Adenomatous Polyposis Coli Determines Sensitivity to Histone Deacetylase Inhibitor–Induced Apoptosis in Colon Cancer Cells

Histone deacetylases (HDAC), together with histone acetyltransferases, regulate the acetylation of core nucleosomal histones, which is important for the transcription activity of the target genes (1). Abnormal HDAC activity has been associated with the development of many types of cancer (2, 3). Inhibitors of HDAC induce differentiation, growth arrest, and apoptosis in cancer cells, whereas they are relatively nontoxic to normal cells (2, 4, 5). The mechanism of apoptosis induced by HDAC inhibitors has not been clearly defined, although it has been recently reported that activation of death receptor pathway (6, 7), up-regulation of proapoptotic proteins Bmf and Bad (8, 9), down-regulation of antiapoptotic protein Bcl-2 (10), and activation of Ku70-inhibited Bax (11) might be involved in certain types of cancers. Several HDAC inhibitors are in clinical trials as anticancer agents (12). Valproic acid (VPA) is a standard treatment for a variety of forms of epilepsy and has been shown recently to inhibit HDAC at therapeutic concentrations (13–15). VPA is a well-tolerated drug (16) and has been shown to effectively induce apoptosis in several types of cancers (17–19). Suberoylanilide hydroxamic acid (SAHA) is a promising HDAC inhibitor and has shown antitumor activity in solid and hematologic tumors in clinical trials (20, 21).

The tumor suppressor gene adenomatous polyposis coli (APC) is frequently mutated in colorectal cancers (22, 23). APC is a key component of the β-catenin destruction complex (consisting of GSK-3β, axin, and APC) and involved in the Wnt signaling pathway (24). Axin and APC form a structural scaffold that allows GSK-3β to phosphorylate β-catenin. Phosphorylated β-catenin is subsequently degraded by the proteasome (24). Loss of wild-type APC expression results in the nuclear accumulation of β-catenin, which interacts with Tcf-4/Lef1 transcription factors to cause aberrant gene transcription and formation of cancer (25, 26). The role of APC as a tumor suppressor has been further supported in mice with APC mutations that develop multiple intestinal neoplasia (Min mice; ref. 27). Mutations of β-catenin are also frequently observed in colon cancers, including those expressing wild-type APC (28), providing a complementing mechanism of disruption of this important tumor-suppressing pathway.

Survivin is an antiapoptotic protein of the inhibitor of apoptosis family. Survivin blocks apoptosis by inhibiting caspases and antagonizing mitochondrial-dependent apoptosis (29). Survivin also regulates cell division through interaction with INCENP and Aurora B kinase (30). Survivin is expressed at high level in many types of cancers, but not in normal tissues from the same organs (29). Loss of p53 tumor suppressor has been associated with the deregulation of survivin expression in cancer cells (31). Survivin is a potential target gene regulated by APC; mutations in APC gene may cause up-regulation of survivin and initiation of colon cancer (32).

In the present study, we found that APC protein determines the relative sensitivity to HDAC inhibitor–induced apoptosis in colon cancer cells. Cells expressing wild-type APC are more sensitive to apoptosis-induced by VPA and SAHA, whereas cells expressing mutant APC are relatively resistant. APC mediates HDAC inhibitor–induced apoptosis through down-regulation of survivin. We also showed that cells expressing mutant APC can be sensitized to VPA-induced apoptosis through down-regulation of survivin with Flavopiridol. These studies have identified important regulators of HDAC inhibitor–induced apoptosis in colon cancer cells. The findings have clear clinical implications for colon cancer therapy with HDAC inhibitors.

Cells and transfection. Human colon cancer cell lines HCA-7 and SW620 were purchased from European Collection of Cell Cultures (Wiltshire, United Kingdom) and American Type Culture Collection (Manassas, VA), respectively. The cells were cultured in RPMI 1640 containing 10% fetal bovine serum. Human colon cancer HT-29/APC and HT-29/β-Gal cells were generously provided by Dr. Vogelstein and cultured as described (33). For transient transfection, plasmids were transfected into cells using LipofectAMINE Plus Reagent (Invitrogen) following the protocol of the manufacturer.

Drugs and chemicals. VPA was purchased from Sigma (St. Louis, MO) and dissolved at 2 mol/L in PBS. SAHA was purchased from Biovision (Mountain View, CA) and dissolved at 5 mmol/L in DMSO. In all studies, an equivalent amount of diluent (DMSO or PBS) was added to culture medium as a negative control. Proteasome inhibitor MG-132 was purchased from Calbiochem (San Diego, CA) and dissolved in DMSO at concentration of 10 mmol/L. Flavopiridol was obtained from Sanofi-Aventis Pharmaceuticals, Inc. and Drug Synthesis and Chemistry Branch of the National Cancer Institute, Bethesda, MD), and dissolved at 10 mmol/L in DMSO.

Plasmid construction. Human cDNA encoding full-length survivin gene was obtained by PCR amplification using an expressed sequence tag clone (I.M.A.G.E. clone ID 4477581) as template. Survivin cDNA was subcloned into phrGFP-C (Stratagene, La Jolla, CA) vector to express a hrGFP fusion protein. The fragment of Tcf-4 cDNA containing the β-catenin binding site, but lacking the DNA binding site, was obtained by reverse transcription-PCR using total RNA isolated from HT-29 cells. Two primers, GCCAAGCTTAATGCCGCAGCTGAACGGC and CGGCTCGAGCCTTTTTGGAGTCCTGATGC, were used for PCR amplification. The Tcf-4 cDNA fragment was subcloned into pCMV-Tag-2A mammalian expression vector (Stratagene) to express the dominant-negative Tcf-4 as a Flag-tagged protein.

Detection of apoptosis. In terminal deoxyribonucleotide transferase–mediated nick-end labeling assay (TUNEL), cells were fixed with 1% paraformaldehyde on ice for 1 hour and then washed with PBS and permeabilized with 70% ethanol at −20°C for 12 hours. Cells were then labeled with Guava TUNEL Assay reagents and analyzed on the Guava PCA microcytometer following the protocol of the manufacturer. In 4′,6-diamidino-2-phenylindole (DAPI) staining, green fluorescent protein (GFP), or GFP-survivin–expressing cells were fixed with PBS containing 3.7% formaldehyde, and stained with 0.5 μg/mL DAPI in PBS. The percentages of apoptotic cells were determined by confocal microscopy, counting GFP-positive cells having nuclear fragmentation and/or chromatin condensation.

Western blot analysis. Cells were lysed in radioimmunoprecipitation assay buffer (1% NP40, 0.5% sodium deoxycholate, 0.1% SDS in PBS). Complete protease inhibitor cocktail (Roche, Alameda, CA) was added to lysis buffer before use. Protein concentration was determined by Bio-Rad detergent-compatible protein assay (Bio-Rad, Hercules, CA). Protein samples were subjected to SDS-PAGE and transferred to nitrocellulose membrane. The membrane was blocked in 5% nonfat milk in PBS overnight and incubated with primary antibody and subsequently with appropriate horseradish peroxidase–conjugated secondary antibody. Signals were developed with enhanced chemiluminescence reagents (Amersham, Piscataway, NJ) and exposure to X-ray films.

For detection of the high molecular weight APC, electrophoresis was done in a 3% low melting point agarose gel in 0.1% SDS in Tris/borate buffer (89 mmol/L Tris-HCl/89 mmol/L boric acid/2 mmol/L EDTA). Proteins were transferred by capillary action to a polyvinylidene difluoride membrane in TBS/0.04% SDS as described (33). Antisurvivin monoclonal antibody and anti-β-catenin polyclonal antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-β-tubulin monoclonal antibody was purchased from Sigma. Anti-APC monoclonal antibody was purchased from Calbiochem.

Small interfering RNAs and transfection. Silencer validated small interfering RNAs (siRNA) and negative control siRNA were purchased from Ambion (Austin, TX). The sequences for survivin siRNA are as follows: sense 5′-GGAACAUAAAAAGCAUUCGtt-3′ and antisense 5′-CGAAUGCUUUUUAUGUUCCTC-3′. The sequences for APC siRNA are as follows: sense 5′-GGAAGUAUUGAAGAUGAAGtt-3′ and antisense 5′-CUUCAUCUUCAAUACUUCCtt-3′. Survivin siRNA was transfected into SW620 cells using X-tremeGENE siRNA transfection reagent (Roche) following the protocol of the manufacturer. APC siRNA was transfected into HCA-7 cells using OligofectAMINE transfection reagent (Invitrogen) following the protocol of the manufacturer. Cells were cultures and transfected in six-well plates (1 × 10⁵ per well) and the final siRNA concentrations were 100 nmol/L for survivin siRNA and 30 nmol/L for APC siRNA. Protein samples were collected at 48 hours after transfection for Western blot analysis. For apoptosis assay, drugs were added at 24 hours after siRNA transfection and cells were collected at 72 hours posttransfection.

Real-time PCR. The mRNA level of survivin was measured by real-time PCR using TaqMan Gene Expression assay from Applied Biosystems (Foster City, CA). Total RNA was isolated from HCA-7 cells using RNeasy kit (Qiagen, Valencia, CA). Five micrograms of total RNA were used in reverse transcription reaction. The cDNAs were used as templates to perform PCR on an Applied Biosystems 7500 real-time PCR system following the protocol of the manufacturer.

GSK-3β activity assay. HCA-7 cells were treated with VPA for 48 hours and lysed in immunoprecipitation buffer [20 mmol/L Tris/HCl, (pH 7.4); 50 mmol/L NaCl; 0.5% Triton X-100; protease inhibitors cocktail; Roche]. The lysates were immunoprecipitated with polyclonal anti-GSK-3β antibody (Santa Cruz Biotechnology). The immunoprecipitates were washed thrice with immunoprecipitation buffer. GSK-3β activity was measured by mixing the immunoprecipitates with recombinant Tau protein (10 ng/mL) in GSK-3β assay buffer [40 mmol/L HEPES (pH 7.2), 5 mmol/L MgCl₂, 5 mmol/L EDTA, 100 μmol/L ATP, 50 μg/mL heparin]. The mixture was incubated at 30°C for 1 hour with constant mixing rotations. Phosphorylation of Tau Ser³⁹⁶ was then measured with Tau [pS396] phosphoELISA kit (Invitrogen) following the protocol of the manufacturer.

Cell cycle analysis. Cell cycle distribution was analyzed using a Guava PCA microcytometer (Guava Technologies, Hayward, CA). Cells were fixed in 70% ethanol for 1 hour and labeled with Guava Cell Cycle Assay reagent and analyzed following the protocol of the manufacturer.

β-catenin/Tcf-4 signaling. TCF activity was measured with TCF reporter plasmid kit (Upstate, Charlottesville, VA). Cells were transfected with TOPFLASH or FOPFLASH reporter plasmids with LipofectAMINE Plus Reagent. Luciferase activity was measured at 24 hours after transfection or after drug treatments with a SpectraMax M5 Multi-Detection Microplate Reader. TCF activity was calculated as the ratio of TOP/FOP.

VPA induces apoptosis in cells expressing wild-type APC. Mutations in APC gene contribute to colorectal tumorigenesis and are frequently observed in sporadic colon cancers. To determine whether APC plays a role in colon cancer cell apoptosis in response to HDAC inhibitors, we selected HCA-7 cells (which express wild-type β-catenin and APC proteins; ref. 28) and SW620 cells (which express mutant APC protein; ref. 28) to compare their response with VPA. Both HCA-7 and SW620 cells express mutant p53 proteins. VPA induced significant level of apoptosis in HCA-7 cells, whereas SW620 cells were resistant when treated at the same doses (Fig. 1A). HT-29 colon cancer cells express two COOH-terminal–truncated mutant APC proteins and mutant p53 protein. Like SW620 cells, HT-29 cells are also relatively resistant to VPA-induced apoptosis (data not shown). To further determine if APC plays a critical role in VPA-induced apoptosis, we used genetically engineered HT-29 cells in which wild-type APC is expressed from a Zn²⁺-inducible transgene (33). Expression of APC induces apoptosis in HT-29 cells (33). However, when APC induction is incomplete due to deliberate addition of a suboptimal concentration of zinc (50 μmol/L instead of the usual 100 μmol/L), cells do not undergo apoptosis (Fig. 1B; ref. 33). To avoid apoptosis induced by APC expression alone, we used 50 μmol/L zinc to induce APC expression. After induction of wild-type APC, apoptosis was observed in HT-29/APC cells when treated with VPA (Fig. 1B). In contrast, the HT-29/β-Gal cells were relatively resistant. When Zn²⁺ was not added to the culture medium to induce APC expression, HT-29/APC cells showed comparable resistance to VPA-induced apoptosis (data not shown).

Knocking down of endogenous APC protein by siRNA blocks VPA-induced apoptosis. If APC is important to VPA-induced apoptosis, knocking down of endogenous APC expression should render HCA-7 cells to become resistant to VPA. We used an APC targeting siRNA to successfully knock down APC protein expression in HCA-7 cell and observed a reduction of apoptosis after VPA treatment (Fig. 1C). Cells transfected with a negative control siRNA remained sensitive to VPA-induced apoptosis.

Down-regulation of survivin is involved in APC-mediated apoptosis. It has been shown that APC regulates the expression of survivin (32). To further understand the mechanism of APC-mediated apoptosis after VPA treatment, we examined the expression of survivin. Exposure to 2 mmol/L VPA for 48 hours significantly down-regulated the level of survivin protein in HCA-7 cells, but not in SW620 cells (Fig. 2A). Down-regulation of survivin was also observed in HT-29/APC cells after induction of APC expression and treatment with VPA, but not in HT-29/β-Gal cells (Fig. 2B). We did not observe any changes in the levels of other apoptosis-regulatory proteins, such as Bax, XIAP, and tumor necrosis factor–related apoptosis-inducing ligand receptor DR-5, in VPA-treated HCA-7 cells (data not shown). Because all four cell lines express mutant p53 proteins, the down-regulation of survivin seems to be p53 independent. To determine whether down-regulation of survivin is responsible for VPA-induced apoptosis in cells expressing APC, we used a survivin-specific siRNA to selectively knock down endogenous survivin protein in SW620 cells. Down-regulation of survivin protein by siRNA sensitized SW620 cells to VPA-induced apoptosis, whereas the negative control siRNA showed no effects (Fig. 2C). To further confirm the role of survivin in VPA-induced apoptosis, we overexpressed survivin as a GFP fusion protein in HCA-7 cells. Overexpression of GFP-survivin protected the cells against VPA-induced apoptosis, whereas GFP alone had no effects (Fig. 2D).

GSK-3β/β-catenin/Tcf-4 pathway and proteasome mediate the down-regulation of survivin. As APC promotes β-catenin degradation, we further investigated whether the APC/GSK-3β/β-catenin/Tcf-4 signal transduction pathway is involved in the down-regulation of survivin in VPA-induced apoptosis. First, we examined whether down-regulation of survivin by VPA occurred at transcription level. Using real-time PCR, we found that treatment with 2 mmol/L VPA for 48 hours decreased survivin mRNA level by 56% in HCA-7 cells, compared with untreated cells (Fig. 3A). Next, we determined whether proteasome-dependent degradation is involved in the down-regulation of survivin in response to VPA. As shown in Fig. 3B, cotreatment with a proteasome inhibitor MG-132 (10 μmol/L) completely blocked the VPA-induced down-regulation of survivin protein in HCA-7 cells. To determine if VPA promotes the degradation of survivin protein, HCA-7 cells were treated with 20 μmol/L cycloheximide or 20 μmol/L cycloheximide plus 2 mmol/L VPA; degradation of survivin was determined by Western blotting. VPA exposure promoted survivin degradation in HCA-7 cells (Fig. 3C). GSK-3β phosphorylates β-catenin and promotes β-catenin degradation. Unlike RKO cells, which express wild-type APC and have very low level of β-catenin/TCF signaling (34), HCA-7 cells have relatively significant level of β-catenin/TCF activity (Fig. 4A). VPA exposure increased GSK-3β activity in HCA-7 cells (Fig. 4C). To determine the role of GSK-3β activity in survivin expression, we inhibited GSK-3β by lithium (15 mmol/L), a widely used selective GSK-3β inhibitor (35). Cotreatment with lithium blocked VPA-induced GSK-3β activity (Fig. 4C), VPA-induced down-regulation of survivin mRNA (Fig. 3A), and VPA-induced decrease of survivin and β-catenin proteins (Fig. 4B). VPA exposure also decreased β-catenin/TCF signaling and lithium cotreatment blocked this effect of VPA (Fig. 4C). Furthermore, transfection of a dominant-negative mutant of Tcf-4 in HCA-7 cells led to a significant reduction of β-catenin/TCF signaling and survivin protein level (Fig. 4D). These data indicate that VPA down-regulates survivin through GSK-3β, promoting β-catenin degradation (a process requires wild-type APC), and subsequently decreasing β-catenin/Tcf-4–regulated expression of survivin.

APC determines apoptosis induced by SAHA. To evaluate whether APC is critical for apoptosis induced by other HDAC inhibitors, we treated HT-29 cells with SAHA. Expression of wild-type APC sensitized HT-29/APC cells to SAHA-induced apoptosis (Fig. 5A), whereas the HT-29/β-Gal cells were relatively resistant. We also observed the down-regulation of survivin by SAHA (Fig. 5B) in HT-29/APC cells, but not in HT-29/β-Gal cells. These data indicate that like VPA, SAHA also induces apoptosis in colon cancer cells through the action of APC.

Flavopiridol enhances VPA-induced apoptosis in cells expressing mutant APC. It has been shown that Flavopiridol, an inhibitor of cyclin-dependent kinases, can down-regulate survivin protein by inhibiting its phosphorylation on Thr³⁴ and promoting its degradation (36). To overcome resistance to VPA-induced apoptosis in colon cancer cells expressing mutant APC, we tested whether Flavopiridol can down-regulate survivin in these cells and sensitize the cells to VPA-induced apoptosis. Treatment with Flavopiridol alone did not reduce the level of survivin protein in SW620 cells, even at high doses that killed over 50% of cells (Fig. 6A and data not shown). However, when cells were treated with VPA for 24 hours and followed by 24 hours of coexposure of 100 nmol/L Flavopiridol plus VPA, the level of survivin protein was significantly reduced (Fig. 6A). Sequential treatment with VPA first followed by Flavopiridol also significantly enhanced VPA-induced apoptosis in SW620 cells (Fig. 6B). Treatment with 100 nmol/L Flavopiridol alone did not induce significant level of apoptosis. In contrast, combination treatments with reverse sequence (100 nmol/L Flavopiridol for 24 hours followed by Flavopiridol plus VPA for 24 hours) or with both drugs added at the same time for 48 hours did not reduce the level of survivin protein, and did not enhance apoptosis in SW620 cells (Fig. 6C). This is consistent with a previous report describing the effect of Flavopiridol on survivin in MCF-7 cells (36), in which the sequential treatment with Adriamycin followed by Flavopiridol, rather than Flavopiridol alone, is effective at suppression of survivin. To understand the mechanism behind the sequential requirement for combination treatment, we examined the changes in cell cycle distribution after VPA and Flavopiridol exposures. As shown in Fig. 6D, treatment with VPA first followed by Flavopiridol caused significant G₁ arrest in SW620 cell, whereas treatment with Flavopiridol first followed by VPA did not change cell cycle distribution comparing with the control cells. Because survivin is expressed at high level only at G₂-M phase (37), arresting cells at G₁ phase should facilitate Flavopiridol-mediated degradation of survivin protein.

This study has shown that APC plays an important role in determining colon cancer apoptosis in response to HDAC inhibitors. Among the HDAC inhibitors, SAHA is at the most advanced stage of clinical trials for treatment of cancer (38, 39), having shown significant anticancer activity with little toxicity to normal tissues. VPA is the treatment of choice for a variety of types of epilepsy since the 1970s (13). VPA has been identified as a HDAC inhibitor and effective inducer of differentiation and apoptosis in cancer cells at concentrations well tolerated in epilepsy patients (20). The molecular mediators of HDAC inhibitor–induced apoptosis have not been clearly defined. In the present study, we found that wild-type APC protein is a critical determinant of apoptosis in colon cancer cells treated with VPA or SAHA. The colon cancer cells expressing wild-type APC are more sensitive to HDAC inhibitor–induced apoptosis, whereas the cells expressing mutant APC are more resistant.

Up to 85% of all sporadic colon cancers have mutations in the APC gene (24). Most of the patients who have wild-type APC often have mutations in β-catenin that results in constitutively active β-catenin/Tcf-4 signaling. Based on our data, the majority of the colon cancers that contain mutations in APC or β-catenin will be more resistant to apoptosis after treatment with HDAC inhibitors. Our data indicate that down-regulation of survivin plays an important role in APC-mediated apoptosis. Knocking down survivin protein by siRNA successfully bypassed the requirement of wild-type APC expression and induced apoptosis after VPA treatment (Fig. 2C). As efforts have already been developed to inhibit survivin by antisense and ribozymes (40–42) and by using survivin mutants to suppress wild-type survivin (41, 43), it is possible to enhance the therapeutic effects of HDAC inhibitors in colon cancer by targeting survivin. To this end, we have successfully overcome APC mutation–mediated resistance using a combination treatment strategy with VPA plus Flavopiridol, a small-molecule agent that promotes survivin degradation and sensitizes SW620 cells to VPA-induced apoptosis.

We have also identified the mechanisms of APC-mediated down-regulation of survivin in response to VPA treatment. GSK-3β phosphorylates β-catenin depending on the formation of a complex including APC and axin (24). Phosphorylation of β-catenin promotes its degradation by the proteasome (44). We have observed increased GSK-3β activity and a decrease of β-catenin protein and β-catenin/TCF signaling after VPA treatment and these effects of VPA were blocked by cotreatment with GSK-3β inhibitor lithium. Moreover, we observed a direct involvement of β-catenin in the regulation of survivin expression. Using a dominant-negative form of Tcf-4, we have shown that β-catenin regulates survivin expression through Tcf-4. This is in agreement with a previous report (32). Thus, the VPA-induced decrease of survivin transcription (Fig. 3A) is caused by the increased degradation of β-catenin. Because survivin is a short-lived protein with a half-life of ∼30 minutes and subjected to proteasome degradation (45), the net result is the decrease of survivin protein after VPA treatment.

Decrease of survivin alone is not enough to cause apoptosis in colon cancer cells, as we have shown in SW620 cells transfected with survivin-targeting siRNA. Apparently, other changes have occurred in colon cancer cells (presumably up-regulation of proapoptotic molecules) in response to HDAC inhibitors. However, these other molecular changes must require the decrease of survivin to lower the threshold for apoptosis induction.

In conclusion, the expression status of wild-type APC determines the relative sensitivity for colon cancer cells to undergo apoptosis in response to HDAC inhibitors. APC mediates HDAC inhibitor–induced apoptosis through down-regulation of survivin. Whether this is a common molecular pathway of HDAC inhibitor–induced apoptosis in other types of cancers remain to be determined. Taken together, the present findings have important implications for the clinical use of HDAC inhibitors in treating colon cancer. Colon cancer patients whose cancers express wild-type APC and β-catenin might respond more favorably due to the apoptosis induced by HDAC inhibitors. Meanwhile, agents targeting survivin, preferably small-molecule drugs like Flavopiridol, should be developed to enhance the activity of HDAC inhibitors. A more detailed understanding of the molecular pathways of APC-mediated apoptotic response to HDAC inhibitors will provide novel targets (such as GSK-3β) to improve therapeutic efficacy of HDAC inhibitors.

Grant support: Eagles Art Ehrmann Cancer Fund (B. Guo) and NIH Center of Biomedical Research Excellence grant 5P20RR015566-05.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Sheri T. Dorsam (Center for Protease Research, North Dakota State University, Fargo, ND) for help with real-time PCR and Dr. Bert Vogelstein (Johns Hopkins University, Baltimore, MD) for providing HT-29/APC and HT-29/β-Gal cells.