Ineffectiveness of histone deacetylase inhibitors to induce apoptosis involves the transcriptional activation of NF-kappa B through the Akt pathway.

Histone deacetylase (HDAC) inhibitors are emerging as a new class of anticancer agents for the treatment of solid and hematological malignancies. Although HDAC inhibitors induce cell death through an apoptotic process, little is known about the molecular events that control their effectiveness. In this study, we demonstrate that HDAC inhibitors are limited in their ability to induce apoptosis in non-small cell lung cancer (NSCLC) cell lines despite their ability to effectively inhibit deacetylase activity. Because the anti-apoptotic transcription factor NF-kappa B has been shown to be under the control of HDAC-mediated repression, we analyzed whether HDAC inhibitors activated NF-kappa B in NSCLC cells. HDAC inhibitors effectively stimulated endogenous NF-kappa B-dependent gene expression by up-regulating IL-8, Bcl-XL, and MMP-9 transcripts. The ability of HDAC inhibitors to increase NF-kappa B transcriptional activity was not associated with signaling events that stimulated nuclear translocation, but rather modulated the transactivation potential of the RelA/p65 subunit of NF-kappa B. The inhibition of HDAC activity was associated with the recruitment of the p300 transcriptional co-activator to chromatin in an Akt-dependent manner. Moreover, Akt directly phosphorylated p300 in vitro and was required for stimulating the transactivation potential of the co-activator following the addition of HDAC inhibitors. Selective inhibition of either the phosphoinositide 3-kinase/Akt pathway, or NF-kappa B itself blocked the ability of HDAC inhibitors to activate NF-kappa B and dramatically sensitized NSCLC cells to apoptosis following of the addition of HDAC inhibitors. Our study indicates that the ineffectiveness of HDAC inhibitors to induce apoptosis in NSCLC cancer cells is associated with the ability of these molecules to stimulate NF-kappa B-dependent transcription and cell survival.

Histone deacetylase (HDAC) 1 inhibitors, as their name implies, are compounds that block deacetylase activity in eukary-otic cells. The loss of HDAC activity results in diminished methylation-associated gene silencing and hyperacetylation of core histones, transcription factors, and proteins involved in transcription (1)(2)(3)(4)(5). Because transient acetylation of core histones has been shown to be a critical prerequisite for chromatin remodeling and gene transcription (6,7), HDAC inhibitors would be predicted to promote cell cycle transition and proliferation. Paradoxically, the loss of deacetylase activity following the addition of HDAC inhibitors has been shown to inhibit proliferation and to stimulate differentiation and apoptosis in transformed cells in vitro and in vivo (1,3,4). These observations would suggest that HDAC activity is important for maintenance of the transformed phenotype. This idea is supported by the fact that several of the class I HDAC molecules have been shown to be overexpressed in hematological and solid cancers, including lung cancer (8 -11). Moreover, oncogenic mutations or fusions involving retinoic acid receptors and retinoid X receptor greatly increase the affinity of HDAC-containing complexes, thus promoting the development of promyelocytic leukemias (1,3,4). Therefore, the maintenance of an HDAC activity may provide a survival and growth advantage to some types of cancers (12). Based on this evidence, a number of structurally different HDAC inhibitors are being evaluated as novel anti-neoplastic agents to treat solid and hematological malignancies (4). To date, the precise molecular mechanisms governing the cellular response and effectiveness of HDAC inhibitors are poorly understood and remain areas of intense research.
The transcriptional activity of nuclear factor B (NF-B), like many immediate early transcription factors, is regulated through acetylation (13). NF-B is an inducible transcription factor that plays a role in the expression of a variety of genes involved in immune and inflammatory responses, adhesion, cell cycle, and survival (14 -16). Classic NF-B, composed of a p50 and RelA/p65 heterodimer, is the most abundant form of NF-B in cells. In unstimulated cells, the IB inhibitor is physically associated with NF-B, thus blocking nuclear translocation; however, upon cellular stimulation, IB is phosphorylated by the IB kinase (IKK) complex (17). Once phosphorylated, IB is ubiquitinated and degraded by the 26 S proteasome, allowing NF-B to translocate to the nucleus to transcriptionally activate gene targets (14 -17). In addition to phosphorylating IB, IKK is also responsible for directly phosphorylating serine 536 within the transactivation domain of the RelA/p65 subunit of NF-B (18). Multiple laboratories, including our own, have shown that a variety of inducers of NF-B transcriptional activity, such as tumor necrosis factor (TNF), interleukin-1 (IL-1)␤, and the Ha-Ras oncoprotein activates NF-B through pathways involving phosphoinositide 3-kinase (PI3K), Akt, and IKK (19 -27). Importantly, these pathways were found to be critical for NF-B-dependent cell survival in response to TNF and the Ha-Ras oncoprotein (27)(28)(29). Although the exact mechanisms of how Akt stimulates the transactivation domain of RelA/p65 is not completely understood, IL-1␤ has been shown to activate pathways involving Akt, the histone acetyl transferases (HATs), p300 and CREBbinding protein (CBP), and the mitogen-activated protein kinase (MAPK) p38 (25).
The RelA/p65 subunit of NF-B has been shown to activate transcription by interacting with co-activators containing HAT activity, including p300/CBP, the steroid receptor-coactivator-1, and the p300/CBP-associated factor (p/CAF) (30 -35). In addition to recruiting HAT activity, NF-B also plays a role in chromatin remodeling by recruiting the ATP-dependent SWI/ SNF complex (36,37). Besides acetylating the N-terminal tails of core histones surrounding NF-B-regulated genes (38), HATs are also responsible for acetylating RelA/p65 directly (13). Recently, it has been proposed that acetylation of the RelA/p65 subunit is responsible for sustaining NF-B-dependent transcription (13). In addition to co-activators, RelA/p65 has also been shown to recruit co-repressor complexes that are believed to basally or actively repress NF-B-dependent transcription. To date, three members of the HDAC class I family of proteins have been shown to modulate NF-B transcriptional activity, HDAC-1, -2, and -3 (39 -41). HDACs are tethered to NF-B, in part, through their interaction with steroid mediator repressor of transcription or nuclear co-repressor (41,42). Therefore, the differential association of NF-B with co-activator and co-repressor proteins would be predicted to have profound consequences on many cellular processes, including transcription and cell survival.
Work from our laboratory and others have established that NF-B provides an anti-apoptotic function in many different cancer cell lines (15), including non-small cell lung cancer (NSCLC) (43,44). NSCLC, which constitutes over 80% of all newly diagnosed lung cancer, is extremely resistant to chemotherapeutic agents (45). Thus, novel treatment strategies are clearly needed for this deadly disease. Currently, there are several Phase I and II clinical trials involving the use of HDAC inhibitors to treat patients with lymphoma and advanced solid tumors, including lung cancer (8,47,48). Because NF-B has been shown to be transcriptionally regulated by pathways involving deacetylase-mediated repression (13, 39 -41), and because we have previously shown that NF-B provides a cell survival role in NSCLC cells, studies were undertaken to determine if NSCLC cells were susceptible to apoptotic pathways induced by HDAC inhibition. In this report, we present evidence that the pro-apoptotic action of HDAC inhibitors is blunted by their ability to stimulate NF-B gene expression in an Akt-dependent manner. Our work supports the hypothesis that the inhibition of HDAC activity modulates NF-B transcription through mechanisms involving Akt-dependent phosphorylation of p300, recruitment of p300 to chromatin, and modulation of transactivation potential of this co-activator complex. Inhibition of NF-B nuclear translocation or transcriptional activity significantly increased apoptosis in NSCLC cells following the addition of HDAC inhibitors. To our knowledge, we report herein the first evidence that NF-B provides a significant protective role against HDAC inhibitor-mediated apoptosis and suggest that a combined molecular targeting that inhibits both NF-B and HDAC activities may provide a significant anti-neoplastic therapy for patients with NSCLC and, potentially, other malignancies.

MATERIALS AND METHODS
Cell Culture, Reagents, and Plasmid Constructs-Human NSCLC lines (NCI-H157, NCI-H358, and NCI-A549) were obtained from the ATCC and grown in RPMI 1640 (Invitrogen, Carlsbad, CA) supplemented with 10% fetal calf serum (HyClone Laboratories, Logan, UT) and penicillin/streptomycin. The 3x-B luciferase (3x-B-Luc) reporter construct contains NF-B DNA binding consensus sites originally identified in the major histocompatibility complex class I promoter, fused upstream to firefly luciferase (49). Plasmids encoding FLAG-tagged p65 were previously described (50). The Gal-4 luciferase construct (Gal4-Luc) contains four Gal-4 DNA consensus binding sites, derived from the yeast GAL-4 gene promoter, cloned upstream of luciferase cDNA. The Gal4-p65 fusion protein has the yeast Gal-4 DNA-binding domain fused to the transactivation domain (TAD) 1 of RelA/p65-(521-551) and was previously described (51). Gal4-p300 and the GST-p300 fusion proteins containing various segments of p300 were previously described (52). Expression plasmids encoding dominant negative Akt (K 3 A) have been described previously (27). The IB␣ (C-21) antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The p65-and p300specific antibodies were obtained from Upstate Biotechnologies (Lake Placid, NY), the RNA polymerase II antibody was from Santa Cruz Biotechnology (Santa Cruz, CA), and the M2 FLAG-epitope tag and ␣-tubulin (T9026) antibodies were obtained from Sigma-Aldrich (St. Louis, MO). Phosphospecific antibodies to Akt(S473) and pan-Akt antibodies were obtained from Cell Signaling Technology (Beverly, MA). Sodium butyrate and trichostatin A (TSA) were obtained from Sigma-Aldrich.
Histone Deacetylase Assays-NSCLC cells, plated in 100-mm dishes 18 h prior to the start of the experiments, were left untreated or were cultured in the presence of either TSA (500 nM) or sodium butyrate (1 M). Eighteen hours following the addition of HDAC inhibitors, cells were harvested and nuclear extracts were isolated using previously described protocols (43,49). Nuclear protein samples were quantitated using the Pierce BCA protein assay dye reagent (Pierce, Rockford, IL). Extracts (50 g) were analyzed for HDAC activity using the histone deacetylase kit from BIOMOL (Plymouth Meeting, PA) as directed by the manufacturer. To ensure that the deacetylase activity observed in untreated cells was due to Class I and II HDAC-mediated activity, TSA was also added to nuclear extracts in vitro. Although not shown, all HDAC activity, observed in untreated NSCLC cells (Fig. 1A), was inhibited in vitro by the addition of TSA.
Transfection and Luciferase Reporter Assays-NSCLC cells at 40 -60% confluency were transiently transfected using Polyfect reagent (Qiagen, Valencia, CA) according to the manufacturer's instructions. Briefly, plasmid constructs (0.75-1 g of DNA/well of a 12-well plate) were diluted in serum-free media and mixed with the Polyfect reagent (4 l/well). Complexes were allowed to form for 10 min before the addition of complete media containing 10% fetal bovine serum. The cells were washed once with 1ϫ phosphate-buffered saline (PBS), and Polyfect-DNA complexes were added to the cells and placed in a humidified incubator at 37°C with 5% CO 2 . Six hours following the start of transfection, additional complete media containing the appropriated pharmacological agent for the experimental conditions were added to the cells. Twenty-four hours post transfection, cells were washed once with 1ϫ PBS and lysed in luciferase reporter buffer (Promega, Madison, WI). Cells were then rapidly frozen and thawed in liquid nitrogen and a 37°C water bath, respectively. Extracts were collected and cleared by centrifugation at 14,000 rpm. Protein concentrations were determined with the Pierce BCA protein assay dye reagent. Luciferase assays were performed using the substrate D-Luciferin, and relative light units were measured using an AutoLumat LB953 luminometer (Berthold Analytical Instruments). Luminescence was normalized to protein concentrations.
Adenovirus Infection-Ad-CMV and Ad-SRIB␣ is a replication-defective E1-deleted adenovirus expressing transgenes under the control of the CMV promoter. Recombinant SR-IB␣ adenovirus was constructed as previously described (50). Adenovirus was amplified in 293 cells and purified by banding in a cesium chloride density gradient. NSCLC cells (ϳ60% confluent) were infected with 100 plaque-forming units/cell in complete media overnight. Eighteen hours later media was replaced with fresh complete media alone or containing HDAC inhibitors.
Electrophoretic Mobility Shift Assays and Western Blot Analysis-Preparation of nuclear and cytoplasmic extracts and EMSAs were performed as described previously (43,49). Briefly, nuclear extracts were prepared at indicated times and incubated with [ 32 P]dCTP-labeled, double-stranded probe containing an NF-B consensus site from the class I major histocompatibility complex promoter. Labeled probe/nuclear complexes were incubated for 10 min at room temperature and separated on a 5% polyacrylamide gel. Subsequently, the gel was dried and exposed to x-ray film. Western blot analysis was performed by either analyzing 0.25 M Tris-HCl lysed cell extracts or cytoplasmic and nuclear proteins on a 10% SDS-polyacrylamide gel. Total protein (50 g) was separated by SDS-PAGE and transferred to nitrocellulose membranes (Schleicher & Schuell, Keene, NH). The indicated primary antibodies were incubated for 30 min, washed, and visualized by incubation with horseradish peroxidase-conjugated secondary antibodies and ECL chemiluminescent reagents (Amersham Biosciences, Piscataway, NJ).
Northern Blot Analysis-Logarithmically growing NSCLC cells were either left alone or were treated with HDAC inhibitors for 18 h. Twentyfour hours later, cells were either left alone ("No Add" in figures) or treated with HDAC inhibitors. Total RNAs were isolated using TRIzol reagent (Invitrogen). RNAs (15 g/lane) were resolved on a denaturing 1.8% agarose-formaldehyde gel, transferred to Hybond membrane (PerkinElmer Life Sciences, Boston, MA) and crossed-linked. Gene expression was determined by analyzing Northern blots with 32 P-labeled random probes generated from IL-8, Bcl-X L , MMP-9, or GAPDH cDNAs, and blots were hybridized with radiolabeled probes in Quickhyb (Stratagene, Cedar Creek, TX). After a 2-h hybridization, the blots were washed twice in 2ϫ SSC/0.1% SDS for 15 min at room temperature and twice in 0.1ϫ SSC/0.1% SDS for 15 min at 60°C. Northern blots were analyzed by autoradiography.
Caspase-3 and Apoptosis Assays-NSCLC cells, plated at 5 ϫ 10 5 cells per well in a six-well plate the night before, were left alone or treated with various HDAC inhibitors. After 24 h of incubation with HDAC inhibitors, cells were harvested, and the extent of apoptosis was determined by quantitation of nucleosomes released into the cytoplasm using the Cell Death Detection ELISA Plus kit (Roche Applied Science, Indianapolis, IN) according to the manufacturer's directions. Caspase-3 assays were performed as previously described (44).

HDAC Inhibitors Fail to Induce Apoptosis in NSCLC Cells-
Histone deacetylase enzymes reside within the nucleus where they affect numerous pathways involved with transcription, cell cycle regulation, differentiation, and apoptosis (1,3,4). Although it is currently not understood how the disruption of HDAC activity initiates apoptosis, HDAC inhibitors have been shown to induce eukaryotic cell death in vitro and in vivo (1,4). To determine whether selected doses of HDAC inhibitors effectively block deacetylase activity, the NSCLC cell lines H157, H358, and A549 were treated with the HDAC inhibitors trichostatin A (TSA) or sodium butyrate. Eighteen hours following the addition of the HDAC inhibitors, cells were harvested and nuclear extracts were analyzed for HDAC activity. Both TSA and sodium butyrate effectively inhibited (Ն99%) HDAC activity in the NSCLC lines tested at the drug doses selected (Fig.  1A). To determine if HDAC inhibitors were effective at inducing apoptosis in NSCLC cells, the same concentrations of these pharmacological agents were added to the culture media, and cellular nucleosome formation was assayed as a measurement of apoptosis. Although HDAC inhibitors effectively inhibited HDAC activity in NSCLC cells (Fig. 1A), these pharmacological molecules failed to significantly induce apoptosis (Fig. 1B). NSCLC cells were sensitive to apoptotic-inducing agents, be- cause the addition of the protein kinase C inhibitor staurosporine significantly induced nucleosome formation in these cells (Fig. 1B). These results suggest that HDAC inhibitors fail to significantly induce apoptosis in NSCLC cell lines despite their ability to effectively inhibit HDAC activity.
HDAC Inhibitors Activate Endogenous NF-B-dependent Gene Expression-Based on findings described in Fig. 1, we were interested in determining whether the addition of HDAC inhibitors up-regulated cell survival signals that abolished their potential pro-apoptotic effects. Because the anti-apoptotic transcription factor NF-B has been shown to be up-regulated following the inhibition of HDAC activity (13,39,40), we analyzed whether HDAC inhibitors activated NF-B in our system. NSCLC cells were transiently transfected with an NF-B-responsive reporter, and cells were treated with HDAC inhibitors. The addition of HDAC inhibitors consistently elevated NF-B-dependent gene expression, as measured by transient reporter gene assays ( Fig. 2A).
Consistent with these results, we found that HDAC inhibitors stimulated endogenous IL-8 gene transcription in the NSCLC cells analyzed (Fig. 2B). To ensure that the addition of HDAC inhibitors up-regulated IL-8 gene expression in an NF-B-dependent manner, H157 cells were treated with an adenovirus encoding either the dominant negative, super-repressor IB␣ (SR-IB␣) protein, or the green fluorescent protein control. The SR-IB␣ protein contains alanine residues rather than serine residues at amino acid positions 32 and 36 and therefore cannot be phosphorylated or degraded in response to cellular stimuli. Expression of the SR-IB␣ protein in cells completely sequesters NF-B in the cytoplasm, thus inhibiting NF-B-dependent gene expression (43,50,54). Cells expressing the dominant negative SR-IB␣ protein no longer displayed elevated IL-8 transcripts following the addition of sodium butyrate, suggesting that this HDAC inhibitor up-regulated the IL-8 gene in an NF-B-dependent manner (Fig. 2C). In addition to the IL-8 gene, which has been previously show to be under HDAC-mediated repression (13,40), we found that two other NF-B-regulated genes, namely the anti-apoptotic gene Bcl-X L and the metastasis-inducing gene MMP-9 were up-regulated following the addition of HDAC inhibitors (Fig. 2D). Interestingly, the addition of HDAC inhibitors did not up-regulate all NF-B-regulated genes. For example, the addition of HDAC inhibitors failed to stimulate the expression of A1/Bfl-1 or Bcl-3 (data not shown). Collectively, we found that several NF-Bregulated genes, including the anti-apoptotic gene Bcl-X L , were up-regulated following the inhibition of deacetylase activity.
HDAC Inhibitors Activate NF-B by Targeting the Transactivation Domain of the RelA/p65 Subunit-To determine the mechanisms by which HDAC inhibitors activate NF-B pathways, we first evaluated whether these pharmacological molecules stimulated signaling pathways that lead to IB␣ degradation and nuclear translocation of RelA/p65. In contrast to IL-1␤, a cytokine known to activate IB␣ degradation and RelA/p65 nuclear translocation, the addition of TSA or sodium butyrate failed to affect these proteins in a similar manner in H358 cells (Fig. 3, A and B). HDAC inhibitors also failed to increase NF-B DNA binding activity, as measured by electrophoretic gel mobility shift assays (EMSAs, Fig. 3C). Similar results were observed in H157 and A549 cells (data not shown). These results suggest that HDAC inhibitors fail to activate NF-B through mechanisms involving IB␣ degradation, RelA/ p65 nuclear translocation, or increased DNA binding.
Previously, our laboratory has shown that the transactivation (TAD)-1 region of RelA/p65 is critical for Ras and cytokineinduced activation of NF-B (25,27). Therefore, we analyzed whether HDAC inhibitors required the TAD1 region of RelA/ p65 to activate NF-B-dependent transcription. H157 and H358 cells co-expressing plasmids encoding p50 and the fulllength RelA/p65-(1-551) protein displayed elevated NF-B transcriptional activity after the addition of sodium butyrate or TSA ( Fig. 4A and data not shown). In contrast, cells co-expressing plasmids encoding p50 and the RelA/p65-(1-521), which lacks the TAD1 region, no longer activated NF-B transcription in response to sodium butyrate (Fig. 4A). These results suggest that the TAD1 region of RelA/p65 is critical for HDAC inhibitors to stimulate NF-B-dependent transcription.
To elucidate the importance of the TAD1 region of RelA/p65 for transactivation function in response to HDAC inhibitors, we utilized a plasmid encoding a Gal4-p65 fusion protein (51). This fusion protein contains sequences encoding the DNA-binding domain of the yeast Gal-4 transcription factor fused to se- quences encoding the TAD1 region of RelA/p65-(521-551). The Gal4-p65 fusion protein allowed us to focus specifically on the ability of HDAC inhibitors to target the TAD1 domain of RelA/ p65 without involving other parameters of NF-B regulation. Cells were co-transfected with an expression plasmid encoding Gal4-p65-(521-551) and with a Gal4-responsive luciferase reporter (Gal4-Luc). As shown in Fig. 4B, the Gal4-p65-(521-551) construct, containing only the TAD1 region, demonstrated significant transactivation potential in response to either TSA or sodium butyrate, as compared with untreated cells. These results suggest that HDAC inhibitors activated NF-B transcriptional activity by targeting the transactivation domain of the RelA/p65 subunit of NF-B and that the TAD1 region of RelA/ p65 is necessary for this effect. These results were specific to the p65 TAD1 region, because co-transfections with the Gal4 DNA-binding domain expression construct and the Gal4-Luc reporter were not responsive to butyrate treatment (data not shown).
Inhibition of HDAC Activity Stimulates the Transactivation Potential of RelA/p65 in an Akt-dependent Manner-Our laboratory has demonstrated that this serine/threonine kinase is required for up-regulating the transactivation potential of RelA/p65 in response to the Ras oncoprotein, TNF, and IL-1␤ (25,27). To determine if HDAC inhibitors stimulated NF-B transcriptional activity in an Akt-dependent manner, we used plasmids encoding dominant negative proteins to Akt. As shown in Fig. 5A, cells transfected with plasmids encoding the dominant negative Akt(KϾA) protein were unable to up-regulate the transactivation potential of the Gal4-p65-(521-551) construct following the addition of sodium butyrate. To expand these experiments, we repeated the Gal4-p65 experiments using a pharmacological inhibitor to the PI3K pathway, LY294002. The addition of LY294002 inhibited the sodium butyrate-induced transactivation potential of RelA/p65 (Fig.  5B). However, the addition of a pharmacological inhibitor to the mitogen-activated protein kinase p38 (SB203580) failed to block the ability of sodium butyrate to stimulate the transactivation potential of the Gal4-p65 protein (Fig. 5B). The inability of SB203580 to inhibit the transactivation potential of the Gal4-p65-(521-551) was surprising because we had previously shown that activated Akt required p38, in part, to mediate the transactivation potential of NF-B in response to TNF and IL-1␤ (25). In support of our findings, sodium butyrate was recently shown to up-regulate the G␣ i2 gene promoter through an MEK/ERK-dependent, but p38-independent pathway (55). The ability of sodium butyrate to up-regulate NF-B in an MEK/ERK-dependent manner was not evaluated, because we have previously demonstrated that the transactivation potential of the Gal4-p65-(521-551) TAD1 region does not require MEK/ERK activity (25). Our results suggest that HDAC inhibitors activate the transactivation potential of NF-B through mechanisms involving Akt that are distinct from signaling pathways requiring the p38 MAPK.

HDAC Inhibitors Stimulate the Recruitment of the p300 Histone Acetyltransferase to Chromatin and Elevate the Transactivation Potential of this Co-activator in an Akt-dependent
Manner-It has already been established that co-expression of Akt and either CBP or p300 co-activators synergistically increase the transactivation potential of NF-B following the addition of IL-1␤ (25). Based on this understanding, we were interested in determining whether there was a direct correlation between Akt and p300 and whether this effect could be modulated following the addition of HDAC inhibitors. As shown in Fig. 5C, the addition of sodium butyrate significantly increased the recruitment of p300 histone acetytransferase to the NF-B-regulated IL-8 promoter, as detected by chromatin immunoprecipitation assays. Importantly, the addition of the PI3K inhibitor LY294002 completely blocked the recruitment of p300 to the IL-8 promoter following the addition of sodium butyrate (Fig. 5C). This effect was specific to the PI3K pathway, because adenoviral expression of the Phosphoinositide 3-phosphatase phosphatase and tensin homolog (PTEN) tumor

FIG. 3. TSA and sodium butyrate fail to activate NF-B through IB␣ degradation, RelA/p65 nuclear translocation, or increased DNA binding.
A and B, H358 cells were either left untreated (No Add), treated with TSA (500 nM), or treated with sodium butyrate (500 M) for 18 h. As a positive control, cells were also treated with IL-1␤ (5 g/ml) for 10 min. Cells were harvested, and cytoplasmic and nuclear proteins were analyzed by Western blot analysis. A, cytosolic proteins were analyzed for IB␣ protein and for the ␤-tubulin-loading control. B, nuclear extracts were also analyzed by Western blot to assess nuclear levels of NF-B components. Analysis of RNA polymerase II was used as a nuclear protein loading control. C, nuclear extracts were assessed for enhanced NF-B activity by performing EMSA. Nuclear extracts, described in B, were incubated with a 32 P-labeled double-stranded oligonucleotide probe containing cis-elements recognized by either NF-B or Oct-1, and DNA-protein complexes were resolved on a non-denaturing polyacrylamide gel. Note: only the Oct-1-specific complex is shown while the free probe is not. NF-Band Oct-1-DNA complexes are indicated by arrows.
suppressor protein also inhibited sodium butyrate-induced recruitment of p300 to chromatin (data not shown). LY294002 also significantly decreased the chromatin-associated levels of RNA polymerase II, as compared with untreated cells (Fig. 5C). Consistent with data presented in Fig. 3C, the addition of HDAC inhibitor failed to increase chromatin-associated RelA/ p65 levels across the IL-8 promoter (Fig. 5C). Chromatin-associated Akt was detected in H358 cells in the absence of stimulus, but was not observed following the addition of sodium butyrate or following the addition of butyrate and LY294002 (Fig. 5C). Although we detected chromatin-associated Akt in unstimulated cells, the precise role of chromatin-associated Akt in the recruitment of p300 is currently not known. Moreover, from these studies it is not clear whether Akt-dependent signals culminate from the cytosol or from the nucleus. Because the addition of LY294002 blocked p300 from interacting with the IL-8 promoter, it suggests that HDAC inhibitors might require Akt to modulate p300 activity by regulating the recruitment of the co-activator to chromatin.
To determine if HDAC inhibitors could modulate p300 transactivation potential, we transiently co-transfected H358 cells with the Gal4-luciferase reporter and with expression plasmids encoding for Gal4-p300 fusion proteins spanning over various regions of the p300 protein (52). As shown in Fig. 5D, the addition of butyrate potentiated the transactivation potential of the Gal4-p300- (19 -2414) protein. Moreover, the butyratesensitive regions of p300 were associated with either the Nterminal region (1-596) or the C-terminal most region of p300 (1734 -2414) (Fig. 5D). To determine whether these elements were regulated by butyrate in a PI3K-dependent manner, cotransfections were repeated in the presence of the pharmacological inhibitor LY294002. As shown in Fig. 5D, the addition of LY294002 blocked the ability of butyrate to stimulate the transactivation potential of the Gal4-p300-(1734 -2414) protein but did not significantly affect the ability of butyrate to activate the Gal4-p300-(1-596) fusion protein. These results suggest that the PI3K signaling pathway might directly regulate the transactivation potential of the p300 co-activator following the inhibition of HDAC activity.
To determine if Akt could directly phosphorylate p300, we performed in vitro kinase assays on GST-p300-purified proteins. As shown in Fig. 5E, activated Akt kinase was able to phosphorylate p300-(1709 -1913) protein as compared with inactive Akt kinase. The ability of activated Akt kinase to phosphorylate GST-p300 was not due to the GST tag, because Akt could not phosphorylate GST-purified protein alone but strongly phosphorylated the glycogen synthase kinase 3-positive control (Fig. 5E). In conclusion, these experiments support the hypothesis that the inhibition of HDAC activity modulates NF-B transcription through mechanisms involving Akt-dependent phosphorylation of p300, recruitment of p300 to chromatin, and modulation of the transactivation potential of this co-activator complex.
NF-B Provides Protection from Apoptosis Induced by HDAC Inhibitors-Because we found that HDAC inhibitors significantly up-regulated NF-B transcriptional activity involving the PI3K/Akt pathway (Figs. 2, 4, and 5), we were interested in determining whether the increase in NF-B activity following the loss of deacetylase activity could protect cells from the pro-apoptotic potential of the HDAC inhibitors. This is an important question, because little is known about the molecular responses within cancer cells that control cell fate following HDAC inhibition. To experimentally address this, we utilized H157 cells stably expressing the dominant negative SR-IB␣ protein (H157I) or cells expressing the empty vector control (H157V) (Fig. 6A) (43). Expression of the dominant negative IB␣ protein significantly inhibited IL-1␤-induced nuclear translocation and DNA binding of NF-B in H157I cells but not in the H157V control cells (Fig. 6A). Consistent with the inhibition of NF-B translocation to the nucleus, we found that H157I cells were no longer capable of activating NF-B transcription following the addition of either IL-1␤ or in response to HDAC inhibitors (Fig. 6B). Importantly, the addition of TSA to H157I cells significantly increased the ability of the HDAC inhibitors to induce apoptosis, as compared with vector control cells (Fig. 6C, right panel). Moreover, our ability to detect DNA fragmentation following the addition of TSA correlated with an increase in caspase-3 activity in these cells (Fig. 6C, left panel). Similar results were obtained when cells were treated with sodium butyrate (data not shown). Collectively, these results suggest that the ability of NSCLC cells to up-regulate NF-B transcriptional activity in response to HDAC inhibitors results in enhanced cell survival following loss of deacetylase activity.
Inhibition of NF-B Directly or with Pharmacological Inhibitors to the PI3K Pathway Sensitizes Cells to Apoptosis following the Addition of HDAC Inhibitors-To ensure that the results found in Fig. 6 were not due to clonal effects observed following the generation of the H157I cell line, experiments were expanded using several NSCLC cell lines. To inhibit NF-B transcriptional activity, we used adenovirus encoding the SR-IB␣ protein (Ad-SRIB␣) or the vector control (Ad-CMV). Although the inhibition of NF-B alone was not proapoptotic in either H157 or A549 cells, H358 cells were more sensitive and displayed elevated nucleosome activity following the expression of the SR-IB␣ protein alone (Fig. 7A). Despite this effect, the inhibition of NF-B activity significantly enhanced apoptosis in all NSCLC cell lines following the addition of HDAC inhibitors, as compared with control nontreated cells (Fig. 7A). The differences in cellular response to the SR-IB␣ was not due to differences in transgene expression, because all NSCLC cells analyzed effectively displayed the SR-IB␣ protein (Fig. 7A). To ensure that the inhibition of NF-B activity sensitized NSCLC cells to apoptosis following the addition of HDAC inhibitors, experiments performed in Fig. 7A were repeated in the presence of the broad caspase-inhibitor Boc-D. Boc-D blocked sodium butyrate-induced apoptosis in H157 cells, following the expression of the SR-IB␣ protein (Fig. 7B). These results strongly suggest that the inhibition of NF-B activity sensitizes NSCLC cells to caspase-mediated apoptosis following the addition of HDAC inhibitors.
Because we showed that the PI3K/Akt pathway is critical for the ability of HDAC inhibitors to activate the transactivation potential of RelA/p65 (Fig. 5), we were interested in determining if the inhibition of Akt-dependent NF-B transcriptional activity using the LY294002 compound would sensitize NSCLC cells to apoptosis following the addition of HDAC inhibitors. As shown in Fig. 7C, the inhibition of the PI3K pathway using the LY294002 compound significantly enhanced sodium butyrateinduced apoptosis. Collectively, these results indicate that the ability of HDAC inhibitors to activate NF-B-dependent transcription significantly diminishes the ability of these agents to induce apoptosis in NSCLC cells. Moreover, pharmacological inhibitors to the PI3K pathway are as effective as the SR-IB␣ protein at blocking the ability of HDAC inhibitors to stimulate the transactivation potential of NF-B and sensitizing cells to undergo apoptosis. DISCUSSION In this study we demonstrate that the treatment of NSCLC cells with HDAC inhibitors stimulates NF-B transcriptional activity and that NF-B is required to overcome apoptosis following the inhibition of deacetylase activity. To our knowledge this is the first study to demonstrate that the inhibition of NF-B activity greatly facilitates apoptosis in response to HDAC inhibitors. To better understand how the inhibition of HDAC activity stimulates NF-B-dependent cell survival pathways, we show that this process stimulates the transactivation potential of the RelA/p65 subunit of NF-B. This is biologically significant, because HDAC inhibitors up-regulated endogenous NF-B gene expression. The inhibition of HDAC activity in NSCLC cells up-regulated NF-B transcriptional activity through mechanisms involving Akt and chromatin-associated recruitment of p300 to NF-B-regulated promoters. Moreover, the inhibition of the PI3K/Akt pathway, which our group has shown to be required to stimulate the transactivation potential of NF-B (25,27,56), was required to overcome apoptosis following the addition of the HDAC inhibitors. Finally, the inhibition of NF-B and subsequent treatment of NSCLC cells with HDAC inhibitors greatly potentiated apoptotic cell death through a caspase-dependent mechanism. Our work supports the hypothesis that the ineffectiveness of HDAC inhibitors to induce apoptosis can be attributed to the ability of cancer cells to respond to these agents by up-regulating NF-B-dependent cell survival pathways. This mechanism may explain why the anti-neoplastic effects of HDAC inhibitors in vitro are highly variable in many different cancer cell lines (1,4).
Although HDAC inhibitors have been shown to induce apoptosis both in vitro and in vivo, the precise mechanisms governing the pro-apoptotic action of these molecules is currently unknown. One of the best-supported hypotheses includes the up-regulation of genes involved in growth control or apoptosis, through the modification of the chromatin condensation and loss of DNA hypermethylation (1). Because the addition of HDAC inhibitors would be predicted to de-repress transcriptional targets, it has been suggested that HDAC inhibitors might function to up-regulate gene products involved in cell cycle arrest and/or apoptosis. In support of this, the p53 tumor suppressor protein becomes transcriptionally elevated in vitro and in vivo following the addition of HDAC inhibitors, and TSA-induced apoptosis has been shown to involve p53 (55,57,58). Despite this evidence supporting p53-dependent apoptosis governing HDAC inhibitor action, it is clear that p53-independent pathways exist (59). In our model systems, we did not observe significant differences in apoptosis induced by the addition of sodium butyrate or TSA in H157, H358, or A549 cells ( Fig. 1B and 6A), despite the fact that only A549 cells express functional wild-type p53 protein (60). Therefore, although p53 status may play a significant role in the response of primary malignancies to HDAC inhibitors in vivo (58), our studies suggest that the status of p53 does not augment apoptosis in FIG. 6. HDAC inhibition increases caspase-3 activity and DNA fragmentation in NSCLC cells lacking NF-B activity. A, left panel, whole cell proteins were isolated from H157 cells stably expressing a plasmid encoding the SR-IB␣ protein (H157I) or vector control (H157V). Western blot analysis was performed to detect the FLAG-tagged SR-IB␣ protein.
Right panel, IL-1␤-stimulated nuclear extracts isolated from H157V and H157I cells were assessed for NF-B DNA binding using EMSA. B, H157V and H157I cells were transfected with 3x-B-Luc reporter, and luciferase activities were determined following a 12-h stimulation with either TSA or sodium butyrate. C, H157V and H157I cells were incubated in the presence or absence of TSA (500 nM). Cells were harvested 24 h following the addition of TSA and analyzed for caspase-3 activity and DNA fragmentation levels. Cell extracts were analyzed for caspase-3 activity using fluorogenic substrate and assaying at A 405 . Histoneassociated DNA fragments were detected using the Cell Death Detection ELISA Plus kit (Roche Applied Science, Indianapolis, IN). NSCLC cells following the loss of deacetylase activity.
Recent evidence indicates that the addition of HDAC inhibitors to cancer cells stimulates the re-expression of tumor suppressor genes repressed by DNA methylation (1). It is now recognized that the methyl-CpG-binding protein (MeCP2) and methyltransferases interact directly with HDACs to transcriptionally inactivate chromatin (61,62). The idea that HDAC inhibitors function in this manner is supported by the observation that TSA-mediated apoptosis is greatly augmented in the presence of methyltransferase inhibitors in several lung cancer cell lines (62). Although it has been estimated that less than 2% of genes become activated following the addition of HDAC inhibitors (4), several genes encoding tumor suppressor proteins and pro-apoptotic proteins have been shown to be under CpG methylation and HDAC-mediated repression, including p16 INK4a , p14 ARF , p15 INK4b , hMLH1, p73, pRB, BRCA1, VHL, APC, and others (63). Therefore, it is not surprising that DNA methyltransferases and HDACs are commonly overexpressed in solid malignancies, including lung cancer (8 -11, 58). These findings support the hypothesis that maintenance of deacetylase activity provides a selective cell growth and survival advantage to cancer cells.
Similar to p53, CBP and p300 HAT proteins are important for maintaining genome integrity and are molecular targets of HDAC inhibitors (64). This is the first report that provides evidence that HDAC inhibitors modulate p300 activity in a PI3K/Akt-dependent manner. Our work suggests that Akt can directly phosphorylate p300 in vitro (Fig. 5E) and that Akt activity is required for modulation of the transactivation potential of this co-activator following the addition of HDAC inhibitors (Fig. 5D). It is not clear how Akt up-regulates the transactivation potential of p300. However, Akt-mediated phosphorylation of the p300 protein may be important for the subsequent assembly of transcriptional co-activators, because the Akt-responsive region is in close proximity to the zinc finger domain responsible for recruitment of p/CAF, TFIIB, and RNA helicase A (64). In support of this hypothesis, the same segment of p300 that we found to be Akt-responsive maps to the adenoviral E1A protein-interacting domain of p300, which is critical for NF-B-mediated transcription (46). Phosphorylation-dependent regulation of p300 within this same region may be a common mechanism for up-regulating the transactivation potential of p300, because the MAPK/ERK kinase kinase (MEKK)-1 has recently been shown to phosphorylate p300 and to modify the transactivation potential within this domain (52). Therefore, it may be that Akt-dependent phosphorylation of p300 is responsible for stimulating the subsequent recruitment of both basal transcription factors, as well as other HAT-containing complexes such as p/CAF and members of the p160 family. Because of this newly discovered link between Akt and p300, future experiments will elucidate the mechanisms by which Akt up-regulates the transactivation potential of the p300 co-activator as a way in which HDAC inhibitors up-regulate NF-B transcriptional activity.
In conclusion, HDAC inhibitors are one of the novel molecular-based anti-cancer therapies currently undergoing clinical trial evaluation for patients with NSCLC (48). In this report we demonstrated that HDAC inhibitors fail to induce apoptosis in vitro in NSCLC cells, in part by the ability of these agents to stimulated NF-B-dependent transcription and cell survival. FIG. 7. Loss of NF-B by the expression of the SR-IB␣ protein or by the PI3K inhibitor LY294002 sensitizes NSCLC cells to butyrate-induced apoptosis. A, H157, H358, and A549 cell lines were transiently infected with virus encoding the mutant IB␣ protein (Ad-SR-IB␣) or a vector control (Ad-CMV). Cells were either left untreated (No Add) or treated with TSA (500 nM) or sodium butyrate (500 M) for 24 h. As described earlier, apoptosis was measured by the Cell Death Detection ELISA Plus kit. Western blot analysis was performed to detect the HA-tagged SR-IB␣ protein at 24 h post-infection. Lane 1, whole cell extracts from cells that received Ad-CMV virus; lane 2, whole cell extracts form cells infected with Ad-SRIB␣ virus. B, experiments described in Fig. 6A were repeated, however, H157 cells were left untreated or were treated with the caspase inhibitor Boc-D at the same time that sodium butyrate was added to the cells. Apoptosis was assayed as described before. C, H157 cells were pretreated with either LY294002 or SB203580 2 h before the addition of sodium butyrate. DNA fragmentation was analyzed 24 h following the addition of the HDAC inhibitor.
The observation that the pro-apoptotic affects of HDAC inhibitors are inhibited by the concomitant activation of the prosurvival transcription factor NF-B may have significant clinical relevance. Finally, these studies provide a better understanding of the molecular mechanisms by which HDAC inhibitors activate NF-B-dependent cell survival and suggest that the inhibition of NF-B in conjunction with HDAC inhibitors may provide a novel approach for the treatment of human cancers.