Cell type-specific anti-cancer properties of valproic acid: independent effects on HDAC activity and Erk1/2 phosphorylation

Cells were grown in a humidified atmosphere at 37°C, 5% CO₂. The rat glioma cell lines BT4C and BT4Cn, the human glioma cell line U87MG (a gift from Dr. Nina Pedersen, Copenhagen University Hospital Rigshospitalet, Copenhagen, Denmark), the mouse neuroblastoma cell line N2a, the mouse adenocarcinoma cell lines CSML0 and CSML100, the human adenocarcinoma cell line HeLa, and the mouse fibroblastoid cell lines Swiss 3T3 and L929 were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) heat-inactivated fetal calf serum (FCS), 2 mM GlutaMAX, 100 U/ml penicillin, 100 μg/ml streptomycin and 2.5 μg/ml fungizone (all from Invitrogen, Taastrup, Denmark). Cells were dislodged with trypsin/EDTA in modified Puck's saline (Invitrogen).

The rat pheochromocytoma cell line PC12-E2 was grown in DMEM supplemented with 5% (v/v) FCS, 10% (v/v) horse serum, 2 mM GlutaMAX, 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen). Cells were dislodged by tapping the culture flask.

Stable transfection of L929 with the pGV16 vector encoding constitutively active rat H-Ras (G12V; a gift from Prof. Berthe Willumsen, Department of Biology, Faculty of Science, University of Copenhagen, Copenhagen, Denmark) was obtained using Lipofectin (Invitrogen). After transfection, cells were grown in medium containing 0.75 mg/ml geneticin (Invitrogen) for 3 weeks. Six geneticin-resistant clones with high H-RasG12V expression were selected, propagated, pooled and used as stock for subsequent experiments. Ten randomly selected clones stably transfected with the empty pGV16 vector were selected, propagated, pooled and used as control cells.

Transient transfections of cells with the pGV16 vector encoding dominant-negative rat H-Ras (S17N; a gift from Dr. Berthe Willumsen), the pRK5 vector encoding constitutively active rat MEK2 (S222E/S226E; a gift from Dr. Klaus Seedorf, Hagedorn Research Institute, Gentofte, Denmark), or the corresponding empty vectors were performed using FuGene 6 (Roche Diagnostics, Mannheim, Germany) or Lipofectin 2000 (Invitrogen). Transient transfections were performed as co-transfections with low amounts of the pEGFP-N1 vector (Clontech, Saint-Germain-en-Laye, France). Consequently, transfected cells were identified from their expression of enhanced green fluorescent protein. Cells were replated for subsequent experiments ~24 h after transfection.

A stock solution of 3 M VPA (Sigma-Aldrich, Copenhagen, Denmark) was prepared in dimethylsulfoxide (DMSO), and all assays were performed in the presence of 0.1% (v/v) DMSO in the presence or absence of VPA.

Cells were plated in 60 mm culture dishes (Nunc, Roskilde, Denmark) at a density of 0.25-0.75 × 10⁶ cells/dish and grown in medium containing 0-3 mM VPA for up to 48 h. In experiments in which cells were exposed to a MEK inhibitor (PD98059; Cell Signaling, Danvers, MA), the compound was added to the cultures 1 h before cell lysis. Following incubation, cells were rinsed in ice-cold PBS and collected in radioimmunoprecipitation assay lysis buffer supplemented with Complete, EDTA-free Protease Inhibitor Cocktail (Roche Diagnostics) and Phosphatase Inhibitor Cocktail Set III (Calbiochem). Proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to Immobilon-P membranes (Millipore, Billerica, MA). The membranes were blocked in PBS containing 0.1% (v/v) Tween-20 and 5% (w/v) bovine serum albumin.

Following incubation with primary antibodies, all membranes were incubated with horseradish peroxidase-conjugated goat anti-rabbit antibody (Cell Signaling), visualized by chemiluminescence (SuperSignal West Dura Extended Duration Substrate, Pierce Biotechnology, Rockford, IL) using a GeneGnome (Syngene, Cambridge, UK), and quantified using the accompanying GenTools software.

For estimation of Erk1/2 phosphorylation, membranes were probed with anti-phospho-Erk1/2 (Thr202/Tyr204) antibody (Cell Signaling). After visualization, the membranes were stripped (30 min, 50°C) in 62.5 mM Tris-HCl, pH 6.8, 2% (w/v) SDS and 50 mM 1,4-dithio-L-threitol, blocked, and reprobed with anti-pan-Erk1/2 antibody (Cell Signaling). Ras expression was determined using anti-H-, K- and N-Ras antibodies (Cell Signaling). Histone acetylation was determined using anti-acetyl histone H3 (Lys9/Lys14) antibody (Cell Signaling). Ras and histone H3 acetylation immunoblots were stripped and reprobed with anti-actin antibody (Sigma-Aldrich).

Cell growth was quantified from crystal violet stainings. 5 × 10³ cells/well were plated in 96-well microwell plates (Nunc) and grown in medium containing 0-3 mM VPA for 48 h, with each treatment replicated in 6 wells/plate. Cells were fixed in 3.7% (v/v) formalin and 1% (v/v) methanol in PBS, and stained with 0.04% (w/v) crystal violet in 4% (v/v) ethanol. Finally, the cell-bound crystal violet was solubilized using 1% (w/v) SDS, and optical density was measured at 600 nm.

6-15 × 10³ cells/well were plated in 6-well culture plates (Nunc) and grown in medium containing 0-3 mM VPA. Time-lapse video-recordings were performed using a computer-assisted microscope workstation as previously described [18]. Simultaneous, sequential recordings from 8-20 microscopic fields/well were performed at 1-10 min intervals for 20-80 min. Evaluation of individual cell motility was performed using the image processing software PRIMA (Protein Laboratory, Copenhagen University, Denmark). 50-200 cells from each well were tracked, and the data were used for the calculation of the mean-cell speed and the mean-squared cell displacement (<d ² >) as previously described [19].

Results are expressed as mean ± SEM, calculated on the basis of the number of experiments. Unless indicated otherwise, statistical analyses were performed on non-normalized data using one-way repeated-measures analysis of variance (ANOVA) followed by the Tukey-Kramer multiple comparison test. Immunoblots were evaluated using the Wilcoxon signed rank test. Estimates of IC ₂₅ and IC ₅₀ values were based on the interpolation of data from log dose-response curves. Correlations were performed as Pearson correlations. All p-values refer to two-tailed calculations. *, ** and *** indicate p < 0.05, 0.01 and 0.001, respectively.

Erk1/2 activity was investigated in 10 cell lines, including BT4C, BT4Cn, U87MG, N2a, PC12-E2, CSML0, CSML100, HeLa, Swiss 3T3 and L929, by estimating the degree of Erk1/2 phosphorylation by immunoblotting. In the absence of VPA, BT4C, BT4Cn, U87MG, and PC12-E2 exhibited high Erk1/2 phosphorylation levels. L929, Swiss 3T3 and CSML0 exhibited intermediate levels, and N2a, CSML100 and HeLa exhibited low levels (Figure 1a). Figure 1b and 1c show the relative changes in the degree of Erk1/2 phosphorylation in the respective cell lines in response to VPA (3 mM, 48 h). Erk1/2 phosphorylation was significantly decreased in CSML0 and L929, significantly increased in CSML100, BT4Cn and N2a, and not significantly affected in BT4C, U87MG, PC12-E2, HeLa and Swiss 3T3.

Bild vergrößern

VPA is a known HDAC inhibitor. Figure 1d shows the degree of acetylation of histone H3 in the 10 cell lines under control conditions determined by immunoblotting. All cell lines exhibited low acetylation levels under control conditions (Figure 1d).

VPA significantly increased the degree of acetylation in 8 of the 10 cell lines (Figure 1e and 1f). The changes in acetylation ranged from an approximately three-fold increase (U87MG) to an ~13-fold increase (BT4C; Figure 1e).

In conclusion, VPA has highly cell type-specific effects on the degree of HDAC inhibition and on changes in the degree of Erk1/2 phosphorylation. Furthermore, the effects of VPA on the degree of Erk1/2 phosphorylation and HDAC inhibition were not related to the degree of Erk1/2 phosphorylation and acetylation in the absence of the drug. Finally, no relationship was found between the effects of VPA on HDAC inhibition and the degree of Erk1/2 phosphorylation. Thus, VPA-induced changes in the degree of Erk1/2 phosphorylation cannot be explained by the effects of the drug on HDAC activity, although HDAC inhibition may contribute to the observed effects.

Figure 2 shows that VPA dose-dependently changed the growth of all investigated cell lines. Nine cell lines exhibited IC₂₅ values ranging from 0.43 to 1.83 mM (Table 1), whereas the growth of a single cell line, U87MG, was significantly elevated upon exposure to 0.75 mM VPA.

Bild vergrößern

A correlation analysis of the data presented in Figure 1 and 2 revealed no significant correlations between the effects of VPA on HDAC inhibition and cell growth or between the degree of Erk1/2 phosphorylation and cell growth (Figure S4). However, four of the five cell lines that did not demonstrate significant changes in the degree of Erk1/2 phosphorylation in response to VPA (Figure 1b) did not have IC₅₀ values for growth within the tested concentration range (below 3 mM; Table 1), whereas all five cell lines demonstrating significant changes in the degree of Erk1/2 phosphorylation in response to VPA had IC₅₀ values for growth below 3 mM. These results demonstrate a significant relationship between the effects of VPA o the degree of Erk1/2 phosphorylation and cell growth (p < 0.05, Fisher's exact test).

In conclusion, VPA induced highly cell type-specific effects on cell growth. Furthermore, no apparent relationship was observed between the effects of VPA on HDAC inhibition and cell growth, whereas VPA was likely to affect the growth of a given cell type if it affected the degree of Erk1/2 phosphorylation in that cell type at physiologically relevant concentrations.

Erk1/2 are known to regulate cell motility [20]. In L929 cells, VPA inhibits the degree of Erk1/2 phosphorylation (Figure 1) and cell speed [21]. Therefore, a potential relationship between the effects of VPA on the degree of Erk1/2 phosphorylation and cell motility was investigated further in L929 cells.

Figure 3a shows time-dependent effects of 3 mM VPA on the mean-cell speed of L929 cells. The speed was significantly reduced after 1 h exposure to the drug. After 24-48 h exposure, further reductions in cell speed were observed. Figure 3b shows data from the time-response experiments, in which VPA was directly added to the culture medium during the motility recordings. Using this approach, a reduction in cell displacement was observed for VPA-treated cells ~20 min after drug addition compared with control cultures. These observations demonstrate that VPA caused biphasic, time-related inhibition of the cell speed of L929 cells, characterized by an initial, rapid inhibition followed by a further inhibition at later time-points.

Bild vergrößern

Figure 3c and 3e show time-response effects of VPA on the degree of Erk1/2 phosphorylation. A 1 h exposure to 3 mM VPA significantly decreased the degree of Erk1/2 phosphorylation. After 6 and 24 h, the degree of Erk1/2 phosphorylation decreased further. Figure 3d and 3f show dose-response curves of the effects of VPA on Erk1/2 phosphorylation. VPA dose-dependently inhibited the degree of Erk1/2 phosphorylation, with estimated IC₂₅ and IC₅₀ values of ~0.24 and 0.58 mM, respectively. A comparison of Figure 3a and 3d revealed that the time-course of the effects of VPA on the degree of Erk1/2 phosphorylation resembled that observed for the inhibition of cell speed, suggesting that VPA partially modulated cell speed through modulation of the degree of Erk1/2 phosphorylation.

Following the studies of L929 cells, the motility of the remaining nine cell lines was investigated. Figure 4 shows that VPA significantly and dose-dependently inhibited the speed of N2a, U87MG and PC12-E2 cells, stimulated the speed of BT4Cn cells, and did not affect the speed of the remaining five cell lines (BT4C, CSML0, CSML100, HeLa and Swiss 3T3).

Bild vergrößern

The GTPase Ras is an upstream activator of the MAPK pathway [11]. To determine whether VPA modulated the MAPK pathway upstream or downstream of Ras, the effects of VPA were investigated in L929 cells expressing constitutively active Ras (caRas). Figure 5a shows micrographs of L929 cells not expressing or expressing caRas and grown in the presence or absence of VPA. Control-transfected cells adopted a stellate phenotype with an increased area when treated with VPA (Figure 5a, two leftmost images), whereas caRas-expressing cells were more round and loosely attached (Figure 5a, two topmost images), a typical phenotype of Ras-transformed cells [22]. However, exposure of caRas-expressing cells to VPA caused them to adopt a phenotype similar to control-transfected cells treated with VPA (Figure 5a, two rightmost images). The immunoblot presented in Figure 5b shows that L929 cells expressing caRas exhibited higher degrees of Erk1/2 phosphorylation than control-transfected cells when grown in the absence of VPA (Figure 5b, lanes 2 and 1). However, VPA caused equally strong inhibition of Erk1/2 phosphorylation in control- and caRas-transfected cells (Figure 5b, lanes 3 and 4). These results demonstrate that exposure to VPA reversed the effects of caRas expression on L929 cell morphology and the degree of Erk1/2 phosphorylation. Cells expressing caRas did not move significantly faster than control-transfected cells when grown in the absence of VPA (Figure 5c). The speed of control- and caRas-transfected cells was similarly inhibited in the presence of VPA (Figure 5c).

Bild vergrößern

Figure 5d shows the effects of VPA exposure on the speed of L929 cells expressing constitutively active MEK2 (caMEK2). Cells expressing caMEK2 did not move significantly faster than control-transfected cells grown in the absence of VPA (Figure 5d, p < 0.063). However, in the presence of VPA, caMEK2-expressing cells moved significantly faster than control-transfected cells exposed to VPA (Figure 5d, p < 0.029). Moreover, the speed of caMEK2-expressing cells was not significantly reduced after exposure to VPA (Figure 5d, column 4 vs. column 2).

To determine whether VPA affected the speed of BT4Cn cells through effects on the degree of Erk1/2 phosphorylation, the effects of VPA were investigated in BT4Cn cells expressing dominant-negative Ras (dnRas). The expression of dnRas was verified by immunoblotting using a polyclonal anti-Ras antibody. As shown in Figure 5e, Ras immunoreactivity was increased in cells transfected with the dnRas expression vector, indicating that the cells expressed dnRas. Figure 5f shows that exposure of BT4Cn cells to 3 mM VPA for 24 h, but not 4 h, significantly increased the speed of both control- and dnRas-transfected cells. Thus, the expression of dnRas was not able to prevent the VPA-induced increase in the speed of BT4Cn cells.

Figure 5g shows an immunoblot of BT4Cn cells treated or untreated with VPA and/or the MEK inhibitor PD98059. As expected, VPA increased the degree of Erk1/2 phosphorylation compared with untreated cells (Figure 5g, lanes 1 and 3). However, this stimulation was prevented by treating the cells with PD98059 (Figure 5g, lane 4). Similarly, treatment of BT4Cn cells with VPA increased the cell speed compared with untreated cells (Figure 5h, columns 1 and 3). However, the VPA-induced increase of the cell speed was completely prevented by PD98059 (Figure 5h, column 3 vs. column 4), which reduced the speed of VPA-treated cells to the level of control cells untreated with VPA and PD98059 (Figure 5h, columns 1 and 4). The same concentration of PD98059 did not significantly reduce the speed of cells untreated with VPA (Figure 5h, columns 1 and 2).

In conclusion, the data presented in Figure 5a-h demonstrate that VPA in L929 and BT4Cn cells modulates cell motility, cell morphology and the degree of Erk1/2 phosphorylation by altering signaling through the MAPK pathway downstream of Ras but upstream of MEK (i.e., at the level of Raf).