Proteomic Analysis of the EWS-Fli-1 Interactome Reveals the Role of the Lysosome in EWS-Fli-1 Turnover.

Ewing sarcoma is a cancer of bone and soft tissue in children that is characterized by a chromosomal translocation involving EWS and an Ets family transcription factor, most commonly Fli-1. EWS-Fli-1 fusion accounts for 85% of cases. The growth and survival of Ewing sarcoma cells are critically dependent on EWS-Fli-1. A large body of evidence has established that EWS-Fli-1 functions as a DNA-binding transcription factor that regulates the expression of a number of genes important for cell proliferation and transformation. However, little is known about the biochemical properties of the EWS-Fli-1 protein. We undertook a series of proteomic analyses to dissect the EWS-Fli-1 interactome. Employing a proximity-dependent biotinylation technique, BioID, we identified cation-independent mannose 6-phosphate receptor (CIMPR) as a protein located in the vicinity of EWS-Fli-1 within a cell. CIMPR is a cargo that mediates the delivery of lysosomal hydrolases from the trans-Golgi network to the endosome, which are subsequently transferred to the lysosomes. Further molecular cell biological analyses uncovered a role for lysosomes in the turnover of the EWS-Fli-1 protein. We demonstrate that an mTORC1 active-site inhibitor, torin 1, which stimulates the TFEB-lysosome pathway, can induce the degradation of EWS-Fli-1, suggesting a potential therapeutic approach to target EWS-Fli-1 for degradation.

Introduction

Ewing sarcoma is the second most common malignancy of bone and soft tissues in children and young adults and is characterized by a chromosomal translocation that generates a fusion oncogene between EWS and an Ets family transcription factor, most commonly Fli-1.[1−5] EWS-Fli-1 fusion accounts for 85% of Ewing sarcoma cases. Ewing sarcoma is an aggressive tumor with relatively poor long-term outcome. Overall survival is approximately 60%, and the five-year survival of recurrent cases is less than 10%. Considering that current cytotoxic chemotherapies used for Ewing sarcoma are not improving the survival of metastatic or recurrent disease, a new approach for targeted therapy needs to be developed.[1−5] The growth and survival of Ewing sarcoma cells critically depend on the EWS-Fli-1 fusion oncoprotein.[1−6] Therefore, targeting EWS-Fli-1 is a promising approach to treat Ewing sarcoma. However, despite a number of attempts, an EWS-Fli-1-targeted therapy has not materialized to date and EWS-Fli-1 continues to be “the perfect target without a therapeutic agent”.[7]

EWS-Fli-1 is a transcription factor that controls the expression of a number of genes important for cell proliferation and transformation.[1−4] Transcriptional regulation by EWS-Fli-1 has been studied extensively, but little is known about the biochemical properties of the EWS-Fli-1 protein. To gain insight into the biochemical nature of the EWS-Fli-1 protein, we undertook proteomic analyses of the EWS-Fli-1 interactome. The result from the interactome analyses was used to provide leads for subsequent molecular biological analyses. Using a tandem affinity purification approach, we identified known EWS-Fli-1 interactors such as EWS[8] and RNA helicase A.[9] Using a proximity-dependent biotinylation technique, BioID,[10] we identified cation-independent mannose 6-phosphate receptor (CIMPR) as a protein located in the vicinity of EWS-Fli-1 within a cell. CIMPR is a cargo that mediates the sorting of lysosomal hydrolase precursors from the trans-Golgi network to endosomes.[11] Additional molecular cell biological analyses revealed that the EWS-Fli-1 protein turns over by a lysosome-dependent mechanism. We show that torin 1, which is an active-site inhibitor of mTORC1 that was shown to stimulate the TFEB-lysosome pathway, can reduce EWS-Fli-1 protein levels in Ewing sarcoma cells, suggesting a potential utility of mTORC1 active-site inhibitors as therapy for Ewing sarcoma.

Experimental Procedures

Reagents

Chloroquine and pepstatin A were purchased from MP Biomedicals. Doxorubicin was purchased from Sigma-Aldrich. Rapamycin and MG-132 were purchased from Calbiochem/EMD Biosciences. Cytosine arabinoside was from Tocris Bioscience. Torin 1 was from Cayman Chemical. The target sequences for shRNAs are as follows: human CIMPR shRNA, CTACCTGTATGAGATCCAA; human VPS26A shRNA, CTCTATTAAGATGGAAGTG; luciferase shRNA, GCACTCTGATTGACAAATACGATTT. Cathepsin D and firefly luciferase–EWS-Fli-1 fusion cDNAs were cloned into pCDF1 lentiviral vector (System Biosciences).

Cell Culture

293 cells and 293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% calf serum. A673 cells and HeLa cells were cultured in DMEM supplemented with 10% fetal calf serum. TC71 cells were cultured in RPMI1640 medium supplemented with 10% fetal calf serum. Calcium phosphate coprecipitation was used for transfection of 293 and 293T cells. Lentiviruses were prepared by transfection in 293T cells following System Biosciences’ protocol, and the cells infected with lentiviruses were selected with 2 μg/mL puromycin for 48 h as described.[12,13] Luciferase assays were done as described.[13]

Protein Sample Preparation and Mass Spectrometry

Tandem Affinity Purification of FLAG-His-EWS-Fli-1-Interacting Proteins

Forty 15-cm plates of 293T cells were transfected with FLAG-His-EWS-Fli-1 (type 1 fusion); 48 h after transfection, the cells were lysed in TN buffer (10 mM Tris pH 7.4/150 mM NaCl/1% NP-40/1 mM AEBSF/10 μg/mL aprotinin/10 μg/mL Leupeptin/1 μg/mL Pepstatin A/20 mM sodium fluoride). The lysate was incubated with Ni-NTA agarose (Qiagen), FLAG-His-EWS-Fli-1 and its interacting proteins were collected by centrifugation, washed three times with TN buffer, and eluted with 50 mM sodium phosphate buffer pH 8.0/150 mM NaCl/250 mM imidazole. The eluted sample was immunoprecipitated with anti-FLAG antibody (M2, Sigma-Aldrich), the immunoprecipitate was eluted with FLAG peptide (Sigma-Aldrich), and the eluted protein sample was processed with an Amicon Ultra 0.5 3k centrifugal filter device (Millipore) for concentration and buffer exchange to 50 mM Tris pH 8.5. Proteins were digested at 37 °C overnight with trypsin (Promega; 1:10, enzyme/substrate) in the presence of 10% acetonitrile (ACN). The pH of the digestion solution was adjusted to 7.5 with 1 mM ammonium bicarbonate, if necessary. The resulting tryptic peptides were analyzed by HPLC-ESI-tandem mass spectrometry (HPLC-ESI-MS/MS) on a Thermo Fisher LTQ Orbitrap Velos mass spectrometer fitted with a New Objective Digital PicoView 550 NanoESI source. Online HPLC separation of the digests was accomplished with an Eksigent/AB Sciex NanoLC-Ultra 2-D HPLC system: column, PicoFrit (New Objective; 75 μm i.d.) packed to 15-cm with C18 adsorbent (Vydac; 218MSB5 5 μm, 300 Å); mobile phase A, 0.5% acetic acid (HAc)/0.005% trifluoroacetic acid (TFA); mobile phase B, 90% ACN/0.5% HAc/0.005% TFA; gradient 2–42% B in 120 min; flow rate, 0.4 μL/min. Precursor ions were acquired in the Orbitrap in centroid mode at 60,000 resolution (m/z 400); data-dependent collision-induced dissociation (CID) spectra of the 10 most intense ions in the precursor scan above a threshold of 3,000 were acquired at the same time in the linear trap (isolation window for MS/MS, 3; relative collision energy, 30). Ions with a 1+ or unassigned charge state were not fragmented. Dynamic exclusion settings were: repeat count, 1; repeat duration, 30 s; exclusion list size, 500; exclusion duration, 30 s.

BioID Proximity-Dependent Biotinylation Proteomics

Three 15-cm plates of 293 cells were transfected with BioID-EWS-Fli-1 (Myc tag and BirA R118G mutant fused to the N-terminus of EWS-Fli-1). Twenty-four hours after transfection, biotinylation of proteins in the vicinity of BioID-EWS-Fli-1 within the cells was induced for 24 h by the addition of 50 μM biotin to the culture medium. The cells were lysed by boiling in a lysis buffer (50 mM Tris, pH 7.4/500 mM NaCl/0.4% SDS/5 mM EDTA/1 mM DTT/1 mM AEBSF/10 μg/mL aprotinin/10 μg/mL Leupeptin/1 μg/mL Pepstatin A/20 mM sodium fluoride). The viscosity of the sample was reduced by passing it through an 18-gauge needle followed by sonication. Triton X-100 was added to 2% final concentration, and the biotinylated proteins were purified using streptavidin agarose (Pierce/Thermo Fisher) and eluted in an SDS-PAGE sample buffer. The proteins in each sample were fractionated by SDS-PAGE and visualized by Coomassie blue. Each gel lane was divided into six slices, and the proteins in each slice were digested in situ with trypsin (Promega modified) in 40 mM NH4HCO3 overnight at 37 °C. The resulting tryptic peptides were analyzed by HPLC-ESI-MS/MS as described above, except that a 30 min HPLC gradient was employed and the six most intense ions in the precursor scan were fragmented.

Mass Spectrometry Data Analysis

The Xcalibur raw files were converted to mzXML format using ReAdW (http://tools.proteomecenter.org/wiki/index.php?title=Software:ReAdW) and were searched against the IPI human protein database (v 3.24; 66,923 protein entries) using X! Tandem. Methionine oxidation was considered as a variable modification in all searches, and lysine biotinylation was included for the BioID experiments. Up to one missed tryptic cleavage was allowed. The X! Tandem search results were analyzed by the Trans-Proteomic Pipeline[14] version 4.3. Peptide/protein identifications were validated by Peptide/ProteinProphet.[15,16] A ProteinProphet score of 0.9 was used as a cutoff, which corresponded to false identification rates of 1.1% and 0.7% in the FLAG-His-EWS-Fli-1 and BioID-EWS-Fli-1 data sets, respectively.

Immunoblotting

Immunoblotting was performed as described.[12,13] The following antibodies were used: rabbit polyclonal anti-CIMPR (ab32815, Abcam); mouse monoclonal anti-cyclin D1 (2926, Cell Signaling Technologies); mouse monoclonal anti-FLAG (M2, Sigma-Aldrich); rabbit polyclonal anti-FLAG (Immunology Consultants Laboratory, Inc.); rabbit polyclonal anti-Fli-1 (ab15289, Abcam); mouse monoclonal anti-HA (16B12, Covance); mouse monoclonal anti-LAMP2 (55803, BD Biosciences); rabbit polyclonal anti-mSin3A (K-20, Santa Cruz Biotechnology); rabbit polyclonal anti-Myc (N262, Santa Cruz Biotechnology); mouse monoclonal anti-nucleolin (C23, Santa Cruz Biotechnology); mouse monoclonal anti-p62/SQSTM1 (610832, BD Biosciences); and mouse monoclonal anti-tubulin (DM1A, Lab Vision).

Preparation of the Lysosomes

A673 cells were treated with 100 μM chloroquine for 12 h or left untreated. Lysosomes were prepared using the Lysosome Enrichment Kit for Tissue and Cultured Cells (#89839, Pierce/Thermo Scientific) following the manufacturer’s protocol. Briefly, cells were lysed by sonication in the manufacturer’s lysis buffer and centrifuged at 500g for 10 min. The resulting supernatant was placed on top of a density gradient comprising of 17% - 30% iodixanol -5,5′-[(2-hydroxy-1,3-propanediyl)-bis(acetylamino)] bis [N,N′-bis(2,3 dihydroxypropyl-2,4,6-triiodo-1,3-benzenecarboxamide)] in the manufacture’s gradient dilution buffer, with an aliquot of the supernatant saved as the input fraction. The lysosome enrichment gradient was centrifuged at 145,000g using a SW60Ti rotor (Beckman-Coulter) for 2 h at 4 °C. The top layer of the gradient, which contains the lysosomes, was collected, diluted with two volumes phosphate-buffered saline, and centrifuged at 16,000g for 30 min at 4 °C. The resulting lysosome pellet was washed once in gradient dilution buffer and dissolved in SDS-PAGE sample buffer. For immunoblotting analysis, the lysosome fraction was loaded at 100x compared to the input.

Results and Discussion

Proteomic Analysis of the EWS-Fli-1-Interacting Proteins

To dissect the EWS-Fli-1 interactome, we initially employed a tandem affinity purification approach. We expressed FLAG-His-tagged EWS-Fli-1 in human embryonic kidney 293T cells and isolated the EWS-Fli-1-containing protein complex by nickel agarose chromatography followed by anti-FLAG immunoprecipitation, which was subsequently analyzed by HPLC-ESI-MS/MS as described in the Experimental Procedures (Figure 1A). At a Protein Prophet probability score of 0.9 or higher, 105 different proteins were identified (Table S1). To exclude the false positive identifications, we employed the data sets obtained using two unrelated nuclear proteins, p21 (CDKN1A) and histone macroH2A. The proteins commonly identified by FLAG-His-EWS-Fli-1 and FLAG-His-p21 or FLAG-His-macroH2A with a Protein Prophet probability score of 0.9 or higher were considered as false positive identifications (the list of these proteins is shown in Table S2). After subtracting these false positive identifications, 54 proteins remained in the FLAG-His-EWS-Fli-1 data set (shown in Table S3), including known EWS-Fli-1 interactors such as EWS (Figure 1B; Note that two EWS C-terminal peptides, which are absent in EWS-Fli-1, were identified)[8] and RNA helicase A (Figure 1C).[9] While the tandem affinity purification approach identified known interactors for EWS-Fli-1, we noticed that the majority of FLAG-His-EWS-Fli-1 expressed in 293T cells was not solubilized under the non-denaturing solubilization conditions used for tandem affinity purification (Figure 1D).

Figure 1

Tandem affinity purification analysis of the EWS-Fli-1-interacting proteins. (A) Tandem affinity purification procedure. 293T cells were transfected with FLAG-His-EWS-Fli-1. Forty-eight hours after transfection, FLAG-His-EWS-Fli-1 and its interacting proteins were isolated by nickel affinity chromatography followed by anti-FLAG immunoprecipitation and the protein sample was analyzed by tandem mass spectrometry. (B) EWS peptides assigned with high confidence. (C) RNA helicase A peptides assigned with high confidence. (D) FLAG-His-EWS-Fli-1 is mostly insoluble under the lysis conditions used for tandem affinity purification. The abundance of FLAG-His-EWS-Fli-1 in whole cell lysate (lane 1 and 4), tandem affinity purification lysate (lane 2 and 5), and postlysis pellet (lane 3 and 6) was determined by anti-FLAG immunoblotting. Tubulin serves as a loading control.

Therefore, as an alternative approach to dissect the EWS-Fli-1 interactome, we used the proximity-dependent biotinylation technique, BioID.[10] In the BioID approach, a bait protein is fused to a mutated BirA biotinylase (BirA R118G) which promiscuously biotinylates the lysine residues of proteins in the vicinity (within 20–30 nm). The biotinylated vicinal proteins are purified by streptavidin affinity chromatography and are identified by mass spectrometry. The BioID approach does not require the purification of a stable protein complex under non-denaturing conditions and is useful for the analysis of insoluble protein complexes or transient low-affinity interactions. It has been successfully applied to identify interactors for nuclear lamin A,[10] whose insolubility has hampered the analysis of its interactors, and to identify the protein components of bilobe,[17] an insoluble cytoskeletal structure in Trypanosoma brucei.

We expressed BioID-tagged EWS-Fli-1 in human embryonic kidney 293 cells, induced the biotinylation of the proteins in the vicinity of BioID-EWS-Fli-1 by 50 μM biotin, and isolated the biotinylated proteins by streptavidin affinity chromatography (Figure 2A). The biotinylated proteins were analyzed as described in the Experimental Procedures. At a Protein Prophet probability score of 0.9 or higher, 561 different proteins were identified (Table S4). To exclude the false positive identifications, we employed the CRAPome database,[18] which is a “contaminant repository for affinity purification”. Using CRAPome version 1.1 (http://www.crapome.org/), 12 control proximity-dependent biotinylation experiments performed in 293 cells were compiled. The proteins detected with five or more spectral counts per control experiment were considered as false positive identifications by the proximity-dependent biotinylation approach. The list of these false positive identifications is shown in Table S5, which contains 656 proteins. After subtracting these false positive identifications from Table S4, 366 proteins remained (Table S6), including EWS, which was shown to form a hetero-oligomer with EWS-Fli-1.[8]

Figure 2

Proximity-dependent biotinylation analysis using BioID-EWS-Fli-1. (A) Proximity-dependent biotinylation procedure. 293 cells were transfected with BioID-EWS-Fli-1 (Myc tag and BirA R118G mutant fused to EWS-Fli-1), and the biotinylation of proteins in the vicinity of BioID-EWS-Fli-1 was induced by the addition of 50 μM biotin to the culture medium. The biotinylated vicinal proteins were purified by streptavidin agarose chromatography and were analyzed by tandem mass spectrometry. (B) CIMPR peptides assigned with high confidence. (C) Co-immunoprecipitation of EWS-Fli-1 and CIMPR. 293T cells were transfected with FLAG-EWS-Fli-1 or FLAG-vector. The transfected cells were treated with 100 μM chloroquine for 12 h where indicated. The physical interaction of FLAG-EWS-Fli-1 and CIMPR was examined by anti-FLAG immunoprecipitation followed by anti-CIMPR immunoblotting. The immunoprecipitation of FLAG-EWS-Fli-1 was verified by anti-FLAG immunoblotting.

After excluding the false positive identifications, 54 proteins were identified from FLAG-His-EWS-Fli-1 affinity purification and 366 proteins were identified from BioID-EWS-Fli-1 analysis, and of these, four proteins were in common: EWS-Fli-1, HNRNPA3, U2AF1, and EWS. Additionally, different isoforms of SUMO proteins were identified in both FLAG-His-EWS-Fli-1 and BioID-EWS-Fli-1 experiments. We note that there is a possible SUMO-binding motif (LELLSDS, residues 340–346) in EWS-Fli-1, which could mediate the interaction with SUMO proteins. EWS-Fli-1 does not contain a sumoylation motif (hydrophobic-K-X-E), and we have not been able to detect its sumoylation (data not shown).

Among the high-scoring proteins identified by the BioID approach, cation-independent mannose 6-phosphate receptor (CIMPR) caught our attention because the detection of CIMPR, which is a cargo that mediates the sorting of lysosomal hydrolase precursors from the trans-Golgi network to endosomes,[11] using BioID-EWS-Fli-1 suggested a possible new link between EWS-Fli-1 and the endosome–lysosome system. Molecular biological characterization of EWS-Fli-1, performed in parallel with the proteomic analyses, demonstrated that EWS-Fli-1 is a relatively stable protein and does not turn over by a proteasome-dependent mechanism (described below), which led us to consider a possibility that EWS-Fli-1 turns over by a lysosome-dependent mechanism.

In the BioID-EWS-Fli-1 experiment, 17 unique and 19 total peptides from CIMPR were identified (Figure 2B and Table S6). Consistent with the identification of CIMPR by the BioID approach, we observed the coimmunoprecipitation of FLAG-EWS-Fli-1 and endogenous CIMPR upon treatment with chloroquine, an inhibitor of lysosomal degradation (Figure 2C). It is noteworthy that CIMPR was identified by the BioID approach using only three 15-cm plates of cells whereas it was not identified by the tandem affinity purification approach using 40 15-cm plates of cells (even though the latter used 293T cells which generally result in higher protein expression levels than 293 cells employed in the former). We believe this is related to the insolubility of EWS-Fli-1 under the solubilization conditions used for tandem affinity purification (Figure 1D). 348 peptides derived from EWS-Fli-1 were identified by the BioID approach using three 15-cm plates (Table S6) whereas 120 peptides derived from EWS-Fli-1 were identified by the tandem affinity purification approach using 40 15-cm plates (Table S3), suggesting the efficient solubilization of EWS-Fli-1 by the BioID lysis buffer which contains 0.4% SDS. Our result as well as two previous BioID studies[10,17] suggest the utility of the BioID approach for the dissection of protein–protein interactions involving insoluble proteins.

EWS-Fli-1 Turnover Occurs via a Lysosome-Dependent Mechanism

The protein transport function of CIMPR is regulated by the retromer complex, which redirects CIMPR from the endosome to the trans-Golgi network.[19,20] Interestingly, knockdown of CIMPR or VPS26A, an essential component of the retromer, resulted in reduced EWS-Fli-1 protein expression (Figure 3A). This raised a possibility that EWS-Fli-1 is transported to the late endosome and degraded by the lysosome, especially when CIMPR and retromer functions are compromised. Importantly, we found that coexpression of TFEB, a potent inducer of lysosomal biogenesis,[21] resulted in striking degradation of EWS-Fli-1 (Figure 3B). Conversely, inhibition of lysosomal degradation by chloroquine stabilized EWS-Fli-1 (Figure 3C). Lysosomes contain many hydrolases that degrade various biomolecules, including proteins. We found that one of the lysosomal proteases, cathepsin D, can degrade EWS-Fli-1, which was inhibited by chloroquine or a cathepsin D inhibitor, pepstatin A (Figure 3D). Cathepsin D did not degrade p53 (Figure 3D), indicating that cathepsin D does not degrade proteins non-selectively. Furthermore, we found that endogenous EWS-Fli-1 in A673 Ewing sarcoma cells is degraded upon expression of cathepsin D (Figure 3E) and is stabilized by chloroquine, an inhibitor of lysosomal degradation (Figure 3F, left). Endogenous EWS-Fli-1 in A673 cells was not stabilized by a proteasome inhibitor, MG-132 (Figure 3F, right). Using subcellular fractionation, we detected endogenous EWS-Fli-1 in the lysosomal fraction, which increased upon chloroquine treatment (Figure 3G). An abundant lysosomal glycoprotein, LAMP2, was readily detectable in the lysosomal fraction whereas a nuclear transcriptional corepressor mSin3A, which is not known to be located in the lysosome, was absent (Figure 3G). p62/SQSTM1, a known substrate of lysosomal degradation, displayed increased lysosomal location upon chloroquine treatment (Figure 3G). These results indicate that EWS-Fli-1 is degraded by the lysosome.

Figure 3

EWS-Fli-1 turns over by a lysosome-dependent mechanism: (A) Knockdown of CIMPR or VPS26A results in destabilization of FLAG-EWS-Fli-1. 293 cells were cotransfected with FLAG-EWS-Fli-1 and shRNA against luciferase (control), CIMPR, or VPS26A. Forty-eight hours after transfection, the levels of FLAG-EWS-Fli-1 were examined by anti-FLAG immunoblotting. Nucleolin serves as a loading control. (B) TFEB induces EWS-Fli-1 degradation in 293 cells. 293 cells were cotransfected with FLAG-EWS-Fli-1 and HA-TFEB or empty vector. Forty-eight hours after transfection, the levels of FLAG-EWS-Fli-1 were examined by anti-FLAG immunoblotting. Tubulin serves as a loading control. (C) Chloroquine stabilizes EWS-Fli-1 in 293 cells. 293 cells were transfected with FLAG-EWS-Fli-1. Transfected cells were left untreated (control) or treated with 100 μM chloroquine for 12 h. The levels of FLAG-EWS-Fli-1 were examined by anti-FLAG immunoblotting. Tubulin serves as a loading control. (D) Cathepsin D degrades EWS-Fli-1, but not p53, in 293 cells. 293 cells were cotransfected with FLAG-EWS-Fli-1 and cathepsin D or empty vector. Transfected cells were left untreated or treated with 100 μM chloroquine for 12 h or 100 nM pepstatin A for 12 h. 293 cells were cotransfected with FLAG-p53 and cathepsin D or empty vector. The levels of FLAG-EWS-Fli-1 and FLAG-p53 were examined by anti-FLAG immunoblotting. Nucleolin serves as a loading control. (E) Cathepsin D degrades endogenous EWS-Fli-1 in A673 Ewing sarcoma cells. A673 cells were infected with a lentivirus vector expressing cathepsin D or an empty vector, the infected cells were selected with puromycin, and the levels of endogenous EWS-Fli-1 were examined by anti-Fli-1 C-terminus antibody immunoblotting at 4 days after infection. Nucleolin serves as a loading control. (F) Chloroquine stabilizes endogenous EWS-Fli-1 in A673 cells. A673 cells were left untreated, treated with 100 μM chloroquine for 12 h, or treated with 10 μM MG-132 for 12 h. The levels of EWS-Fli-1 were examined by anti-Fli-1 C-terminus immunoblotting. While chloroquine increased the levels of endogenous EWS-Fli-1, MG-132 had no effect on the EWS-Fli-1 protein levels, suggesting that EWS-Fli-1 turns over by a lysosomal, but not proteasomal mechanism. (G) Endogenous EWS-Fli-1 in A673 cells displays increased lysosomal location upon chloroquine treatment. A673 cells were treated with 100 μM chloroquine for 12 h or left untreated, and the whole cell extract (WCE) and lysosomal fraction were isolated. The abundance of EWS-Fli-1, LAMP2, p62/SQSTM1, and mSin3A in each fraction was determined by immunoblotting.

Targeting EWS-Fli-1 for Degradation

We employed a translational inhibitor, cycloheximide, to inhibit the new protein synthesis in Ewing sarcoma cells and analyzed the turnover of endogenous EWS-Fli-1 protein. While the EWS-Fli-1 protein levels did not exhibit any significant decrease after 24 h treatment with cycloheximide, the sensitivity of Ewing sarcoma cells to cycloheximide did not allow us to continue the cycloheximide treatment to observe the turnover of EWS-Fli-1, which is consistent with a previous report.[22] The toxicity of translation inhibitors such as cycloheximide was attributed to the depletion of ubiquitin.[23] Therefore, we exogenously expressed ubiquitin in Ewing sarcoma cells, which made the cells less sensitive to cycloheximide, and analyzed EWS-Fli-1 turnover upon prolonged cycloheximide treatment. We observed some decay of EWS-Fli-1 after 72 h treatment with cycloheximide (Figure 4A). In contrast, c-Myc [half-life = ∼30 min[24]] and cyclin D1 [half-life <30 min[25]] displayed the expected rapid decay upon cycloheximide treatment (Figure 4A). Our data suggest that endogenous EWS-Fli-1 is a relatively stable protein, which agrees with the previous findings on transfected EWS-Fli-1[26] and Fli-1.[27]

Figure 4

Targeting EWS-Fli-1 for degradation: (A) The effects of cycloheximide on endogenous EWS-Fli-1 levels in Ewing sarcoma cells. To reduce the toxicity of cycloheximide, ubiquitin was exogenously expressed in A673 and TC71 cells. Subsequently, A673 and TC71 cells were treated with 100 μg/mL cycloheximide for 24, 48, or 72 h and the levels of EWS-Fli-1, c-Myc, and cyclin D1 were determined by anti-Fli-1 C-terminus, anti-c-Myc, and anti-cyclin D1 immunoblotting, respectively. Thirty μg of whole cell lysate was loaded in each lane. (B) The effects of cytosine arabinoside on endogenous EWS-Fli-1 levels in Ewing sarcoma cells. A673 and TC71 cells were treated with 300 nM cytosine arabinoside for 48 h, and the levels of endogenous EWS-Fli-1 were determined by anti-Fli-1 C-terminus immunoblotting. Nucleolin serves as a loading control. (C) The effects of doxorubicin on endogenous EWS-Fli-1 levels in Ewing sarcoma cells. A673 and TC71 cells were treated with 60 or 120 nM doxorubicin for 72 h, and the levels of endogenous EWS-Fli-1 were determined by anti-Fli-1 C-terminus immunoblotting. Nucleolin serves as a loading control. (D) The effects of rapamycin on endogenous EWS-Fli-1 levels in Ewing sarcoma cells. A673 and TC71 cells were treated with 10 ng/mL rapamycin for 48 h, and the levels of endogenous EWS-Fli-1 were determined by anti-Fli-1 C-terminus immunoblotting. Nucleolin serves as a loading control. (E) Torin 1 reduces the EWS-Fli-1 protein levels in Ewing sarcoma. A673 and TC71 cells were left untreated or treated with 150 or 300 nM torin 1 for 24 or 48 h. The levels of endogenous EWS-Fli-1 were determined by anti-Fli-1 C-terminus immunoblotting. The experiment was repeated three times with similar results. Nucleolin serves as a loading control. (F) Luminescent monitoring of EWS-Fli-1 protein levels. 293 cells were infected with a lentivirus vector expressing luciferase–EWS-Fli-1 fusion protein, and the infected cells were selected with puromycin. The cells were treated with 100 μM chloroquine for 12 h or left untreated, and the luciferase activity was determined using the same amount of protein lysate (left). The cells were treated with 300 nM torin 1 for 24 h or left untreated, and the luciferase activity was determined using the same amount of protein lysate (right).

Since EWS-Fli-1 is a stable protein, there is a large therapeutic window to enhance its degradation. There are a few previously reported compounds that reduce EWS-Fli-1 protein levels. Rapamycin, an mTOR allosteric inhibitor, was reported to diminish EWS-Fli-1 protein levels in several Ewing sarcoma cell lines.[28] A screening for chemical compounds that inhibit the EWS-Fli-1-mediated gene expression signature identified cytosine arabinoside, which was later shown to reduce the EWS-Fli-1 protein levels in Ewing sarcoma cells.[29] The same study also demonstrated that doxorubicin, one of the standard chemotherapeutic agents for treating Ewing sarcoma, can reduce EWS-Fli-1 protein levels in Ewing sarcoma cells. However, using the conditions described in refs (28) and (29), we have been unable to reproduce the reported effects of cytosine arabinoside, doxorubicin, and rapamycin on EWS-Fli-1 protein levels (Figure 4B–D).

A transcription factor TFEB recently emerged as a master regulator of lysosomal biogenesis.[21] The activity of TFEB is controlled by cytoplasmic sequestration, which is regulated by mTOR-mediated phosphorylation.[30−32] A potent mTOR active-site inhibitor, torin 1, was shown to efficiently induce the nuclear translocation of TFEB.[30−32] In contrast, rapamycin, an mTOR allosteric inhibitor that only incompletely inhibits mTOR activity, did not induce TFEB nuclear translocation at any of the concentrations that are routinely used (10 nM – 10 μM).[30,31] We found that torin 1 treatment of Ewing sarcoma cells resulted in reduced EWS-Fli-1 protein levels (Figure 4E), suggesting a potential therapeutic utility of mTOR active-site inhibitors against Ewing sarcoma.

We also devised the fusion of EWS-Fli-1 and firefly luciferase to monitor the EWS-Fli-1 protein levels. The luciferase activity derived from the luciferase–EWS-Fli-1 fusion protein increased upon chloroquine treatment and decreased by torin 1 treatment (Figure 4F), further supporting that EWS-Fli-1 turns over by a lysosome-dependent mechanism, which can be enhanced by torin 1. In addition, this luciferase reporter can be used in the future to screen for compounds that target EWS-Fli-1 for degradation.

Since the discovery of chromosomal translocation generating the EWS-Fli-1 fusion oncogene and the pivotal role played by the transcriptional activity of EWS-Fli-1 in Ewing sarcoma, several attempts have been made to target the transcriptional activity of EWS-Fli-1. Stegmaier et al. screened a small molecule library for compounds that inhibit the gene expression signature mediated by EWS-Fli-1 in A673 Ewing sarcoma cells and identified cytosine arabinoside as an EWS-Fli-1 modulator.[29] Erkizan et al. employed surface plasmon resonance screening for compounds that bind EWS-Fli-1 and identified a small molecule that blocks the interaction of EWS-Fli-1 and RNA helicase A, leading to suppression of EWS-Fli-1 transcriptional activity and Ewing sarcoma growth.[33] Grohar et al. employed a high throughput screen (luciferase reporter screen followed by a gene signature secondary screen) to evaluate over 50,000 compounds for inhibition of EWS-Fli-1 transcriptional activity and identified mithramycin as an EWS-Fli-1 inhibitor displaying anti-Ewing sarcoma activity.[34] Boro et al. used four EWS-Fli-1 transcriptional target genes as readout to screen for compounds that abrogate EWS-Fli-1 transcriptional activity and identified a kinase inhibitor, midostaurin, which induced apoptosis in Ewing sarcoma cells.[35] While these seminal attempts provided important insights into Ewing sarcoma biology, an EWS-Fli-1-targeted therapy has not reached clinical translation and EWS-Fli-1 remains “the perfect target without a therapeutic agent”.[7] Our findings that EWS-Fli-1 turns over by a lysosome-dependent mechanism and that an mTORC1 active-site inhibitor can reduce the EWS-Fli-1 protein levels in Ewing sarcoma cells suggest a potential therapy by targeting EWS-Fli-1 for degradation.

Conclusions

Proteomic analysis of the EWS-Fli-1 interactome led to the discovery of the role for the lysosome in EWS-Fli-1 protein turnover. We demonstrated that EWS-Fli-1 is a stable protein, which provides a large therapeutic window to enhance its degradation. We found that an mTORC1 active-site inhibitor, torin 1, which stimulates the TFEB-lysosome pathway, can induce the degradation of EWS-Fli-1 in Ewing sarcoma cells. mTORC1 active-site inhibitors could target both the dependence of Ewing sarcoma on IGF-mTOR signaling and EWS-Fli-1 protein turnover and are potentially more effective than mTOR allosteric inhibitors as therapy for Ewing sarcoma.