NF-κB activation enhances cell death by antimitotic drugs in human prostate cancer cells

Background NF-κB is a transcription factor that promotes inhibition of apoptosis and resistance to chemotherapy. It is commonly believed that inhibition of NF-κB activity can increase sensitivity of cancer cells to chemotherapy. However, there is evidence that NF-κB activation can sensitize cells to apoptosis and that inhibition of NF-κB results in resistance to chemotherapy. In prostate cancer, it is not clear in the different cell types (androgen-dependent and castration-resistant) if activation or inhibition of NF-κB is required for stimulation of apoptosis by chemotherapy. Results Our data indicate that the response of prostate cancer (PC) cells to the antimitotic drugs docetaxel (Doc) and 2-methoxyestradiol (2ME2) is dependent on the levels of NF-κB activity. In androgen-dependent LNCaP cells, Doc and 2ME2 treatment increased the low basal NF-κB activity, as determined by Western blot analysis of phospho-IκBα/p65, NF-κB promoter reporter assays, and p65 localization. Treatment of LNCaP cells with parthenolide, a pharmacologic inhibitor of NF-κB, or introduction of dominant-negative IκBα, or an shRNA specific for p65, a component of the NF-κB heterodimer, blocked apoptosis induced by Doc and 2ME2. In castration-resistant DU145 and PC3 cells, Doc and 2ME2 had little effect on the high basal NF-κB activity and addition of parthenolide did not enhance cell death. However, the combination of Doc or 2ME2 with betulinic acid (BA), a triterpenoid that activates NF-κB, stimulated apoptosis in LNCaP and non-apoptotic cell death in DU145 and PC3 cells. Increased sensitivity to cell death mediated by the Doc or 2ME2 + BA combination is likely due to increased NF-κB activity. Conclusions Our data suggest that the combination of antimitotic drugs with NF-κB inhibitors will have antagonistic effects in a common type of PC cell typical of LNCaP. However, combination strategies utilizing antimitotic drugs with BA, an activator of NF-κB, will universally enhance cell death in PC cells, notably in the aggressive, castration-resistant variety that does not respond to conventional therapies.


Introduction
NF-κB, originally discovered as a transcription factor that regulates the immune system, is now known to be widely expressed in almost all cells and mediates multiple signaling pathways including cell proliferation and survival [1]. A common form of NF-κB is a heterodimer consisting of p65 (RelA) and p50 proteins that exists as an inactive IκBα-bound form in the cytoplasm of unstimulated cells. Activation of the NF-κB pathway by a variety of inducers including cytokines, growth factors, UV light and DNAdamaging drugs often occurs by increasing the phosphorylation of IκBα by the IκB kinase (IKK) complex. This results in ubiquitination and rapid degradation of IκBα by the 26S proteasome, allowing for the increase of nuclear NF-κB DNA binding activity and transcriptional activation of its target genes, including Inhibitor of Apoptosis (IAP) family members XIAP, IAP-1, IAP-2 and anti-apoptotic Bcl-2 family members Bcl-2, Bcl-xL [1]. Therefore, NF-κB activation is often associated with increased survival of cancer cells and resistance to chemotherapy. Accordingly, there are several candidate inhibitors of NF-κB activity that are in development as anti-cancer therapy [2][3][4].
However, the role of NF-κB in cancer progression and in anti-cancer therapeutics is complex, as there is also evidence to suggest that NF-κB activation can sensitize cells to apoptosis [5,6]. For example, inhibition or loss of NF-κB activity blocks p53-mediated apoptosis, suggesting that inhibition of NF-κB in p53 positive tumors may weaken the therapeutic response [7]. Furthermore, acti-vation of NF-κB by UV light and doxorubicin converts it into an active repressor of the anti-apoptotic genes XIAP and Bcl-xL [8]. Additional evidence supporting a proapoptotic role for NF-κB in cancer chemotherapy comes from the observation that the retinoid-related compounds 3-Cl-AHPC and CD437 require activation of NF-κB in order to induce apoptosis in DU145 and PC3 castration-resistant prostate cancer (CRPC) cells [9,10]. Exposure of CRPC cells to 3-Cl-AHPC or CD437 enhances the expression of the pro-apoptotic Death Receptor (DR) 4 and 5 genes. An NF-κB binding site located in intron 1 of the DR5 promoter is important for positive regulation by NF-κB [11]. Activation of NF-κB by betulinic acid (BA), a naturally occurring pentacyclic triterpenoid small molecule with anticancer properties, is also required for induction of apoptosis in tumor cells [12].
The clinical progression of prostate cancer (PC) involves the transition from androgen-dependent cancer, which can be successfully treated with androgen-ablation therapy, to a castration-resistant cancer with few treatment options [13]. One of the critical factors in the progression to CRPC is the increased activity of NF-κB and its promotion of apoptotic inhibition [14][15][16]. It is not clear in the different types of PC cells (androgen-dependent and castration-resistant) whether activation or inhibition of NF-κB is required for stimulation of apoptosis by chemotherapy. Since PC consists of a heterogeneous mixture of cell types, it is important to better understand the mechanisms of the effect of chemotherapy on NF-κB activity in different PC cell lines in order to increase therapeutic response [17,18].
In this report, we investigate the effects of the antimitotic drugs docetaxel (Doc) and 2-methoxyestradiol (2ME2) on NF-κB activity and induction of cell death in androgen-responsive and castration-resistant PC cell lines. Doc is now one of the most effective anti-cancer drugs and is FDA approved for the treatment of prostate, breast, gastric, head and neck, and non-small cell lung cancers [19,20]. 2ME2, an endogenous metabolite of estradiol, is being investigated in clinical trials as an anticancer agent [21]. Both Doc and 2ME2 have been reported to increase NF-κB activity in tumor cell lines including PC, but whether this stimulates or antagonizes apoptosis appears to be dependent on the specific tumor cell type [22,23].
Studies have shown constitutive NF-κB DNA binding and transcriptional activity in DU145 and PC3 CRPC cells but not in androgen-dependent LNCaP cells [24][25][26]. Our data indicated that both Doc and 2ME2 increased NF-κB activity in LNCaP cells and that inhibition of NF-κB was able to block treatment-induced apoptosis. Doc and 2ME2 treatment had little effect on NF-κB activity in DU145 and PC3 cells and the addition of an NF-κB inhib-itor did not stimulate cell death in these cells. In contrast, addition of BA increased NF-κB activity and stimulated Doc-and 2ME2-mediated apoptosis in LNCaP and caspase-independent cell death in DU145 and PC3 cells.

NF-κB Reporter Assay
To measure NF-κB transcription activity, we used a plasmid containing the luciferase reporter gene regulated by four copies of NF-κB cis-acting elements linked to TATA box from the thymidine kinase promoter (NF-κB-TA/luc; Clontech). TA/Luc is the negative control plasmid without NF-κB elements. Plasmids were co-transfected with CMV/β-galactosidase (Clontech) into LNCaP and PC3 cells using FuGene 6 HD transfection reagent (Roche), as previously described [29]. After 24 h, transfected cells were grown in the presence or absence of 5 μM 2ME2 for 24 and 72 h and the luciferase and β-gal activities determined. The NF-κB luciferase activity relative values as light units/β-gal were divided by TA/luc values and expressed as fold above TA/luc from 3 independent experiments done in duplicate.

DAPI Apoptosis and Trypan Blue Exclusion Assays
The DAPI staining apoptosis assay was done as previously described [28]. Changes in apoptosis in cells treated with drugs were determined as percentage of apoptotic cells (densely stained and fragmented chromatin) from 3-6 independent experiments done in duplicate. Minimal apoptosis was detected in control treated cells (<0.5%). For the trypan exclusion assay, treated and control pros-tate cancer cells were harvested, resuspended in growth media, diluted 1:1 in 0.4% trypan blue, dead blue and live non-blue cells immediately counted using a hemacytometer, and the % dead blue cells determined from at least 3 independent experiments done in duplicate.

Annexin-FITC/Propidium Iodide (PI) Flow Cytometry
For the annexin apoptosis assay, we used the ApoAlert Annexin V-FITC Apoptosis kit (Clontech). LN-AI and DU145 cells were resuspended in binding buffer followed by the addition of annexin V-FITC and PI. After 20 min., cells were analyzed by flow cytometry using a Coulter XL flow cytometer and the percentage of annexin+ and PI+ cells determined using WinMDI version 2.8.

Stable Transfection of Dominant Negative IκBα
To inhibit endogenous NF-κB activity, we obtained the pCMV-IκBαM plasmid (Clontech) expressing dominant negative IκBα containing Ser to Ala mutations at positions 32 and 36, which cannot be phosphorylated and degraded [30]. LN-AI cells (90% confluent) were cotransfected with pCMV-IκBαM and pCMVneo (for drug selection) using FuGene 6 HD following the manufacturer's instructions. The negative control was transfection with pCMVneo alone. Cells were initially grown in media with 400 μg/ml G418 (Invitrogen), colonies selected, and clones that express dominant negative-IκBα (dnI) compared to pCMVneo negative control cells clones were identified by Western blot (migrates faster than endogenous wild type IκBα).

Statistical Analysis
Statistical differences between drug-treated and control PC cells were determined by two-tailed Student's t-test with P < 0.05 considered significant.

Results
To evaluate the effect of the antimitoic drugs 2ME2 and Doc on NF-κB activity in human PC cells, we used androgen-dependent LNCaP and castration-resistant LN-AI, DU145, and PC3 cells. LNCaP and LN-AI cells contain wild-type p53 and exhibit higher sensitivity to 2ME2 and Doc apoptosis relative to DU145 and PC3, which are p53 mutated or null and thus are more resistant to apoptosis [27,28,[35][36][37].

Antimitotic Drugs Activate NF-κB in LNCaP Cells
To determine the effect of 2ME2 and Doc on NF-κB, we analyzed phosphorylation of IκBα and of p65 at the serine 536 position, which can enhance the transcriptional activity of NF-κB and lower its affinity to IκBα [1]. Western blot analysis of LNCaP cells treated with 5 μM 2ME2 for 24 h resulted in increased phospho-IκBα (decrease in total IκBα) and phospho-p65 (no change in total p65) relative to control treated cells (Fig. 1A). In DU145 and PC3 cells, which unlike LNCaP cells have constitutively active NF-κB, there were no differences in phospho-IκBα or  phospho-p65 in 2ME2-treated cells relative to control cells (Fig. 1A). In addition, total IκBα protein decreased in 2ME2 or Doc-treated LNCaP and LN-AI cells but not in DU145 cells (Fig. 1B).
To determine the effect of 2ME2 on NF-κB transcriptional activity, we used a plasmid containing the luciferase reporter regulated by NF-κB cis-acting elements. The results showed that 2ME2 increased NF-κB activity ~2-fold in LNCaP cells at 24 and 72 h compared to control treated cells (Fig. 1C). In contrast, 2ME2 slightly decreased NF-κB activity 1.5-fold in PC3 cells at 24 h but not at 72 h. The results also showed a 7.3-fold higher basal NF-κB activity in PC3 compared to LNCaP cells.

Nucleolar Localization of p65 in LNCaP cells Treated with Anitmitotic Drugs
We used fluorescence immunocytochemistry to determine if p65 localizes to the nucleus after 2ME2 treatment of LNCaP cells. In control cells, p65 localized in the cytoplasm but after 24 h treatment with 5 μM 2ME2, some cells appeared to have p65 localization to the nucleolus, similar to a previous report [38]. Double fluorescence immunocytochemistry confirmed the presence of several cells with localization of p65 to the nucleolus, as revealed by merged images of the nucleolar marker nucleolin (Texas Red) with p65 (FITC), and DAPI nuclear stain (blue) (Fig. 2). There was also co-localization of p65 and nucleolin in the cytoplasm (yellow), likely due to the disruption of the nucleolus and nuclear membrane by antimitotic drugs. The nucleolus is normally disassembled during mitosis and reassembled after cell division [39]. These results suggest that in LNCaP cells treated with antimitotic drugs, p65 can localize to the nucleolus, which has previously been shown to be important in increasing apoptosis in colon cancer cells treated with aspirin [38].

NF-κB Inhibition Blocks Apoptosis Induced by Antimitotic Drugs
To determine whether 2ME2-or Doc-mediated activation of NF-κB in LNCaP cells is important for stimulating apoptosis, we used the NF-κB inhibitor parthenolide [40]. Treatment of LNCaP cells with 10 μM parthenolide lowered apoptosis induced by 2ME2 or Doc, as determined by the DAPI apoptosis assay and decreased levels cleaved PARP protein (Fig. 3). Parthenolide lowered the 2ME2-or Doc-mediated increase in phospho-IκBα and phospho-p65, suggesting inhibition of NF-κB activity. These results indicate that 2ME2-or Doc-mediated increase in NF-κB activity is important for induction of apoptosis in LNCaP cells. Similar results were obtained in LN-AI cells (result not shown).
To further investigate molecular changes involved in why inhibition of NF-κB reduced 2ME2-or Doc-mediated apoptosis, we analyzed the expression of p53 and XIAP. p53, the most commonly mutated gene in human cancers, can mediate the apoptosis response to chemotherapy [41]. Overexpression of IAP family members such as XIAP blocks apoptosis and increases drug resistance [42]. Similar to our previous results, 2ME2 and Doc increased p53 and decreased XIAP proteins in LNCaP In control LNCaP cells, p65 (green) is cytoplasmic, nucleolin (red) is in the nucleolus, DAPI (blue) stains the nucleus, and the merge of p65/nucleolin/DAPI shows no co-localization of p65 and nucleolin or DAPI. In 2ME2 treated LNCaP cells, p65 remains mostly cytoplasmic but some is localized within the nucleus; nucleolin is localized to the cytoplasm and some remains in the nucleolus or nucleus; DAPI shows mitotic cells. The merge of p65/nucleolin/DAPI shows some cells (arrows 1) with co-localized p65/nucleolin within the nucleolus (white) and within the cytoplasm (yellow). Also shown is a cell (arrow 2) with no co-localization of p65 and nucleolin. (×200). and LN-AI cells [28,35]. However, parthenolide blocked the 2ME2 and Doc-induced changes in p53 and XIAP relative to control levels (Fig. 3). These results suggest that the 2ME2-or Doc-mediated increase in NF-κB activity correlates with increased p53 and decreased XIAP, conditions that favor the induction of apoptosis.

Dominant Negative IκBα and p65 Knockdown Inhibit 2ME2 and Doc Apoptosis
To further determine if there is a role for activation of NF-κB in 2ME2-or Doc-mediated apoptosis, we isolated  [30]. LN-AI/dnI clones 7 and 20 and negative control LN-AI/neo cells were treated with 2ME2 or Doc for 48 h and the effect on apoptosis determined by DAPI and cleaved PARP protein levels. Results showed decreased apoptosis in LN-AI/dnI-7 and 20 compared to LN-AI/neo cells (Fig. 4). As expected, there was an increase in endogenous phospho-IκBα in all cells treated with 2ME2 or Doc for 24 h but phospho-dnI was not detected (faster migrating band in total IκBα blot; Fig.  4B). Similar results were obtained in LNCaP cells stably expressing dnI (result not shown) or shRNA targeting p65 (Fig. 4C). These results further support an important role for NF-κB activation-mediated increase in apoptosis by treatment with 2ME2 or Doc in LN-AI and LNCaP cells.

Betulinic Acid (BA), an NF-κB Activator, Stimulates Cell Death in All Prostate Cancer Cells Treated with Antimitotic Drugs
Our results showed that the NF-κB inhibitor parthenolide antagonized apoptosis mediated by antimitotic drugs in LNCaP and LN-AI cells. In DU145 or PC3 cells, parthenolide had little effect on 2ME2-or Doc-mediated cell death, as shown by the trypan blue exclusion assay (Fig.  5). Thus, we determined whether addition of an NF-κB activator could increase apoptosis by antimitotic drugs in all types of PC cells. BA is a natural pentacyclic triterpenoid proposed to activate NF-κB [12]. Combination of 2ME2 or Doc with 10 μM BA increased apoptosis in LN-AI cells compared to the single drugs, as determined by DAPI, cleaved PARP protein levels, and annexin V-FITC/ PI flow cytometry (Figs. 6, 7). BA further increased phospho-IκBα and phospho-p65 and decreased total IκBα compared to 2ME2 or Doc alone in LN-AI and DU145 cells, suggesting enhancement of NF-κB activity (Fig. 6). Similar results were obtained in LNCaP cells (results not shown). In addition, BA increased NF-κB DNA binding in LNCaP and DU145 cells as determined by EMSA (result not shown). In DU145 and PC3 cells, despite increased fragmented and condensed nuclei (DAPI), BA +2ME2 or Doc did not increase cleaved PARP (substrate for active caspase) ( Fig. 6 and not shown), suggesting caspase-independent cell death. Further supporting a role for activation of NF-κB in the 2ME2/Doc + BA combination is the observation that the LN-AI/dnI clones underwent less apoptosis compared to the LN-AI/neo negative control cells (Fig. 8A). Knockdown of p65 in LNCaP and DU145 also lowered cell death and cleaved PARP (Fig. 8B). In addition, pretreatment of DU145 cells with parthenolide (10 μM) for 24 h followed by 2ME2/Doc + BA/parthenolide lowered cell death compared to DU145 cells treated without parthenolide (Fig. 8C). Furthermore, PC cells treated with BA alone or in combination with 2ME2 or Doc decreased IκBα and XIAP protein levels (Fig. 6). These results suggest that enhancement of NF-κB activity by BA plays a role in increasing cell death in PC cells treated with antimitotic drugs.
Because cleaved PARP levels are not increased in DU145 or PC3 cells treated with the 2ME2/Doc + BA combination, we assessed whether there were any differences in total cell death by the trypan blue exclusion assay. Results indicated a greater extent of cell death in the 2ME2/Doc + BA combination compared to 2ME2, Doc, or BA treatment alone (Fig. 8D). This result further suggests that BA increases caspase-independent cell death in DU145 and PC3 cells.

Apoptosis-Inducing Factor (AIF) Increases Cell Death by 2ME2/Doc + BA
To investigate potential downstream effectors of increased cell death mediated by BA, we analyzed the expression of apoptosis-inducing factor (AIF), the first identified protein involved in caspase-independent cell death [43]. Upon cytotoxic insult, AIF is cleaved, released from the mitochondria to the cytoplasm and translocates to the nucleus where it causes caspase-independent DNA fragmentation and cell death [32,43]. Western blotting indicated that DU145 and LNCaP cells treated with 2ME2/Doc + BA increased nuclear AIF compared to single drug treatment and control (Fig. 9A). Increased nuclear AIF was also detected by immunofluorescent cell staining of DU145 cells treated with 2ME2/Doc + BA and compared to control cells (Fig. 9C).
In addition, we investigated the expression of AIFshort (AIFsh) protein, an alternative transcription start site coding for AIF corresponding only to the C-terminal prodeath portion that translocates to the nucleus to induce cell death [32]. Western blot and RT-qPCR analysis showed that LNCaP cells treated with 2ME2/Doc + BA or BA alone increased AIFsh protein and mRNA (2-5-fold) compared to 2ME2 and Doc alone (Fig. 9B, D). IκBα, A20, and DR5 mRNAs, genes known to be regulated by NF-κB [1] also increased in LNCaP cells treated with the 2ME2/ Doc + BA combination (Fig. 10). In DU145 and PC3 cells, there was also an increase in AIFsh protein after treatment with 2ME2/Doc + BA or BA alone (Fig. 9B).
Finally, we analyzed if shRNA knockdown of AIF has any effect on cell death induced by the 2ME2/Doc + BA combination in LNCaP and DU145 cells. In LNCaP cells, AIF shRNA reduced both total AIF and AIFsh protein and lowered cell death compared to control LNCaP/ shGFP cells (Fig. 9E). In DU145 cells, AIF knockdown also lowered cell death and cleaved PARP by the 2ME2/ Doc + BA combination (result not shown). We suggest that the NF-κB activator BA increases expression of AIFsh and stimulates caspase-independent cell death in apoptosis resistant PC cells such as DU145.

Discussion
Treatment of cancer cells with chemotherapeutic drugs often results in substantial heterogeneity in the response to NF-κB activity. In some cases, NF-κB activation by chemotherapeutic drugs elicits a pro-survival cellular response and combination with inhibitors of NF-κB improves efficacy [1,3,4]. However, depending on the type of drug or cancer cell, activation of NF-κB can elicit a pro-death response [6]. Our results indicated that improving the cell death response to 2ME2 and Doc in PC cells depends on stimulating rather than inhibiting NF-κB activity. In contrast to what was observed with NF-κB inhibitors, combination of 2ME2 or Doc with BA, an activator of NF-κB, increased cell death in androgenresponsive as well as castration-resistant PC cell lines. Therefore, our data suggests that a chemotherapy combination strategy utilizing antimitotic drugs with BA is likely to be a more universally effective chemotherapeutic strategy for PC. Our data suggest that the combination of antimitotic drugs with NF-κB inhibitors will have antagonistic effects in a common type of PC cell typical of LNCaP and LN-AI. Support for this observation comes from a report demonstrating that bortezomib, a proteasome inhibitor that lowers NF-κB activity by blocking degradation of IκBα, inhibits Doc-induced apoptosis in LNCaP cells [44]. More importantly, recent clinical trials indicate that patients with CRPC have no added benefit from bortezomib above Doc in one study and some antitumor activity in another study [45,46]. It is not yet known, however, if more specific inhibitors of NF-κB in combination with antimitotic drugs will have a better therapeutic effect clinically, especially since constitutive NF-κB activity is very prominent in CRPC [14][15][16].
Our results are similar to a previous report showing that inhibition of NF-κB with dnI or the NF-κB inhibitor BAY 117082 blocks 2ME2-induced apoptosis in LNCaP cells [23]. Others have shown that 2ME2 can inhibit NF-κB in PC3 and medulloblastoma/glioma cell lines and blocking the Doc increase in NF-κB can enhance apoptosis in a variety of cancer cell lines [47][48][49][50][51]. Overall, the heterogeneity implicated in the NF-κB response to anticancer drugs is dependent on the specific type of drug and cancer cell.
In LNCaP and LN-AI cells, the requirement of 2ME2 and Doc to activate NF-κB and increase apoptosis may depend upon the p53 tumor suppressor protein [37]. There is evidence suggesting a link between activation of NF-κB and the ability of p53 to induce apoptosis [7,52]. Our results in LNCaP cells indicated that 2ME2 increased nucleolar localization of p65 (Fig. 2). Nucleolar localization of p65 has been previously reported in colon cancer cells treated with aspirin and a model is proposed that the nucleolus sequesters p65 and inhibits its antiapoptotic functions [38]. Interestingly, the ARF tumor suppressor protein is localized to the nucleolus and after activation by oncogenes can prevent Hdm2 from targeting p53 for degradation and therefore increases the stability of p53 [53]. In addition, ARF can modulate p65 transcriptional activity to repress antiapoptotic genes in a p53 independent manner [54]. One of the effects of antimitotic drugs is the disruption of the nucleolus, the release of ARF, and sequestration of Hdm2, which then leads to the stabilization of p53 and subsequent induction of apoptosis [55]. Our results suggest that investigating the mechanistic basis of p65 nucleolar localization is likely to yield significant insights regarding how to optimize the cytotoxic antitumor action of antimitotic drugs.
It is known that the ability of BA to kill cancer cells occurs by multiple signaling pathways including through activation of NF-κB [12,56]. One potential mechanism for NF-κB activation and increase in apoptosis by BA is the degradation of IκBα and XIAP (Fig. 6). Activation of selective proteasome-dependent degradation of Sp1, 3, and 4 transcription factors controlling the proangiogenic gene VEGF and the antiapoptosis gene survivin by BA has been recently reported [57].
The pro-death effects of BA are independent of p53, which is a desirable characteristic for any agent utilized for the treatment of advanced PC, which frequently lacks functional p53 [58]. Our data show that despite a decrease in p53 protein, BA can still increase apoptosis or cell death in all PC cells (Fig. 6). Our results also suggest that BA combined with 2ME2 or Doc increases cell death in a caspase-independent manner (Figs. 6, 7B, 8D). We suggest that one of the factors that allow DU145 cells to overcome the defect in the apoptosis pathway (mutant p53/null Bax) is the increased nuclear translocation AIF/ AIFsh to mediate non-apoptotic cell death.

Conclusions
Combination chemotherapy is required to further improve the survival of patients with CRPC. A prevailing strategy has been to inhibit the NF-κB response in order to block its pro-survival effect and improve drug efficacy.
In this study, we demonstrated that in PC cells such as LNCaP and LN-AI, activation of NF-κB by the antimitotic agents 2ME2 or Doc is important for increasing apoptosis. In addition, when 2ME2 or Doc is combined with an NF-κB activator such as BA, there is effective induction of cell death in all the PC cells analyzed. We are currently investigating whether other NF-κB activators will also mediate increased cell death by antimitotic drugs. The combination of antimitotic agents with NF-κB activators may promote the pro-death responses in a greater variety of PC cells, a requirement for increased therapeutic efficacy.

Competing interests
The authors declare that they have no competing interests.