Metastatic renal cell carcinoma (RCC) remains refractory to therapies. The nuclear factor-κB (NF-κB) transcription factor is involved in cell growth, cell motility, and vascularization. We evaluated whether targeting NF-κB could be of therapeutic and prognostic values in human RCC. The activation of the NF-κB pathway in human RCC cells and tumors was investigated by Western blot. In vitro, the effects of BAY 11-7085 and sulfasalazine, two NF-κB inhibitors, on tumor cell growth were investigated by cell counting, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide analysis, terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling, and fluorescence-activated cell sorting. Their specificity toward NF-κB was analyzed by Western blot, confocal microscopy, NF-κB small interfering RNA, and NF-κB transcription assay. In vivo, the effects of BAY 11-7085 on the growth of human RCC tumors were investigated in nude mice. A tissue microarray (TMA) containing 241 cases of human RCC with 12 to 22 years of clinical follow-up and corresponding normal tissues was built up to assess prognostic significance of activated NF-κB. NF-κB is constitutively activated in cultured cells expressing or not the von Hippel-Lindau (VHL) tumor suppressor gene as a consequence of Akt kinase activation and in tumors. In vitro and in vivo NF-κB inhibition blocked tumor cell growth by inducing cell apoptosis. On the TMA, NF-κB activation was correlated with tumor dimension but was not found to be an independent prognostic factor for patient survival. This report provides strong evidence that the mechanisms responsible for the intrinsic resistance of RCC cells to apoptosis converge on NF-κB independently of VHL expression and that targeting this pathway has great anticancer potential. [Cancer Res 2007;67(24):11668–76]
Renal cell carcinoma (RCC) accounts for 3% of adult malignancy and >90% of adult renal neoplasms. RCC is among the first 10 leading cause of cancer-related death worldwide (1). Metastatic RCC is resistant to radiotherapy and to systemic therapy (2). However, it should be stressed that recent advances in understanding the biology of human RCC have led to novel targeted therapeutic approaches with higher response rates, especially inhibitors of tyrosine kinase receptors such as sunitinib or sorafenib that have been approved by the Food and Drug Administration for the treatment of advanced kidney cancer (3, 4). Unfortunately, the clinical response to these agents is limited in time due to the development of tumor resistance by still unknown mechanisms. New therapeutic options, including combination regimens, have still to be uncovered.
Biallelic inactivating mutations of the von Hippel-Lindau (VHL) tumor suppressor gene occur in patients with the VHL syndrome and in most patients with sporadic RCC (5, 6). The VHL gene products are involved in the degradation of hypoxia-induced transcription factors (HIFs), leading to the down-regulation of several angiogenic and growth factors, such as vascular endothelial growth factor (VEGF) and transforming growth factor-β, which contribute to RCC tumorigenesis (7).
The nuclear factor-κB (NF-κB) transcriptional pathway is involved in many fundamental biological processes, including immunity, inflammation, angiogenesis, cell migration, cell proliferation, and apoptosis (8). In resting conditions, NF-κB is maintained in an inactive state in the cytoplasm through binding to the endogenous inhibitor IκBα. On stimulation by G protein–coupled receptors or tyrosine kinase receptor ligands, the IκB kinase (IKK) tripartite complex may be activated, leading to IκBα phosphorylation and degradation by a proteasome-mediated process. This releases NF-κB from the negative regulation of IκB, exposing its nuclear localization signal and allowing it to translocate to the nucleus where it exerts its transcriptional activity (8). The activation of the IKK complex may be achieved through different ways, including phosphorylation by mitogen-activated protein kinase (MAPK) or Akt (8). The classic pathway is responsible for inhibition of programmed cell death in most conditions (8, 9).
RCC is characterized by a high resistance to tumor cell apoptosis both intrinsic and induced by radiation or systemic therapies, including chemotherapy and immunotherapy. The mechanisms of this resistance are not elucidated. Recent data in other tumors, including tumors from the pancreas, bladder, ovary, breast, or lung, suggest that survival signaling pathways, such as the phosphoinositide 3-kinase (PI3K)/Akt or MAPK-extracellular signal-regulated kinase 1/2 pathways, are involved in the resistance of these tumors to current chemotherapies using apoptotic compounds (10–12). The results obtained by various investigators have ruled out the involvement of multidrug resistance gene in human RCC resistance (2, 13). We have recently shown the critical role played by the PI3K/Akt pathway in human RCC growth (14). In this study, this pathway was found to be constitutively activated and found to promote inhibition of tumor cell apoptosis both in vitro and in vivo. However, the downstream target(s) of Akt responsible for its survival effect was not investigated. In their recent study, Oka et al. (15) have shown that inhibition of NF-κB phosphorylation/activation with parthenolide slows down RCC tumor (OUR-10 cells) growth in nude mice through induction of tumor cell apoptosis. This study is interesting because the activity of NF-κB is regulated by Akt.
The present study was conducted to characterize the intracellular pathways involved in human RCC tumorigenesis and to identify molecular targets that might be used to design efficient, targeted, and safe therapies for this refractory disease. We found that the NK-κB signaling pathway plays a fundamental role in promoting growth and that the intrinsic resistance of RCC to cell apoptosis converges on NF-κB independently on the VHL status. Targeting NF-κB or one of its downstream target genes specific for RCC may thus constitute potential targets for therapeutic intervention.
Materials and Methods
Cells and Cell Culture
Human clear RCC cell lines either deficient in VHL (786-0, UOK-126, UOK-128, and A498) or expressing VHL (ACHN, Caki-1, and Caki-2) were maintained in DMEM (Invitrogen) supplemented with 10% fetal bovine serum and used at 80% confluence, unless otherwise specified. These cell lines were obtained from the American Type Culture Collection, except UOK-126 and UOK-128 that were generously given by Dr. P. Anglard (Institut National de la Sante et de la Recherche Medicale U575, Centre de Neurochimie, Strasbourg, France; ref. 16).
Human Tumor Biopsies
The tumor and normal corresponding tissues of three patients with sporadic RCC were obtained from the Department of Urology, University Hospital of Strasbourg (Strasbourg, France). Informed consent was obtained from all patients. The tumors were staged pT3bN0M0 according to tumor-node-metastasis (TNM) classification (17, 18). Immediately after surgical resection, tissues were fresh frozen and kept in liquid nitrogen until protein expression analysis.
Small Interfering RNA Transfection
Small interfering RNA (siRNA) duplexes specific for human NF-κB p65 subunit and control nonsilencing siRNA were obtained from Ozyme (Cell Signaling local distributor). Transient transfection of RCC cells in 25 cm2 plates (20,000/mL) was performed according to the manufacturer's instructions.
Western Blot Analysis
Tumor tissue or whole-cell lysates were prepared in lysis buffer (14, 19). Protein concentrations were determined according to the method of Lowry et al. (20). The membranes were incubated for 24 h at 4°C with 1:250 dilution of the following primary antibodies (Ozyme): polyclonal rabbit anti-NF-κB p65 antibody, polyclonal rabbit anti-phospho-NF-κB p65 (S536) antibody, polyclonal rabbit anti-IκBα antibody, and monoclonal mouse anti-phospho-IκBα (S32/36; 5A5) antibody. For visualization of protein gel loading, a monoclonal mouse anti-β-actin antibody (Sigma-Aldrich) was used at 1:5,000 dilution. The appropriate horseradish peroxidase–conjugated secondary antibody was used (14, 19). Immunoreactivity was visualized with the enhanced chemiluminescence Western blotting detection kit (Amersham). Phospho-NF-κB to total NF-κB ratios and phospho-IκB to total IκB ratios were calculated using Adobe Photoshop version 7.0 analysis software.
Cell Proliferation Measurements
RCC cell proliferation was assessed by counting adherent cells as described (14, 19). RCC cells were seeded in 24-well plates (20,000/mL), grown for 48 h, and then treated for 48 h with various concentrations of the NF-κB inhibitors BAY 11-7085 [(E)3-[(4-t-butylphenyl)sulfonyl]-2-propenenitrile; 0–30 μmol/L; VWR International] or sulfasalazine (0.1–10 mmol/L; Sigma-Aldrich).
Cell Viability Measurements
RCC cell viability was assessed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Sigma-Aldrich) in cells treated as above (14, 21).
Terminal Deoxynucleotidyl Transferase–Mediated dUTP Nick End Labeling Staining
RCC cells were seeded in four-well Tissue-Tek chamber slides (20,000/mL), grown for 48 h, and then treated with either NF-κB inhibitor or DMSO alone (control) for 24 or 48 h. RCC cell death was assayed as described (14) by the terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) method. Total and stained cells in 10 fields (0.25 mm2 each) were counted and cell death was expressed as a percentage of stained cells to total cells.
Fluorescence-Activated Cell Sorting Analysis
RCC cells were seeded in 25 cm2 plates (20,000/mL) and treated as above. Fluorescence-activated cell sorting (FACS) analysis was performed exactly as described (14) using Annexin V-FITC and propidium iodide, except for sulfasalazine experiments that were performed using Annexin V alone because sulfasalazine has the same color than propidium iodide in solution. To ensure that cell death was due to cell apoptosis, we also performed DNA fragmentation assay in cells treated with sulfasalazine. In some experiments, the cell-permeable pan-caspase inhibitor Boc-Asp(Ome)-fluoromethylketone (B-D-FMK; Enzyme Systems Products) was used. In this case, RCC cells were treated for 48 h with 20 μmol/L BAY 11-7085 or 0.2 mmol/L sulfasalazine in the presence or absence of 40 μmol/L B-D-FMK.
DNA Fragmentation Assay
Experiment was performed as described previously (19).
RCC cells were plated in four-well chamber slides (20,000/mL) for 48 h and then treated with either NF-κB inhibitor or DMSO alone (control) for 8 h. Tumor cells were fixed with paraformaldehyde at 4% for 20 min at 4°C and then permeabilized with Triton X-100 at 0.5% for 5 min. Slides were saturated in bovine serum albumin (BSA) at 3% for 30 min and incubated with monoclonal mouse anti-NF-κB p65 (Tebu-Bio, Santa Cruz Biotechnology local distributor) and polyclonal rabbit anti-phospho-NF-κB p65 (S276; Abcam) primary antibodies diluted at 1:50 and 1:25, respectively, in 1.5% BSA at 4°C overnight in a wet atmosphere. Slides were then washed thrice for 10 min in PBS and incubated with polyclonal mouse and rabbit secondary antibodies coupled, respectively, to A555 and A488 (Invitrogen) diluted at 1:500 in 1.5% BSA at room temperature for 1 h in the dark. Slides were washed twice for 10 min in 1× PBS and once for 10 min in bidistilled water and mounted with Aquamount and analyzed by confocal microscopy (Zeiss LSM 510 inverted microscope).
NF-κB Transcriptional Activity Measurement
It was assayed using the nonradioactive NF-κB p50/p65 transcription factor assay (Chemicon) according to the manufacturer's protocol.
RCC Tumor Model
Tumor implantation and growth. All animal studies were in compliance with the French animal use regulations. Xenografts with 10 millions of 786-0 or Caki-1 cells were performed on twenty 7-week-old male Swiss nu/nu nude mice each (Iffa-Credo; ref. 14). One group was treated with BAY 11-7085 (one injection daily of 5 mg/kg, i.p.) dissolved in DMSO/PBS (diluent, 1:1, v/v). The other group was treated with the diluent alone (control). The average volume of the tumors was 155.56 ± 38.34 mm3 (control) and 160.36 ± 36.05 mm3 (BAY 11-7085; not significant) for 786-0 tumors and 117.03 ± 36.76 mm3 (control) and 125.42 ± 47.11 mm3 (BAY 11-7085; not significant) for Caki-1 tumors. The tumor volumes and animal weight were measured twice weekly during the treatment period.
The tumors were harvested, paraffin embedded, and cut in 4-μm-thick sections for subsequent immunohistochemical analysis (19). A tissue microarray (TMA) was built up by taking a 1-mm cylindrical sample from each tumor. To assess the possible relationship between Akt (14) and NF-κB activations, we built up a TMA including samples from tumors grown in mice treated with the PI3K/Akt inhibitor LY294002 (14). Staining with polyclonal rabbit anti-NF-κB p65 and anti-phospho-NF-κB p65 (S536) antibodies at 1:250 dilution was performed.
Immunohistochemistry. TMA sections were processed as described (14) using the following primary polyclonal rabbit antibodies obtained from Ozyme: polyclonal rabbit anti-phospho Akt (S473) and anti-phospho NF-κB p65 (S276) at 1:50 dilution and polyclonal rabbit anti-phospho-glycogen synthase kinase (GSK)-3β (S9) at 1:25 dilution. Endometrial and prostatic adenocarcinoma samples were incorporated in the TMA and used as positive tissue controls in all TMAs used. Negative control reaction was performed by omitting the primary antibody.
The scoring was based on cytoplasmic staining, nuclear staining, or both for phospho-GSK-3β, phospho-NF-κB, and phospho-Akt expression, respectively. Staining intensity was graded as weak, moderate, intense, or negative (1+, 2+, 3+, or 0) and plotted as shown in the corresponding figures.
Proliferative and apoptotic index. Both indexes were determined as described (14) using a mouse monoclonal anti-human Ki67 antibody (Mib-1; Dako) and the TUNEL method (Roche Diagnostics), respectively. Stainings were quantified in a blinded fashion by an experienced urologic histopathologist (V.L.).
Neovascularization. Tumor sections were stained for endothelial cells with a rabbit polyclonal anti-human factor VIII antibody (Dako) using a standard immunohistochemical method described previously (19). Quantifications of both vessel intersecting points and the total number of vessels were performed as described previously (14, 19, 22).
TMA of Human RCC
Patients and clinical data. From January 1980 to December 1990, 255 RCC patients were subjected to radical nephrectomy in the Department of Urology of the University Hospital of Strasbourg. The tumors were staged according to TNM classification (17, 18) and ranged from pT1a to pT3bN0M0. The characteristics of the population are summarized in Supplementary Table S1. Four patients were lost during follow-up. Death occurred in 141 cases.
In a multicentric and retrospective study, these tumors were graded according to the Fuhrman's classification by three independent urologic histopathologists (23).
TMA and immunohistochemistry. RCC tumor biopsies were fixed in buffered formalin and paraffin embedded. For each patient, normal kidney tissue was available in the renal biopsies. After selection of morphologically representative regions of individual paraffin-embedded renal tumors, core biopsies of 1-mm-diameter tumor and corresponding normal tissue were taken and transferred to a recipient paraffin block using a custom-built microarrayer. Sections (4 μm thick) were prepared and used for subsequent immunohistochemical analysis.
Sections were processed as described above and using a polyclonal rabbit anti-phospho-NF-κB (S276; Ozyme) at 1:50 dilution in PBS-Tween 20 buffer overnight at 4°C. Standard indirect immunoperoxidase procedures were used for visualization.
The scoring was based on nuclear staining and scored as described above. The distribution was graded in percentage (%) of stained cells among the total number of cells.
All values are expressed as mean ± SEM. Values were compared using multifactorial ANOVA followed by the Student-Newman-Keul's test for multiple comparisons. P < 0.05 was considered significant.
For analysis of the patient TMA, overall survival was defined as the time between nephrectomy and patient death or censoring. Nonparametric Mann-Whitney and Kruskal-Wallis test were performed to assess significativity of activated NF-κB staining with patient characteristics (i.e., sex, Fuhrman grade, microvascular invasion, tumor dimension, and TNM stage). Correlation coefficient was computed using Spearman test. To assess whether activated NF-κB expression is associated with poor patient prognosis, we used Kaplan-Meier (with log-rank test) method. Multivariate prognosis analysis was performed using Cox proportional hazard model. Tumor stage and grade were used as variables in this analysis. Computations were done using Statistical Package for the Social Sciences 13.0. A P value of <0.05 was considered significant.
NF-κB expression in human RCC tumors treated with the PI3K/Akt inhibitor LY294002. We have previously shown that the PI3K/Akt pathway is constitutively activated in human RCC both in vitro and in vivo (14) and NF-κB is activated through Akt-dependent IKK phosphorylation. In this previous studies, we identified GSK-3β as being regulated by Akt. Using the TMA, phospho-Akt (S473) was substantially decreased by >30% (Fig. 1A). We confirmed the Akt-dependent GSK-3 phosphorylation and inactivation (Fig. 1B). Now, we extend this finding by showing that the NF-κB transcription factor is also a downstream target of Akt in human RCC (Fig. 1C). Representative immunostainings are shown in Supplementary Fig. S1A to C, respectively.
NF-κB and IκB expression in human RCC tumor biopsies. The activation state of NF-κB and IκB was thus investigated in human RCC tumors compared with corresponding normal kidney tissues. For that, we analyzed both NF-κB p65 subunit and IκB expression as well as their phosphorylation status in tumor and normal tissue protein samples. Both proteins are expressed in the nonphosphorylated state at a similar level in tumors as well as in normal corresponding tissues (Fig. 2A). In contrast, high levels of phosphoproteins are found in tumors compared with corresponding normal tissues (Fig. 2A), indicating that NF-κB is constitutively activated in human RCC tumors. The ratios of phospho-NF-κB to total NF-κB were 0.25, 0.14, 0.05, 0.43, 0.9, and 0.23 for N1, N2, N3, T1, T2, and T3, respectively. The ratios of phospho-IκB to total IκB were 0.76, 0.75, 0.75, 0.85, 0.92, and 1.09 for N1, N2, N3, T1, T2, and T3, respectively.
NF-κB and IκB expression in human RCC cell lines. A panel of human RCC cell lines either expressing the VHL tumor suppressor or deficient in VHL was used to analyze the expression of both NF-κB p65 subunit and IκB in their nonphosphorylated or phosphorylated states and the possible relationship with the VHL status. No differences in phosphorylation states or in the levels of expression of both proteins were observed, indicating a constitutive activation of NF-κB regardless of the VHL status (Fig. 2B). The ratios of phospho-NF-κB to total NF-κB were 0.65, 0.63, 0.74, 1.01, 1.01, 1.01, and 0.86 for 786-0, UOK-126, UOK-128, A498, ACHN, Caki-1, and Caki-2 cells, respectively. The ratios of phospho-IκB to total IκB were 0.7, 0.37, 0.38, 1.0, 0.44, 0.79, and 1.1 for 786-0, UOK-126, UOK-128, A498, ACHN, Caki-1, and Caki-2 cells, respectively.
Effect of NF-κB inhibitors on cultured RCC cell growth and death. We used two NF-κB inhibitors, BAY 11-7085 and sulfasalazine, to investigate the involvement of NF-κB activation on RCC growth. Both inhibitors act by maintaining IκB bound to NF-κB, thus inhibiting the activation and nuclear translocation of NF-κB (24, 25). We used 786-0 and Caki-1 cells, respectively deficient in normal VHL tumor suppressor gene products or expressing them, in further experiments. BAY 11-7085 decreased cell density by up to 100% in a concentration-dependent manner. No difference in the effectiveness of the inhibitors was noted with the VHL status of the cells (Supplementary Figs. S2A and S3A). Testing cell viability by MTT (Supplementary Figs. S2B and S3B) and cell death by TUNEL staining (Fig. 3A; Supplementary Fig. S4A) and FACS analysis (Fig. 3B; Supplementary Fig. S4B) in response to either inhibitor in both cell lines strongly suggests that the effects of the inhibitor are achieved through induction of cell apoptosis independently of VHL expression. In addition, DNA laddering characteristic of cell apoptosis was observed in response to the inhibitors (Supplementary Fig. S4C). To further ensure that the effects of both NF-κB inhibitors are achieved through induction of cell apoptosis, the effects of either NF-κB inhibitor were also measured by FACS in the presence of B-D-FMK (26), which reduced significantly the apoptotic effects of both NF-κB inhibitors (Fig. 3B; Supplementary Fig. S4B). The NF-κB pathway thus seems to be turned toward cell survival in human RCC.
Specificity toward NF-κB of the effects of the inhibitors. To ascertain that the apoptotic effects of BAY 11-7085 and sulfasalazine were obtained through inhibition of the NF-κB pathway, we followed three experimental procedures.
The effect of both inhibitors on NF-κB p65 phosphorylation was then studied in 786-0 and Caki-1 cells by treating them for 0 min, 30 min, 1 h, 2 h, 5 h, 8 h, 24 h, and 48 h with a maximally effective concentration of either inhibitor (i.e., BAY 11-7085 at 20 μmol/L and sulfasalazine at 0.2 mmol/L). Whereas unphosphorylated proteins remained unchanged during this time course, phospho-NF-κB p65 (S256) was rapidly undetectable (30 min in cells treated with BAY 11-7085 and 1 to 2 h in cells treated with sulfasalazine) and returned to basal value only after 24 to 48 h of treatment (Fig. 4A; Supplementary Fig. S5).
The effect of a maximally effective concentration of either inhibitor on phospho-NF-κB p65 (S276) expression and subcellular localization was also assessed by confocal microscopy. In 786-0 cells treated in control, phospho-NF-κB p65 was observed in the nucleus of virtually all cells (Fig. 4B), argumenting in favor of the constitutive activation of NF-κB in human RCC. After treatment with BAY 11-7085, phospho-NF-κB p65 was no longer detected. Similar results were obtained in Caki-1 cells and in cells treated with sulfasalazine (data not shown).
Finally, NF-κB p65 was specifically knock down in RCC cells using specific siRNA. Apoptosis was then assessed by FACS analysis in the presence or absence of a maximally effective concentrations of either inhibitor, as above. In 786-0 cells transfected with p65-specific siRNA, p65 expression was significantly reduced, whereas no difference was noted between untransfected cells and cells transfected with transfection reagent. In 786-0 cells treated with BAY 11-7085, we obtained similar results as the ones shown in Figs. 3B and 4C. The transfection with p65-specific siRNA at the concentration used (100 nmol/well) alone induced apoptosis by ∼20%, further argumenting the critical role of NF-κB in tumor cell survival (Fig. 4C). The efficiency of BAY 11-7085 to induce tumor cell apoptosis was substantially decreased in cells transfected with p65-specific siRNA by 30% (Fig. 4C). Because similar results were obtained with sulfasalazine in 786-0 cells and in Caki-1 cells with both inhibitors.
Taken together, these results strongly suggest that the apoptotic effects of NF-κB inhibitors were attributable to NF-κB inhibition.
Transcriptional activity of NF-κB. NF-κB activity was evaluated using a NF-κB p50/p65 transcription factor kit as detailed in Materials and Methods in resting cells and in cells treated with either inhibitor at a maximally effective concentration. Because the effects of both inhibitors on NF-κB phosphorylation were maximal after 30 min of exposure and sustained for at least 8 h (Fig. 4D), we choose to measure NF-κB activity in cells treated for 2 h with either inhibitor. NF-κB is constitutively activated in human RCC, and this activity is inhibited by 40% to 50% by BAY 11-7085 and sip65 in both cell lines (Fig. 4D). Similar results were obtained with sulfasalazine (data not shown). No significant difference was observed in 786-0 versus Caki-1 cells. Thus, tumor cell apoptosis mediated by NF-κB inhibitors is achieved through inhibition of NF-κB transcriptional activity.
Effects of NF-κB inhibition on RCC tumor growth in vivo. The treatment of xenograft athymic mice with BAY 11-7085 (5 mg/kg) inhibited significantly tumor growth by ∼80% (Fig. 5). This concentration of the NF-κB inhibitor was chosen from previous studies by other investigators in nude mice (27, 28). No difference was observed in the overall efficiency of the NF-κB inhibitor with the VHL status of the implanted tumors. In some mice, partial regression was observed following the treatment.
All the mice maintained body weight, and no difference was observed between mice treated in control or with BAY 11-7085 (Supplementary Table S2). The treatment seems to be well tolerated; indeed, plasma concentrations of electrolytes, creatinine, albumin, and urea showed no difference between mice bearing 786-0 or Caki-1 tumors treated in control or with BAY 11-7085 (Supplementary Table S2).
Immunohistochemical analysis of tumors revealed that BAY 11-7085 exerts its antitumoral effect through induction of tumor cell apoptosis (786-0 tumors: 1.4 ± 0.2% versus 2.4 ± 0.3% of TUNEL-stained cells in tumors of control-treated and BAY 11-7085–treated mice, respectively; n = 8; P < 0.05; Caki-1 tumors: 1.2 ± 0.3% versus 2.5 ± 0.3% of TUNEL-stained cells in tumors of control-treated and BAY 11-7085–treated mice, respectively; n = 7; P < 0.05; Fig. 6A), confirming the in vitro data. No effects were observed on tumor cell proliferation (Fig. 6B) and neovascularization (data not shown) in either tumor.
To assess the efficiency of the treatment on NF-κB activation, a TMA was built up with the tumors harvested from the mice to avoid difference in staining due to the manipulation of many slides. Immunostaining clearly showed a substantial diminution of activated NF-κB in tumors harvested from mice treated by the NF-κB inhibitor compared with tumors harvested from mice treated in control regardless of the VHL status of the tumors (Fig. 6C). This finding confirms appropriate drug targeting in vivo and also the constitutive activation of NF-κB in RCC tumors.
These results strongly suggest that NF-κB and its downstream targets participate in the overall intrinsic resistance of human RCC to tumor cell death and that the treatment by specific inhibitor seems to be safe.
Clinicopathologic tumor variables and activated NF-κB expression in human RCC. Clinicopathologic tumor variables and patient characteristics are detailed in Supplementary Table S1. Different views of the patient TMA are shown in Supplementary Fig. S6A and B. Some normal tissues showed weak activated NF-κB staining that were systematically less than corresponding tumor tissues (data not shown). Analysis of tumor tissues showed that, as already known, sex, age, tumor dimension, and Fuhrman grade are independent prognostic factor for patient survival. Variables that were tested about activated NF-κB are sex, RCC subtype, Fuhrman grade, microvascular invasion, tumor dimension, TNM stage, and death. Among these variables, significativity was obtained for clear RCC having higher expression of activated NF-κB compared with the other types taken together (P = 0.002) and for tumor dimension (P = 0.02). About the significativity with RCC types, expression values were 28.3 ± 2.4 (median 5%) for clear RCC and 6.8 ± 2.7 (median 0%) for the other RCC types. About tumor dimension, the Spearman nonparametric correlation coefficient was −0.153. Activated NF-κB did not appear as an independent prognostic factor for patient survival (Supplementary Fig. S6C).
The NF-κB signal transduction pathway is misregulated in a variety of hematologic and solid tumor malignancies due either to genetic changes, such as chromosomal rearrangements, amplifications, and mutations, or to chronic activation of the pathway. Constitutive activation of the NF-κB pathway can contribute to the oncogenic state in several ways (e.g., by driving proliferation, enhancing cell survival, and/or promoting angiogenesis or metastasis).
At the nonphosphorylated state, NF-κB and IκBα were found at similar levels in tumors compared with normal corresponding tissues, suggesting that DNA duplication or amplification events are not present in human RCC, although such assumption will need to be confirmed by specific genetic studies. However, phosphorylation of both NF-κB and IκBα was only observed in tumors and not in corresponding normal tissues, as well as in cultured tumor cells, suggesting that this pathway is constitutively activated in human RCC. Our results also strongly suggest that the constitutive activation of NF-κB is most essentially the result of the constitutive activation of Akt in human RCC. Thus, besides GSK-3, Akt also regulates NF-κB activity in this tumor type.
To date, and in contrast to other tumor types, the possible involvement of NF-κB in human RCC tumorigenesis has only received little attention, and no NF-κB–regulated genes or set of genes that might be involved in this disease have been described. In addition, in vivo data are missing and the few data available on cultured cells are in most cases difficult to interpret or contradictory (29–33). The VHL gene products have been shown to suppress NF-κB activity in human RCC. However, the role VHL may play in tumor cell survival remains unknown at present in human RCC (32–34). Evidence has been presented that certain tumor suppressors can block NF-κB activation, such as Arf, CYLD, and, in some conditions, p53 (35–37). Our results did not reveal any difference in the efficiency and mechanism of action of the NF-κB inhibitors depending on the VHL status of the cells. The absence of VHL dependency was already observed in our previous studies dealing with the involvement of the PI3K/Akt pathway in RCC growth. Thus, VHL gene products do not seem to control the NF-κB signaling pathway in human RCC, although HIF is regulated by NF-κB (38). Consistent with our findings, Oka et al. (15) have recently shown that inhibition of NF-κB phosphorylation by sesquiterpene lactone parthenolide slows down OUR-10 RCC tumor growth in nude mice through induction of tumor cell apoptosis. However, the VHL status of OUR-10 cells has not been reported. Thus, additional experiments focusing on that particular point are needed to more precisely define the role, if any, of the VHL tumor suppressor gene in the sensitivity of human RCC to NF-κB inhibition.
Cancer-relevant NF-κB–dependent genes include those encoding cytokines and chemokines (such as TNF-α, IL1, IL8, and MCP-1), proliferative regulators (such as cyclin D1), antiapoptotic proteins (such as Bcl-2, Bcl-XL, and IAP), and modulators of invasion and angiogenesis (such as MMP and VEGF; refs. 39, 40). Besides VEGF, which is mainly controlled by the VHL/HIF system in human RCC, there are no consistent data allowing to know whether some of these downstream targets are involved in this disease. Part of our ongoing work using target gene array and two-dimensional difference gel electrophoresis aims at identifying genes controlled by NF-κB in human RCC.
An exciting feature of NF-κB is the demonstration in various reports that this transcription factor is activated in response to chemotherapies and to radiation and that it functions to suppress the apoptotic potential of that cancer therapy in various tumor types (41–43). Some clinical trials using certain chemotherapies in conjunction with NF-κB inhibitors, such as thalidomide, are presently under way to assess whether NF-κB blockade promotes cancer therapy efficacy. Because chemotherapeutic agents, such as daunorubicin or vinblastine, have been shown to activate the NF-κB signaling pathway in human RCC cells in culture, the possibility exists that such association may have therapeutic potential in human RCC as well. Part of our ongoing work explores this possibility in human RCC.
Oya et al. (44) have shown that, out of 45 cases of human RCC they investigated, 15 cases showed an increase of >200% in NF-κB activity compared with corresponding normal renal tissue. Such increase was more often observed in locally advanced cases than in localized cases. They concluded that an increase in NF-κB activity may be related to tumor development. However, the relation, if any, of the increase in NF-κB activity with patient survival was not investigated. In the present study, we used a TMA composed of 241 cases of human RCC that have been harvested between 1980 and 1990. NF-κB activation was found to be higher in clear RCC versus other subtypes and to be related to tumor dimension but did not seem to be an independent prognostic factor for patient survival. We did not find significant differences in NF-κB activation between localized and locally advanced cases. The reason for this apparent discrepancy between the study of Oya et al. (44) and the present study is not known but may be related to the number of cases that were analyzed.
This report provides strong evidence that the mechanisms responsible for the intrinsic resistance of RCC to cell apoptosis converge on NF-κB independently on the VHL status and that NF-κB or its downstream targets have potential therapeutic value in this refractory disease. The identification of these downstream targets is currently on the way in our laboratory.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
Grant support: Institut National de la Sante et de la Recherche Medicale, Université Louis Pasteur of Strasbourg, Strasbourg School of Medicine (T. Massfelder); French Ligue Contre le Cancer, Comités du Bas-Rhin, du Haut-Rhin et Comité National (T. Massfelder); and Association pour la Recherche sur le Cancer (T. Massfelder).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank the Institut de Recherche contre les Cancers de l'Appareil Digestif and Université Louis Pasteur EA 3430 (Dr. F. Raul, Institut de Recherche contre les Cancers de l'Appareil Digestif, Strasbourg, France) for housing nude mice, Dr. Bouissac for help in confocal microscopy, Dr. Crémel for allowing us to perform FACS analysis in his laboratory (Institut National de la Sante et de la Recherche Medicale U575), and F. Reymann (Department of Pathology, Strasbourg University Hospital) for technical assistance in immunohistochemical studies.