Skip to main content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Mol Oncol. 2014 Dec; 8(8): 1379–1392.
Published online 2014 May 28. doi: 10.1016/j.molonc.2014.05.001
PMCID: PMC4646083
NIHMSID: NIHMS602555
PMID: 24954856

Histone deacetylase inhibitor‐mediated cell death is distinct from its global effect on chromatin

Victoria L. Luchenko, 1 Thomas Litman, 2 Arup R. Chakraborty, 1 Aaron Heffner, 3 Christopher Devor, 3 Julia Wilkerson, 1 Wilfred Stein, 1 , 4 Robert W. Robey, 1 Lois Bangiolo, 1 David Levens, 3 and Susan E. Batescorresponding author 1
Author information Article notes Copyright and License information PMC Disclaimer

Associated Data

Supplementary Materials

Abstract

Romidepsin and vorinostat are histone deacetylase inhibitors (HDACis) that have activity in T‐cell lymphomas, but have not gained traction in solid tumors. To gain deeper insight into mechanisms of HDACi efficacy, we systematically surveyed 19 cell lines with different molecular phenotypes, comparing romidepsin and vorinostat at equipotent doses. Acetylation at H3K9 and H4K8 along with 22 other histone lysine acetylation and methylation modifications were measured by reverse phase proteomics array (RPPA), and compared with growth inhibition (IC50), and cell cycle arrest. These assays typically used to assess HDACi effect showed that acetylation and methylation of specific lysine residues in response to HDACis were consistent across cell lines, and not related to drug sensitivity. Using a treatment duration more reflective of the clinical exposure, cell death detected by annexin staining following a 6 h drug exposure identified a subset of cell lines, including the T‐cell lymphoma line, that was markedly more sensitive to HDAC inhibition. Kinetic parameters (Km values) were determined for lysine acetylation and for cell cycle data and were themselves correlated following HDACi exposure, but neither parameter correlated with cell death. The impact on cell survival signaling varied with the molecular phenotype. This study suggests that cellular response to HDACis can be viewed as two distinct effects: a chromatin effect and a cell death effect. All cells undergo acetylation, which is necessary but not sufficient for cell death. Cells not primed for apoptosis will not respond with cell death to the impact of altered histone acetylation. The divergent apoptotic responses observed reflect the variable clinical outcome of HDACi treatment. These observations should change our approach to the development of therapeutic strategies that exploit the dual activities of HDACis.

Keywords: HDAC inhibitors, Cell context, Cell cycle arrest, Histone modification, Apoptosis

Highlights

  • HDIs cause two distinct effects: a chromatin effect and a cell death effect.
  • Acetylation after HDI is monotonous, found at multiple lysine residues.
  • The impact on chromatin triggers apoptosis in only selected cell types.
  • Short drug exposure models the divergent clinical responses.
  • Further clinical development should take into account this differential effect.

1. Introduction

Histone deacetylase inhibitors (HDACis) constitute a class of targeted anticancer agents that inhibit histone deacetylases (HDACs) and effect chromatin modulation through altered binding of transcriptional co‐activator complexes. Studies have shown that HDACis alter approximately 7–10% of genes including those that control growth, differentiation, and apoptosis; genes are both up‐ and down‐regulated (Peart et al., 2005; Xu et al., 2007). In addition, HDACis have been shown to impair angiogenesis and metastasis; inhibit DNA repair while increasing the generation of reactive oxygen species (ROS); and modulate the mitogen‐activated protein kinase (MAPK) pathway (Khan and La Thangue, 2012; Rosato and Grant, 2003). While unrestrained histone acetylation results from HDAC inhibition, a myriad of cellular effects have been described. HDACis also promote acetylation of cytoplasmic proteins, altering function in some and in others, including several oncogenic proteins, inciting their degradation (Bali et al., 2005; Chen et al., 2009; Gu and Roeder, 1997; Gupta et al., 2012; Lane and Chabner, 2009; Zhang et al., 2003). Sorting out which of these effects is most important to accomplish cell death has been a perplexing problem for those working in the field.

Gene induction by HDACis has been tied to both cell cycle arrest and apoptosis. G1 arrest occurs via activation of the cyclin‐dependent kinase inhibitor p21 in a p53‐independent manner (Blagosklonny et al., 2002; Sandor et al., 2000) and downregulation of cyclins D and A. HDACis repress two genes involved in DNA synthesis, CTP synthase and thymidylate synthetase (Glaser et al., 2003), and as a result may block S‐phase progression. Finally, HDACis can mediate G2/M phase arrest by activating a G2‐phase checkpoint when the centromere is aberrantly acetylated (Robbins et al., 2005). Similarly, HDACis have been shown to induce apoptosis via the extrinsic and the intrinsic pathways. Some studies indicate a death receptor pathway, such as the induction of Fas or Fas ligand, TRAIL, DR‐4, DR‐5 (Kaminskyy et al., 2011; Nebbioso et al., 2005; Yeh et al., 2009); while other studies suggest that the mitochondrial apoptotic pathway is involved through decreasing expression of antiapoptotic proteins BCL‐2, BCL‐XL, MCL‐1 (Jiang et al., 2007; Newbold et al., 2008), and upregulating expression of the BH3‐domain proapoptotic proteins Bim, Bad, Bid, Bik, Puma, Noxa (Baumann et al., 2012; Xargay‐Torrent et al., 2011).

It has been difficult to reconcile all the promising in vitro data with the clinical results. To date, the clinical data for HDACis have proven efficacy in T‐cell lymphomas, but not in various solid tumors in which they have been tested (Venugopal and Evans, 2011). The U.S. Food and Drug Administration has approved two HDACis, romidepsin and vorinostat, for the treatment of cutaneous T‐cell lymphoma and romidepsin for the treatment of peripheral T‐cell lymphoma (Bates et al., 2010; Olsen et al., 2007; Piekarz and Bates, 2009; Piekarz et al., 2011). In this study, we sought to identify differential activity between cell lines that might yield insights into the clinical observations, and also to determine potential differences between the HDACis romidepsin and vorinostat after correction for the well‐known difference in potency. We examined global histone changes, cell cycle arrest, and apoptosis after HDACi treatment in a series of nineteen cell lines with differing genetic lesions. We conclude that epigenetic effects on histone acetylation are homogenous between drugs and across cell lines, but that the ability to undergo rapid cell death in response to acetylation is cell context specific.

2. Materials and methods

2.1. Cell lines and drugs

Cell lines were obtained from American Type Culture Collection and the NCI Anticancer Drug Screen; the p21‐deficient HCT116 subline (HCT116 p21−/−) was a gift from Dr. Bert Vogelstein (Johns Hopkins University). Cell line validation was performed and DNA fingerprinting confirmed their identities. Cultures were replaced in less than 3 months. Cells were cultured in RPMI 1640 or IMEM (GIBCO, Grand Island, NY) supplemented with 10% fetal bovine serum (GIBCO, Grand Island, NY), 2 mM glutamine (BioFluids, Rockville, MD), and 100‐units/L penicillin‐streptomycin (BioFluids). MCF‐10A was grown in DMEM‐F12 medium (Mediatech, Inc., Herndon, VA) supplemented with 5% horse serum, 10 μg/ml insulin, 20 ng/ml epidermal growth factor, 0.5 μM/ml hydrocortisone (Sigma, St. Lois, MO), and 100‐units/L penicillin‐streptomycin (BioFluids). HDACis romidepsin and vorinostat were obtained from the Anticancer Drug Screen (Cancer Therapy Evaluation Program, NCI, NIH, Bethesda, MD) and Cayman Chemical (Ann Arbor, MI), dissolved in DMSO at 100 μg/ml and 100 mM, respectively, and stored in aliquots at −20 °C. Caspase Inhibitor Q‐VD‐OPh was obtained from R&D Systems (R&D Systems, Minneapolis, MN) and dissolved in DMSO at 10 mM.

2.2. Cell sensitivity assay

Cells were treated with the indicated concentrations of HDACis for 96 h. For suspension cells analysis of growth inhibition was performed by MTS assay using the Cell Titer 96 Aqueous One Solution (Promega, Madison, WI, USA), and for adherent cells growth inhibition was performed by MTT assay or using sulforhodamine B stain (Sigma, St. Lois, MO) according to the manufacturer's protocols.

2.3. Cell cycle analysis by flow cytometry

Cell cycle distribution was determined by analyzing DNA content after propidium iodide staining. Cells were harvested, fixed in 70% ice‐cold ethanol, washed with PBS, and stained with 50 μg/ml propidium iodide containing 200U/ml RNase A. DNA content was analyzed using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Data were collected with Cell Quest Pro software from no fewer than 10,000 cells and analyzed using FlowJo software (Tree Star, Inc, Ashland, OR). Once cell cycle parameters were determined, the delta or difference between treated and untreated control was calculated at each dose and for each parameter, and used in further analyses.

2.4. Annexin V assay

Apoptosis was measured using the annexin V‐fluorescein isothiocyanate (annexin V‐FITC) Apoptosis Detection Kit (BD Biosciences, San Diego, CA) according to the manufacturer's instructions. Cells were treated with 25 ng/ml (45 nM) romidepsin or 25 μM vorinostat for 6 h, washed, and incubated an additional 42 h in drug‐free medium. Following treatment, annexin positive cells were quantitated using FlowJo Software. The percentage of annexin positive cells was derived, and the delta between treated and untreated control determined.

2.5. Reverse phase protein array (RPPA)

Cells were lysed in 10 mM HEPES, pH 7.4, 1% LDS, containing protease (Sigma, St. Louis, MO) and phosphatase inhibitors (Roche Diagnostic, Indianapolis, IN), and then sonicated. Protein concentrations were determined using BCA protein assay (Pierce, Rockford, IL); the samples were adjusted to equal protein concentrations, and then printed onto nitrocellulose‐coated glass slides (Grace Bio‐Labs, Bed, USA) using a solid‐pin arrayer (Aushon Biosystems, Billirica, MA). Triplicate four‐point serial deposition curves were printed for each lysate. Primary antibody staining was done overnight at 4 °C, and probed using a secondary‐antibody amplification‐and‐detection system FITC tyramide (Perkin–Elmer, USA) at room temperature. For the origin and description of all antibodies used in this study see Supplementary Table S1.

Slides were imaged with the FLA‐8000 scanner (Fujifilm), and relative signal intensities were calculated using MicroVigene software (VigeneTech, Carlisle, USA). The average staining of the triplicate prints was used for statistical analysis. After adjusting for the negative control signal, each histone modification value was normalized to the total histone H3 level.

2.6. Immunoblot analysis

Cell pellets were suspended in RIPA buffer with protease inhibitor cocktail (Sigma, St. Louis, MO), phosphatase inhibitors (Roche Diagnostic, Indianapolis, IN), and trichostatin A inhibitor, and then sonicated for 40 s. Protein concentration was measured using the Bio‐Rad Protein Assay (Bio‐Rad Inc., USA). Denaturated protein (20 μg) was loaded onto precast 4–12% Bis‐Tris NuPAGE gels, subjected to electrophoresis, and transferred onto 0.2 μM pore size nitrocellulose membranes (Invitrogen, Carlsbad, CA). Membranes were stained with 0.1% Ponceau S (Sigma, St. Louis, MO) and checked for comparable loading. After blocking with Odyssey blocking buffer (LI‐COR Bioscience, Lincoln, NE) for 1h at room temperature, membranes were incubated with following primary antibodies: H3K9ac (Upstate Biotechnology, Lake Placid, NY), H3K18ac, H3K36me2, H3K79me2, H3pan, PARP, cleaved‐PARP, p‐Rb, p‐MEK 1/2, MEK 1/2, p‐ERK 1/2, ERK 1/2, p‐AKT, AKT, Bim, Cyclin D1 (Cell Signaling Technology, Danvers, MA), H3K4me3, H3K9me3 (Abcam, Cambridge, MA), H4K8ac, c‐Myc (Santa Cruz Biotechnology, CA), p21 (Calbiochem‐Millipore, Billerica, MA), Ac‐α‐tubulin, α‐tubulin (Sigma, St. Lois, MO), GAPDH (American Research Products, MA) overnight at 4 °C. Membranes were probed with the IRDye 800CW goat anti‐mouse or IRDye 680 goat anti‐rabbit secondary antibodies (LI‐COR, Lincoln, NE), visualized and quantified using Odyssey System (LI‐COR).

2.7. Data analysis

Data visualization, including dynamic principle component analysis (PCA), heat maps and unsupervised hierarchical clustering was performed in Qlucore Omics Explorer v.2.2 (Qlucore AB, Lund, Sweden). Correlations between protein marker expression and all other parameters (including cell cycle, growth, apoptosis, drug concentration, time, and phosphorylation) were analyzed using both the non‐parametric Spearman's rank correlation coefficient test for ordinal data, and linear regression for continuous data. (GraphPad Prism 5.0). P‐values less than 0.05 were considered significant. Graphical presentation and regression analyses were made using Sigmaplot 11 software (Systat Software, Inc. 1735 Technology Drive, Suite 430 San Jose, CA 95110 USA). For both cell cycle and proteomic histone modification data, dose response curves were used to determine the Km values. The data were fitted to a hyperbolic function, descending or ascending as appropriate. In some cases of the histone modification data, the fitting procedure failed to give a convergent solution since the data were too few or too scattered. The Km values from the convergent solutions for the individual histone modifications were then averaged to generate a final mean “Km marks” value that was used in correlation analyses. In a few cases, the cell cycle changes were insufficiently large to allow the extraction of the kinetic parameters.

3. Results

To study the effect of the HDACis romidepsin and vorinostat we selected immortalized breast cell line MCF‐10A, and 18 cancer cell lines derived from breast (MCF‐7, SK‐BR‐3, MDA‐MB‐231), prostate (PC3), lung (A549, H460, EKVX, H146, H526), and colon cancer (HCT116, HCT116 p21−/−, S1, SW620), melanoma (LOX‐IMVI, UACC‐62, MDA‐MB‐435), glioma (SF‐295), and T‐cell lymphoma (HUT‐78) with different molecular phenotypes. Known mutations are shown in Table 1.

Table 1

Cell line characteristics.

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g008.jpg

3.1. Analysis of growth inhibition by 96‐h assay shows a similar profile across cell lines

One of the most widely used tests for drug efficacy is the 96‐h growth inhibition/cytotoxicity assay. This assay is widely used to assess HDACi effects and we thus utilized it to examine differences between romidepsin and vorinostat and among the cell lines. The concentration at which growth is inhibited by 50% (IC50) was determined after 96 h incubation in equipotent dose ranges for both drugs. Table 1 shows IC50 values for the various cell lines, with the median IC50 for romidepsin (1.5 nM) and for vorinostat (2 μM) indicating an approximate 1000‐fold difference in potency. With the P‐glycoprotein‐over‐expressing S1 cell line excluded (Pgp is a known mechanism of resistance for romidepsin) (Robey et al., 2006; Xiao et al., 2005), the scatter plot in Figure 1 shows a similar profile for the cell lines (R = 0.75, P = 0.005), demonstrating indiscriminate anticancer activity for both drugs in the 96 h continuous exposure format, without differentiating between the two HDACis, after correction for the well‐known difference in potency between the two drugs.

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g001.jpg

Analysis of growth inhibition by 96‐h assay in 18 cell lines treated with HDACis. The 96‐h cytotoxicity assay was performed to detect sensitivity of the cells to romidepsin or vorinostat using sulforhodamine for adherent cell lines and MTS‐based assay for suspension cell lines. Cells were treated with various concentrations of HDACis, starting with high concentrations of 10 μM for romidepsin and 30 μM for vorinostat. The data represent the mean of 3–6 independent experiments in 18 cell lines, and regression of IC50 shows remarkable similarity between two drugs (R = 0.75, p < 0.001). S1 is not shown as its IC50 for romidepsin is out of scale (8 nM) due to Pgp expression. Color key: small cell lung cancer (SCLC), light blue; non‐small cell lung cancer (NSCLC), dark blue; melanoma, gray; CNS, purple; T‐cell lymphoma, red; prostate, green; colon, brown; and breast, pink. The IC50 data for all cell lines including S1 are presented in Table 1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.2. Reverse phase proteomics array shows similar patterns for histone marks across cell lines

We next looked for differential effects at the histone level examining histone marks using a custom‐designed, reverse phase protein array (RPPA), a high‐throughput dot‐blot protein microarray containing 4992 spots per slide. The lysines on histone tails undergo post‐translational modification including acetylation and methylation. Acetylation, resulting from the unrestrained histone acetyltransferase activity that follows inhibition of the histone deacetylases, is classically an activating modification. Methylation occurs with one, two or three groups added to the lysine, and although typically repressive, methylation at H3K4 is known to be activating. We screened the protein array for 24 histone acetylation and methylation modifications in separate experiments for romidepsin and vorinostat at six equipotent concentrations (3‐fold serial dilutions of 0.1–30 ng/ml (0.2–54 nM) romidepsin, 0.1–30 μM vorinostat) at four different time points (8, 24, 48, 72h). Figure 2A displays graphically the induction of H3K9 acetylation in every cell line, while Figure 2B presents a heat map showing the common global effect on acetylation following romidepsin, despite minor differences in acetylation intensity among the cell lines. Results here are not clustered but arranged by cell line and concentration.

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g002.jpg

RPPA of histone acetylation modifications. Heat maps of thirteen acetylated marks sorted according to cell lines and concentration of HDACis. (A) The H3K9 acetylation as a function of romidepsin concentration (X‐axis, 0–30 ng/ml, log‐scale) is shown for all cell lines following 24 h treatment. The Y‐axis shows the pan‐H3 normalized values for H3K9ac histone modification. (B) Heat map showing increase in histone acetylation following 24 h treatment with romidepsin (Romi) and vorinostat (Vor). Cell line identifiers in panel A also apply to the heat map. The drug concentration ranges are depicted on a gray scale from 0 (black) to high concentrations (white) of 30 ng/ml (54 nM) for romidepsin or 30 μM vorinostat. The color bars on the top indicate tissue of origin: SCLC, light blue; NSCLC, dark blue; melanoma, gray; CNS, purple; lymphoma, red; prostate, green; colon, brown; breast, pink. Data are shown for 17 cell lines (the immortalized MCF‐10A and the transfected HCT116‐p21−/− cell lines have been excluded). The color green indicates relatively low and the color red indicates relatively high protein expression. (C) Representative plots of histone modifications as a function of romidepsin concentration. The lines drawn are the best‐fit predictions to ascending or descending hyperbolas using Sigmaplot software. From these plots, Km values were determined, as described in Methods. Color and symbols key: H2AK5ac, black circle; H2BK20ac, red circle; H2BK46ac, light green triangle; H2BK120ac, yellow triangle; H3K9ac, blue square; H3K18ac, pink square; H3K23ac, turquoise diamond; H3K56ac, grey diamond; H3K79ac, brown triangle; H4K5ac, dark green triangle; H4K8ac, dark green hexagon; H4K16ac, dark blue hexagon. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

In Figure 2C, dose responses for several acetylated lysines are plotted, from which Km values (the concentrations required to achieve half‐maximal acetylation) were derived. Due to background staining in the RPPA, there was variability in the number of acetylation sites for which an accurate Km could be attained for each cell line. The Km values from the convergent solutions for the individual histone modifications were then averaged to generate a final value with its SD that was used in correlation analyses. The n and mean acetylation Km for each cell line is provided in Supplementary Table S2. Notably, Km values for the histone modifications were not at all correlated with the IC50 obtained from the growth inhibition assays (R = 0.0437, P = 0.863, n = 18 for romidepsin; and R = 0.095, P = 0.747, n = 14 for vorinostat).

Next, unsupervised clustering analyses were performed. Since the romidepsin and vorinostat RPPA experiments were performed separately, clustering analyses were performed separately. Unsupervised two‐way clustering of histone acetylation modifications revealed similar patterns across cell lines after 24 h of either romidepsin or vorinostat treatment (Supplementary Figure S2A and S2B). When methylation modifications were included and the data organized by concentration and subjected to one‐way clustering, the repressive modifications, as one might expect, generally separated from the activating modifications (Supplementary Figure S2C and S2D). Upon examining the time course of HDACi treatment on histone modifications (also referred to as marks) (Supplementary Figure S3), activating marks were found to have the highest expression after 8 and 24 h following HDACi treatment, decreasing with continued exposure to romidepsin. Minimal or no change was observed in repressive marks at 8 and 24 h. Unlike RNA induction, which increases with longer exposure to HDACis (Kitazono et al., 2001), the global acetylation increase appears to be self‐limited.

To validate the proteomic data and to look for further insight into the mechanism of HDACi effect, we performed immunoblot analysis of selected histone modifications in three cell lines with different molecular phenotypes treated for 24 h with romidepsin and vorinostat as shown in Figure 3A. As in the RPPA, the acetyl marks and the methyl activating mark H3K4me3 were induced with both HDACis, while repressive marks such as H3K9me3 did not change significantly (Figure 3B). The correlation between RPPA and Western blot analysis of histone proteins is consistent for both drugs with R = 0.78 for romidepsin and with R = 0.74 for vorinostat, P < 0.001, as shown in Figure 3C.

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g003.jpg

Analysis of histone modification proteins in three cell lines shows the induction of activating marks and no or little change in repressive marks. (A) Exponentially growing cells were treated either with romidepsin (Romi) or vorinostat (Vor) for 24 h as indicated. Data are representative example of three independent experiments. (B) Bar graph shows quantitative analysis of histone marks from immunoblots as in (A). Color key: MDA‐MB‐231, pink; A549, blue; PC3, green. Data were normalized to total histone proteins and to untreated cells for each of three cell lines (control = 1). Results are mean ± SD from three independent experiments. Data are presented for both HDACis as indicated. (C) Correlation between RPPA and Western blot analysis. The graphs show regression for romidepsin (left) and vorinostat (right) with P < 0.001 for both treatments. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.3. Analysis of cell cycle distribution reveals cell type‐specific responses to HDACI treatment

We performed cell cycle analysis on the nineteen cell lines after a 24 h treatment using 3‐fold serial dilutions of romidepsin (0.1–30 ng/ml, 0.2–54 nM) and vorinostat (0.1–30 μM), again adjusting for the difference in potency between romidepsin and vorinostat. As shown in Figure 4 and Supplementary Figures S4A–C, the effects of HDACI treatment were different among the cell lines, ranging from only minor changes in G1/S/G2‐M distribution as in MDA‐MB‐435 to mostly apoptosis in H146. As shown in Figure 4C, we found three basic patterns of cell cycle response to HDACis: (1) predominant G1 arrest (MCF‐10A, UACC‐62), (2) predominant G2 arrest (PC3, LOX‐IMVI, MDA‐MB‐231, MDA‐MB‐435, SW620, S1), including a subset with an increasing Sub‐G1 population indicative of apoptosis (MCF‐7, A549, EKVX, HCT116 p21−/−), and (3) cells undergoing apoptosis either without appearing to have entered cell cycle arrest or leaving cell cycle arrest very rapidly (H146, H526, HuT‐78, SK‐BR‐3, H460, HCT116). It is interesting to note that the cell cycle effect most often mentioned, G1 arrest, was least common among the 19 cell lines. Figure 4D shows that a cell cycle arrest pattern emerged when the general caspase inhibitor Q‐VD‐OPh was added to prevent cell death.

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g004.jpg

Effect of HDACis on cell cycle distribution in nineteen cell lines. (A and B) Exponentially growing cells were treated with romidepsin (A) or vorinostat (B) in concentrations 0.1–30 ng/ml or μM, respectively, for 24 h and cell cycle analysis was performed after propidium iodide staining using a FACS with FlowJo software. The colors indicate different concentrations of HDACis. A representative example of three or more independent experiments is shown here. Cell cycle data for each cell line are shown in Supplementary Figure S4. (C) Three different patterns of cell cycle response to HDACis were observed among the individual analyses shown in Supplementary Figure 4B and C, representing cell cycle distribution in the cell lines over a range of concentrations from 0.1 to 30 0.1–30 ng/ml or μM, as shown in Panel A. Stacked bar graphs depict the cell cycle parameters in control cells and following treatment with romidepsin at 10 ng/ml. Data are representative of three or more independent experiments. (D) Cell cycle analysis was performed for 3 cell lines from the group of those undergoing apoptosis rather than cell cycle arrest. When romidepsin (10 or 25 ng/ml) was simultaneously added with the 10 μM of general caspase inhibitor Q‐VD‐OPh, a G2 arrest emerged.

Notably, as in the RPPA, romidepsin and vorinostat again appear to have similar effects in the equipotent dose range used, a result supported by the high correlation coefficients for all cell cycle parameters (Supplementary Figure S4D). For example, the Pearson correlation coefficients between the two drugs, for all cell line parameters, exceeds 0.54 for G1 arrest in concentrations ≥ 3 ng/mL (5.4 nM) for romidepsin and ≥ 3 μM for vorinostat.

We also determined the Km – concentration of HDACi required to produce a half‐maximal cell cycle effect – in some cases measurable only for G2‐M arrest or Sub‐G1 increase. Examples of the plots for these Km determinations are shown in Figure 5A and Supplementary Figure 5. Often the decrease in G1 content that accompanied the G2‐M arrest could be fit to a curve and the Km calculated. Providing an internal validation of this quantitation, the Km's for the effect on G1 and G2‐M content were correlated for both romidepsin and vorinostat (R = 0.92, P = 0.0014, n = 8 for romidepsin; R = 0.95, P = 0.015, n = 5 for vorinostat). When the Km for drug effects on G2‐M and on Sub‐G1 were combined in plots against the Km's for G1 (Figure 5B), the overall correlation was R = 0.80, P = 0.0006, n = 13 for romidepsin. Similar observations were made for vorinostat (Figure 5C) when one outlier was excluded (R = 0.89, P = 0.002, n = 9). We also specifically evaluated the sub‐G1 region of DNA content in cells treated with the highest concentration of the drugs for 24 h, which shows a strong correlation in sub‐G1 between romidepsin and vorinostat (R = 0.96), but also discriminates between cell lines in that there are two major groupings of the cell lines in sub‐G1 effect (Figure 5D).

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g005.jpg

Kinetic analysis of cell cycle parameters (A) Two representative plots of cell cycle parameters determined by flow cytometry (see Methods) as a function of romidepsin (left) or vorinostat (right) concentrations. The lines drawn are the best‐fit predictions to ascending or descending hyperbolas using Sigmaplot software (see Methods). Km values (the concentrations required to reach half‐maximal effect on cell cycle) were obtained where the line could be adequately fit to the data. Additional plots are shown in Supplementary Figure 5. (B) Correlation of the Km values for cell cycle parameters G2‐M and Sub‐G1 with Km for G1 phase following 24 h romidepsin. The data arise from plots such as those in Fig 5A. The line drawn is the line of identity. The regression through the combined points (not shown) has R = 0.80, P = 0.0006, n = 13. (C) As in B above, but for vorinostat. The regression through the combined points (not shown) has R = 0.48, P = 0.16, n = 9. Leaving out the very deviant point, the regression (again not shown) has R = 0.89, P = 0.002, n = 8. (D) Analysis of the delta Sub‐G1 population for 30 ng/ml (54 nM) romidepsin and 30 µM vorinostat. Results are mean from three independent experiments. The data cluster in two groups: a “low” Sub‐G1 and a “high” Sub‐G1 group. The inset shows the distribution of the “low” Sub‐G1 responses (between 0 and 4.5%). The H146 cell line exhibits an extreme Sub‐G1 response compared to the rest of the cell lines. Small cell lung cancer (SCLC), light blue; non‐small cell lung cancer (NSCLC), dark blue; melanoma, gray; CNS, purple; T‐cell lymphoma, red; prostate, green; colon, brown; and breast, pink. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.4. An assay that discriminates among the cell lines and between drugs

Given that romidepsin is administered over 4 h and vorinostat administered in a single daily oral dose, and that both drugs have half‐lives under 3 h, we concluded that long exposures do not mirror clinical HDACi dosing and utilized a short‐term assay for apoptosis detection, with cells exposed to HDACi for 6 h, and harvested after an additional 42 h incubation in drug‐free medium before staining for annexin positivity (Yu et al., 2007). A representative histogram quantitating annexin staining is shown in Figure 6A; in Figure 6B results for all cell lines treated at 25 ng/ml (45 nM) romidepsin and 25 μM vorinostat are graphed (excluding two cell lines that do not stain for annexin). This method provided for the first time a clear spectrum of HDACi sensitivity across the cell lines and identified some differential activity between the two HDACis. Although romidepsin's and vorinostat's efficacy by annexin assay were correlated (R = 0.69, P = 0.0022, n = 17), romidepsin was generally four times more potent in the induction of apoptosis, despite the use of concentrations assessed as equipotent as in the cell cycle and growth inhibition assays (Figure 6C). As might be expected from the uniform pattern of acetylation observed in Figure 2, romidepsin's and vorinostat's efficacy, as determined in this annexin assay, were not correlated with the absolute difference in H3K9ac, P = 0.94 and P = 0.56, respectively (Figure 6D and 6E), nor was annexin staining correlated with median Km values for the histone modifications (For romidepsin, P = 0.46, for vorinostat, P = 0.87). Including in the analysis only those 12 cell lines for which the annexin signal for romidepsin was less than 30% improved the correlation between the Km for histone modification and those for the cell cycle parameters, for example the Pearson correlation coefficient between the Km for the histone marks and the Km for the G2‐M arrest was R = 0.75, P = 0.019, n = 9 (a tighter correlation than the R = 0.629, P = 0.05 correlation noted in the entire group) (Supplementary Figure 6A and B), suggesting, as did the sub‐G1 fraction data, that a separate subset of cell lines exists in which rapid cell death occurs apart from the epigenetic effects.

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g006.jpg

Annexin assay shows variation among the cell lines for both romidepsin and vorinostat. (A) A short‐term exposure and annexin staining method was utilized for detection of apoptosis. Cells were treated with 25 ng/ml (45 nM) romidepsin or 25 μM vorinostat for 6 h, followed by washing, and incubation in drug‐free medium for additional 42 h. Analysis of apoptotic cell death was performed as described in methods, and percentage of annexin V positivity was calculated as shown on the dot plot. (B) Analysis of apoptotic cell death in seventeen cell lines following HDACI treatment. Bar graphs display results for romidepsin in black and vorinostat in gray, and colors indicate tissue of origin. The data for annexin staining are shown as the difference between control and treated. Results are mean ± SD from three to six independent experiments. Two cell lines, H146 and H526, are not shown because they did not stain for annexin under any condition. (C) Correlation of annexin responses (shown as in B, in delta values) following treatment with romidepsin and vorinostat. The regression line has R = 0.69, P < 0.05. (D) and (E) As in (C), except that the Y‐axis records the histone mark H3K9ac signal obtained by subtracting the untreated control value from the 10 ng/ml romidepsin or 10 μM vorinostat values. Color legends in the bottom indicate cell lines for plots C, D, and E.

3.5. Differential sensitivity of cancer cell lines to HDACi treatment depends on cell context

Because we have previously noted an interaction of HDACi and the MAPK pathway (Chakraborty et al., 2013; Ieranò et al., 2012), we examined the effect of HDACis on selected signaling proteins by immunoblot analysis using the same cells, doses, and exposures as in Figure 3. Two of the cell lines, MDA‐MB‐231 and A549, share common mutations, KRAS and CDKN2A, while the PC3 cell line has a PTEN mutation, and is relatively resistant to HDACi effect (Table 1). We compared the ratio of phosphorylated and total MEK1/2, ERK1/2, and AKT in MDA‐MB‐231, A549 and PC3 cells. Because phosphorylated AKT was relatively lower in MDA‐MB‐231 cells, results were calibrated to the MDA‐MB‐231 cells as shown in Figure 7A and B. Although basal levels of phosphorylated MEK1/2, ERK1/2 and AKT were different across these cell lines, both drugs inhibited phosphorylation in all of these cells. The analysis of other signaling proteins showed decreasing expression of p‐Rb, c‐Myc, cyclin D1, but induction of p21 and the apoptosis‐related molecules c‐PARP and Bim following treatment with HDACis (Figure 7C). Altogether, these results indicate that the cell death response to HDACi is more variable than the molecular effects of the HDACi. There is one point of divergence between romidepsin and vorinostat. It known that romidepsin has a relatively lower potency against HDAC6, which acetylates non‐histone proteins such as α‐tubulin, β‐catenin, Hsp90 and others (Crabb et al., 2008; Robey et al., 2011; Zhang et al., 2003). We confirmed this, showing a marked induction of α‐tubulin acetylation following treatment with vorinostat but not with romidepsin (Figure 7C).

An external file that holds a picture, illustration, etc. Object name is MOL2-8-1379-g007.jpg

Association of response to HDACi with cell context. (A) Cells were treated for 24 h with romidepsin (Romi) or vorinostat (Vor) as indicated and then examined for cell cycle and signaling molecules by immunoblot analysis. Data are representative example of three independent experiments. (B) The different levels of p‐MEK, p‐ERK and p‐AKT following HDACi treatment are shown in three different cell lines (MDA‐MB‐231 indicated in pink, A549 – in blue, PC3 – in green). Graph bar represents the expression of indicated signaling molecules after normalization to total proteins and to MDA‐MB‐231‐ untreated cells (control = 1). Data are mean ± SD from three independent experiments for each HDACi as indicated. (C) Quantitative analysis of selected signaling proteins from immunoblots in (B) shows similarity across cell lines and between romidepsin (left) and vorinostat (right). Numbers were normalized to GAPDH‐loading control values, and the effect of HDACis‐treatment in each cell line was compared to untreated cells as shown in the graph bars. Data are mean ± SD from three independent experiments. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

4. Discussion

We report a systematic approach to understanding the mechanism of action of HDACis. We examined 18 human cancer cell lines with different molecular phenotypes, and one immortalized breast cell line, and analyzed the effect of romidepsin or vorinostat treatment at multiple concentrations and at different time points on growth inhibition and cell death, cell cycle distribution, histone modification, and protein expression. We selected two FDA‐approved HDACis, with different structures and HDAC affinity profiles and compared them across cell lines and with each other, generating insights into mechanism of HDACi activity. These studies show that HDACis induce similar global effects, including growth inhibition and acetylation – across cell lines and between drugs. Cell cycle perturbations were largely similar for romidepsin and vorinostat. Cell death assays detecting annexin expression demonstrated that a subset of cell lines is markedly more sensitive to death following a short exposure to the HDACi.

The most widely used assays for testing HDACi efficacy are a continuous (96‐h) cytotoxicity/growth inhibition assay, histone acetylation, and cell cycle effects. In the 96‐h cytotoxicity assays, we found the HDACis had nearly identical effects after taking into account the 1000‐fold lower IC50 values for romidepsin compared to vorinostat. When histone modifications were examined by reverse phase proteomic assay, it appeared that the global effect across cell lines was similar, with almost all cell lines responding with acetylation of the 13 different marks tested, suggesting that cell lines respond similarly to both HDACis with comparable degrees of histone acetylation. In vitro studies of inhibition of the most prevalent HDAC enzymes showed that romidepsin is by far the most potent agent in development and that its most prominent effect was inhibition of the Class I HDACs: HDAC 1, 2, 3, and 8 (Bradner et al., 2010). Vorinostat, in contrast, is able to inhibit Class II HDACs at concentrations near those required to inhibit the Class I enzymes. We were able to confirm the effect of vorinostat but not romidepsin on the Class II enzymes by showing α‐tubulin acetylation following the former, but not the latter. This suggests that the HDAC6 inhibition by vorinostat, as manifested by α‐tubulin acetylation in vitro, is not responsible for the clinical activity of romidepsin and vorinostat against T‐cell lymphomas. Indeed, HDAC6 inhibition may effectively be an off‐target effect for vorinostat in the therapy of in T‐cell lymphoma. However, in a different clinical context in which the HDAC6‐regulated Hsp90 chaperone function may be critical, such as HER2 amplification, HDAC6 inhibition by vorinostat could be of value.

Given the differences in HDAC specificity, we were surprised to find how similar were the effects of the two HDACis on cell cycle beginning at 3 ng/ml (5.4 nM) romidepsin and 3 μM vorinostat. However, not all cell lines responded similarly to treatment with the HDACis. In contrast to previously published observations, most of the cell lines fail to show a clear G1 cell cycle arrest (Sandor et al., 2000; Zhang et al., 2005). These variations are likely due to the different status of cell cycle checkpoints in different cell lines (Kastan and Bartek, 2004). Consistent with this, inhibition of caspase to prevent cell death revealed a G2 arrest in cells typically responding with apoptosis rather than cell cycle arrest.

The consistent effects on acetylation and cell cycle are similar to numerous reports in the literature showing HDACi activity in virtually any cell line tested (Crabb et al., 2008; Gravina et al., 2012; Imre et al., 2006; Marrocco‐Tallarigo et al., 2009; Mitić and McKay, 2005; Palmieri et al., 2009). Yet, this is in conflict with clinical observations in which HDACis as single agents are active to date exclusively in T‐cell lymphomas and other hematologic malignancies. We hypothesized that this discordance was due to the long exposure durations typically tested in cell culture, very different from the 2–3 h half‐life of the HDACis in patients (Piekarz et al., 2011, 2009). We showed that a short exposure to romidepsin followed by measurement of cell death by annexin positivity was a more discriminating assay in showing differences between cell lines than the continuous 96‐h assay. Some cell lines were completely insensitive in this assay, something not observed with the 96‐h growth inhibition assay. We also found romidepsin to be a more effective inducer of apoptosis in the annexin assay than vorinostat, despite equalization of potency based on cell cycle analysis. Interestingly, when we examined correlations between annexin and cell cycle arrest or histone modification, such correlations were only readily observed in those cell lines that did not show rapid cell death.

From these studies emerged a recognition of two distinct HDACi effects. The ubiquitous global effect on acetylation is in contrast to the differential annexin sensitivity observed after a 6 h exposure and also in contrast to the differential sensitivity seen in the clinic. The data suggest that the effects of HDACis, while in a spectrum, can be viewed as of two principle types. The first HDACi effect is a Type I effect that primarily affects chromatin and includes histone acetylation, altered gene expression, cell cycle arrest, and growth inhibition. This effect will lead to cell death if the exposure to the HDACi persists long enough. This is the effect studied when HDAC inhibitors were viewed primarily as differentiating agents (Richon et al., 1996). If we need to optimize expression of a target gene for a drug combination, it is likely that lower doses and longer exposure will be most effective (Amiri‐Kordestani et al., 2013; Kitazono et al., 2001).

At the other end of the spectrum, the second HDACi effect is the ability of HDACis to rapidly invoke cell death in some cell types – in our assay, induced by a brief 6‐h exposure to the HDACis. It may be useful to separate this Type II cell death effect from the Type I primarily chromatin effect, and make a goal of future drug development shifting from Type I to Type II, unless we specifically seek a particular gene induction effect. The HDAC enzyme target remains the same in both response types, and our data suggest that it is not the magnitude of acetylation but rather the susceptibility to apoptosis that determines the sensitivity. We propose that the 6‐h exposure, which more closely replicates the clinical exposure duration for most HDACis, induces a burst of global acetylation that we hypothesize is the initiating event that induces cell death in apoptosis‐susceptible cells, such as T‐cell lymphomas. This burst of acetylation may be recognized by cells as DNA damage – Conti et al. reported stalling of replication forks (and hypothesized a comparable effect on transcription) that resulted in DNA double strand breaks (Conti et al., 2010). If this is a key mechanism, then the DNA damage is triggering apoptosis in susceptible cells, based on cell context. This would explain not only the clinical results with rapid destruction of malignant T‐cells (Piekarz et al., 2001), but would also explain the widely observed susceptibility to combinations with agents that target various apoptosis proteins. Multiple studies have shown that HDACi sensitivity is impacted by the presence or absence of proteins modulating apoptosis – presence of Bak and Bax (Ieranò et al., 2012; Shao et al., 2010), a requirement for Bim (Chakraborty et al., 2013; Yang et al., 2009), and downregulation of proapoptotic BCL2A1 (Bolden 2013). Different molecular contexts will differentially impact the expression and function of the proteins involved in apoptosis.

Taken together, the data suggest that different cellular contexts determine the response to a ubiquitous and fairly consistent epigenetic effect. Supporting this argument are correlations observed in a pathway analysis of gene expression data correlating with panobinostat sensitivity in 500 cell lines of the Cancer Cell Line Encyclopedia. As shown in Supplementary Figure S7, 124 genes were correlated with panobinostat sensitivity. A search for upstream regulators for those 124 genes found multiple proteins associated with the MAPK pathway among the top hits. In turn, we compared HDACi sensitivity to mutations conferring MAPK pathway activation among the 19 cell lines and found those to be more sensitive to HDAC inhibitor than cell lines with mutations in the PI3K or p53 pathways (Supplementary Figure S7). If acetylation‐induced DNA damage is confirmed as the dominant mechanism of action of HDACis, clinical studies can then focus on approaches to modify sensitivity in specific molecular phenotypes. Together these studies point to a need to select specific molecular phenotypes for sensitivity to the HDACis and then to identify relevant combinations as a way forward in the development of these epigenetic therapies.

Conflicts of interest

The authors have no conflicts of interest to declare and no financial interests to report. Dr. Bates received research funding from Celgene Pharmaceuticals (CRADA #01683) through a Cooperative Research and Development Agreement with the National Cancer Institute.

Supporting information

The following is the supplementary data related to this article:

Supplementary data

Acknowledgments

This work was supported by the Intramural Research Program of the National Cancer Institute. The authors would like to thank Dan Sackett for advice over many phases of this project, and Julian Bahr for assisting in data acquisition.

Supplementary data 1. 

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.molonc.2014.05.001.

Notes

Luchenko Victoria L., Litman Thomas, Chakraborty Arup R., Heffner Aaron, Devor Christopher, Wilkerson Julia, Stein Wilfred, Robey Robert W., Bangiolo Lois, Levens David, Bates Susan E., (2014), Histone deacetylase inhibitor‐mediated cell death is distinct from its global effect on chromatin, Molecular Oncology, 8, doi: 10.1016/j.molonc.2014.05.001. [PMC free article] [PubMed] [Google Scholar]

References

  • Amiri-Kordestani, L. , Luchenko, V. , Peer, C.J. , Ghafourian, K. , Reynolds, J. , Draper, D. , Frye, R. , Woo, S. , Venzon, D. , Wright, J. , 2013. Phase I trial of a new schedule of romidepsin in patients with advanced cancers. Clin. Cancer Res.. 19, 4499–4507. [PMC free article] [PubMed] [Google Scholar]
  • Bali, P. , Pranpat, M. , Bradner, J. , Balasis, M. , Fiskus, W. , Guo, F. , Rocha, K. , Kumaraswamy, S. , Boyapalle, S. , Atadja, P. , 2005. Inhibition of histone deacetylase 6 acetylates and disrupts the chaperone function of heat shock protein 90: a novel basis for antileukemia activity of histone deacetylase inhibitors. J. Biol. Chem.. 280, 26729–26734. [PubMed] [Google Scholar]
  • Bates, S. , Zhan, Z. , Steadman, K. , Obrzut, T. , Luchenko, V. , Frye, R. , Robey, R. , Turner, M. , Gardner, E. , Figg, W. , 2010. Laboratory correlates for a phase II trial of romidepsin in cutaneous and peripheral T-cell lymphoma. Br. J. Haematol.. 148, 256–267. [PMC free article] [PubMed] [Google Scholar]
  • Baumann, P. , Junghanns, C. , Mandl-Weber, S. , Strobl, S. , Oduncu, F. , Schmidmaier, R. , 2012. The pan-histone deacetylase inhibitor CR2408 disrupts cell cycle progression, diminishes proliferation and causes apoptosis in multiple myeloma cells. Br. J. Haematol.. 156, 633–642. [PubMed] [Google Scholar]
  • Blagosklonny, M.V. , Robey, R. , Sackett, D.L. , Du, L. , Traganos, F. , Darzynkiewicz, Z. , Fojo, T. , Bates, S.E. , 2002. Histone deacetylase inhibitors all induce p21 but differentially cause tubulin acetylation, mitotic arrest, and cytotoxicity. Mol. Cancer Ther.. 1, 937–941. [PubMed] [Google Scholar]
  • Bradner, J.E. , West, N. , Grachan, M.L. , Greenberg, E.F. , Haggarty, S.J. , Warnow, T. , Mazitschek, R. , 2010. Chemical phylogenetics of histone deacetylases. Nat. Chem. Biol.. 6, 238–243. [PMC free article] [PubMed] [Google Scholar]
  • Chakraborty, A.R. , Robey, R.W. , Luchenko, V.L. , Zhan, Z. , Piekarz, R.L. , Gillet, J.P. , Kossenkov, A.V. , Wilkerson, J. , Showe, L.C. , Gottesman, M.M. , 2013. MAPK pathway activation leads to Bim loss and histone deacetylase inhibitor resistance: rationale to combine romidepsin with a MEK inhibitor. Blood. [PMC free article] [PubMed] [Google Scholar]
  • Chen, S. , Dai, Y. , Pei, X.Y. , Grant, S. , 2009. Bim upregulation by histone deacetylase inhibitors mediates interactions with the Bcl-2 antagonist ABT-737: evidence for distinct roles for Bcl-2, Bcl-xL, and Mcl-1. Mol. Cell Biol.. 29, 6149–6169. [PMC free article] [PubMed] [Google Scholar]
  • Conti, C. , Leo, E. , Eichler, G. , Sordet, O. , Martin, M. , Fan, A. , Aladjem, M. , Pommier, Y. , 2010. Inhibition of histone deacetylase in cancer cells slows down replication forks, activates dormant origins, and induces DNA damage. Cancer Res.. 70, 4470–4480. [PMC free article] [PubMed] [Google Scholar]
  • Crabb, S.J. , Howell, M. , Rogers, H. , Ishfaq, M. , Yurek-George, A. , Carey, K. , Pickering, B.M. , East, P. , Mitter, R. , Maeda, S. , 2008. Characterisation of the in vitro activity of the depsipeptide histone deacetylase inhibitor spiruchostatin A. Biochem. Pharmacol.. 76, 463–475. [PubMed] [Google Scholar]
  • Glaser, K.B. , Staver, M.J. , Waring, J.F. , Stender, J. , Ulrich, R.G. , Davidsen, S.K. , 2003. Gene expression profiling of multiple histone deacetylase (HDAC) inhibitors: defining a common gene set produced by HDAC inhibition in T24 and MDA carcinoma cell lines. Mol. Cancer Ther.. 2, 151–163. [PubMed] [Google Scholar]
  • Gravina, G.L. , Marampon, F. , Giusti, I. , Carosa, E. , Di Sante, S. , Ricevuto, E. , Dolo, V. , Tombolini, V. , Jannini, E.A. , Festuccia, C. , 2012. Differential effects of PXD101 (belinostat) on androgen-dependent and androgen-independent prostate cancer models. Int. J. Oncol.. 40, 711–720. [PubMed] [Google Scholar]
  • Gu, W. , Roeder, R.G. , 1997. Activation of p53 sequence-specific DNA binding by acetylation of the p53 C-terminal domain. Cell. 90, 595–606. [PubMed] [Google Scholar]
  • Gupta, M. , Han, J.J. , Stenson, M. , Wellik, L. , Witzig, T.E. , 2012. Regulation of STAT3 by histone deacetylase-3 in diffuse large B-cell lymphoma: implications for therapy. Leukemia. 26, 1356–1364. [PMC free article] [PubMed] [Google Scholar]
  • Ieranò, C. , Basseville, A. , To, K.K. , Zhan, Z. , Robey, R.W. , Wilkerson, J. , Bates, S.E. , Scala, S. , 2012. Histone deacetylase inhibitors induce CXCR4 mRNA but antagonize CXCR4 migration. Cancer Biol. Ther.. 14, [PMC free article] [PubMed] [Google Scholar]
  • Imre, G. , Gekeler, V. , Leja, A. , Beckers, T. , Boehm, M. , 2006. Histone deacetylase inhibitors suppress the inducibility of nuclear factor-kappaB by tumor necrosis factor-alpha receptor-1 down-regulation. Cancer Res.. 66, 5409–5418. [PubMed] [Google Scholar]
  • Jiang, X. , Tsang, Y.H. , Yu, Q. , 2007. c-Myc overexpression sensitizes Bim-mediated Bax activation for apoptosis induced by histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA) through regulating Bcl-2/Bcl-xL expression. Int. J. Biochem. Cell Biol.. 39, 1016–1025. [PubMed] [Google Scholar]
  • Kaminskyy, V.O. , Surova, O.V. , Vaculova, A. , Zhivotovsky, B. , 2011. Combined inhibition of DNA methyltransferase and histone deacetylase restores caspase-8 expression and sensitizes SCLC cells to TRAIL. Carcinogenesis. 32, 1450–1458. [PubMed] [Google Scholar]
  • Kastan, M.B. , Bartek, J. , 2004. Cell-cycle checkpoints and cancer. Nature. 432, 316–323. [PubMed] [Google Scholar]
  • Khan, O. , La Thangue, N.B. , 2012. HDAC inhibitors in cancer biology: emerging mechanisms and clinical applications. Immunol. Cell Biol.. 90, 85–94. [PubMed] [Google Scholar]
  • Kitazono, M. , Robey, R. , Zhan, Z. , Sarlis, N.J. , Skarulis, M.C. , Aikou, T. , Bates, S. , Fojo, T. , 2001. Low concentrations of the histone deacetylase inhibitor, depsipeptide (FR901228), increase expression of the Na(+)/I(-) symporter and iodine accumulation in poorly differentiated thyroid carcinoma cells. J. Clin. Endocrinol. Metab.. 86, 3430–3435. [PubMed] [Google Scholar]
  • Lane, A. , Chabner, B. , 2009. Histone deacetylase inhibitors in cancer therapy. J. Clin. Oncol.. 27, 5459–5468. [PubMed] [Google Scholar]
  • Marrocco-Tallarigo, D.L. , Centenera, M.M. , Scher, H.I. , Tilley, W.D. , Butler, L.M. , 2009. Finding the place of histone deacetylase inhibitors in prostate cancer therapy. Expert Rev. Clin. Pharmacol.. 2, 619–630. [PubMed] [Google Scholar]
  • Mitić, T. , McKay, J.S. , 2005. Immunohistochemical analysis of acetylation, proliferation, mitosis, and apoptosis in tumor xenografts following administration of a histone deacetylase inhibitor–a pilot study. Toxicol. Pathol.. 33, 792–799. [PubMed] [Google Scholar]
  • Nebbioso, A. , Clarke, N. , Voltz, E. , Germain, E. , Ambrosino, C. , Bontempo, P. , Alvarez, R. , Schiavone, E.M. , Ferrara, F. , Bresciani, F. , 2005. Tumor-selective action of HDAC inhibitors involves TRAIL induction in acute myeloid leukemia cells. Nat. Med.. 11, 77–84. [PubMed] [Google Scholar]
  • Newbold, A. , Lindemann, R.K. , Cluse, L.A. , Whitecross, K.F. , Dear, A.E. , Johnstone, R.W. , 2008. Characterisation of the novel apoptotic and therapeutic activities of the histone deacetylase inhibitor romidepsin. Mol. Cancer Ther.. 7, 1066–1079. [PubMed] [Google Scholar]
  • Olsen, E.A. , Kim, Y.H. , Kuzel, T.M. , Pacheco, T.R. , Foss, F.M. , Parker, S. , Frankel, S.R. , Chen, C. , Ricker, J.L. , Arduino, J.M. , Duvic, M. , 2007. Phase IIb multicenter trial of vorinostat in patients with persistent, progressive, or treatment refractory cutaneous T-cell lymphoma. J. Clin. Oncol.. 25, 3109–3115. [PubMed] [Google Scholar]
  • Palmieri, D. , Lockman, P.R. , Thomas, F.C. , Hua, E. , Herring, J. , Hargrave, E. , Johnson, M. , Flores, N. , Qian, Y. , Vega-Valle, E. , 2009. Vorinostat inhibits brain metastatic colonization in a model of triple-negative breast cancer and induces DNA double-strand breaks. Clin. Cancer Res.. 15, 6148–6157. [PMC free article] [PubMed] [Google Scholar]
  • Peart, M.J. , Smyth, G.K. , van Laar, R.K. , Bowtell, D.D. , Richon, V.M. , Marks, P.A. , Holloway, A.J. , Johnstone, R.W. , 2005. Identification and functional significance of genes regulated by structurally different histone deacetylase inhibitors. Proc. Natl. Acad. Sci. U S A. 102, 3697–3702. [PMC free article] [PubMed] [Google Scholar]
  • Piekarz, R. , Bates, S. , 2009. Epigenetic modifiers: basic understanding and clinical development. Clin. Cancer Res.. 15, 3918–3926. [PMC free article] [PubMed] [Google Scholar]
  • Piekarz, R.L. , Frye, R. , Prince, H.M. , Kirschbaum, M.H. , Zain, J. , Allen, S.L. , Jaffe, E.S. , Ling, A. , Turner, M. , Peer, C.J. , 2011. Phase 2 trial of romidepsin in patients with peripheral T-cell lymphoma. Blood. 117, 5827–5834. [PMC free article] [PubMed] [Google Scholar]
  • Piekarz, R.L. , Frye, R. , Turner, M. , Wright, J.J. , Allen, S.L. , Kirschbaum, M.H. , Zain, J. , Prince, H.M. , Leonard, J.P. , Geskin, L.J. , 2009. Phase II multi-institutional trial of the histone deacetylase inhibitor romidepsin as monotherapy for patients with cutaneous T-cell lymphoma. J. Clin. Oncol.. 27, 5410–5417. [PMC free article] [PubMed] [Google Scholar]
  • Piekarz, R.L. , Robey, R. , Sandor, V. , Bakke, S. , Wilson, W.H. , Dahmoush, L. , Kingma, D.M. , Turner, M.L. , Altemus, R. , Bates, S.E. , 2001. Inhibitor of histone deacetylation, depsipeptide (FR901228), in the treatment of peripheral and cutaneous T-cell lymphoma: a case report. Blood. 98, 2865–2868. [PubMed] [Google Scholar]
  • Richon, V.M. , Webb, Y. , Merger, R. , Sheppard, T. , Jursic, B. , Ngo, L. , Civoli, F. , Breslow, R. , Rifkind, R.A. , Marks, P.A. , 1996. Second generation hybrid polar compounds are potent inducers of transformed cell differentiation. Proc. Natl. Acad. Sci. U S A. 93, 5705–5708. [PMC free article] [PubMed] [Google Scholar]
  • Robbins, A.R. , Jablonski, S.A. , Yen, T.J. , Yoda, K. , Robey, R. , Bates, S.E. , Sackett, D.L. , 2005. Inhibitors of histone deacetylases alter kinetochore assembly by disrupting pericentromeric heterochromatin. Cell Cycle. 4, 717–726. [PubMed] [Google Scholar]
  • Robey, R.W. , Chakraborty, A.R. , Basseville, A. , Luchenko, V. , Bahr, J. , Zhan, Z. , Bates, S.E. , 2011. Histone deacetylase inhibitors: emerging mechanisms of resistance. Mol. Pharm.. 8, 2021–2031. [PMC free article] [PubMed] [Google Scholar]
  • Robey, R.W. , Zhan, Z. , Piekarz, R.L. , Kayastha, G.L. , Fojo, T. , Bates, S.E. , 2006. Increased MDR1 expression in normal and malignant peripheral blood mononuclear cells obtained from patients receiving depsipeptide (FR901228, FK228, NSC630176). Clin. Cancer Res.. 12, 1547–1555. [PubMed] [Google Scholar]
  • Rosato, R. , Grant, S. , 2003. Histone deacetylase inhibitors in cancer therapy. Cancer Biol. Ther.. 2, 30–37. [PubMed] [Google Scholar]
  • Sandor, V. , Senderowicz, A. , Mertins, S. , Sackett, D. , Sausville, E. , Blagosklonny, M.V. , Bates, S.E. , 2000. P21-dependent g(1)arrest with downregulation of cyclin D1 and upregulation of cyclin E by the histone deacetylase inhibitor FR901228. Br. J. Cancer. 83, 817–825. [PMC free article] [PubMed] [Google Scholar]
  • Shao, W. , Growney, J.D. , Feng, Y. , O'Connor, G. , Pu, M. , Zhu, W. , Yao, Y.M. , Kwon, P. , Fawell, S. , Atadja, P. , 2010. Activity of deacetylase inhibitor panobinostat (LBH589) in cutaneous T-cell lymphoma models: defining molecular mechanisms of resistance. Int. J. Cancer. 127, 2199–2208. [PubMed] [Google Scholar]
  • Venugopal, B. , Evans, T.R. , 2011. Developing histone deacetylase inhibitors as anti-cancer therapeutics. Curr. Med. Chem.. 18, 1658–1671. [PubMed] [Google Scholar]
  • Xargay-Torrent, S. , López-Guerra, M. , Saborit-Villarroya, I. , Rosich, L. , Campo, E. , Roué, G. , Colomer, D. , 2011. Vorinostat-induced apoptosis in mantle cell lymphoma is mediated by acetylation of proapoptotic BH3-only gene promoters. Clin. Cancer Res.. 17, 3956–3968. [PubMed] [Google Scholar]
  • Xiao, J.J. , Huang, Y. , Dai, Z. , Sadée, W. , Chen, J. , Liu, S. , Marcucci, G. , Byrd, J. , Covey, J.M. , Wright, J. , 2005. Chemoresistance to depsipeptide FK228 [(E)-(1S,4S,10S,21R)-7-[(Z)-ethylidene]-4,21-diisopropyl-2-oxa-12,13-dithia-5,8,20,23-tetraazabicyclo[8,7,6]-tricos-16-ene-3,6,9,22-pentanone] is mediated by reversible MDR1 induction in human cancer cell lines. J. Pharmacol. Exp. Ther.. 314, 467–475. [PubMed] [Google Scholar]
  • Xu, W. , Parmigiani, R. , Marks, P. , 2007. Histone deacetylase inhibitors: molecular mechanisms of action. Oncogene. 26, 5541–5552. [PubMed] [Google Scholar]
  • Yang, Y. , Zhao, Y. , Liao, W. , Yang, J. , Wu, L. , Zheng, Z. , Yu, Y. , Zhou, W. , Li, L. , Feng, J. , 2009. Acetylation of FoxO1 activates Bim expression to induce apoptosis in response to histone deacetylase inhibitor depsipeptide treatment. Neoplasia. 11, 313–324. [PMC free article] [PubMed] [Google Scholar]
  • Yeh, C.C. , Deng, Y.T. , Sha, D.Y. , Hsiao, M. , Kuo, M.Y. , 2009. Suberoylanilide hydroxamic acid sensitizes human oral cancer cells to TRAIL-induced apoptosis through increase DR5 expression. Mol. Cancer Ther.. 8, 2718–2725. [PubMed] [Google Scholar]
  • Yu, X.D. , Wang, S.Y. , Chen, G.A. , Hou, C.M. , Zhao, M. , Hong, J.A. , Nguyen, D.M. , Schrump, D.S. , 2007. Apoptosis induced by depsipeptide FK228 coincides with inhibition of survival signaling in lung cancer cells. Cancer J.. 13, 105–113. [PubMed] [Google Scholar]
  • Zhang, C. , Richon, V. , Ni, X. , Talpur, R. , Duvic, M. , 2005. Selective induction of apoptosis by histone deacetylase inhibitor SAHA in cutaneous T-cell lymphoma cells: relevance to mechanism of therapeutic action. J. Invest. Dermatol.. 125, 1045–1052. [PubMed] [Google Scholar]
  • Zhang, Y. , Li, N. , Caron, C. , Matthias, G. , Hess, D. , Khochbin, S. , Matthias, P. , 2003. HDAC-6 interacts with and deacetylates tubulin and microtubules in vivo. EMBO J.. 22, 1168–1179. [PMC free article] [PubMed] [Google Scholar]
Articles from Molecular Oncology are provided here courtesy of Wiley
-