Background
Sarcomas are a diverse group of malignant mesenchymal neoplasms with over 50 histologically distinct subtypes [
1]. They can be broadly classified in two groups: those containing simple karyotypic defects including recurrent translocations or those with complex cytogenetic lesions characterised by multiple altered genes, such as leiomyosarcoma (LMS) [
1,
2]. LMS accounts for 11 % of soft tissue sarcomas (STS) and although its genetic basis is not fully characterized, common genetic abnormalities include loss of function mutations in p53 or PTEN and activating mutations in the PI3K/AKT/mTOR pathways [
3‐
6]. Conventional treatment of LMS often involves surgical resection, chemotherapy and/or radiation [
2]. Unfortunately, current 5-year disease specific survival for resectable, non-metastatic LMS is 60 % for retroperitoneal and 75 % in extremity patients [
7,
8]. Since the main patterns of failure are metastatic disease and multifocal local recurrence, effective chemotherapeutic options are essential to improve more durable disease control.
First-line LMS chemotherapy currently consists of doxorubicin (Dox), an anthracycline that inhibits topoisomerase II thereby disrupting DNA repair, in combination with ifosfamide [
9]. However, locally advanced, recurrent or metastatic uterine LMS shows only a 25 % response rate to Dox when administered at a dose of 60–80 mg/m
2 IV every 3 weeks [
10]. The use of Dox in cancer therapy is limited by cardiotoxicity, leucopenia, thrombocytopenia and the development of resistance [
10]. Another commonly used regimen is gemcitabine (900 mg/m
2 day 1 and 8) and docetaxel (100 mg/m
2 day 8), which has response rates ranging from 27 to 53 % in uterine and non-uterine leiomyosarcoma [
11]. Therefore, overall poor response rates and significant toxicity in LMS patients implores that more effective, less toxic selective therapies be developed to improve patient outcomes.
In an effort to increase efficacy of treatment, selective inhibitors are being widely developed to target tumor-specific molecular pathways. Recently, STS and other solid tumors including breast, lung and colon cancer have been characterized as exhibiting PI3K/AKT/mTOR dysregulation [
12‐
16]. The PI3K/AKT/mTOR pathways are activated by receptor tyrosine kinases (RTK) that transmit extracellular signals from the tumor microenvironment. These pathways can be dysregulated not only by hyperactivation of growth factor signalling but also through activating or loss of function mutations affecting key molecules such as AKT and PTEN. Targeting these pathways has become a rapidly expanding field in drug development as several small molecule inhibitors targeting these proteins have recently been studied in clinical trials (BKM120, BEZ235 and MK2206) [
17‐
20]. However, complex regulation of the PI3K/mTOR pathways includes feedback loops allowing targeted proteins to be circumvented. Mounting evidence of cross-talk/redundant functions between pathways MAPK and ERK have been linked to the development of adaptive resistance when targeting the PI3K/mTOR pathways [
21,
22]. Thus, it has been suggested that small molecules targeting multiple pathways or combining these agents with conventional cytotoxic agents may thwart the development of resistance and result in more durable drug therapy [
23].
To identify novel selective therapies for LMS that are effective and potentially less toxic, we utilized a selective drug library of 480 small molecule kinase inhibitors. Based on our initial screen with validation studies in vitro, we report the novel finding that PI3K/AKT/mTOR pathways are functionally important for LMS cell survival. Furthermore, we discovered that BEZ235, a dual PI3K/mTOR inhibitor, significantly inhibited LMS growth in vivo. Finally, this small molecule is synergistic with the current standard of care, doxorubicin, suggesting a promising combination therapy for LMS patients.
Methods
Cell line culture conditions and characterisation
Following Research Ethics Board consent at Mount Sinai Hospital (Creation of a Sarcoma Cell Line Biorepository from Human Tissue MSH REB# 10-0310-C), the LMS cell line, STS39, was derived by manual dissociation of a patient sample (pelvic LMS) which was incubated overnight at 37 °C in supplemented DMEM/F-12 10 % FBS (Life Technologies, Burlington, CA) with 2 μg/ml collagenase (SigmaAldrich, Oakville, CA). Cells were repeatedly aspirated to create single cell suspension, centrifuged at 1000 rpm for 5 min, and plated to a T75 flask. SKLMS1 cells obtained from ATCC were cultured in DMEM (Life Technologies), 10 % FBS media. Cells were serially passaged with standard conditions and analysed for copy number variation using array comparative genomic hybridization (aCGH) on the Genome-Wide Human SNP Array 6.0, and mapped using short tandem repeat (STR) analysis (The Centre for Applied Genomics (TCAG) Sick Kids, Toronto, CA). Immunocytochemistry was performed by staining for desmin [clone DE-U-10] (AbCam, Toronto ON, CAN) 1:200, and smooth muscle actin (SMA) (clone 1A4) (Dako, Burlington, CA) 1:200, mouse IgG (Santa Cruz Biotechnology, Dallas, US) 1:400 was used to control for background. To ensure the cell line was representative of the tumor of origin, immunocytochemistry of STS39 cells was matched to patient tumor immunohistochemistry performed by the Department of Pathology at Mount Sinai Hospital. Ki67 and TUNEL staining were quantitated by ImageJ (148v), while p-AKTS473 staining was assessed by staining distribution and intensity. Sequencing of exon 9 and 20 in the p85 subunit of PI3K (Ensembl:ENSG00000145675), and exon 14 and 15 of mTOR (Ensembl:ENSG00000198793) was performed by TCAG.
Kinase library
A 480 compound kinase library was assembled by the drug discovery group at the Ontario Institute for Cancer Research (OICR). All drugs were used as 100 % solutions in DMSO. The drugs encompassed a variety of targets including but not limited to PI3K/AKT, EGFR/ErbB2, CDKs, and GSK3. BEZ235 and BKM120 were supplied by the Medicinal Chemistry Platform or purchased from ChemieTek (Indianapolis, US).
Primary screen
A primary drug screen was performed at the SMART Facility in the Lunenfeld–Tanenbaum Research Institute (LTRI) by seeding SKLMS1 and STS39 cells at a density of 600 cells/well and 700 cells/well respectively, in 384 well plates (VWR). Plates were incubated for 6 h to ensure cell attachment, at which time drugs were pinned at 3 concentrations, 0.2, 1 and 5. Following incubation at 37 °C for 96 h, 45 μl of AlamarBlue® (LifeTechnologies) was added. The fluorescence intensity was then measured after 5 h on a BMG Pherastar plate reader. Cells grown in 0.1 % dimethylsulfoxide (DMSO) (SigmaAldrich) alone served as a control to which the results were normalized, while media with no cells was used for background correction.
Drug selection criteria and hit validation
Hits from the primary screen were defined as the top 10 % of drugs that caused a signal decrease as compared to controls across all three dose groups in both cell lines. The top 10 % of hits were validated by performing a 10-point, threefold serial dilution EC50 curve. Stock compounds in the form of lyophilized powder were reconstituted with DMSO to a concentration of 50 mM. SKLMS1 and STS39 cells were seeded at a density of 5000 c/well and 10,000 c/well respectively, into 96 well plates (Fisher, Ottawa, CA) and compounds were added at final concentrations ranging from 0.25 nM to 5 μM and incubated for 72 h. An ATPlite assay (PerkinElmer) was then performed according to manufacturer’s description and fluorescence read on a PerkinElmer Enspire 2300 multimode plate reader. EC50 curves were generated and analysed using GraphPad Prism 4.0.
Immunoblots
SKLMS1 and STS39 cells were seeded in six well plates at a density of 250,000 c/well and 500,000 c/well respectively. Cells were treated with BKM120 or BEZ235 (5–1000 nM) for 72 h prior to harvest. Following a cold PBS (Sigma) rinse, cells were lysed for 20 min on ice with RIPA buffer (50 mM pH 7.4 Tris–HCl, 150 mM NaCl, 1 % NP-40, 1 mM EDTA) supplemented with phosphatase and protease inhibitors (Sigma). Protein concentration was measured with DC Protein concentration assay (BioRad). Electrophoresis was performed using MiniProtean TGX gels (Bio-Rad) and transferred to PVDF by wet transfer. Immunoblots were performed with the following antibodies: p-AKTS473, total AKT, p-S6KT389, total S6K, p-4EBP1T37/46, total 4EBP1, PARP-1, all from Cell Signaling Technology (Denver, US), and tubulin [clone DM1A] (Sigma, St. Louis, USA). All immunoblots shown are representative of at least three independent experiments.
Combination studies
In vitro combination studies were performed by seeding 96 well plates with SKLMS1 and STS39 cells (5000 c/well and 10000 c/well respectively). Cells were treated with BEZ235 and Dox at multiple concentrations: 4 × IC50, 2 × IC50, IC50, 0.5 × IC50, 0.25 × IC50 according to the following schedule: (1) concurrent treatment with Dox and selective inhibitor (SI) for 72 h, (2) SI alone for 24 h, followed by Dox for an additional 48 h and (3) Dox alone for 24 h followed by SI for an additional 48 h. Cell viability was quantified using the ATPlite assay on a PerkinElmer Enspire 2300 multimode plate reader and analyzed as described below.
Treatment of LMS xenografts with BEZ235 and/or Dox
All in vivo work was carried out in accordance with the Animal Care Committee at the Toronto Centre for Phenogenomics, Toronto, Canada. Female NOD.Cg-
Prkdc
scid
IL2rg
tm1Wjl
/SzJ mice were purchased from JAX Laboratories. At 6–8 weeks of age, animals received an intramuscular (i.m.) injection in the right hind limb with 5 × 10
6 SKLMS1 cells. Once palpable, tumors were measured with callipers and tumor volumes calculated using the following formula: length × width × height × 0.5236 [
24]. When tumors reached approximately 0.5 cm
3 (~4–5 weeks for SKLMS1 cell lines), animals were assigned to four groups: Group 1. vehicle alone consisting of 10 % NMP (1-methyl-2-pyrrolidone)/PEG300 90 % daily by oral gavage and biweekly intraperitoneal (i.p) injection of PBS, Group 2. BEZ235 alone (25 mg/kg of BEZ235 daily by oral gavage dissolved in 10 % NMP/PEG300 90 % as described previously [
25]), Group 3. Dox alone (1.2 mg/kg biweekly i.p. injection of Dox dissolved in PBS) and Group 4. concurrent treatment of BEZ235 and Dox. Body weight and tumor measurements were recorded twice a week for the duration of the study (12 days of treatment). Animals were euthanized by CO
2 asphyxiation when tumors reached approximately 1.5 cm
3 in size.
All tissues were stored either in liquid nitrogen or 10 % formalin. Xenograft tumors were stained for desmin (Dako), SMA [clone 1A4] (Dako), Ki-67 [clone SP6] (ThermoScientific), TUNEL (Roche, TdT) and p-AKTS473 (Cell Signaling Technology) at 1:100. Immunohistochemistry (IHC) was completed using standard protocols. For mouse primary antibodies the Mouse-on-Mouse Peroxidase Kit (Vector Laboratories, Burlington, ON, CAN) was used to prevent false positives to mouse-derived antigens.
Statistics
Analysis of drug synergy was performed by calculating the combination index (CI) as a measure of interaction between two drugs. The CI was calculated according to the median- effect principle of the Chou and Talalay method using the CalcuSyn software 2.1 (BioSoft, UK) [
26,
27]. Mann–Whitney test was performed to determine tumor volume differences between treatment groups. Differences of p < 0.05 were considered statistically significant. One-way ANOVA was performed to determine if there was any difference in mouse weight between treatment groups.
Conclusions
In summary, we demonstrate that inhibition of the PI3K and mTOR pathways impairs LMS growth in vitro and in vivo. Also, we describe that the use of PI3K and mTOR inhibitors may have a synergistic effect with doxorubicin, standard chemotherapy for this disease. Thus, future studies with inhibitors targeting these pathways are warranted.
Authors’ contributions
YB carried out immunoblots, drug synergy and xenograft experiments, performed and interpreted supplemental data and drafted the manuscript with figures. LK performed the drug screen, generated EC50 curves, performed immunohistochemistry experiments and drafted the manuscript with figures. AD supervised and aided in analysis of the drug screen. DU, MI, RA and MP generated the kinase library, aided in data interpretation and selection of drugs of interest. RXS and PCB performed bioinformatics analyses. RV performed mouse imaging and drafted the manuscript. BD performed the pathology review of all immunocytochemistry and immunohistochemistry experiments. RG conceived of the study, participated in its design and coordination and edited the manuscript. All authors read and approved the final manuscript.