Development of a histone deacetylase 6 inhibitor and its biological effects

Significance

We discovered a hydroxamic acid-based small-molecule N-hydroxy-4-(2-[(2-hydroxyethyl)(phenyl)amino]-2-oxoethyl)benzamide selectively inhibits histone deacetylase 6 catalytic activity in vivo and in vitro.

Abstract

Development of isoform-selective histone deacetylase (HDAC) inhibitors is important in elucidating the function of individual HDAC enzymes and their potential as therapeutic agents. Among the eleven zinc-dependent HDACs in humans, HDAC6 is structurally and functionally unique. Here, we show that a hydroxamic acid-based small-molecule N-hydroxy-4-(2-[(2-hydroxyethyl)(phenyl)amino]-2-oxoethyl)benzamide (HPOB) selectively inhibits HDAC6 catalytic activity in vivo and in vitro. HPOB causes growth inhibition of normal and transformed cells but does not induce cell death. HPOB enhances the effectiveness of DNA-damaging anticancer drugs in transformed cells but not normal cells. HPOB does not block the ubiquitin-binding activity of HDAC6. The HDAC6-selective inhibitor HPOB has therapeutic potential in combination therapy to enhance the potency of anticancer drugs.

Histone deacetylase 6 (HDAC6) is unique among the eleven zinc-dependent HDACs in humans. HDAC6 is located in the cytoplasm, and it has two catalytic domains and an ubiquitin-binding domain at the C-terminal region (1–3). This study focused on the development of a HDAC6-selective inhibitor and its biological effects. The substrates of HDAC6 include nonhistone proteins such as α-tubulin, peroxiredoxin (PRX), cortactin, and heat shock protein 90 (Hsp90) but not histones (4–7). HDAC6 plays a key role in the regulation of microtubule dynamics including cell migration and cell–cell interactions. The reversible acetylation of Hsp90, a substrate of HDAC6, modulates its chaperone activity and, accordingly, the stability of survival and antiapoptotic factors, including epidermal growth factor receptor (EGFR), protein kinase AKT, proto-oncogene C-RAF, survivin, and other factors. HDAC6, through its ubiquitin-binding activity and interaction with other partner proteins, plays a role in the degradation of misfolded proteins by binding polyubiquitinated proteins and delivering them to the dynein and motor proteins for transport into aggresomes which are degraded by lysosomes (8–10). Thus, HDAC6 has multiple biological functions through deacetylase-dependent and -independent mechanisms modulating many cellular pathways relevant to normal and tumor cell growth, migration, and death. HDAC6 is an attractive target for potential cancer treatment.

There are several previous reports on the development of HDAC6-selective inhibitors (11–15). The most extensively studied is tubacin (16, 17). Tubacin has non–drug-like qualities, high lipophilicity, and difficult synthesis and has proved to be more useful as a research tool rather than as a potential drug (18). We and others (12–15, 19) have developed HDAC6-selective inhibitors whose pharmacokinetics, toxicity, and efficacy make them potentially more useful than tubacin as therapeutic agents. ACY-1215, 2-(Diphenylamino)-N-(7-(hydroxyamino)-7-oxoheptyl)pyrimidine-5-carboxamide, a HDAC6-selective inhibitor, is currently being evaluated in clinical trials (http://clinicaltrials.gov).

HDAC inhibitors, such as suberoylanilide hydroxamic acid (SAHA), consist of three structural domains: a metal-binding domain, a linker domain, and a surface domain (20). The catalytic pocket of HDAC1 is deeper and narrower than the catalytic pocket of HDAC6 (14). To develop HDAC6-selective inhibitors, we synthesized small molecules with bulkier and shorter linker domains than the pan-HDAC inhibitor SAHA (20, 21). A hydroxamic acid-based small-molecule N-hydroxy-4-(2-[(2-hydroxyethyl)(phenyl)amino]-2-oxoethyl)benzamide (HPOB) was synthesized that selectively inhibits HDAC6. We report the effects of this HDAC6-selective inhibitor on normal and transformed cells. Further, we found that selective inhibition of HDAC6 increases the effectiveness of anticancer agents, etoposide, doxorubicin, and SAHA in inducing cell death of transformed cells but not normal cells.

Results

Synthesis of the HDAC6-Selective Inhibitor.

HPOB was synthesized from commercially available materials in five steps with an overall yield of 36% (Fig. 1A). (i) Reaction of aniline with glycolaldehyde in dichloroethane yielded an imine intermediate, which was subsequently reduced with sodium triacetoxyborohydride to give 2-(phenylamino)ethanol, compound 2. (ii) The reactive hydrophilic hydroxyl group of compound 2 was protected with tert-butyldimethylsilyl-chloride (TBDMS-Cl) to give N-(2-[(tert-butyldimethylsilyl)oxy]ethyl)aniline, compound 3. (iii) Compound 8 was obtained from oxidation of commercially available 4-(Hydroxymethyl)phenylacetic acid with calcium hypochlorite in the presence of methanol in acetonitrile, using 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide to yield methyl 4-(2-[(2-[(tert-butyldimethylsilyl) oxy]ethyl)(phenyl)amino]-2-oxoethyl) benzoate. (iv) Compound 3 was then coupled with 2-[4-(methoxycarbonyl) phenyl]acetic acid, compound 8. (v) Further, the hydroxamic acid functional group was introduced to compound 4 by reacting with aqueous hydroxylamine and a catalytic amount of potassium cyanide to yield 4-(2-[(2-[(tert-butyldimethylsilyl)oxy]ethyl) (phenyl)amino]-2-oxoethyl)-N-hydroxybenzamide, compound 5. (vi) Finally, removal of the TBDMS group from compound 5 using 2% (vol/vol) HCl in ethanol resulted in compound 6, HPOB.

Fig. 1.

HPOB is a HDAC6-selective inhibitor. (A) Synthesis of a HDAC6-selective inhibitor, HPOB. (B) IC₅₀ values of HDAC6-selective inhibitors. IC₅₀ values of HPOB and tubacin. (C) IC₅₀ values of HPOB and SAHA for the 11 zinc-dependent HDAC enzymes. HPOB selectively inhibits HDAC6 in vitro compared with pan-HDAC inhibitors. The table shows the average values of two runs for each case, with differences between duplicate runs within less than 10%. P < 0.001. P value was derived from the Student’s one-tail t test.

HPOB Is a Selective Inhibitor of HDAC6.

To determine if HPOB is a selective inhibitor of HDAC6, it was assayed for inhibition of recombinant HDAC1 compared with HDAC6. HPOB has an IC₅₀ inhibitory activity for HDAC6 of 0.056 μM compared with HDAC1 of 2.9 μM (Fig. 1B). HPOB inhibitory activity against the 11 zinc-dependent HDACs ranged from 0.056 μM for HDAC6 to ≥1.7 μM for the other enzymes (Fig. 1C). By comparison, SAHA is a potent inhibitor of Class I HDACs and HDAC6 (Fig. 1C). Tubacin, a HDAC6-selective inhibitor discovered by Schreiber and coworkers (16) has an IC₅₀ inhibitory activity for HDAC6 of 0.045 μM and for HDAC1 of 0.193 μM (Fig. 1B). HPOB is a more potent inhibitor of HDAC6 than tubacin.

HPOB Inhibits Growth, However, Not Viability, of Normal or Transformed Cells.

We next determined the effect of HPOB on cell growth and viability of normal human foreskin fibroblast (HFS) and transformed (LNCaP, human prostate adenocarcinoma; A549, lung adenocarcinoma; and U87, glioblastoma) cells cultured with 8, 16, or 32 μM HPOB for up to 72 h. HPOB inhibited cell growth of normal and transformed cells in a concentration-dependent manner (Fig. 2A) but did not induce cell death of normal or transformed cells (Fig. 2B).

Fig. 2.

Effects of HPOB on cell growth and viability and acetylated patterns of proteins and histones in normal and transformed cells in culture. Normal (HFS) and transformed (LNCaP, A549, and U87) cells were cultured with indicated doses of HPOB for 72 h. Five micromolars SAHA is a positive control. (A) Cell growth. (B) Cell viability. Inhibition of cell growth of normal and transformed cells is concentration-dependent. Viable cells were evaluated by trypan blue staining. Data are represented as mean ± SD, P < 0.001. P value was derived from the two-way ANOVA. (C) HPOB causes accumulation of acetylated α-tubulin but not acetylated histone H3 in normal (HFS) and transformed (LNCaP, U87, and A549) cells. Cells were cultured with 5 µM SAHA, 4 µM tubacin, or 4, 8, and 16 µM compound HPOB for 24 h as indicated. SAHA, the pan-HDAC inhibitor, and tubacin, the HDAC6 relative inhibitor, were controls. Cell lysates were prepared for immunoblot analysis of acetylated α-tubulin (Acet-Tub), acetylated peroxiredoxin (Acet-PRX), and acetylated histone H3 (Acet-H3). GAPDH and total H3 are loading controls. (D) SAHA and tubacin induce accumulation of γH2AX, an early marker of DNA damage in LNCaP cells. HPOB did not induce accumulation of γH2AX in LNCaP cells. Cells were cultured with 5 µM SAHA, 4 µM tubacin, or 8 or 16 µM HPOB for 24 h. Immunoblots are of phosphorylated H2AX (γH2AX). Total H3 is the loading control.

HPOB Induces Acetylation of α-Tubulin, However, Not Histones, in Normal and Transformed Cells.

In normal (HFS) and transformed (LNCAP, U87, and A549) cells, HPOB causes accumulation of acetylated α-tubulin and acetylated peroxiredoxin, substrates of HDAC6 (4, 5), but not of acetylated histones (Fig. 2C). As previously reported (20, 22, 23), SAHA induced the accumulation of acetylated α-tubulin and histone H3, and tubacin induced accumulation of acetylated α-tubulin but not of histone H3 (Fig. 2C).

HPOB Does Not Induce Accumulation of γH2AX, an Early Indicator of DNA Double-Strand Breaks.

We previously found that the HDAC6-selective inhibitor, tubacin, causes accumulation of phosphorylated histone H2AX (γH2AX), an early indicator of DNA double-strand breaks (DSB), in transformed cells (23). Tubacin and SAHA induced the accumulation of γH2AX in LNCaP cells. HPOB did not induce detectable accumulation of γH2AX in LNCaP cells (Fig. 2D). These data suggest that HPOB, unlike tubacin and SAHA, does not cause DNA damage in transformed cells.

HPOB Does Not Inhibit Trehalose-Induced Autophagy.

We next determined if HPOB inhibited HDAC6 ubiquitin-binding complex formation induced by the complex carbohydrate, trehalose. It has been shown that HDAC6 has a role in ubiquitin-binding-complex–forming aggresomes in the autophagic pathway of cell death (8, 9, 24, 25). Trehalose induces aggresome formation and autophagy (26). Trehalose-induced inhibition of cell growth was not blocked by HPOB in normal and transformed cells (Fig. 3A). Consistent with these findings, HPOB was found not to alter trehalose-induced LC3-II (microtubule-associated protein 1A/1B-light chain 3-II) accumulation, an indication of autophagosome formation which is responsible for degradation of polyubiquitinated complexes (Fig. 3B).

Fig. 3.

HPOB does not block trehalose-induced autophagy and cell death in LNCaP. (A) Trehalose, an inducer of aggresome autophagy, inhibits cell growth in HFS and LNCaP cells and induces cell death in LNCaP. HPOB does not block trehalose-induced LNCaP cell death. Cells were cultured with 16 µM HPOB, 200 mM trehalose, or combination for 72 h. Cell growth and viability were determined by trypan blue staining. Data are represented as mean ± SD, P < 0.001. P value was derived from the two-way ANOVA. (B) Trehalose-induced accumulation of LC3 is not inhibited by HPOB in HFS and LNCAP cells. Cells were cultured with 8 or 16 µM HPOB, 200 mM trehalose, or combination for 24 h and were prepared for immunoblot analysis. Immunoblot is of LC3. GAPDH is the loading control. (C) HPOB, SAHA, or tubacin, inhibitors of HDAC6 catalytic activity, do not block polyubiquitin complex formation in HFS and LNCaP cells. Cells were cultured with 5 µM SAHA, 4 µM tubacin, or 8 or 16 µM compound HPOB for 24 h. Immunoblots are of polyubiquitin complex. GAPDH is the loading control. (D) Immunoprecipitation of HDAC6 brings down ubiquitin complex, which is not blocked by SAHA or HPOB. LNCaP cells were cultured with 5 µM SAHA or 16 µM HPOB for 24 h. Cell lysates were immunoprecipitated with HDAC6 or ubiquitin antibody and prepared for immunoblot analysis. Immunoblots are of HDAC6, polyubiquitin complex, and acetylated α-tubulin (Acet-Tub).

To further evaluate the effect of HPOB on formation of the HDAC6 ubiquitin-binding complex, we assayed polyubiquitin complex accumulation in normal and transformed cells cultured with SAHA, tubacin, or HPOB (Fig. 3C). There was no detectable difference in the accumulation of the polyubiquitin complex in normal or transformed cells cultured with SAHA, tubacin, or HPOB. These results indicate that HPOB may not inhibit HDAC6 activity which is important in proteasome inhibition.

To further evaluate whether ubiquitin-binding activity of HDAC6 is inhibited by HPOB, we performed an immunoprecipitation assay with HDAC6 antibody in LNCaP cells. HPOB causes an increase in the accumulation of acetylated α-tubulin, but does not block HDAC6 binding to the ubiquitin complex (Fig. 3D, Right). Immunoprecipitation assay with HDAC6 antibody showed that HDAC6 binds to mono- and di-ubiquitin complex (<75 kDa) rather than polyubiquitin complex. Immunoprecipitation assay with ubiquitin antibody showed that ubiquitin complex binds to HDAC6 protein. The binding of ubiquitin complex with acetylated tubulin did not change in LNCaP cells cultured with HDAC inhibitors compared with the control. These data indicate that HPOB inhibits the deacetylase activity of HDAC6 but not its ubiquitin-binding activity.

HPOB Enhances Transformed Cell Death Induced by the Anticancer Drugs Etoposide, Doxorubicin, or SAHA.

We previously reported that inhibition of HDAC6 by either si-RNA or tubacin potentiates the cytotoxicity of anticancer drugs in transformed but not in normal cells (23). To assess whether selective inhibition of HDAC6 by HPOB enhances cell death of normal and transformed cells in culture with anticancer agents, cells were cultured with HPOB and the topoisomerase II inhibitors etoposide or doxorubicin or the pan-HDAC inhibitor SAHA for 72 h. In HFS cells, HPOB alone or in combination with doxorubicin, etoposide, or SAHA inhibited cell growth but did not induce loss of cell viability (Fig. 4A).

Fig. 4.

HPOB enhances etoposide-, doxorubicin-, and SAHA-induced transformed cell death but not normal cell death. Cell growth and viability of (A) HFS, (B) LNCaP, (C) U87, and (D) A549 cells cultured with HPOB with 50 µM etoposide, 400 nM doxorubicin, or 5 µM SAHA alone and in combination with HPOB for 72 h. Cell growth and viability were determined as previously described. Data are represented as mean ± SD, P < 0.001. P value was derived from the two-way ANOVA. (E) LNCaP and HFS cells were cultured with HPOB in combination with etoposide for 24 h. Immunoblots show PARP degradation and acetylated α-tubulin. GAPDH is a loading control. (F) Immunoblots for phosphorylated H2AX (γH2AX) and acetylated histone H3 (Acet-H3). Total H3 is a loading control.

LNCaP cells cultured with 50 μM etoposide and 8 μM HPOB demonstrated inhibition in cell growth and loss of cell viability to a greater extent than LNCaP cells cultured with etoposide alone (Fig. 4B). LNCaP cells cultured with 400 nM doxorubicin and 8 μM HPOB had increased cell death compared with cultures with 400 nM doxorubicin alone. LNCaP cell death was enhanced in cultures with HPOB and 2.5 μM SAHA compared with cultures with SAHA alone (Fig. 4B).

In U87 cells, combination treatment with HPOB and 400 nM doxorubicin resulted in an increase in cell death compared with cultures with doxorubicin alone (Fig. 4C). There was no enhanced cell death in U87 cells cultured with 8 μM HPOB and 50 μM etoposide or SAHA. In A549 cells, HPOB in combination with 50 μM etoposide or 400 nM doxorubicin showed enhanced cell death compared with cultures with either drug alone (Fig. 4D).

HPOB Induces Increased Apoptotic Cell Death of Transformed Cells Cultured with Anticancer Drugs.

To investigate the pathway of cell death in transformed cells cultured with the combination of HPOB with etoposide, we determined the levels of poly(ADP ribose) polymerase (PARP) and its cleavage fragments. In LNCaP cells cultured with HPOB and etoposide, there was an increase in cleaved PARP, a marker of apoptosis (27) (Fig. 4E). There was no increase in accumulation of cleaved PARP in HFS cells cultured with HPOB alone and in combination with etoposide. These findings are consistent with HPOB enhancing etoposide-induced transformed cell death via the apoptotic pathway.

We next examined whether selective inhibition of HDAC6 with HPOB activates a DNA damage response in combination with anticancer drugs. Combination of HPOB with etoposide increased the accumulation of DNA damage compared with etoposide alone as evidenced by accumulation of γH2AX in LNCaP cells (Fig. 4F).

HPOB is Well-Tolerated in Mice and Enhances Cytotoxicity of an Anticancer Drug.

We next determined the toxicity of HPOB. HPOB was intraperitoneally injected daily for 5 d with 100, 200, or 300 mg/kg HPOB. There was no weight loss in these mice, suggesting that HPOB is well-tolerated in these animals (Fig. 5A). The effect of HPOB on the acetylation of α-tubulin and histones in the spleen isolated from mice treated with HPOB was analyzed at three time points after the administration of the drug. At 1.5 h after injection of HPOB, an increased accumulation of acetylated tubulin was found in the spleen (Fig. 5B). By 5 h after injection of HPOB, the accumulation of acetylated tubulin was reduced to the level seen in vehicle-treated controls. There was no detectable accumulation of acetylated histones in the spleen from the mice receiving HPOB. One hundred milligrams/kilogram SAHA increased the accumulation of acetylation of both α-tubulin and histones which persisted up to 3 h, but the increased acetylated levels were reduced to the level seen in vehicle-treated controls by 5 h. These data are consistent with previous reports (11).

Fig. 5.

HPOB enhances anticancer effects of SAHA in mice bearing human prostate cancer CWR22 xenograft. (A) HPOB is well-tolerated in animals. Mice were injected with indicated doses of HPOB and SAHA intraperitoneally daily for 5 d. Data are represented as mean ± SD, P < 0.001. P value was derived from the two-way ANOVA. (B) HDAC6 selectivity of HPOB in spleen isolated from immune-deficient mice. Spleens were isolated from mice injected with indicated drugs at 1.5, 3, and 5 h after the last injection on day 5. Immunoblots for acetylated α-tubulin (Acet-Tubulin) and acetylated histone H3 (Acet-H3). Hsp90 and total H3 are loading controls. (C) Mice injected with HPOB in combination with SAHA showed significant shrinkage of CWR22 tumors. There was no weight loss in the animals. Data are represented as mean ± SD, *P < 0.05, ***P < 0.001. P value was derived from the two-way ANOVA. (D) HDAC6 selectivity of HPOB in spleen, brain, and tumors isolated from mice bearing human prostate cancer CWR22 xenograft. Mice were injected with SAHA, HPOB, or combination of SAHA and HPOB intraperitoneally daily for 18 d. Tissues were isolated on day 25 and prepared for immunoblot analysis. Immunoblots are shown for acetylated α-tubulin (Acet-Tubulin), acetylated peroxiredoxin (Acet-PRX), and acetylated histone H3 (Acet-H3). Hsp90 and total H3 are loading controls.

Next, we examined the effects of HPOB in combination with an anticancer drug, SAHA, in nude mice with the androgen-dependent CWR22 human prostate cancer xenograft, which was grown s.c. Daily administration of either 300 mg/kg HPOB or 50 mg/kg SAHA alone for 18 d caused no significant suppression of the growth of established CWR22 tumors and no weight loss (Fig. 5C). Daily administration of HPOB and SAHA caused suppression of the growth of established CWR22 tumors, such that doses of 300 mg/kg/d HPOB in combination with SAHA 50 mg/kg/d caused reductions of 50% in the mean final tumor volume compared with vehicle-treated control animals. Tumors, spleen, and brain were removed from the animals, and histones and proteins were extracted for the detection of acetylated lysine patterns. There was increased accumulation of acetylated α-tubulin in CWR22 tumors and spleen from mice treated with HPOB, SAHA, or combination of HPOB and SAHA. In the brains of mice treated with HPOB, there was increased accumulation of acetylation of PRX1, a substrate of HDAC6 (Fig. 5D). Increased levels of accumulation of histones were found in tumors of mice injected with SAHA or a combination of SAHA and HPOB, but not with HPOB alone. These data indicate that HPOB is a selective inhibitor against HDAC6 in vivo, and HPOB can enhance the antitumor effect of chemotherapeutic agents.

Discussion

We report the discovery of a HDAC6-selective inhibitor, HPOB, and its biological effects in normal and transformed cells. HPOB inhibits HDAC6 in vitro with ∼50-fold selectivity against HDAC6 over HDAC1 enzyme. Concentrations as high as 16 µM of HPOB induce accumulation of acetylated α-tubulin and acetylated PRX, substrates of HDAC6, but not of acetylated histones, not a substrate of HDAC6, in both normal and transformed cells. HPOB in concentrations ≤16 µM does not induce normal cell death. HPOB enhances etoposide, doxorubicin, or SAHA-induced transformed cell death. These findings (12, 25, 28) provide evidence that selective inhibition of HDAC6 in combination with anticancer drugs may be an important avenue to enhance the therapeutic efficacy of such drugs in treating human cancers.

HPOB selectively inhibits the catalytic activity of HDAC6 but does not block HDAC6 binding to form a polyubiquitinated protein complex. The levels of LC3-II, a marker of autophagosome formation, do not change in cells cultured with HPOB. Combination of HPOB and trehalose, an inducer of autophagy, causes cell growth inhibition but not cell death of normal cells.

HPOB does not induce cell death in normal or transformed cells. Culture with HPOB in transformed cells enhances the cytotoxicity of DNA-damaging anticancer drugs through increased induction of apoptosis and accumulation of DNA damage.

HPOB is well-tolerated in animals. HPOB in combination with SAHA significantly enhances the antitumor effect of SAHA against the androgen-dependent CWR22 human prostate cancer xenograft in nude mice.

In summary, we have discovered a HDAC6-selective inhibitor, HPOB, that has the potential to enhance anticancer drug efficacy in combination therapy of human cancers, suggesting the promise of drugs targeting HDAC6 to improve therapeutic strategies in cancers.

Experimental Procedures

The section discussing materials and methods is included in SI Experimental Procedures. This section describes preparation of cells, reagents, proteins, and histone extracts used in this study. Assay procedures for determination of in vitro enzymatic assay for histone deacetylases and animal experiments are also detailed in SI Experimental Procedures. Animal studies were carried out under protocol 12-02-003, approved by the Memorial Sloan–Kettering Cancer Center Institutional Animal Care and Use Committee. Institutional guidelines for the proper, humane use of animals in research were followed.