Introduction

Ewing sarcoma (EwS) is a malignant tumor of bone and soft tissue predominantly affecting children and adolescents1. Since specific treatment options do not exist, current therapy concepts comprise local surgery combined with conventional poly-chemotherapy and irradiation1. Despite such intense conventional therapy, prognosis of patients with metastatic disease still remains poor2. Thus specific and, in particular, less toxic treatment options are urgently required.

EwS is characterized by gene fusions involving the EWSR1 gene on chromosome 22 (chr22) and various members of the ETS family of transcription factors—most commonly FLI1 on chr11 (85% of cases)1. EWSR1-FLI1 can arise either through balanced chromosomal translocations or through complex genomic breakage/fusion events3,4. Notably, EWSR1-FLI1 encodes an aberrant chimeric transcription factor, which binds DNA at ETS-binding site-like GGAA-motifs and furthermore at GGAA-microsatellites consisting of multiple sequential GGAA-motifs5. While EWSR1-FLI1 binding at single ETS binding site-like motifs in gene promoters either activates or represses gene transcription, EWSR1-FLI1 binding at GGAA-microsatellites creates de novo enhancers, whose activity correlates positively with the number of consecutive GGAA-repeats1,6,7.

Recent sequencing efforts revealed EWSR1-ETS translocations being virtually the only highly recurrent somatic mutation in EwS8,9. Although EwS is genetically well characterized, its precise cell of origin remains controversial. Transcriptome profiling and functional studies suggested that EwS may arise from mesoderm- or neural crest-derived mesenchymal stem cells10,11. Owing to this histogenic uncertainty, there is currently no bona fide genetically engineered animal model available for EwS, which hampers the development of new therapeutic strategies1,12. Like many other ligand-independent transcription factor oncoproteins, EWSR1-FLI1 also proved to be notoriously difficult to target1,13. However, the EWSR1-FLI1-induced transcriptomic signature may harbor specific changes that could be exploited therapeutically.

To explore such EWSR1-FLI1 surrogate targets, we focused in this study on the putative EWSR1-FLI1 target gene CALCB (calcitonin related polypeptide β; alias CGRP2, calcitonin gene-related peptide 2), which encodes a neuropeptide that was already described in 1987 to be highly expressed in EwS cell lines14,15. Nevertheless, its functional effects in EwS have remained unexplored until now.

The CALCB gene is located next to its homolog CALCA (calcitonin related polypeptide α) on chr11p15.2 and encodes a secretory neuropeptide composed of 37 amino acids16,17. CALCB is predominantly expressed in the central nervous system and causes potent vasodilatation18,19. Signaling of both CALCA and CALCB is mediated through G protein-coupled receptor complexes present on the cell surface. There is a variety of different receptors, formed by heterodimerization, which recognize both peptides. Most importantly they are recognized by the so called CGRP receptor, which is formed by the calcitonin receptor-like receptor (CLR, encoded by the CALCRL gene) and RAMP1 (receptor activity-modifying protein 1). RAMP1 makes the receptor complex specific for the binding of CALCA and CALCB20,21. Receptor–ligand interaction leads to G protein-mediated increase in intracellular cAMP levels22. Apart from the above-described CGRP receptor, CALCB also binds to a receptor complex consisting of RAMP1 and the calcitonin receptor (CTR, encoded by the CALCR gene), which is called AMY1 (amylin subtype 1) receptor. However, this receptor is not specific for CALCA and CALCB but is also activated by binding of islet amyloid polypeptide (IAPP). Since the biological role of AMY1 is not fully understood, and given that both CALCR and IAPP are not or only barely expressed in EwS (Supplementary Figure S1), we focused in this study on CALCB and the CGRP receptor containing CLR and RAMP121.

Here we show that CALCB is an EWSR1-FLI1 target gene highly overexpressed in EwS as compared to normal tissues and other childhood malignancies and that its high expression is likely mediated through EWSR1-FLI1 binding to an enhancer-like GGAA-microsatellite. Proteomic and functional analyses revealed that CALCB, but not CALCA, is secreted by EwS cells and that suppression of either CALCB or its receptor’s component RAMP1 significantly reduced proliferation and clonogenic/spheroidal growth of EwS cells in vitro, as well as tumor growth in vivo, which can be mimicked in vitro by application of the small molecule CGRP receptor inhibitors MK-3207 and BIBN-4096 (Olcegepant).

Materials and methods

Analysis of microarray data

The microarray datasets for cancer and normal tissues were downloaded from public repositories and processed as described previously23. Data generated on Affymetrix HG-U133Plus2.0 microarrays were normalized simultaneously by Robust Multi-chip Average (RMA) using brainarray chip description files (CDF; ENTREZg, v21) yielding one optimized probe-set per gene24,25. Accession codes of used datasets are given in Supplementary Table 1.

Cell culture and provenience of cell lines

A673, HEK-293T, and SK-PN-DW cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA; CRL-1598, CRL-1573, and CRL-2139, respectively). RDES, SK-ES1, SK-N-MC, and MHH-ES1 cells were provided by the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany). TC-71 cells were kindly provided by the Children’s Oncology Group and ES7, EW-1, EW-3, EW-7, EW-16, EW-18, EW-22, EW-24, LAP35, MIC, ORS, POE, SKNPLI, and STA-ET1 cells were provided by O. Delattre (Institute Curie, Paris, France). SB-KMS-KS1 was established in the Department of Pediatrics at the TU Munich (Munich, Germany) and described previously26. A673/TR/shEF1 cells, which contain a doxycycline (dox)-inducible short hairpin RNA (shRNA) against EWSR1-FLI1, were kindly provided by J. Alonso (Madrid, Spain)27. All cell lines were grown at 37 °C and 5% CO2 in a humidified atmosphere. RPMI 1640 medium with stable glutamine (Biochrom, Berlin, Germany), 10% tetracycline-free fetal calf serum (FCS; Biochrom), 100 U/ml penicillin (Biochrom), and 100 µg/ml streptomycin (Biochrom) was used as standard medium to grow the cells. For cell lines, which tend to grow in suspension, TPP cell culture flasks (Faust, Klettgau, Germany) were coated with 1:40 phosphate-buffered saline (PBS)-diluted (Biochrom) collagen solution (Sigma-Aldrich/Merck Millipore, Darmstadt, Germany) to enable adherent growth. Cells were routinely checked for mycoplasma infection by nested PCR. Cell line purity was confirmed by Short Tandem Repeat profiling.

RNA extraction, reverse transcription, and quantitative real-time PCR (qRT-PCR)

RNA for analysis of gene expression with qRT-PCR from cell lysates and frozen tumor tissue was extracted using the NucleoSpin RNA Kit (Macherey-Nagel, Düren, Germany). Subsequent reverse transcription was performed with the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA) utilizing 1 µg RNA per reaction and following the manufacturers’ protocols of both kits. qRT-PCRs were performed using SYBR green (Applied Biosystems, Waltham, MA, USA) with a total volume of 15 µl. cDNA was diluted 1:10 and concentration of primers was 0.5 µM. Oligonucleotides were purchased from MWG Eurofins Genomics (Ebersberg, Germany). Expression levels were determined with the CFX Connect Real time PCR Cycler (Bio-Rad Laboratories, Hercules, CA, USA) in a two-step protocol: initial enzyme activation and denaturation at 95 °C for 2 min, denaturation at 95 °C for 10 s, and annealing and extension at 60 °C for 20 s (repeating the last two steps 49 times), followed by a melting curve starting at 55 °C and increasing by 0.5 °C every 10 s until a temperature of 95 °C was reached. Expression levels were calculated according to the 2−ΔΔCT method28. RPLP0 served as housekeeping gene. Primer sequences were as follows:

RPLP0 forward, 5’-GAAACTCTGCATTCTCGCTTC-3’;

RPLP0 reverse, 5’-GGTGTAATCCGTCTCCACAG-3’;

EWSR1-FLI1 forward, 5’-GCCAAGCTCCAAGTCAATATAGC-3’;

EWSR1-FLI1 reverse, 5’-GAGGCCAGAATTCATGTTATTGC-3’;

CALCB forward, 5’-GCTCTCAGTATCTTGGTCCTG-3’;

CALCB reverse, 5’-CACATAGTCCTGCACCAGTG-3’;

RAMP1 forward, 5’-CCCAGTTCCAGGTAGACATG-3’;

RAMP1 reverse, 5’-CCAGCTTCTCCGCCATGTG-3’.

Quantification of CALCB gene expression levels in EwS cell lines

Different EwS cell lines were cultured under standard conditions. After a minimum of 48 h of culture, cells were harvested at a confluency of approximately 80%, and RNA extraction was carried out. CALCB expression was determined using qRT-PCR as described. CALCB expression levels were calculated relative to that of the A673 EwS cell line.

Quantification of EWSR1-FLI1-dependent CALCB gene expression in vivo

For analysis of in vivo CALCB expression dependence on EWSR1-FLI1, 5 × 106 A673/TR/shEF1 EwS cells, which harbor a dox-inducible shRNA against EWSR1-FLI1, were injected subcutaneously in the flanks of immunocompromised NSG (Nod/scid/gamma) mice. When tumors reached an average volume of 180 mm3, mice were randomized and either received 2 mg/ml dox (Beladox, bela-pharm, Vechta, Germany) and 5% sucrose (Sigma-Aldrich/Merck Millipore) in the drinking water (dox+) or only 5% sucrose (dox−). Mice were sacrificed 96 h after beginning of dox treatment, and tumors were collected for RNA analysis. Total RNA was extracted using the ReliaPrep miRNA Cell and Tissue Miniprep System (Promega, Madison, WI, USA). Knockdown of EWSR1-FLI1 was confirmed by qRT-PCR and proved EWSR1-FLI1 expression to be downregulated to 15% of the control (dox−). Tumor purity (>95%) was confirmed in routine histology (hematoxylin and eosin [H&E] stains). The transcriptomes of 3 tumors of each group were profiled on Affymetrix Clariom D arrays (RNA integrity number >9). Microarray data were simultaneously normalized on gene level using Signal Space Transformation RMA and Affymetrix CDF as described29.

Analysis of chromatin-immunoprecipitation followed by sequencing (ChIP-Seq) data

Publicly available data were retrieved from the Gene Expression Omnibus (GEO; GSE61944)7 and the ENCODE project30, processed as described31, and displayed in the UCSC genome browser. The used samples are available under the following accession codes:

GSM736570 ENCODE_SKNMC_hg19_DNAseHS_rep2;

GSM1517546 SKNMC.shGFP96.FLI1;

GSM1517555 SKNMC.shFLI196.FLI1;

GSM1517548 SK-N-MC_shGFP_96h_H3K4me1;

GSM1517557 SK-N-MC_shFLI1_96h_H3K4me1;

GSM1517547 SKNMC.shGFP96.H3K27ac;

GSM1517556 SKNMC.shFLI196.H3K27ac;

GSM1517569 A673.shGFP48.FLI1;

GSM1517572 A673.shFLI148.FLI1;

GSM1517571 A673.shGFP96.H3.k27ac;

GSM1517574 A673.shFLI196.H3K27ac.

Dual luciferase reporter assays

A 359 base pairs (bp) fragment around the CALCB-associated GGAA-microsatellite was cloned from two different EwS cell lines (TC-71 and MHH-ES1) by PCR into the pGL3 Luciferase Reporter Vector (Promega) using the In-Fusion HD Cloning Kit (Clontech, Takara Bio USA, CA, USA). The DNA from the cell lines was extracted using the NucleoSpin Tissue genomic DNA Prep Kit (Macherey-Nagel) and digested with the restriction enzymes Eco-RV and SphI (New England Biolabs [NEB], Ipswich, MA, USA), followed by a fragment separation by gel electrophoresis, and fragment purification of the band at 8200 bp using the NucleoSpin Gel and PCR Clean-up Kit (Macherey-Nagel). A touch-down PCR was performed with 100 ng of the purified DNA fragment using the Infusion-Primers:

forward: 5’-ctagcccgggctcgagGAGCCCTTTAGTATCCCCTTTG-3’;

reverse: 5’-gatcgcagatctcgagACCCTTGTACTAACATGCTTCG-3’,

(MWG Eurofins Genomics) at a concentration of 0.5 µM, and the components of the Go Taq Hot Start Polymerase Kit (Promega) according to the manufacturer’s instructions. The thermal protocol was as follows: activation at 95 °C for 2 min, followed by 20 cycles of 98 °C for 10 s, 59-49 °C for 30 s (decreasing 0.5 °C in each cycle), and 72 °C for 1 min, finalizing at 72 °C for 5 min. The PCR product was separated by gel electrophoresis, and the desired fragment of 359 bp was purified with the NucleoSpin Gel and PCR Clean-up Kit (Macherey-Nagel). The reaction with the In-Fusion HD Cloning Kit (Clontech) was initiated using 10 ng of the linearized pGL3 Luciferase Reporter Vector (Promega), digested with XhoI (NEB) and prepared according to the manual, and 20 ng of the purified DNA fragment. The reaction was incubated at 50 °C for 15 min, stopped on ice for 5 min, and Escherichia coli Stellar competent cells (Clontech) were transformed with the DNA of the infusion mixture. Bacteria were grown on agar plates supplemented with ampicillin at a final concentration of 100 µg/ml (Sigma-Aldrich/Merck Millipore) overnight at 37 °C. Colonies were picked the next morning and checked for clones containing the plasmid with the insert by colony PCR. Positive clones were incubated overnight in LB broth (Miller)-medium with 100 µg/ml ampicillin (Sigma-Aldrich/Merck Millipore) at 37 °C, and plasmids were purified using the PureYield Plasmid Midiprep System 2 (Promega). The identity of the fragment was validated by Sanger sequencing (Eurofins GATC Biotech, Konstanz, Germany).

For the luciferase reporter assay, 2 × 105 A673/TR/shEF1 EwS cells, harboring a dox-inducible shRNA against EWSR1-FLI1, were plated in a well of a 6-well plate (TPP, Faust) in 1.8 ml of growth medium and transfected with the microsatellite-containing pGL3-luc vector and Renilla pGL3-Rluc vector (ratio 100:1) using Lipofectamine LTX with Plus Reagent (Thermo Fisher Scientific). Transfection medium was replaced by media with/without dox (1 μg/ml; VWR/Merck, Radnor, PA, USA) 4 h after transfection. After 72 h, the cells were lysed and assayed with a dual luciferase assay system (Berthold Technologies, Bad Wildbad, Germany). Firefly luciferase activity was normalized to Renilla luciferase activity.

Gene set enrichment analysis (GSEA)

To identify gene signatures and biological processes associated with CALCB in normalized gene expression data from 166 primary EwS32, GSEA was performed on lists of genes ranked by their correlation coefficient with CALCB (MSigDB, c2.cpg.v6.2). GSEA was carried out with 1000 permutations in default settings33.

Generation of cell lines with dox-inducible shRNAs

For generation of EwS cell lines with dox-inducible constructs (here in A673 and RDES), either a non-targeting negative control shRNA (MWG Eurofins Genomics) or specific shRNAs targeting CALCB or RAMP1 (MWG Eurofins Genomics) were cloned in the pLKO-Tet-on-all-in-one system (Addgene plasmid # 21915, Cambridge, MA, USA) as described previously34. Lentivirus production was performed in HEK-293T cells. A673 and RDES EwS cells were infected with the respective lentiviruses and selected with 1.5 µg/ml puromycin (Invivogen, Toulouse, France). Single-cell cloning was performed, and knockdown efficacy of individual clones was assessed by qRT-PCR 48 h after addition of dox (1 µg/ml; VWR/Merck). The shRNA target sequences were as follows:

shControl, 5’-CAACAAGATGAAGAGCACCAA-3’;

shCALCB1, 5‘-AAGGAATGAAACTGAATGCAA-3’;

shCALCB4, 5’-AACCTTGGTGATGCATTACAA-3’;

shRAMP1_3, 5’-GCGCACTGAGGGCATTGTGTA-3’;

shRAMP1_4, 5’-TGCCTGCCAGGAGGCTAACTA-3’.

Proliferation assays

A total of 1–5 × 105 cells were seeded in 6-well plates (TPP, Faust) in 1.5 ml of standard growth medium. Gene knockdown was induced by addition of 1 µg/ml dox (VWR/Merck) to the growth medium (refreshed every 48–72 h) of cells harboring an inducible shRNA against CALCB or by serial transfections with 25 nM small interfering RNA (siRNA) directed against CALCB (Hs_CALCB_1 FlexiTube siRNA or Hs_CALCB_4 FlexiTube siRNA, QIAGEN, Hilden, Germany) or a scrambled control siRNA (MISSION siRNA Universal Negative Control #1, Sigma-Aldrich/Merck Millipore) following the manufacturer’s handbook of the transfection reagent HiPerfect (QIAGEN). After 2–4 h, 3 ml of standard growth medium was added to prevent toxic effects of the transfection reagent. After 24 h, the medium was exchanged and after another 24 h a second transfection was performed. Cell counts were determined by using standardized hemocytometers (C-chips, Biochrom) and Trypan-blue (Sigma-Aldrich/Merck Millipore) exclusion 72 h after seeding in the short-term proliferation assay and 6–9 days after seeding in the long-term proliferation assay.

Colony-forming assays

A673 and RDES EwS cells harboring a dox-inducible shRNA construct against CALCB or RAMP1 or RDES wild-type cells were seeded at concentrations of 100–1 000 cells per well of 12-well plates and grown in standard culture medium for 12–14 days. Cells were treated with/without 1 µg/ml dox (VWR/Merck) and RDES wild-type cells were serially transfected with a non-targeting siRNA or siRNAs against CALCB (as described above). Twice gently PBS-washed (Biochrom) colonies were stained with 500 µl crystal violet (Sigma-Aldrich/Merck Millipore) and the number of colonies was quantified using ImageJ.

Sphere-formation assays

For analysis of three-dimensional (3D) sphere formation, A673 and RDES EwS cells harboring a dox-inducible shRNA construct against CALCB or RAMP1 were seeded at a density of 1000 cells per well of ultra-low attachment 96-well plates (Corning, NY, USA) in 80 µl standard cell culture medium with/without dox (1 µg/ml; VWR/Merck). The culture medium was refreshed by adding 10 µl medium with/without dox on top every second day. Spheroidal growth was monitored for 14 days. Thereafter, phase-contrast imaging and morphological analyses of spheres were carried out with an inverted Zeiss Axiovert 25 microscope (Jena, Germany) equipped with a Zeiss Axiocam 105 color camera (Aptina CMOS Color Sensor, square pixels of 2.2 μm side length, 2560 × 1920 pixel resolution). Sphere numbers and diameters were analyzed with ImageJ.

Analysis of tumor growth in vivo

A total of 2.5 × 106 A673 EwS cells harboring a dox-inducible shRNA construct against CALCB or RAMP1 or a non-targeting control shRNA (shControl) were injected subcutaneously in NSG mice. After 10–14 days, when tumors were first palpable, mice were randomized and thereafter received either 2 mg/ml dox (bela-pharm) dissolved in sterile water containing 5% sucrose (Sigma-Aldrich/Merck Millipore) (dox+) or sterile water with 5% sucrose alone (dox−). Tumor growth was monitored with a caliper every other day and mice were sacrificed by cervical dislocation when the tumors exceeded an average diameter of 15 mm (prior start of the experiment defined as “event”). Experiments were approved by the government of Upper Bavaria and conducted in accordance with ARRIVE guidelines and recommendations of the European Community (86/609/EEC) and UKCCCR (guidelines for the welfare and use of animals in cancer research).

Small-molecule inhibitor assays

A673 and A673/TR/shRAMP1_4 EwS cells, the latter harboring a dox-inducible shRNA against RAMP1, were seeded at a density of 1500 cells per well of a 96-well plate (TPP, Faust) in 50 µl standard growth medium with/without dox (1 µg/ml; VWR/Merck). After 24 h of incubation, treatment was started by addition of 50 µl standard growth medium containing either different concentrations of the CGRP receptor inhibitor MK-3207 (AdooQ Bioscience, Irvine, CA, USA) dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich/Merck Millipore) or the corresponding concentration of DMSO alone and dox refreshment for dox+ wells. After 72 h, read-out was performed by addition of 20 µl of 1:10 dissolved Resazurin (1 g/l, Sigma-Aldrich/Merck Millipore) to the cells and measurement of fluorescence with a plate reader (Thermo Fisher Scientific) after 7 h of incubation.

For analysis of two-dimensional (2D) colony-formation capacity under inhibitor treatment, A673 and RDES EwS cells were seeded at a density of 100 cells per well of 12-well plates (TPP, Faust) in 1 ml culture medium. Forty-eight hours after seeding, inhibitors were added at final concentrations of 100 µM for BIBN-4096 (Olcegepant; R&D systems, Minneapolis, MN, USA) and 20 µM for MK-3207 (AdooQ Bioscience). DMSO (Sigma-Aldrich/Merck Millipore) served as control. After 1–3 weeks, colonies were gently washed twice with PBS (Biochrom) and stained with 500 µl crystal violet solution (Sigma-Aldrich/Merck Millipore). Colonies were photographed, and the number of colonies was counted using Image J.

For analysis of 3D sphere-formation capacity under inhibitor treatment, A673 and RDES EwS cells were seeded at a density of 1000 cells per well in 80 µl culture medium in wells of 96-well ultra-low attachment culture plates (Corning). After 24 h of incubation, 20 µl culture medium containing either inhibitor or DSMO (control; Sigma-Aldrich/Merck Millipore) was added to the wells resulting in a final concentration of 100 µM BIBN-4096 (Olcegepant; R&D systems) or 20 µM MK-3207 (AdooQ Bioscience), respectively. After 14 days, spheres were photographed, and their number and size were analyzed using ImageJ.

Human samples and immunohistochemistry (IHC)

Available tissue microarrays (TMA) of primary EwS tumors containing 2 cores of each sample, with a diameter of 1 mm, as well as internal controls were stained for CALCB. Analysis were carried out with approval from LMU Munich ethics committee.

For IHC, 4-μm sections were cut and antigen retrieval was carried out by heat treatment using target unmasking fluid (PanPath, Budel, Netherlands). Slides were incubated for 60 min at room temperature with a rabbit polyclonal anti-CALCB antibody (bs-0791R, Bioss Antibodies Inc., MA, USA; dilution 1:120). Then slides were incubated with a secondary anti-rabbit IgG antibody (Vectastain ABC-Kit Elite Universal, Vector laboratories, Burlingame, CA, USA) followed by target detection using DAB plus (Agilent Technologies, Santa Clara, CA, USA). Counterstaining was performed with Hematoxylin Gill’s Formula (Vector). Intensity of CALCB staining was scored independently by two researchers on a scale from 0 to 2 (0 = majority of the cells is negative for CALCB staining, 1 = majority of the cells shows moderate CALCB staining, and 2 = majority of the cells shows strong CALCB staining). Specificity of the anti-CALCB-antibody was assured by determination of immunoreactivity scores (IRSs) using the Remmele and Stegner scoring system35 in 6–11 representative high-power fields (×40) per xenograft with/without qRT-PCR-confirmed knockdown of CALCB.

CD31 staining and evaluation of microvessel density

For CD31 staining, 4-µm sections of formalin-fixed and paraffin-embedded tumor tissue derived from EwS xenografts in mice were cut and heat treated using the Target Retrieval Solution (Agilent Technologies). Thereafter, tissue slides were stained with a primary monoclonal rat anti-CD31-antibody (DIA-310, Dianova, Hamburg, Germany; dilution 1:150, 60 min incubation at room temperature). As secondary antibody, a biotinylated and mouse-absorbed anti-rat-IgG-antibody (BA 4001, Vector; dilution 1:100) was used. After Streptavidin horseradish peroxidase (Novocastra Laboratories, Newcastle upon Tyne, United Kingdom) treatment, DAB plus (Agilent Technologies) was used for target detection. The slides were counterstained with Hematoxylin Gill’s Formula (Vector).

For evaluation of the microvessel density in the CD31-stained slides, the Chalkley-grid method was used36. To this end, the number of overlaps of a CD31-positive cell with a dot of the Chalkley-grid in each quarter of the grid in four independent regions of the CD31-stained slide was counted, and the mean vessel density of the tumor was extrapolated.

Evaluation of mitosis and necrosis in xenografts

The average number of mitotic cells per high-power field was determined in 22 representative A673/TR/shCALCB xenografts and 10 A673/TR/shRAMP1 xenografts in H&E-stained slides with/without knockdown of CALCB or RAMP1, respectively. The average area of necrotic tissue over the total tissue area as well as the number of mitoses per tumor sample was determined by a data-blinded resident pathologist through evaluation of 10 high-power fields (×40) per slide.

Mass spectrometric analyses

A673 EwS cells were seeded at a density of 4 × 106 cells per T150 flask (TPP, Faust) in 20 ml of standard culture medium. After 48 h, the supernatant from the cells was removed, and the cells were washed with PBS (Biochrom) twice. Thereafter, the cells were grown for further 24 h in 20 ml Opti-MEM (Thermo Fisher Scientific) only. Afterwards, the supernatants were collected and immediately frozen at −80 °C until the mass spectrometry was performed at the Antibody Engineering and Proteomics facility of the Immunity and Infection Research Centre (Jack Bell Bldg., Vancouver, Canada). For mass spectrometric analysis, samples were lyophilized and resuspended in 50mM ammonium bicarbonate. In total 200 μg of protein was reduced and alkylated using 10 mM dithiothreitol (Thermo Fisher Scientific) and 100 mM iodoacetamide (Sigma-Aldrich/Merck Millipore), respectively. Next, samples were digested using 20 ng/μl trypsin (NEB) for 18 h at 37 °C. Samples were separated in a Nano-HPLC (NanoLC-2D, Eksigent, Sciex, CA, USA) using a C18 column and a gradient composed of solvent A (5% acetonitrile) and solvent B (95% acetonitrile). The program was: 5% acetonitrile for 5 min, 5–100% for 50 min, and 100% for 10 min. Eluted samples were spotted (Eksigent) on a 384-well plate and 1 μl of the matrix α-cyano-4-hydroxycinnamic acid (10 mg/ml in 50% acetonitrile and 0.1% trifluoroacetic acid) was added. The mass spectrometric analysis was performed on a MALDI-TOF/TOF 4800 (Sciex) using positive mode. The data were analyzed with the Trans-Proteomic Pipeline (Seattle Proteome Center, WA, USA). To identify the peptide profile, a full-length synthetic CALCB polypeptide (Peptides & Elephants, Henningsdorf, Germany) was processed as a standard and analyzed. The peptide 106SNFVPTNVGSK116 (m/z 1149.5898, monoisotopic) was used to identify the presence of CALCB in the samples.

Results

CALCB is an EWSR1-FLI1 target gene highly but heterogeneously expressed in EwS

In the search of potential EWSR1-FLI1 surrogate targets, we analyzed publicly available gene expression microarray data comprising 71 normal tissue types and 50 tumor entities. Thereby, we identified CALCB as being highly overexpressed in EwS compared to most tumor entities and all normal tissues except for trigeminal ganglia (Fig. 1a, b; Supplementary Fig. S2). The high expression of CALCB in EwS was validated by IHC staining of a TMA of primary EwS tumors, of which 44% (39/89) displayed a high and 37% (33/89) an intermediate IRS (2 or 1, respectively) for CALCB expression (Supplementary Fig. S3).

Fig. 1: CALCB is an EWSR1-FLI1 target gene specifically expressed in Ewing sarcoma (EwS).
figure 1

a Analysis of CALCB expression in EwS (n = 50) and normal tissues (71 tissue types, n = 929 samples). Data are represented as box plots in log2 scale. Horizontal bars indicate median expression levels, boxes the interquartile range. Whiskers indicate the 2.5th and 97.5th percentiles, respectively. Unpaired two-tailed Student’s t test. b Analysis of CALCB expression in EwS (n = 50) and different (pediatric) tumors (49 types, n = 1 699 samples). Publicly available microarray data are represented as dot plots in log2 scale with mean and SEM. Each dot represents one sample. The number of samples is given in parentheses. EwS highlighted in blue color. c Time-course analysis of CALCB and EWSR1-FLI1 expression in A673/TR/shEF1 EwS cell harboring a dox-inducible shRNA against EWSR1-FLI1 by qRT-PCR in vitro after dox application. Given are mean expression levels and SEM (n = 7). d Analysis of CALCB expression in xenografts derived from A673/TR/shEF1 cells with/without dox treatment for 96 h in vivo. Gene expression levels were determined by Affymetrix Clariom D microarrays as previously described29. Expression levels are shown in log2 scale; horizontal bars indicate mean expression level (n = 3); unpaired two-tailed Student’s t test. e Analysis of CALCB expression in a published dataset (GSE64686)40 with ectopic dox-inducible EWSR1-FLI1 expression in human embryoid bodies. Data were generated on Affymetrix HG-U133Plus2.0 microarrays and normalized simultaneously by RMA and brainarray CDF (ENTREZg; v21). Horizontal bars indicate mean expression levels; unpaired two-tailed Student’s t test. **P ≤ 0.01; ***P ≤ 0.001

Full size image

This EwS-specific expression pattern suggested a potential regulatory relationship between EwS-specific EWSR1-ETS fusion oncogenes and CALCB. To test this hypothesis, we performed time-course knockdown experiments of EWSR1-FLI1 in the EwS cell line A673/TR/shEF1, which harbors a dox-inducible shRNA against EWSR1-FLI1 and measured the expression of EWSR1-FLI1 and CALCB at different time points after start of dox treatment (0–96 h) by qRT-PCR. The results showed that the expression of CALCB is tightly linked to that of EWSR1-FLI1 (Fig. 1c), which was confirmed in xenografts derived from A673/TR/shEF1 cells in vivo (Fig. 1d). Conversely, ectopic expression of EWSR1-FLI1 in human embryoid bodies strongly induced CALCB expression (Fig. 1e).

To further assess this regulatory relationship, we explored available FLI1 ChIP-seq data from two EwS cell lines (A673 and SK-N-MC) and found strong EWSR1-FLI1 binding at intron 5 of the longest isoform (isoform 3) of the CALCB gene, which mapped to a GGAA-microsatellite that showed epigenetic characteristics of an active enhancer (Fig. 2a). Knockdown of EWSR1-FLI1 in both cell lines abolished the EWSR1-FLI1 signal at this GGAA-microsatellite and markedly reduced the signals for acetylated H3K27 (H3K27ac), indicating abrogated enhancer activity upon EWSR1-FLI1 silencing (Fig. 2a).

Fig. 2: CALCB expression is regulated through EWSR1-FLI1 binding to a nearby enhancer-like GGAA-microsatellite.
figure 2

a Integrative genomic view of published ChIP-seq7 and DNAse-seq data30 of the CALCB locus. Data were generated in A673 and SK-N-MC Ewing sarcoma cells, stably transfected with either shRNA targeting GFP (shGFP; negative control) or EWSR1-FLI1 (shEF1). The blue box marks the location of the CALCB gene-associated GGAA-microsatellite, consisting of 16 repeats of the GGAA-motif in the human reference genome (hg19). b Luciferase reporter assays in A673/TR/shEF1 cells with/without knockdown of EWSR1-FLI1 (Dox+/−) 72 h after transfection with plasmids containing a 359-bp fragment around the CALCB-associated GGAA-microsatellite as displayed in a. Data are presented as mean (n = 3–5) and SEM; unpaired two-tailed Student’s t test. **P ≤ 0.01; ***P ≤ 0.001

Full size image

To confirm its EWSR1-FLI1-dependent enhancer activity, we cloned a 359-bp fragment containing this GGAA-microsatellite from two EwS cell lines (TC-71 and MHH-ES1) into the pGL3 luciferase reporter vector and performed reporter assays in A673/TR/shEF1 cells with/without silencing of EWSR1-FLI1. In these assays, we observed strong enhancer activity of the GGAA-microsatellite, which was significantly diminished upon EWSR1-FLI1 knockdown (Fig. 2b). In accordance with the higher number of consecutive GGAA-repeats and higher CALCB expression levels in TC-71 EwS cells (12 repeats) as compared to MHH-ES1 EwS cells (9 repeats), we noted a higher enhancer activity of the GGAA-microsatellite derived from TC-71 as compared to the microsatellite derived from MHH-ES1 in luciferase assays (Fig. 2b, Supplementary Fig. S4). Collectively, these data provide evidence that CALCB is a direct EWSR1-FLI1 target gene, whose high but heterogeneous expression in EwS is regulated by EWSR1-FLI1 binding to an intronic, polymorphic, and enhancer-like GGAA-microsatellite.

CALCB expression in primary EwS correlates with proliferation signatures

To obtain first clues on the potential functional role of CALCB in EwS, we performed GSEA on CALCB co-expressed genes in a transcriptome dataset of 166 primary EwS32. GSEA revealed that CALCB is co-expressed with other EWSR1-FLI1 target genes (ZHANG_TARGETS_OF_EWSR1-FLI1_FUSION)37 and with gene signatures involved in stemness and proliferation (Fig. 3a, b).

Fig. 3: CALCB expression in primary Ewing sarcoma (EwS) correlates with proliferation signatures.
figure 3

a Heatmap of CALCB correlated genes (rPearson ≥ ∣0.3∣) in 166 primary EwS tumors. b Results of the gene set enrichment analysis on the ranked list of CALCB correlated genes as in Fig. 3a. NES normalized enrichment score

Full size image

CALCB signaling in EwS cells contributes to growth of EwS

To test the bioinformatic predictions from our GSEA in primary EwS, we carried out several functional experiments in EwS models. Mass spectrometric analysis showed that CALCB was readily detectable in FCS-free cell culture supernatants conditioned by A673 EwS cells, whereas it was not detectable in FCS-free cell culture medium not conditioned by EwS cells (Supplementary Table 2), suggesting that CALCB is indeed secreted by EwS cells. Notably, in accordance with the low expression levels of CALCA in primary EwS cells (Supplementary Fig. S1), CALCA was not detectable in cell culture supernatants of EwS cells (Supplementary Table 2).

To investigate the functional role of CALCB in EwS, we performed RNA interference experiments in two EwS cell lines (RDES and A673), which showed relatively high or moderate CALCB expression levels as compared to 20 other EwS cell lines (Supplementary Fig. S4). While the short-term knockdown of CALCB for 3 days had no effect on cellular proliferation (Supplementary Fig. S5), its long-term knockdown for 6–9 days significantly reduced proliferation and clonogenic growth in vitro (Fig. 4a, b). As no differences in the relative number of dead cells was detectable by Trypan-blue-labeled cell counting, the reduced cell count appeared to be not mediated by an increase in cell death (Fig. 4a).

Fig. 4: Knockdown of CALCB or RAMP1 inhibits proliferation of Ewing sarcoma (EwS) cells in vitro.
figure 4

a Analysis of cell growth and cell death of A673 and RDES EwS cells with/without siRNA- or shRNA-mediated knockdown of CALCB. Left panel A673 and RDES: Given is the mean of relative cell count compared to Co (Control), which either received according doses of a non-targeting siControl or did not receive dox in assays with dox-inducible shRNAs (n = 3–5). SEM and unpaired two-tailed Student’s t test of relative total cell count. Right panel A673 and RDES: Knockdown of CALCB was verified by qRT-PCR. Given is the mean gene expression and SEM; unpaired two-tailed Student’s t test. b Colony-forming assays of A673 and RDES EwS cells with/without siRNA- or dox-induced shRNA-mediated knockdown of CALCB or RAMP1. Mean colony number and SEM normalized to control, which either received according doses of a non-targeting siControl or did not receive dox in assays with dox-inducible shRNAs (n = 3–6). Unpaired two-tailed Student’s t test. Representative images of each condition are shown. Knockdown of CALCB and RAMP1 were verified by qRT-PCR and are displayed in Fig. 4a and Supplementary Fig. S6. c Sphere-formation assays of A673 and RDES EwS cells with/without dox-induced shRNA-mediated knockdown of CALCB and RAMP1. Sphere index was calculated by addition of diameter of all existing spheres in one well divided by diameter of spheres in the control well. Mean and SEM (n = 3–6); unpaired two-tailed Student’s t test. n.s. P > 0.05; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001

Full size image

Interestingly, knockdown of RAMP1—the crucial component of the CALCB receptor complex—phenocopied the effect of CALCB knockdown in clonogenic growth assays (Fig. 4b, Supplementary Fig. S6) as well as in 3D sphere-formation assays (Fig. 4c).

We next investigated whether knockdown of either gene could alter growth of xenografted EwS cells in vivo. To this end, we injected A673 cells, harboring a dox-inducible shRNA against either CALCB or RAMP1, subcutaneously in NSG mice. When tumors were palpable, we induced the knockdown of the respective gene by addition of dox to the drinking water. In both settings, knockdown of the corresponding gene significantly delayed tumor growth, which led to a delayed achievement of a mean tumor diameter of 15 mm, the defined termination criteria, and therefore allowed a prolonged survival of the animals (Fig. 5a, b). However, dox treatment of mice carrying tumors with dox-inducible expression of a non-targeting control shRNA did not alter tumor growth compared to mice not receiving dox (data not shown), as also described previously for other xenografted EwS cell lines31. Although CALCB and RAMP1 were knocked down to low levels as confirmed by qRT-PCR of tumor tissue (Fig. 5a, b), the growth-inhibiting effect was more pronounced in the group of the RAMP1 knockdown. Histological analysis of the xenografts revealed significantly higher mitotic activity in tumors without CALCB or RAMP1 knockdown, respectively, compared to tumors with shRNA-induced knockdown of either gene (Fig. 5c). In contrast, no differences in tumor necrosis was observed (Fig. 5d).

Fig. 5: Knockdown of CALCB or RAMP1 inhibits proliferation of Ewing sarcoma (EwS) cells in vivo.
figure 5

a Left panel: Analysis of tumor growth of A673 EwS cells with/without dox-induced knockdown of CALCB in NSG mice (n = 33). Event was defined as average diameter of 15 mm. Event-free survival time of mice was analyzed by the Kaplan–Meier method and a log-rank test. Right panel: Knockdown of CALCB in the tumors of dox-treated mice was verified by qRT-PCR. Given are mean normalized gene expression levels and SEM; unpaired two-tailed Student’s t test. b Left panel: Analysis of tumor growth of A673 EwS cells with/without dox-induced knockdown of RAMP1 in NSG mice (n = 10). Event was defined as average diameter of 15 mm. Event-free survival time of mice was analyzed by the Kaplan–Meier method and a log-rank test. Right panel: Knockdown of RAMP1 in the tumors of dox-treated mice was verified by qRT-PCR. Given are mean normalized gene expression levels and SEM; unpaired two-tailed Student’s t test. c Histological analysis of the number of mitoses in tumor tissue of EwS xenografts with/without dox-induced knockdown of CALCB or RAMP1. Given is the mean number of mitoses and SEM per high-power filed (HPF) of 22 representative A673/TR/shCALCB xenografts shown in a and 10 A673/TR/shRAMP1 xenografts shown in b. Unpaired two-tailed, Student’s t test. d Histological analysis of necrosis in tumor tissue of EwS xenografts with/without dox-induced knockdown of CALCB or RAMP1. Given is the average percentage of necrotic area and SEM of 22 representative A673/TR/shCALCB xenografts shown in a and 10 A673/TR/shRAMP1 xenografts shown in b. Unpaired two-tailed, Student’s t test. n.s. P > 0.05; **P ≤ 0.01; ***P ≤ 0.001

Full size image

Taken together, these data suggest that CALCB is a secreted peptide in EwS and that the CALCB/RAMP1 axis promotes growth of EwS cells.

Pharmacological inhibition of the CALCB/RAMP1 axis decreases growth of EwS cells

To test whether the CALCB/RAMP1 axis could also be exploited therapeutically in EwS, we treated EwS cells with the small molecule CGRP receptor inhibitor MK-3207 for 3 days and quantified cell viability with a Resazurin assay. For these assays, we used dox-inducible CALCB or RAMP1 knockdown EwS cells and applied increasing doses of MK-3207. We observed a dose-dependent reduction of cell viability (Fig. 6a), which could be partially abrogated by knockdown of RAMP1—the central component of the inhibitor’s target structure (Fig. 6b). These data suggest that, albeit relatively high doses of MK-3207 were applied to reduce viability of EwS cells, its effect was specific for the CALCB/RAMP1 axis. To validate these findings, we performed colony- and sphere-formation assays under MK-3207 treatment and replicated these experiments with another small molecule CGRP inhibitor (Olcegepant, BIBN-4096) (Fig. 6c, d). In both assays and for both inhibitors, we noted a significant reduction of 2D colony-formation and 3D sphere-formation capacity of EwS.

Fig. 6: Blockage of the calcitonin gene-related peptide (CGRP) receptor by small molecule inhibitors mimics the effect of CALCB and RAMP1 knockdown in vitro.
figure 6

a Analysis of cell viability (normalized Resazurin signal to DMSO control) in A673 Ewing sarcoma (EwS) cells treated for 72 h with the indicated concentrations of MK-3207. The graph shows the dose-dependent relative Resazurin signal. Data are represented as mean and SEM (n = 12); unpaired two-tailed Student’s t test. b Comparison of relative Resazurin signal of A673 cells carrying a dox-inducible shRNA against RAMP1 treated with 150 µM of MK-3207 with/without knockdown of RAMP1 by additional addition of 1 µg/ml dox to the growth medium. Data are represented as mean and SEM (n = 3); unpaired two-tailed Student’s t test. c Analysis of colony-forming capacity of A673 (left panel) and RDES (right panel) EwS cells under treatment with the small molecule CGRP receptor inhibitors MK-3207 (20 µM) or BIBN-4096 (Olcegepant; 100 µM). DMSO served as control for treatment. Representative images of the colonies are shown below. Data are represented as mean and SEM (n = 3); unpaired two-tailed Student’s t test. d Analysis of sphere-formation capacity of A673 (left panel) and RDES (right panel) EwS cells under treatment with the small molecule CGRP receptor inhibitors MK-3207 (20 µM) or BIBN-4096 (Olcegepant; 100 µM). DMSO served as control for treatment. Sphere index was calculated by addition of diameters of all existing spheres in one well divided by diameter of spheres in the control well. Data are represented as mean and SEM (n = 3); unpaired two-tailed Student’s t test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001

Full size image

Together, these data provide further evidence for a functional role of the CALCB/RAMP1 axis in growth of EwS, which could potentially be exploited therapeutically.

Discussion

Albeit EwS is genetically well characterized, standard therapy still comprises only unspecific cytotoxic approaches. As it is not possible to directly target the action of EWSR1-ETS transcription factors, current strategies to find new and more specific treatment options include the investigation of target genes of EWSR1-ETS and determination of their potential as possible new therapy targets. To this end, we investigated the potential of the CALCB/RAMP1 axis as new target for treatment of EwS and explored its functional role in EwS cells by combining a series of in situ, in vitro, and in vivo experiments. We found that CALCB is a secreted peptide in EwS that shows a highly specific expression pattern among malignant and normal tissues. The high expression of CALCB in EwS is likely driven by EWSR1-FLI1 binding to a GGAA-microsatellite at the CALCB locus. Since this GGAA-microsatellite is transformed into a de novo enhancer upon EWSR1-FLI1 binding but does not show enhancer activity in the absence of EWSR1-FLI1, we speculate that different mechanisms may operate in other tissue types such as trigeminal ganglia to upregulate CALCB expression, which remain to be elucidated.

In our long-term knockdown experiments, we observed that silencing of CALCB or RAMP1 reduced growth of EwS cells in vitro and in vivo. To the best of our knowledge, this is the first report of a functional role of CALCB in growth of cancer cells to date. However, we noted a weaker effect of the CALCB knockdown as compared to that of RAMP1 on tumor growth in in vivo experiments, which may be caused by residual CALCB expression in the EwS cells (around 5% remaining expression), or alternatively by circulating murine Calcb, which might have compensated at least in part for the loss of human CALCB.

In our drug–response assays, we found that inhibition of CGRP receptors with two different small molecules had a similar antiproliferative effect on EwS cells to that of the CALCB or RAMP1 knockdown. CGRP receptor inhibitors already showed high efficacy in the treatment of migraine, which is presumably caused via CALCA- or CALCB-mediated vasodilatation in the vicinity of the trigeminal ganglia—the only normal tissue type with physiologically high CALCB expression levels found in our analyses. We speculate that repurposing and further optimization of such “migraine drugs” (e.g., newly developed antibodies directed against CALCB or RAMP1)38 could perhaps offer novel therapeutic options for EwS patients in the future. As we did not observe differences in density of tumor-associated blood vessels quantified by staining for murine CD31 in IHC of our xenograft experiments (Supplementary Fig. S7), we assume that the growth-promoting effect of the CALCB/RAMP1 axis is conferred via different mechanisms than one would expect from the vasodilatory effect of CALCB known from the literature39.

Collectively, we identified CALCB as a highly specifically expressed EWSR1-FLI1 target gene encoding a secreted peptide that promotes growth of EwS cells, and show that targeting the CALCB/RAMP1 axis in EwS may offer a new therapeutic approach. Future studies will have to dissect the precise downstream signaling and how the CALCB/RAMP1 axis promotes proliferation of EwS cells to further explore its therapeutic potential.