SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Disclosure Statement
  7. References

Histone deacetylase inhibitors are a group of recently developed compounds that modulate cell growth and survival. We evaluated the effects of the histone deacetylase inhibitor MGCD0103 on growth of pancreatic carcinoma models following single agent treatment and in combination with gemcitabine. MGCD0103 inhibited tumor cell growth and acted synergistically with gemcitabine to enhance its cytotoxic effects. Gene expression analysis identified the cell cycle pathway as one of the most highly modulated gene groups. Our data suggest that MGCD0103 + gemcitabine might be an effective treatment for gemcitabine-refractory pancreatic cancer. (Cancer Sci 2011; 102: 1201–1207)

Pancreatic cancer (PC) is a devastating and universally fatal disease, with a 5-year survival rate of 5%. Despite recent efforts to improve early diagnosis, staging, and surgical and chemotherapeutic management of PC, survival remains poor.(1) This is due to the non-detectable development of tumors, aggressive phenotype, rapid and severe patient debilitation, and radio- and chemoresistance to many drugs. Although systemic chemotherapy with gemcitabine has led to modest clinical benefits in patients, the combined use of other cytotoxics with gemcitabine has failed to improve survival.(2) This underlies the need to explore new targets and molecular events in order to treat PC patients in a multimodal and multitargeted manner.

Histones are small proteins that complex with DNA to form the nucleosome core and chromatin that package the human genome.(3) Histone acetylation is a posttranslational modification that allows for “relaxation” of chromatin structure, subsequent accessibility of transcription factors to DNA and transcriptional activation. Deacetylation of histone tails by histone deacetylase (HDAC) induces transcriptional inactivation through chromatin condensation.(4) There are 18 known human HDACs classified into four groups based on phylogenetic and functional criteria. MGCD0103 inhibits the class I HDACs (1–3 and 8) that typically associate with multiprotein repressor complexes and are thought to play a role in cell survival and proliferation.(5)

Alterations in HDACs are found in many human cancers including pancreatic adenocarcinoma.(5) Abnormally high levels of HDAC1 in up to 40% of cancer cells and their surrounding “normal” tissues was observed(6) and high HDAC1 expression correlates with advanced stage lung and pancreatic cancer.(7,8) In many tumor types including PC, aberrant transcriptional silencing of tumor suppressor genes contributes to transformation and malignant growth. Reactivation of these epigenetically silenced genes by HDAC inhibitors (HDACi) has resulted in antitumor effects such as apoptosis, cell cycle arrest, differentiation, and senescence.(9) Histone deacetylases have also been shown to induce re-expression of a number of proteins such as the cell cycle inhibitor p21, estrogen receptor, and pro-apoptosis proteins such as caspase-3, PARP, and BAX.(5)

Conventional chemotherapeutic drugs target both normal and tumor cells, but HDACi appear to be cytotoxic to malignant cells, sparing the normal cells.(10) Because of their diverse antitumor effects and largely manageable side-effects, HDACi are candidates for use in combination with cytotoxics, targeted therapy, or other epigenetic modifiers. For example, entinostat (MS-275) and vorinostat (SAHA) have been shown to chemosensitize breast and pancreatic tumor cells to adriamycin and gemcitabine, respectively.(9,11,12) Currently, there are two HDACi that are clinically approved: vorinostat and romidepsin for the treatment of cutaneous T-cell lymphoma.

MGCD0103, an isotype-selective HDACi, has been clinically evaluated for the treatment of hematologic malignancies and advanced solid tumors, alone and in combination with standard-of-care agents. MGCD0103, a compound with favorable pharmacokinetic/pharmacodynamic profiles, showed mechanism-based antileukemia activity in a recent phase I trial and was also deemed tolerable in a subsequent advanced solid tumor trial.(13,14) MGCD0103 was evaluated for activity in combination with gemcitabine in refractory solid malignancies and in locally advanced or metastatic pancreatic adenocarcinoma. Findings from this clinical study suggest that the combination might have clinical activity in patients with refractory pancreatic cancer.(15) To better understand the potential mechanism of MGCD0103 in pancreatic cancer, we evaluated the ability of MGCD0103 to synergize with gemcitabine in pancreatic tumor models.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Disclosure Statement
  7. References

Cell lines and explants.  Pancreatic carcinoma cell lines, AsPC-1, BxPC-3, MiaPaca-2, and Panc-1 were purchased from American Type Culture Collection (Manassas, VA, USA) and maintained at 37°C, 5% CO2 in cell culture media supplemented with 10% FBS (RPMI [AsPC-1 and BxPC-3] and DMEM [MiaPaca-2 and Panc-1]). The cell lines vary in terms of differentiation state/grade and K-ras mutation status, and although most of the cell lines originated from primary tumors, AsPC-1 was isolated from ascites (see Table 1).

Table 1.   MGCD0103 and gemcitabine EC50 values
Cell lineSiteK-ras statusEC50 (μM)
MGCD0103Gemcitabine
  1. †Oncotest primary tumor explants. Values are from one single experiment representative of two additional experiments. 1°, primary tumor; Asc, ascites; Mut, mutant; WT, wild type.

BxPC-3WT1.12.6
AsPC-1AscMut3.912.9
MiaPaca-2Mut0.60.2
Panc-1Mut1.88.0
PAXF 546L†AscMut1.50.1
PAXF 1657L†Mut0.30.1

Oncotest pancreatic carcinoma explants PAXF 546 and PAXF 1657 were derived from K-ras-mutant (G12V) patient tumors. PAXF 546 was isolated from a well-differentiated peritoneal adenocarcinoma and PAXF 1657 was isolated from a moderately well-differentiated lung metastasis. These models differ from the cell lines as they were generated from primary tumors that were implanted and passaged in immune-compromised mice, thereby retaining most of the histological and molecular characteristics of the original patient tumor. Cultured cell lines were established from both explants and designated PAXF 546L and PAXF 1657L, respectively. Authenticity of cell lines and corresponding tumor xenografts was proven by short tandem repeat analysis.

Drugs and reagents.  MGCD0103 was provided by Methylgene (Montreal, Quebec, Canada) and gemcitabine (Gemzar) purchased from Eli Lilly (Indianapolis, IN, USA). The p21(12D1) rabbit mAb, produced by immunizing animals with synthetic peptide corresponding to residues near the carboxy terminus of human p21, was purchased from Cell Signaling Technology (Danvers, MA, USA). Furthermore, this antibody has been validated and does not cross-react with other cyclin-dependent kinase (CDK) inhibitors. A secondary antibody (affinity purified, HRP-conjugated goat anti-rabbit IgG) was also purchased from Cell Signaling Technology and used in Western blotting procedures.

Cell viability assay.  Cell viability was measured using the CellTiter-Glo Luminescent assay (Promega, Madison, WI, USA). An optimal seeding density of 5000 cells was plated in triplicate in 96-well opaque microtiter plates and after a 24-h incubation, five concentrations of drug were added to the cells for 72 h and luminescence measured using a SpectraMax M5 (0.5 s/well integration time; Molecular Devices, Sunnyvale, CA, USA). All viability assays were repeated at least twice and growth inhibition expressed as treated/control × 100 (% viability). Relative EC50 values were determined using the software GraphPad Prism, Prism 5 for Windows, version 5.01 (GraphPad Software, La Jolla, CA, USA).

Monolayer assay (Oncotest).  A modified propidium iodide assay was used, with 4000–6000 cells plated in 96-well flat bottom microtiter plates. After 24 h, test compounds were added to the cell cultures. MGCD0103 and gemcitabine were tested alone at 10 concentrations in triplicate to determine the individual EC50 values of each test compound. Following 72 h of continuous test compound exposure, cell culture medium with or without drug was replaced by 200 μL aqueous propidium iodide solution (7 μg/mL). Fluorescence was measured using the CytoFluor 4000 (http://www.mtxlsi.com/Cytofluor4000.htm) microplate reader (excitation λ = 530 nm, emission λ = 620 nm) and growth inhibition expressed as treated/control × 100 (%T/C). Relative EC50 values were determined as above.

Clonogenic assay (Oncotest).  Solid human tumor xenografts growing s.c. in athymic nude mice were removed, then mechanically and enzymatically disaggregated prior to passing through sieves of 200 and 50 μm mesh size.

The clonogenic assay was carried out in a modified two-layer soft agar assay in which 0.8 × 104–5 × 104 cells were plated in 24-well dishes. Cells were continuously exposed to six concentrations of test compounds and corresponding controls. Cultures were incubated at 37°C and 7.5% CO2 in a humidified atmosphere for up to 20 days and monitored closely for colony growth using an inverted microscope.

Drug combination studies.  The concentration–effect analysis of drug combinations was carried out according to Chou–Talalay using CalcuSyn (Biosoft, Cambridge, UK) and potential synergy or antagonism determined at a constant, equipotent ratio of the two drugs. A combination index (CI) was calculated based on doses that have equivalent effects and is a measure of the degree of interaction between the two drugs: CI = 1 indicates additivity; CI < 1 indicates synergy; and CI > 1 indicates antagonism. Three different drug combination schedules were tested: (i) concurrent addition of MGCD0103 and gemcitabine for 72 h; (ii) MGCD0103 pretreatment for 24 h followed by gemcitabine for 48 h; and (iii) pretreatment with gemcitabine for 24 h followed by MGCD0103 for 48 h. Cell viability assays were run as described above. Tests were carried out in duplicate or triplicate followed by up to five independent reproductions.

Immunoblot analysis.  Cells were trypsinized and resuspended in RIPA buffer (Cell Signaling Technology) supplemented with protease inhibitors (Complete Protease Inhibitor Cocktail Tablets; Roche, Nutley, NJ, USA). Protein concentrations were determined using the BCA protein estimation assay (Thermo Scientific, Rockford, IL, USA).

Protein was separated on 4–20% SDS–polyacrylamide mini gels and blotted onto PVDF membranes (Bio-Rad Laboratories, Hercules, CA, USA). Filters were blocked in TBS containing 5% non-fat milk and Tween-20 and incubated in 5% BSA with primary antibody (1:1000 dilution) overnight at 4°C. After washing with TBS-Tween, filters were incubated in 5% milk/TBS-Tween plus secondary antibody (1:10 000 dilution) for 1 h at room temperature. Blots were washed and proteins detected using a SuperSignal chemiluminescence detection kit (Pierce, Rockford, IL, USA).

Gene expression analysis.  PAXF1657L cells (4.4 × 105) were seeded on day 0 and treated with 0.1, 0.5, and 1,0 μM MGCD0103 for 24 h on day 1. Cells were washed with PBS and lysed with RNeasy RLT buffer (Qiagen, Valencia, CA, USA) with 10% beta-mercaptoethanol added. Following QIAshredder homogenization, total RNA was isolated using Qiagen’s RNeasy mini kit protocol with DNase treatment. Using the MessageAmp Premier RNA amplification kit (Ambion, Austin, TX, USA) double-stranded cDNA and biotin-labeled cRNA was synthesized from 100 ng total RNA. Thereafter, 10 μg cRNA was fragmented and hybridized to each human U133A 2.0 GeneChip (Affymetrix, Santa Clara, CA, USA). The GC-RMA algorithm was used for analysis and all analyses were done using GeneSpring 7.3 (Agilent, Santa Clara, CA, USA). Averaged signals from biological duplicate samples were used to determine fold change (treated versus untreated), with absolute fold change of ≥2.0 defining regulated genes. NextBio (Cupertino, CA, USA) was used to identify regulated biogroups (based on the Gene Ontology consortium) from lists of regulated genes.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Disclosure Statement
  7. References

MGCD0103 decreases viability of pancreatic carcinoma cells and explants.  In order to evaluate the effect of MGCD0103 on cell viability, Oncotest proprietary tumor explant-derived models PAXF 546L and PAXF 1657L and four human ductal pancreatic carcinoma cell lines were treated with MGCD0103 for 72 h. PAXF 546L and PAXF 1657L were both responsive to MGCD0103 with EC50 values of 1.5 and 0.3 μM, respectively (Fig. 1A, Table 1). MGCD0103 also inhibited growth of pancreatic tumor cell lines. Extent of inhibition varied approximately 10-fold across models, with EC50 levels ranging from 0.3 to 4.0 μM (Fig. 1B, Table 1). PAXF1657L and MiaPaca-2 were the most sensitive models and near complete inhibition of cell viability was observed at the two highest drug concentrations. In contrast, AsPC-1, Panc-1, and PAXF 546L were only partially growth inhibited despite treatment with up to 5 μM MGCD0103. K-ras status in these models did not appear to affect response to MGCD0103.

image

Figure 1.  Histone deacetylase inhibitor MGCD0103 dose-dependently inhibits growth of explant-derived cell lines PAXF 546L and PAXF 1657L (A) and pancreatic carcinoma cell lines AsPC-1, BxPC-3, MiaPaca-2, and Panc-1 (B) (72 h). Cells were grown in monolayer and viability measured using CellTiter-Glo reagent or propidium iodide.

Download figure to PowerPoint

Inhibition of anchorage-independent growth by MGCD0103.  In addition to the studies described above, we also investigated the ability of MGCD0103 to inhibit pancreatic tumor cell colony formation in a semi-solid medium. This assay evaluates anchorage-independent growth, a widely accepted hallmark of cellular transformation and metastasis that occurs when malignant cells gain the ability to proliferate independently of attachment to a substratum.(16,17) This assay is thought to better predict in vivo tumor growth and response to drug treatment than the monolayer assay in which cell growth and survival are artificially dependent upon attachment to plastic.

Disaggregated tumor samples prepared from the two Oncotest patient-derived human pancreatic cancer explants, PAXF 546 and PAXF 1657, were treated with increasing doses (0–10 μM) of MGCD0103 for 20 days. Concentration-dependent inhibition was observed and tumor colony formation was inhibited by approximately 90% in both tumor models at a dose of 1.0 μM MGCD0103 (Fig. 2). EC50 values were also calculated and were in the submicromolar range. These EC50 values were approximately 10-fold lower than those obtained from cell viability assays in monolayer format, consistent with the notion that growing tumor cells anchorage-independently reduces substrate adhesion-related survival signals and increases sensitivity to drug.(18,19)

image

Figure 2.  Histone deacetylase inhibitor MGCD0103 dose-dependently inhibits growth of pancreatic carcinoma primary tumor-derived explants PAXF 1657 and PAXF 546 (20 days). Clonogenic assays were carried out according to a two-layer soft agar assay supplemented with 20% FCS and colonies quantified using an automatic image analysis system. %T/C = treated cells/control cells × 100.

Download figure to PowerPoint

MGCD0103 synergizes with gemcitabine to enhance cytotoxicity.  Prior to carrying out combination studies, pancreatic tumor cell lines were treated with gemcitabine in order to evaluate sensitivity to the chemotherapeutic agent and to determine appropriate concentrations to use in the combination experiments. As shown in Figure 3(A), two of the cell lines, AsPC-1 and Panc-1, were resistant to gemcitabine, as confirmed by the EC50 values 12.9 and 8.0 μM, respectively (Table 1). The other two cell lines, BxPC-3 and MiaPaca-2, were somewhat more sensitive to gemcitabine (EC50 levels shown in Table 1); however, the maximum growth inhibition, approximately 50%, was not surpassed even following treatment with high drug concentrations. This plateau effect was also observed in the primary tumor explant-derived cell lines PAXF 546L and PAXF 1657L. Although they responded to gemcitabine (Fig. 3B, Table 1), growth inhibition beyond 50% was unattainable despite increasing drug concentrations. These findings are consistent with previously published work(20,21) and reflect the drug-resistant nature of pancreatic carcinoma.

image

Figure 3.  Gemcitabine has limited growth inhibitory effects in pancreatic carcinoma models. Cells were grown in monolayer, treated with gemcitabine for 72 h, and their viability measured using CellTiter-Glo. Maximal growth inhibition by gemcitabine was approximately 50% (A). Inhibition of explant-derived cells by gemcitabine was slightly higher (B).

Download figure to PowerPoint

Next, the tumor cell lines were evaluated for MGCD0103 and gemcitabine combination effects. Concentrations of MGCD0103 and gemcitabine were chosen based on doses needed to reach single agent EC50 plus dilutions above and below this value. Three dosing schedules were evaluated: concurrent treatment; pretreatment with MGCD0103; and pretreatment with gemcitabine (24 h pretreatment). Figure 4 shows the combination dose–effect curves for the concurrent schedule and Table 2 summarizes EC50 CI values for all three schedules in the four pancreatic tumor cell lines. When treated concurrently, MGCD0103 strongly synergized with gemcitabine in the cell lines AsPC-1 and Panc-1 (CI = 0.204 and 0.102, respectively) and in the case of BxPC-3, combination treatment resulted in almost complete inhibition of cell survival at the higher drug concentrations (CI = 0.394). A smaller effect was observed in MiaPaca-2, which was previously shown to be the most gemcitabine-sensitive cell line (CI = 0.654). The CI values also confirmed that MGCD0103 augmented the cytotoxic effects of gemcitabine in all three treatment schedules (CI < 1) and, furthermore, that these effects were observed at sub-EC50 (submicromolar) doses of MGCD0103 (Table 2). Interestingly, slightly stronger synergy was observed when the cell lines were either concurrently treated or pretreated with gemcitabine, suggesting that the continuous presence of the cytotoxic agent is important in order to achieve an optimal drug combination effect in these tumor cell lines.

image

Figure 4.  Combining histone deacetylase inhibitor MGCD0103 and gemcitabine augments the effect of single agent treatment in pancreatic cancer cell lines and explant-derived models. Cells were grown in monolayer and viability measured using CellTiter-Glo. This figure shows the growth inhibition of each cell line following concurrent drug treatment (72 h).

Download figure to PowerPoint

Table 2.   Combination index (CI) values for MGCD0103 + gemcitabine combination
 CI†CI‡Effect
  1. †Combination index value at EC50 MGCD0103 dose. ‡Combination index value at sub-EC50 MGCD0103 dose (0.25-fold of EC50). NE, not evaluated.

Concurrent treatment
 AsPC-10.2040.163Synergy
 BxPC-30.3940.131Synergy
 MiaPaca-20.6540.586Synergy
 Panc-10.1020.074Synergy
 PAXF 1657L0.190NESynergy
 PAXF 546L0.349NESynergy
MGCD0103 pretreatment
 AsPC-10.5490.663Synergy
 BxPC-30.6860.483Synergy
 MiaPaca-20.4780.810Synergy
 Panc-10.6070.443Synergy
Gemcitabine pretreatment
 AsPC-10.0940.043Synergy
 BxPC-30.1080.053Synergy
 MiaPaca-20.3200.598Synergy
 Panc-10.2080.223Synergy

Combination effects of gemcitabine and MGCD0103 were also evaluated in the explant models PAXF 546L and PAXF 1657L. Results following concurrent treatment with MGCD0103 and gemcitabine are summarized in Figure 4 and Table 2. Although both PAXF 546L and PAXF 1657L were more sensitive to gemcitabine than most of the cultured cell lines, combination effects with MGCD0103 were nevertheless observed and CI values were below 0.5, suggesting synergistic effects. PAXF 546L was almost maximally inhibited by the drug combination.

Finally, in a preliminary experiment, Panc-1 xenografts were treated with a combination of MGCD0103 and gemcitabine to evaluate whether our observed effects would translate to an in vivo setting. Although tumor volume was not significantly different between various treatment groups, xenograft growth was delayed in the animals treated with both MGCD0103 and gemcitabine versus animals receiving either of the single agents (data not shown).

Cell cycle genes induced by MGCD0103.  The MGCD0103 mechanism of action was explored by gene expression profiling in one of the pancreatic tumor explant models (PAXF 1657L). PAXF1657L cells were treated with a dose range of MGCD0103 (0.1–1.0 μM) for 24 h. All drug concentrations induced gene regulation, with the highest concentration, 1.0 μM, regulating the greatest number of genes (2936 genes regulated at least 1.7-fold or higher). As the MGCD0103 EC50 for this model was 0.3 μM, we focused on analysis of the genes modulated following treatment with 0.5 μM drug. A list of these 1673 genes was analyzed using NextBio to identify Gene Ontology biogroups that were regulated. Among the top five most significantly regulated biogroups, the cell cycle group had the highest number of modulated genes (Table 3). A complete list of all the genes found in this biogroup is shown in Table 4. Two of the most significantly regulated genes are cell cycle inhibitors p21 (CDKN1A) and p15 (CDKN2B), both upregulated by MGCD0103 approximately threefold. To begin to evaluate MGCD0103 effects on p21 expression at the protein level, we treated four pancreatic cancer cell lines with drug and assessed p21 protein expression 24 h after treatment. As shown in Figure 5, p21 was induced by MGCD0103 in all four cell lines. This finding is consistent with previous studies showing that the p21 gene is often silenced in tumor cells but can be re-expressed following treatment with HDAC inhibitors.(5)

Table 3.   Top five most significant biogroups regulated by 0.5 μm MGCD0103 in PAXF 1657L pancreatic tumor model
BiogroupNo. of genes (up + down)P-value
  1. Down, downregulated; up, upregulated.

Cell cycle68 (33 + 35)2.40E-09
Response to external stimulus65 (56 + 9)2.20E-09
Cytoskeletal protein binding53 (43 + 10)9.70E-11
Response to wounding44 (37 + 7)2.50E-09
Cell junction39 (33 + 6)2.90E-09
Table 4.   MGCD0103 modulates cell cycle genes in pancreatic tumor model PAXF1657L
GeneFold changeGeneFold change
  1. Cells were treated with 0.5 μM MGCD0103 for 24 h.

RASSF25.2SMARCB1−1.7
SEPT.63.7MPHOSPH9−1.7
CDKN2B3.4AKAP8−1.7
PDGFD3.1CDC25A−1.7
CDKN1A3.1AXL−1.8
RASSF52.9PNN−1.8
IL82.8E2F8−1.8
MLF12.8CDKN1C−1.8
RECK2.6CCNB1−1.8
CCNYL12.5ASPM−1.8
BMP22.5NF2−1.9
LLGL22.3CUL5−1.9
BCL62.3HK2−1.9
SESN32.3SUV39H1−1.9
DUSP12.2CLIP1−1.9
DUSP62.2ILF3−1.9
UBE2I2.0CDKN1B−1.9
HPGD2.0TPD52L1−1.9
AHR2.0NFYC−2.0
GADD45A2.0ZMYND11−2.0
MAPRE31.9GAS1−2.0
MN11.9PLK1−2.1
RHOB1.8STAG1−2.1
EVI51.8NBN−2.1
JUN1.8GTSE1−2.1
PTP4A11.8MCM5−2.2
SEPT.81.8MCRS1−2.2
MAPRE21.8TLK1−2.3
JAG21.8HCFC1−2.3
CYLD1.7PBRM1−2.5
MYC1.7CDKN2C−2.7
MAD2L21.7STAG2−2.8
RCBTB11.7ATF5−2.9
HEXIM2−1.7FOXC1−5.7
image

Figure 5.  Regulation of p21 protein expression by histone deacetylase inhibitor MGCD0103. Cell lines were plated in monolayer and treated with drug for 24 h.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Disclosure Statement
  7. References

MGCD0103 is an orally available, isotype-specific HDACi that is being evaluated as a single agent and in combination with chemotherapy for treatment of both hematologic and solid malignancies. In vitro, MGCD0103 has potent antiproliferative activity against human tumor cell lines but not against normal cells at submicromolar to micromolar concentrations.(22–25) In pharmacological models, MGCD0103 showed significant antitumor activity at well-tolerated doses in a variety of tumor types including colon, non-small-cell lung, pancreatic, and prostate cancer xenografts. Notably, drug was detected in the plasma up to 24 h post-dose and for the first 8 h, plasma concentrations were at or above the EC50 values needed to inhibit HDACs 1 and 2 in vitro.(22,24)

MGCD0103 has also been evaluated in several clinical trials. Pharmacokinetic profile was favorable with dose-dependent exposure and a half-life of approximately 10 h regardless of dosing schedule.(13,14,23,26) MGCD0103 as a single agent displays antitumor activity in hematological malignancies as well as in lymphoproliferative disease.(13,27,28) However, single agent effects in solid tumor patients have been less straightforward. In a phase I study of patients with advanced solid tumors who were treated with up to 45 mg/m2 (equivalent to a fixed dose of 90 mg) MGCD0103 three times per week, a Cmax of approximately 172 ng/mL (0.43 μM) was reached. Drug was tolerated and the most frequently observed side-effects were fatigue and gastrointestinal -related symptoms. Stable disease, but no objective tumor response, was observed. In a second, as yet unpublished phase I/II clinical study, patients with refractory and/or advanced solid tumors were treated with ascending doses (fixed dose of 50–110 mg; 90 mg was the maximum tolerated dose and recommended dose) of MGCD0103 in combination with gemcitabine. Dose-limiting toxicities included fatigue, diarrhea, vomiting, and nausea; all were expected with such a combination treatment. Interestingly, a study update reported two partial responses out of five pancreatic carcinoma patients and two additional partial responses in patients with nasopharygeal carcinoma and cutaneous T-cell lymphoma. Although this is only preliminary data from a very small number of patients, it is promising and suggests that solid tumor efficacy may be reached by combining MGCD0103 with gemcitabine.(15)

In the current study, we evaluated the in vitro effects of the HDACi, MGCD0103, on human pancreatic carcinoma cell lines and explant growth. Micromolar concentrations of MGCD0103 inhibited cell proliferation in a dose-dependent manner, both on plastic as well as in a 3-D clonogenic assay and the monolayer EC50 values are consistent with those determined in previous in vitro studies.(22,24,25) However, most of the pancreatic cell lines and explants were not completely inhibited by MGCD0103 at concentrations comparable to levels attainable in clinical dosing regimens. Therefore, we next evaluated a dose range of MGCD0103 (including clinically achievable, submicromolar concentrations) in combination with gemcitabine in the tumor cell lines at three different treatment schedules (concurrent treatment, MGCD0103 pretreatment, and gemcitabine pretreatment) and found that, in all cases, MGCD0103 synergized with gemcitabine to induce cytotoxicity. Our drug combination data indicates that clinically attainable concentrations of MGCD0103 could be effective in synergizing with gemcitabine and that sensitization by MGCD0103 may be independent of treatment sequence, potentially allowing for more flexible dosing schedules in the clinic.

Resistance to the cytotoxic effect of gemcitabine could be associated with multiple mechanisms including alteration of apoptosis genes, changes in drug transport and cellular turnover, as well as decreased expression or sensitivity of drug targets.(29–32) If agents such as HDACi can enhance gemcitabine-associated cytotoxicity pathways or activate novel pathways leading to cell death, gemcitabine refractory tumors could be rendered responsive to treatment. It has been previously shown that SAHA and trichostatin A, two HDACi, induce the upregulation of cyclin-dependent kinase inhibitor p21,(33,34) facilitating apoptosis of pancreatic carcinoma cell lines. Both of these inhibitors also potentiate the effects of gemcitabine on growth inhibition. MGCD0103 has been shown to induce cell cycle arrest (G1 and G2/M) in various human cancer cell lines(22) and in the current study, we propose that one of the mechanisms by which MGCD0103 could induce pancreatic carcinoma cell growth arrest is through upregulation of p21 expression. Unlike gemcitabine, which primarily leads cells into G1 arrest, p21 is thought to promote arrest in both the G1 and G2/M phase of the cell cycle.(35–38) Therefore, as we observed in our experiments, treatment of tumor cells with an HDACi such as MGCD0103 is expected to enhance growth inhibition compared to treatment with gemcitabine alone. To further elucidate potential mechanisms of synergy between MGCD0103 and gemcitabine, additional studies should focus on evaluating apoptotic/autophagic proteins, markers of antitumor immunity, and cell differentiation in MGCD0103-treated pancreatic cells to identify unique pathways that are stimulated by MGCD0103 and which might lead to synergy with gemcitabine. Indeed, preliminary data showed that combined treatment with MGCD0103 plus gemcitabine in the Panc-1 pancreatic cancer cell line synergistically induced apoptosis (Nguyen H., Li Z., Martell R. E., Besterman J. M., 2005, unpublished data).

Chemotherapeutic treatment options for advanced and metastatic pancreatic cancer are currently limited and an efficient and well-tolerated option is urgently needed. Most of the single chemotherapy agents used to treat pancreatic cancer yield low response rates and only a minimal impact on survival. Although gemcitabine is currently the standard-of-care for advanced disease, it appears to improve the quality of life for a short time only and clinical efficacy is poor. Our study showed that MGCD0103 has inhibitory effects on pancreatic cancer cell growth both as a single agent and in combination with the standard-of-care gemcitabine, and that cell cycle inhibition is one potential pathway through which MGCD0103 acts. Although additional studies are necessary in order to better understand its multiple mechanisms of action, these data suggest that MGCD0103 may be an effective treatment for pancreatic cancer patients, even those who have previously been shown to be refractory to gemcitabine.

Disclosure Statement

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Disclosure Statement
  7. References

Contributing authors are full-time employees of either Celgene or Oncotest.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Disclosure Statement
  7. References