted in a 16-to 18-fold up-regulation of antiapoptotic c-IAP mRNA and an up-regulation of antiapoptotic FLIP of ϳ4-fold, respectively, but only a 2-fold up-regulation of TNF␣. In contrast, stimulation with IL-1 plus UVB caused only a 3-fold up-regulation of c-IAP and down-regulation of FLIP, whereas the TNF␣ gene was 5-fold up-regulated (Table I).
When screening for other genes being differentially regulated by NFB, we found genes encoding TRAF-1 and TRAF-6 to be critically down-regulated upon IL-1 plus UVB treatment (Table I). Because TRAF proteins are involved in TNF-R1 as well as IL-1R-mediated signal transduction, differences in their expression rates may influence signal transduction path-FIG. 1. IL-1 causes reduction of TRAIL-induced apoptosis but enhancement of UVB-induced apoptosis. A, KB cells were left untreated or stimulated with 10 ng/ml IL-1 alone, with IL-1 plus UVB (350 J/m 2 ), or IL-1 plus TRAIL (80 ng/ml). After 16 h, apoptosis was determined using a cell death detection ELISA. The rate of apoptosis is reflected by the enrichment of nucleosomes in the cytoplasm shown on the y-axis (mean Ϯ S.D. of triplicate samples). B, supernatants from the cells tested in A were subjected to a TNF␣ ELISA to determine the amount of TNF␣ secreted in an autocrine fashion. C, either untreated or UVB-irradiated cells were stimulated with TNF␣ (100 pg/ml), with supernatants from IL-1 plus UVB-treated cells or with IL-1 (10 ng/ml). After 16 h, apoptosis was quantified by a cell death detection ELISA, displaying the mean Ϯ S.D. of triplicates. ways triggered by the respective receptors, thereby influencing the fate of the cell.
The results concerning the different expression rates of TRAF genes obtained from GeneChip® data were verified by real-time PCR (Table II). Accordingly, TRAF-1 was shown to be ϳ16-fold up-regulated upon IL-1 or IL-1 plus TRAIL stimulation, whereas it was only 7-fold up-regulated in case of IL-1 plus UVB treatment, in comparison to untreated cells. TRAF-6 expression was only minimally influenced by different proapoptotic costimuli, but down-regulated by the combination of IL-1 plus UVB. The TRAF-2 gene, which was not included in the GeneChip®, was at least 2-fold up-regulated by IL-1 or IL-1 plus TRAIL, respectively, but 2-fold down-regulated by IL-1 plus UVB (Fig. 3). Taken together, IL-1-mediated up-regulation of TRAF-1, -2, and -6 genes remains unaffected by TRAIL treatment but is antagonized by UVB radiation.
FLIP is known to be a major inhibitor of death receptordriven apoptosis, because it prevents activation of the proximally acting procaspase-8. Efficient blockade of downstream effector caspases, like caspase-3, is procured by c-IAP protein. Intracellular FACS analysis was performed to show that differential c-IAP and FLIP regulation is also reflected at the protein level. 6 h after stimulation FLIP protein was found to be up-regulated by IL-1, whereas it took 8 h to significant increase the protein level of c-IAP by IL-1. Along this line, cotreatment with IL-1 counteracted TRAIL-mediated downregulation of both FLIP and c-IAP. In contrast, IL-1 stimulation even enhanced the UVB-mediated down-regulation of FLIP and c-IAP (Fig. 4). Accordingly, the differential regulation of FLIP and c-IAP follows the same regulation pattern as indicated at the mRNA expression level.
To elucidate whether decreased c-IAP and FLIP protein levels contribute to the enhancement of UVB-induced apoptosis by IL-1, cells were transiently transfected with plasmids encoding c-IAP and FLIP, respectively. 16 h after IL-1 plus UVB or IL-1 plus TRAIL treatment cell death detection ELISA revealed that overexpression of either c-IAP or FLIP counteracts IL-1mediated enhancement of UVB-induced apoptosis. Overexpression of neither FLIP nor c-IAP influenced IL-1 plus TRAILinduced apoptosis, possibly because in this case these antiapoptotic proteins were up-regulated anyway (Fig. 5).
TRAF adapter proteins are involved in signal transduction pathways triggered by TNF-R1 and IL-1R, respectively, mediating cell survival upon NFB activation. Consequently, downregulation of TRAF proteins may support a pronounced activation of the alternative proapoptotic pathway triggered by TNF-R1. Intracellular FACS analysis of cells 6 h after stimulation confirmed that stimulation with IL-1 alone leads to up-regulation of TRAF-1 as well as of TRAF-2 and TRAF-6 proteins. Furthermore, TRAIL-mediated down-regulation of all three TRAF proteins was counteracted by costimulation with IL-1. In contrast, costimulation of cells with IL-1 plus UVB resulted in down-regulation of each TRAF-1, -2, and -6 in an IL-1-dependent manner. In accordance with the mRNA expression rates regulation was strongest for TRAF-1, indicating that TRAF-1 regulation plays a major role in this scenario (Fig. 6A). All the effects were very likely to be mediated by NFB, because upregulation of TRAF proteins by IL-1 alone or IL-1 plus TRAIL as well as their down-regulation by IL-1 plus UVB stimulation was abolished in cells transiently transfected with the IB-DN variant described above (Fig. 6B).
Because the RNA expression rate of TRAF-1 as demonstrated in Table II did not correlate with the protein expression level upon IL-1 plus UVB stimulation (Fig. 6A), we studied whether post-translational proteolysis of TRAF-1 may be responsible for this discrepancy. In fact, inhibition of proapoptotic caspases with the pancaspase inhibitor zVAD significantly enhanced the TRAF-1 protein levels as demonstrated by intracellular FACS analysis performed 6 h after stimulation (Fig. 7).
To elucidate whether a reduction of TRAF protein level might promote proapoptotic pathways triggered by TNF␣ through activation of TNF-R1 and may thereby contribute to enhancement of UVB-induced apoptosis by IL-1, cells were transiently transfected with plasmids encoding TRAF-1, TRAF-2, and TRAF-6, respectively. 16 h after IL-1 plus UVB or IL-1 plus TRAIL stimulation, cell death detection ELISA revealed that cells overexpressing either one of these TRAF proteins became more resistant against IL-1 plus UVB-induced apoptosis, whereas IL-1 plus TRAIL-induced apoptosis remained unaffected (Fig. 8). These results strongly indicate that NFB-dependent down-regulation of TRAF-1, TRAF-2, and TRAF-6, respectively, contributes to IL-1-mediated enhancement of UVB-induced apoptosis. DISCUSSION The death ligand TRAIL as well as UVB radiation are potent inducers of apoptosis in transformed keratinocytes and epithelial cells. Previous work from our laboratory revealed that costimulation with IL-1 protects transformed keratinocytes from the cytotoxic effect of TRAIL in an NFB-dependent manner (24). Because IL-1 can be released by a variety of tumor cells and is also released by inflammatory cells participating in the tumor-host immune response (32), cancer cells under these conditions may escape the therapeutic effect of TRAIL through NFB activation.
The NFB family of transcription factors (Rel (c-Rel), Rel A (p65), Rel B, NFB-1 (p105/p50), and NFB-2 (p100/p52)) is involved mainly in stress-induced, immune, and inflammatory responses. Together, these proteins regulate the expression of TABLE I IL-1 differentially influences the expression of either pro-or antiapoptotic genes, depending on the apoptotic costimulus Results of GeneChip® experiments using Affymetrix® Hg-U95Av2 chips, reflecting the expression rate of different genes upon treatment with IL-1, IL-1 plus TRAIL and IL-1 plus UVB, respectively, taken 1.5 h after stimulation. Groups contain 4 replicates for IL-1 stimulation and 5 replicates for the three other groups. Genes with high n-fold scores and significant p values are highlighted in gray.
* n-fold regulation was calculated for each stimulation group against the control group. § Students t test was used to determine statistical significance of mean n-fold, reflected by the p value. genes encoding cytokines, chemokines, adhesion molecules, and antiapoptotic proteins (33). For a long time, NFB was considered to act exclusively as a mediator of cell survival, antagonizing cell death by up-regulation of antiapoptotic proteins. Correspondingly, the development and progression of a number of malignancies like Hodgkin's disease (21), squamous cell carcinoma from head and neck (23), and primary breast cancers (22) appear to be linked to constitutive NFB activation. Thus, strategies that interfere with the signal transduction pathways activating NFB are major targets for future anticancer interventions (34,35).
The observation that NFB is capable of rescuing cells from death induced by TNF␣ (14,20), TRAIL (24), CD95L (25), and ionizing radiation and chemotherapeutic drugs (36) gave rise to the speculation that NFB protects from apoptosis universally. Thus, it was quite surprising to observe that IL-1 did not protect from, but even enhanced, apoptosis induced by UVB radiation. Enhancement of UVB, as well as reduction of TRAIL-induced apoptosis, seem to be NFB-dependent, because prevention of NFB activation in cells overexpressing a super suppressor of its inhibitor IB reversed both effects (Fig.  2). Furthermore, enhancement of UVB-induced apoptosis was shown to coincide with a pronounced release of TNF␣, which as an additional apoptotic stimulus triggers TNF-R1, indicating that enhancement of UVB-induced apoptosis by IL-1 correlates with TNF␣ release and is NFB-dependent (25). Surprisingly, addition of the same amount of TNF␣ in the range of 100 pg/ml to UVB-irradiated cells did not significantly enhance UVBinduced apoptosis neither did addition of supernatants from IL-1 plus UVB-treated cells. These two observations clearly rule out the possibility that either TNF␣ alone or other poten-tial proapoptotic mediators released from the cells are sufficient to enhance UVB-induced apoptosis. In turn, these findings suggest that other intracellular mechanisms have to be involved as well, most likely mechanisms that intensify the proapoptotic effects mediated by activation of TNF-R1 through binding of autocrine released TNF␣.
One of these mechanisms includes NFB-mediated repression of c-IAP and FLIP genes. Both c-IAP and FLIP are very potent inhibitors of almost any proapoptotic pathway, because they interfere with the activation of initiator caspase-8 (FLIP) and effector caspases (c-IAP), respectively (11)(12)(13)37). It is tempting to propose that UVB drives NFB to exert its proapoptotic effect by a general repression of antiapoptotic genes that NFB usually activates. A modified phosphorylation pattern of NFB upon UVC or daunorubicin treatment was recently shown to manipulate NFB to interact with chromatin silencing histone deacetylase molecules, thereby mediating specific gene repression of antiapoptotic genes, like c-IAP and bcl-2 (38). These findings strongly support our concept that UVB radiation changes the working pattern of NFB. However, this may not be the only mechanism responsible, because NFB upon UVB irradiation strongly accelerates the expression of the proapoptotic TNF␣ gene, which is not effectively induced by IL-1 alone. As displayed in Table I, TNF␣ was also 2-fold up-regulated upon IL-1 or IL-1 plus TRAIL stimulation but no TNF␣ was released (Fig. 1). This phenomenon is due to the fact that, after a short initial induction of the TNF␣ gene upon any stimulation involving IL-1, transcription is completely repressed again after 2 h in the case of IL-1 and IL-1 plus TRAIL stimulation, respectively. This short term TNF␣ expression is obviously not sufficient to result in appropriate protein synthesis and autocrine secretion. In contrast, transcriptional activation stays stable for several hours in the case of IL-1 plus UVB treatment, as documented by semiquantitative PCR analysis (data not shown).
A similar way of NFB-dependent gene repression upon UVB irradiation appears to apply for TRAF genes TRAF-1 and TRAF-6, whereas TRAF-3, TRAF-4, and TRAF-5 expression remains unaffected (data not shown). TRAF-2 is one of six TRAF protein members and functions as an activator of NFB (39,40). TRAF-1-and -6-like TRAF-2 are involved in NFB activation (41). TRAF-1 was shown to heterodimerize with TRAF-2, thereby enhancing its antiapoptotic effects, i.e. NFB activation and mediation of cell survival signals (42,43). TRAF-6 mediates interleukin-1 receptor-induced activation of NFB (44). Because TRAF-2 was not present on the Gene-Chip®, real-time PCR experiments were performed that confirmed the assumption that besides TRAF-1 and TRAF-6 also the TRAF-2 gene is differentially regulated by NFB, depending on the apoptotic costimulus. All three TRAF genes contain TRAF proteins is critically influenced by the apoptotic costimulus KB cells were either left untreated or stimulated with IL-1 (10 ng/ml) alone, with IL-1 plus UVB (350 J/m 2 ), or with IL-1 plus TRAIL (80 ng/ml). 3 h later cDNA was generated from total RNA and subjected to real-time PCR analyses utilizing assays-on-demand probes from Applied Biosystems®. Expression rates of TRAF1, TRAF2, and TRAF6, respectively, were normalized to glyceraldehyde-3-phosphate dehydrogenase expression levels. Genes with high scores and significant results are highlighted in gray.
* ⌬ CT values were calculated by software algorithms of AB-protocols. The table shows the mean of the ⌬ CT values for each gene and each group with the associated standard deviation.
§ The corresponding p value was calculated with Student's t test for significance analysis of three independent performed experiments.  Table II) leads to n-fold regulation of the respective genes.
NFB consensus sequences within their promoters and have previously been shown to become activated in an NFB-dependent manner (45). Here we could demonstrate that TRAF proteins cannot only be up-regulated but can also be downregulated upon costimulation with IL-1, indicating again that NFB exerts different regulatory functions depending on the proapoptotic costimulus applied (Fig. 6, A and B). The expression rate determined for TRAF-2 and TRAF-6 was only slightly enhanced upon treatment with IL-1 alone or IL-1 plus TRAIL. Nevertheless, in comparison to this up-regulation the transcription efficacy was repressed even below the control level by IL-1 plus UVB stimulation. The differences in TRAF-1 expression were even more pronounced, revealing very strong induction by IL-1 and IL-1 plus TRAIL, respectively, but only minor by IL-1 plus UVB treatment. Taken together, the transcription efficacy of all three TRAF members was significantly impaired upon IL-1 plus UVB treatment.
Based on the previous observation of either up-or downregulation of c-IAP by different apoptotic stimuli after 16 h (25), we determined whether early changes of c-IAP and FLIP and TRAF adapter proteins after 6 h may influence the fate of the cell upon costimulation of either UVB or TRAIL with IL-1. In accordance with the GeneChip® data, intracellular FACS analysis confirmed down-regulation of FLIP and c-IAP as well as of TRAF proteins upon costimulation with IL-1 plus UVB, whereas the proteins were up-regulated by IL-1 in the absence as well as in the presence of TRAIL. IL-1-mediated differential TRAF regulation was shown to be clearly NFB-dependent, because inhibition of NFB with an IB super-suppressor almost completely reversed both IL-1 mediated up-as well as down-regulation of TRAF proteins. In the case of TRAF-1, the protein expression rate did not correlate with the pronounced enhancement of the transcription rate. This might be due to a general low mRNA protein turnover. Nevertheless, costimulation with IL-1 plus UVB resulted in a further reduction of protein amounts, although real-time PCR data still showed a 7-fold up-regulation of TRAF-1 mRNA. This phenomenon is obviously due to caspase-dependent proteolytic degradation of TRAF-1, because addition of the pancaspase inhibitor zVAD significantly enhanced the protein level of TRAF-1. Accordingly, previous studies already identified TRAF-1 as a cleavage substrate for caspases during induction of death receptor-induced apoptosis (46 -48). In contrast, upon stimulation with IL-1 plus TRAIL up-regulation of antiapoptotic FLIP and c-IAP proteins seems to prevent TRAF-1 cleavage by blocking caspase activation.
The data shown imply the existence of a biological correlation between NFB-dependent differential gene regulation and a bivalent influence on apoptosis induced by either death receptor activation or UVB radiation. Accordingly, overexpression of either of the antiapoptotic proteins c-IAP and FLIP or the adapter proteins TRAF-1, TRAF-2, and TRAF-6, respectively, counteracted the enhancing effect of IL-1 on UVB-induced apoptosis completely, whereas it had no significant impact on IL-1 plus TRAIL-induced apoptosis, in which all the antiapoptotic proteins mentioned become up-regulated in an IL-1-dependent manner.
The present study clearly demonstrates that down-regulation of antiapoptotic proteins and adapter proteins together determine the sensitivity of the cell toward UVB radiation, In fact this is a secondary effect mediated via TNF-R1 that becomes activated by autocrine release of sublethal TNF␣ doses. Because treatment of cells with IL-1 plus UVB additionally results in repression of antiapoptotic FLIP and c-IAP genes and NFB-dependent down-regulation of TRAF-1, -2, and -6 proteins, the following scenario may take place within the cell. The quickly released TNF␣ (release starts after 1 h, data not shown) activates the TNF-R1, resulting first of all in TNF receptor-associated protein with death domain recruitment. Due to down-regulation of TRAF-1 and TRAF-2, the balance at the TNF-R1 shifts toward binding of the proapoptotic adapter protein FADD, which transduces the proapoptotic signal, resulting in induction of cell death. Furthermore, down-regulation of TRAF-6 mitigates NFB activation triggered by IL-1 through IL-1R. Additionally, inhibition of caspase-8 by FLIP and of effector caspases by c-IAP is strongly impaired due to NFB-dependent down-regulation of these antiapoptotic proteins. In summary, the different molecular mechanisms de-FIG. 6. IL-1-dependent differential regulation of TRAF-1, TRAF-2, and TRAF-6 proteins is NFB-dependent. A, KB cells were left untreated or stimulated with UVB (350 J/m 2 ) or TRAIL (80 ng/ml) alone and in combination with IL-1 (10 ng/ml). After 6 h intracellular FACS analysis was performed utilizing PE-labeled antibodies directed against TRAF-1, -2, and -6, respectively. Histograms show fluorescence intensity (x-axis) versus cell number (y-axis) of cells treated without IL-1 (gray shaded histogram) and with IL-1 (black line), respectively, in comparison to isotype control (dashed line). B, experiments were performed exactly as described in A using cells transiently transfected with an IB supersuppressor. Data presented show one representative of three independently performed experiments. scribed shift the balance at the TNF-R1 toward proapoptotic signaling, resulting in enhancement of UVB-induced apoptosis (Fig. 9). In this scenario NFB mediated TNF␣ release, which exclusively occurs upon IL-1 plus UVB treatment, serves as the initial trigger of apoptosis enhancement via activation of TNF-R1 but is not sufficient alone. Down-regulation of TRAF proteins then sets the course for initiating the apoptotic pathway. In this context down-regulation of TRAF-1 seems to play the major role. Finally down-regulation of apoptosis inhibitors FLIP and c-IAP intensify apoptosis triggered by TNF-R1. Here down-regulation of FLIP is the predominant mediator, because it inhibits receptor-driven apoptosis more upstream and its half-life seems to be shorter that the one of c-IAP, as deduced from intracellular FACS analysis.
Thus, the present findings imply that upon UVB radiation NFB has the potential to trigger physiologically proapoptotic instead of antiapoptotic pathways within the cell. Although a mechanism underlying UV-mediated and NFB-dependent gene repression has been proposed (38), the mechanism by which NFB, upon UVB radiation, induces genes that it usually does not induce is still unclear. Studying the role of NFB under different conditions on the physiological status of the cell is of primary importance, because the combination of phototherapy and activators of NFB may be of potential practical value for the treatment of malignancies especially of the skin. The present study may contribute to a change in the dogma that NFB exclusively mediates antiapoptotic effects. This observation should be carefully taken into account when designing new concepts for anticancer therapy, especially for those utilizing NFB inhibitors.