NF- κ B Promotes Survival During Mitotic Cell Cycle Arrest

data suggest that NF- κ B activation is linked to cell survival through the activation of NF- κ B-responsive anti-apoptotic genes and may represent an important survival mechanism during mitotic cell cycle arrest. nuclear localization occurring 6h and the number of cells showing a nuclear localization of p65 although this phenotype remained above basal levels. Control cells treated with DMSO alone show any change in the distribution of the p65 protein during the time-course examined. In mitotic cells (collected at 12-24h after nocodazole treatment) p65 reported interaction between I κ B α and a microtubule-associated dynein light chain protein (56) may provide the link between changes in microtubule dynamics and activation of NF- κ B. It can that the pool of microtubule-bound I κ B α may act as a sensor of microtubule dynamics that is phosphorylated and degraded upon microtubule depolymerisation/stablilzation. It is less clear how anti-microtubule drugs stimulate the activation of the IKK’s. NF- κ B-inducers such as TNF- α , IL-1 or Fas activate the IKK’s through upstream protein kinases such as NIK and the MAP3K kinases including MEKK1, 2 and 3 There are data to indicate that MEKK1 is activated in response to microtubule depolymerisation However, using immunecomplex kinase assays we have not observed the activation of MEKK1, MEKK2 or MEKK3 in HeLa cells upon nocodazole treatment (data not shown). Whether it is NIK or other protein kinases that activate the IKK’s in response to the anti-microtubule drugs remains to be determined.


INTRODUCTION
By perturbing microtubule dynamics certain anticancer drugs cause mammalian cells to arrest in mitosis through activation of the spindle checkpoint (or mitotic checkpoint) (1). Generally, mitotic arrest is then followed by apoptosis (2,3).
However, the biochemical mechanisms by which the anticancer drugs induce toxicity are not clearly understood. We have recently demonstrated that anticancer drugs such as nocodazole, vincristine and vinblastine (microtubule depolymerising agents) and taxol (a microtubule stabilising agent) activate intracellular signalling pathways that have opposing effects on cell survival (4). Another signal transduction pathway that is activated by these anticancer drugs involves the transcription factor nuclear factor (NF)-κB (5)(6)(7). However, the functional consequence of NF-κB activation by either microtubule depolymerising or stabilising drugs remains unknown.
Recent studies provide compelling evidence that NF-κB suppresses cell death in response to a variety of apoptotic stimuli (9). For example, the disruption of genes encoding components of the NF-κB signalling pathway results in early embryonic lethality with the mouse embryos displaying increased apoptosis in various organs and tissues (9). Candidate anti-apoptotic genes targeted by NF-κB include the mitochondrial membrane stabilising proteins Blf-1 (24) and Bcl-xL (25,26), the caspase 8 inhibitor FLIP (27), the immediate early response gene IEX-1L (28) and the inhibitor of apoptosis proteins (IAP's) cIAP1, cIAP2 and XIAP (29,30). Recent studies suggest that NF-κB can also inhibit TNFα-induced apoptosis by inhibiting activation of the c-Jun NH 2 -terminal kinase (JNK) via a poorly defined mechanism involving increased expression of either GADDβ or X-IAP (31,32). In this report we show that NF-κB is activated in HeLa cells in response to both microtubule depolymerising and microtubule stabilising anticancer agents. We also show that inhibition of NF-κB using three independent approaches selectively increases the susceptibility of mitotically-arrested cells to anticancer drug-induced apoptosis. Our

Cell culture and preparation of cell extracts
Human cervical carcinoma cells (HeLa) were cultured as described previously (33).

Immunoprecipitation and Western blot analysis
Immunoprecipitations were performed as described previously (35). The immunoprecipitated proteins and the cell extracts were resolved by SDS-PAGE and electroblotted onto Hybond-C nitrocellulose membrane using a Hoeffer semi-dry blotting apparatus (Amersham Biosciences, Piscataway, NJ). Immunoreactive proteins were visualised using ECL (Amersham Biosciences).

Protein kinase assays
Immunecomplex kinase assays for the IKK's were performed as described previously (35) using GST-IκBα as substrate.

Reporter assays
For transient transfection of reporter plasmids, cells were plated in 12-well plates and the following day cells were co-transfected with a luciferase reporter plasmid containing either 3xNF-κB binding sites (Stratagene, CA, USA) in combination with the internal control CMV Renilla luciferase plasmid (provided by Anne Willis, University of Leicester). Luciferase activities of reporter plasmids were measured using the Dual-Luciferase Reporter Assay System (Promega UK, Southampton, UK).

Peptide synthesis and purification
The amino acid sequence of both the wild type and the mutant (W739A, W741A) NEMO-binding domain (NBD) peptide fused to the Antennapedia homeodomain was 10 as described (36). Crude synthetic NBD peptides (wild type and mutant) were synthesised by Pepceuticals Ltd., (Leicester, UK). They were further purified using reverse-phase HPLC chromatography (Agilent Technologies UK Ltd., Stockport, Cheshire, UK) and concentrated by freeze-drying. The quality of the purified peptides was assessed using an analytical HPLC column (Perkin Elmer Instruments, Buckinghamshire, UK) and by electrospray mass spectrometry on a Micromass Platform-II instrument (Waters Ltd., Manchester, UK).

Densitometry
Autoradiograms were scanned and quantified using a β-Imaging Computing Densitometer (Molecular Dynamics, Little Chalfont, UK) with MD ImageQuant software (version 3.3).

NF-κB activation in response to anti-microtubule drugs is mediated by the IKK
and IκB signalling pathway.
To investigate whether microtubule depolymerization activates NF-κB signalling we treated exponentially growing HeLa cells with nocodazole, a reversible microtubule depolymerising compound (1). Cytoplasmic extracts were assayed for the activation of both IKKα and IKKβ using immunecomplex kinase assays. As shown in Fig.1A, nocodazole treatment rapidly activated (within 15 min) both IKKα (1.8-fold activation above the basal level) and IKKβ (2.1-fold above the basal level). Both IKKα and IKKβ were increasingly activated for upto 6h after nocodazole addition. From 12h onwards we were able to separate the nocodazole-treated cells into a mitotic population (as assessed by increased phosphorylation of histone H3; [37]) and an attached, nonmitotic population (as assessed by low phosphorylation of histone H3). As shown in To further investigate the relationship between NF-κB activation and the survival of nocodazole-arrested mitotic cells we transfected cells with either wild type IκBα or with a mutant, non-degradable IκBα (S32A/S36A, [38]). As shown in Fig.6A nondegradable IκBα reduced nocodazole-stimulated luciferase activity by 82% in the attached cells and by 78% in the mitotically-arrested cells when compared to wild type IκBα. The reduction in luciferase activity by the non-degradable IκBα could not be attributed to variations in the level of expression of the mutant IκBα when compared to the wild type protein (Fig.6B). Expression of non-degradable IκBα selectively increased the level of apoptosis in the nocodazole-arrested mitotic cells (79% apoptosis compared to 41% apoptosis in cells expressing wild type IκBα), without affecting the survival of the nocodazole-treated attached cells (Fig.6C).

The NBD peptide inhibits nocodazole-induced activation of NF-κB and reduces the survival of nocodazole-blocked mitotic cells.
As a second approach to suppress nocodazole-induced activation of NF-κB we used the NBD peptide, a known inhibitor of the IKK's (36). Treatment of exponentially growing HeLa cells with nocodazole for 3h caused p65 to translocate to the nucleus in approximately 32% of the cells (Fig.7A). However, treatment of cells with 100µM NBD peptide reduced nocodazole-induced p65 translocation into the nucleus by 80% to near basal levels, whereas treatment of cells with a mutant NBD peptide caused a small, but statistically significant (p<0.05) increase in nocodazole-induced p65 translocation into the nucleus (Fig.7A). The inhibition of p65 nuclear translocation by the NBD peptide was not restricted to nocodazole treatment but was also observed with other NF-κB-inducers such as TNF-α (data not shown). To determine the effect of the NBD peptide on the survival of the nocodazole-treated cells, exponentially growing HeLa cells were treated with nocodazole in the presence of varying concentrations of either the wild type or the mutant NBD peptide. As shown in Fig.7B the wild type NBD peptide, but not the mutant peptide, caused a dose-dependent increase in apoptosis in the mitotically-arrested cells but not in the attached cell population.

Proteasome inhibitors block nocodazole-induced activation of NF-κB and reduce the survival of nocodazole-blocked mitotic cells.
Finally, we have used inhibitors of the 26S proteasome to inhibit nocodazole-induced degradation of IκBα and suppress activation of NF-κB (38)(39)(40)(41). As shown in Fig.8A treatment of exponentially growing HeLa cells with nocodazole resulted in the timedependent degradation of IκBα. The nocodazole-induced degradation of IκBα was inhibited by the proteasome inhibitors MG132 or ALLN (42) but not the protease inhibitors leupeptin or PMSF (Fig.8A). The inhibition of IκBα degradation by MG132 either in the absence (Fig.8B, lane 4) or presence of nocodazole (Fig.8B, Lane 3) correlated with the appearance of higher molecular weight forms of IκBα when compared to control, untreated cells (Fig.8B, lane 1) or following treatment of cells with nocodazole only (Fig.8B, lane 2). The low molecular weight bands (indicated by an asterisk) observed in the IκBα immunoblot have also been observed in a previous study (38) and are thought to represent intermediates in the degradation of IκBα.
These intermediate products of IκBα degradation were lost following nocodazole treatment (Fig.8B, lane 2) but their levels were maintained in cells treated with MG132 either in the presence or absence of nocodazole (Fig.8B, lanes 3 and 4). To further demonstrate that MG132 inhibits degradation of ubiquitinated IκBα, we co-  (Fig.8C, lane 2) that was suppressed by MG132 (Fig.8C, lanes 3 and 4).
To determine the effect of MG132 on nocodazole-induced apoptosis HeLa cells were treated with either nocodazole or with both nocodazole and MG132. As shown in Similarly, ALLN but not PMSF or leupeptin, also enhanced nocodazole-induced apoptosis preferentially in the mitotic cell population. Treatment of cells with the inhibitors alone did not increase apoptosis significantly above basal levels. Together these results suggest that activation of NF-κB by microtubule depolymerising agents and the corresponding increase in NF-κB-dependent gene expression may be a mechanism by which NF-κB contributes to the survival of mitotically-arrested cells.

DISCUSSION
Previous work has clearly demonstrated that microtubule depolymerisation leads to the activation of NF-κB and NF-κB-dependent gene expression (5). What has remained unclear is the functional consequence of activating this particular signalling pathway. In this study we have identified a number of components in the NF-κB signalling pathway that are activated by anti-microtubule drugs and demonstrate that one functional consequence of activating this signalling pathway is to selectively aid the survival of mitotically-arrested cells.
We have shown in this study that a range of compounds that either depolymerise or stabilise microtubules can activate the NF-κB signalling pathway. Consistent with previous work (5)  activation of the IKK's that we observe specifically in the mitotically-arrested cells is more difficult to reconcile with changes in gene expression as both transcription and translation are generally suppressed during mitosis (48,49). The identity of the protein kinase(s) that activate the IKK's in mitotically-arrested cells and their potential function remains to be identified although there are data to suggest that the activation of the IKK's in mitotically-arrested cells may be mediated by caspases. In mammalian cells the three death domain-containing proteins caspase-8, caspase-10 and MRIT (Mach-related inducer of toxicity) have been shown to activate IKK by a mechanism that involves an interaction between the prodomain of caspases and the IKK's (50). This caspase-dependent mechanism of IKK activation would be consistent with the observation that caspase 8 is known to be active in taxol-arrested mitotic cells (51,52).
Although microtubule depolymerisation and not microtubule stabilization was originally reported to be the stimulus for NF-κB activation (5), subsequent studies have shown that the microtubule stabilising drug taxol can also activate NF-κB in mammalian cells (6,7,53,54). In the present study we have also shown that taxol can activate NF-κB as assessed by the increase in the phosphorylation and degradation of IκBα. Therefore, we suggest that it is not microtubule depolymerisation or stabilisation per se that may lead to NF-κB activation but rather a perturbation of microtubule dynamics. Another explanation for the activation of NF-κB by both microtubule depolymerising and microtubule stabilising drugs is that these compounds are thought to activate a common enzyme, such as PKC (6) that subsequently activates NF-κB. The mechanism by which PKC stimulates NF-κB is unclear but may involve the activation of the IKK's (55).

21
The reported interaction between IκBα and a microtubule-associated dynein light chain protein (56) may provide the link between changes in microtubule dynamics and activation of NF-κB. It can be envisaged that the pool of microtubule-bound IκBα may act as a sensor of microtubule dynamics that is phosphorylated and degraded upon microtubule depolymerisation/stablilzation. It is less clear how anti-microtubule drugs stimulate the activation of the IKK's. NF-κB-inducers such as TNF-α, IL-1 or Fas activate the IKK's through upstream protein kinases such as NIK (12) and the MAP3K kinases including MEKK1, 2 and 3 (15)(16)(17)(18). There are data to indicate that MEKK1 is activated in response to microtubule depolymerisation (57)(58). However, using immunecomplex kinase assays we have not observed the activation of MEKK1, MEKK2 or MEKK3 in HeLa cells upon nocodazole treatment (data not shown). Whether it is NIK or other protein kinases that activate the IKK's in response to the anti-microtubule drugs remains to be determined.
While the precise biochemical mechanism by which perturbation of microtubule dynamics leads to the activation of NF-κB remains unclear the current study has established a role for NF-κB in regulating cell survival during mitotic cell cycle arrest.
As outlined in the Introduction the prevalent hypothesis is that NF-κB functions as an anti-apoptotic factor. In some instances, however, NF-κB is reported to have a proapoptotic function (53,54,(59)(60)(61)(62) suggesting that the precise effect of NF-κB may be cell and stimulus-specific. Our data are consistent with the majority view and assign a survival function for NF-κB. Anti-microtubule drugs cause cell to arrest in mitosis through activation of the mitotic checkpoint and as demonstrated in this study and in earlier work (1-4) the mitotically-arrested cells then undergo apoptosis. Cells that 22 have been treated with the anti-microtubule drugs for an equivalent time but that are not in M-phase do not undergo apoptosis, suggesting that activation of the mitotic checkpoint initiates the apoptotic response. The mechanism by which activation of the mitotic checkpoint then leads to cell death is not yet understood. We have recently demonstrated that both pro-apoptotic signals, such as p38 MAP kinase, and anti-apoptotic signals such as p21-activated kinase (PAK) are preferentially activated in the mitotically-arrested cells (4). Other work has shown that the anti-apoptotic protein bcl-2 becomes hyperphosphorylated during mitotic cell cycle arrest and may constitute a pro-apoptotic signal (63). In contrast, the rapid activation of NF-κB in response to microtubule perturbation may represent a survival signal that is activated prior to activation of the mitotic checkpoint. Whether the activation of NF-κB and the resulting transcription of anti-apoptotic genes are fortuitous, or a programmed response to microtubule perturbation that pre-empts the requirement for survival factors during subsequent mitotic arrest is presently unclear. As outlined in the Introduction there are a number of anti-apoptotic genes that are targeted by NF-κB.
The identity of the NF-κB-regulated anti-apoptotic gene product(s) that contribute to cell survival in response to anti-microtubule drugs is currently under investigation.        were pretreated with 5µM MG132 or an equivalent volume of DMSO for 1h prior to nocodazole treatment (3µM) for 12h. The mitotic cells were recovered by 'shake-off' and reattached onto poly-D-Lysine-coated coverslips. The mitotic were fixed and immunostained with the M30 antibody to identify the apoptotic cells as described in