Pharmacodynamic Response and Inhibition of Growth of Human Tumor Xenografts by the Novel Histone Deacetylase Inhibitor PXD101

Tumor cells are characterized by aberrant gene expression. Although this is often the result of genetic change (1), it also occurs as a consequence of changes in the epigenetic regulation of gene expression (2). In eukaryotic cells, levels of histone acetylation play a pivotal role in chromatin remodeling and the regulation of gene expression. Acetylation levels are thought to result from an equilibrium between the activities of histone acetyltransferase and histone deacetylase enzymes (3). In general, deacetylated histones are associated with inactive nontranscribed DNA (4). Aberrant epigenetic transcriptional repression has been demonstrated in a wide variety of tumor types and involves genes in multiple cellular processes involved in tumorigenic transformation and maintenance of the transformed state (5). Consequently, attention has focused on the development of inhibitors of histone deacetylase enzymes as a novel approach to cancer treatment, and several compounds are now undergoing evaluation in Phase I clinical trials (6).

Generically, HDAC inhibitors have been described as consisting of three parts: (a) a zinc-chelating group; (b) a spacer group, which is generally hydrophobic; and (c) an “enzyme binding” group that confers specificity and is generally aromatic in character (7, 8). All three moieties have been explored in our studies. Starting from the structures of natural product inhibitors, we have designed novel, low molecular weight, achiral, and synthetically accessible inhibitors of histone deacetylase activity. It is known that low molecular weight hydroxamic acid containing molecules, such as trichostatin A, oxamflatin, and SAHA,² are potent histone deacetylase inhibitors (9–11), and we have concentrated on the hydroxamic acid moiety as the zinc-chelating group because this group confers high levels of potency. PXD101 is the lead compound from this series (12), and here we describe the potent antitumor activity of PXD101 both in vitro against a number of human tumor cell lines and in vivo in human tumor xenografts.

The human ovarian cell line A2780 and cisplatin (A2780/cp70) and doxorubicin (2780AD) resistant derivatives were originally obtained from Dr. R. F. Ozols (Fox Chase Cancer Center, Philadelphia, PA). They were grown in RPMI 1640 supplemented with glutamine (2 mm) and FCS (10%). The human colon (HCT116 and HT29), melanoma (HS852), prostate (PC3), lung (CALU-3), and breast (MCF7) cell lines were obtained from American Type Culture Collection (Rockville, MD). MCF7 was grown in RPMI 1640 and the rest in DMEM supplemented as above. The human non-small cell lung cancer cell line WIL was originally obtained from the Ludwig Institute for Cancer Research (Sutton, United Kingdom) and was grown in DMEM supplemented as above.

Drug sensitivity was determined by a clonogenic assay (13). Briefly, cells were plated in 5 ml of medium at a density of 8 × 10⁴ cells/25 cm² flask and allowed to attach and grow for 48 h. Cells were exposed to drug (five concentrations from 0.016 to 10 μm) for 24 h. The medium was removed, and 1 ml of trypsin/EDTA was added to each flask. Once the cells had detached, 1 ml of medium was added, the cells were resuspended, and those from the control untreated flask were counted. Cells were diluted and plated into 6-cm Petri dishes (three per flask) at a density of 500-2000 cells/dish depending on the cell line. Cells from the drug-treated flasks were diluted and plated as for the control flasks. Dishes were incubated for 10–15 days at 37°C. Cells were washed with PBS, fixed in methanol, and stained with crystal violet, and colonies that contained ≥50 cells counted. Sensitivity is expressed as the IC₅₀ (mean ± SE of three experiments) defined as the concentration of drug required to reduce the number of colonies to 50% of that of the control untreated cells.

Subconfluent cultures were harvested and washed twice in ice cold PBS and pelleted by centrifugation at 200 × g for 5 min. The cell pellet was resuspended in two volumes of lysis buffer [60 mm Tris buffer (pH 7.4) containing 30% glycerol and 450 mm NaCl] and lysed by three freeze (dry ice) thaw (30°C water bath) cycles. Cell debris was removed by centrifugation at 12,000 × g for 5 min, and the supernatant was stored at −80°C.

Histone H4 peptide (sequence SGRGKGGKGLGKGGAKRHRK corresponding to the 20 NH₂-terminal residues) was acetylated by a recombinant protein containing the hypoxanthine-aminopterin-thymidine domain of p300, using [³H]acetyl CoA as a source of acetate. H4 peptide (100 μg) was mixed with hypoxanthine-aminopterin-thymidine buffer (50 mm Tris HCl pH 8.0, 5% glycerol, 50 mm KCl, and 0.1 mm EDTA), 1 mm DTT, 1 mm 4-(2-aminoethyl) benzenesulfonylfluoride, 1 × complete protease inhibitors (Amersham Pharmacia Biotech), 50 μl of purified p300, and 1.85 m [³H]acetyl CoA (Amersham Pharmacia Biotech; 4.50Ci/mmol) in a final volume of 300 μl and incubated at 30°C for 45 min. The p300 protein was removed by incubation with 20 μl of 50% Ni-agaroase beads (Qiagen) for 1 h at 4°C and centrifugation. The supernatant was applied to a 2-ml Sephadex G15 column, and the flow through was collected. One milliliter of distilled H₂O was gently applied, and three drop fractions were collected; this was repeated until 4–5 ml of distilled H₂O had been added, and ∼40 fractions were collected. Three microliters of each fraction were diluted in 2 ml of scintillation fluid and counted in a scintillation counter to identify the fractions containing the labeled peptide. These fractions were pooled, and 1 μl of the combined sample was measured to assess the radioactivity in every peptide batch (3000–7000 cpm/μl).

For activity assays, the reaction was carried out in a total volume of 150 μl of buffer [60 mm Tris (pH 7.4) containing 30% glycerol] containing 2 μl of cell extract and, where used, 2 μl of PXD101. The reaction was started by the addition of 2 μl of [³H]labeled substrate (acetylated histone H4 peptide corresponding to the 20 NH₂-terminal residues). Samples were incubated at 37°C for 45 min, and the reaction stopped by the addition of HCl and acetic acid (0.72 and 0.12 m final concentrations, respectively). Released [³H]acetate was extracted into 750 μl of ethyl acetate, and samples were centrifuged at 12,000 × g for 5 min. The upper phase (600 μl) was transferred to 3 ml of scintillation fluid and counted.

Cells were plated out 48 h before drug treatment in 75-cm² flasks such that cultures were 75% confluent on the day of treatment. They were exposed to drug for various times in 25 ml of growth medium. Cells were trypsinized, resuspended in medium, and centrifuged, and the pellet was washed in ice cold PBS. Histone proteins were isolated as described by Yoshida (9). Briefly, cells were resuspended in 1 ml of lysis buffer [10 mm Tris (pH 6.5), 50 mm sodium bisulfite, 10 mm MgCl₂, 8.6% sucrose, and 1% Triton X-100], homogenized, and centrifuged at 1000 × g for 10 min at 4°C. The pellet was washed three times in lysis buffer and then once in 10 mm Tris pH 7.4 containing 13 mm EDTA. The pellet was resuspended in 100 μl of ice cold water, and H₂SO₄ was added to give a final concentration of 0.4N. After incubation for ≥1 h on ice, samples were centrifuged at 12,000 × g for 5 min at 4°C in an Eppendorf centrifuge. The supernatant was transferred to a tube containing 1 ml of acetone, incubated overnight at −20°C, and then centrifuged as before. The pellet was air dried and then resuspended in 50 μl of water.

Tissue samples were homogenized in ice cold lysis buffer and processed as above. For blood samples, red cells were lysed by the addition of three volumes of cell lysis solution (Promega). After incubation for 10 min at room temperature, samples were centrifuged at 2000 × g for 10 min. Histones were isolated from the white cell pellet as described above.

Proteins (5 μg) were separated by the NuPage electrophoresis system (Invitrogen) on 10% Bis-Tris gels with MES SDS running buffer. The “Novex Xcell II” blotting apparatus (Invitrogen) was used to transfer proteins onto Immobilon polyvinylidene difluoride membrane (Millipore). The membrane was blocked for 1 h in Tris-buffered saline containing 0.02% Tween 20 and 5% powdered milk and then incubated overnight at 4°C with the primary antibody [antiacetylated histone H3 (Lys 9 and 14) or H4 (Lys 5, 8, 12, and 16) from Upstate Biotechnologies]. The membrane was then washed and incubated for 1 h at room temperature with the secondary antibody (goat antirabbit horseradish peroxidase; Amersham). After washing, protein bands were visualized by enhanced chemiluminescence (ECL, Amersham). All samples were analyzed on two separate gels to exclude loading variations.

Cells were plated at a density of 10⁵ cells in a 25-cm² flask and allowed to attach and grow for 48 h. Drug was added at a range of concentrations for 24 h. Both adherent cells and those in the medium were harvested and washed twice with ice cold PBS. They were resuspended in 200 μl of lysis buffer [50 mm HEPES (pH 7.0), 250 mm NaCl, and 0.5% NP-40] supplemented with protease inhibitors (Complete from Roche Diagnostics Ltd., Lewes, United Kingdom) and incubated on ice for 20 min. Samples were centrifuged at 12,000 × g for 5 min at 4°C to remove debris.

Proteins were separated on 4–12% Bis-Tris gels with 4-morpholinepropanesulfonic acid SDS running buffer and processed as above. The primary antibody was either anti-PARP (PharMingen from BD Biosciences) or anti-Cip1/WAF1 (Transduction Laboratories from BD Biosciences), and the secondary antibody was sheep antimouse horseradish peroxidase (Amersham).

Animal studies were carried out under an appropriate United Kingdom Home Office Project License, and all work was conformed to the UKCCR guidelines for the welfare of animals in experimental neoplasia.

For the human tumor xenograft studies, monolayer cultures were harvested with trypsin/EDTA (0.25%/1 mm in PBS) and resuspended in PBS. About 10⁷ cells were injected s.c. into the right flank of athymic nude mice (CD1 nu/nu mice from Charles River). After 10–15 days when the mean tumor diameter was ≥0.5 cm, animals were randomized into groups of six for experiments. PXD101 was dissolved in DMSO and then diluted in water to give a final concentration of DMSO of 10% and was administered i.p. at the times specified. This formulation gave sufficient solubility for doses of ≤40 mg/kg. Mice were weighed daily, and tumor volumes were estimated by caliper measurements assuming spherical geometry (volume = d³ × π/6). Mice were also observed daily for any changes in behavior or condition according to the Morton and Griffiths scoring system (14). Significant differences between groups were identified by ANOVA, and the significance level of individual differences was determined by Student’s t test.

The canine pharmacokinetic studies were carried out by Covance Laboratories, Ltd. (Harrogate, United Kingdom). Beagle dogs received an i.v. infusion of PXD101 given at a rate of 2.5 ml/min over 45 min. The drug was formulated in 0.1 m Tris buffer (pH 8.5) containing 10% volume for volume ethanol and 5% polyethylene glycol 200. Blood samples (n = 10) were taken at various times for measurement of PXD101, and a second 3-ml sample was taken for estimation of histone acetylation. Histones were extracted immediately as described above, the pellet was resuspended in water, and samples were stored at −70°C and transferred to the Department of Medical Oncology, Glasgow, in dry ice for analysis by Western blot.

PXD101 was quantified in plasma by liquid chromatography coupled with tandem mass spectrometric detection by Covance Laboratories. Samples were mixed with water:methanol:formic acid (90:10:1) and the internal standard (oxamflatin) and applied to a Waters Oasis MAX 30-mg cartridge using automated solid phase extraction (Gilson 215). Elution was performed with methanol:water:formic acid (90:10:2), and the extract dried and reconstituted in 200 μl of water:acetonitile:acetic acid (65:35:1) for injection. Samples were analyzed on a Micromass Quattro LC with a Luna 3 μm C18(2) 50 × 3-mm column. The mobile phase was water:acetonitile:acetic acid (65:35:1), the flow rate was 0.5 ml/min, and the retention time of PXD101 was 1.8 min. Results were acquired and quantified using MassLynx software (versions 3.3 and 3.1).

Within the core sulfonamide template exemplified by PXD101, a number of important structure activity relationship points have been elucidated (Fig. 1). Modifications to the hydroxamic acid, a key zinc-chelating moiety, lead to large decreases in activity. Acid and amide analogues are inactive (e.g., Compound 1), and substitution on the hydroxamic acid nitrogen or oxygen also leads to inactive compounds (e.g., Compound 2). In the central region of the molecule, replacing the cinnamic acid moiety with an alkyl chain leads to a reduction in activity (e.g., Compound 3), and substitution of the sulfonamide nitrogen also reduces activity (e.g., Compound 4). The position of substitution on the cinnamic acid phenol group is important; ortho substitution leads to inactive compounds (e.g., Compound 5). There is greater flexibility in the “enzyme binding group” region of the molecule, where substitutions are tolerated, and larger groups can be introduced (e.g., Compound 6). On the basis of its potent antiproliferative and histone deacetylase inhibitor activity, PXD101 was selected as the lead compound for additional studies.

Histone acetylation was determined with antibodies specific for acetylated histones H3 and H4 by Western blot of histones isolated from A2780 ovarian tumor cells treated for various times with PXD101 (Fig. 2a). Markedly increased histone acetylation was observed after drug exposure (1 μm) for 30 min, and acetylation of both histone H3 and H4 was maintained for 36 h with continuous drug exposure (Fig. 2a). After removal of PXD101, acetylation was maintained for 30 min but markedly reduced after incubation in drug-free medium for 1 h and had returned to basal levels by 3 h (Fig. 2b). There was a clear concentration-dependent increase in levels of acetylated histone H4 in A2780 cells treated for 3 h with PXD101 (Fig. 1c). Histone H4 acetylation was also observed in a cisplatin (A2780/cp70) and doxorubicin (2780AD)-resistant variant of A2780, although only at the higher drug concentrations.

PXD101 inhibited the growth of a number of human tumor cell lines in vitro with IC₅₀s determined by a clonogenic assay in the range 0.2–3.4 μm (Table 1). There was no effect on colony size. Sensitivity to PXD101 is not related to the total HDAC activity of the cell line or inhibition of this activity in the cell lysate. There was no correlation between sensitivity to PXD101 and sensitivity to a DNA-damaging agent, such as cisplatin (r² = 0.01). The cisplatin- (A2780/cp70) and doxorubicin (2780AD, p-glycoprotein positive)-resistant derivatives of the human ovarian tumor cell line A2780 showed low fold cross-resistance to PXD101.

PXD101 induced apoptosis as determined by measurement of PARP cleavage after drug incubation for 24 h (Fig. 3a). PARP cleavage was detected in all cell lines examined except for PC3 and 2780AD (Fig. 3a). Interestingly, these two cell lines are not the most resistant to PXD101, and the lines did not show PARP cleavage when incubated with the DNA-damaging agent cisplatin (50 μm; data not shown). The colon tumor cell line HCT116 was markedly sensitive to induction of PARP cleavage, which was observed after drug incubation for between 18 and 24 h and at a PXD101 concentration as low as 0.16 μm (Fig. 3b). Acetylation of histones H3 and H4 was also clearly apparent after incubation for 1 h with PXD101 at these concentrations.

PXD101 induced expression of the cyclin-dependent kinase inhibitor p21^(Cip1/WAF1) in HCT116 colon tumor cells (Fig. 3b). The time course for expression was similar to that observed for PARP cleavage, but induction of p21^(Cip1/WAF1) was observed at lower concentrations of PXD101. Induction of p21^(Cip1/WAF1) was observed in all cell lines included in Table 1 (data not shown).

Tumor-bearing mice were treated i.p. with PXD101 once daily for 7 days. Significant (P < 0.01) growth delay was observed at a dose of 10 mg/kg/day in xenografts of A2780 (Fig. 4a) and in the cisplatin derivative (A2780/cp70; Fig. 4,c). For both tumors, growth delay increased with increasing dose of PXD101 up to a dose of 40 mg/kg/day. Drug treatment had no effect on the body weight of the mice (Fig. 4b), and there were no apparent signs of toxicity to the mice. Growth inhibition was also observed in xenografts of the human colon tumor cell line HCT116 (Fig. 4d).

Acetylated histone H4 was detected in peripheral blood mononuclear cells at 1 and 2 h after a single i.p. injection of PXD101 (40 mg/kg) to A2780 tumor-bearing mice and had returned to baseline levels by 3 h (Fig. 5a). This effect of PXD101 on histone acetylation was dose dependent with marked acetylation apparent at doses of ≥10 mg/kg in both peripheral blood mononuclear cells (blood) and tumor (Fig. 5b). Plasma drug concentrations were determined at 0.5 and 2 h after a single i.p. injection of PXD101 (20 mg/kg) in mice. At 0.5 h, the mean drug concentration was 3.3 ± 0.7 μm (n = 3), and it had decreased to 0.042 ± 0.002 μm by 2 h.

Acetylation of histones H3 and H4 was also determined in peripheral blood mononuclear cells in canine blood taken at various times after an i.v. infusion of PXD101 given over 45 min (total dose 50 mg/kg). At the end of the infusion, the plasma concentration of PXD101 was between 20 and 30 μm, and the plasma half-life was estimated as ∼40 min. Blood samples were taken pretreatment (t = 0) and at various times (t = 10–240 min) after the end of the infusion. Acetylation was quantified by Western blotting of extracted histones. A marked increase in histone acetylation was observed at 10 min (Fig. 6). For H4, acetylation was maximal after 60 min and was still elevated at 240 min (a). For H3, acetylation was maximal at 10 min and returned to basal levels by 4 h (b).

The histone deacetylase inhibitor PXD101 inhibits growth and induces apoptosis in human tumor cell lines in vitro. It inhibits growth of human tumor xenografts in vivo at concentrations that are without apparent toxicity to the mice. Growth inhibition both in vitro and in vivo is associated with a marked increase in the level of acetylation of histone proteins.

PXD101 inhibited histone deacetylase activity in cell lysates with an IC₅₀ in the range 9–100 nm (Table 1). The potency is slightly less than that seen for trichostatin A (Ki 3.4; Ref. 9) but comparable with that reported for other hydroxamate type inhibitors, such as SAHA (IC₅₀ 10 nm for HDAC1 and 20 nm for HDAC3; Ref. 15) and oxamflatin (IC₅₀ 16 nm; Ref. 10). PXD101 was clearly able to inhibit the deacetylase activity in whole cells as evidenced by the appearance of acetylated histones H3 and H4 within 30 min of drug addition (Fig. 2a). Enzyme inhibition was stable in the presence of drug and maintained for ≥36 h. However, acetylation was reversible as seen by a marked reduction in the level of acetylated histones 1 h after drug removal (Fig. 2b).

PXD101 inhibited the growth of a range of human tumor cell lines in vitro, including ovarian, non-small cell lung, breast, and prostate (Table 1). Sensitivity to PXD101 was not related either to the total histone deacetylase activity of the cell line or the ability of the drug to inhibit the enzyme activity in cell extracts (Table 1). These are crude estimates because they do not distinguish between the different classes of histone deacetylases, but there is no evidence that the histone deacetylase inhibitors currently available exhibit any selectivity between the enzyme classes (16). We do not know the substrate specificity if any of PXD101. There was no correlation between sensitivity to PXD101 and the DNA-damaging agent cisplatin (Table 1), which suggests that in the cell lines examined, cellular factors affecting PXD101 sensitivity are distinct from those affecting cisplatin sensitivity.

The basis of the selective antitumor activity of histone deacetylase inhibitors is not clear. The effects of histone deacetylase inhibitors on gene expression appear to be specific, and trichostatin A has been shown to affect transcription of only ∼7% of genes in CACO-2 colon tumor cells (17). Histone deacetylase inhibitors have been reported to induce growth arrest, differentiation, and apoptosis (7), and the effects may depend on the cell line model used (18). Cytotoxic concentrations of PXD101 induced apoptosis as determined by measurement of PARP cleavage in the majority of the cell lines examined (Fig. 3a). Clearly, this is not the only apoptotic pathway induced by PXD101 because PARP cleavage was not detected in PC3 prostate and 2780AD ovarian tumor cells, yet both cell lines were relatively sensitive to growth inhibition by PXD101. Furthermore, these two cell lines did not show PARP cleavage when treated with cytotoxic concentrations of cisplatin.

In addition to induction of apoptosis, treatment of cells with PXD101 resulted in increased expression of the cyclin-dependent kinase inhibitor p21^(Cip1/WAF1) (Fig. 3b). The time course of expression of p21^(Cip1/WAF1) and induction of apoptosis was similar with both apparent after incubation with drug for between 18 and 24 h (Fig. 3a). p21^(Cip1/WAF1) is a negative regulator of growth and thought to be involved in the regulation of a number of processes, including induction of differentiation (19). A number of histone deacetylase inhibitors have been reported to induce expression of p21^(Cip1/WAF1) and cell cycle arrest in G1, and this is consistent with the observation that they can induce differentiation in some cell lines (20–22). For HCT116 cells, histone acetylation, induction of p21, and PARP cleavage were apparent at around the IC₅₀ of PXD101 (Fig. 3b). However, for A2780, which is equally sensitive to PXD101, PARP cleavage was only apparent at a PXD101 concentration 10-fold higher than the IC₅₀ (Table 1 and Fig. 3 a). This suggested that growth inhibition of A2780 by PXD101 could be explained by cell cycle arrest rather than induction of apoptosis. However, we did not see any effect on the cell cycle distribution of A2780 cells (measured by fluorescence-activated cell sorter analysis of propidium iodide-stained cells after BrdUrd labeling) when treated with an IC₅₀ of PXD101 (0.2 μm; data not shown).

PXD101 showed antitumor activity in vivo in human tumor xenografts. For xenografts of the human ovarian cell line A2780, there was a clear dose response with activity at a dose of 10 mg/kg daily (Fig. 4a). Daily treatment with PXD101 at ≤40 mg/kg had no effect on body weight and was without any apparent toxicity to the mice (Fig. 4b). A similar lack of toxicity to mice has been reported for the histone deacetylase inhibitors SAHA (11) and MS-27–275 (23). Antitumor activity of PXD101 was also observed for the cisplatin-resistant derivative of A2780 (A2780/cp70; Fig. 4,c), which showed a 5-fold cross-resistance to PXD101 in vitro, and for xenografts of the colon cell line HCT116 (Fig. 4d). This is a potentially clinically relevant observation, because xenografts of A2780/cp70 and HCT116 are relatively resistant to current cytotoxic chemotherapeutic drugs, and both are resistant to the maximum tolerated dose of cisplatin (24). Treatment of tumor-bearing mice with PXD101 resulted in the appearance of acetylated histone H4 in both peripheral blood mononuclear cells and in the tumor (Fig. 5). Histone acetylation in peripheral blood mononuclear cells was apparent between 1 and 2 h after a single i.p. injection and was dose dependent within the range at which antitumor activity was observed. (Fig. 5). This is consistent with the observation that plasma concentrations of PXD101 (3 μm 30 min after treatment of mice with PXD101 20 mg/kg) were comparable with those required to detect histone acetylation in vitro.

In view of the lack of apparent toxicity associated with inhibition of histone deacetylase activity in mice, it is possible that optimal anticancer activity will not equate with the maximum tolerated dose in the clinic. Phase I trials of the histone deacetylase inhibitors phenylbutyrate (25) and depsipeptide (26) achieved plasma concentrations comparable with those required for activity in vitro in cell lines with minimal toxicity to patients. It is therefore important to identify a means of monitoring drug activity in patients, and our observation suggests that measurement of the levels of histone acetylation in peripheral blood mononuclear cells might provide such a measure. Inhibition of tumor growth is associated with a marked increase in histone acetylation in the tumor, and this is mirrored by increased acetylation of histones in peripheral blood mononuclear cells. A detailed study of the changes in acetylation of histones H3 and H4 in peripheral blood mononuclear cells in canine blood demonstrates that this is a reliable and reproducible measurement. Histone acetylation levels are not merely a reflection of the pharmacokinetics of the drug because the kinetics of the changes in acetylation of histones H3 and H4 differ and do not relate to changes in plasma drug concentrations. Although H3 acetylation appears to be maximal by the end of the infusion, H4 acetylation does not reach a maximum until 30–60 min later. In contrast, plasma concentrations are maximal at the end of infusion, followed by a rapid (half-life of ∼40 min) decrease in concentration.

In summary, we have shown that the novel histone deacetylase inhibitor PXD101 is active in vivo in human tumor xenografts with no obvious toxicity to the mice. Significantly, PXD101 shows activity in cisplatin-resistant human ovarian tumor xenografts. Our observations suggest that PXD101 has potential as a novel antitumor agent. Furthermore, measurement of histone acetylation in blood cells could provide a suitable pharmacodynamic end point to monitor drug activity in patients.