NF- k B Activation by Camptothecin A LINKAGE BETWEEN NUCLEAR DNA DAMAGE AND CYTOPLASMIC SIGNALING EVENTS*

Activation of the transcription factor NF- k B by extracellular signals involves its release from the inhibitor protein I k B a in the cytoplasm and subsequent nuclear translocation. NF- k B can also be activated by the anti-cancer agent camptothecin (CPT), which inhibits DNA topoisomerase (Topo) I activity and causes DNA double-strand breaks during DNA replication to induce S phase-dependent cytotoxicity. Here we show that CPT acti-vates NF- k B by a mechanism that is dependent on initial nuclear DNA damage followed by cytoplasmic signaling events. NF- k B activation by CPT is dramatically diminished in cytoplasts and in CEM/C2 cells expressing a mutant Topo I protein that fails to bind CPT. This response is intensified in S phase cell populations and is prevented by the DNA polymerase inhibitor aphidicolin. In addition, CPT activation of NF- k B involves degradation of cytoplasmic I k B a by the ubiquitin-proteasome pathway in a manner that depends on the I k B kinase complex. Finally, inhibition of NF- k B activation aug-ments CPT-induced apoptosis. These findings elucidate the progression of signaling events that initiates in the nucleus with CPT-Topo derived from

tion sequence of NF-B and retains it in the cytoplasm (4,5). Dissociation from IB␣ is essential for NF-B to enter the nucleus and to activate gene expression. Several signaling cascades that control NF-B activation converge at an IB kinase (IKK) 1 complex, responsible for site-specific phosphorylation of IB␣ at serines 32 and 36 (6 -10). Phosphorylation of IB␣ induces multiubiquitination of IB␣ and its subsequent degradation by the ubiquitin-dependent 26 S proteasome (11,12). This sequence of events can be induced without de novo protein synthesis by multiple extracellular stimuli, including tumor necrosis factor ␣ (TNF␣), interleukin-1, phorbol ester (PMA), bacterial lipopolysaccharide (LPS), and others. However, NF-B activation can also be achieved through mechanisms that are distinct from the above IKK-dependent model. These include phosphorylation-independent yet proteasome-mediated IB␣ degradation induced by ultraviolet irradiation (13,14), calpain-dependent degradation of IB␣ by silica and TNF␣ (15,16), and tyrosine phosphorylation-induced dissociation of IB␣ from NF-B following hypoxia and reoxygenation (17). Thus, depending on the stimuli, NF-B can be activated through multiple distinct regulatory pathways.
Activation pathways of NF-B typically originate from ligand-receptor interactions on the cell membrane. However, NF-B can also be activated by a group of agents that damage DNA in the nucleus. A paradox confounding our current understanding of the mechanism of NF-B activation by agents that damage DNA is that the major source of damaged DNA is in the nucleus, whereas latent NF-B complex is in the cytoplasm. It was previously hypothesized that a signal may transfer from the nucleus to the cytoplasm (18). In support of this model, a recent study by Piette and Piret (19) provides evidence that NF-B activation by DNA-damaging agents correlates with their capacity to induce DNA breaks. However, the requirement of damaged DNA in the nucleus has not been directly demonstrated. In contrast, Devary et al. (20) showed that enucleated cells (i.e. cytoplasts) retained full capacity to activate NF-B following UV irradiation, indicating that nuclear DNA damage is not necessary for NF-B activation by UV irradiation. There is now substantial evidence to support the notion that UV activation of NF-B involves activation of cell surface receptors by ligand-independent oligomerization (14,(21)(22)(23)(24) and/or oxidative stress-mediated inactivation of receptor tyrosine phosphatases, which ultimately leads to ligandindependent activation of receptor tyrosine kinases (25). Whether nuclear DNA damage can directly activate an intracellular NF-B signaling pathway without involving cell surface receptors remains an important question yet to be resolved.
We and others have observed that an anti-cancer agent, camptothecin (CPT), can activate NF-B in pre-B or T cell lines (19,26). CPT inhibits the activity of DNA topoisomerase (Topo) I (27)(28)(29). Topo I changes the supercoiling of DNA and therefore plays critical roles in DNA replication, in RNA transcription, and, indirectly, in DNA damage repair (30). CPT selectively binds to and stabilizes a covalent Topo I-DNA reaction intermediate, referred to as the cleavable complex, which contains a single-strand DNA break (SSB) (31,32). DNA double-strand breaks (DSBs) are then generated during DNA replication when the replication fork collides with the cleavable complex (33). In the present study, our objective was to determine whether or not nuclear events associated with the DNA-damaging action of CPT and a clinically utilized derivative of CPT, topotecan (TPT) (34,35) were required for activation of cytoplasmically localized NF-B complexes. We also examined whether CPT activation of NF-B modulated an apoptotic response. Our findings elucidate a series of nuclear events induced by CPT/TPT that converge with cytoplasmic signaling events responsible for the activation of NF-B, which can provide anti-apoptotic function.

EXPERIMENTAL PROCEDURES
Cell Culture-Culture conditions for 70Z/3 and 70Z/3-CD14 murine pre-B cells have been described (36). CEMp and CEM/C2 human T cell lines were kindly provided by Dr. Y. Pommier (National Institutes of Health) and maintained in RPMI 1640 medium (Cellgro, Mediatech) supplemented with 10% fetal bovine serum (HyClone Laboratory, Inc.), 1000 units of penicillin G (Sigma), and 0.5 mg/ml streptomycin sulfate (Sigma) in an humidified 5% CO 2 -95% air incubator (Forma Scientific). HeLa human cervical carcinoma cells and PC-3 human prostate carcinoma cells were maintained in Dulbecco's modified Eagle's medium (Cellgro) supplemented with 10% fetal bovine serum and antibiotics as above in a 10% CO 2 -90% air incubator. The human kidney embryonic fibroblast 293 (HEK293) was maintained in the latter medium on 0.1% (w/v) gelatin-coated plastic culture dishes.
Reagents-CPT, VP16, calpain inhibitor I (ALLN), Me 2 SO, bacterial LPS, PMA, cycloheximide, aphidicolin, and cytochalasin B were purchased from Sigma. TPT was a gift from SmithKline Beecham. Lactacystin was generously provided by Dr. E. J. Corey (Harvard University). Stock solutions were prepared in Me 2 SO at 10 mM (CPT), 10 mM (VP16), 30 mM (aphidicolin), cytochalsin B (10 mg/ml), and 25 mM (lactacystin, 25% Me 2 SO-75% H 2 O). TPT stock was made in water at 30 mM. LPS was prepared in RPMI growth medium at 1 or 10 mg/ml. Human recombinant TNF␣ was from CalBiochem and resuspended in phosphate-buffered saline containing 0.1% bovine serum albumin (fraction V, Sigma). A 50 M PMA stock was made in 100% ethanol. A cycloheximide stock solution was made in water at 50 mg/ml. In each experiment, all samples received the same amounts of Me 2 SO to control for potential Me 2 SO effects. All stock solutions were stored in aliquots at either Ϫ70°C or Ϫ20°C. IgGs against IB␣ (C21), c-Rel (C), p65 (A and C20), Rel-B (C-19), p52 (I-18), and p50 (NLS) were purchased from Santa Cruz Biotechnology. A monoclonal anti-FLAG antibody was purchased from Kodak, and horseradish peroxidase-conjugated anti-rabbit and anti-mouse antibodies and protein A were obtained from Amersham Pharmacia Biotech. Prestained protein molecular weight markers were purchased from Life Technologies, Inc. Cell preparation and Western immunoblots were performed as described (36) and developed using the ECL procedure according to the manufacturer (Amersham Pharmacia Biotech). Blots were then exposed to x-ray film (Kodak).
Electrophoretic Mobility Shift and Supershift Assays-The Ig-B oligonucleotide probe and conditions for EMSA were previously described (36). For supershift assays, 1 g of IgG antibodies specific to members of the NF-B proteins (Santa Cruz Biotechnology) were added to nuclear extracts for 20 min on ice prior to addition of radiolabeled probe. The AP-1 site used for control EMSA reactions was obtained from Promega.
Enucleation Procedure-Enucleation was performed as described (37) with the following modifications. PC3 or HeLa cells were seeded in 30-mm dishes, grown overnight, and exposed to cytochalasin B (10 g/ml) for a total of 30 min at 37°C. Plates were placed upside down in 400-ml centrifuge bottles and bathed in growth medium containing cytochalasin B at the same concentration. Plates were secured at the bottom of the centrifuge bottles by appropriately sized styrofoam. Samples were then centrifuged at 10,000 rpm for 15 min at 37°C using a Beckman JA-14 rotor. The rotor and centrifuge chamber were prewarmed to ϳ37°C by prior centrifugation without samples. Plates with enucleated cells (i.e. cytoplasts) were then removed from the centrifuge bottle, gently rinsed with phosphate-buffered saline, and incubated with prewarmed growth medium for 30 min at 37°C. Samples of enucleated cells were fixed in 3:1 methanol/acetic acid, stained with Hoechst dye 33258, and photographed under a fluorescent microscope equipped with a 4Ј6,-diamidino-2-phenylindole-specific filter. The enucleation efficiency varied from ϳ75 to 95% for PC-3 and HeLa cells.
Transient Transfection Assay-Cells (HEK293, HeLa, or PC-3) were transiently transfected using a standard calcium phosphate precipitation method (38). CEMp and CEM/C2 cells were transfected with DEAE-Dextran method (39). An NF-B-dependent luciferase reporter (3xB-Luc) was constructed by inserting a HindIII-BglII fragment of 3xB-CAT into the HindIII-BglII sites of the tk-Luc plasmid (kindly provided by Dr. R. Evans, Salk Institutite). 3xMB-Luc was constructed using a similar cloning strategy starting with the 3xMB-CAT (40). 24 h following transfection, cells were treated with TPT (30 M) or CPT (10 M) for 2 h, rinsed twice with growth medium, and further incubated without drugs for 6 h before termination of the cultures. Positive control samples were treated with TNF␣ (10 ng/ml) for a total of 8 h. Control samples were transfected with the LacZ cDNA under the control of the cytomegalovirus promoter in the pCMX vector (CMV-LacZ). For CEMp and CEM/C2 cells, total proteins were used for normalization.
Full-length human IKK␣ cDNA was provided by Dr. I. M. Verma (Salk Institute). Full-length human IKK␤ cDNA was cloned by screening a human kidney cDNA library in ZAPII with a polymerase chain reaction-amplified DNA fragment using a HeLa cDNA library (CLON-TECH) and TO237 (5-CTCAGCAGCTCAAGGCCAAG-3Ј) and TO240 (5Ј-CCAGAGCTCCTTCTGCCGC-3Ј) primers. IKK␣ and IKK␤ with a Lys-to-Ala substitution at the conserved ATP binding site were generated by polymerase chain reaction mutagenesis and confirmed by DNA sequencing. The mutant genes were placed under the control of the cytomegalovirus promoter in the pcDNA3.1(ϩ) expression vector (CLONTECH). HEK293 cells were transfected with these constructs by calcium phosphate precipitation and then treated with either TNF␣ (10 ng/ml) or TPT (30 M) for 2 h. Nuclear extracts were analyzed by EMSA as described above. Cytoplasmic extracts were analyzed by Western blotting using the anti-FLAG monoclonal antibody (Kodak) to determine expression levels of respective dominant-negative mutants in each condition. An horseradish peroxidase-conjugated donkey anti-mouse antibody (Amersham Pharmacia Biotech) was used for secondary antibody followed by ECL development.
Retrovirus Construction and Infection-Production and infection of HA-tagged wild-type and HA-tagged S32A/S36A mutant IB␣ expression constructs were described (36). Other IB␣ deletion mutants were generated by polymerase chain reaction-mediated mutagenesis and confirmed by sequencing. Stable pools were selected with hygromycin (1 mg/ml, Roche Molecular Biochemicals), and the expression levels of the corresponding proteins were examined by either anti-IB␣ (C21, a C-terminal epitope) or anti-HA antibodies. For experiments shown in Figs. 7B and 9, HA-S32A/S36A clone 5 that expressed a relatively high level of the mutant protein was used. Similar but less pronounced effects were also seen with pooled cultures and in five isolated clones expressing varying levels of mutant protein (not shown).
Generation of a Green Fluorescent Protein-IB␣ Fusion Construct-N-terminally fused GFP-IB␣ was generated by subcloning polymerase chain reaction amplified human IB␣ (MAD3) into HindIII-BamHI sites of the pEGFP vector (CLONTECH), such that the entire MAD3 coding sequence was in-frame with the GFP coding sequence. Stable HEK293 cell clones were generated by G418 selection and subsequent FACS sorting. Cells were photographed using a Zeiss Axioplan microscope equipped for fluorescence with the aid of a fluorescein-specific filter.
FACS Analyses-For FACS sorting of G 1 , S, and total cell fractions for EMSA analyses, 70Z/3 cells untreated or treated with CPT, TPT, or LPS for appropriate durations were stained with Hoechst 33342 (stock at 10 mg/ml in water) at the final concentration of 10 g/ml for 15 min at 37°C in RPMI growth medium, followed by cell isolation using FACStar PLUS (Becton Dickinson) at 4°C. 10 6 cells each were purified, and total cell extracts were prepared for EMSA analyses. The status of the cell cycle of purified fractions was confirmed by propidium iodide staining followed by analysis with FACSCalibur (Becton Dickinson). Detailed protocols for apoptosis analyses using FACS have been published (41). Briefly, cells were fixed in ethanol, treated with a citric acid buffer to release fragmented DNA out of the cells, stained with propidium iodide, and analyzed using FACSCalibur.

CPT Induces Transient NF-B Activation in the Absence of de
Novo Protein Synthesis-NF-B activity is dictated by its ability to bind cognate B sites present in responsive genes. We utilized a B site from the immunoglobulin intronic enhancer in EMSA analyses to evaluate NF-B activation by CPT or TPT treatments. CPT induces dose-dependent (Fig. 1A, saturating at 10 M) and transient (Fig. 1B, peaking at 1-2 h) NF-B binding activity in 70Z/3-CD14 pre-B cells. Addition of 50-fold excess specific and nonspecific oligonucleotides (Fig. 1C, lanes [1][2][3] shows that the binding activity is specific to NF-B. Specificity is further demonstrated by the interaction of binding complex with antibodies to p50, RelA, and c-Rel (Fig. 1C). Antibodies to other NF-B family members, p52 and RelB, did not alter binding, indicating that these proteins are not components of the NF-B complex induced by CPT in 70Z/3-CD14 cells. Pretreatment with cycloheximide ( Fig. 1D, lanes 3 and 4) did not interfere with this pathway, which shows that CPT action does not require de novo protein synthesis. This activation is not limited to lymphoid cells because it was also induced by both CPT and TPT in CEM T leukemic, PC-3 prostate cancer, HEK293 embryonic kidney fibroblast, and HeLa cervical cancer cell lines (see below, others not shown). Induction of M CPT, terminated at various times and analyzed as described above. The LPS positive control (lane 9) was as described above. C, specificity of NF-B complexes induced by CPT. A nuclear extract isolated from cells exposed to 10 M CPT for 2 h was incubated with 50-fold excess wild-type or mutated oligonucleotides or with antibodies specific to p50 (NLS), p65/RelA (A), c-Rel (C), p52 (I-18), or RelB (C-19). Supershifted bands for anti-p65 and anti-c-Rel can be seen as slower migrating bands. Antibody against p50 causes reduced DNA binding. D, NF-B activation by CPT occurs in the absence of de novo protein synthesis. Cells were treated with or without 20 g/ml cycloheximide (CX) for 30 min to block de novo protein synthesis and treated with TPT (30 M) for 1 h and analyzed as described above. The above data are representative of experiments performed at least three times.
NF-B DNA binding activity by CPT or TPT treatment resulted in increased NF-B-dependent transcription of a luciferase reporter gene (see below). Thus, CPT or TPT activation of NF-B occurs without de novo protein synthesis and may utilize preexisting regulatory component(s).
Interaction of CPT with Nuclear Topo I Is Necessary for NF-B Activation-The primary molecular target of CPT or TPT is nuclear Topo I enzyme (27,32). However, mitochondria also contain CPT-sensitive Topo I (42). It is also possible that CPT activation of NF-B may involve molecular target(s) other than nuclear Topo I. To evaluate the requirement of a nuclear event in NF-B activation by CPT, we enucleated PC-3 and HeLa cells by the cytochalasin B-mediated enucleation procedure (37). This protocol produced enucleated cells with approximately 90% efficiency as determined by nuclear staining with Hoechst dye (Fig. 2A). Consistent with a previous report (43), EMSA of total cell extracts prepared from intact and enucleated cells demonstrated that NF-B activation by activators, such as PMA (Fig. 2B) or TNF (80), does not require an intact nucleus. By contrast, the NF-B response by CPT and TPT was dramatically diminished in the enucleated cells (Fig. 2B). Modest activation by TPT in enucleated cells is likely due to low numbers of intact cells present in the enucleated cell population (Fig. 2B). NF-B (p65), and upstream kinases in the signaling pathway, IKK␣ and IKK␤, are still present in the cytoplasts (Fig. 2B, lanes 2, 4, and 6), demonstrating that the lack of NF-B activation response in certain enucleated samples (lane 4) is not due to potential leakage of these signaling components. Thus, these results are consistent with the hypothesis that a nuclear event is necessary for NF-B activation by CPT-related compounds.
Although the above data are consistent with the notion that an intact nucleus is required for NF-B activation by CPT, it is unknown whether Topo I-induced DNA damage is also required for this process. Its DNA-damaging function requires CPT to interact with Topo I-DNA cleavable complexes (33). To address whether direct interaction of CPT and a Topo I-DNA complex is necessary for activation of NF-B, we examined human CEM/C2 cells, which exclusively express a mutant Topo I enzyme (44). This mutant Topo I enzyme contains two amino acid substitutions, Met 370 to Thr and Asp 722 to Ser. The latter mutation alone makes Topo I enzyme ϳ1000-fold more resistant to CPT (or TPT)-mediated inhibition of the religation of DNA nicks, making it incapable of efficiently inducing DNA damage by CPT treatment in vivo (45). We compared CPTinduced NF-B activity in CEM/C2 and the parental CEMp cells by EMSA. Time course and dose response studies (Fig. 3,  A and B, respectively), as well as B-dependent luciferase reporter assay (Fig. 3C), clearly demonstrated that CEM/C2 cells could not mount the NF-B response by CPT treatment, whereas the parental cell line retained the ability to activate NF-B. Efficient activation of NF-B in CEM/C2 cells by TNF (Fig. 3B) or other DNA-damaging agents, such as VP16 and ionizing radiation (Fig. 3D, others not shown), revealed that the lack of NF-B activation was specific to CPT treatment. Thus, these results together provide strong evidence that Topo I-mediated nuclear DNA damage is necessary for NF-B activation by CPT treatment.
NF-B Activation by CPT Depends on DNA Replication and Is Concentrated in S phase of the Cell Cycle-CPT inhibition of the religation step during the Topo I reaction induces stabilization of the cleavable complexes, resulting in generation of SSB. These SSB are reversible but can be converted into lethal DSB during S phase, when the replication fork collides with the cleavable complex (33). It has been suggested that aphidicolininduced inhibition of DNA polymerase activity prevents DSB liberation (34). Aphidicolin prevents S phase-specific toxicity of CPT (46). To evaluate whether a SSB or DSB is critical for NF-B activation by CPT, we examined the influence of aphidicolin on CPT induction of NF-B. FACS analysis confirmed that ϳ50% 70Z/3-CD14 cells were in S phase of the cell cycle at the time of CPT treatment (see below). The EMSA demonstrated that CPT activation of NF-B was efficiently blocked by this treatment (Fig. 4A, lanes 3-5). Aphidicolin, however, did not block NF-B activation by bacterial LPS (lanes 10 -12). Aphidicolin also did not directly block NF-B DNA binding activity (lanes 6 -8).
These results are consistent with the hypothesis that the generation of DSB, not SSB, is necessary for efficient NF-B activation by CPT treatment (19). These data also imply that this activation pathway may occur only in the S phase of the cell cycle. We therefore enriched 70Z/3-CD14 cells in the S phase by FACS sorting after cells were stimulated with CPT or LPS for 2 h (Fig. 4B). Compared with a similarly obtained G 1 cell population, NF-B activation was 2.8-fold higher in the S phase The asterisk indicates an NF-B complex whose appearance was not consistently seen. The above data are representative of three independent experiments. population when equivalent amounts of cell extracts were analyzed by EMSA (Fig. 4C). LPS stimulation did not show any significant differences between S and G 1 cells. These findings demonstrate that CPT activation of NF-B is cell cycle coupled and predominantly takes place during the S phase of the cell cycle in a DNA-polymerase-dependent fashion. This also can explain why NF-B activation by CPT or TPT is relatively lower in virtually all cell types examined when compared with LPS or TNF␣. NF-B activation by CPT or TPT is dependent on the percentage of replicating cells in S phase, whereas activation by either LPS or TNF␣ is not cell cycle coupled.
CPT Induces Degradation of IB␣ by a Ubiquitin-Proteasome Pathway-The regulatory events governing NF-B activity induced by cytokines and LPS are well characterized and involve activation of cytoplasmic signaling cascades (2). The primary regulator is the inhibitory protein, IB␣, which maintains NF-B in the cytoplasm. Release of NF-B to the nucleus depends on degradation of IB␣. To determine whether CPT activation of NF-B is solely dependent on nuclear events or whether cytoplasmic events are also required, IB␣ protein levels were monitored following treatment with CPT by Western immunoblot analyses. CPT treatment caused a reduction in IB␣ protein levels (Fig. 5A, compare lanes 6 and 7), consistent with induction of IB␣ degradation. This degradation was prevented by the proteasome inhibitors, ALLN and lactacystin (Fig. 5A, lanes 8 and 9). A longer exposure of the film (Fig. 5B,  lanes 8 and 9) revealed an accumulation of characteristic high molecular mass multiubiquitinated IB␣ ladders (11,12). Proteasome inhibitors not only prevented IB␣ degradation but also NF-B activation by CPT treatment (Fig. 5C, compare  lanes 2 and 3). TPT induced similar IB␣ degradation (Fig. 6A). Inhibition of IB␣ degradation by ALLN resulted in accumulation of IB␣ in the cytoplasm, as visualized by an IB␣ protein N-terminally tagged with the green fluorescent protein (GFP-IB␣) (Fig. 6B, right panel). Control coimmunoprecipitation experiments with RelA-specific antibodies confirmed that the addition of the GFP tag did not interfere with its association with NF-B (80). The GFP tag also did not affect TPTinduced proteolysis (Fig. 6A, lanes 2-5). Thus, induction of IB␣ degradation by CPT or TPT is similar to that induced by LPS in 70Z/3 cells or TNF␣ in multiple cell types (1, 2), which utilizes a ubiquitin-proteasome pathway. Moreover, we also found that IB␣ degradation by TPT was markedly reduced in enucleated PC3 cells (not shown). These data demonstrate that the progression of events initiated in the nucleus by TPT or CPT treatment is continued in the cytoplasm.
IB␣ Degradation Induced by CPT or TPT Is Ser 32/36 -dependent-Although cytokines and LPS cause IB␣ degradation by a ubiquitin-proteasome pathway that requires intact Ser 32 and Ser 36 residues, UV irradiation causes IB␣ degradation by a ubiquitin-proteasome pathway independent of these Ser residues (13,14). To evaluate whether CPT or TPT-induced IB␣ degradation requires intact Ser 32/36 residues, the S32A/S36A mutant protein was N-terminally tagged with an HA epitope (HA-S32A/S36A) (36), stably introduced in 70Z/3-CD14 cells, and analyzed for sensitivity to degradation by TPT treatment. The S32A/S36A mutant protein was completely resistant to degradation induced by TPT treatment (Fig. 7A). This was not due to the presence of the HA tag because the control HA-WT IB␣ protein was efficiently degraded. Moreover, a Ser 32/36 deletion mutant without the HA tag also failed to be efficiently degraded (Fig. 7A, ⌬30 -40). Stable expression of the HA-S32A/ S36A mutant, but not HA-WT, selectively eliminated the appearance of NF-B DNA binding in the nucleus after treatment with CPT or TPT (Fig. 7B, lanes 9 and 10). Consistent with the formation of multiubiquitinated IB␣ ladders (Fig. 5B, lanes 8  and 9), substitution of the primary ubiquitination sites Lys 21 and Lys 22 (11,50), with Arg resulted in retardation of degradation following TPT treatments (Fig. 7A, HA-K21/22R). These results are similar to those obtained with LPS (Fig. 7, A,   FIG. 4. CPT activation of NF-B requires S phase-dependent generation of DSB. A, DNA replication is required for NF-B activation by CPT treatment. 70Z/3-CD14 cells were treated with 30 M TPT or 1 g/ml LPS in the presence or absence of the indicated concentration of aphidicolin (in vivo treatments) and nuclear extracts were analyzed by EMSA. In vitro refers to addition of the drug direct to cell extracts as in lane 2 prior to electrophoresis. B, FACS enrichment of G 1 and S phase cell population. Exponentially growing 70Z/3-CD14 cells were FACS purified based on DNA content as measured by Hoechst staining into total, G 1 , and S populations. C, NF-B activation by CPT treatment is concentrated in S phase of the cell cycle. NF-B activation in cell fractions isolated as in B was analyzed by EMSA using equal protein loading. Phosphorimage analysis showed that CPT activation was 2.8fold higher in S fraction than in G 1 . The data are representative of at least two independent experiments. Relative molecular mass (in kDa) is shown on the right. C, NF-B activation induced by CPT is blocked by proteasome inhibitors. Nuclear extracts prepared from cells that were treated as in A were analyzed by EMSA for NF-B binding activity as in Fig. 1A. The data are representative of two or more independent experiments. lane 5, and B, lanes 7 and 8) or TNF␣ (1, 2) but distinct from data obtained with UV irradiation (13,14). Of note, LPS caused efficient degradation of HA-K21/22R (Fig. 7A, K21/22R, lane  2), which is consistent with the observations that other Lys residue(s) can compensate for the lack of Lys 21/22 sites when cells are exposed to potent NF-B inducers (50).
The IKK Complex Is Essential for NF-B Activation by CPT-To further elucidate the events upstream of IB␣ degradation that are involved in CPT activation of NF-B, we evaluated the NF-B response by EMSA in HEK293 cells transiently expressing dominant-negative IKK␣ and ␤ proteins. The use of EMSA analysis to investigate potential inhibitory responses was possible because transfection efficiency was consistently Ͼ90% in this cell type (Fig. 8A), and thus almost all cells in the transfected population expressed the IKK mutant proteins. Both IKK mutants (N-terminally tagged with a FLAG epitope) reduced the level of NF-B activation by TPT in a dose-dependent manner (Fig. 8B, upper panel). By contrast, these mutants did not appreciably alter the DNA binding levels of AP-1 complex (Fig. 8B, lower panel). Dose-dependent expression of IKK mutant proteins is shown by Western blot analysis of cell extracts using monoclonal anti-FLAG antibody (Fig. 8C).

NF-B Activation by CPT or TPT Is an Anti-apoptotic Cell Survival
Response-Recent studies demonstrated that NF-B activation by certain death inducing agents can provide an anti-apoptotic function (57)(58)(59). To evaluate whether NF-B activation affects CPT induced apoptotic responses, the levels of apoptosis were estimated by FACS analysis based on sub-G 0 /G 1 DNA content in WT and S32A/S36A expressing 70Z/3-CD14 cells. Untreated cells showed an expected pattern of cell cycle distribution for these cells with more than half of the cell population in S phase of the cell cycle ( Fig. 9, OT, WT). Expression of S32A/S36A mutant protein did not significantly affect the cell cycle status of untreated cells (OT, S32A/S36A). After treatment with 1 M CPT for 24 h, however, most of the cells appeared in either a G 2 /M or sub-G 0 /G 1 apoptotic peak (Fig. 9,  24 h, WT). The fraction of apoptotic peak was approximately twice as great in S32A/S36A expressing cells as WT expressing cells (24 h, S32A/S36A). Similar results were obtained with higher CPT doses or TPT treatments (not shown). Thus, these observations indicate that activation of NF-B retards some cancer cells from undergoing apoptosis. These findings demonstrate that CPT activation of NF-B can provide an anti-apoptotic activity. DISCUSSION The activity of NF-B depends on a series of reactions that releases it from an inhibitory complex in the cytoplasm and allows it to migrate to the nucleus. The elucidation of the individual steps within NF-B signaling cascades induced by a variety of structurally and functionally unrelated stimuli has revealed the use of both shared and unique components that may contribute to the diverse functions of this ubiquitous transcription factor (38, 60 -68). DNA-damaging agents represent a unique group of NF-B activators because their primary site of action is in the nucleus. In this study, we demonstrate that nuclear events arising from the DNA-damaging function of CPT and TPT are components of a NF-B signaling pathway that converges in the cytoplasm with events associated with signaling pathways induced by cytokines or LPS stimulation.
The DNA-damaging function of CPT in replicating cells is a multi-step process that initiates with intercalation of CPT into a covalent Topo I-DNA reaction intermediate. CPT stabilizes this transient intermediate, forming the cleavable complex with a SSB. The cleavable complexes and associated SSBs are mostly reversible until the cell undergoes replication, during which the replication fork collides with the cleavable complex and yields a DSB. Our results obtained by utilizing mutant CEM/C2 cells, pharmacological agents, and FACS enrichment of S phase cells indicate that S phase-dependent generation of DSB is essential for NF-B activation. Of note, however, is that CPT or TPT activation of NF-B is transient in all cell lines tested thus far. Similar dose-dependent and time course responses in lymphoid, fibroblastic, and epithelial cell lines suggest that a conserved activation mechanism may be involved. It has been demonstrated that CPT can induce degradation of Topo I enzyme by the ubiquitin-proteasome pathway causing marked reduction of Topo I enzyme within 2-4 h (69). Although this correlates well with the reduction of NF-B activation in the continual presence of CPT in the present study, substantial levels of DSBs induced by CPT can persist for as long as 24 h in SV40-transformed human skin fibroblast cells (SV40MRC5VI) and EJ30/8D human bladder carcinoma cells (70). Because the critical DNA lesion (i.e. DSBs) may remain despite declining levels of Topo I enzyme, it is unlikely that the reduction of Topo I enzyme is solely responsible for transient NF-B activation. It further implies that the mere presence of DSBs is insufficient to maintain NF-B activation. It is thus possible that event(s) downstream of DSBs or those coupled to cell cycle may be responsible for transient NF-B activation by CPT-related compounds. A recent study has implicated the involvement of the ataxia telangiectasia mutant protein in sustained activation of NF-B by CPT (81).
Enucleation studies demonstrated that the process of NF-B activation induced by CPT or TPT requires an intact nucleus. To our knowledge, this is the first demonstration of the lack of NF-B activation in enucleated cells. Although this may be implied from the demonstration that events associated with DNA damage are required for NF-B activation, mitochondria also contain DNA and CPT-sensitive Topo I enzyme (42). Studies utilizing L929 fibrosarcoma cells deficient for mitochondrial (DNA) and antimycin A, which increases the generation of reactive oxygen intermediates by inhibiting the electron transport chain, indicate that NF-B activation by TNF␣ requires reactive oxygen intermediates derived from mitochondria (71). Because NF-B is implicated as an important mammalian oxidative stress-responsive transcription factor (72), determination of the contribution of nuclear versus potential mitochondrial events was crucial for elucidating the NF-B activation mechanism induced by CPT and TPT. Our findings provide direct evidence that CPT-or TPT-induced DNA damage in the nucleus is a primary component of the signaling events required for NF-B activation. Although recent studies that utilized UV-C treatment of Xeroderma pigmentosa group A fibroblasts suggested the involvement of DNA damage in late stage NF-B activation (14), whether or not an intact nucleus is required for this late activation was not investigated. Previous studies demonstrated that UV activation of NF-B could efficiently take place in enucleated cells (20).
CPT induction of DNA damage translates into activation of a cytoplasmic signaling cascade that liberates active NF-B from the inhibitor protein, IB␣. We utilized well characterized mutants of signaling components within the cytokine-inducible NF-B signaling pathway to dissect the signaling cascade activated by CPT and TPT. Ser-to-Ala mutations at positions 32 and 36 of IB␣ disrupt inducible phosphorylation and prevent subsequent ubiquitination and degradation by the proteasome pathway (47)(48)(49). We showed that these mutants also prevent IB␣ degradation and activation of NF-B induced by CPT and TPT. We additionally demonstrated that dominant-negative IKK mutants that inhibit phosphorylation of IB␣ at these sites also prevent NF-B activation by TPT. IKK␣-, ␤-, or ␥-deficient cells fail to activate NF-B by CPT treatment. Mutation of Lys residues critical for the attachment of ubiquitin moieties further disrupts CPT-inducible IB␣ degradation. Together with pharmacological evidence using proteasome inhibitors, our findings show that CPT and TPT induction of IB␣ degradation is similar to that induced by cytokines, LPS, and several other signals (7, 8, 10, 11, 47-50, 73, 74). Our findings therefore demonstrate that nuclear DNA damage causes IKKdependent degradation of IB␣ in the cytoplasm. This activation may involve signal transfer from the nucleus to the cytoplasm. This type of nuclear-to-cytoplasmic signaling has also been suggested for the late stage NF-B activation induced by UV irradiation, which involves the production of an autocrine/ paracrine factor, interleukin-1␣ (14). A recent study has also implicated the involvment of the DNA-dependent protein kinase in NF-B activation by certain DNA damaging agents (82). Further definition of signaling components and reactions will help to determine whether NF-B activation by CPT indeed involves a nuclear-to-cytoplasmic signal transduction pathway.
CPT derivatives, including TPT, are utilized clinically as part of cancer therapy regimes (34,35). Recently, several studies have reported that NF-B may control expression of genes involved in the regulation of apoptosis (57)(58)(59)(75)(76)(77)(78). In particular, Wang et al. (58) have shown that NF-B activation by ionizing radiation and daunarubicin may have anti-apoptotic effects in HT1080 human fibrosarcoma cells. The same group recently showed that NF-B activation by CPT-11 can display similar anti-apoptotic effects in the above cell line (79). We have also shown that CPT activation of NF-B provides an anti-apoptotic function. NF-B-dependent survival of even a fraction of cancer cells after treatment with DNA-damaging agents, such as TPT or ionizing radiation, will likely lead to increased mutation rates and accelerated manifestation of malignancy. Moreover, it may also contribute to transformation of normal cells during the therapy. Understanding the mechanism(s) of NF-B activation, therefore, may help improve the current methods of cancer therapy by defining a resistance mechanism to Topo I inhibitors and potentially other clinically important DNA-damaging and NF-B-activating agents, such as ionizing radiation and Topo II inhibitors.