Radiation-triggered Tumor Necrosis Factor (TNF) α-NFκB Cross-signaling Favors Survival Advantage in Human Neuroblastoma Cells*

Induced radioresistance in the surviving cancer cells after radiotherapy could be associated with clonal selection leading to tumor regrowth at the treatment site. Previously we reported that post-translational modification of IκBα activates NFκB in response to ionizing radiation (IR) and plays a key role in regulating apoptotic signaling. Herein, we investigated the orchestration of NFκB after IR in human neuroblastoma. Both in vitro (SH-SY5Y, SK-N-MC, and IMR-32) and in vivo (xenograft) studies showed that IR persistently induced NFκB DNA binding activity and NFκB-dependent TNFα transactivation and secretion. Approaches including silencing NFκB transcription, blocking post-translational NFκB nuclear import, muting TNF receptor, overexpression, and physiological induction of either NFκB or TNFα precisely demonstrated the initiation and occurrence of NFκB → TNFα → NFκB positive feedback cycle after IR that leads to and sustains NFκB activation. Selective TNF-dependent NFκB regulation was confirmed with futile inhibition of AP-1 and SP-1 in TNF receptor muted cells. Moreover, IR increased both transactivation and translation of Birc1, Birc2, and Birc5 and induced metabolic activity and clonal expansion. This pathway was further defined to show that IR-induced functional p65 transcription (not NFκB1, NFκB2, or c-Rel) is necessary for activation of these survival molecules and associated survival advantage. Together, these results demonstrate for the first time the functional orchestration of NFκB in response to IR and further imply that p65-dependent survival advantage and initiation of clonal expansion may correlate with an unfavorable prognosis of human neuroblastoma.

Neuroblastoma (NB), 2 the most frequent extra cranial solid tumor in children (aged Յ5 years) accounts for 8 -10% of all childhood cancers (1) and 15% of childhood cancer fatalities. To that note, NB recurrences remain high (20.2%), and a substantial fraction (46.8%) of those develop metastatic disease. With only 13 months from first diagnosis to recurrence, the survival ratio was 43% for local and 10% for systemic recurrences. Clinical and laboratory evidence suggests that several human cancers contain populations of rapidly proliferating clonogens that can have substantial impact on local control following chemoradiation or radiotherapy (RT) (2). Tumor cell repopulation may arise from remnant cells of the original neoplasm that have escaped therapeutic intervention and later become visible at the original site. RT is now widely used for high risk NB patients after chemotherapy, and the survival rate is significantly improved (76%) by using RT with chemotherapy, compared with chemotherapy alone (46%) (3). Overall goal of the RT is to damage as many cancer cells as possible, while limiting harm to nearby healthy tissue, a major complication with RT. Conversely, ionizing radiation (IR)-induced neoplasms occur at the edges of the irradiated field, where the IR does not cause cell death but is sufficient to induce malignant transformation (4). Unfortunately, these treatment-induced tumors are often more aggressive than primary tumors and are highly refractory to therapy (5).
IR has been shown to activate various transcription factors including NFB (6), and studies have suggested their influential role in tumorigenesis (7). NFB is a member of the c-rel protooncogene family found within the promoter and enhancer region of a wide variety of genes involved in proliferation, apoptosis, inflammation, differentiation, and cell cycle control (8,9). Unlike other inducible transcription factors, a multitude of conditions/agents can activate NFB, and elevated NFB activity has been linked with tumor resistance to chemotherapy and IR (10). Soon after we first reported that clinically relevant doses of IR induces NFB (11,12), innumerable studies both in vitro and in vivo demonstrated that IR specifically activates NFB. We identified that IR profoundly activates NFB in human NB cells (13,14), leading to induced radioprotection, and further that forced inhibition of NFB enhanced IR-induced cell death. To that end, disruption of aberrantly regulated survival signaling mediated by NFB has recently become an important task in the therapy of several chemoresistant/ radioresistant cancers (15). However, mechanistic orchestration of NFB after clinical doses of IR and its functional role in induced survival advantage and/or tumor recurrence is poorly understood.
TNF␣ has been demonstrated to induce NFB via receptor activation (16). Details of the NFB pathways responding to TNF␣ have been well established (17). Mutual activation of NFB and TNF␣ required for the inflammatory response induced by IR has also been suggested (18). TNF␣ can activate NFB through TNF receptor associated factors that in turn interact with NFB-interacting kinase, which plays a key role in cytokine-induced NFB activation in irradiated cells. Furthermore, ERK activated by TNF␣ regulates NFB activation (19,20) through IB kinase phosphorylation. To that extent, blocking NFB has been demonstrated to sensitize cancer cells to TNF␣-induced killing (21). Recent evidence suggests that endogenous production of TNF␣ is a potent trigger of NFB activation by IR. In addition, molecular cloning analysis has disclosed the presence of one or more putative binding sites for NFB in the promoter/enhancer region of TNF␣ (17,22). Accordingly, we investigated whether the cells of the original neoplasm that have escaped IR insult result in the development of concurrent radioadaptation and survival advantage mediated by persistent activation of NFB through positive feedback (NFB 3 TNF␣ 3 NFB) cycle (PFC). Our data suggests that at least in human NB cells, clinical doses of IR results in the (i) occurrence of NFB 3 TNF␣ 3 NFB PFC; (ii) feedback cycledependent sustained activation of NFB; (iii) NFB-dependent regulation of prosurvival IAP1, IAP2, and Survivin; and (iv) and NFB-mediated radioprotection and survival advantage.
Inhibition/Overexpression and Irradiation Experiments-For IR-induced sustained activation of NFB and initiation of NFB-TNF␣ PFC, the cells were exposed to 2 Gy using Gamma Cell 40 Exactor (Nordion International Inc., Ottawa, Canada) at a dose rate of 0.81 Gy/min. Mock irradiated cells were treated identical except that the cells were not subjected to IR. Irradiated cells were incubated at 37°C for an additional 1, 3, 6, 24, 48, and 72 h. All of the experiments were repeated at least three times in each group. For NFB inhibition studies, the cells were treated for 3 h prior to IR or 1 and 24 h after IR with 50 nM of SN50 NFB cell-permeable peptide (Calbiochem, La Jolla, CA). Likewise for TNF␣ inhibition, 100 ng/ml TNFR1 antibody (Santa Cruz Biotech, Santa Cruz, CA) was used. Conversely, for physiological induction of NFB, the cells were treated with 1 mM SNP (Sigma-Aldrich) or 20 ng/ml of endotoxin-free exogenous human recombinant TNF␣ (ProSpec-Tany Ltd., Ness-Ziona, Israel).
Plasmid Preparation, DNA Transfection, and Luciferase Reporter Assay-Transient transfection of NFB p65 and p50 subunits was carried out by the lipofection method using Effectene TM reagent (Qiagen) as described in our earlier studies (24). NFB inhibition was achieved using 150 ng of siRNAs (Qiagen) targeting RelA, NFB1, NFB2, and Rel. siRNA mixed with 12 l of HiPerfect transfection reagent (Qiagen) was incubated for 15 min and slowly added to 80% confluent cells grown in 30-mm plates. After 18 h, transfection medium was replaced with growth medium before IR. Moreover, NFB inhibition was also accomplished using transient transfection of S32A/ S36A double mutant IB␣ (⌬IB␣; Upstate Biotechnology, Lake Placid, NY). The mutated form of IB␣ with a serine-toalanine mutation at residues 32 and 36 does not undergo signalinduced phosphorylation and thus remains bound to NFB, subsequently preventing nuclear translocation and DNA binding. In addition, alterations in IR-induced NFB promoter activation were investigated in SH-SY5Y and IMR-32 cells treated with TNFR1 Ab or transfected with RelA siRNA. The pNFB-Luc plasmid construct was amplified and purified as described earlier (22). Cell lysates were assayed for luciferase activity as per the manufacturer's protocol (Biovision Research Products, Mountain View, CA).
Electrophoretic Mobility Shift Assay-Nuclear protein extraction and electrophoretic mobility shift assay were performed as described in our earlier studies (13,14). For the competition assay, the nuclear extract was preincubated with unlabeled homologous NFB oligonucleotide followed by the addition of [␥-32 P]ATP-labeled NFB probe. Supershift analysis was performed as described earlier (25).
Immunoblotting-Total protein extraction and immunoblotting were performed as described in our earlier studies (26). For this study, the protein-transferred membranes were incubated with either mouse monoclonal anti-pIB␣ antibody; rabbit polyclonal anti-IB␣, -cIAP1, -cIAP2, or -TNF␣ antibody; or Survivin antibody (Santa Cruz). The blots were stripped and reblotted with mouse monoclonal anti-␣-tubulin antibody (Santa Cruz) to determine equal loading of samples.
QPCR-IR-induced NFB-dependent regulation of TNF␣ mRNA expression and persistent activation of NFB-dependent transcriptional response of cIAP1, cIAP2, and survivin were analyzed by real time QPCR as described earlier (13). Likewise, inhibition of IR-induced p65 transcriptional levels in RelA siRNA transfected SH-SY5Y and IMR-32 cells were validated using QPCR. We used ␤-actin as a positive control, and a negative control without template RNA was also included. Each experiment was carried out four times, and the ⌬⌬ Ct values were calculated by normalizing the gene expression levels to ␤-actin, and the relative expression level was expressed as a fold change. Group-wise comparisons were made using analysis of variance with Tukey's post-hoc correction.
ELISA-ELISA was performed as described in our earlier studies (27). In this study, conditioned medium from cells either mock irradiated or exposed to IR was recovered after 10 min through 72 h and concentrated using 9KD ICON concentrators (Thermo Scientific, Rockford, IL). Samples and standards coated on high binding microwell plates were blocked and labeled with human TNF␣ antibody (Santa Cruz) and tagged with anti-rabbit IgG HRP conjugate (Alpha Diagnostics, San Antonio, TX). TMB substrate was used as a detection system, and the reaction was stopped using 1 N HCl. The absorbance at 450 nm was read on a Synergy II multi-detection microplate reader (Biotek Instruments, Winooski, VT) and group-wise comparisons were made using analysis of variance with Tukey's post-hoc correction. A p value of Ͻ0.05 is considered statistically significant.
Cell Survival by MTT and Clonogenic Assay-Cell survival was analyzed using MTT and clonogenic assays as described in our previous studies (23). For colony forming assay, NB cells exposed to 2 Gy were allowed to incubate for 24, 48, and 72 h. Cells from the respective plates were then seeded at a density of 2500 cells/30-mm plate and incubated for an additional 14 days. Plates from all groups including mock IR controls were transported to the irradiation facility and handled similarly to normalize any variations between groups. The colonies were fixed, stained with 0.5% crystal violet, counted using computed colony counting (Image Quant; GE Healthcare), implying standard sensitivity, noise factor, and background. For MTT assay, the cells (1000 cells/300 l in a 24-well plate) that were either mock irradiated, exposed to IR, treated with TNFR1 Ab, or transfected with RelA siRNA and exposed to IR were treated with MTT (30 l/well from 5 mg/ml stock) for 4 h after 24, 48, and 72 h of post-IR. Solubilization of converted purple formazan dye was accomplished by acid-isopropanol with continuous shaking at 37°C. The reaction product was quantified by measuring the absorbance at 570 nm using Synergy II micro plate reader (Biotek). Cell survival response was compared using analysis of variance with Tukey's post-hoc correction.
In Vivo Xenograft Experiments-All of the experiments conformed to American Physiological Society standards for animal care and were carried out in accordance with guidelines laid down by the National Research Council and were approved by the University of Oklahoma Health Sciences Center Institutional Animal Care and Use Committee. Seven-week-old athymic NCr-nu/nu nude mice (National Cancer Institute, Frederick, MD) received subcutaneous injections of SK-N-MC cells (5 ϫ 10 6 ) suspended in 100 l of medium into their right flank. Tumor growth was periodically monitored, tumor volume was calculated using the formula V ϭ [(/6) ϫ L ϫ W 2 ] (28), and tumors were allowed to grow to a size of ϳ500 mm 3 . Xenografts were exposed to fractionated irradiation (FIR) of 2 Gy/day for 5 days. FIR dose regimen was used to effectively delineate the induced response in the xenograft. A specially designed cerrobend shield was used to encase the body of the mice, and the exposed flank tumor (ϳ500 mm 3 ) was irradiated using Gamma Cell. For inhibition studies, the animals received intraperitoneal injection of 20 mg/kg NFB selective pharmacological inhibitor (29), N-acetyl-leucinyl-leucinyl-norleucinal-H (ALLN; Calbiochem) or 200 g of TNFR1 Ab (Santa Cruz) 3 h before first day IR. Likewise, for radioresponse simulation studies, the animals received intraperitoneal injection of 20 mg/kg of NFB physiological inducer, SNP (Sigma), or endotoxin-free recombinant TNF␣ (ProSpec) (30). Xenografts were collected 72 h after FIR and examined for modulations in NFB DNA binding activity and TNF␣ as discussed above.  Fig. S1). Specific binding of NFB to its sequence-specific oligonucleotide was confirmed with competition binding assay (supplemental Fig. S2A). Moreover, supershift assay revealed that the major subunits in all three cell types are p50 and p65 (supplemental Fig. S2, B-E). The differences seen between the cell types are the magnitude of p50/p65 heterodimers and p50/p50 homodimers. The distance between the heterodimer and homodimer band is varied between different cell types because the electrophoretic runs were performed separately at different time points and because the levels of NFB in a given amount of total protein varied and changed the exposure time to properly show both bands. Furthermore, immunoblotting revealed relatively reduced levels of constitutive IB␣ levels after 1, 3, 24, 48, and 72 h in irradiated SK-N-MC, IMR-32, and SH-SY5Y cells (Fig. 1B). The expression of IB␣ decreased at 1 h and continues to be at a reduced level until 24 h after exposure. The IB␣ expression again started increasing at 48 and 72 h. This correlates well with the reduced activation of NFB at that later time points, i.e. 48 and 72 h (Fig. 1A, left panel). Increasing IB␣ at a later time point and the corresponding decrease in NFB activity indicate the feedback cycling of NFB. Conversely, we observed a consistent induction of IB␣ phosphorylation in these cells after IR. To that note, we observed an induction in IB␣ phosphorylation immediately (1 h) after IR, and this induced phosphorylation remained consistent at least up to 72 h in all three NB cell lines investigated (Fig. 1B). Once IB␣ is phosphorylated, phospho-IB␣ undergoes ubiquitination and subsequent degradation. It was not apparent in neuroblastoma cells. These findings prompt us to determine in future studies the possibility of deregulated ubiquitination pathway in these cell types. Because the occurrence of phosphorylation of IB␣ results in an immediate release of active NFB and the active NFB translocates into the nucleus (confirmed by the DNA binding activity of NFB in nuclear extracts), its downstream function is not affected by the sustained levels of phosphor-IB␣ in the cytosol.

IR-induced Persistent Activation of NFB-To
IR-induced TNF␣ Transactivation and Sustained Secretion of TNF␣-To delineate whether IR-induced alterations in TNF␣ actively contribute to the induced persistent NFB activity and subsequent survival advantage, we elucidated whether IR modulates TNF␣. The cells exposed either to mock IR or 2 Gy and harvested after 10 min through 72 h were analyzed for alterations in TNF␣ mRNA. Compared with mock IR, 2 Gy significantly induced TNF␣ transactivation after 10, 15, and 30 min and 1, 3, 6, 12, and 24 h (Fig. 1C). Furthermore, we observed a sustained and significant (p Ͻ 0.001) TNF␣ mRNA induction after 48 and 72 h in SK-N-MC cells. Secondly, concentrated medium recovered after 10 min through 72 h from SK-N-MC, IMR-32, and SH-SY5Y cells exposed to either mock IR or 2 Gy were examined for secreted levels of TNF␣. IR significantly increased secreted levels of TNF␣ after 10, 15, 30, or 45 min and 1, 3, and 12 h post-IR in SK-N-MC cells. More importantly, significant and sustained induction of secreted TNF␣ was evident after 24, 48, and 72 h post-IR in all three NB cell lines (Fig. 1D).
IR-induced NFB-TNF␣-NFB PFC-To precisely identify and validate the occurrence of IR-induced NFB-TNF␣-NFB PFC, we investigated whether (i) inhibition of TNF␣ or NFB disrupts IR-induced PFC and (ii) simulating the IR response by activating NFB or TNF␣ initiates PFC. First, to accomplish the disruption of IR induced PFC, IR-induced TNF␣ was studied after inhibiting NFB. Inhibition of NFB in SK-N-MC cells was accomplished using SN50 cell-permeable peptide (50 nM for 3 h), whereas SH-SY5Y and IMR-32 cells were transfected with RelA siRNA. The cells were then exposed to mock IR or 2 Gy and harvested after 15 min through 72 h. Treating SK-N-MC cells with SN50 markedly inhibited IR-induced NFB DNA binding activity as early as 15 min, and this induced inhibition remained significant consistently at all time points at least up to 72 h ( Fig. 2A). More importantly, induced inhibition of NFB concordantly inhibited secreted TNF␣ (Fig. 2B). To that end, inhibiting NFB significantly (p Ͻ 0.001) inhibited sustained elevation of NFB-dependent secreted TNF␣ in SK-N-MC cells. Coherently, parallel cultures of RelA siRNA transfected SH-SY5Y and IMR-32 cells showed significant inhibition of IR-induced NFB activity ( Fig. 2A). Autorads were overexposed to capture the reduced activities that were lower than mock IR controls. QPCR analysis revealed a significant inhibition of IR-induced NFB transcriptional activation as early as 1 h and remained consistent at least up to 72 h in RelA siRNA transfected SH-SY5Y and IMR-32 cells (Fig. 3A). Consequently, when IR-induced persistent NFB activation is muted, we observed a significant (p Ͻ 0.001) decrease in secreted TNF␣ (Fig. 2B). These results clearly elucidate that inhibition of IRinduced NFB resulted in the sustained suppression of TNF␣ and thereby demonstrates the initiation of NFB-TNF␣ link in NB cells after clinical doses of IR.
Furthermore, to delineate the occurrence of a TNF␣-NFB feedback, we adopted two approaches. First, we investigated the IMR-32 cells exposed to IR. The cells were exposed to IR (2 Gy) and analyzed for total and phospho-IB␣ after 1, 3, 24, 48, and 72 h. ␣-Tubulin expression was determined to validate equal sample loading. C, real time QPCR analysis showing TNF␣ mRNA expression in human NB cells exposed to IR. ␤-actin was used as the positive control. The histogram shows the fold change in relation to mock irradiated controls. IR significantly induced TNF␣ transcription in SK-N-MC, SH-SY5Y, and IMR-32 cells. D, results of ELISA showing induced intercellular levels of TNF␣ in NB cells exposed to IR. Used medium from cells either mock irradiated or exposed to IR were recovered after 10, 15, 30, or 45 min or 1, 3, 6, 12, 24, 48, or 72 h and were concentrated (9KD concentrators) and subjected to ELISA. Group-wise comparisons were made using analysis of variance with Tukey's post-hoc correction. ANOVA, analysis of variance; Neg., negative.
effect of post-IR-induced NFB in TNF␣ transactivation and secretion. To achieve this, human SK-N-MC, SH-SY5Y, and IMR-32 cells were exposed to mock IR or IR (2 Gy) and were incubated at 37°C for additional 1 or 24 h. The cells were then treated with SN50 and examined for TNF␣ transactivation (at 1 and 3 h) and secretion (at 24, 48 and 72 h). EMSA analysis showed complete suppression of IR-induced NFB after both 1h and 24 h post-IR SN50 treatment (Fig. 3B). Further, this post-IR SN50-induced inhibition of NFB was sustained at least up to 72 h (Fig. 3B). Coherently, in all three cell lines, we observed a significant inhibition of IR-induced TNF␣ transactivation after both 1 and 24 h post-IR SN50 treatment (Fig. 3C). Consistently, ELISA analysis revealed a significant (p Ͻ 0.001) and sustained (at least up to 72 h) inhibition of TNF␣ intercellular secretion in all three of the cell lines investigated (Fig. 3D).
Next, we investigated the alterations in IR-induced NFB after blocking TNF␣. Blocking TNF␣ was achieved by treating the cells with TNFR1 Ab and validated with ELISA. Treatment with TNFR1 Ab completely (p Ͻ 0.001) suppressed the IR-induced secreted TNF␣ in SK-N-MC cells as early as 15 min after IR, and this induced inhibition remained consistent up to 72 h after IR (Fig. 2C). More importantly, blocking TNF␣ inhibited IRinduced NFB activity at all time points investigated (Fig. 2D). Autorads were overexposed to capture the reduced activities that were lower than mock IR. Similarly, compared with IRexposed cells, both SH-SY5Y and IMR-32 cells treated with TNFR1 Ab and exposed to IR showed complete (p Ͻ 0.001) inhibition of secreted TNF␣ after 24, 48, and 72 h (Fig. 2C). Concordantly, silencing TNF␣ significantly inhibited IR-induced NFB activity after 1 h through 72 h in both cell lines investigated (Fig. 2D). The specificity of NFB inhibition with blocking secreted TNF␣ (with TNFR1 antibody) was examined by analyzing the alterations in AP1 and SP1-DNA binding activity. SK-N-MC, SH-SY5Y, and IMR-32 cells exposed to both mock IR and IR (2 Gy) with or without TNFR1 Ab pretreatment were analyzed for SPI and AP1 DNA binding activity after 1-72 h. Blocking TNF␣ did not reveal any significant inhibition of both SP1 and AP1 DNA binding activity even after 72 h (Fig. 2E) in all three of the cell lines investigated and serves as the positive control for the study. Taken together, these results clearly portray the occurrence of specific TNF␣-NFB feedback in all of the NB cell lines examined.
Moreover, to substantiate the occurrence of PFC, IR simulation experiments were performed. First, IR-induced molecular response was simulated with increased NFB by overexpressing NFB or by treating the cells with physiological inducer, SNP (50 g/kg). Overexpressing NFB subunit significantly induced NFB DNA binding activity (Fig. 4A) after 24 h and remained at elevated levels after 48 and 72 h. Immunoblotting revealed an increased p65 and p50 expression in these transfected cells (supplemental Fig. S2F). Consistently, overexpressing NFB profoundly induced secreted TNF␣ levels in these cells and remained high up to 72 h (Fig. 4B). Likewise, cells exposed to SNP and harvested after 15 min through 72 h showed robust NFB DNA binding activity in SK-N-MC cells (Fig. 4C). Concordantly, SNP-induced NFB resulted in a marked and significant (p Ͻ 0.001) up-regulation in secreted TNF␣ levels (Fig. 4D) as early as 15 min and remained consistent up to 72 h. These results validate that persistent activation of NFB in response to a stimuli (in our case clinical doses of IR) initiate a NFB-TNF␣ link and subsequent feedback cycle. Conversely, occurrence of TNF␣-dependent feedback was investigated by exposing the cells to exogenous human recombinant TNF␣. SK-N-MC cells treated with rH-TNF␣ and harvested after 15 min through 72 h were first validated for significant (p Ͻ 0.001) induction of secreted TNF␣ (Fig. 4E). Interestingly, TNF␣ induction significantly activated NFB in NB cells as early as 30 min (Fig. 4F) and remained persistent up to 72 h. Taken together, these results clearly delineate that IRinduced NFB initiates the activation of TNF␣, and the secreted TNF␣ in turn activates NFB and thereby promotes a NFB-TNF␣-NFB PFC. Moreover, as an end product of this IR-induced PFC, elevated NFB activity remains persistent and mediates the IR-induced downstream survival response.
NFB Regulates IR-induced cIAP1, cIAP2, and survivin-SK-N-MC cells exposed to 2 Gy and harvested after 15 min through 24 h were analyzed for the IR-induced transcriptional alterations in cIAP1, cIAP2, and survivin. Compared with mock IR, QPCR analysis revealed a robust and significant induction of cIAP1 mRNA at all time points investigated (Fig. 5A). Furthermore, we observed an elevated level of cIAP2 and survivin in SK-N-MC cells. Consistently, immunoblotting analysis showed an induced level of cIAP1, cIAP2, and Survivin as early as 30 min (Fig. 5B) and reached maximum at 90 min. We observed a unrelenting induction of these proteins at least up to 72 h. Likewise, SH-SY5Y cells exposed to IR showed a significant (p Ͻ showing elevated levels of intercellular TNF␣ in cells exposed to SNP and harvested after 15, 30, 45 and min and 1, 3, 6, 24, 48, and 72 h. E, ELISA analysis confirming the increased levels of intercellular TNF␣ levels in SK-N-MC cells. The cells were exposed to 20 ng/ml endotoxin-free exogenous human recombinant TNF␣. F, representative autoradiogram from EMSA analysis demonstrating the induced levels of NFB DNA binding activity in human SK-N-MC cells exposed to rH-TNF␣. ANOVA, analysis of variance. 0.001) increase in cIAP1 mRNA after 3 h post IR and remained high even after 24, 48, and 72 h (Fig. 5C). Compared with the mock irradiated controls, we observed a significant increase in cIAP2 as early as 1 h, and cIAP2 remained at elevated levels up to 24 h. Only a marginal difference in cIAP2 mRNA was observed after 48 h in these cells. However, after 72 h post IR, the cIAP2 levels were significantly (p Ͻ 0.001) induced in SH-SY5Y cells. Similarly, a significant (p Ͻ 0.001) and profound increase in survivin mRNA levels was observed after 1, 3, 24, 48, and 72 h (Fig. 5C). Coherent with the mRNA expression data, immunoblotting analysis showed a robust induction of Survivin, cIAP1, and cIAP2 proteins after 24, 48, and 72 h (Fig. 5D). Consistent with expression patterns observed in SK-N-MC and SH-SY5Y cells, QPCR analysis showed a significant induction of cIAP1, cIAP2, and survivin mRNA levels in IMR-32 cells at all time points investigated (Fig. 5E). Interestingly, both cIAP2 and survivin reached almost baseline levels after 48 h while reaching the maximal levels after 72 h. Conversely, compared with the mock irradiated controls, Western blot analysis showed a robust induction of cIAP1 after 24, 48, and 72 h and of cIAP2 and Survivin after 48 and 72 h post IR in IMR-32 cells (Fig. 5D). Taken together, these results precisely demonstrate the up-regulation of pro-survival cIAP1, cIAP2, and Survivin after clinical doses of IR in human NB cells.
Furthermore to throw light on the mechanism of IR-induced survival signaling, we investigated whether IR-induced persis- , and IMR-32 (E) cells. NB cells exposed to IR were incubated in a CO 2 /air incubator for additional 15,30,45,60, and 90 min and 3, 6, 12, 24, 48, and 72 h and subjected to QPCR analysis. ␤-actin was used as the positive control. The histogram shows the fold change in relation to mock irradiated controls. IR significantly induced cIAP1, cIAP2, and survivin transcription in all three cell lines investigated. B, Western blot analysis showing the induced levels of cIAP1, cIAP2, and Survivin in SK-N-MC cells exposed to IR and incubated for additional 15,30,45,60, and 90 min and 3, 6, 12, 24, 48, and 72 h. ␣-Tubulin expression was determined to validate equal sample loading. D, Western blot analysis showing the induced levels of cIAP1, cIAP2, and Survivin in SH-SY5Y and IMR-32 cells exposed to IR and incubated for additional 24, 48, and 72 h. Forced inhibition of NFB with RelA siRNA transfection or TNF␣ with TNFR1 Ab resulted in complete inhibition of cIAP1, cIAP2, and Survivin in both cell lines. F, luciferase reporter assay. SH-SY5Y and IMR-32 cells were transfected with pNFB-Luc construct and then either mock irradiated, exposed to 2 Gy, treated with TNFR1 Ab, and exposed to 2 Gy or co-transfected with RelA siRNA and exposed to 2 Gy. The cells were harvested at 24 and 72 h post-IR, and equal amount of lysates were subjected to luciferase assay. The data shown represent the means and S.D. of three independent experiments. G, representative autoradiogram of EMSA analysis showing NFB DNA binding activity in SK-N-MC cells exposed to 2 Gy, transfected with ⌬IB␣, or transfected with ⌬IB␣ and exposed to 2 Gy. ⌬IB␣ completely inhibited IR induced NFB DNA binding activity. H, QPCR analysis showing complete suppression and overexpression of cIAP1, cIAP2, and survivin transcription in SK-N-MC cells transfected with ⌬IB␣ and p65, respectively, after 3 and 6 h. ANOVA, analysis of variance. tent activation of NFB mediates the expression of cIAP1, cIAP2, and Survivin. First, human SH-SY5Y and IMR-32 cells were transfected with pNFB-Luc construct that expresses the luciferase reporter gene in an NFB-dependent manner. It contains four tandem copies of the NFB consensus sequence fused to a TATA-like promoter region of the herpes simplex virus thymidine kinase promoter. Binding of NFB to the promoter activates transcription, allowing the Luc reporter gene to be expressed. Transfected cells were then mock irradiated, exposed to 2 Gy, treated with TNFR1 Ab, and exposed to 2 Gy or co-transfected with RelA siRNA and exposed to 2 Gy. The luciferase assay was performed in extracts obtained from cells harvested after 24 and 72 h. Compared with mock IR, cells exposed to 2 Gy showed a significant (2.0 -9.0-fold) increase in luciferase activity, indicating that IR could specifically initiate transcriptional activation of NFB in both cell lines investigated (Fig. 5F). Conversely, cells co-transfected with RelA siRNA showed complete inhibition of IR-induced NFB transcriptional activity. Consistently, immunoblot analysis revealed a significant suppression of IR-induced IAP1, IAP2, and Survivin in both SH-SY5Y and IMR-32 cells (Fig. 5D). Moreover, luciferase assay in cells pretreated with TNFR1 Ab showed marked inhibition of IR-induced NFB transcriptional activity (Fig. 5F). Consequently, inhibition of IR-induced NFB transcriptional activity with TNFR1 Ab resulted in significant suppression of IAP1, IAP2, and Survivin in both cell lines (Fig. 5D). As an alternative approach, SK-N-MC cells either (i) mock irradiated, exposed to 2 Gy, transfected with ⌬IB␣, and exposed to IR or (ii) transfected with p65 subunit were harvested after 3 and 6 h and examined for modulations in cIAP1, cIAP2, survivin, and XIAP mRNA levels. First, inhibition of NFB activity with ⌬IB␣ transfection was validated using EMSA analysis. Compared with the mock irradiated controls, ⌬IB␣ significantly inhibited the NFB DNA binding activity in human SK-N-MC cells. More importantly, ⌬IB␣ completely suppressed the IR-induced NFB DNA binding activity (Fig. 5G). Concordantly, compared with the IR exposure, QPCR analysis revealed a significant suppression of cIAP1 (p Ͻ 0.001), cIAP2 (p Ͻ 0.001), survivin (p Ͻ 0.001), and XIAP (p Ͻ 0.05) after both 3 and 6 h in ⌬IB␣ transfected cells that were exposed to IR. Moreover, compared with the 2 Gy exposed cells, forced activation of NFB (using p65 transfection) markedly induced FIGURE 6. IR induced NFB dependent survival advantage. A and B, MTT analysis showing survival response in NB cells exposed to 2 Gy after 24, 48, and 72 h (A) or after forced inhibition of IR-induced NFB or TNF␣ (B). C, clonogenic analysis of SK-N-MC cells either mock irradiated or exposed to 2 Gy. D, histograms showing colony forming capacity of SK-N-MC cells exposed to IR, transfected with ⌬IB␣, and exposed to IR or treated with TNFR1 Ab and exposed to IR. Compared with 2 Gy, silencing IR-induced NFB completely suppressed clonal expansion. Moreover, inhibition of IR-induced NFB-dependent TNF␣ revealed robust inhibition of clonogenic activity in SK-N-MC cells. ANOVA, analysis of variance. cIAP1, cIAP2, survivin, and XIAP mRNA levels after 3 and 6 h (Fig. 5H). These results demonstrate that (i) IR induces prosurvival IAP1, IAP2, and Survivin; (ii) IR-induced transcriptional activation of NFB mediates the IAP1, IAP2, and Survivin levels; and (iii) blocking TNF␣ attenuates IR-induced NFB-dependent survival molecules in human NB cells.
NFB Regulates IR-induced Survival Advantage in NB Cells-To assess the IR-induced survival advantage and to delineate the influence of NFB activity in mediating the induced survival response, we examined the induced modulations in both metabolic and clonogenic activity of the cells. To determine the changes in cell survival, SK-N-MC, SH-SY5Y, and IMR-32 cells were exposed to 2 Gy and examined using MTT analysis after 24, 48, and 72 h. Compared with the mock irradiated cells, IR significantly reduced cell survival after 24, 48, and 72 h post IR in all three cell lines investigated (Fig. 6A). However, compared with the 2 Gy exposed cells, forced inhibition of NFB (through RelA siRNA transfection) completely (p Ͻ 0.001) inhibited the cell survival after 24, 48, and 72 h in all three cell lines investigated, demonstrating that inhibition of IR-induced NFB significantly enhances IR-induced cell killing (Fig. 6B). Further, we investigated whether inhibition of other subunits of NFB has similar effect in cell killing. To achieve this, SH-SY5Y, IMR-32, and SK-N-MC cells transfected with NFB1, NFB2, or Rel siRNA were exposed to 2 Gy and examined for cell survival after 24, 48, and 72 h. Interestingly, we did not observe any significant difference in cell survival in these transfected cells as opposed to 2 Gy exposure (Fig. 7). Furthermore, inhibition of IR-induced NFB-dependent TNF␣ (with TNFR1 Ab) profoundly enhanced the IR-induced cell killing after 24, 48, and 72 h in SK-N-MC, SH-SY5Y, and IMR-32 cells, suggesting that IR-induced NFB-initiated NFB-TNF␣-NFB PFC-dependent persistent activation of NFB mediates survival advantage, and disruption of the induced feedback cycle inhibits the survival advantage and promotes radiosensitivity. To substantiate these findings, human SK-N-MC cells either mock irradiated or exposed to 2 Gy were seeded after 24, 48, and 72 h to assess the colony forming capacity (Fig. 6C). Compared with the mock irradiated cells, IR significantly (p Ͻ 0.001) induced clonogenic activity after 24, 48, and 72 h (Fig. 6D). Conversely, compared with the 2 Gy exposed cells, forced inhibition of IR-induced NFB completely suppressed (p Ͻ 0.001) the colony forming capacity, demonstrating that NFB regulates IR-induced survival advantage (Fig. 6D). Moreover, inhibition of IR-induced NFB-dependent TNF␣ revealed robust inhibition of clonogenic activity in human SK-N-MC cells, flaunting that disruption of IR-induced feedback cycle inhibits induced survival advantage.
IR-induced NFB-TNF␣-dependent Persistent Activation of NFB Mediates Survival Advantage in Xenograft-To substantiate our in vitro findings, human SK-N-MC xenografts developed in athymic nude mice (Fig. 8A) that were either mock irradiated, exposed to FIR (2 Gy five times), treated with SNP and exogenous rH-TNF␣, treated with ALLN or TNFR1 Ab, and exposed to FIR were examined for modulation in NFB DNA binding activity, TNF␣ expression, and alterations in prosurvival cIAP1, cIAP2, and Survivin. EMSA analysis revealed that compared with mock irradiated control, FIR significantly induced NFB DNA binding activity in FIR-exposed xenografts after 3 days (Fig. 8B). Conversely, FIR-induced NFB activity was significantly suppressed in ALLN-treated animals. Similarly, forced inhibition of IR-induced NFB-dependent TNF␣ with TNFR1 Ab profoundly inhibited IR-induced NFB DNA binding activity in the tumors. Consistently, QPCR analysis revealed a marked increase in cIAP1, cIAP2, and survivin mRNA in the xenografts exposed to FIR (Fig. 8C). Similarly we observed a significant induction of cIAP1, cIAP2, survivin, and TNF␣ mRNA levels in xenografts exposed to exogenous rH-TNF␣ or NFB physiological inducer SNP, suggesting that NFB induction in tumor initiates a link with downstream TNF␣ and subsequently induces survival advantage. Conformingly, when we inhibited the FIR-induced NFB with ALLN or TNF␣ with TNFR1 antibody, FIR-induced cIAP1, cIAP2, and survivin were significantly inhibited in the xenografts (Fig. 8C). Furthermore, immunoblotting analysis revealed a consistent increase in TNF␣ expression in xenografts exposed to FIR, SNP, or exogenous rH-TNF␣ (Fig. 8D). To that end, compared with FIR-exposed xenografts, inhibition of FIR-induced NFB with ALLN significantly inhibited TNF␣.

DISCUSSION
NB remains a major therapeutic challenge in pediatric oncology despite the high response rates. The radiobiological consid- FIGURE 7. Role of other NFB subunits in regulating IR-induced inhibition of survival advantage. MTT analysis shows survival response in SH-SY5Y, IMR-32, and SK-N-MC cells exposed to 2 Gy or transfected with NFB1, NFB2, or Rel siRNA and exposed to 2 Gy after 24, 48, and 72 h. erations predict that the combination of radio and chemotherapy administered as initial treatment for NB would be the optimum clinical strategy. Convincingly, IR delivered to the local NB sites has several well recognized applications including respiratory distress relief for stage IV patients apart from its direct tumoricidal effect. To that note, functional links between cellular signal transduction responses and DNA damage recognition, repair, and cell death have been well recognized. IR is known to induce signal transduction pathways that lead to apoptosis. However, most tumors including NB (31) respond to the effects of IR oppositely by inducing pro-survival signal transduction pathways. Furthermore, we (14,23,32) and others (31) have reported an elevated constitutive level of NFB in NB cells and enhanced NFB DNA binding activity as an IR response. More importantly, studies have causally linked the induced NFB activity to the responsiveness to therapy and survival of NB cells (33)(34)(35). Together, these observations tie NFB responsiveness to the tumorigenic behavior of NB, particularly with treatment resistance and relapse. Here, we provide insight into the mechanistic regulation of NFB in response to IR and its direct role in subsequent survival advantage. The results of the present study clearly elucidated an induced level of NFB activity as a response to IR, and further we have shown that the induced levels were persistent with no recovery from the altered activation at least for 3 days. However, sustained elevation of NFB activity has been reported in normal lung tissue in response to IR (36); to our knowledge, for the first time, this study provides evidence of persistent activation of NFB in tumor cells, in particular, NB cells exposed to single clinically relevant dose of IR.
Moreover, our results implicate IR-induced NFB mediated initiation of a TNF␣-dependent positive feedback mechanism. Consistent with our NFB activity data, sustained activation of TNF␣ transcription and intercellular TNF␣ levels were evident in all three NB cell lines investigated. TNF␣, produced originally by activated T cells and macrophages, has been demonstrated to induce NFB via receptor activation (16). Details of the NFB pathways responding to TNF␣ have been well estab- FIGURE 8. A, representative photograph showing NB xenograft in 7-week-old athymic nude mouse. B, representative autoradiogram of EMSA analysis showing NFB DNA binding activity. Xenografts were exposed to FIR of 2 Gy/day for 5 days. For inhibition of FIR-induced NFB, the animals received intraperitoneal injection of ALLN, and for TNF␣ inhibition, the animals received TNFR1 Ab. C, QPCR showing cIAP1, cIAP2, survivin, and TNF␣ transcription in xenografts either exposed to FIR with or without NFB/TNF␣ inhibition, SNP, or with rH-TNF␣. D, immunoblot showing TNF␣ levels in xenografts exposed to FIR with or without NFB/TNF␣ inhibition, with or without SNP, or with rH-TNF␣.
lished (17,37). A mutual activation of NFB and TNF␣ required for the inflammatory response induced by IR has also been suggested (18). TNF␣ can activate NFB through TNF receptor associated factors (38) that in turn interact with the downstream NFB-interacting kinase (39), which plays a key role in cytokine-induced NFB activation in irradiated cells. Furthermore, ERK activated by TNF␣ regulates NFB activation (19,20,40) through IB kinase phosphorylation. To that extent, blocking NFB has been demonstrated to sensitize cancer cells to TNF␣-induced killing (21,41). Similarly, studies, both in vitro and in vivo, have demonstrated that pretreatment of cells with TNF␣ resulted in increased NFB activation (42,43). Recently, it has been demonstrated that endogenous production of TNF␣ is a potent trigger of NFB activation by IR. In addition, molecular cloning analysis has disclosed the presence of one or more putative binding site for NFB in the promoter/ enhancer region of TNF␣. Using five independent strategies, our results show that IR-induced NFB-dependent TNF␣ is required for the persistent activation of NFB in NB cells and clearly portrayed the initiation and occurrence of NFB-TNF␣-NFB feedback mechanism. Regulation of expression of TNF␣ is a complex process. TNF␣ is transcriptionally active within minutes and protein production begins within few hours after being initially stimulated with an inducer (in this case radiation exposure). The secreted TNF␣ lasts as a soluble factor in the culture supernatant for less than an hour. The second synthesis involves receptor binding, activation of NFB, transcriptional initiation of its promoter, regulation of message splicing, regulation of message turnover, and regulation of translational product as a mature protein (44), which accounts for a lapse in the availability of TNF␣ in the culture supernatant examined. This pattern of initial induction, lapse in the availability of TNF␣, and again secretion of the second synthesis is shown as two-phase induction. Although this type of time-dependent biphasic induction is not significantly demonstrated in NFB activation, findings of TNF␣ inhibition and NFB blocking studies clearly demonstrated that the activation of NFB and the production of TNF␣ are interdependent regulation that maintains the feedback loop. In addition, NFB measure by EMSA may deceit a clear bi-phasic induction because of the intranuclear availability of pre-existing and newly synthesized NFB after initiating the cycle. Conformingly, results of in vivo xenograft studies provide evidence that TNF␣-dependent feedback mechanism plays a definite role in the sustained induction of NFB in NB. As shown in supplemental Fig. S3, we outline a pathway in which NFB activation leads to TNF␣-dependent MAPK activity, resulting in NFB activation through the IB kinase phosphorylation of IB␣.
Constitutive activation of cell survival signaling pathways is a general mechanism underlying tumor development and resistance to therapy and constitutes a major clinical problem in cancer. Disruption of aberrantly regulated survival signaling mediated by NFB has recently become an important task in the therapy of several chemoresistant and radioresistant cancers (15). IAPs are expressed at high levels in many tumors and have been reported to contribute to the resistance of cancers to therapy including resistance to radiotherapy (45). In the current study, we provide evidence that clinical doses of IR signif-icantly induced IAPs and Survivin in NB cells. Furthermore, our results showed a good correlation in the pattern of increased IAPs and Survivin to that of induced NFB activity in response to IR. More importantly, our results obtained from two independent strategies demonstrated that IR-induced NFB mediates the expression of these pro-survival proteins. Because inhibition of caspases by IAPs occurs at the core of the apoptotic machinery (45), therapeutic modulation of IAPs could target a key control point in deciding cell fate. To that end, there is mounting evidence that IAPs determine sensitivity to radiotherapy in human cancers (31, 46 -48). Inhibition of Survivin or IAP using antisense oligonucleotides was shown to enhance the efficacy of radiotherapy by reducing survival and increasing apoptosis of lung cancer cells (46,47). NFB enhances cell survival by switching on cIAP-1, cIAP-2, XIAP, the FLICE inhibitory protein cFLIP, and members of the Bcl-2 family (Bcl-XL and A1/Bfl-1), as well as TNF receptor associated factors 1 and 2, which dampen pro-apoptotic signals and attenuate the apoptotic response to anticancer drugs and IR (49,50). Conformingly, we showed that IR-induced NFB is both necessary and sufficient to activate prosurvival proteins and to limit IR-induced cell death in NB cells. It is tempting to speculate that its sustained and pro-survival activity in NB explains the direct correlation between NFB expression and unfavorable outcomes in patients receiving radiochemotherapy. It is intriguing that TNF␣ is generally understood to promote tumor cell death but to possess pro-survival role in NB. These results support the need to define the function of these signaling proteins (in this case TNF␣) according to disease and cell type before applying therapeutic strategies that target these proteins.
With respect to NB, for the first time, our data suggest that clinical doses of IR induce NFB, which initiates a TNF␣-dependent positive feedback effect that in turn maintains the elevated levels of NFB activity for an extended period. This persistent activation of NFB mediates the IR-induced expression of the cell death inhibitors that enhance the cellular proliferation and survival advantage, thereby limiting the therapeutic potential of IR. Therefore, therapeutic measures that selectively target NFB-TNF␣ link and disrupt the IR-induced PFC may mitigate local failure of NB control after radiotherapy.