Potential conflict of interest: Nothing to report.
Histone deacetylase (HDAC) inhibitors exhibit a unique ability to degrade topoisomerase (topo)IIα in hepatocellular carcinoma (HCC) cells, which contrasts with the effect of topoII-targeted drugs on topoIIβ degradation. This selective degradation might foster novel strategies for HCC treatment in light of the correlation of topoIIα overexpression with the aggressive tumor phenotype and chemoresistance. Here we report a novel pathway by which HDAC inhibitors mediate topoIIα proteolysis in HCC cells. Our data indicate that HDAC inhibitors transcriptionally activated casein kinase (CK)2α expression through increased association of acetylated histone H3 with the CK2α gene promoter. In turn, CK2 facilitated the binding of topoIIα to COP9 signalosome subunit (Csn)5 by way of topoIIα phosphorylation. Furthermore, we identified Fbw7, a Csn5-interacting F-box protein, as the E3 ligase that targeted topoIIα for degradation. Moreover, knockdown of CK2α, Csn5, or Fbw7 reversed HDAC inhibitor-induced topoIIα degradation. Mutational analysis indicates that the 1361SPKLSNKE1368 motif plays a crucial role in regulating topoIIα protein stability. This motif contains the consensus recognition sites for CK2 (SXXE), glycogen synthase kinase (GSK)3β (SXXXS), and Fbw7 (SPXXS). This study also reports the novel finding that topoIIα may be a target of GSK3β phosphorylation. Evidence suggests that CK2 serves as a priming kinase, through phosphorylation at Ser1365, for GSK3β-mediated phosphorylation at Ser1361. This double phosphorylation facilitated the recruitment of Fbw7 to the phospho-degron 1361pSPKLpS1365 of topoIIα, leading to its ubiquitin-dependent degradation. Conclusion: This study shows a novel pathway by which HDAC inhibitors facilitate the selective degradation of topoIIα, which underlies the complexity of the functional role of HDAC in regulating tumorigenesis and aggressive phenotype in HCC cells. (Hepatology 2011;)
Hepatocellular carcinoma (HCC) is a leading cause of cancer death worldwide. The clinical management of HCC is complicated by typically late-stage disease at presentation and prevalent underlying liver dysfunction that can render patients ineligible for potentially curative surgical therapies, which are generally suitable for only 20%-30% of HCC patients.1 Although regional therapies, such as transarterial embolization and percutaneous treatments, are used in patients with nonresectable disease, their success is curtailed by recurrence as locally advanced or metastatic disease.2 For these patients, systemic therapies are indicated but have been largely unsuccessful, in part, due to cellular resistance to conventional cytotoxic agents.3, 4 Thus, a clear need exists to develop effective, life-prolonging therapeutic strategies for the large number of HCC patients with advanced disease.5
Previously, we demonstrated that the novel phenylbutyrate-derived histone deacetylase (HDAC) inhibitor AR42 (formerly OSU-HDAC42) exhibited high in vivo potency in suppressing HCC tumor growth, which was attributable to its ability to target both histone acetylation-dependent and -independent pathways.6 In addition to HDAC inhibition, AR42 also blocked the phosphorylation/expression level of a series of apoptotic regulators, including Akt, Bcl-xL, survivin, cIAP1, and cIAP2. Here we show that AR42 facilitates the proteasomal degradation of topoisomerase (topo)IIα without disturbing topoIIβ expression in HCC cells, which was also noted with MS-275, a class I HDAC inhibitor, and, to a lesser extent, vorinostat (suberoylanilide hydroxamic acid).
The unique ability of HDAC inhibitors to degrade topoIIα contrasts with the selective effect of topoII-targeted drugs on topoIIβ degradation,7, 8 and may foster novel strategies for HCC treatment considering the correlation of topoIIα overexpression with the aggressive tumor phenotype and chemoresistance.9, 10 Moreover, topoIIβ may underlie many of the side effects associated with topoII-targeted drugs, such as doxorubicin-induced cardiotoxicity11 and etoposide-induced secondary malignancies.12 From a mechanistic perspective, HDAC inhibitors provide a useful tool to elucidate the pathways governing topoIIα degradation, which represents the focus of this study.
PLC5 and HepG2 cells were obtained from the American Type Culture Collection (Manassas, VA), and Huh7 cells were from the Health Science Research Resources Bank (Osaka, Japan). These HCC cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Invitrogen). All cells were cultured at 37°C in a humidified incubator containing 5% CO2. The HDAC inhibitors vorinostat, MS-275, and AR42 (OSU-HDAC42)6, 13, 14 were synthesized in our laboratory with purities exceeding 99%. MG132, wortmannin, PD98059, SB202190, SB216763, and DMAT were purchased from Sigma-Aldrich (St. Louis, MO). Bay11-7082, cycloheximide and GF-109203X were from Calbiochem (San Diego, CA). Antibodies against various proteins were from the following sources: topoIIα, BD Transduction (San Diego, CA); topoIIβ, casein kinase (CK)2α, Ets-1, ubiquitin, hemagglutinin, HDAC1, and HDAC6, Santa Cruz Biologicals (Santa Cruz, CA); Fbw7, Bmi1 and Skp2, Invitrogen; Fbx4, Rockland (Gilbertsville, PA); Fbx7, ProteinTech (Chicago, IL); acetylated lysine, HDAC4, HDAC5 and GSK3β, Cell Signaling Technology (Danvers, MA); Flag, Sigma-Aldrich; β-actin, MP Biomedicals (Irvine, CA); COP9 signalosome subunit (Csn)5, GeneTex (Irvine, CA); p-Ser/Thr, Abcam (Cambridge, MA); acetyl-histone H3 and HDAC2, Millipore (Billerica, MA). Goat antirabbit and rabbit antimouse immunoglobulin G (IgG)-horseradish peroxidase conjugates were from Jackson Laboratories (West Grove, PA).
Transient Transfection and Immunoblotting.
PLC5 cells were transfected with Lipofectamine 2000 (Life Technologies, Gaithersburg, MD) according to the manufacturer's protocol. Plasmids and RNA interference were obtained from the following sources: short-hairpin RNA (shRNA) constructs against HDAC1, HDAC2, HDAC6, and CK2α, and plasmids encoding CK2α and Csn5, Origene (Rockville, MD); small interfering RNAs (siRNAs) against Csn5, HDAC4, and HDAC5, Invitrogen; Fbw7 shRNA and hemagglutinin-GSK3β plasmid, Addgene. Immunoblotting was performed as described.14
Cells were treated with AR42 for 48 hours and lysed by buffer B (5 mM HEPES, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, 26% glycerol [v/v], 300 mM NaCl, pH 7.9) on ice for 1 hour. After centrifugation at 13,000g for 20 minutes, one-tenth volume of supernatant was stored at 4°C for use as input and the remainder was incubated with protein A/G-Sepharose beads for 1 hour to eliminate nonspecific binding. The mixture was centrifuged at 1,000g for 5 minutes and the supernatants were incubated with anti-topoIIα antibodies and protein A/G Sepharose overnight. The immunocomplexes were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and proteins were detected with indicated antibodies.
Chromatin Immunoprecipitation (ChIP) Assay.
PLC5 cells were treated with AR42 for 36 hours and fixed in 1% formaldehyde for 15 minutes to immobilize histone to DNA. Cross-linking was stopped with 125 mM glycine for 5 minutes. ChIP was performed as previously described6 using antibodies against acetyl-histone H3 or Ets-1 with nonspecific rabbit IgG as negative control. Primers spanning the proximal promoter regions of CK2α were used for amplification by reverse-transcription polymerase chain reaction (RT-PCR): 5′-GGGGATTCCTTCCATTTTGC-3′/5′-ATG GAGGAGGAGACACACGG-3′.
Total RNA was isolated from drug-treated cells with Trizol reagent (Invitrogen) and chloroform extraction. Aliquots of 2 μg of total RNA were reverse-transcribed to cDNA with the iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA) according to the manufacturer's instructions. PCR products were resolved by agarose (1.2%) gel electrophoresis and visualized by ethidium bromide staining. The sequences of primers used were as follows: topoIIα, 5′-CGTCAGAACATGGACCCAGA-3′/5′-AGC AGATTCAGCACCAAGCA-3′; CK2α, 5′-TGAGGAT AGCCAAGGTTCT-3′/5′-CAGCAGCAATCACTGG TGA-3′; GAPDH, 5′-AGGGGTCTACATGGCAAC TG-3′/5′-CGACCACTTTGTCAAGCTCA-3′; HDA C1, 5′-CCACATCAGTCCTTCCAATA-3′/5′-TTCT CCTCCTTGGTTTTCTC-3′; HDAC2, 5′-GCTAT TCCAGAAGATGCTGTTC-3′/5′-GTTGCTGAGCTG TTCTGATTTG-3′; HDAC4, 5′-CAGATGGACTTTC TGGCCG-3′/5′-CTTGAGCTGCTGCAGCTTC-3′; HDAC5, 5′-AGAATGGCTTTACTGGCTCAG-3′/5′-CATGAGCACATCCTCTATTCC-3′; HDAC6, 5′-CAACTGAGACCGTGGAGAG-3′/5′-CCTGTGCG GAGACTGTAGC-3′.
Plasmid Construction and Site-Directed Mutagenesis.
Plasmids encoding various topoIIα mutations were generated from Flag-TopoIIα (GeneCopoeia, Rockville, MD) by site-directed mutagenesis using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Primers used to generate topoIIα mutations were as follows: S1361A, 5′-TGCTAGTCCAC CTAAGACCAAAACTGCCCCAAAACTTAG-3′/5′-C TAAGTTTTGGGGCAGTTTTGGTCTTAGGTGGA CTAGCA-3′; S1365A, 5′-GAC-CAAAACTTCCCCA AAACTTGCTAACAAAGAACTGAAACCACAG-3′/5′-CTGTG-GTTTCAGTTCTTTGTTAGCAAGTT TTGGGGAAGTTTTGGTC-3′; E1368A, 5′-CCC CAAAACTTAGTAACAAAGCACTGAAACCACAGA AAAGTGT-3′/5′-ACAC-TTTTCTGTGGTTTCAGT GCTTTGTTACTAAGTTT-TGGGG-3′; S1393A, 5′-GGGC-AGTGTACCACTGTCTTCAGCCCCTCCT GCTAC-3′/5′-GTAGCAGGAGGGGCTGA-AGACAG TGGTACACTGCCC-3′; T1397A, 5′-CTTCAAGCC CTCCTGCTGCACATT-TCCCAGATGAA-3′/5′-TTC ATCTGGGAAATGTGCAGCAGGAGGGCTTGAAG-3′.
In Vivo Mechanistic Validation.
Female athymic nude mice (5-6 weeks of age) were obtained from Harlan Laboratories (Indianapolis, IN). All experimental procedures were done according to protocols approved by the Ohio State University Institutional Laboratory Animal Care and Use Committee. Each mouse was injected subcutaneously with 1 × 106 PLC5 cells in 0.1 mL serum-free medium containing 50% Matrigel. Mice with established tumors (mean starting tumor volume, 223 ± 75 mm3) were randomized to two groups (n = 5) that received the following treatments daily by gavage (10 μL/g body weight) for 3 or 6 days: (1) methylcellulose/Tween 80 vehicle, and (2) AR42 at 25 mg/kg. At the study endpoint, tumors were snap-frozen and stored at −80°C for subsequent coimmunoprecipitation analysis.
Differential Suppression of TopoIIα Expression by HDAC Inhibitors.
Pursuant to our finding that AR42 exhibits high in vivo efficacy against PLC5 tumor growth,6 we examined the effects of AR42 on various biomarkers pertinent to the aggressive phenotype of HCC, among which the concentration- and time-dependent suppression of topoIIα expression was noteworthy (Fig. 1A). As AR42 inhibited topoIIα expression at concentrations well below its median inhibitory concentration (IC50) of 0.72 μM in inhibiting cell viability,6 this down-regulation was not consequent to drug-induced cell death. This topoIIα repression was also noted with MS-275 and, to a lesser extent, vorinostat; however, at an order-of-magnitude higher concentration. This drug-induced suppression was topoIIα-selective because these HDAC inhibitors did not cause dramatic changes in topoIIβ expression. The suppressive effect of these HDAC inhibitors on topoIIα expression was also demonstrated in Huh7 and HepG2 cells (Fig. 1B).
Published reports of the effects of other HDAC inhibitors on topoIIα expression indicate a cell type- and/or context-specificity. For example, treatment of D54 glioblastoma cells with trichostatin A or vorinostat had no effect on topoIIα expression.15 Although sodium butyrate was reported to sensitize leukemia cells to etoposide by increasing topoIIα gene expression,16 treatment of MCF-7 cells with valproic acid led to transcriptional repression of topoIIα.17 To clarify this issue, we assessed the concentration-dependent effect of sodium butyrate on topoIIα expression in PLC5 cells. Our data show that treatment with a range of concentrations of sodium butyrate revealed a biphasic effect on topoIIα expression levels, i.e., up-regulation at low concentrations (≤0.25 mM) and down-regulation at higher concentrations (≥0.5 mM), without disturbing topoIIβ expression (Fig. 1C). These concentrations are consistent with those of sodium butyrate (0.4 mM) and valproic acid (2 mM) that up-regulated and down-regulated topoIIα expression, respectively, in the aforementioned studies. This dichotomous effect may typify the complex mode of action of short-chain fatty acids in regulating topoIIα expression relative to other HDAC inhibitors examined.
HDAC Inhibitors Promote TopoIIα Degradation.
The finding that MS-275 was able to suppress topoIIα expression suggests the involvement of class I HDACs in the drug response. Thus, we assessed the effect of shRNA or siRNA-mediated knockdown of class I (HDAC1 and 2) vis-à-vis class II isozymes (HDAC4-6) on topoIIα messenger RNA (mRNA) and protein expression in PLC5 cells. Silencing of HDAC1 caused a sharp decrease in the topoIIα protein level, whereas the mRNA expression was not altered (Fig. 2A). However, the knockdown of other isozymes had no effect on the mRNA or protein expression of topoIIα. Evidence indicates that this topoIIα down-regulation was attributable to proteasomal degradation. First, exposure of PLC5 cells to AR42 or MS-275 did not cause appreciable changes in topoIIα mRNA levels as determined by RT-PCR (Fig. 2B). Second, the proteasome inhibitor MG132 protected cells against the suppressive effect of AR42, MS-275, and vorinostat on topoIIα expression (Fig. 2C). Third, in the presence of cycloheximide, AR42 promoted the elimination of topoIIα relative to the DMSO control (Fig. 2D). Together, these data suggest a pivotal role of HDAC1 in the regulation of topoIIα protein stability.
CK2 Is Involved in Ubiquitin-Dependent Degradation of TopoIIα.
It is well documented that ubiquitin-dependent protein degradation is preceded by phosphorylation.18 As shown in Fig. 3A, concentration-dependent topoIIα repression by AR42 was accompanied by parallel increases in p-Ser/Thr phosphorylation and ubiquitination. However, no appreciable acetylation of topoIIα was noted in response to AR42 treatment, suggesting that topoIIα stability is not influenced by HDAC-regulated acetylation. Thus, to shed light on the mechanism by which HDAC inhibitors facilitated topoIIα proteolysis, we first investigated the identity of the kinase involved in AR42-mediated topoIIα repression by examining the abilities of a panel of kinase inhibitors to block this cellular response. As CK2,19, 20 protein kinase Cδ,21 and extracellular signal-regulated protein kinase22 have been reported to target topoIIα, we assessed the effects of their respective inhibitors, DMAT, GF-109203X, and PD98059, on AR42-induced topoIIα repression. Also, inhibitors of phosphoinositide 3-kinase (wortmannin), IκB kinase (Bay11-7082), and p38 MAP kinase (SB202190) were used as controls. Among them, DMAT exhibited a unique ability to block AR42-facilitated topoIIα repression, whereas the other inhibitors showed no appreciable protective effect (Fig. 3B). This finding suggests a mechanistic link between CK2, a tetrameric kinase comprised of two catalytic subunits (α and α′) and two identical regulatory subunits (β),23 and HDAC inhibitor-mediated topoIIα proteolysis.
CK2 forms a stable, catalytically active complex with topoIIα20 and has been implicated in the modulation of topoIIα trafficking.24 Here we obtained three lines of evidence to corroborate the role CK2 in promoting HDAC inhibitor-induced topoIIα degradation.
First, AR42 and MS-275 treatment led to concentration-dependent increases in CK2α protein and mRNA expression in PLC5 cells (Fig. 4A), suggesting the transcriptional activation of CK2α expression by HDAC inhibitors. ChIP analysis revealed that AR42 treatment caused a concentration-dependent increase in the association of CK2α promoter DNA with acetylated histone H3 (Fig. 4B), which in turn was associated with the enhanced recruitment of the transcription factor Ets-1, a key regulatory element of the CK2α gene,25 to the promoter, without altering the expression level of Ets-1 (Fig. 4C). Moreover, shRNA-mediated HDAC1 knockdown led to increased CK2α expression like that observed with topoIIα repression (Fig. 4D). Together, these findings provide direct evidence of the involvement of HDAC inhibition in the observed increase in CK2α expression. Second, overexpression of CK2α mimicked the suppressive effect of HDAC inhibitors on topoIIα expression without disturbing topoIIβ (Fig. 4E). Third, shRNA-mediated CK2α knockdown protected PLC5 cells from AR42- and MS-275-mediated inhibition of topoIIα expression (Fig. 4F).
Role of Csn5 in HDAC Inhibitor-Mediated TopoIIα Degradation.
Csn5 (aka, Jun-activation domain-binding protein-1 [Jab1]), a component of the COP9 signalsome complex, plays a critical role in the degradation of a number of signaling proteins.26 We hypothesized that Csn5 plays an intermediary role between increased CK2α expression and topoIIα degradation based on the following published data: (1) Csn5 facilitates topoIIα degradation in response to glucose starvation by interacting with topoIIα's glucose-regulated destruction domain.27 (2) Csn5-mediated degradation of its target proteins can be prevented by the pharmacological inhibition of CK2, a Csn complex-associated kinase.28 These data, together with our findings, prompted us to investigate the involvement of Csn5 in the HDAC inhibitor-induced topoIIα degradation.
As shown in Fig. 5A, treatment of PLC5 cells with AR42 had no effect on Csn5 expression (input), but led to a concentration-dependent increase in the association of topoIIα with CK2α and Csn5 (right panel), which is noteworthy in that physical interaction with Csn5 is reported to be a prerequisite for the degradation of its target proteins.27 This increase in the amount of CK2α associated with the Csn5-topoIIα complex paralleled the increase in total cellular CK2α levels in AR42-treated cells. Moreover, the ectopic expression of Csn5 dose-dependently mimicked the suppressive effect of HDAC inhibitors on topoIIα expression (Fig. 5B), whereas siRNA-mediated knockdown of Csn5 protected against the drug-induced down-regulation of topoIIα in AR42- and MS-275-treated PLC5 cells (Fig. 5C). These results are consistent with the putative role of Csn5 in HDAC inhibitor-mediated topoIIα degradation.
Fbw7 Acts as an E3 Ligase That Targets TopoIIα for Csn5-Induced Degradation.
The Csn complex facilitates the proteasomal degradation of target proteins by functioning as a docking platform for recruitment of the target's specific kinase and E3 ligase.29 Consequently, we sought to identify the E3 ligase that targets topoIIα in the Csn5 complex. Csn5 is known to maintain the stability of a number of the F-box proteins of the Skp1-Cul1-F-box-protein family, including Skp2, Fbw7, Fbx4, and Fbx7, as the silencing of Csn5 led to the down-regulation of these F-box proteins.30 Thus, using these Csn5-interacting F-box proteins as candidates for the topoIIα-targeted E3 ligase, we assessed the concentration-dependent effects of AR42 on the binding of these F-box proteins to topoIIα. The E3 ligase Bmi1 was also assessed in light of a recent report that Bmi1 controlled topoIIα degradation in response to glucose starvation.31
PLC5 cells exhibited robust expression of Skp2, Fbw7, and Bmi1, but had low abundance of Fbx4 and Fbx7 (Fig. 6A, input). Coimmunoprecipitation revealed a concentration-dependent increase in the binding of Fbw7 to topoIIα by AR42 (right panel). This AR42-induced association was highly selective because the other F-box proteins were undetectable or present in extremely low levels, relative to Fbw7, in the complex formation with topoIIα. The functional role of Fbw7 as the topoIIα-targeted E3 ligase was further supported by the protective effect of shRNA-mediated knockdown of Fbw7 on AR42- and MS-275-mediated topoIIα ablation (Fig. 6B).
Above, we showed that, in addition to Csn5, CK2α also associated with topoIIα in response to AR42 (Fig. 5A). Thus, we hypothesized that phosphorylation of topoIIα by CK2 facilitated the association of topoIIα with the Csn5-Fbw7 complex in AR42-treated cells. Results in support of this hypothesis are shown in Fig. 6C, where the CK2 inhibitor DMAT abrogated the interaction of topoIIα with Csn5 and Fbw7. Exposure of PLC5 cells to AR42 induced a concentration-dependent increase in topoIIα phosphorylation, accompanied by parallel increases in its association with Csn5 and Fbw7, culminating in topoIIα proteolysis (left panel). However, pharmacological inhibition of CK2 by DMAT prevented increases above basal levels of AR42-induced topoIIα phosphorylation and its consequent association with Csn5 and Fbw7, thereby protecting topoIIα from drug-induced degradation (Fig. 6C, right panel).
Glycogen Synthase Kinase (GSK)3β-Dependent Binding of TopoIIα to Fbw7 Through a Recognition Motif at the C-Terminus.
Fbw7 recognizes the Cdc4 phosphodegron (CPD) motif of (S/T)PXX(S/T) (X, any amino acid) in many of its target proteins, including cyclin E, Myc, Jun, SV40 large T antigen, and the sterol regulatory element binding protein.32 Within this CPD motif, phosphorylation at the Thr residue by GSK3β in conjunction with that at the Ser residue by a priming kinase is required for binding. Analysis of the topoIIα sequence revealed two plausible Fbw7 recognition motifs, 1361SPKLS1365 and 1393SPPAT1397 in the C-terminal domain (Fig. 6D, boxed). It is especially noteworthy that the former motif encompasses a well-characterized GSK3β phosphorylation motif (SXXXS) and overlaps with a putative CK2 recognition site 1365SNKE1368 (consensus sequence, [S/T]XX[D/E]; all mapped CK2 sites are underlined),33 suggesting that CK2 might be the priming kinase for GSK3β-mediated phosphorylation of topoIIα.
The involvement of GSK3β in AR42-mediated topoIIα degradation was corroborated by several lines of evidence. First, pharmacological inhibition of GSK3β by SB-216763 protected cells against the suppressive effect of AR42 on topoIIα expression (Fig. 7A). Second, coimmunoprecipitation indicates that AR42 led to a concentration-dependent increase in the association of topoIIα with GSK3β (Fig. 7B). Third, ectopic GSK3β expression dose-dependently mimicked the effects of AR42 on the levels of topoIIα expression (Fig. 7C, left panel) and phosphorylation (right panel), and its association with Fbw7 (right panel).
The involvement of the 1361SPKLSNKE1368 motif in regulating topoIIα protein stability through interactions with Fbw7, GSK3β, and CK2 was supported by mutational analyses. Flag-tagged topoIIα mutants were created by replacing the Ser1361, Ser1365, Glu1368, Ser1393, or Thr1397 residue with Ala by way of site-directed mutagenesis, and then expressed in PLC5 cells in the presence or absence of ectopically expressed CK2α. Ectopic CK2α expression was used to mimic HDAC inhibitor-induced CK2α up-regulation and consequent topoIIα degradation because treatment with AR42 and other HDAC inhibitors induced the expression of the transfected Flag-topoIIα (data not shown), presumably through the epigenetic activation of transcription. Of these five mutants, only S1361A, S1365A, and E1368A abrogated the suppressive effect of CK2α overexpression on topoIIα expression (Fig. 7D, input). Coimmunoprecipitation analysis indicates that this reversal of drug action was attributable to the inability of the S1361A, S1365A, and E1368A mutants to bind Fbw7 (Fig. 7D, lower panel). In contrast, S1393A and T1397 did not confer protection against CK2α-induced degradation or binding to Fbw7, indicating that the 1393SPPAT1397 motif did not play a role in mediating topoIIα degradation in the presence of ectopically expressed CK2α.
The premise that CK2 might be the priming kinase for GSK3β-mediated phosphorylation of topoIIα was supported by coimmunoprecipitation analysis of the effect of CK2 and GSK3β inhibitors, DMAT and SB-216763, respectively, on AR42-induced association of topoIIα with CK2α and GSK3β. Cotreatment with DMAT abrogated the ability of AR42 to facilitate the complex formation (Fig. 7E). In contrast, although SB-216763 blocked the association of topoIIα with GSK3β, it exhibited only a modest suppressive effect on topoIIα-CK2α interactions.
In Vivo Mechanistic Validation.
To confirm our in vitro findings of a functional role for the CK2α-Csn5-Fbw7 signaling axis in mediating HDAC inhibitor-induced topoIIα degradation, we conducted an in vivo study in a xenograft model. PLC5 tumor-bearing mice were treated for 3 or 6 days with a tumor suppressive dose of AR42 (25 mg/kg daily).6 AR42 down-regulated topoIIα and increased CK2α expression levels in xenograft tumors, without changing those of Csn5 or Fbw7 (Fig. 8A, input). Moreover, coimmunoprecipitation analysis revealed that AR42 enhanced the intratumoral association of topoIIα with CK2α, Csn5, and Fbw7, reminiscent of that observed in vitro.
In the literature, a number of stress conditions have been reported to induce the proteasomal degradation of topoIIα, including G1 arrest,34 glucose starvation,35 hypoxia,35 and adenovirus E1A-induced apoptosis,36 although the underlying mechanism remains unclear. Here we report a novel mechanism by which HDAC inhibitors stimulate the selective degradation of topoIIα in HCC cells. As shRNA-mediated knockdown of HDAC1 (but not other HDAC isozymes examined) and could mimic the suppressive effect of AR42 and MS-275 on topoIIα expression, this drug-induced topoIIα degradation was, at least in part, attributable to the inhibition of HDAC1. Although HDAC1 has been reported to be associated with both the α and β isoforms of topoII,37 the significance of this binding in the effect of HDAC inhibitors on topoIIα degradation remains to be investigated.
We obtained evidence that transcriptional activation of CK2α expression represents a key driver for HDAC inhibitor-mediated topoIIα proteolysis. For example, ectopic expression of CK2α led to topoIIα repression, whereas pharmacological inhibition of CK2 kinase activity or shRNA-mediated silencing of CK2α expression protected cells from the suppressive effect of HDAC inhibitor on topoIIα expression. CK2 is known to bind and phosphorylate topoIIα on several serine and threonine residues near the nuclear export or localization signal.19, 20, 24 It was reported that CK2 could stabilize topoIIα against thermal inactivation in a phosphorylation-independent manner.38 Thus, this study provides a new insight into the role of CK2 in regulating the function/stability of topoIIα. Our data suggest that CK2-mediated phosphorylation of topoIIα, followed by GSK3β phosphorylation, facilitated its inclusion in the formation of a multiprotein complex with Csn5 and the Fbw7 E3 ligase, leading to its ubiquitin-dependent degradation (Fig. 8B). For instance, the silencing of either binding partner abolished the ability of HDAC inhibitors to deplete topoIIα, and pharmacological inhibition of CK2 kinase activity blocked both the formation of this complex and the drug-induced reductions of topoIIα levels. It is well documented that the Csn complex functions as a master docking platform to bring together a target substrate with its specific kinase and E3 ubiquitin ligase, which, in conjunction with the proteasome, facilitates the ubiquitin-dependent degradation.26, 29 The functional role of Csn5 in mediating CK2-facilitated topoIIα degradation is further corroborated by the reports that CK2 regulates the activity of Csn in mediating ubiquitin-dependent protein degradation,28 and that Csn5 is involved in topoIIα degradation in response to glucose starvation.27
Fbw7, the substrate recognition component of the SCF complex, is recognized as a tumor suppressor because of its ability to target a number of dominant oncogenes.32 In this study we used coimmunoprecipitation and shRNA-mediated knockdown of Fbw7 to demonstrate the functional role of Fbw7 as an E3 ligase targeting topoIIα. These findings reveal an additional layer of complexity in the regulation of topoIIα degradation and/or activity. Other E3 ligases have also been implicated in the degradation of topoIIα. It has been reported that Bmi1 is involved in topoIIα degradation in response to glucose starvation or the topoII trapping agent teniposide (VM-26).31 In the present report the role of Bmi1 in HDAC inhibitor-induced topoIIα degradation, however, was refuted by its decreased expression and lack of association with topoIIα in response to AR42 treatment (Fig. 6A). In other studies, Mdm239 and BRCA140 have been implicated in the ubiquitination of topoIIα, the former in the context of etoposide-mediated topoII degradation and the latter in the context of its participation in DNA decatenation. In addition, teniposide caused conjugation of small ubiquitin-related modifier-1 to topoIIα in HeLa cells, although its role in regulating topoIIα stability remains to be defined.41 The involvement of these pathways in HDAC inhibitor-induced topoIIα degradation remains to be investigated.
This study also reported the novel finding that topoIIα is a target of GSK3β phosphorylation, presumably at Ser1361, which might be primed by CK2-mediated phosphorylation at Ser1365, consistent with the reported mechanism underlying Fbw7-targeted protein degradation.32 Our data suggest that this double phosphorylation facilitated the recruitment of Fbw7 to the recognition motif 1361pSPKLpS1365 at the C-terminus of topoIIα, leading to its ubiquitin-dependent degradation.
In conclusion, our report shows a novel pathway by which HDAC inhibitors facilitate the selective degradation of topoIIα, which underlies the complexity of the functional role of HDAC in regulating tumorigenesis and aggressive phenotype in HCC cells. Previously, we demonstrated the efficacy of oral AR42 in the in vitro and in vivo models of HCC through the inhibition of HDAC and modulation of multiple aspects of cancer cell survival signaling,6 which, as we now have shown, includes topoIIα degradation. As AR42 has entered Phase I clinical trials, the present finding may be of translational value for the use of AR42 as a component of therapeutic strategies for advanced HCC, in which systemic therapies have largely been unsuccessful.
The authors thank Dr. Jack C. Yalowich (University of Pittsburgh) for critical reading of the article and valuable comments and suggestions.