Kinase profiling of liposarcomas using RNAi and drug screening assays identified druggable targets

Eleven human liposarcoma cell lines were used in the study: SW872 (undifferentiated liposarcoma) was purchased from American Tissue Type Culture Collection (ATCC, Rockville, MD, USA); LP6 cells were provided by Dr. Christopher DM Fletcher; SA-4 (liposarcoma) cells were a kind gift from Ola Myklebost; LiSa-2 (metastatic poorly differentiated pleomorphic liposarcoma) was kindly provided by Dr. Moller [17]; and FU-DDLS-1 [18] and LPS141 [19] (dedifferentiated liposarcoma) were gifts from Dr. Nishio and Dr. Fletcher, respectively. GOT-3 [20] (recurrence of a myxoid variant of a well-differentiated liposarcoma) and MLS-402 (myxoid liposarcoma) [16] were generous gifts from Dr. Åman. T778 and T1000 (recurrent well-differentiated liposarcoma) were generously provided by Dr. Pedeutour. All these liposarcoma cell lines were maintained in RPMI medium supplemented with fetal bovine serum in a humidified incubator at 37 °C with 5% CO₂ [14]. STR profiling was done on all the cell lines.

Liposarcoma cell lines were transfected with a siRNA kinase library (including 94 kinase gene target) as described earlier [21]. Liposarcoma cells were manually transfected using electroporation technology with a pool of four siRNAs targeting different regions of individual kinase RNAs per well (Dharmacon) along with control non-specific siRNAs. All transfections were performed in triplicates, three times, and cell viability was assessed using CellTiter 96 AQ_ueous One Solution Cell Proliferation Assay (Promega). Cell viability values were calculated by normalizing to mean of non-specific siRNA control values. Kinase genes were considered as a significant target only the P value was less than 0.05, and mean viability value was less than 70% of non-specific siRNA control value.

High-throughput drug inhibitor screen (using 120 drugs) was performed as previously published [22]. List of drugs used in the study provided are in Additional file 1: Table S1A. Briefly, each drug was put in replicates in 96-well plates, and 50,000 cells were added per well diluting the drug to the final desired concentration. Cells were treated with the drug for 3 days at 37 °C with 5% CO₂, and cell proliferation was measured with CellTiter 96 AQ_ueous One Solution Cell Proliferation Assay (Promega). Wells without drug were set up as controls and used to normalize the data. Previously published algorithm was implemented to obtain automated IC₅₀ calculations and identification of therapeutic target [22].

Human PTK2 and KIT gene-specific SMARTpool ON-TARGETplus siRNA containing four pairs of siRNAs including non-targeting control siRNA pool were purchased from Dharmacon (CO, USA). LPS141 and MLS402 cells were transfected with 20 nM PTK2 and KIT siRNA pool, respectively, using Lipofectamine RNAi Max according to the manufacturer’s protocol along with non-targeting siRNA. Transfection efficiency was around 80–90%. Cells were harvested for protein expression analysis 48 h after transfection.

We generated stable knockdown of PTK2 and KIT in LPS141 and MLS402 cells, respectively, using gene-specific short hairpin RNAs (shRNAs) and non-targeting shRNA in a lentiviral vector system [14]. Sequences of shRNA used are listed in Additional file 1: Table S1B. Knockdown cells were analyzed for protein inhibition by western blotting and for cell proliferation by MTT assay.

Short guide RNAs (sgRNAs) for CRISPR/Cas9 were designed at BROAD sgRNA design website (https://​portals.​broadinstitute.​org/​gpp/​public/​analysis-tools/​sgrna-design) targeting PTK2 and KIT kinase genes along with sgRNA targeting bacterial lacZ as control sgRNA. Sequences for guide RNAs available in Additional file 1: Table S1B. Complementary sgRNAs were annealed and cloned in lentiCRISPRv2 vector (a gift from Feng Zhang: Addgene plasmid # 52961) [23]. Lentiviral particles were prepared; cells were infected, and after 24 h of infection, cells were selected with puromycin (0.5 μg/ml) for 3–5 days [24]. After two additional weeks, cells were harvested and analyzed for silencing of the target gene. This was followed by analyzing cell viability at different time points using MTT assay.

Full-length PTK2 cDNA clone without the 3′UTR was generated in lentiviral pLX303 [25] (a gift from David Root; Addgene plasmid # 25897) vector using pDONR223-PTK2 vector [26] (gift from William Hahn and David Root; Addgene plasmid # 23902) by Gateway cloning system. Lentiviral particles expressing PTK2 cDNA were infected in PTK2 shRNA3 knockdown LPS141 cells (shRNA directed against 3′UTR of PTK2 gene) and selected with blasticidin for a week. After selection, infected cells were analyzed for protein expression and cell viability.

The coding region of KIT without the 3′UTR with a GFP tag in pCMV6-AC-GFP vector was purchased from OriGene (MD, USA). The vector expressing KIT-GFP was transfected in KIT shRNA3 knockdown stable MLS402 cells (shRNA directed against 3′UTR of KIT gene) using Lipofectamine 2000 following manufacturer’s instructions. Cells were analyzed for protein expression and cell proliferation at 72 h after transfection.

Reduction in cell viability was measured by MTT cell proliferation assay as previously described [14]. Briefly, 5000 cells per well were seeded in a 96-well plate and incubated for 24 h. Serial dilutions of the drug were added in triplicates to the cells along with control diluent and incubated for 3 days at 37 °C with 5% CO₂. MTT reagent (10 μl/well) was added and incubated for 2–4 h. The reaction was stopped using 100 μl/well MTT stop solution. Absorbance was measured at 570 nm.

Soft agar colony formation assays were used to study the anchorage-independent growth of cells before and after drug treatment as described earlier [14]. In brief, a base agarose layer (0.8%) was prepared in a 24-well plate containing different dilutions of ponatinib or control diluent. Cells were harvested, counted, and 1500 cells per well resuspended in top agarose layer (0.5%) with ponatinib or diluent dilutions and layered on the base agarose layer. After 10–30 days of culture, colonies were counted using an inverted microscope.

Anchorage-dependent colony formation assays were performed by seeding cells (2000 cells/well) in a 6-well plate. After 24 h, media changed to media containing different concentrations of the drug. Media was changed twice a week, and colonies were allowed to grow for 10–15 days. Colonies were washed with PBS, fixed with methanol, stained with crystal violet, and quantified.

Cells (4 × 10⁶) were seeded in a10-cm dish; after 24 h, cells were treated with different drug concentrations for different durations. After treatment, cells were harvested, washed with ice-cold PBS, and lysates were prepared using Protein Extraction Reagent (Thermo Scientific) containing a protease inhibitor. Proteins were separated on SDS-PAGE and transferred to polyvinylidene difluoride membranes, which were further incubated with the indicated antibodies, and detection was performed using Chemiluminescent HRP Substrate. List of antibodies used is provided in Additional file 1: Table S1C.

Cells were cultured with either ponatinib or diluent, washed with ice-cold PBS and fixed in cold 70% ethanol while vortexing, and washed in cold PBS. Propidium iodide/RNase solution was added and analyzed by flow cytometer.

Cells were seeded overnight and treated with either ponatinib or control diluent for 16 h. After treatment, cells were collected, washed, and stained with Annexin V-APC and Propidium iodide (BD Biosciences) according to manufacturer’s instructions. Stained cells were analyzed by flow cytometry.

All animal experiments were performed according to the ethical regulations of Institutional Animal Care and Use Committee of the National University of Singapore. Male NOD SCID gamma mice (5–6 weeks old) were transplanted with the liposarcoma cells. LPS141 cells (2 × 10⁶ cells per mice in 200 μl volume) combined with matrigel were subcutaneously injected into the flank of the mice. After tumors reached about ~ 100 mm³, mice were randomly divided into two groups of nine mice each. Experimental cohort received oral ponatinib (10 mg/kg), and control mice received diluent, both by oral gavage. Drug treatment continued every day for 3 weeks. At the end of the experiment, mice were sacrificed; tumors were dissected and weighed. Half of each tumor was snap frozen, and the other half was formalin fixed for immunohistochemical analysis. Tumor tissues investigated for protein expression by preparing tumor lysates using RIPA lysis buffer. After bicinchoninic acid assay protein estimation, lysates were analyzed by SDS-PAGE and western blotting assays.

Immunohistochemistry was performed on sections from the xenografts tumor tissue using the Ki-67 antibody as described earlier [14].

Statistical significance was determined using the paired two-tailed Student’s t test. A P value of ≤ 0.05 was considered statistically significant. All in vitro experiments were done three times in triplicates, and data expressed as mean ± standard error.

Loss of function screening was performed using 11 liposarcoma cell lines against a tyrosine kinase siRNA library (94 target genes) to identify targetable kinases. List of siRNA target genes is provided in Additional file 1: Table S1D. Liposarcoma cell lines of different subtypes were assayed including well-differentiated: T449, T1000, T778, and GOT3; dedifferentiated: LPS141, FU-DDLS-1, LP6, and SA4; myxoid: MLS402; pleomorphic: LiSa-2; and undifferentiated: SW872. Transfected cells received four individual siRNA molecules against each kinase gene. After 96 h of transfection, cell viability was assessed using MTS assay, and data was normalized. Cell viability was calculated as a mean percent normalized to non-specific control wells for each of the kinase genes and plotted as a bar graph (Fig. 1a and Additional file 1: Table S1D). Kinase genes showing a significant reduction in cell viability and proliferation are shown as red bars. List of kinase targets sensitized by siRNAs that caused a significant reduction in cell viability in at least one cell line is provided in Table 1. Several kinases were associated with the liposarcoma cell viability, and some of them are subtype specific. Two kinase genes which were the top hits and suppressed the growth of more than one cell line are protein tyrosine kinase 2 (PTK2) (LPS141, T778) and KIT (T449, GOT-3, and MLS402). In further experiments, PTK2 and KIT were silenced in LPS141 (dedifferentiated) and MLS402 (myxoid) cells, respectively, using siRNA and shRNA separately (Fig. 1b). Knockdown of PTK2 and KIT significantly reduced the cell viability compared to controls as assessed using the MTT assay (Fig. 1b). We also generated CRISPR-Cas9 mediated PTK2 and KIT gene knockout in LPS141 and MLS402 cells, respectively, using two sgRNA against each gene along with control LacZ sgRNA. Western blotting analysis established inhibition of the expression of the target protein in cells transfected with target sgRNA1 and sgRNA2 compared to control LacZ sgRNA. Significant growth reduction in kinase knockout cells compared to control cells in both cell lines was observed using MTT assay (Fig. 1c). In addition, we performed rescue experiments. We generated stable liposarcoma cell lines containing shRNA targeting 3′UTR of PTK2 and KIT gene and rescued these cells by expressing cDNA of the respective protein (not containing 3′UTR of these genes). Exogenous expression of PTK2 and KIT protein in PTK shRNA3-LPS141 and KIT shRNA3-MLS402 cells, respectively, resulted in significant rescue of cell proliferation signifying that inhibition of proliferation was specific to knockdown of target gene (Additional file 1: Figure S1). These results further validated and strengthened the potential of PTK2 in dedifferentiated liposarcoma and KIT in myxoid liposarcoma as targetable kinases.

In addition, a small-molecule inhibitor screen was performed to identify kinase targets in liposarcoma cells. A library of small-molecule kinase inhibitors (120 drugs FDA approved or in clinical trials) was tested for their anti-proliferative activity against our panel of 11 liposarcoma cell lines. Most of these drugs targeted kinases, and a complete list is provided in Additional file 1: Table S1A. Response to these small molecule inhibitors is shown as a heat map (Fig. 2). Complete data containing IC₅₀ values of all drugs is provided in Additional file 1: Table S1A. Most of the cell lines were highly sensitive to growth inhibition by a proteasome inhibitor (Velcade), a protein kinase inhibitor (Staurosporine), and Hsp90 inhibitors (17-AAG and Elesclomol). PI3K, mTOR, and CDK inhibitors also were potently cytotoxic against the liposarcoma cell lines. Inhibitors targeting ERBB and MAPK family members were inactive, hence ruling out therapeutically inhibiting these signaling pathways in liposarcoma. Interestingly, the myxoid liposarcoma cells (MLS402) showed a robust sensitivity towards all four tested Aurora kinase inhibitors. In contrast, the IGF-1R inhibitor NVP-ADW742 selectively inhibited the growth of many well-differentiated/dedifferentiated liposarcomas suggesting possible subtype-specific therapeutic targets.

Among the receptor tyrosine kinase inhibitors, ponatinib (trade name Iclusig, previously AP24534), dasatinib, and sunitinib (previously known as SU11248) were the most potent drugs affecting the viability of liposarcoma cells lines irrespective of subtypes (Fig. 3a), and each is FDA approved for multiple other malignancies. These are multi-targeted tyrosine kinase inhibitors affecting various targets including BCR/ABL, KIT, SRC, VEGFR, and FGFR (Fig. 3b). KIT is a common target of these three inhibitors (Fig. 3b). Significantly, KIT was also a top hit in our RNAi screen with the KIT siRNA prominently inhibiting cell viability of three liposarcoma cell lines. Taken together, the screening assays revealed and highlighted KIT as an important and potentially druggable kinase for liposarcoma therapy.

Furthermore, we validated the potent suppressive effects of the two KIT kinase inhibitors: ponatinib (IC₅₀ range from 33 to 525 nM) and dasatinib (IC₅₀ range from 0.20–105 nM) against all the liposarcoma cell lines (Fig. 3c, d). A recent phase 2 trial of dasatinib in patients with advanced sarcoma (not specifically liposarcoma) found that the drug had little clinical benefit [27]. Therefore, we focused on ponatinib, which is an orally bioavailable, multi-targeted receptor tyrosine kinase inhibitor with antiangiogenic and antineoplastic activities.

Ponatinib (16 h exposure) inhibited the phosphorylation of KIT (Y719) and its downstream signaling molecules including phosphorylation of AKT, ERK1/2, STAT3, mTOR, and P70S6K in a dose-dependent fashion in dedifferentiated (LPS141) and a myxoid (MLS402) cell lines (Fig. 4a). Growth inhibitory effect of ponatinib against these liposarcoma cells was confirmed using two additional in vitro assays: foci formation (Fig. 4b) and clonogenic growth in soft agar (Fig. 4c).

Ponatinib (16 h, 1000 nM) caused a decrease in S phase and an increase in sub-G1 phase (apoptotic cells) in the liposarcoma cells (Fig. 5a). Ponatinib induced a dose-dependent decreased levels of cyclin D1, total CDK4, and phosphorylated RB (Fig. 5b). In addition, apoptosis measured by Annexin V/propidium iodide staining showed that ponatinib caused a significant increase in apoptotic liposarcoma cells compared to diluent control (Fig. 5c). Also, ponatinib decreased levels of phosphorylated BAD, BCL-X_L, and MCL-1 and increased the levels of cleaved caspase 9 as well as proapoptotic BAX (Fig. 5d). These results show that the anti-proliferative activity of ponatinib was associated with cell cycle arrest and apoptosis of the liposarcoma cells.

Tumors in experimental mice (ponatinib treatment) were significantly (P = 0.004) smaller than in control mice (vehicle treatment) (Fig. 6a, b). Body weights and overall activity of both murine groups were stable during the experiment (data not shown). Tumors of the ponatinib-treated mice had a significantly lower percentage of proliferating cells (less Ki-67 staining) compared to the tumors from the control mice (Fig. 6c). Further investigation revealed slightly reduced levels of phosphorylated KIT and FGFR in ponatinib-treated tumors compared to vehicle-treated group (Fig. 6d). Tumor cell lysates revealed activation of pro-apoptotic proteins (BAX and cleaved caspase 9) and reduction of anti-apoptotic proteins (MCL-1, BCL-X_L, p-BAD) in tumors of the ponatinib-treated mice compared to tumors of the vehicle-treated mice (Fig. 6e). Thus, the xenograft study demonstrated ponatinib significantly decreased the tumor growth by inhibiting cellular proliferation in vivo and inducing apoptosis.