|Home | About | Journals | Submit | Contact Us | Français|
Enhanced proliferative signaling and loss of cell cycle regulation are essential for cancer progression. Increased mitogenic signaling through activation of the mTOR pathway, coupled with deregulation of the Cyclin D/retinoblastoma (Rb) pathway is a common feature of lymphoid malignancies, including plasmacytoma (PCT), multiple myeloma (MM), Burkitt's lymphoma (BL), and mantle cell lymphoma (MCL). Here we evaluate the synergy of pharmacologically affecting both of these critical pathways using the mTOR inhibitor sirolimus and the histone deacetylase inhibitor entinostat. A dose‐matrix screening approach found this combination to be highly active and synergistic in a panel of genetically diverse human MM cell lines. Synergy and activity was observed in mouse PCT and human BL and MCL cell lines tested in vitro, as well as in freshly isolated primary MM patient samples tested ex vivo. This combination had minimal effects on healthy donor cells and retained activity when tested in a co‐culture system simulating the protective interaction of cancer cells with the tumor microenvironment. Combining sirolimus with entinostat enhanced cell cycle arrest and apoptosis. At the molecular level, entinostat increased the expression of cell cycle negative regulators including CDKN1A (p21) and CDKN2A (p16), while the combination decreased critical growth and survival effectors including Cyclin D, BCL‐XL, BIRC5, and activated MAPK.
Cancer is a multigenic disease often arising from a complex interplay of environmental factors with primary disease alleles, modifier loci encoding efficiency alleles, and somatic events which themselves may be affected by genetic background. As such, it is not particularly surprising that single agents with high target specificity ultimately fail to control cancer growth and metastases. Drug combinations designed to target multiple pathways are more likely to control this complex disease process, and may help ameliorate drug resistance mechanisms so common with single agent use (Dancey and Chen, 2006; Stewart, 2009; Zimmermann et al., 2007).
Two central pathways frequently dysregulated in cancer are the PI3K/Akt/mTOR/p53 (mTOR) and Cyclin/CDK/CDKI/Rb (CDK) pathways. mTOR and CDK pathway dysregulation is common in B cell neoplasias, including MCL (de Boer et al., 1995; Rizzatti et al., 2005), MM (Dilworth et al., 2000; Harvey and Lonial, 2007; Kuehl and Bergsagel, 2012; Peterson et al., 2009; Zhan et al., 2006), BL (Schmitz et al., 2012) and PCT where genetic predisposition is determined in part by alleles of Mtor and Cdkn2a (Bliskovsky et al., 2003, 1998, 2001).
mTOR pathway dysregulation can mechanistically involve mutations, activation by growth factor receptor pathways, PTEN loss, and amplification of AKT and DEPTOR (Guertin and Sabatini, 2007; Harvey and Lonial, 2007; Laplante and Sabatini, 2009; Meric‐Bernstam and Gonzalez‐Angulo, 2009; Peterson et al., 2009; Zhang et al., 2011; Zoncu et al., 2011). mTOR is a serine‐threonine kinase that forms two complexes, mTORC1 (mTOR, RAPTOR, PRAS40, mLST8, DEPTOR) and mTORC2 (mTOR, RICTOR, PROTOR, mLST8, SIN1, DEPTOR), which phosphorylate a number of downstream targets (most notably S6K1, 4EBP1,2, AKT, SGK1) to effect regulatory roles in transcription and translation, cell proliferation and survival, and immune response, metabolism, and autophagy (Guertin and Sabatini, 2007; Laplante and Sabatini, 2009; Meric‐Bernstam and Gonzalez‐Angulo, 2009; Zhang et al., 2011; Zoncu et al., 2011). Rapamycin (sirolimus) is a relatively specific inhibitor of mTORC1, but can also affect mTORC2 following prolonged exposure (Sarbassov et al., 2006). Clinical investigations using rapamycin or its analogs as single agents have shown only modest long‐term benefit despite initial antitumor activity in some patients (Dancey, 2010).
Similarly, CDK pathway dysregulation often involves tumor suppressor gene (Rb, cyclin‐dependent kinase inhibitors (CDKI) including p16, p21) loss or mutation, and Cyclin/cyclin dependent kinase (CDK) amplification (Fernandez et al., 2005; Malumbres and Barbacid, 2009). The benzamide entinostat (MS‐275) is predominantly a selective Class I HDAC inhibitor with many activities, one of which is the reactivation of tumor suppressor gene (CDKI) pathways, which can ultimately affect CDK activity and lead to apoptosis (Bantscheff et al., 2011). Entinostat has strong activity against HDAC1, weak activity for HDACs2 and 3, and no activity against HDAC8 (Bantscheff et al., 2011; Witt et al., 2009); it has relatively strong activity against HDAC9, which is a Class IIA histone deacetylase (Khan et al., 2008). Recent studies have shown that entinostat has low affinity for binding to HDAC1/2‐Sin3 complexes, and higher associations with HDAC3‐NCOR1 and HDAC1/2‐CoREST complexes (Bantscheff et al., 2011). Use of HDAC inhibitors (vorinostat and entinostat) in MM cell lines has shown decreased phospho‐Rb, decreased cyclin D1 and E2f1 expression, enhanced p53 activity, and increased p21 and p27 expression (Lee et al., 2010, 2004, 2003). Combining HDAC inhibitors with other targeted agents, radiation, or chemotherapeutics has shown efficacy in clinical trials for MM (Badros et al., 2009), and breast cancer (Huang et al., 2011), despite relatively modest benefit as single agents (Federico and Bagella, 2011; Gojo et al., 2007; Gore et al., 2008; Hess‐Stumpp et al., 2007; Kummar et al., 2007).
In this study, we have investigated the synergistic effects of combining mTOR and HDAC inhibition to limit the growth of a variety of mature B cell neoplasms in vitro and in vivo. Further, we have extended these findings into an examination of activity in primary patient cells, in a model system simulating the pro‐tumor effects of the microenvironment upon drug treatment, and in long‐term animal studies.
Human MM cell lines L363, U266, EJM, RPMI‐8226, FR‐4, JK‐6L, OCI‐MY5, LP‐1, MM‐1R, KMS‐11, KMS‐26, KMS‐20, and KMS‐28BM were cultured and authenticated as previously described (Gabrea et al., 2008). MOPC265 (IL6 dependent, p16(+)), and MOPC460(IL6 dependent, p16+ve, p53 partial deletion) cells were derived from pristane‐induced PCTs from BALB/c mice. RMDPC 107‐403 cells (p16(−), deleted) were cloned from a myc‐ras retroviral‐induced PCT from DBA/2 mice. MM cell lines were cultured in RPMI‐1640 (2 mM l‐glutamine, 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 μg/ml streptomycin: Invitrogen). Mouse cell lines MOPC265, MOPC460 and 107403 were cultured in RPMI‐1640 (+50 μM β‐mercaptoethanol and 10 ng/mlIL6), in a humidified incubator (5%CO2 at 37 °C). MCL lines SP53 and JEKO were kindly provided by Lou Staudt (Center for Cancer Research (CCR), National Cancer Institute (NCI), Bethesda, MD) and Burkitt's lymphoma lines CB‐32, Namalwa, BJAB, and DG‐75 were kindly provided by Giovanna Tosato (CCR, NCI, Bethesda, MD).
Bone marrow aspirate was collected from patients with newly diagnosed multiple myeloma enrolled in clinical trials at the NCI/NIH. All patients reviewed and signed informed consent forms prior to admission. Bone marrow mononuclear cells and primary MM cells were isolated using Ficoll‐Hypaque density gradient sedimentation from bone marrow aspirates according to manufacturer's protocol. CD138+ cells were further separated from bone marrow samples by antibody‐mediated positive selection using anti‐CD138 magnetic‐activated cell separation microbeads (Miltenyi Biotech, Cambridge, MA). The percentage of CD138+ cells in the positive fraction was quantified by flow cytometric analysis using FlowJo software and found to be greater than 98%.
For in vitro studies entinostat (MS‐275) was purchased from Sigma–Aldrich and sirolimus (rapamycin) was provided by the Drug Synthesis and Chemistry Branch (DSCB), Developmental Therapeutics Program (DTP), Division of Cancer Treatment and Diagnosis (DCTD), NCI, NIH). The drugs were dissolved in dimethylsulfoxide (DMSO; Sigma) at a concentration of 10 mM and stored at −80 °C.
Assessment of activity and synergy was performed with a dose matrix comprised of five single agent concentrations for each compound, and the 25 combinations thereof (sirolimus: 0.1–100 nM; entinostat: 125–2000 nM). MM cells were seeded in 96‐well plates at 50,000 cells per well in 200μ media with the matrix duplicated on each plate. Viability was assessed after 48 h of treatment with CellTiter Aqueous MTS reagent (Promega). Subsequent single agent and combination dose response curves were repeated with at least quadruplicate wells in each experiment. Cell viability graphs depict the mean of at least three experimental replicates with error bars showing standard error of the mean.
Two methods for evaluating drug synergy were applied: Excess over Highest Single Agent (EOHSA) and Combination Index (CI). EOHSA is a standard measure of synergy used by the FDA for evaluation of drug combinations and is calculated as the difference of the effect produced by the drug combination and the greatest effect produced by each of the combination's single agents at the same concentrations as when combined (Borisy et al., 2003). Combination Index (CI) derived by Chou and Talalay (Chou, 2010) from the mass‐action law principle allows quantifying drug interactions in terms of synergy (CI > 1), antagonism (CI < 1) and additivity (CI = 1) based on the median–effect equation. CI computations for the dose matrices were done using CompuSyn software (http://www.combosyn.com/). Heatmaps and CI plots for the dose matrices were generated with R version 2.15.1 (Team, 2011). The activity of the drug combination on MM cells in the presence of bone marrow stromal cells was determined using an L363 cell line generated to stably express a luciferase construct as previously described (McMillin et al., 2010) in quadruplicate wells, and the experiment was performed three times (representative experiment shown). The MM line was co‐cultured with the non‐labeled immortalized bone marrow stromal line HS‐5 (Roecklein and Torok‐Storb, 1995). After 48 h of drug treatment, luciferin substrate was added to the medium and luminescence measured as a read‐out of MM cell viability.
Cells (2 × 106) were cultured in 6‐well plates in 6 ml media/well with either DMSO or varying concentrations of single agents and in combination for 24 or 48 h. Cells were washed with phosphate buffered saline (PBS) and fixed with 70% ethanol overnight at −20 °C. Cells were washed with PBS and stained with propidium iodide/RNase staining buffer (Pharmingen™/BD Biosciences; San Jose, CA) for 30–40 min. Cells were analyzed using the CellQuestTMPro v.5.2.1 (BD Biosciences) on a flow cytometer (FACSCalibur, Becton Dickinson, San Jose, CA) and percentages of cells in sub‐G1, G0G1, S and G2M phases quantified using ModFitLT™3.1 (Verity Software House, Inc., Topsham, ME). Apoptosis was analyzed using the Annexin V‐PE/7AAD (7‐AminoactinomycinD) apoptosis detection kit I (Pharmingen™/BD Biosciences; San Jose, CA). Cells (2 × 106) were cultured in 6‐well plates as above. Collected cells were washed with cold PBS, resuspended in 1× binding buffer, stained with Annexin V‐PE/7‐AAD and analyzed using BD CellQuestTMPro or FACScan flow cytometry. These experiments were performed at least three times and a representative experiment is shown.
The quantification of mRNA (Trizol) for p16 and p21 was done by real‐time PCR (Taqman RT) using SYBR GREEN and the following primers:
|Hum p21 (F):Hum p21 (R):||5′‐AGGTGGACCTGGAGACTCTCAGG‐3′5′‐GCTTCCTCTTGGAGAAGATCAGCC‐3′|
|Hum p16 (F):Hum p16 (R):||5′‐TGCCCAACGCACCGA‐3′5′‐CGCTGCCCATCATCATGA‐3′|
|Hum β‐actin (F):Hum β‐actin(R):||5′‐CCTGGCACCCAGCACAAT‐3′5′‐GCCGATCCACACGGAGTACT‐3′|
|Mus p21 (F):Mus p21 (R):||5′‐CCAGACATTCAGAGCCACAGG‐3′5′‐GGTCGGACATCACCAGGATT‐3′|
|Mus p16 (F):Mus p16 (R):||5′‐CCCAACGCCCCGAACT‐3′5′‐GTGAACGTTGCCCATCATCA‐3′|
|Mus β‐actin (F):Mus β‐actin (R):||5′‐CTTCTTGGGTATGGAATCC‐3′5′‐GGCATAGAGGTCTTTACG‐3′|
Immunoblot analyses were performed on cells lysed with RIPA buffer, electrophoresed on 4–20% Tris‐Glycine SDS‐PAGE gels (Novex), and blotted to nitrocellulose using iBlot (Invitrogen). For western blots of lysates from xenografted tumors, lysates were made from tumors in each treatment group and then pooled for blotting. Each experiment was repeated at least three times, and a representative blot is shown in the figure. Antibodies for β‐actin, Acetyl‐Histone H3/H4, phospho‐S6 and S6‐ribosomal protein, mTOR, ERK1/2, pERK1/2, pAKT, Survivin, p21, p16, and cyclin D1,3 were obtained from Cell Signaling and used at 1:1000 dilutions.
For in vivo studies, sirolimus was provided by the NCI (DSCB/DTP/DCTD), and entinostat was generously provided by Syndax Pharmaceuticals and NCI, NIH. Athymic, NCr‐nu/nu mice (Frederick, MD) were tested using institutionally approved (LCBG‐009, ACUC, NCI) animal protocols. For visualization, MM cells were infected with pSicoLV‐luciferase‐green fluorescent protein fusion gene (Day et al., 2009). In the L363 experiment, 5 × 106 cells were inoculated onto each flank (2 tumors per mouse) and allowed to grow for 11 days prior to randomization into treatment groups (7 groups; 5 mice per group). For the long‐term treatment experiment, 8 × 106 U266 cells were inoculated onto each flank and allowed to grow for 3 weeks prior to randomization into treatment groups (4 groups; 4–5 mice/group). Growth of luc/GFP positive cells was measured weekly by bioluminescence using a XenogenIVIS®100 system. A 50 mg/ml stock solution of sirolimus was prepared in ethanol (stored at −20 °C) and on the day of injection, drug was diluted to final concentration in 5% Tween‐80, 5% polyethylene glycol‐400 (Sigma, St Louis, MO). Entinostat (MS‐275) suspension was made using 20% hydroxypropyl β‐cylodextrin (Sigma). Sirolimus and entinostat (200 μl of each) were administered daily, 5 days per week (wk) for 4 wks (L363) or 12 wks (U266) by i.p. injection and oral gavage, respectively.
Multiple myeloma is a genetically heterogeneous disease comprised of several molecular subtypes with prognostic and therapeutic characteristics (Broyl et al., 2010; Decaux et al., 2008; Zhan et al., 2006). The diverse genetic lesions contributing to myeloma pathogenesis include recurrent IgH translocations into several oncogenic loci, RAS mutations, MYC deregulation, p53 mutations, genetic or epigenetic loss of Rb pathway control, and aberrant activation of the NFκB pathway (Annunziata et al., 2007; Bergsagel and Kuehl, 2005; Fonseca et al., 2004). As drug response is intricately linked to the molecular events leading to tumor formation, thirteen MM cell lines with diverse genetic aberrations were selected for evaluation across a range of dose combinations. Included within this panel are lines from the various molecular subtypes, with RAS mutations (L363, EJM, 8226, and MM‐1R), MMSET translocations (KMS‐11, LP‐1), Cyclin D translocations (U266, FR4), as well as lines with MYC amplification and NFκB pathway activation (Annunziata et al., 2007; Decaux et al., 2008; Dib et al., 2008; Keats et al., 2007; Lombardi et al., 2007; Moreaux et al., 2011).
A 6 × 6 matrix combination dose response screen of sirolimus and entinostat at five individual concentrations, and all combinations thereof, was performed in 13 MM cell lines to assess activity and synergy across a dose spectrum. Heatmaps show the percent viable cells compared to vehicle control across the dose matrix after 48 h of drug exposure (Figure 1A and B). A surface plot of the excess inhibition over highest single agent (EOHSA) is also shown, with positive values (yellow) indicating the additional inhibition of viability elicited by adding the second drug. In a matrix dose response screen, plotting EOHSA across the matrix can provide a visual indicator of potential synergy by depicting the difference in viability between the highest single agent dose in the row or column and the viability of each combination with that dose. For example, in the L363 line (Figure 1A), all combination dose points inhibit viability more than either single agent. Synergy was analyzed across all 25 combination dose points by calculating the combination index (CI) based on the median effect principle as proposed by Chou and Talalay (Chou, 2006). Synergy was observed at a majority of dose points in all of the lines. Inhibition of at least 50% of cell viability was observed across the lines at concentrations considered pharmacologically achievable (i.e. 1–10 nM sirolimus and 500 nM entinostat). These dose combinations were also found to be synergistic in all lines (Figure 1A and B; blue circles and diamonds). The broad activity and synergy across dose points in a diverse array of molecular backgrounds are indicative of potential broad clinical application across MM molecular subtypes.
Matrix dose response testing for activity and synergy. A) Heatmap of cell viability relative to control for L363 (MM) across the dose matrix. B) Topological graph of excess over highest single agent (EOHSA) shows the amount of additional activity achieved ...
Pharmacologically relevant doses of sirolimus (1–10 nM)(Cloughesy et al., 2008) and entinostat (500 nM)(Kummar et al., 2007) were selected for follow‐up evaluation (Figure 2A). Sirolimus inhibited phosphorylation of ribosomal protein S6, a downstream target of mTOR (Figure 2B and C). The increased acetylation of histone H3/H4 lysine residues in the presence of entinostat is indicative of its activity on HDACs (Figure 2B and C). We also examined the efficacy of the combination on other B cell neoplasms (plasmacytomas, lymphomas), which frequently share similar genetic alterations in the CDK/Rb and PI3k/mTOR pathways (Giulino‐Roth et al., 2012; Perez‐Galan et al., 2011; Schmitz et al., 2012). The dose combination of 10 nM sirolimus and 500 nM entinostat was active in mouse PCT lines (Figure 2D), human MCL lines (Figure 2E), and human Burkitt's lymphoma lines (Figure 2F).
Combination treatment for 48 h with entinostat and sirolimus suppressed growth of A) MM cells, D) mouse PCT cells, E) MCL cells and F) BL cells. Cells were either untreated (black), treated with 0.5 μM of entinostat (green), 10 nM ...
The accessory cells of the tumor microenvironment have been shown to contribute to therapeutic resistance in many tumor types, including myeloma. To assess the effectiveness of this combination in the presence of bone marrow stromal cells, the cell‐specific bioluminescence (CS‐BLI) platform (McMillin et al., 2010) was utilized. MM cells stably expressing luciferase, were co‐cultured with non‐labeled immortalized bone marrow stromal cells (BMSCs) for 48 h. BMSCs showed some protection of tumor cells treated with single agents; however, this protective effect was partially overcome by the combination treatment. The drug combination decreased MM cell viability by 70% even in the presence of BMSCs (Figure 3A).
Combination activity in stromal co‐culture and ex vivo‐treated primary MM cells. A) Viability of luciferase‐tagged L363 cells co‐cultured with bone marrow stromal cells treated with single agents and the combination. ...
The combination treatment was largely non‐toxic for normal peripheral blood mononuclear cells from two individuals cultured ex vivo with the drugs for 48 h (Figure 3B). The large discrepancy between the cytotoxic effects of the combination on normal and tumor cells at drug concentrations that are achievable in patient plasma strongly suggests a favorable therapeutic window.
To ensure the efficacy of the combination was not limited to cultured cell lines, the combination effect of entinostat and sirolimus on viability was assessed in freshly isolated patient myeloma cells. CD138+ plasma cells were isolated from two newly diagnosed MM patients and cultured in human serum supplemented media ex vivo for 48 h in the presence of vehicle or the drug combination. By 48 h, the combination treatment decreased the viability of both samples by at least 50% (Figure 3C and D). The drug combination was more effective than either single agent in the MM cells from the patients, in contrast to the response seen in freshly isolated PBMCs, suggestive of a favorable therapeutic window for combination treatment.
Sirolimus is not known to induce substantial apoptosis when used as a single agent, while HDAC inhibition has been shown to induce apoptosis in a variety of cell types. Using Annexin V‐7AAD staining of MM line L363, increased apoptosis was seen with entinostat alone but not with sirolimus alone (Figure 4A). The addition of sirolimus to entinostat enhanced the amount of apoptosis induced by entinostat alone by ~30% (Figure 4A). The presence of cleaved PARP in the entinostat and combination treated cells (Figure 4B) is consistent with the Annexin V results. Similar effects on apoptosis were observed in the MCL line Jeko (see cleaved caspase levels in Figure 5F).
Apoptosis and survival pathways. A) The percent of L363 cells positive for apoptosis by Annexin V/PI staining after 48 h with 10 nM sirolimus, 500 nM entinostat, or the combination. B) Western blot of survival proteins ...
Cell cycle. Proportion of cells in G2/M, G0/G1, and S‐phase after 48 h of drug treatment. Human MM lines L363 (A), U266 (B); mouse plasmacytoma line 107403 (C); and human MCL line SP53 (D). Relative mRNA expression, normalized ...
The combination of entinostat and sirolimus was found to reduce the expression of critical tumor cell survival proteins BCL‐xL and Survivin (BIRC5) (Figure 4B). Further, engagement of the pro‐survival MAPK pathway is a frequent pathological feature of myeloma (Annunziata et al., 2011). As a read‐out of MAPK activation, p‐ERK1/2Thr202/Tyr204 was assessed in L363 4 h after drug treatment (Figure 4C). The combination had a cooperative effect in keeping p‐AKT levels from increasing (Figure 4D).
In addition to apoptosis (Figure 4) being induced by the combination, there was substantial inhibition of cell cycle progression (Figure 5). Cell cycle analysis was performed with the MM cell lines L363 and U266 (Figure 5A and B respectively), as well as with the mouse PCT line 107403 (Figure 5C) and the human MCL line SP53 (Figure 5D). The proportion of cells in S phase and mitosis was decreased by both individual drug treatments, and to a greater extent by combination treatment in all lines tested (Figure 5A–D). Sirolimus, entinostat, and the combination greatly increased the proportion of cells blocked in G0/G1 in all of the lines, indicative of cell cycle arrest. Entinostat alone increased mRNA expression of the cell cycle negative regulators CDKN1A (p21) and CDKN2A (p16) in both MCL and PCT lines (Figure 5E–G). CDKN1A and CDKN2A have both been shown to be inactivated in many Non‐Hodgkin lymphomas including BL (Klangby et al., 1998), MCL (Pinyol et al., 1998) and MM (Gonzalez‐Paz et al., 2007). The combination decreased the expression of Cyclin D3 and CDK4 below that of single agent use (Figure 5G). Since cyclin dependent kinases are involved in phosphorylating the Rb tumor suppressor to allow for cell cycle progression, we also observed decreased levels of pRb as early as 18 h after the combination treatment (Figure 5H). A western blot for phosphorylated Histone H3 (pHH3) was performed as an indicator of cells actively dividing and undergoing mitosis. Levels of pHH3 were decreased by 18 h of treatment, and were absent after 48 h (Figure 5H).
As an initial assessment of the in vivo activity of the sirolimus/entinostat combination, a four‐week drug treatment study was performed in nude mice bearing xenografts of the L363 MM line. Briefly, luciferase‐positive L363 cells were inoculated in the flanks of nude mice and allowed to grow for 11 days, at which point the luminescence was measured. Five mice per group were randomized to seven groups (control, entinostat alone (10 or 20 mg/kg), sirolimus alone (2.5 or 5 mg/kg), and two combination groups (10 or 20 mg/kg of entinostat plus 2.5 mg/kg sirolimus), and treatment was initiated (Day 0). Tumor growth was monitored by photon flux with a Xenogen Imager and was substantially lower in mice receiving combination treatment compared to vehicle controls and single agent treatment groups (Figure 6A: day 28). The control vehicle‐only group developed palpable tumors as early as 1 week after inoculation, with tumors at day 28 averaging 850 mm3 in volume and 1 g in weight. Palpable tumors developed in all of the mice in the single agent arms of the study, but no palpable tumors were present in either combination arm at study completion (Figure 6C).
In vivo activity. A) Xenogen imaging of L363 xenografts (n = 5 mice/group) at the end of 4 weeks of daily (Mon‐Fri) treatment. B) Time course of photon flux imaging (Xenogen) of L363 xenografts in vehicle, single agent, ...
To further test the utility of the combination, a second xenograft experiment was performed using the U266 MM cell line, which had a lower in vitro response than L363 (Figure 2A). To more accurately mimic progressive disease, xenografts were allowed to grow for three weeks into palpable tumors prior to randomization to treatment arms (4–5 mice/group). Euthanasia of the control arm was required by treatment week 4 because of tumor burden. In each of the single agent groups, tumor progression was delayed, but tumor outgrowth eventually occurred (Figure 6D). By contrast, the combination treatment prevented tumor growth for 3 months, with no or small tumors present at necropsy (Figure 6E). None of the mice in either experiment experienced treatment‐related illness or significant weight loss (not shown).
Tumor lysates from these in vivo experiments were analyzed for inhibition of mTOR and HDAC targets. The in vivo effects of the drugs on the xenografts were similar to their effects in vitro with respect to phospho‐S6. In contrast to in vitro experiments, rapamycin increased the amount of histone acetylation (AcH3) leading to enhanced AcH3 in the combination treated tumors (Figure 6F). Furthermore, expression of pro‐survival proteins such as cyclin D3, Survivin, and to some extent bcl‐xl, was decreased by the combination treatment (Figure 6F). p‐ERK1/2 (MAPK) levels were noticeably lower in the xenografts treated with the combination (Figure 6F).
Therapeutic strategies that target multiple arms of growth and survival pathways represent a rational approach for treating B cell neoplasias. The RB and mTOR signaling pathways have a prominent role in deregulated cell growth and proliferation in plasma cell dyscrasias and a number of other B cell malignancies. The prospect of integrating HDAC and mTOR inhibitors, each of which cause cell cycle arrest via different signaling events, is attractive. The combination of HDACi (entinostat/MS‐275, 0.1–0.5 μM) and mTORi (sirolimus, 1 nM–10 nM) studied here was effective at inhibiting growth of cell lines from mouse and human B cell malignancies. Cell lines treated with the combination had reduced cell proliferation, together with increased cell cycle arrest and apoptosis. Although the lines displayed varying degrees of sensitivity, the mTORi/HDACi combination treatment was greater than additive in its inhibition of the cell lines tested, with CI values well below 1. This synergy permits pharmacologically achievable doses of each drug (Cloughesy et al., 2008; Kummar et al., 2007), when given in combination. Notably, single agent and combination treatment, when tested at therapeutic doses, had minimal toxicity on normal human PBMC, as assessed by viability.
MM xenograft tumors responded in a dose‐dependent fashion to both drugs. When given as single agents in vivo, they produced cytostatic responses initially, but after 5–8 weeks, xenografts began to grow; this effect was particularly evident in tumors treated with rapamycin alone. By contrast, the combination treatment resulted in a cytoreductive effect for small tumors and a cytostatic effect for larger ones, as seen in the two in vivo experiments. Although bioluminescent imaging could visualize small foci of tumor cells in some of the mice given the combination treatment soon after tumor engraftment, tumor masses were not detectable upon dissection after 4 weeks of treatment. In the U266 delayed‐treatment trial, the combination resulted in consistently smaller tumors than single agent treatment in xenografts at 12 weeks.
The D type cyclins, which are growth promoting, and myc, which can affect both growth and apoptosis, are frequently dysregulated in MM and PCT (Gabrea et al., 2008). The ability to normalize or keep the levels of these proteins in proper balance may be a benefit of the combined treatment. As mTOR can upregulate the translation of proteins from genes with highly structured 5′ UTRs such as myc and the D type cyclins, the inhibition of mTOR by sirolimus presents a rational drug choice (Frost et al., 2004; Gera et al., 2004; Nishioka et al., 2008). Similarly, entinostat has been shown to upregulate levels of CDKIs (Gojo et al., 2007; Nishioka et al., 2008; Verheul et al., 2008), which, via their effects on CDK/cyclin complexes, regulate cell cycle progression. Our studies have also shown that entinostat and the combination treatment can increase or keep CDKI levels normal similar to the studies of entinostat treatment in prostate (Verheul et al., 2008) and AML (Gojo et al., 2007; Nishioka et al., 2008). Thus, the mTORi/HDACi combination may lower cell cycle progressors such as the cyclins and cyclin dependent kinases and increase cell cycle negative regulators to negatively regulate cell cycle progression via different routes.
As has been seen in other human cancers, rapamycin treatment of MM cells may lead to AKT (Shi et al., 2005) and MAPK (Carracedo et al., 2008) activation and enhanced tumor growth/survival through negative feedback loops stemming from S6K1 or IGF‐1 through RTK/IRS1/PI3K signaling. An important aspect of the mTORi/HDACi combination is that pERK1/2Thr202/Tyr204 and pAKTSer473 were not elevated by this treatment, and their levels appeared to be slightly lower in cells treated with the combination as compared to control cells, or those treated with rapamycin alone. This observation is consistent with lower pAKTSer473 levels seen in AML (Nishioka et al., 2008) and diffuse large B cell lymphoma (Gupta et al., 2009) cells treated with in combination with mTOR and HDAC inhibitors, and with suppression of autocrine production of IGF‐1 in MM cells treated with vorinostat (Mitsiades et al., 2004). It is therefore possible that the negative feedback loops (Carracedo et al., 2008; Shi et al., 2005; Wan et al., 2007) induced by rapamycin treatment leading to activation of MAPK and AKT, are attenuated by the combination treatment.
In summary, these studies provide an experimental rationale to consider the combination of mTORi/HDACi in the treatment of human MM and further support this option for BL (Dong et al., 2013) and MCL (Yazbeck et al., 2008). The combination altered RB and PI3K signaling to retain the beneficial effects of single mTOR and HDAC inhibitors, overcame the enhanced MAPK and AKT activation often associated with mTORi treatment (Carracedo et al., 2008; Shi et al., 2005), inhibited cell growth, and promoted apoptosis in association with increases in CDKI and with decreases in cyclins and the pro‐survival proteins, Bcl‐xl and Survivin.
This research was supported by the Intramural Research Program of the National Institutes of Health, National Cancer Institute, Center for Cancer Research. Comments and suggestions from Doug Lowy, Peter Blumberg, Lee Helman, Ke Zhang, Peter Ordentlich, Wyndham Wilson, Lou Staudt, Kevin Camphausen, Brigitte Widemann, Val Bliskovsky, Richard Robinson and the NCI JDC contributed significantly to the project and the manuscript. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.
Simmons John K., Patel Jyoti, Michalowski Aleksandra, Zhang Shuling, Wei Bih-Rong, Sullivan Patrick, Gamache Ben, Felsenstein Kenneth, Kuehl W. Michael, Simpson R. Mark, Zingone Adriana, Landgren Ola and Mock Beverly A., (2014), TORC1 and class I HDAC inhibitors synergize to suppress mature B cell neoplasms, Molecular Oncology, 8, doi: 10.1016/j.molonc.2013.11.007.