Tumor and Stem Cell Biology TNF-a Promotes c-REL / D Np 63 a Interaction and TAp 73 Dissociation from Key Genes That Mediate Growth Arrest and Apoptosis in Head and Neck Cancer

Inflammation-induced activation of proto-oncogenic NF-kB/REL and dysfunction of tumor suppressor TP53/ p63/p73 family transcription factors are key events in cancer progression. How inflammatory signaling coordinates dysregulation of these two transcription factor families during oncogenesis remains incompletely understood. Here, we observed that oncoprotein c-REL and tumor suppressor TAp73 are coexpressed and complex with DNp63a in the nucleus of a subset of head and neck squamous cell carcinoma (HNSCC) cell lines with mutant (mt)TP53. TNF-a, a proinflammatory cytokine, promoted c-REL nuclear translocation, c-REL/ DNp63a interaction, and dissociation of TAp73 from DNp63a and the nucleus to the cytoplasm, whereas c-REL siRNA knockdown attenuated this effect. Overexpression of c-REL or a c-REL kB-site DNA-binding mutant enhanced protein interaction with DNp63a and TAp73 dissociation, implicating c-REL/DNp63a-specific interactions in these effects. We discovered that TNF-a or genetic alteration of c-REL expression inversely modulates DNp63a/TAp73 interactions on distinct p63 DNA-binding sites, including those for key growth arrest and apoptotic genes p21WAF1, NOXA, and PUMA. Functionally, c-REL repressed these genes and the antiproliferative effects of TNF-a or TAp73. Conversely, c-REL siRNA depletion enhanced TAp73 promoter interaction and expression of genes mediating growth arrest and apoptosis. Similar to TNF-a–treated HNSCC lines, human HNSCC tumors and hyperplastic squamous epithelia of transgenic mice overexpressing DNp63a that exhibit inflammation also show increased nuclear c-REL/DNp63a and cytoplasmic TAp73 localization. These findings unveil a novel and reversible dynamic mechanism whereby proinflammatory cytokine TNF-a–induced c-REL/ DNp63a interactions inactivate tumor suppressor TAp73 function, promoting TNF-a resistance and cell survival in cancers with mtTP53. Cancer Res; 71(21); 6867–77. 2011 AACR.


Introduction
TP53, a tumor suppressor that mediates growth arrest and apoptosis of critically damaged cells, is the most frequently mutated gene in cancer, including head and neck squamous cell carcinomas (HNSCC; refs.1,2).The TP53 family also includes p63 and p73, which have overlapping functions in cell growth and apoptosis, as well as development of squamous epithelia of the skin and mucosa (3,4).When TP53 is mutated or inactivated, transactivating (TA) isoforms TAp63 and TAp73 can potentially replace the tumor suppressor function of TP53.TAp63 and TAp73 have full-length N-terminal domains which share homology, transactivating, and tumor suppressor function with TP53.Although TAp63 and TAp73 are rarely mutated, expression of alternatively transcribed DN isoforms can differentially affect their transactivation and tetramerization.Recent evidence suggests that the DNp63a isoform is predominantly overexpressed together with TAp73 in major subsets of HNSCC, breast, and other cancers (5).Nuclear interactions between DNp63a and TAp73 proteins on regulatory promoters have been implicated in repression of genes mediating growth arrest and apoptosis (4,5).However, what factor(s) regulate these interactions between overexpressed DNp63a and TAp73, and how their interaction leads to inactivation of TAp73 in cancers with mtTP53, remains unclear.
Among candidate regulatory factors, TNF-a is a cytotoxic cytokine which is expressed by cancer and infiltrating inflammatory cells in the tumor microenvironment of many cancers, including HNSCC (6,7).Interestingly, we found that HNSCC are paradoxically resistant to TNF-a-mediated growth arrest and apoptosis, and that such resistance involves aberrant activation of NF-kB/REL transcription factors (8).Among these, TNF-a has been shown to signal via a canonical pathway to promote nuclear translocation and transactivation of RELA (p65) and c-REL (9).RELA (p65) promotes expression of genes that enhance cell proliferation and survival of cancers, including HNSCC (10)(11)(12).However, inhibition of RELA and these prosurvival genes by siRNA in vitro or by proteasome inhibitor in a phase I clinical trial in vivo showed limited cytotoxic and clinical activity (11)(12)(13).Amplification and nuclear localization of c-REL has also been previously detected and found to be relatively unaffected by proteasome inhibition in HNSCC (13,14).These observations raised the question whether TNFa-regulated or overexpressed c-REL may contribute to inhibition of growth arrest and apoptosis of HNSCC by a distinct mechanism.
We recently discovered that overexpressed DNp63a colocalizes and forms novel nuclear complexes with c-Rel in murine keratinocytes or ortholog c-REL in human HNSCC (15).Herein, we explored the hypothesis that TNF-a and c-REL activation is linked to modulation of the aforementioned DNp63a/TAp73 interactions and dysregulation of growth arrest and apoptosis in cancer.We reveal a dynamic and reversible mechanism whereby TNF-a promotes c-REL nuclear translocation and interaction with DNp63a, dissociation of TAp73 from distinct p63 promoter sites of key growth arrest and apoptotic genes, and TAp73 translocation to the cytoplasm.These results can help explain how inflammatory factor TNF-a and c-REL, through DNp63a, inhibit the compensatory ability of TAp73 to activate genes that mediate growth arrest and apoptosis in HNSCC with mutant TP53.

Cell lines
The characteristics, TP53 genotype, and culture of HNSCC (UMSCC) cell lines obtained from the University of Michigan were previously described (16,17).UMSCC lines used herein were obtained in 2008.Authentication was done at the University of Michigan by DNA genotyping of alleles for 9 loci, D3S1358, D5S818, D7S820, D8S1179, D13S317, D18S51, D21S11, FGA, vWA, and the amelogenin locus in 2008 and 2010, as recently described (18).Cell line stocks were preserved at À80 C and cultured for fewer than 6 months before use.

Western blot
Western blot analysis was done as previously (17).Primary antibodies are listed in Supplementary Methods.

Real-time RT-PCR
RNA isolation and cDNA synthesis were done as previously (19).Real-time PCR primers and probes for DNp63 and TAp63 validated previously (20) were synthesized by Applied Biosystems.Primers to amplify TAp73 and DNp73 are listed in Supplementary Methods.

Plasmid, siRNA transfection, and reporter gene assays
These experiments were conducted as described previously (19).DNp63-specific siRNA (21) was synthesized by Integrated DNA Technologies (IDT).TAp73-specific siRNA was purchased from Dharmacon.The 2.6-kb p21-luc reporter was provided by Dr. Bert Vogelstein laboratory (22).Each sample was assayed in triplicate and data were presented as mean AE SD.

Coimmunoprecipitation analysis
Coimmunoprecipitation (co-IP) was done as described previously (15).Antibodies used are listed in Supplementary Methods

Chromatin immunoprecipitation analyses
Chromatin immunoprecipitation (ChIP) assays were done using the Magna ChIP G kit (Upstate) as previously (24).For quantitative PCR primers see Supplementary Methods.

DNA-based cell-cycle and apoptotic flow cytometric analysis
DNA flow cytometric analysis was done as previously (12).Samples were run on fluorescence-activated cell analyzer (FACS Canto, BD Biosciences, San Jose, CA) within 1 hour and analyzed using DIVA flow cytometric analysis software.
Additional experimental detail, methods, and figures are included in Supplementary Information

Both c-REL and TAp73 exhibited basal interaction with
DNp63a in whole-cell extracts from 2 independent cell lines by co-IP analysis (Fig. 1A; UMSCC-22A or -46).However, c-REL and TAp73 exhibited weak or negligible interaction with one another, seen only with supersensitive substrate and prolonged exposure.Furthermore, c-REL, but not RELA(p65), interacted with DNp63a (although RELA formed an independent complex with c-REL; Fig. 1B), providing evidence for distinct interactions by these RELs.The complexes between c-REL, DNp63a, and TAp73, identified in whole-cell extracts above, were found in nuclear, but not cytoplasmic extracts (Fig. 1C).No interaction with mtTP53 was observed (Fig. 1B and C).In the absence of further stimulation, a significant portion of nuclear DNp63a co-IPed with either c-REL or TAp73, when compared with input (Fig. 1D).

TNF-a or genetic modulation of c-REL affects nuclear localization and interactions between c-REL, DNp63, and TAp73
The inflammatory cytokine TNF-a is a canonical inducer of c-REL and is expressed in the HNSCC microenvironment in vivo, but not by these cell lines in vitro (7,25).We examined the effect of exogenous TNF-a on the cellular distribution of c-REL, DNp63a, and TAp73 in the UMSCC-22A cell line, which has lower endogenous c-REL expression, making it amenable to TNF-a or genetic modulation (Supplementary Fig. S1).TNF-a stimulation increased nuclear localization of c-REL within 30 minutes (Fig. 2A, left panel).Remarkably, TNF-a not only induced c-REL but also decreased TAp73, without significantly altering DNp63a expression in nuclear extracts (Fig. 2A, right panel).Moreover, TNF-a stimulation increased the interaction between nuclear c-REL and DNp63a and a corresponding dissociation of nuclear DNp63 and TAp73 in co-IP (Fig. 2B).This dissociation corresponded with a decrease in TAp73 in the nuclear fraction and appearance in the cytoplasmic fraction (Fig. 2C).Conversely, c-REL knockdown by siRNA enhanced DNp63a/TAp73 interaction without or with TNF-a (Fig. 2D).
Similar to TNF-a-induced c-REL, overexpressing c-REL-Flag in UMSCC-22A resulted in increased interaction between c-REL and DNp63a in the nuclear fraction (Fig. 3A, top panels).Concurrently, TAp73 dissociated from DNp63a in the nuclear fraction and appeared in the cytoplasmic fraction (Fig. 3A, bottom panels).Further supporting that c-REL and TAp73 form alternative complexes with DNp63a, no significant interaction between c-REL and TAp73 was detected in co-IP.Overall, the above findings indicate that TNF-a-induced or overexpressed c-REL is involved in the dissociation of TAp73 from DNp63a in the nucleus to the cytoplasm.
Previous studies have indicated that c-REL can bind kB motifs bound by other NF-kB/REL members, as well as distinct promoter sites and synthetic oligonucleotides (26).We compared the effect of overexpressing a wild-type (wt)c-REL and a N-terminal RxxRxR to AxxAxA mutant reported to abrogate c-REL binding to classical REL/kB DNA motifs (27).Overexpression and TNF-a enhanced nuclear translocation of exogenous wt and AxxAxA mutant c-REL (Fig. 3B).Interestingly, both wt and AxxAxA mutant c-REL reduced total nuclear TAp73 and DNp63-TAp73 interaction, and this reduction was preferentially enhanced with TNF-a-stimulated AxxAxA mutant c-REL (Fig. 3C and D).These results suggested that TNF-a-induced To explore whether TNF-a modulates interactions between c-REL, DNp63a, and TAp73 on p63/p73 regulatory sites, we examined their binding to an oligonucleotide encoding a wellcharacterized p63/p73-binding site (À2,283 to À2,273; Supplementary Fig. S2A) from the promoter of the p63/p73-regulated growth arrest gene p21WAF1 (15,23).Using this p63 sitespecific radiolabeled oligonucleotide probe, a major band was revealed by EMSA in human UMSCC-22A cells (Fig. 4A, left panel, MB).Binding specificity was confirmed by unlabeled probe competition, and the complex contained c-REL, DNp63a, or TAp73, shown by supershift with specific antibodies.Similar to our findings in nuclear extracts, TNF-a stimulated c-REL binding, while decreasing binding of TAp73, without significantly affecting DNp63a.
Next, we examined the endogenous binding of c-REL, DNp63, and TAp73 to p21WAF1 promoter DNA using ChIP assays.All 3 factors exhibited basal binding to the p63-binding site region of the p21WAF1 promoter in UM-SCC 22A, -46, or -1 cells (Fig. 4A, right panel; Supplementary Fig. S2B-D).TNF-a induced c-REL binding and decreased TAp73 binding, without significantly affecting DNp63a binding to the p63 site in 2 different cell lines (Fig. 4B; Supplementary Fig. S2C).Similarly, c-REL overexpression also induced the same effects in either UMSCC-22A or -1 lines (Fig. 4A, right panel, Supplementary Fig. S2D).Conversely, c-REL depletion by siRNA significantly decreased c-REL and DNp63 binding, and reciprocally enhanced TAp73-binding activity in UM-SCC-22A cells (Supplementary Fig. S3E).Importantly, DNp63a siRNA knockdown attenuated basal and TNF-a inducible binding of c-REL, DNp63a, and TAp73, implicating DNp63a as a key anchor for c-REL-and TAp73-binding activity (Fig. 4C).Together, these findings in 3 different UMSCC lines indicate that TNF-a or genetic modulation of c-REL results in reciprocal modulation of TAp73 binding, with DNp63a on the established p63 regulatory site of the p21WAF1 promoter.
To examine the broader role of TNF-a-induced c-REL, DNp63, and TAp73 interactions, we carried out ChIP assays for p63-binding sites identified by bioinformatic analysis (28; Supplementary Methods) on the promoters of 2 key proapoptotic genes, NOXA and PUMA (Supplementary Fig. S2A, arrows).As with p21WAF1, stimulation with TNF-a increased c-REL and decreased TAp73 binding without affecting DNp63 bound to the predicted p63 sites in the PUMA and NOXA promoters (Fig. 4D).No significant binding by RELA(p65) or TP53 was detected on the p63 promoter sites of any of these 3 genes (Fig. 4A, right panel; D), consistent with co-IP results above.Together, our results indicate that TNF-a or genetically modulated c-REL reciprocally affects binding of DNp63a with TAp73 on p63 sites of multiple gene promoters.

c-REL differentially modulates p21WAF1, NOXA, and PUMA gene expression in UMSCC cell lines differing in expression of TP53 family members
We next examined how c-REL-modulated DNp63a-TAp73 interactions affect expression of growth arrest mediator p21WAF1.The UM-SCC cell subsets express lower or higher baseline levels p21WAF1 mRNA and protein (Supplementary Fig. S3A and B), similar to relative differences in expression of wt or mt TP53, DNp63a, and TAp73 shown above (Supplementary Fig. S1A-D).Knockdown of endogenous c-REL by siRNA significantly reduced levels of c-REL mRNA in 2 cell lines, UMSCC-1 and UMSCC-22A (Fig. 5A, left panel), from those subsets that differ in baseline expression of p21WAF1 and TP53 family proteins.However, c-REL depletion in UMSCC-22A cells with abundant mtTP53, DNp63, and TAp73 resulted in a strong increase in p21WAF1 mRNA and protein  We further examined the function of c-REL in proliferation and survival of HNSCC cells in vitro.c-REL depletion with siRNA had no significant effect on proliferation of UMSCC-1 cells that express lower levels of wtTP53 and other family members, but inhibited proliferation and density of UMSCC-22A cells that exhibit high protein levels of DNp63a, TAp73 and mtTP53 and inducible p21WAF1 (Fig. 6A and B; Supplementary Fig. S4A).c-REL siRNA depletion also increased sub-G 0 DNA-positive dead cells in UMSCC-22A, but not in UMSCC-1 cells (Supplementary Fig. S4B), whereas c-REL overexpression reduced the sub-G 0 fraction in UMSCC-22A cells (Supplementary Fig. S4C).The results showing that c-REL promotes cell proliferation and survival are consistent with the effect of c-REL in repressing p21WAF1, NOXA, and PUMA gene expression (Fig. 5), which mediate growth arrest and apoptosis in HNSCC (29).
To further examine the effects of c-REL on the suppressor function of TAp73, UMSCC-1 lacking TAp73 was transfected with control vector, c-REL, TAp73, or a combination of c-REL and TAp73.c-REL enhanced and TAp73 inhibited proliferation, whereas cotransfection of c-REL with TAp73 abrogated the inhibitory function of TAp73 (Fig. 6C).As HNSCC are relatively resistant to TNF-a-mediated inhibition of growth and survival (8), we examined the functional effect of TNF-a treatment or c-REL transfection, alone or in combination, on UM-SCC-22A cells (Fig. 6D).TNF-a alone weakly inhibited and c-REL enhanced proliferation, but the combination of TNF-a and c-REL overexpression further enhanced proliferation.Together, these data support the role of endogenous or exogenously overexpressed c-REL in repressing the antiproliferative function of TNF-a and TAp73.To further confirm the biological significance in human HNSCC, we compared the expression and cellular distribution of these proteins in human HNSCC and squamous epithelia.Human HNSCC tumors displayed increased (2þ to 3þ) nuclear costaining for c-REL (24/24; 100%) and DNp63a (20/24; 83%) and decreased TAp73 staining distributed in the cytoplasm and nucleus throughout the malignant epithelia, when compared with a basilar pattern of nuclear staining for these proteins in nonmalignant squamous mucosae (5/5; 100%; Fig. 7A; Supplementary Fig. S5A and B).Interestingly, greater nuclear c-REL and cytoplasmic TAp73 were detected in HNSCC tumors than in unstimulated UMSCC cultured in vitro (Supplementary Fig. S1A), consistent with overexpression of inflammatory factors such as TNFa, previously shown in the HNSCC tumor environment (7).
We also examined a recently developed transgenic mouse model in which overexpression of DNp63a targeted to skin under a tet-K5 promoter results in inflammation, overexpression of TNF-a, other proinflammatory cytokines, and squamous hyperplasia (28, 30, J.Du and Z. Chen, unpublished data).By 1 month, these mice developed cutaneous hyperplasia, with  increased suprabasilar nuclear c-Rel and DNp63a, and both cytoplasmic and nuclear staining for TAp73 in squamous epithelia (Fig. 7B).Thus, overexpression of DNp63a results in suprabasilar enhancement of nuclear c-REL and DNp63a, and cytoplasmic/nuclear distribution of TAp73, similar to that of orthologous proteins throughout malignant human squamous epithelia.

Discussion
Here we unveil evidence for a new paradigm (Fig. 7C), whereby an inflammatory factor, TNF-a, can promote nuclear interaction of c-REL with DNp63a, and TAp73 translocation to the cytoplasm, inhibiting the compensatory ability of TAp73 to activate key genes that mediate growth arrest and apoptosis in HNSCC with mutant TP53.Our findings establish a novel link between TNF-a expression and resistance (6)(7)(8)10), nuclear activation of proto-oncogene c-REL (13)(14)(15), and dysregulation of p63/p73 family transcription factors (4,5,23,28,29), which have been independently implicated in cancer progression.TNF-a-induced c-REL/DNp63a interactions and TAp73 dissociation were shown at unique p63 DNA regulatory sites within the promoters of important growth arrest and proapoptotic genes.The role of TNF-a induced c-REL in these specific interactions with DNp63a/TAp73 did not seem to involve RELA, nor c-REL transactivation domain residues involved in binding to kB enhancers shared with other NF-kB/REL family members.These findings suggest that TNFa-induced c-REL has a distinct function in inactivating TAp73 in p63/p73 promoter-regulated proapoptotic gene expression, complementing that previously shown for RELA in promoting kB-regulated prosurvival genes important in the malignant  A novel finding of this study is the demonstration of the role of TNF-a in coordinating interactions between nuclear c-REL/ DNp63a and reciprocal dissociation of DNp63a/TAp73 complexes, linking these previously reported components and interactions into a common dynamic mechanism.Previously, we discovered nuclear interactions between murine c-Rel and overexpressed DNp63a in keratinocytes and human c-REL and DNp63a in HNSCC (15).In murine keratinocytes, c-Rel and DNp63a were shown to prevent growth arrest, suggesting that corresponding c-REL/DNp63a complexes in HNSCC may also contribute to loss of growth control and promote the malignant phenotype.Independently, overexpressed DNp63a was shown to interact with TAp73 and promote survival in HNSCC (29).However, the relationship between these components, factor(s) modulating them, and basis for TAp73 inactivation were unknown.Here, we found that either c-REL or TAp73 interact with DNp63a, but we detected minimal interaction between c-REL and TAp73, suggesting that c-REL or TAp73 alternately complex with DNp63a.Supporting this model, we further established that TNF-a dynamically promotes nuclear c-REL/DNp63a interaction and reciprocal dissociation of DNp63a/TAp73.This is accompanied by nuclear-cytoplasmic translocation of TAp73, thus providing a basis for the inactivation of TAp73.TNF-a-induced c-REL was an important component of this mechanism, as c-REL siRNA attenuated dissociation of DNp63a/TAp73, although c-REL overexpression had a similar promoting effect on the dissociation and cytoplasmic translocation of TAp73.These effects of TNF-a or c-REL in promoting increased nuclear c-REL/DNp63a and cytoplasmic TAp73 in a subset of HNSCC lines may help explain the similar distribution we observed in HNSCC tumors, in which increased TNF-a expression as well as amplification and nuclear localization of c-REL have been detected (7,13,14).
These dynamic nuclear c-REL-DNp6a-TAp73 interactions, modulated by TNF-a or c-REL, were mirrored on established and novel predicted p63 regulatory sites in the promoters of several growth arrest and apoptotic genes.EMSA results indicated TNF-a induces reciprocal modulation between c-REL or TAp73 in binding to an established p63 site sequence of the p21WAF1 promoter (15,23).DNp63a binding remained relatively unchanged by TNF-a, but DNp63a knockdown inhibited ChIP binding of both c-REL and TAp73 to the p21WAF1 promoter, indicating that DNp63a bound to this p63 site is a key anchor for c-REL and TAp73 transcription factors.Although NOXA and PUMA were also previously identified as p63/p73 target genes (29), specific binding sites for p63/p73 were not previously resolved.Using a bioinformatics approach, we predicted p63-binding sites in these genes.We observed similar specificity in binding and modulation by TNFa of c-REL and TAp73 cobound with DNp63a on p21WAF1, NOXA, and PUMA promoters by ChIP assay.We have recently shown the capability of TNF-a to modulate c-REL/DNp63a interaction on p63 sites of additional genes by ChIP and EMSA Together, these findings suggest that the dynamic modulation of nuclear interactions involving these transcription factors observed in co-IP analysis are likely related to their specific interactions on p63/p73 sites of multiple gene promoters.
Our results further revealed the important and reversible function of nuclear c-REL in attenuating the compensatory ability of TAp73 to promote expression of these key growth arrest and apoptotic genes.Knockdown of c-REL potentiated the expression of p21WAF1, NOXA, and PUMA, further supporting the biological and potential therapeutic relevance of c-REL-DNp63a-TAp73 interactions observed on their promoters.Moreover, the modulation of these growth arrest and apoptotic genes was specifically observed in UMSCC-22A from a subset overexpressing TAp73 and DNp63a with mtTP53.This effect was not seen in UMSCC-1 from a subset with attenuation of expression and function of both TAp73 and wtTP53, which we have shown can result from other therapeutically reversible mechanisms (17,32).Consistent with these findings, c-REL modulation affected proliferation and apoptosis in UMSCC-22A but not UMSCC-1.Conversely, overexpressed c-REL inhibited the expression of p21WAF1 and antiproliferative effects when TAp73 was reexpressed in TAp73-deficient UMSCC-1, supporting an important role for c-REL in inhibiting the compensatory ability of TAp73.Together, these results highlight the functional importance and potentially reversible nature of c-REL-mediated inhibition of TAp73 function in HNSCC overexpressing TAp73, DNp63a, and mtTP53.
Our findings in UMSCC lines are likely to be of broader relevance in HNSCC and other cancers.Increased nuclear c-REL/DNp63a and nuclear and cytoplasmic TAp73 was observed in a majority of HNSCC tumor specimens.Similarly, increased DNp63a and TAp73 was previously seen in an independent panel of HNSCC tumors and lines and linked with inactivation of TAp73 (5,29).Breast cancer specimens also show increased DNp63 and TAp73 immunostaining, and this is most often seen in specimens with mtTP53 status (33).Subsets exhibiting increased expression of DNp63 in HNSCC and breast cancer have also been reported to be more sensitive to the chemotherapy drug cisplatin (33,34), underscoring the potential clinical relevance of identifying and selecting agents active in these tumor subsets.
Inflammatory mediator TNF-a is identified as a key modulator of dynamic interactions of c-REL with DNp63a and inactivation of TAp73.This mechanism could therefore contribute to the acquired resistance and promoting effects of TNF-a produced by inflammatory cells during tumorigenesis and metastatic tumor progression of HNSCC and other cancers (6)(7)(8)10).Supporting this hypothesis, we show here that combining TNF-a with overexpression of c-REL not only negated the inhibitory effect of TNF-a but also enhanced proliferation over that observed with either alone.Furthermore, human HNSCC tumors and epithelia of K5-DNp63a transgenic mice that exhibit increased TNF-a expression, inflammation, and epithelial proliferation (7, 28, J. Du and Z. Chen, unpublished observations) showed greater nuclear c-REL/DNp63a and distribution of TAp73 between the nucleus and the cytoplasm.Previous findings support targeting TNF-a or the canonical signal pathway which activates c-REL, for prevention or therapy of SCC.Knockout of TNF-a or TNF receptor-1, or TNF-a inhibitors, have been shown to reduce chemical carcinogenesis and malignant progression of SCC of the skin and other epithelial cancers (6).We previously showed that blocking canonical pathway signaling increased TNF-a cytotoxicity in HNSCC in vitro and induced apoptosis and regression of established murine and human SCC in vivo (8).However, early-phase clinical trials with TNF-a or proteasome inhibitors, which inhibit canonical pathway activation, have shown limited potential to slow disease progression in patients with advanced cancers (6,13).
Consequently, other molecular requirements for coordinated modulation of c-REL, DNp63a, and TAp73 interactions on the promoters of target genes merit investigation as targets for therapy.Unique structural characteristics of c-REL have previously been implicated in TNF-a-induced, TBK and IKKepsilon phosphorylation, dimerization, DNA binding, or transactivation (35,36).One of our laboratories found that the DNp63 a-domain, critical for oligomerization (5), is necessary for interaction with c-Rel (15).Cisplatin-induced IKKa and b activation have been reported to promote degradation of DNp63a and stabilization of p73, suggesting DNA damaging agents, and these kinases may be important modulators of these important components of the mechanism described herein (37,38)

Figure 1 .
Figure 1.Interactions among c-REL, DNp63a, TAp73, and RELA(p65) in HNSCC with mtTP53.A, left, whole-cell lysates from UM-SCC 22A cells or, right, 46 cells, were immunoprecipitated, then immunoblotted with antibodies as indicated.B, whole-cell lysates of UM-SCC 22A were immunoprecipitated with c-REL or p65 antibodies, then immunoblotted with the antibodies as indicated.C, cytoplasmic and nuclear fractions were immunoprecipitated with IgG control or c-REL antibody, then probed for c-REL, DNp63a, TAp73, and TP53.D, whole-cell lysates of UM-SCC 22A were immunoprecipitated with c-REL, p63 and TAp73 antibodies, and immunoprecipitated proteins and total lysates (input) were then immunoblotted with p63 antibody.IB, immunoblotting; IP, immunoprecipitation.

Figure 2 .
Figure 2. TNF-a-induced and overexpressed c-REL interacts with DNp63a and modulates TAp73.A, left, distribution of c-REL detected by IF staining in nuclear and cytoplasmic cellular counterparts, nontreated (NT), or with 20 ng/mL TNF-a for 30 minutes (original magnification, 200Â).Right, nuclear c-REL, DNp63a, and TAp73 proteins were detected before and after 20 ng/mL TNF-a treatment for 30 and 60 minutes.OCT-1: loading control.B, increased nuclear interaction between c-REL and DNp63a, and decreased nuclear interaction between TAp73 and DNp63a were observed 30 and 60 minutes after TNF-a treatment.C, decreased nuclear and increased cytoplasmic TAp73 were observed 30 and 60 minutes after TNF-a treatment.D, knockdown of c-REL increased the interaction between TAp73 and DNp63a without and with TNF-a stimulation.IB, immunoblotting; IP, immunoprecipitation.

Figure 3 .
Figure 3. Overexpressed c-REL-and c-REL kB-binding mutant interacts with DNp63a and modulates TAp73.A, after transfection with increasing amount of the expression plasmid for Flag-c-REL, interaction between exogenous c-REL and endogenous DNp63a was increased (top), and interaction between endogenous DNp63a and TAp73 was decreased, with redistribution of TAp73 from the nucleus to the cytoplasm (bottom).B, detection of endogenous c-REL 48 hours after transfection with empty pCDNA3 vector, recombinant Flag-c-REL wild-type, and Flag c-REL with AxxAxA mutations of the RxxRxR N-terminal kB DNA-binding residues, in cytoplasmic and nuclear fractions (CF, NF).b-Actin and OCT-1 were used as loading controls.C, nuclear TAp73 expression detected 48 hours after pCDNA3 and c-REL plasmid transfection AE TNF-a treatment (20 ng/mL, 60 minutes).D, overexpression of recombinant c-REL wild-type and c-REL AxxAxA mutant enhanced interaction between c-REL and DNp63a, and decreased interaction between DNp63a and TAp73.TNF-a treatment enhanced the effect of wild-type and AxxAxA mutant c-REL in decreasing DNp63a-TAp73 interactions.IP, immunoprecipitation.

Figure 4 .
Figure 4. c-REL, DNp63a, and TAp73 complex binding to the p21WAF1, NOXA, and PUMA gene promoters.A, left, EMSA with nuclear extracts from UM-SCC 22A cells showing protein bound to oligonucleotide containing the established p63-binding site sequence (À2,283 bp) for the p21WAF1 promoter.MB: Main band.Cold: unlabeled oligonucleotide competition (200Â); arrows, protein supershifts using antibodies against DNp63a, TAp73, or c-REL.c-REL binding gradually increased whereas TAp73 binding decreased after TNF-a treatment (20 ng/mL for 30 and 60 minutes).Probe only (no extract), indicates the antibodies lack inherent probe shifting activity.A, right, ChIP assay indicates c-REL, DNp63a, TP53, and TAp73 bound to (À2,283 bp) p63 site of p21WAF gene promoter, whereas p65 showed no significant binding activity.TNF-a (20 ng/mL for 30 minutes) increased the binding of c-REL but decreased the binding of TAp73.B, overexpression of c-REL increased c-REL but decreased TAp73 binding to the p21WAF1 promoter.C, DNp63 siRNA decreased the binding activities of c-REL, DNp63a, and TAp73.D, left, PUMA and right, NOXA p63 site ChIP binding for antibodies indicated, AE TNF-a 20 ng/mL for 30 minutes.Ã , P < 0.05.

Figure 5 .
Figure 5. c-REL modulates p21WAF1, PUMA, and NOXA expression in HNSCC lines.A, left, c-REL siRNA significantly decreased c-REL mRNA expression in UM-SCC 1 (deficient wtTP53) and 22A (mtTP53) cells.A, right, significant induction of p21WAF1 mRNA was detected only in UM-SCC 22A cells.B, Western blots, left, c-REL protein expression decreased after c-REL siRNA knockdown in UM-SCC 1 and 22A cells, but p21WAF1 protein increased only in UM-SCC 22A.Right, transfection with plasmid to overexpress c-REL decreased p21WAF1 protein expression in UM-SCC 22A cells.C, left, PUMA and right, NOXA, mRNA expression after c-REL siRNA depletion in UM-SCC 1 and 22A cells, with significant induction of PUMA and NOXA in UM-SCC 22A cells.The mean AE SD is shown.Ã , P < 0.05.
ctrl phenotype(10,31).Furthermore, our findings suggest that targeting TNF-a signaling, c-REL, or other regulators of these c-REL-DNp63a-TAp73 interactions could enhance TAp73 function, potentially helping in prevention or treatment of cancers with altered TP53 status.

Figure 7 .
Figure 7. Immunohistochemical staining and protein expression of TP53, p63, p73, and c-REL in HNSCC and DNp63a transgenic mice.Proposed model for interaction of c-REL, DNp63a, and TAp73.A, immunostaining for c-REL, DNp63, p73, and TP53 in representative matched squamous mucosa and HNSCC tumor specimens.Pancytokeratin staining was used as a positive control to highlight epithelia.Original, 400Â.B, dorsal skin sections from adult transgenic (TG) K5-DNp63a mice were stained with specific antibodies for c-REL, DNp63a, and TAp73 by immunofluorescence as indicated.Ki67, marker of proliferation and keratin 14, marker of epidermal differentiation.DAPI, 4 0 ,6diamidino-2-phenylindole.Bars, 37.5 mmol/L.C, model of TNFa-modulated c-REL, DNp63a, and TAp73 interactions in HNSCC.In response to TNF-a, c-REL undergoes nuclear translocation, binding with DNp63a, whereas TAp73 dissociates from the promoters of growth arrest and apoptotic genes, and the nucleus, enhancing cell proliferation and survival.c-REL-mediated effects are pronounced in the subset of the HNSCC with increased expression of c-REL, DNp63a, and TAp73 with mtTP53.