Histone deacetylase inhibitor, Trichostatin A induces ubiquitin-dependent cyclin D1 degradation in MCF-7 breast cancer cells

In previous studies we demonstrated that TSA induces the rapid degradation of cyclin D1in MCF-7 cells. Co-culture of TSA-treated cells with the proteasomal inhibitor MG132 inhibited cyclin D1 degradation, demonstrating a role for the ubiquitin-dependent degradation pathway [15]. In MCF-7 cells transiently expressing GFP-cyclin D1, the recombinant protein localizes to both the cytoplasm and nucleus of most cells (Figure 1, top panel). Co-culture with TSA and MG132 however, promoted the localization of GFP-cyclin D1 within the cytoplasm of most cells. (Figure 1, gray arrows). In contrast, co-culture of MCF-7 cells with TSA and leptomycin B (LMB), an inhibitor of CRM1-dependent nuclear export, promoted the nuclear accumulation of GFP-cyclin D1 (Figure 1, bottom panel). We wished to determine the role of GSK3β in mediating this effect of TSA on cyclin D1 localization. Site-directed mutagenesis was used to mutate wild type GFP-cyclin D1 residues Thr286 and/or Thr288 to alanine. MCF-7 cells transiently expressing wild type or mutant GFP-cyclin D1 were co-cultured with TSA and MG132 for 6 h and subsequently examined by quantitative fluorescence microscopy. Although the expression of the recombinant protein was very high in some cells, a distinct cytoplasmic, nuclear or nucleo-cytoplasmic localization could be observed in most transfected cells (Figure 1). GFP-cyclin D1 localization did not appear to be dependent on the level of recombinant protein expression (Figure 1). Cells with a >80 % cytoplasmic or nuclear localization of the recombinant protein were scored as cytoplasmic and nuclear respectively. All other cells were scored as nucleo-cytoplasmic. More than 200 cells from at least eight separate fields were scored from three separate experiments for quantitation. Wild type and mutant GFP-cyclin D1 localization was predominantly nucleo-cytoplasmic in the majority of cells examined (Figure 1 (gray arrows), Figure 2A). Treatment with either TSA or MG132 alone did not affect GFP-cyclin D1 localization. Co-culture with TSA and MG132 however, resulted in a significant increase in the number of cells in which GFP-cyclin D1 localization was predominantly cytoplasmic. Mutation of Thr286 alone or together with Thr288 abolished the effect of TSA on GFP-cyclin D1 localization. No effect on localization was observed when cells expressing GFP-cyclin D1_T288A were co-cultured with TSA and MG132. While these experiments did not allow us to distinguish clearly between increased nuclear degradation, nuclear export and cytoplasmic sequestration, they none the less suggested a possible role for GSK3β-dependent Thr286 phosphorylation in mediating the effect of TSA on GFP-cyclin D1 localization. We previously reported, that TSA enhances the accumulation of polyubiquitylated GFP-cyclin D1 when 26S proteasomal degradation is inhibited by MG132 [15]. It is possible that TSA treatment results in the cytoplasmic accumulation of polyubiquitylated cyclin D1 when 26S proteasomal degradation is inhibited.

Our previous observation that residue Thr286 was required for the apparent TSA-induced effect on GFP-cyclin D1 localization (Figure 1 and Figure 2A), suggested a role for GSK3β in facilitating TSA-induced cyclin D1 degradation in MCF-7 cells. We therefore investigated the effect of inhibiting GSK3β activity or CRM1-dependent nuclear export on the cyclin D1 degradation response to TSA. Endogenous levels of GSK3β were reduced by siRNA to >70% of the level in mock transfected MCF-7 cells (Figure 3A). GSK3α levels were unchanged confirming the specificity of the siRNA oligonucleotide pool. The cellular levels of cyclin D1 mRNA and protein in GSK3β knockdown cells were comparable to those in mock transfected cells. TSA-induced cyclin D1 degradation was partially inhibited in GSK3β knockdown cells. In contrast, MG132 completely abolished the cyclin D1 degradation response to TSA. Total GSK3α and GSK3β levels were unaffected by TSA, even after 24 h of treatment (Figure 3B). Specific inhibition of GSK3 activity with SB216763 also resulted in only a partial abrogation of TSA-induced cyclin D1 degradation. SB216763 activity was monitored by immunoblot analysis of the cellular levels of β-catenin, an endogenous GSK3β substrate protein. Although pretreatment of MCF-7 cells with up to 20 μM SB216763 for 24 h failed to completely abolish TSA-induced cyclin D1 degradation, β-catenin stabilization was observed at concentrations as low as 5 μM (Figure 3C). The inability of SB216763 to abrogate TSA-induced cyclin D1 depletion did not therefore result from a loss of activity or inappropriately low concentration of inhibitor. These observations suggested that while GSK3β activity could enhance TSA-induced cyclin D1 degradation in MCF-7 cells, this activity was not indispensable for cyclin D1 ubiquitin-dependent degradation. TSA induces cyclin D1 degradation in MCF-7 cells prior to repressing its transcription and without inhibiting protein synthesis [15] The stabilization of β-catenin and partial inhibition of TSA-induced cyclin D1 degradation by SB216763 may be due to the existence of basal GSK3β activity in the asynchronously growing MCF-7 cells used in this study.

We also investigated further, the effect of inhibiting CRM1-dependent nuclear export on TSA-induced cyclin D1 degradation. Treatment of MCF-7 cells with 5–10 ng / ml of leptomycin B alone for 6 h, did not lead to a significant increase in total cyclin D1 levels above those of vehicle control treated cells (Figure 3E). Nonetheless, pretreatment of MCF-7 cells with leptomycin B partially inhibited TSA-induced cyclin D1 degradation and the effect was maximal at 5 ng/ ml (Figure 3E). Inhibition of GSK3β did not result in changes in the level of cyclin D1 mRNA (Figure 3D). We noted that mutation of Thr286 to alanine (Figure 2B), "knockdown" of GSK3 by siRNA or nuclear export with leptomycin B (Figures 3A and 3E), all failed to induce significant cyclin D1 accumulation. Nonetheless, the ability of SB216763 and leptomycin B to partially inhibit TSA-induced cyclin D1 degradation demonstrates the active involvement of GSK3β-mediated nuclear export. These observations suggest that GSK3β-mediated nuclear export enhances the rate of TSA-induced cyclin D1 degradation in MCF-7 cells. The GSK3β-independent degradation of cyclin D1 has been described previously [8]. In addition, we have demonstrated that the inhibition of 26S proteasomal degradation by MG132 results in the accumulation of both wild type and Thr286 mutant GFP-Cyclin D1 in MCF-7 cells (Figure 2B). It is thus possible that TSA can induce cyclin D1 degradation independently of GSK3β in this cell line.

The lack of specificity inherent to many inhibitors has previously led to proteins being linked mistakenly to one degradation pathway or another. Indeed, the peptide aldehydes ALLN and MG132 have been shown to inhibit cysteine and serine proteases such as calpains and cathepsins [24]. We thus investigated the possibility that the stabilization of cyclin D1 in MCF-7 cells by MG132 did not result from the inhibition of 26S proteasome activity alone. Lactacystin, a more specific inhibitor of the 26S proteasome than MG132 [25, 26], was used to investigate the possibility that cyclin D1 degradation occurs in the absence of 26S proteasome activity. Lactacystin partially abolished TSA-induced cyclin D1 degradation (Figure 4A). The partial inhibition of TSA-induced cyclin D1 degradation by lactacystin was observable at concentrations as low as 0.5 μM and did not increase even at concentrations as high as 10 μM (Figures 4A,B and 4C). In contrast, inhibition of calpain 1 and calpain 2 by ALLM (Figure 3C) or the inhibition of lysosomal degradation with NH₄Cl (Figure 4B) or cathepsin inhibitor-1 (CATI-1) (not shown) had no effect on TSA induced loss of cyclin D1 protein expression in MCF-7 cells. It remains unclear why lactacystin is less effective than MG132 at inhibiting TSA-induced cyclin D1 degradation. This differential effect on stability was also observed when cyclin D1 was ablated by treatment with cycloheximide. Similarly, the peptide aldehyde inhibitors ALLN and MG132 but not lactacystin or NLVS (a peptide vinyl sulfone proteasome inhibitor) inhibited cycloheximide-induced cyclin D1 ablation in MCF-7 cells (see Additional file 1). Interestingly, treatment of asynchronous populations of MCF-7 cells with 10 μM lactacystin or 100 μM ALLM for 6 h resulted in decreased cyclin D1 levels (Fig. 4D). Co-treatment with MG132 inhibited lactacystin- or ALLM-induced cyclin D1 loss (Fig. 4D). The inhibition of TSA-induced cyclin D1 degradation by both MG132 and lactacystin confirms our previous observation, that the 26S proteasome plays an important role in TSA-induced cyclin D1 degradation.

Carbobenzoxy-leucyl-leucyl-leucinal (Z-LLL-CHO, MG132), N-acetyl-leucyl-leucyl-norleucinal (LLnL, ALLN, MG101), lactacystin, Z-Phe-Gly-NHO-Bz (Z-FG-NHO-Bz, Cathepsin inhibitor I, CATI-1), NIP-Leu₃-vinyl sulphone (NLVS) and N- acetyl-leucyl-leucyl-methional (ALLM, Calpain Inhibitor II) (Calbiochem, VWR International Ltd., Lutterworth, United Kingdom) were dissolved in DMSO at appropriate concentrations and aliquots stored at -80°C. Stock solutions of TSA in ethanol, aqueous ammonium chloride (NH₄Cl) and leptomycin B in 70% (v/v) methanol (Sigma-Aldrich; Dorset, United Kingdom) were stored at -20°C. The GSK3-specific inhibitor SB216763 (Tocris Bioscience, Avonmouth, United Kingdom) was dissolved in DMSO and stored at -20°C. Antibodies to cyclin D1, actin, Sp1 and GFP (Santa Cruz Biotechnology, Santa Cruz, CA), and β-Catenin (Transduction Laboratories, BD Biosciences Ltd., Oxford, United Kingdom) were used.

MCF-7 cells (American Type Culture Collection, Rockville, MD) were cultured in DMEM supplemented with 10% (v/v) fetal calf serum, 2 mM L-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin at 37°C in humidified 5% CO₂.

The pEGFP-cyclin D1 plasmid vector was mutated to create pEGFP-cyclin D1_T286A, pEGFP-cyclin D1_T288A and pEGFP-cyclin D1_{T286A, T288A}. Site-directed mutagenesis was performed using a QuikChange^® II kit (Stratagene) according to the manufacturer's instructions. Mutation of T286A was achieved using the primer pair forward, 5'-TGGACCTGGCTTGCGCACCCACCGA-3', and reverse, 5'-TCGGTGGGTGCGCAAGCCAGGTCCA-3', and T288A using forward, 5'-TTGCACACCCGCCGACGTGCGGGAC-3', and reverse, 5'-AACGTGTGGGCGGCTGCACGCCCTG-3'. The T286A, T288A double mutant was generated by mutating the GFP-Cyclin D1_T286A plasmid construct with the T288A primer pair.

Cells treated as indicated were harvested in 5 ml of medium, pelleted by centrifugation (1,000 × g for 5 min at 4°C), washed twice with ice-cold PBS and lysed in ice-cold HEPES buffer [50 mM HEPES (pH 7.5), 10 mM NaCl, 5 mM MgCl₂, 1 mM EDTA, 10% (v/v) glycerol, 1% (v/v) Triton X-100 and a cocktail of protease inhibitors] on ice for 30 min. Lysates were clarified by centrifugation (15,000 × g for 10 min at 4°C) and the supernatants then either analyzed immediately or stored at -80°C. Equivalent amounts of protein (20–50 μg) from total cell lysates were resolved by SDS-PAGE using precast 4–12% Bis-Tris gradient gels (Invitrogen Ltd., Paisley, United Kingdom) and transferred onto polyvinylidene difluoride (PVDF) membranes (Hybond P; Amersham Biosciences United Kingdom Limited, Little Chalfont, United Kingdom) with a Novex XCell system (Invitrogen). Membranes were blocked overnight at 4°C in blocking buffer [5% (w/v) nonfat dried milk, 150 mM NaCl, 10 mM Tris (pH 8.0) and 0.05% (v/v) Tween 20]. Proteins were detected by incubation with primary antibodies at appropriate dilutions in blocking buffer overnight at 4°C. Blots were then incubated at room temperature with horseradish peroxidase-conjugated secondary antibody. Bands were visualized by enhanced chemiluminescence (Supersignal West Pico; Perbio Science UK Ltd., Cheshire, United Kingdom) followed by exposure to autoradiography film (Kodak BioMax ML-light or MR-1).

For RT-PCR analyses, cells were washed with PBS buffer and lysed directly in the culture plate using RTL buffer (Oiagen). RNA was extracted using an RNeasy Kit (Qiagen) followed by cDNA synthesis and a 25 cycle PCR reaction using a One-Step RT-PCR kit (Qiagen). The annealing temperatures of the primers for Cyclin D1 amplification (Forward 5'-aacagaagtgcgaggaggag-3'; Reverse 5'-ctggcattttggagaggaag-3') was 55°C. GSK3β (Forward 5'-cagcaaggtgacaacagtgg-3'; Reverse 5'-ggaacatagtccagcaccaga-3') and CGI-128 (Forward 5'-ggaattccagcggtggtggttccgcg-3'; Reverse 5'-ccccccgggagctgggaccaggat-3') amplification the annealing temperatures were 58°C and 68°C respectively. Optimization reactions were carried out to ensure the reactions were within the exponential amplification window. Reaction products were resolved on 2 % agarose gels containing 10 μg/ml ethidium bromide and visualized with a UV transilluminatior.

MCF-7 cells were transfected with commercially available siRNA oligonucleotide pools (Dharmacon, Perbio) using Oligofectamine transfection reagent (Invitrogen) as previously described [15]. Fugene 6 transfection reagent (Roche Diagnostics Ltd, East Sussex, United Kingdom) was used for DNA plasmid transfection. Asynchronous cell populations at a density of 50–60% in 6-well plates or on coverslips were transfected with 1–2 μg of plasmid DNA, following the formation of lipid-DNA complexes for 20 min at room temperature in Optimem I medium (Invitrogen). Complexes were added directly to cells growing in 2 ml DMEM and incubated for 5 h followed by washing with PBS buffer and addition of fresh DMEM. Cells were normally used in experiments 24 h following transfection and the recombinant proteins detected by immunoblotting, or by fluorescent microscopy as previously described