Relb Inhibits Cell Proliferation and Tumor Growth Through P53 Transcriptional Activation

ORIGINAL ARTICLE RelB inhibits cell proliferation and tumor growth through p53 transcriptional activation

E Jacque1,2,3,4, K Billot1,2,3,5, H Authier1,2,3,5,6, D Bordereaux1,2,3 and V Baud1,2,3

The alternative nuclear factor-kappaB (NF-kB) -activation pathway proceeds via inducible p100 processing, leading to the activation of RelB-containing dimers. This pathway is aberrantly activated in several types of tumors; however, a direct role for RelB in the control of cell proliferation is still largely unexplored. Here, we demonstrate that RelB provides cell proliferation-inhibitory signals in murine fibroblasts. In agreement with these results, RelB ectopic expression inhibits xenograft tumor growth in vivo, whereas RelB knockdown enhances it. Significantly, we show that RelB inhibits cell proliferation and tumor growth in a p53-dependent manner. Mechanistic studies indicate that RelB regulates the transcription of the p53 tumor-suppressor gene through direct recruitment to the p53 promoter, thus increasing both p53 protein levels and expression of p53 target genes such as p21. Our findings define a novel link between NF-kB and growth-inhibitory pathways involving the RelB-dependent transcriptional upregulation of p53. Furthermore, they suggest that inhibition of RelB in some tumor types that retain wild-type p53 may diminish rather than improve therapeutic responses.

Oncogene (2013) 32, 2661–2669; doi:10.1038/onc.2012.282; published online 9 July 2012 Keywords: RelB; NF-kB; cell proliferation; p53

INTRODUCTION cell function and secondary lymphoid organ organization.18,19 In mammals, the nuclear factor-kappaB (NF-kB) transcription RelB À / À mice also spontaneously develop a multiorgan factor family is composed of five members, RelA (p65), RelB, cRel inflammatory syndrome that contributes to their premature (Rel), NF-kB1 (p50 and its precursor p105) and NF-kB2 (p52 and its mortality,19 preventing long-term studies on cancer precursor p100), and forms a collection of various homodimeric development. Thus far, the majority of evidence indicating and heterodimeric complexes.1,2 NF-kB activation is generally growth-regulatory roles for RelB comes from gain- or loss-of- associated with cell survival and proliferation through the function mutations of molecules important for its activation in induction of anti-apoptotic and pro-proliferative proteins.3,4 B-cell lymphoma cell lines and patient samples.20,21 In contrast, Aberrant NF-kB activity has been observed in many cancers, although there are indications suggesting that RelB may including both solid and hematopoietic malignancies, and participate in tumor incidence and invasive phenotype,22,23 how sustained activation of NF-kB can affect all six hallmarks of RelB affects cell growth and tumor promotion in non-lymphoid cancer including insensitivity to growth-inhibitory signals, evasion cells is still debatable. of apoptosis, induction of angiogenesis and metastasis,5–7 leading The p53 tumor suppressor is frequently mutated in human to the concept of NF-kB addiction in cancer cells. However, this cancers. Like NF-kB, p53 has also been extensively implicated in general view is increasingly challenged by recent evidence the regulation of cell growth and involved in the pathogenesis of indicating that NF-kB can, under some circumstances, mediate most, if not all, cancers.24,25 In its wild-type (WT) form, p53 exerts pro-apoptotic effect, growth arrest and inhibition of cancer potent anti-tumor effects through the regulation of genes development.8–13 Therefore, an understanding of how NF-kB involved in growth arrest or induction of apoptosis.24,25 The arbitrates these opposing outcomes has to be addressed before crosstalk between NF-kB and p53 is becoming increasingly NF-kB-based therapy can become successful. Multiple appreciated as an important mechanism that influences the mechanisms have been delineated to achieve the required intricate regulation of cell proliferation and survival of both normal specificity and selectivity of the NF-kB response, including and transformed cells,17,26,27 but the potential participation of RelB regulation of specific target genes by individual NF-kBsubunits14 remains unexplored. and crosstalk between NF-kB and other transcription factors, such In the study presented here, we reveal a role for RelB in limiting as p53,2,15–17 but their influence on cell fate remains poorly cell proliferation and tumor growth. We find that the p53 tumor- understood. suppressor gene is a direct target of RelB in fibroblasts, and show Analysis of RelB-deficient mice has shown that RelB is essential that RelB exerts its anti-proliferative and anti-tumoral activities by for the development of medullary epithelium, mature dendritic functional activation of p53.

1De´partement d’Immunologie-He´matologie, Institut Cochin- INSERM U1016, Paris, France; 2Cnrs, UMR 8104, Paris, France and 3Universite´ Paris Descartes, Sorbonne Paris Cite´, Paris, France. Correspondence: Dr V Baud, De´partement d’Immunologie-He´matologie, Institut Cochin- INSERM U1016, Maternite´ Port-Royal, 123 Boulevard de Port-Royal, 75014, Paris, France. E-mail: [email protected] 4Current address: Division of Immune Cell Biology, National Institute for Medical Research, London, UK. 5These authors contributed equally to this work. 6Current address: IRD UMR152, Universite´ Paul Sabatier Toulouse III, Equipe Polarization des Macrophages et Rećepteurs nucleáires dans les pathologies Inflammatoires et Infectieuses, Toulouse, France. Received 20 January 2012; revised 30 April 2012; accepted 19 May 2012; published online 9 July 2012 RelB inhibits cell proliferation through p53 E Jacque et al 2662 RESULTS lentivirus carrying RelB and analyzed cell growth. Ectopic RelB slows down cell proliferation in murine fibroblasts expression of RelB in WT MEFs resulted in a marked reduction Although, as we previously reported, there is a significant basal of cell growth (Figure 1b). As we have previously reported a cross- level of RelB in the nucleus of primary murine embryonic regulation between RelA and RelB,28 we also examined whether fibroblasts (MEFs),28 we wondered whether RelB might exert a the increase in cell proliferation upon RelB knockdown observed biological activity in the absence of cell stimulation. To address in WT MEFs can still be seen in RelA-deficient MEFs. We found that the effect of RelB on cell growth in primary MEFs, we first used RelA is not required for RelB to exert its anti-proliferative activity in RNA interference (RNAi) to downregulate RelB levels, and analyzed fibroblasts (Supplementary Figure 1A). In addition, p100 proces- the effect of RelB knockdown on three independent WT MEFs sing into p52 being a key step in the control of RelB activation,29 from different mice. As shown in Figure 1a, two different siRNAs we also analyzed cell growth in NF-kB2-deficient MEFs upon (siRNA RelB no. 1 and siRNA RelB no. 2) directed against RelB RelB knockdown, and showed that p100/p52 are not required for efficiently repressed RelB protein levels; and most importantly, the RelB to slow down cell proliferation in murine fibroblasts growth rate of all three WT MEFs was markedly increased by RelB (Supplementary Figure 1B). knockdown compared with that seen with the scrambled control To gain further insight into the mechanisms through which RelB siRNA. To extend the above observations, we next infected three slows down cell growth in fibroblasts, we next examined the effect independent WT MEFs with either an empty lentivirus or a of RelB on cell cycle progression. We first infected three

siRNA controlRelB #1RelB #2 siRNAcontrolRelB #1RelB #2 siRNA controlRelB #1RelB #2 RelB RelB RelB

-actin -actin -actin

14 12 12 siRNA control siRNA control siRNA control 12 siRNA RelB #1 10 siRNA RelB #1 10 siRNA RelB #1 10 siRNA RelB #2 8 siRNA RelB #2 8 siRNA RelB #2 8 6 6 6 4 4 4 Fold increase

2 Fold increase 2 Fold increase 2 0 0 0 0 1 2 3 0 1 2 3 0 1 2 3 Days Days Days

LentiviruscontrolRelB LentiviruscontrolRelB LentiviruscontrolRelB

RelB RelB RelB -actin -actin -actin 10 10 8 control control control RelB 8 RelB 8 RelB 6 6 6 4 4 4 2

2 Fold increase 2 Fold increase Fold increase 0 0 0 0 1 2 3 0 1 2 3 0123 Days Days Days

100 100 80 ** 80 80 60 60 60 * G2-M 40 40 40 S G0-G1 20

20 20 G1 phase (%)

cell cycle phases (%) cell cycle 0 0 0 hours 016 016 016 016 control RelB control RelB RelB RelB control control shRNA shRNA Figure 1. RelB inhibits murine fibroblast proliferation. (a) RelB knockdown by RNAi promotes cell proliferation in WT MEFs. Three independent WT MEFs from different WT mice were transfected with either a scrambled control sequence (siRNA control) or two different siRNA oligonucleotides targeting RelB (siRNA RelB no. 1 or siRNA RelB no. 2). Twenty-four hours after transfection, cell proliferation was assessed by MTS assay. Data are shown as mean±s.e.m. of three independent experiments for each of the three individual WT MEFs. Efficient knockdown of RelB protein levels was analyzed by immunoblotting (upper panels). (b) Ectopic expression of RelB inhibits cell proliferation in MEFs. Three independent WT MEFs from different WT mice were transduced with lentiviruses encoding either the parental empty vector (control) or RelB, and analyzed for cell proliferation by MTS assays as in (a). Efficient ectopic expression of RelB was analyzed by immunoblotting (upper panels). (c and d) RelB slows down G1-S phase progression. Three independent WT MEFs transduced with lentiviruses encoding either the parental empty vector (control) or RelB (left), or transduced with lentiviruses encoding either a shRNA targeting RelB or a scrambled control (right) were synchronized and followed by cell cycle analysis as described in Materials and methods. The fraction of cells in G1, S and G2/M phases of one representative experiment is shown in (c), and the percentage (means±s.e.m. of two independent experiments using three independent WT MEFs) of cells in G1 phase is represented in (d). Asterisks indicate significant effects as compared to control cells, unpaired t-test, *Po0.05, **Po0.01.

Oncogene (2013) 2661 – 2669 & 2013 Macmillan Publishers Limited RelB inhibits cell proliferation through p53 E Jacque et al 2663 independent WT MEFs with either an empty lentivirus or a 16 lentivirus carrying RelB, synchronized the cells by serum depriva- shRNA control 14 tion and analyzed cell cycle progression upon serum exposure. shRNA RelB 12 Ectopic expression of RelB resulted in a larger G1 population after 10 LentivirusshRNA controlshRNA RelB 16 h of serum stimulation (mean: 39 versus 64%) (Figure 1c, left 8 and Figure 1d). Reciprocally, a reduced proportion of cells in G1 RelB 6 phase was observed upon RelB knockdown in WT MEFs (mean: 41 -actin

Fold increase 4 versus 29%) (Figure 1c, right, and Figure 1d). These differences 2 were statistically significant (P ¼ 0.0179 and P ¼ 0.0334, respec- 0 tively). Importantly, apoptosis was not observed upon either RelB 1 2 3 knockdown or ectopic expression, where no apoptotic peaks were Days detected, a finding that was further confirmed by annexin V/ propidium iodide staining (Supplementary Figure 2). Collectively, 2000

) shRNA control our data indicate that RelB acts as a negative modulator of cell 3 proliferation in ﬁbroblasts and has a role in inducing a delay in the 1500 shRNA RelB G1/S-phase transition. 1000 RelB limits tumor growth in vivo As RelB can inhibit cell proliferation in WT MEFs, we speculated 500

that RelB might also affect cell growth in transformed murine (mm volume Tumor ﬁbroblasts and consequently tumor growth in vivo. To test this 0 hypothesis, we used NIH-3T3 cells transformed by the RET 5678910111213 oncoprotein (NIH3T3/RET), described as forming tumors within Days days post-injection into immunodeﬁcient mice,30 infected these cells with either a lentivirus carrying a shRNA targeting RelB or a 2500

) control scrambled control and first analyzed cell growth. RelB knockdown 3 2000 RelB in NIH3T3/RET cells resulted in a marked increase of cell growth (Figure 2a), just as what was seen in primary MEFs. We next asked 1500 LentiviruscontrolRelB whether the growth-suppressive effect of RelB could consequently RelB be seen on tumor growth in vivo. To test this hypothesis, we 1000 injected the NIH3T3/RET infected with a shRNA targeting RelB or a 500 -actin scrambled control into the flank of immunodeficient NOD-SCID Tumor volume (mm volume Tumor mice, and followed tumor development over time. As shown in 0 Figure 2b, RelB knockdown significantly promoted tumor growth 567891011 12 13 in NOD-SCID mice. Conversely, RelB overexpression in NIH3T3/RET Days cells delayed tumor growth (Figure 2c). Taken together, these data Figure 2. RelB limits tumor growth in vivo.(a) RelB knockdown by indicate that RelB can inhibit cell growth in non-transformed and RNAi promotes transformed murine fibroblasts proliferation. NIH- transformed fibroblasts, and, as a result, limit tumor growth in vivo. 3T3/RET tumor cells were stably transduced with lentiviruses encoding either a shRNA targeting RelB (shRNA RelB) or a scrambled p53 is a RelB target gene in murine fibroblasts control (shRNA control) and analyzed for cell proliferation by MTS assays. Data are shown as mean±s.e.m. of three independent Given that p53 is a first-line tumor suppressor associated with cell 24,25 experiments. Efficient knockdown of RelB protein levels was growth arrest, we speculated that RelB might control p53 analyzed by immunoblotting (right panel). (b) RelB knockdown by activation in fibroblasts. To test this hypothesis, we asked whether RNAi promotes tumor growth. 106 NIH-3T3/RET tumor cells stably the forced expression of RelB in WT MEFs leads to the transduced with lentiviruses encoding either a shRNA targeting RelB accumulation of p53 protein. As shown in Figure 3a, the amount (shRNA RelB) or a scrambled control (shRNA control) were injected of p53 protein was markedly and significantly elevated in WT subcutaneously into immunodeficient NOD-SCID mice. Data repre- MEFs infected with a lentivirus carrying RelB compared with what sent means tumor volume±s.e.m. (n ¼ 13 mice per group) from is seen in the empty virus-infected cells. Reciprocally, RelB three independent experiments. (c) Reduced tumor growth upon RelB ectopic expression. NIH-3T3/RET tumor cells (106 cells per mice) knockdown by RNAi in WT MEFs decreased p53 protein expression stably transduced with lentiviruses encoding either the parental levels (Figure 3b). Surprisingly, the cells exhibited no significant empty vector (control) or RelB were injected into NOD-SCID mice differences in the half-life of p53 protein and no interaction and monitored for tumor growth as in (b). between RelB and p53 was observed, suggesting that RelB might instead regulate basal p53 expression at the transcriptional level in fibroblasts. To test this possibility, using quantitative RT–PCR, we first compared p53 mRNA levels in WT MEFs infected with either stimulation induced efficient recruitment of RelA (Figure 3e). an empty lentivirus or a lentivirus carrying RelB. Ectopic expression Importantly, RelB recruitment was further increased upon RelB of RelB in WT MEFs resulted in an increase in p53 mRNA levels overexpression in WT MEFs (Figure 3f, Po0.0001). Taken together, (Po0.0001). In contrast, as a control, no significant effects on RelA these results indicate that RelB regulates p53 expression through expression were seen (Figure 3c). Reciprocally, RelB knockdown direct transcriptional control in fibroblasts. significantly decreased p53 mRNA levels (Figure 3d, P ¼ 0.0004). Further, in vivo evidence for a direct role for RelB in regulating p53 transcription in WT MEFs was obtained by chromatin immuno- RelB regulates constitutive and inducible p53 activity precipitation (ChIP) analysis. As shown in Figure 3e, RelB was Beyond controlling the transcription of the p53 tumor suppressor found to be constitutively associated to the p53 promoter through direct recruitment to the p53 promoter, we investigated (Po0.0001). In contrast, no significant binding of RelA was whether modulation of RelB levels also affected p53 activity. As observed. As controls neither RelB nor RelA were found to several reports have indicated that p53 contributes towards basal constitutively bind to the RANTES promoter, although TNFa expression of many of its target genes in untreated cells,31,32 we

& 2013 Macmillan Publishers Limited Oncogene (2013) 2661 – 2669 RelB inhibits cell proliferation through p53 E Jacque et al 2664 first examined by quantitative RT–PCR the effect of RelB knock- in WT MEFs (Figure 4e). A reduction in p53 serine 15 down on the mRNA levels of p21, Mdm2 and Gadd45b, which are phosphorylation was also observed. Remarkably, loss of RelB known direct p53 transcriptional targets 24 in WT MEFs. As shown resulted in reduced expression of p21 and Mdm2 in response to in Figure 4a, depletion of RelB resulted in a significant reduction in doxorubicin both at the mRNA (Figure 4f, upper panel) and the expression of all three genes (P ¼ 0.0004, P ¼ 0.0008 and protein levels (Figure 4f, lower panel), thus indicating that RelB P ¼ 0.0230, respectively). To extend the above observations, we also contributes to p53 activation in response to DNA damage. In also analyzed the mRNA expression levels of these genes upon addition, RelB knockdown impaired p53 upregulation upon TNFR RelB overexpression in WT MEFs (Figure 4b). Reintroduction of and LTbR ligation (Supplementary Figure 3). Thus, basal and RelB induces an increase in p21, Mdm2 and Gadd45b mRNA levels induced p53 activation is under the control of RelB. (Po0.0001, P ¼ 0.0335 and P ¼ 0.0021, respectively). Importantly, Taken together, these results indicate that RelB is required for RelB was not seen to bind to the p21, Mdm2 and Gadd45b p53-dependent gene expression in fibroblasts. promoters (Figure 4c). To directly assess whether RelB regulation of p21 expression was p53 dependent, we compared p21 mRNA þ / þ RelB inhibits cell proliferation and tumor growth in a p53- levels upon RelB ectopic expression in MEF p53 and MEF dependent manner p53 À / À . Significantly, p21 induction induced by RelB over- We next examined whether the effect of RelB on p53 activation expression was lost in the absence of p53 (Figure 4d), confirming might be responsible for the anti-proliferative activity of RelB. To that RelB is a p53-dependent positive regulator of p21 expression. test this possibility, we infected MEF p53 þ / þ and MEF p53 À / À Next, we examined whether RelB also regulates inducible p53 with lentiviruses carrying either a siRNA oligonucleotide targeting function. The doxorubicin-induced increase in p53 protein was RelB or a scrambled control, and analyzed cell growth. Knockdown markedly reduced following shRNA-mediated knockdown of RelB of RelB in MEF p53 þ / þ cells increased cell proliferation to a level similar to what was seen in primary WT MEFs (Figure 5a, left versus Figure 1a). In contrast, no change in cell growth was observed 7 * when RelB expression was inhibited in MEF p53 À / À cells 6 (Figure 5a, center), indicating that RelB promotes p53-mediated Lentiviruscontrol RelB 5 cell growth arrest. Importantly, RelB knockdown in MEF p53 þ / þ p53 4

-actin resulted in a decrease in basal expression of p53 (Figure 5a, right). 3 RelB As p53 has signiﬁcant roles in the regulation of tumor

p53/ 24 2 progression, we speculated that RelB might exert its negative RelA 1 effect on tumor growth through its control of p53 activity. To test (Densitometric analysis) this hypothesis, we examined the effect of forced expression of -actin 0 Lentivirus control RelB p53 on the tumorigenesis seen when NIH-3T3/RET cells stably

*** 1.0

shRNA control RelB 0.8 Figure 3. p53 is a RelB target gene in ﬁbroblasts. (a) p53 protein level is increased upon ectopic RelB expression in MEFs. Whole-cell p53 0.6 extracts from three independent WT MEFs from different mice -actin

RelB ββ transduced with lentiviruses encoding either the parental empty 0.4 vector (control) or RelB were analyzed by immunoblotting for the RelA p53/ 0.2 indicated proteins. Data of one representative experiment are

β-actin (Densitometric analysis) shown (left), and densitometric analysis of p53 protein level 0.0 normalized to b-actin (means±s.e.m. of three independent experi- shRNAcontrol RelB ments using three independent WT MEFs) is reported (right). Asterisks indicate signiﬁcant effects as compared to control,

6 1.4 shRNA control *Po0.05. (b) Stable RelB knockdown decreased p53 protein control shRNA RelB expression levels. Whole-cell extracts from three independent WT 5 1.2 *** RelB *** MEFs transduced with lentiviruses encoding either a shRNA 1.0 4 targeting RelB or a scrambled control were analyzed as in (a), 0.8 ***P 0.001. (c and d) RelB controls constitutive p53 expression at 3 o 0.6 the mRNA level. Quantitative RT–PCR was performed with speciﬁc 2 0.4 primer pairs for the indicated genes using total RNAs prepared from three independent WT MEFs transduced with either (c) lentiviruses 1 0.2 Relative expression level expression Relative Relative expression level expression Relative encoding either the parental empty vector (control) or RelB, or else 0 0.0 (d) a shRNA targeting RelB or a scrambled control. Results are p53 RelA p53 RelA means±s.e.m. of three independent experiments using three independent WT MEFs normalized to the level HPRT mRNA, 80 12 *** ***Po0.001. (e) RelB is constitutively bound to the p53 promoter *** control 70 control in WT MEFs. Recruitment of RelA and RelB to the p53 and RANTES 10 α-RelB α-RelB promoters was examined in two independent WT MEFs by ChIP 60 α-RelA α-RelA experiments followed by quantitative PCR analysis. Two indepen- 8 50 dent WT MEFs were left untreated or treated with TNF-a for 1 h and recruitment of RelB to the p53 and Rantes promoters was examined 6 40 by ChIP experiments followed by quantitative PCR analysis. Results 30 are means±s.e.m. of three independent experiments, n ¼ 2, normal- 4 ized to inputs that reﬂect relative amount of sonicated DNA 20 fragments present before immunoprecipitation, ***Po0.001. 2 10 (f) Recruitment of RelB to the p53 promoter is increased upon RelB Recruitment (relative units/input) Recruitment (relative Recruitment (relative units/input) Recruitment (relative ectopic expression. Recruitment of RelB to the p53 and RANTES 0 0 promoters was examined in two independent WT MEFs stably α TNF- (hours) 0 101 Lentivirus control RelBcontrol RelB transduced with lentiviruses encoding either the parental empty Promoter p53 RANTES Promoter p53 RANTES vector (control) or RelB as described in (e), ***Po0.001.

Oncogene (2013) 2661 – 2669 & 2013 Macmillan Publishers Limited RelB inhibits cell proliferation through p53 E Jacque et al 2665 transduced with lentiviruses carrying either a shRNA targeting We have found that RelB slows down cell proliferation and RelB (shRNA RelB) or a scrambled control (shRNA control) were tumor growth in fibroblasts. It has been recently reported that injected into NOD-SCID mice. Overexpression of p53 counteracted RelB can promote the more invasive phenotype of ERa-negative the increase in tumor progression observed upon RelB breast cancer cell lines,22 and RelB knockdown reduces the size of knockdown, whereas it had no significant effect on the tumor primary prostate carcinomas in a mice xenograft tumor model,23 growth observed upon expression of the scrambled control while RelB can act as a negative regulator of cell survival in diffuse sequence (Figure 5b). Taken together, our results suggest that large B-cell lymphoma cell lines characterized by constitutive RelB promotes cell cycle arrest and inhibits tumor growth in a p53- Malt1 activity.39 Thus, whether RelB is a tumor suppressor or a dependent manner. tumor promoter is most likely context dependent, which is worth further investigation. It is firmly established that NF-kB activation is associated with cell proliferation through the induction of pro-proliferative DISCUSSION genes.1,6 More recently, the relevance of NF-kB for tumor In the present study, we uncovered a role for RelB in limiting cell maintenance, tumor development and even possibly tumor proliferation and tumor growth in vivo. On the other hand, in initiation is becoming evident.5,6,7 Thus, without calling into numerous experimental systems, unequivocal evidence indicates question this paradigm as a whole, our data support an alternative that NF-kB promotes cell growth and tumorigenesis.4–6 Thus, the mechanism: a likely role for RelB in limiting cell proliferation and data presented here are somewhat provocative with respect to tumor growth. A similar conclusion was made earlier from studies the generally accepted role of NF-kB in the development of of RelA-deficient epidermal cells, showing that NF-kB antagonized human cancers. Furthermore, our findings point to a mechanism proliferative signals derived from TNFR and JNK.40 In human involving direct transcriptional activation of the p53 tumor- keratinocytes, NF-kB triggers cell cycle arrest and blockade of suppressor gene by RelB. Although there is accumulating NF-kB generates malignant human epidermal tissue resembling evidence linking IKK/NF-kB to the p53 pathway,15–17,26,27 a direct squamous cell carcinoma (SCC).10 Similarly, loss of IKKa causes role for RelB as a regulator of p53 expression and activity was not striking epidermal hyperplasia in mice,41 and reduction in IKKa defined previously. expression promotes skin carcinogenesis in mice.42 Most ChIP experiments have shown that RelB is recruited to the importantly, loss of heterozygosity, as well as genetic and region that encompasses the NF-kB-binding sequence of the p53 epigenetic alterations of the IKKa locus have been reported in promoter in fibroblasts.33,34 These data are consistent with the SCC of the skin, esophagus, and head and neck in humans,43,44 suggested kB site serving as a functional NF-kB subunit-binding indicating a novel role of IKKa as a tumor suppressor in SCC. site.33 However, RelA recruitment to the p53 promoter was not Another striking example of the counterintuitive role of NF-kBin significantly seen under the same conditions, suggesting that limiting cell growth has been provided by analysis of IKKb- these effects might be specific to RelB in fibroblasts. As we have deficient MEFs,11 and a study using mice with a hepatocyte- shown that RelB exerts its anti-proliferative activity independently specific deletion in the IKKb gene demonstrated that hepatocyte- of RelA, our data suggest that the anti-proliferative and anti- specific IKKb deficiency promotes chemical carcinogen-induced tumoral activities of RelB are exerted through kB site- and DNA- hepatocarcinogenesis.12 Therefore, a careful re-evaluation is binding-dependent recruitment of RelB to the p53 promoter. certainly required to define the role of NF-kB in the various However, it would be simplistic to attribute all of these negative stages of malignant transformation and tumorigenesis, and effects on cell growth to the control of p53 expression and re-introducing the tumor suppressor form of NF-kB might have function, and RelB may additionally modulate cellular responses great potential for therapeutic intervention. via other target genes that remain to be identified. In conclusion, we established RelB as an important p53-positive In contrast to our findings, it has been recently reported that regulator that exerts its inhibitory role on cell growth and basal p53 expression is under direct RelA-dependent transcrip- tumorigenesis in a p53-dependent manner. Our data are of great tional control in MEFs.35 The reasons for these discrepant results functional importance because they constitute a significant are not clear but could be due to differences in conditions advance in the understanding on RelB physiological function employed for the ChIP analysis. Interestingly, it was shown that and shed light on a novel positive link between NF-kB and p53 is a crucial downstream mediator of RelA in the bioenergetics growth-inhibitory pathways. pathway controlling adaptation to metabolic stress.35 Whether RelB might be involved in the control of adaptation to metabolic stress by controlling p53 expression is worth further investigation. MATERIALS AND METHODS The cross-regulation between RelB and p53 that we report here Antibodies and reagents is likely to be part of a wider complex network intimately linking The antibodies were purchased from Santa Cruz Biotechnology, Santa p53 and the alternative NF-kB pathway. This pathway controls the Cruz, CA, USA (RelA C-20, RelB C-19, p53 FL-393) and Sigma-Aldrich, activation of RelB-containing dimers (that is, RelB-p52 and RelB- St Louis, MO, USA (b-actin, clone AC-15). Doxorubicin and murine TNF-a p50) and relies on the inducible activation of IKKa in response to a were purchased from Sigma-Aldrich, and agonistic anti-LTbR mAb AC.H6 specific subset of stimuli.29 For example, IKKa-mediated was a kind gift from J Browning. phosphorylation of the CBP transcriptional co-activator increases cell proliferation by inhibiting the binding of CBP to p53, while Plasmids 36 promoting the interaction with RelA. In several cell lines, p52 Expression vectors for RelB was obtained from M Ko¨rner (Institut Andre´ is directly recruited by p53 to multiple p53-regulated promoters Lwoff, Villejuif, France), p53 from K Vousden (the Beatson Institute for where it regulates co-activator or co-repressor recruitment, Cancer Research, Glasgow, UK), M Oren (the Weizmann Institute of Science, thus leading to either repression or activation of p53 target Rehovot, Israel) and D Reisman (University of South Carolina, Columbia, SC, genes associated with cycle arrest and cell death.37,38 However, USA), and pSuper vector-containing polymerase III H1 promoter was p52 has no significant effects on p53 expression levels,37 provided by T Tuschl (the Rockfeller University, New York, NY, USA). pTrip- shRNA RelB was generated by subcloning an oligonucleotide designed to indicating that the effects we have observed with RelB are target RelB under control of the H1 promoter into pTRIP-DU3-MND-GFP distinct from those described for p52 in human cancer cells. lentiviral vector.45 pTRIP–shRNA control was generated following the same Although establishment of the significance of these observations approach using a scrambled control oligonucleotide. RelB sense 5’- will require further investigation, they nevertheless underline the GATCCCCGGAAATCATCGACGAATACTTCAAGAGAGTATTCGTCGATGATTTCC complex relationships that exist between NF-kB and p53. TTTTTGAATTCA-30; RelB antisense 50-AGCTTGAATTCAAAAAGGAAATCATC

& 2013 Macmillan Publishers Limited Oncogene (2013) 2661 – 2669 RelB inhibits cell proliferation through p53 E Jacque et al 2666 GACGAATACTCTCTTGAAGTATTCGTCGATGATTTCCGGG-30; control sense Saclay, France), and NF-kB2 þ / þ and NF-kB2 À / À from F Weih and U 50-GATCCCCCGTACGCGGAATACTTCGATTCAAGAGATCGAAGTATTCCGCGTA Siebenlist (the National Institute of Allergy and Infectious Diseases, CGTTTTTGAATTCA-30; control antisense 50-AGCTTGAATTCAAAAACGTACG Bethesda, MD, USA). Cells were grown in Dulbecco’s modiﬁed Eagle CGGAATACTTCGATCTCTTGAATCGAAGTATTCCGCGTACGGGG-30. pTrip-RelB medium (Life Technologies, Grand Island, NY, USA) supplemented with and pTRIP p53 were generated by subcloning, into pTRIP-DU3-EF1a-IRES 10% heat-inactivated fetal bovine serum, 2 mML-glutamine, 1 mM sodium GFP lentiviral vector,45 full-length RelB and p53 cDNA, respectively. pyruvate and 100 units/ml penicillin–streptomycin.

Cell culture Lentiviral production and transduction RelB À / À mouse embryonic ﬁbroblasts (MEFs) were kind gifts from F Weih Production of infectious recombinant lentiviruses was performed by (Fritz Lipmann Institute, Jena, Germany), p53 þ / þ and p53 À / À MEFs from transient transfection of 293 T cells as described in Kieusseian et al.45 For T Jacks (MIT, Cambridge, USA), NIH-3T3/RET from B Tavitian (CEA Saclay, infections, 105 cells in 35 mm dishes were transduced with 5000 ng/ml of

1.4 shRNA control 2.4 control shRNA RelB RelB 1.2 2.0 *** *** * *** * ** 1.0 1.6 0.8 1.2 0.6 0.8 0.4 Relative expression level expression Relative Relative expression level expression Relative 0.2 0.4

0.0 0.0 p21 Mdm2 Gadd45β p21 Mdm2 Gadd45β

12 control 2.0 *** control α-RelB 10 1.8 RelB 1.6 8 1.4 1.2 6 1.0 0.8 4 0.6 0.4 2 level expression Relative

Recruitment (relative units/input) Recruitment (relative 0.2 0 0.0 β Promoter p21 Mdm2 Gadd45 p53+/+ p53-/-

shRNA control RelB Dox (hours) 0 4 8 0 4 8 shRNA control RelB p53 Dox (hours) 048048 Mdm2 PSer15-p53 p21 RelB

RelA β-actin β -actin shRNA control RelB

control Dox (hours) 048048 30 control ** 12 * RelB RelB 25 10 * Mdm2

** -actin

20 β 8 p21 -actin

β 15 6 10 4 β-actin p53/ 5 2

0 PSer15-p53/ 0 (Densitometric analysis) 0h 4h 8h (Densitometric analysis) 0h 4h 8h

Oncogene (2013) 2661 – 2669 & 2013 Macmillan Publishers Limited RelB inhibits cell proliferation through p53 E Jacque et al 2667 p24 (HIV-1 capsid protein). Fourty-eight hours later, cells were washed and mRNA. Primer sequences are as follow: p21 sense 50-TTCCTTGCCACTT fresh medium was added. The culture was then continued as described CTTACCTG-30; p21 antisense 50-TCCTAGCTGGCCTTAGAG-30; Mdm2 sense above. GFP-positive cells were sorted with Cytopea Inﬂux-500 sorter (BD 50-TTGATCCGAGCCTGGGTCTGTG-30; Mdm2 antisense 50-AAGATCCTGATG Biosciences, San Jose, CA, USA). CGAGGGCGTC-30; Gadd45b sense 50-TCTCTCTAGAGGAACGCTGAG-30; Gadd45b antisense 50-GGTATCACGGGTAGGGTA-30; p53 sense 50-CACC siRNA transfections TGCACAAGCGCCTCT-30; p53 antisense 50-TCATTCAGCTCCCGGAACATC-30; RelA sense 50-ATGATCGCCACCGGATTGAA-30; RelA antisense 50-ACAGAA siRNA transfections were performed as described previously.28 All siRNA GTTGAGTTTCGGGTAGGCA-30; HPRT sense 50-GCTGGTGAAAAGGACCTCT-30; duplex oligonucleotides targeting the following proteins were synthesized and HPRT antisense 50-CACAGGACTAGAACACCTGC-3. by Proligo: RelB no. 1 sense 50-GGAAAUCAUCGACGAAUACdTdT-30; RelB no. 2 sense 50-UCAUCACAGAGCAGCCAAAdTdT-30; scrambled control sense 50-CGUACGCGGAAUACUUCGAdTdT-3’. ChIP assays ChIP assays were performed as described.28 Samples were analyzed by Immunoblotting real-time PCR. Sequences of promoter-speciﬁc primers are as follow: p53 0 0 0 Immunoblotting were performed as described in Jacque et al.28 sense 5 -TTGCCCTTACTTGTTATG-3 ; p53 antisense 5 -CCCAATGAACTG AAGCTACAG-30; MCP1 sense TCCAATACTGCCTCAGAATGG; and MCP1 Real time RT–PCR antisense ATCACCATTGCAAAGTGAATTG. Total RNA extraction and RT were performed as previously described.28 Real-time PCR analysis was carried out with LightCycler FastStart DNA Cell proliferation assays Master plus SYBR Green I on a Light Cycler 1.5 (Roche Applied Science, Cells (1500/well) were seeded in 96-well plates and grown under the Indianapolis, IN, USA). All values were normalized to the level of HPRT culture conditions as described above. Cell viability was assessed every

12 MEF p53 +/+ 12 MEF p53 -/- 10 10 siRNA control siRNA control siRNA controlRelB controlRelB siRNA RelB 8 siRNA RelB 8 RelB 6 6 4 4 p53

Fold increase Fold 2 2 β-actin 0 0 0 1 2 3 0 1 2 3 MEF p53 +/+ p53 -/- Days Days

) 1500 shRNA control 3 shRNA control + p53 shRNA RelB 1000 shRNA RelB + p53

shRNAshRNA controlshRNA RelBshRNA control RelB + p53 + p53 500 RelB p53 Tumor volume (mm volume Tumor 0 5 6 7 8 9 10 11 12 13 β-actin Days Figure 5. RelB inhibits cell proliferation and tumor growth in a p53-dependent manner. (a) RelB anti-proliferative activity requires p53. MEF p53 þ / þ and MEF p53 À / À were transfected with either a siRNA oligonucleotide targeting RelB (siRNA RelB) or a scrambled control (siRNA control). Twenty-four hours after transfection, cell proliferation was assessed by MTS assay. Data are shown as mean±s.e.m. of three independent experiments. Protein expression levels were analyzed by immunoblotting for the indicated proteins (right panel). (b)106 NIH- 3T3/RET tumor cells stably transduced with the indicated lentiviruses were injected subcutaneously into NOD-SCID mice. Tumor size was followed over-time as described in Figure 2. Protein expression levels were analyzed by immunoblotting for the indicated proteins (right panel).

Figure 4. RelB regulates constitutive and inducible p53 activity. (a) RelB controls p53-dependent gene expression. Quantitative RT–PCR was performed with speciﬁc primer pairs for the indicated genes using total RNAs prepared from three independent WT MEFs stably transduced with lentiviruses encoding either a shRNA targeting RelB or a scrambled control. Results are means±s.e.m. of three independent experiments using three independent WT MEFs normalized to the level HPRT mRNA, *Po0.05, **Po0.01, ***Po0.001. (b) Quantitative RT–PCR was performed as in (a) using total RNAs prepared from three independent WT MEFs stably transduced with lentiviruses encoding either the parental empty vector (control) or RelB, *Po0.05, **Po0.01, ***Po0.001. (c) RelB does not constitutively bind to the p21, Mdm2 and Gadd45b promoters in WT MEFs. Recruitment of RelB to the p21, Mdm2 and Gadd45b promoters was examined in WT MEFs by ChIP experiments followed by quantitative PCR analysis. The results are means±s.e.m. of three independent experiments normalized to inputs that reﬂect relative amount of sonicated DNA fragments present before immunoprecipitation. (d) RelB regulation of p21 expression is p53 dependent. Total RNAs from MEF p53 þ / þ and MEF p53 À / À were analyzed by RT–PCR. Data are shown as mean±s.e.m. of three independent experiments, ***Po0.001. (d) p53 protein level is increased upon ectopic RelB expression in HCT116 cells. Whole-cell extracts from HCT116 p53 þ / þ cells transduced with lentiviruses encoding either the parental empty vector (control) or RelB were analyzed by immunoblotting for the indicated proteins. Data of one representative experiment are shown, and densitometric analysis of p53 protein level normalized to b- actin (means±s.e.m. of three independent experiments) is reported, *Po0.05. (e) RelB controls inducible p53 expression. Whole-cell extracts from three independent WT MEFs transduced with lentiviruses encoding either a shRNA targeting RelB or a scrambled control and treated with 0.5 mM of doxorubicin for the indicated periods of time were analyzed by immunoblotting for the indicated proteins. Data of one representative experiment are shown, and densitometric analysis of p53 and phospho-p53 Ser 15 protein level normalized to b-actin (means±s.e.m. of three independent experiments) is reported, *Po0.05, **Po0.01. (f) RelB controls inducible p53-dependent gene expression. Whole-cell extracts from WT MEFs transduced with lentiviruses encoding either a shRNA targeting RelB or a scrambled control and treated with doxorubicin for the indicated periods of time were analyzed by immunoblotting for the indicated proteins (lower panel), and total RNA from the same cells were analyzed by RT–PCR (upper panel). Data are representative of three independent WT MEFs.

& 2013 Macmillan Publishers Limited Oncogene (2013) 2661 – 2669 RelB inhibits cell proliferation through p53 E Jacque et al 2668 24 h over a period of 3 days using the Cell Titer 96 Aqueous 11 Chen F, Lu Y, Castranova V, Li Z, Karin M. Loss of Ikkbeta promotes migration and NonRadioactive Cell Proliferation Assay (Promega, Madison, WI, USA), proliferation of mouse embryo fibroblast cells. J Biol Chem 2006; 281: 37142– according to the manufacturer’s instructions. 37149. 12 Maeda S, Kamata H, Luo JL, Leffert H, Karin M. IKKbeta couples hepatocyte death Cell cycle analysis to cytokine-driven compensatory proliferation that promotes chemical hepato- Cells were seeded on 35 mm dishes at a density of 105 cells per well, grown carcinogenesis. Cell 2005; 121: 977–990. in DMEM 0.5% FBS for 12 h before being stimulated with 10% FBS for up to 13 Gapuzan ME, Yufit PV, Gilmore TD. Immortalized embryonic mouse fibroblasts 16 h. Progression through the cell cycle was monitored by flow cytometric lacking the RelA subunit of transcription factor NF-kappaB have a malignantly analysis of DNA content of cell populations stained with propidium iodide transformed phenotype. Oncogene 2002; 21: 2484–2492. using a FACSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ, 14 Saccani S, Pantano S, Natoli G. Modulation of NF-kappaB activity by exchange of USA). Quantification of cell cycle distribution was performed using dimers. Mol Cell 2003; 11: 1563–1574. CellQuest software (Becton Dickinson). 15 Perkins ND. Integrating cell-signalling pathways with NF-kappaB and IKK function. Nat Rev Mol Cell Biol 2007; 8: 49–62. 16 Ak P, Levine AJ. p53 and NF-kappaB: different strategies for responding to stress Statistical analysis lead to a functional antagonism. FASEB J. 2010; 24: 3643–3652. Statistical significance was assessed using unpaired t-tests (Prism 5.0c, 17 Ryan KM, Ernst MK, Rice NR, Vousden KH. Role of NF-kappaB in p53-mediated GraphPad Software, La Jolla, CA, USA). A value of Po0.05 was considerated programmed cell death. Nature 2000; 404: 892–897. as statistically significant with the following degrees: *Po0.05; **Po0.01; 18 Wu L, D’Amico A, Winkel KD, Suter M, Lo D, Shortman K. RelB is essential for the ***Po0.001. development of myeloid-related CD8alpha- dendritic cells but not of lymphoid- related CD8alpha þ dendritic cells. Immunity 1998; 9: 839–847. Mice 19 Weih F, Carrasco D, Durham SK, Barton DS, Rizzo CA, Ryseck RP et al. Multiorgan inflammation and hematopoietic abnormalities in mice with a tar- 1 Â 106 NIH-3T3/RET cells stably transduced with the indicated lentiviral geted disruption of RelB, a member of the NF-kappa B/Rel family. Cell 1995; 80: vectors were injected intracutaneously into 6–8-week-old NOD-SCID mice 331–340. (n ¼ 13/group). Four days post-injection, tumor size was measured every 20 Annunziata CM, Davis RE, Demchenko Y, Bellamy W, Gabrea A, Zhan F et al. other day over a period of 12 days using digital vernier caliper, and tumor Frequent engagement of the classical and alternative NF-kappaB pathways volume was calculated using the formula V ¼ 1/2(L Â W2), where L is length by diverse genetic abnormalities in multiple myeloma. Cancer Cell 2007; 12: (longest dimension) and W is width (shortest dimension). Mice were killed 115–130. either 2 weeks after tumor inoculation or else when the tumor volume 21 Keats JJ, Fonseca R, Chesi M, Schop R, Baker A, Chng WJ et al. Promiscuous exceeded 2500 mm3. Animal studies were conducted in accordance with mutations activate the noncanonical NF-kappaB pathway in multiple myeloma. French regulations for animal experimentation. Cancer Cell 2007; 12: 131–144. 22 Wang X, Belguise K, Kersual N, Kirsch KH, Mineva ND, Galtier F et al. Oestrogen CONFLICT OF INTEREST signalling inhibits invasive phenotype by repressing RelB and its target BCL2. Nat Cell Biol 2007; 9: 470–478. The authors declare no conflict of interest. 23 Xu Y, Josson S, Fang F, Oberley St TD, Clair DK, Wan XS et al. RelB enhances prostate cancer growth: implications for the role of the nuclear factor-kappaB ACKNOWLEDGEMENTS alternative pathway in tumorigenicity. Cancer Res 2009; 69: 3267–3271. 24 Vousden KH, Prives C. Blinded by the light: the growing complexity of p53. Cell We thank F Weih, U Siebenlist, K Vousden, B Vogelstein, T Jacks, J Cleveland, M Oren, 2009; 137: 413–431. D Reisman, G Kroemer, JL Perfettini, M Ko¨ rner, LA Martinez, T Tchenio, J Browning 25 Levine AJ, Oren M. The first 30 years of p53: growing ever more complex. Nat Rev and I Verma for valuable material; G Piton for technical assistance; N Boggetto for cell Cancer 2009; 9: 749–758. sorting (ImagoSeine, Institut Monod, France, Re´gion Ile-de-France grant no. E539); 26 Tergaonkar V, Pando M, Vafa O, Wahl G, Verma I. p53 stabilization is decreased and LL Pritchard for critically reading the manuscript. This work was supported by upon NFkappaB activation: a role for NFkappaB in acquisition of resistance to grants to VB from Agence Nationale pour la Recherche (ANR), Association pour la chemotherapy. Cancer Cell 2002; 1: 493–503. Recherche sur le Cancer, Belgian InterUniversity Attraction Pole, Cance´ropole Ile-de- 27 Xia Y, Padre RC, De Mendoza TH, Bottero V, Tergaonkar VB, Verma IM. Phos- France and Universite´ Paris Descartes, post-doctoral funding from ANR (HA and KB), phorylation of p53 by IkappaB kinase 2 promotes its degradation by beta-TrCP. doctoral fellowships from Ministe`re de la Recherche et des Technologies, Ligue Proc Natl Acad Sci USA 2009; 106: 2629–2634. Nationale contre le Cancer, and Socie´te´ Française du Cancer (EJ), and an award from 28 Jacque E, Tchenio T, Piton G, Romeo PH, Baud V. RelA repression of RelB activity the Bettencourt-Schueller Fundation (EJ). induces selective gene activation downstream of TNF receptors. Proc Natl Acad Sci USA 2005; 102: 14635–14640. 29 Dejardin E. The alternative NF-kappaB pathway from biochemistry to biology: REFERENCES pitfalls and promises for future drug development. Biochem Pharmacol 2006; 72: 1 Oeckinghaus A, Ghosh S. The NF-kappaB family of transcription factors and its 1161–1179. regulation. Cold Spring Harb Perspect Biol 2009; 1: a000034. 30 Mise N, Drosten M, Racek T, Tannapfel A, Putzer BM. Evaluation of potential 2 Hayden MS, Ghosh S. Shared principles in NF-kappaB signaling. Cell 2008; 132: mechanisms underlying genotype-phenotype correlations in multiple endocrine 344–362. neoplasia type 2. Oncogene 2006; 25: 6637–6647. 3 Karin M, Lin A. NF-kappaB at the crossroads of life and death. Nat Immunol 2002; 31 Espinosa JM, Verdun RE, Emerson BM. p53 functions through stress- and pro- 3: 221–227. moter-specific recruitment of transcription initiation components before and after 4 Fan Y, Dutta J, Gupta N, Fan G, Gelinas C. Regulation of programmed cell death by DNA damage. Mol Cell 2003; 12: 1015–1027. NF-kappaB and its role in tumorigenesis and therapy. Adv Exp Med Biol 2008; 615: 32 An W, Kim J, Roeder RG. Ordered cooperative functions of PRMT1, p300, and 223–250. CARM1 in transcriptional activation by p53. Cell 2004; 117: 735–748. 5 Naugler WE, Karin M. NF-kappaB and cancer-identifying targets and mechanisms. 33 Roy B, Reisman D. Positive and negative regulatory elements in the murine p53 Curr Opin Genet Dev 2008; 18: 19–26. promoter. Oncogene 1996; 13: 2359–2366. 6 Basseres DS, Baldwin AS. Nuclear factor-kappaB and inhibitor of kappaB kinase 34 Hellin AC, Calmant P, Gielen J, Bours V, Merville MP. Nuclear factor - kappaB- pathways in oncogenic initiation and progression. Oncogene 2006; 25: 6817– dependent regulation of p53 gene expression induced by daunomycin genotoxic 6830. drug. Oncogene 1998; 16: 1187–1195. 7 Baud V, Karin M. Is NF-kappaB a good target for cancer therapy? Hopes and 35 Mauro C, Leow SC, Anso E, Rocha S, Thotakura AK, Tornatore L et al. NF-kappaB pitfalls. Nat Rev Drug Discov 2009; 8: 33–40. controls energy homeostasis and metabolic adaptation by upregulating mito- 8 Chen F, Castranova V. Nuclear factor-kappaB, an unappreciated tumor suppressor. chondrial respiration. Nat Cell Biol 2011; 13: 1272–1279. Cancer Res 2007; 67: 11093–11098. 36 Huang WC, Ju TK, Hung MC, Chen CC. Phosphorylation of CBP by IKKalpha pro- 9 Perkins ND. NF-kappaB: tumor promoter or suppressor? Trends Cell Biol 2004; 14: motes cell growth by switching the binding preference of CBP from p53 to NF- 64–69. kappaB. Mol Cell 2007; 26: 75–87. 10 Dajee M, Lazarov M, Zhang JY, Cai T, Green CL, Russell AJ et al. NF-kappaB 37 Schumm K, Rocha S, Caamano J, Perkins ND. Regulation of p53 tumour sup- blockade and oncogenic Ras trigger invasive human epidermal neoplasia. Nature pressor target gene expression by the p52 NF-kappaB subunit. EMBO J. 2006; 25: 2003; 421: 639–643. 4820–4832.

Oncogene (2013) 2661 – 2669 & 2013 Macmillan Publishers Limited RelB inhibits cell proliferation through p53 E Jacque et al 2669 38 Barre B, Perkins ND. The Skp2 promoter integrates signaling through the NF- 42 Park E, Zhu F, Liu B, Xia X, Shen J, Bustos T et al. Reduction in IkappaB kinase alpha kappaB, p53, and Akt/GSK3beta pathways to regulate autophagy and apoptosis. expression promotes the development of skin papillomas and carcinomas. Cancer Mol Cell 2010; 38: 524–538. Res 2007; 67: 9158–9168. 39 Hailﬁnger S, Nogai H, Pelzer C, Jaworski M, Cabalzar K, Charton JE et al. 43 Liu B, Park E, Zhu F, Bustos T, Liu J, Shen J et al. A critical role for I kappaB kinase Malt1-dependent RelB cleavage promotes canonical NF-kappaB activation alpha in the development of human and mouse squamous cell carcinomas. Proc in lymphocytes and lymphoma cell lines. Proc Natl Acad Sci USA 2011; 108: Natl Acad Sci USA 2006; 103: 17202–17207. 14596–14601. 44 Maeda G, Chiba T, Kawashiri S, Satoh T, Imai K. Epigenetic inactivation of IkappaB 40 Zhang JY, Green CL, Tao S, Khavari PA. NF-kappaB RelA opposes Kinase-alpha in oral carcinomas and tumor progression. Clin Cancer Res 2007; 13: epidermal proliferation driven by TNFR1 and JNK. Genes Dev 2004; 18: 5041–5047. 17–22. 45 Kieusseian A, Chagraoui J, Kerdudo C, Mangeot PE, Gage PJ, Navarro N et al. 41 Hu Y, Baud V, Oga T, Kim KI, Yoshida K, Karin M. IKKalpha controls formation of the Expression of Pitx2 in stromal cells is required for normal hematopoiesis. Blood epidermis independently of NF-kappaB. Nature 2001; 410: 710–714. 2006; 107: 492–500.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

& 2013 Macmillan Publishers Limited Oncogene (2013) 2661 – 2669