Abstract
There are 18 histone deacetylases (HDAC) generally divided into four classes based on homology to yeast HDACs. HDACs have many protein substrates in addition to histones that are involved in regulation of gene expression, cell proliferation, and cell death. Inhibition of HDACs can cause accumulation of acetylated forms of these proteins, thus altering their function. HDAC inhibitors (HDACi), such as the hydroxamic acid–based vorinostat (suberoylanilide hydroxamic acid), inhibit the zinc-containing classes I, II, and IV, but not the NAD+-dependent class III, enzymes. HDACis are a group of novel anticancer agents. Vorinostat is the first HDACi approved for clinical use in the treatment of the cancer cutaneous T-cell lymphoma. Factors affecting expression of HDACs are not well understood. This study focuses on the effect of the HDACi vorinostat on the expression of class I and class II HDACs. We found that vorinostat selectively down-regulates HDAC7 with little or no effect on the expression of other class I or class II HDACs. Fourteen cell lines were examined, including normal, immortalized, genetically transformed, and human cancer-derived cell lines. Down-regulation of HDAC7 by vorinostat is more pronounced in transformed cells sensitive to inhibitor-induced cell death than in normal cells or cancer cells resistant to induced cell death. Modulation of HDAC7 levels by small interfering RNA–mediated knockdown or by HDAC7 overexpression is associated with growth arrest but without detectable changes in acetylation of histones or p21 gene expression. Selective down-regulation of HDAC7 protein may serve as a marker of response of tumors to HDACi. [Mol Cancer Ther 2007;6(9):2525–34]
Introduction
Histone deacetylases (HDAC) comprise 18 members in humans and are generally classified into class I and class II based on homologies to yeast deacetylases RPD3 and HDA1, respectively, and class III with homology to yeast Sir2 group (1–4). Class I includes HDAC1, HDAC2, HDAC3, and HDAC8; class II includes HDAC4, HDAC5, HDAC6, HDAC7, and HDAC9. HDAC6 and HDAC10 have two catalytic sites and may be called class IIB. HDAC11 has conserved residues in its catalytic core that are shared by both class I and class II enzymes and is sometimes placed in a class IV (2). Classes I, II, and IV HDACs are zinc dependent. In addition to histones, HDACs have many nonhistone protein targets, which are involved in regulation of gene expression, cell proliferation, migration, and death pathways (1–4). Furthermore, phylogenetics evidence indicates that these enzymes are present in organisms that do not have histones (5). It seems that these enzymes may more properly be considered “lysine deacetylases” rather than specifically HDACs (6).
Biological activity of different HDACs are not redundant. Class I HDACs generally have a nuclear localization, whereas class II HDACs may shuttle between cytoplasm and nucleus and have biological activities related to specific cell lineages (1–4, 6). Class III HDACs have homology to yeast Sir2 group of deacetylases, have an absolute requirement for NAD+, and are not inhibited by compounds that inhibit classes I and II enzymes (e.g., trichostatin A or vorinostat). Analysis of HDAC1 knockout phenotype indicates that it is essential for proliferation and viability of embryos despite compensatory increases in HDAC2 and HDAC3 (7). HDAC9 null mice have defects in cardiac muscle development (8, 9). HDAC7 null mouse embryos display defects in blood vessel development and integrity (10). HDAC7 is involved in T-cell differentiation (11, 12). In addition to inhibition of enzymatic activity, valproic acid is reported to selectively induce proteasomal degradation of HDAC2 (13). Vorinostat down-regulates HDAC3 and several other proteins in chronic myeloid cell lines (14).
The expression of certain HDACs is altered in tumor compared with normal cells. Increased expression of HDAC1 has been found in gastric cancers, esophageal squamous cell carcinoma, hormone-refractory prostate cancer, and colon cancer (15–18). HDAC2 expression is increased in human colon cancer, in which there is a loss of APC tumor suppressor (19). HDAC3 expression is increased in human tumor samples of colon cancer compared with normal tissue (18, 20). HDAC5 is down-regulated in colon cancer (21). HDAC proteins are recruited to oncogenic fusion proteins in various leukemias (22).
HDAC inhibitors (HDACi) include a group of novel targeted agents with significant anticancer activity in patients with both hematologic and solid tumors at doses that are well tolerated by patients (1–4, 23). Vorinostat (suberoylanilide hydroxamic acid), the leading HDACi in clinical development, is a hydroxamic acid that inhibits catalytic activities of both class I and class II HDACs (23). Vorinostat has received Food and Drug Administration approval for the treatment of cutaneous T-cell lymphoma, representing the first of the new HDACi to be approved for clinical use (23).
Normal cells are relatively resistant to HDACi-induced cell death. The difference in sensitivity of normal and transformed cells to HDACi seems to be related to “cell context” (1–4, 6). HDACi can alter the structure and function of a broad range of proteins regulating cell proliferation, migration, and death that are substrates of HDACs (1–4, 6). Cancer cells generally have multiple defects in proteins regulating cell proliferation, cell migration, and cell death (1–4, 6, 24). Thus, cancer cell may have less capacity to compensate for the HDACi effects than normal cells (6, 25, 26).
This study focuses on the effects of the HDACi on the expression of class I and class II HDAC proteins. We found that vorinostat selectively suppresses HDAC7 with little or no effect on the expression of other class I and class II HDACs examined. Fourteen cell lines were studied, including three normal, three immortalized, and eight transformed cell lines, of which six were sensitive and two were resistant to vorinostat-induced cell death. Down-regulation of HDAC7 protein by vorinostat was most pronounced in transformed cells, which are sensitive to HDACi-induced cell death. In normal cells and in cancer cells that are resistant to HDACi-induced cell death, there was relatively little or no detectable suppression of HDAC7 protein. Depsipeptide, a structurally different HDACi than vorinostat, had similar effects. The down-regulation of HDAC7 protein by HDACi is associated with down-regulation of HDAC7 mRNA with no alteration in the half-life of HDAC7 protein. Selective down-regulation of HDAC7 by small interfering RNA resulted in growth arrest in normal and transformed cells but did not cause other changes associated with vorinostat, e.g., cell death of transformed cells, induction of p21, and accumulation of acetylated histones or of acetylated tubulin. Overexpression of HDAC7 protein was associated with growth arrest in transformed cells but did not block vorinostat effects on transformed cell death. The present findings indicate that HDACi can selectively and profoundly suppress HDAC7 expression in transformed cells sensitive to HDACi-induced cell death. Expression of HDAC7 may be a marker of response of tumor cells to vorinostat therapy.
Materials and Methods
Cells and Cell Culture
The following normal cells were used: normal lung fibroblast (WI38) obtained from American Type Culture Collection (ATCC), normal foreskin fibroblast cell line (hFS) obtained from Cell Culture Core Facility-Yale Skin Diseases Research Center, and normal prostate epithelium cell line (PrEC) obtained from Clonetics/Cambrex-Cambrex BioScience Rockland, Inc. (Table 1). Genetically manipulated cell lines include lung fibroblast (VA13), foreskin fibroblast (BJE, BJELB, BJELR), and prostate epithelium (PrEC-LE, PRC-LHSR) that were immortalized or transformed by distinct oncogenic elements [cell lines were obtained from ATCC and Weinberg (27) and Hahn (28)]. Human prostate cancer–derived cell lines (LNCaP, DU145, PC3), human bladder carcinoma–derived cell lines (T24), and multiple myeloma–derived cell lines (ARP-1) were obtained from ATCC (Table 1). Normal, immortalized and transformed foreskin cells were maintained in culture as described (27). Human prostate cancer cell lines were maintained in culture as previously described (24). Immortalized human prostate epithelial cell (PrEC-LE) and transformed human prostate epithelial cells (PrEC-LHSR) were kindly provided by Dr. William Hahn (Dana-Faber Cancer Center; ref. 28). All prostate epithelial cells (normal, immortalized, and transformed) were propagated in defined medium (PrEGM), as recommended by Clonetics/Cambrex (28).
Characteristics and sources of cells
Cell line . | Original cell line . | Cell type . | Genetic manipulation . | Mutations . | Reference/source . |
---|---|---|---|---|---|
hFS | Foreskin fibroblast | Normal | None | NA | Yale Skin Diseases Research Center |
BJE | Foreskin fibroblast | Immortalized* | hTERT† | NA | Weinberg laboratory (27) |
BJELB | Foreskin fibroblast | immortalized | hTERT and SV40 oncogenic products | NA | Weinberg laboratory (27) |
BJELR | Foreskin fibroblast | Transformed‡ | hTERT, SV40, H-Ras | NA | Weinberg laboratory (27) |
PrEC | Prostate epithelial | Normal | none | NA | Clonetics/Cambrex |
PrEC-LE | Prostate epithelial | Immortalized | Large T and hTERT | NA | Hahn laboratory (28) |
PrEC-LHSR | Prostate epithelial | Transformed | Large T, hTERT, small T, and H-Ras | NA | Hahn laboratory (28) |
LNCaP | Lymph node metastasis | Prostate cancer§ | N/A | PTEN-mutated | ATCC |
p53-wild type | |||||
Rb-wild type | |||||
DU-145 | Brain metastasis | Prostate Cancer | NA | PTEN-wild type | ATCC |
p53-mutated | |||||
Rb-mutated | |||||
PC-3 | Bone metastasis | Prostate Cancer | NA | PTEN-mutated | ATCC |
P53-mutated | |||||
RB-wild type | |||||
T24 | Bladder carcinoma | Transformed | NA | P53-mutated | ATCC |
Rb-positive | |||||
WI-38 | Lung fibroblast | Normal | NA | NA | ATCC |
VA-13 | Lung fibroblast | Transformed | SV40-transformed | NA | ATCC |
ARP-1 | Multiple myeloma | Transformed | NA | NA | J. Hardoc (Arkansans Cancer Research Center) |
Cell line . | Original cell line . | Cell type . | Genetic manipulation . | Mutations . | Reference/source . |
---|---|---|---|---|---|
hFS | Foreskin fibroblast | Normal | None | NA | Yale Skin Diseases Research Center |
BJE | Foreskin fibroblast | Immortalized* | hTERT† | NA | Weinberg laboratory (27) |
BJELB | Foreskin fibroblast | immortalized | hTERT and SV40 oncogenic products | NA | Weinberg laboratory (27) |
BJELR | Foreskin fibroblast | Transformed‡ | hTERT, SV40, H-Ras | NA | Weinberg laboratory (27) |
PrEC | Prostate epithelial | Normal | none | NA | Clonetics/Cambrex |
PrEC-LE | Prostate epithelial | Immortalized | Large T and hTERT | NA | Hahn laboratory (28) |
PrEC-LHSR | Prostate epithelial | Transformed | Large T, hTERT, small T, and H-Ras | NA | Hahn laboratory (28) |
LNCaP | Lymph node metastasis | Prostate cancer§ | N/A | PTEN-mutated | ATCC |
p53-wild type | |||||
Rb-wild type | |||||
DU-145 | Brain metastasis | Prostate Cancer | NA | PTEN-wild type | ATCC |
p53-mutated | |||||
Rb-mutated | |||||
PC-3 | Bone metastasis | Prostate Cancer | NA | PTEN-mutated | ATCC |
P53-mutated | |||||
RB-wild type | |||||
T24 | Bladder carcinoma | Transformed | NA | P53-mutated | ATCC |
Rb-positive | |||||
WI-38 | Lung fibroblast | Normal | NA | NA | ATCC |
VA-13 | Lung fibroblast | Transformed | SV40-transformed | NA | ATCC |
ARP-1 | Multiple myeloma | Transformed | NA | NA | J. Hardoc (Arkansans Cancer Research Center) |
Abbreviation: NA, not applicable.
Immortalized cells grow indefinitely in suitable media but do not cause tumors on transplant to animal models.
Retroviral infection with indicated gene(s).
Transformed cells will form a tumor when transplanted to animal model.
Prostate cancer cell lines derived from biopsies of human prostate cell metastatic site.
Human bladder carcinoma–derived cell line T24 (29), normal lung fibroblast cell line WI-38, and transformed lung fibroblast cell line VA-13 were obtained from ATCC and cultured as previously described (26, 29). Multiple myeloma–derived cell line ARP-1 was maintained in culture as previously described (30).
For protein analysis, cells were plated at ∼106 per 150-mm plate and harvested at described time points. For cell growth and viability curves, cells were plated at 25,000 to 50,000 cells per 1-mL well in triplicate. Cells were recovered after trypsinization, and cell density and viability were determined (29).
Constructs
Flag-tagged HDAC7-overexpressing construct in pcDNA3.1 vector was a gift from Dr. Eric Verdin, University of California, San Francisco (31).
Drugs and Chemicals
Western Blot Analysis
Cell lysates were prepared, and the protein concentration of extracted lysate was determined by a Bradford protein assay (29). Protein (30 μg) was subjected to SDS-PAGE (Bio-Rad) and transferred to a Hybond-poly (vinylidene difluoride) membrane (Amersham Pharmacia Biosciences). Membranes were then blocked with PBS–0.1% Tween with 5% bovine serum albumin and incubated with specific primary antibody for 18 to 24 h. Horseradish peroxidase–labeled secondary antibody was added and visualized using the Western Lightening Chemiluminescence reagent (Perkin-Elmer Life Sciences).
Antibodies
The primary antibodies used were as follows: polyclonal rabbit anti-mouse HDAC1 at 1:300 (Upstate Biotechnology), polyclonal rabbit anti-human HDAC2 at 1:1000 (Calbiochem), polyclonal rabbit anti-human HDAC3 at 1:300 (Santa Cruz Biotechnology), polyclonal rabbit anti-human HDAC4, HDAC5, HDAC6, HDAC7, and HDAC8 at 1:1,000 (Santa Cruz Biotechnology), polyclonal rabbit anti-HDAC10 antibody at 1:200 (Bio-Vision), monoclonal mouse antitubulin at 1:2,000 (Calbiochem), anti-p21 mouse monoclonal antibody CP74 at 1:1,000 (Neo Marker), mouse monoclonal anti–acetylated tubulin antibody at 1:1,000 (Sigma), polyclonal rabbit anti–acetyl histone H3 at 1:200 (Upstate), and polyclonal rabbit anti–acetyl histone H4 at 1:200 (Upstate).
The secondary antibodies used were as follows: donkey anti-rabbit horseradish peroxidase conjugated at 1:10,000 (Amersham Biosciences) and goat anti-mouse peroxidase conjugated at 1:10,000 (Jackson Immunoresearch).
Small Interfering RNA Transfection
Small interfering RNA transfections were done by electroporating respective cell lines using conditions commercially available from Amaxa, Inc. HDAC7 small interfering RNA was purchased from Qiagen (Hs_HDAC7A-2-HP small interfering RNA). HDAC7 sense small interfering RNA sequence is (GGCUGGAAACAGAAACCCA) dTdT, and HDAC7 antisense small interfering RNA sequence is r(UGGGUUUCUGUUUCCAGCC) dTdT. Control (nonsilencing) small interfering RNA was commercially available from Qiagen. Control (mock) transfections were done with buffer used for dissolving small interfering RNA.
Northern Blot Analysis
mRNA for Northern blot analysis was prepared by harvesting cells and resuspending cellular material in Trizol according to the standard protocol (34). HDAC7 probe was prepared from commercially available partial cDNA clone. (RZPD-GMBH).
Measurement of HDAC7 Protein Half-life
To measure HDAC7 protein half-life, transformed foreskin fibroblast cells BJELR were pulse chased according to a modified protocol described in Xu et al. (35). Cells were grown to 50% confluence in 60-mm dishes and preincubated for 30 min in methionine-free 5% FCS DMEM to deplete cold methionine and cysteine from the cells. Cells were washed once in PBS, then labeled for 3 h with 10 mCi/mL S35 methionine/cysteine in methionine/cysteine-free DMEM–10% FCS. Reaction was terminated by the addition of excess cold methionine/cysteine. Cells were rinsed twice in ice-cold PBS and then cultured for 4 h in the presence and absence of 5 μmol/L vorinostat. Cells were washed in cold PBS, recovered, lysed, and frozen in liquid nitrogen until cells are ready to perform immunoprecipitation. TCA precipitation experiment determined labeled incorporation into the cells. HDAC7 immunoprecipitation with polyclonal anti–HDAC-7 rabbit antibody was done (35). HDAC7 band was detected by drying the gel and placing it overnight in a storage phosphor screen (Amersham Biosciences). The signal of the amplified bands was quantified with a PhosphoiImager (Molecular Dynamics) using ImageQant software.
Results
Vorinostat Inhibits Growth of Normal, Immortalized, and Transformed Cells
Vorinostat inhibited the growth of all cells studied: normal cells (hFS, PrEC, WI38), immortalized cells (BJE, BJELB, PrEC-LE), and transformed cells (BJELR, PrEC-LHSR, ARP-1, VA13, LNCaP, DU-145, PC-3, and T24; Fig. 1; Table 1). Data were not shown for LNCaP, DU-145, PC-3, ARP-1, WI-38,VA-13, and T24, as these were previously published (24, 26, 29, 30, 34). Cells differed in sensitivity to vorinostat-induced cell growth arrest; ARP-1 cell growth was inhibited at 1 μmol/L vorinostat, whereas 5 μmol/L vorinostat were required to completely inhibit the growth of normal (hFS), immortalized (BJE and BJELB), and transformed (BJELR) foreskin cells (Fig. 1A). Vorinostat at 2.5 μmol/L inhibited growth of normal (PrEC), immortalized (PrEC-LE), and transformed (PrEC-LHSR) prostate epithelial cells (Fig. 1B). The effects of vorinostat on normal PrEC cell line growth were only apparent after day 3, and this may reflect slow growth of PrEC normal prostate cells (Fig. 1B). The cell number in control PrEC culture was 2 × 104 cells per milliliter compared with 1.3 × 104 cells per milliliter in cultures with the inhibitor. Vorinostat at 2.5 μmol/L inhibited growth of prostate cancer–derived cell lines LNCaP, DU-145, and PC-3 (24) and WI38 and VA-13 cells (26). Culture with 5 μmol/L vorinostat completely inhibited the growth of T24 bladder carcinoma–derived cell line (29).
Effect of vorinostat on cell growth and viability of normal, immortalized, and transformed cells. A, foreskin fibroblast cells; normal (hFS), immortalized (BJE and BJELB), and transformed (BJELR) cells were cultured with different concentrations of vorinostat (as indicated). B, different prostate epithelium cells; normal (PrEC), immortalized (PrEC-LE), and transformed (PrEC-LHSR) were cultured with different concentrations of vorinostat (as indicated).
Effect of vorinostat on cell growth and viability of normal, immortalized, and transformed cells. A, foreskin fibroblast cells; normal (hFS), immortalized (BJE and BJELB), and transformed (BJELR) cells were cultured with different concentrations of vorinostat (as indicated). B, different prostate epithelium cells; normal (PrEC), immortalized (PrEC-LE), and transformed (PrEC-LHSR) were cultured with different concentrations of vorinostat (as indicated).
Vorinostat-Induced Cell Death of Transformed Cells, But Not of Normal Cells
Normal (hFS,WI38, and PrEC) and immortalized (BJE) foreskin cell lines were relatively resistant to vorinostat-induced cell death compared with immortalized cell lines BJELB and PrEC-LE and transformed cell lines BJELR, PrEC-LHSR, VA13, ARP1, LNCaP, and DU145 (previously published data for WI38, VA-13, ARP-1, LNCaP, DU-145; refs. 24, 26, 30, 34; Fig. 1A and B). Culture with 5 μmol/L vorinostat caused 10% loss in viability in immortalized BJELB and >20% loss in viability in the transformed cell line BJELR by day 3 (Fig. 1A). Vorinostat at 5 μmol/L caused no detectable loss in viability in immortalized cell line BJE or in normal cell lines hFS and WI38 and normal prostate cells PrEC (data previously published; ref. 26; Fig. 1). Vorinostat at 10 μmol/L caused an 80% loss in viability in immortalized PrEC-LE and >90% loss in viability in transformed cell line PrEC-LHSR by day 3 (Fig. 1B). Prostate cancer–derived cell lines LNCaP and DU-145 lost >50% in viability in culture with 5 μmol/L vorinostat, whereas PC-3 cells were resistant to HDACi-induced cell death (24). Bladder carcinoma cell line T24 was resistant to vorinostat-induced cell death at 5 μmol/L (29). The relative resistance of the normal cells (foreskin and prostate) to vorinostat-induced cell death, compared with certain transformed cells, is consistent with previous findings (26, 36).
Effect of Vorinostat on Expression of Class I and Class II HDACs
We examined the effect of vorinostat on expression of class I HDAC1, HDAC2, HDAC3, and HDAC8 and class II HDAC4 and HDAC6 in foreskin fibroblast cells hFS (normal fibroblast cell line), BJE (immortalized cell line), BJELB (immortalized cell line), and BJELR (transformed cell line; Fig. 2A): the effect of vorinostat on the expression of class I HDAC1, HDAC2, HDAC3, and HDAC8 and class II HDAC4, HDAC5, and HDAC6 in prostate epithelium-derived cells, PrEC (normal cells), PrEC-LE (immortalized cell line), and PrEC-LHSR (transformed cell line; Fig. 2B) and class I HDAC1, HDAC2, and HDAC3 and class II HDAC4, HDAC6, and HDAC10 in human prostate cancer cells LNCaP, DU-145, and PC-3 (Fig. 2C). There was little or no consistent change in the level of these class I (HDAC1, HDAC2, HDAC3, and HDAC8) and class II (HDAC4, HDAC5, HDAC6, and HDAC10) HDACs in normal, immortalized, or transformed cells cultured with 5 μmol/L vorinostat for 24 or 48 h (Fig. 2A-C).
Effect of vorinostat on HDAC1, HDAC2, HDAC3, HDAC8 and HDAC4, HDAC5, HDAC6 protein levels in normal, immortalized, and transformed cells. A, Western blot analysis of class I HDAC1, HDAC2, HDAC3, HDAC8 and class II HDAC4 and HDAC6 in normal (hFS), immortalized (BJE and BJELB), and transformed (BJELR) cells cultured without (no treatment) and with vorinostat (5 μmol/L) for the times indicated. α-Tubulin was used as normalization for Western blot analysis. B, Western blot analysis of class I HDAC1, HDAC2, HDAC3, and HDAC8 and class II HDAC4, HDAC5, and HDAC6 of normal (PrEC), immortalized (PrEC-LE), and transformed (PrEC-LHSR) cells cultured without (no treatment) and with vorinostat (5 μmol/L) for times indicated. C, Western blot analysis of class I HDAC1, HDAC2, and HDAC3 and class II HDAC4, HDAC6, and HDAC10 in transformed human prostate cells LNCaP and DU145 sensitive to vorinostat (5 μmol/L)–induced cell death and PC3 resistant to vorinostat (5 μmol/L)–induced cell death.
Effect of vorinostat on HDAC1, HDAC2, HDAC3, HDAC8 and HDAC4, HDAC5, HDAC6 protein levels in normal, immortalized, and transformed cells. A, Western blot analysis of class I HDAC1, HDAC2, HDAC3, HDAC8 and class II HDAC4 and HDAC6 in normal (hFS), immortalized (BJE and BJELB), and transformed (BJELR) cells cultured without (no treatment) and with vorinostat (5 μmol/L) for the times indicated. α-Tubulin was used as normalization for Western blot analysis. B, Western blot analysis of class I HDAC1, HDAC2, HDAC3, and HDAC8 and class II HDAC4, HDAC5, and HDAC6 of normal (PrEC), immortalized (PrEC-LE), and transformed (PrEC-LHSR) cells cultured without (no treatment) and with vorinostat (5 μmol/L) for times indicated. C, Western blot analysis of class I HDAC1, HDAC2, and HDAC3 and class II HDAC4, HDAC6, and HDAC10 in transformed human prostate cells LNCaP and DU145 sensitive to vorinostat (5 μmol/L)–induced cell death and PC3 resistant to vorinostat (5 μmol/L)–induced cell death.
Vorinostat caused a decrease in HDAC7 protein level in a dose- and time-dependent manner in all cells examined (Fig. 3). A decrease in HDAC7 protein was detectable in normal hFS cells only in culture with 5 μmol/L vorinostat for 48 h or 10 μmol/L vorinostat for 24 h. In immortalized cells (BJE and BJELB), the decrease in HDAC7 was detectable at 5 μmol/L vorinostat after 24 h of culture, whereas in transformed cells (BJELR), a decrease was detectable at 1 μmol/L vorinostat within 24 h of culture (Fig. 3A).
Effect of vorinostat on HDAC7 protein level in normal, immortalized, and transformed cells. A, Western blot analysis of HDAC7 level in hFS, BJE, BJELB, and BJELR cells cultured with vorinostat for the time and with concentration of inhibitor indicated. B, Western blot analysis of HDAC7 level in prostate epithelium cells (PrEC, PrEC-LE, and PrEC-LHSR) cultured without (no treatment) and with vorinostat. C, Western blot analysis of HDAC7 in LNCaP, DU145, and PC3. D, Western blot analysis of HDAC7 in bladder carcinoma–derived cell line T24 cultured with vorinostat as indicated. Sensitive and resistant refers to vorinostat-induced cell death. Culture of the cell lines with vorinostat was done for the duration and concentration as indicated. HDAC2 was a loading control in series (A) as well as α-tubulin (from Fig. 2B and C).
Effect of vorinostat on HDAC7 protein level in normal, immortalized, and transformed cells. A, Western blot analysis of HDAC7 level in hFS, BJE, BJELB, and BJELR cells cultured with vorinostat for the time and with concentration of inhibitor indicated. B, Western blot analysis of HDAC7 level in prostate epithelium cells (PrEC, PrEC-LE, and PrEC-LHSR) cultured without (no treatment) and with vorinostat. C, Western blot analysis of HDAC7 in LNCaP, DU145, and PC3. D, Western blot analysis of HDAC7 in bladder carcinoma–derived cell line T24 cultured with vorinostat as indicated. Sensitive and resistant refers to vorinostat-induced cell death. Culture of the cell lines with vorinostat was done for the duration and concentration as indicated. HDAC2 was a loading control in series (A) as well as α-tubulin (from Fig. 2B and C).
PrEC (Fig. 3B) showed relatively little suppression of HDAC7 proteins compared with immortalized (PrEC-LE) and transformed PrEC-LHSR cells cultured with 5 μmol/L vorinostat for 24 h. A decrease in HDAC7 level was detected in PrEC-LHSR prostate cancer cells with 5 μmol/L vorinostat at 24 h of culture (Fig. 3B).
In prostate cancer cell lines LNCaP and DU-145, 5 μmol/L vorinostat caused a marked decrease in HDAC7 within 24 h of cultures (Fig. 3C). In PC-3 cell line resistant to vorinostat-induced cell death, the decrease in HDAC7 was much less pronounced than in either LNCaP or DU145 (Fig. 3C). In the culture T24 cells, a cell line resistant to vorinostat-induced cell death (29), HDAC7 protein decrease is pronounced only in culture with 10 μmol/L vorinostat for 48 h (Fig. 3D). ARP-1, a human multiple myeloma cell line, is very sensitive to vorinostat-induced cell death (26) and showed a marked decrease in HDAC7 protein in culture with 1.0 μmol/L within 24 h of culture (data not shown).
We examined the effect of another HDACi, depsipeptide, that is structurally different than vorinostat on HDAC7 level in cells sensitive (DU145 and LNCaP) and resistant (PC-3 cells) to induced death by this agent. Depsipeptide, similar to vorinostat, caused a selective decrease in HDAC7 in LNCaP and DU-145 to a much greater degree than in PC-3 cells (data not shown).
Effect of Vorinostat on HDAC7 mRNA Expression
We next determined if the selective decrease in HDAC7 protein level induced by vorinostat was associated with a decrease in HDAC7 mRNA. We used HDAC7 partial cDNA as a probe and 18S as a loading control (Fig. 4). Culture with 5 μmol/L vorinostat caused a decrease in HDAC7 mRNA in normal cells at 6 and 14 h, whereas by 6 h, it was undetectable in immortalized foreskin (BJE and BJELB) and transformed foreskin (BJELR) cells that persisted through 48-h duration of culture (Fig. 4). Vorinostat at 5 μmol/L caused a profound decrease in HDAC7 mRNA in prostate cancer cell lines LNCaP, DU-145, and PC-3 (data not shown).
Effect of vorinostat on the level of HDAC7 mRNA message in foreskin cells. Northern blot analysis was done on foreskin fibroblast cell samples collected from cells cultured with 5 μmol/L vorinostat for the times indicated. 18S mRNA was used as a loading control.
Effect of vorinostat on the level of HDAC7 mRNA message in foreskin cells. Northern blot analysis was done on foreskin fibroblast cell samples collected from cells cultured with 5 μmol/L vorinostat for the times indicated. 18S mRNA was used as a loading control.
Vorinostat did not alter HDAC7 protein stability as determined in transformed BJELR cells (data not shown).
Effect of HDAC7 Small Interfering RNA and Transient Overexpression of HDAC7 Protein
We next determined the effect of suppressing HDAC7 protein with HDAC7 small interfering RNA. Mock-treated cells and nonsilencing control oligos were used as controls and compared with the effects of down-regulating HDAC7 protein by small interfering RNA. After electroporation, the cells were allowed to recover for 24 h and then replated. HDAC7 small interfering RNA oligos strongly reduced HDAC7 protein within 24 h after electroporation (day 0) in both normal (hFS) and transformed (BJELR) cells (Fig. 5A). To test the effect of HDACi in mock, nonsilencing, and HDAC7 small interfering RNA–treated BJELR cells, the cells were cultured without and with different concentrations of vorinostat. In cells cultured without inhibitor, a similar level of HDAC7 protein was found in mock and nonsilencing BJELR cells whereas there was no detectable HDAC7 protein in cells electroporated with HDAC7 small interfering RNA oligo (Fig. 5B). In BJELR cells cultured with 2 or 5 μmol/L vorinostat for 24 h, as expected, there was little or no detectable HDAC7 protein. hFS and BJELR cells treated with small interfering RNA oligos for HDAC7 had a decreased rate of growth but no loss in viability compared with control cells. There was no difference in sensitivity to vorinostat-induced cell death in cells with HDAC7 knockdown compared with cells expressing HDAC7 (data not shown).
Effect of small interfering RNA for HDAC7 on HDAC7 expression in normal (hFS) and transformed (BJELR) cells. A, Western blot analysis of HDAC7 level in hFS and BJELR cells after electroporation with mock, nonsilencing, and small interfering RNA for HDAC7 oligos (day 0). B, Western blot analysis of HDAC7 in cells cultured with different concentrations of vorinostat (as indicated) for 24 h. α-Tubulin was used for loading control. C, Western blot analysis of HDAC7, p21, and acetylated tubulin in BJELR cells electroporated with small interfering RNA for HDAC7 to down-regulate HDAC7 and electroporated with HDAC7-expressing vector to overexpress HDAC7. BJELR cells were also cultured with vorinostat at concentrations indicated for 24 h, and HDAC7 levels assayed by Western blot.
Effect of small interfering RNA for HDAC7 on HDAC7 expression in normal (hFS) and transformed (BJELR) cells. A, Western blot analysis of HDAC7 level in hFS and BJELR cells after electroporation with mock, nonsilencing, and small interfering RNA for HDAC7 oligos (day 0). B, Western blot analysis of HDAC7 in cells cultured with different concentrations of vorinostat (as indicated) for 24 h. α-Tubulin was used for loading control. C, Western blot analysis of HDAC7, p21, and acetylated tubulin in BJELR cells electroporated with small interfering RNA for HDAC7 to down-regulate HDAC7 and electroporated with HDAC7-expressing vector to overexpress HDAC7. BJELR cells were also cultured with vorinostat at concentrations indicated for 24 h, and HDAC7 levels assayed by Western blot.
Vorinostat can cause an increase in global histone acetylation, tubulin acetylation, and p21 level in many transformed cells (29, 30). Suppression of HDAC7 by small interfering RNA oligos, as well as transient overexpression of HDAC7 from cytomegalovirus-based construct did not result in change of p21 expression or the level of acetylated tubulin in transformed BJELR cells (Fig. 5C). Levels of global histone H3 and histone H4 acetylation were not also changed by HDAC7 knockdown (data not shown). Vorinostat increased expression of p21 and accumulation of acetylated tubulin and acetylated histones in these transformed cells (Fig. 5C).
Discussion
This study has found that HDACis, vorinostat and depsipeptide, selectively down-regulate HDAC7, a class II HDAC. This is the first report that identifies class II HDAC7 protein level as a target for vorinostat and depsipeptide. We have shown this down-regulation of HDAC7 with 14 different cell lines, including a normal foreskin fibroblast, a normal prostate epithelial and a normal lung fibroblast cell line, two immortalized foreskin fibroblast and immortalized prostate epithelium cell lines, and transformed cell lines derived either by genetic manipulation of foreskin fibroblast and prostate epithelium or derived from prostate cancer, bladder carcinoma, and multiple myeloma. Vorinostat-induced HDAC7 down-regulation is associated with reduction in HDAC7 mRNA level with no change in protein stability. This suggests that vorinostat may have a role in selectively down-regulating HDAC7 gene transcription. Although selective effects of vorinostat on HDAC7 mRNA stability or posttranscriptional processing cannot be ruled out, the kinetics of down-regulation (strong effects on HDAC7 mRNA level by 6 h) suggest that vorinostat may be more likely to exert effects on gene transcription. Efforts to clone the HDAC7 promoter to assay the effect of vorinostat on this gene's promoter were unsuccessful, possibly owing to the inability to identify and clone an active promoter sequence for this gene.
HDACi-induced HDAC7 down-regulation is dependent on “cell context” (molecular profile of cultured cells), dose of inhibitor, and time of exposure. Of the cells tested, the transformed cells sensitive to vorinostat-induced cell death showed the most profound down-regulation of HDAC7 protein whereas normal cells and prostate cancer cells resistant to vorinostat-induced cell death were least sensitive to HDACi induced down-regulation of HDAC7. The present study does not elucidate the mechanisms for this selective effect of vorinostat in the down-regulation of HDAC7 mRNA and HDAC7 protein.
HDAC7 protein has been implicated in several biological processes, including regulation of gene expression either as coactivator or corepressor (31, 37–39), sensitization of cells to apoptotic stimuli by T-cell receptor by inducing Nur77 (12), or maintenance of vascular integrity in developing embryos by repressing matrix metalloproteinase-10 (10). Oncoproteins Plag1 and Plag2 are deacetylated, and their activity was repressed by HDAC7 protein (40). In vitro studies with recombinant HDAC7 protein have suggested that histones may be substrates of HDAC7 deacetylase activity (37). Our studies have shown that HDAC7 knockdown with small interfering RNA did not affect the levels of global histone acetylation.
Down-regulation of HDAC7 by small interfering RNA resulted in slower cell growth with no effect on viability of normal or transformed cells. Although the mechanism of growth inhibition is not clear, our data indicate that this effect is independent of p21.
Vorinostat has been recently approved for clinical treatment of cutaneous T-cell lymphoma, a rare T-cell malignancy (23). There are multiple clinical trials, either alone or in combination with other agents, that are currently under way with vorinostat. HDAC7 may be a useful biomarker predictive of response to therapy with vorinostat.
Grant support: NIH grant P30CA-08748-41, Jack and Susan Rudin Foundation grant, David H. Koch Foundation grant, Experimental Therapeutics Center at Memorial Sloan-Kettering Cancer Center Prostate Cancer Research award, and DeWitt Wallace Research fund.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Acknowledgments
We thank Dr. Eric Verdin (University of California, San Francisco) for the gift of HDAC7-overexpressing plasmid and Joann Perrone and Mabel Miranda for their expertise in literature searches and preparation of this manuscript. Vorinostat and related compounds are the intellectual property of Columbia University and Sloan Kettering Institute, which provided Aton Pharma an exclusive license. Merck acquired Aton Pharma in April 2004. P.A. Marks was a founder of Aton and has a continuing financial interest in Merck's development of vorinostat.