Small Molecule Therapeutics

Afatinib Is a New Therapeutic Approach inChordoma with a Unique Ability to Target EGFRand BrachyuryPaola Magnaghi1, Barbara Salom1, Liviana Cozzi1, Nadia Amboldi1, Dario Ballinari1,Elena Tamborini2, Fabio Gasparri1, Alessia Montagnoli1, Laura Raddrizzani1,Alessio Somaschini1, Roberta Bosotti1, Christian Orrenius1, Fabio Bozzi2, Silvana Pilotti2,Arturo Galvani1, Josh Sommer3, Silvia Stacchiotti2, and Antonella Isacchi1

Abstract

Chordomas are rare bone tumors with no approved therapy.These tumors express several activated tyrosine kinase receptors,which prompted attempts to treat patients with tyrosine kinaseinhibitors. Although clinical benefit was observed in phase IIclinical trials with imatinib and sorafenib, and sporadically alsowith EGFR inhibitors, therapies evaluated to date have shownmodest activity. With the goal of identifying new drugs withimmediate therapeutic potential for chordoma patients, we col-lected clinically approved drugs and other advanced inhibitors ofMET, PDGFRb, and EGFR tyrosine kinases, and assessed theirantiproliferative activity against a panel of chordoma cell lines.Chordoma cell lines were not responsive to MET and PDGFRbinhibitors. U-CH1 and UM-Chor1 were sensitive to all EGFRinhibitors, whereas the remaining cell lines were generally insen-

sitive to these drugs. Afatinib was the only EGFR inhibitor withactivity across the chordoma panel. We then investigated themolecular mechanisms behind the responses observed and foundthat the antiproliferative IC50s correlate with the unique ability ofafatinib to promote degradation of EGFR and brachyury, anembryonic transcription factor considered a key driver of chor-doma. Afatinib displayed potent antitumor efficacy in U-CH1,SF8894, CF322, and CF365 chordoma tumor models in vivo. Inthe panel analyzed, high EGFR phosphorylation and low AXL andSTK33 expression correlated with higher sensitivity to afatinib anddeserve further investigation as potential biomarkers of response.These data support the use of afatinib in clinical trials and providethe rationale for the upcoming European phase II study on afatinibin advanced chordoma. Mol Cancer Ther; 17(3); 603–13. �2017 AACR.

IntroductionChordomas are primary malignant bone tumors that arise

along the axial skeleton, usually in the sacrum or skull-base, butalso with low frequency in the mobile spine. Chordomas arerare tumors, with an incidence below 1:1,000,000, and accountfor 1% to 4% of all primary bone malignancies. They aretypically late onset tumors with a peak incidence between thefifth and sixth decades of life, but can also occur in children andyoung adults.

Chordomas are slow growing tumors, but are characterizedby a high recurrence rate even after complete surgical resectionof the primary tumor. Distant metastases occur in 20% to 30%of cases (1), but local recurrences affect >50% of patients (2). Incase of relapse, surgery or radiotherapy become challenging and

patients usually die of their disease. Due to the location of thesetumors along the neuro-axis, patients commonly experiencephysical dysfunctions and significant pain requiring morphinederivatives and steroids (1, 3). No standard medical therapy iscurrently available, and chordomas are resistant to cytotoxicchemotherapy.

Chordomas arise from embryonic notochordal remnantsand are characterized by expression of the "T" gene product"brachyury", a notochord-specific transcription factor essentialfor mesodermal specification and differentiation during devel-opment (4). The anomalous expression of brachyury in adultnotochordal remnants is believed to play amajor role in the onsetand maintenance of chordoma (5, 6). Brachyury silencing inchordoma cell lines was shown to impair cell proliferation andinduce senescence, and attempts to target brachyury-expressingcells through a vaccine are currently ongoing (7–9). Multiplestudies have shown that chordomas commonly exhibit expres-sion and activation of tyrosine kinase receptors and downstreamsignaling molecules, with MET (HGF receptor), PDGFRb(PDGFRB), and EGFR as the most widely expressed, and HER2(ERBB2), KIT (SCFR), and VEGFR (KDR) also expressed (10–14).

The availability of clinically approved drugs targeting EGFRand PDGFR has prompted the evaluation of imatinib andlapatinib in phase II clinical trials for chordoma patientsselected for expression of corresponding drug targets (15–19). Imatinib demonstrated some clinical benefit, althoughnot achieving dimensional and long-lasting responses (16),

1Oncology, NervianoMedical Sciences, Nerviano,Milan, Italy. 2Fondazione IRCCSIstituto Nazionale dei Tumori, Milan, Italy. 3Chordoma Foundation, Durham,North Carolina.

Note: Supplementary data for this article are available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org/).

Corresponding Author: Paola Magnaghi, Nerviano Medical Sciences, Viale L.Pasteur 10, Nerviano, Milan, Italy. Phone:þ390331581287; Fax:þ390331581267;E-mail: paola.magnaghi@nervianoms.com

doi: 10.1158/1535-7163.MCT-17-0324

�2017 American Association for Cancer Research.

MolecularCancerTherapeutics

www.aacrjournals.org 603

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

whereas lapatinib did not show a real clinical benefit (19).However, anecdotal responses to other EGFR inhibitors havebeen reported, suggesting that EGFR inhibitors might havetherapeutic potential for these tumors (20–24). More recently,a phase II clinical trial with sorafenib was conducted onchordomas, based on the in vitro activity of this drug on VEGFRand PDGFR family members. In this study, patients were notselected or characterized for expression of the correspondingreceptors. Notably, sorafenib actually achieved a longer pro-gression-free survival compared with imatinib (25).

Systematic preclinical studies to identify drugs active inchordomas have been limited in the past by the paucityof biological models, as U-CH1 and U-CH2 cell lines repre-sented the only two available validated chordoma cell lines(26). More recently, a few additional bona fide chordoma celllines have become available through the Chordoma Founda-tion, a patients' advocacy organization (www.chordomafoundation.org).

With the aim of identifying kinase inhibitor drugs that mighthave therapeutic interest for chordoma patients, we assembled arepresentative panel of chordoma cell lines and assessed theirsensitivity to EGFR, PDGFR, and MET inhibitors, includingapproved drugs and other representative compounds.

Materials and MethodsCell lines and culture conditions

U-CH1, U-CH2, JHC7, and UM-Chor1 cell lines were obtainedfrom Chordoma Foundation. MUG-Chor1 and U-CH2 (ATCC)cell lines were purchased from the ATCC. Chor-IN-1 cell line wasestablished in house from a surgical sample (27). Chordoma celllines were cultured in collagen-coated plates with IMDM/RPMI1640 4:1 Medium (Gibco BRL) at 37�C in a humidifiedatmosphere containing 5% CO2. Media were supplemented with10% (v/v) heat-inactivated FBS (Euroclone). All cell lines wereauthenticated by short tandem repeat (STR) analysis (AmpFlSTRIdentifier PCR Amplification Kit, Applied Biosystems) andchecked for mycoplasma presence (MycoAlert MycoplasmaDetection Kit, Lonza LT07) upon arrival and within 6 monthsfrom resuscitation.

Reagents and antibodiesThe following antibodies were used: anti-brachyury (Santa

Cruz Biotechnology; sc-20109 rabbit polyclonal), anti-EGFR(Santa Cruz Biotechnology; sc-03 rabbit polyclonal), anti-PDGFRb (Cell Signaling Technology; 4564 rabbit monoclonal),anti-MET (Santa Cruz Biotechnology; sc10 rabbit polyclonal),anti P-MET (Cell Signaling Technology; 7896 rabbit monoclo-nal), anti–P-STAT3 (Cell Signaling Technology; 9131 rabbit poly-clonal), anti–P-AKT (Cell Signaling Technology; 4060 rabbitmonoclonal), anti-GAPDH (Santa Cruz Biotechnology; 25778rabbit polyclonal), anti–P-Tyr (Millipore; 05321 clone 4G10mouse monoclonal), anti–P-MAPK (Cell Signaling Technology;4377 rabbit monoclonal), anti-Caspase3 (Cell Signaling Tech-nology; 9662 rabbit polyclonal), and anti-LC3B antibody (600-1384) from NB (Novus Biologicals).

Horseradish peroxidase–conjugated secondary antibodieswere used 1:10,000 (Immunopure goat anti-mouse and Immu-nopure Goat anti-rabbit from Thermo-Scientific). The detectionwas performed using SuperSignal West Pico Chemiluminescentsubstrate from Thermo Scientific.

Compound acquisition and managementGefitinib (Iressa), imatinib (Glivec; Gleevec), erlotinib (Tar-

ceva), sunitinib (Sutent), and doxorubicin were synthesized inhouse. Crizotinib (PF-2341066), cabozantinib (BMS-907351;XL-184), and crenolanib (CP-868596; ARO-002) were purchasedfrom Selleck Chemicals. Afatinib (Gilotrif) was purchased fromLC Laboratories and MedChem Express. Neratinib (HKI-272),dacomitinib, rociletinib (CO-1686), osimertinib (AZD-9291),lapatinib (Tykerb), and PHA-665752 were purchased from Med-Chem Express.

Identity and purity of compounds were confirmed by LC-HRMS analysis and 1H-NMR. Compounds were dissolved in100% 10 mmol/L DMSO solutions and stored under controlledatmosphere at –20�C. All compound working stocks were pre-pared by serial dilutions in DMSO and then diluted 1:1,000 intocell culturemedium, to reach a 0.1%finalDMSOconcentration inall wells.

Assessment of antiproliferative activity of the compoundsCells were seeded at 3,600 cells/well in 384-well white clear-

bottom plates (Greiner) 24 hours before treatment. Compounds(6 dilution points in duplicate) were added and cells incubatedfor additional 144 hours. After cell lysis, cell viability was deter-mined by quantifying the ATP content (CellTiter-Glo assay,Promega using an Envision instrument, PerkinElmer). Inhibitoryactivity (IC50) was calculated by comparing treated versus controldata using a sigmoidal fitting algorithm.

Assessment of the mechanism of action of the compoundsExponentially growing cells (0.4� 106) were seeded in 60mm

plates 48 hours before treatment. Compounds were added tothe medium and plates incubated for the indicated times.Cells were then washed twice with ice-cold PBS and lysed inRIPA buffer (50 mmol/L Hepes, pH 7.5, 150 mmol/L NaCl, 1%TritonX-100, 1% sodium deoxycholate, 0.1% SDS, 10 mmol/LEDTA, protease inhibitors Cocktail (Sigma; P8340), and phos-phatase inhibitors Cocktail I and II (Sigma; P2850 and P5725).Lysates were clarified by centrifugation for 10 minutes at16,000 x g, samples (20 mg each) fractionated in 4% to 12%SDS-PAGE, and transferred onto a Nitrocellulose Membrane(Whatman-Protan; 10401196). Immunoblots were performedwith the described antibodies in TBS 1X (Biorad) with 5% milkand 0.1% Tween 20.

RT-qPCR analysisRNA was extracted following the RNeasy Mini Kit (Qiagen) as

per the manufacturer's procedures. RNA yield and purity wereevaluated by 260 nm UV absorption and 260/280 and 260/230ratios using Nanodrop 1000 (Thermo Scientific). RNA wasreverse-transcribed using TaqMan Reverse Transcription Reagents(Applied Biosystems/Life Technologies) and random hexamerpriming. Each sample (10–12 ng cDNA) was assayed in duplicate12.5 mL reactions by real-time qPCR, on a ABI Prism 7900HTApplied Biosystems Sequence Detector (Life Technologies) usingSYBR Green technology, with reagents and materials fromApplied Biosystems/Life Technologies (Power SYBR Green PCRMaster Mix). Target-specific and endogenous reference control(GUSB, PPIA, 18S rRNA) primers (300 nmol/L) are detailed in thetable below. Relative quantification of expression levels wascalculated following theDDCtmethod (Livak; Applied Biosystems

Magnaghi et al.

Mol Cancer Ther; 17(3) March 2018 Molecular Cancer Therapeutics604

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

Sequence Detector User Bulletin 2, https://www2.appliedbiosystems.com). All procedures were performed according to themanufacturer's instructions.

Primer sequences:EGFR: fw 5'-AAGTCCCCCAGTGACTGCT-3', rev 5'-TGGCTT-

CGTCTCGGAATTT-3'T-brachyury: fw 5'-GTATGAGCCTCGAATCCACAT-3', rev 5'-

GATAAGCAGTCACCGCTATGAA-3';STK33: fw 5'-CGACTATAGCCAGCAGTGTGA-3', rev 5'-TCTCT-

TCTGAGCTTGCCAAA-3';AXL: fw 5'-CACCTCCCTGCAGCTTTC-3', rev 5'-CTCCAGCC-

CAACATAGCC-3';GUSB: fw 5'-GCATCCAAAAGACGCACTTC-3', rev 5'-CACC-

CACCACCTACATCGAT-3';PPIA: fw 5'-CCCACCGTGTTCTTCGACAT-3', rev 5'-TTTCT-

GCTGTCTTTGGGACCTT-3';18S rRNA: fw 5'-AGCTGGAATTACCGCGG-3', rev 5'-TGAC-

GAAAAATAACAATACAGGACTC-3'.

In vivo studiesPreclinical studies were conducted through the Chordoma

Foundation Drug Screening Pipeline at South Texas AcceleratedResearch Therapeutics (START) under International Animal Careand Use Committee–approved protocols.

Antitumor activity was tested in:* U-CH1: xenograft of the U-CH1 cell line (28).* SF8894: PDX model (29).* CF322: new chordoma PDX model generated from a clival

recurrent tumor of a 42-year-old male.* CF365: new chordoma PDX model generated from a clival

poorly differentiated metastatic tumor of a 11-year-old male.

For all models, athymic nude mice (Charles River Labs)between 6 and 8 weeks of age were implanted subcutaneouslywith tumor fragments from host animals. Once tumors reachedapproximately 150 to 250mm3, animals were matched by tumorvolume (TV) and randomized to control and treatment groups.All groups were dosed at 20 mg/kg PO daily, U-CH1 and SF9984group for 28 days, and CF322 and CF365 groups to end. Initialdosing began at day 0. Animals were observed daily and weighedtwice a week. TV and animal weight data were collected electron-ically using a digital caliper and scale; tumor dimensions wereconverted to volume using the formula TV (mm3) ¼ width2

(mm2) � length (mm) � 0.52. Endpoints were a mean controlTV of approximately 1 to 2 cm3. Percent tumor growth inhibitionvalues were calculated and reported for treatment versus controlgroups using initial and final tumor measurements. Statisticalanalysis was performed using a two-way ANOVA followed by theDunnett multiple comparisons test.

EGFR immunoprecipitationCell were lysed in 1X Cell Lysis Buffer CLB (20 mmol/L TRIS,

pH7.5, 150mmol/LNaCl, 1mmol/L EDTA, 1mmol/L EGTA, 1%Triton X-100, Protease and Phosphatase inhibitor cocktail-EZBlock Protease Inhibitor Cocktail EDTA-free BV-K272, EZBlockPhosphatase Inhibitor Cocktail II, BV-K275-1EAEZ, Block Phos-phatase Inhibitor Cocktail III, BV-K276-1EA-Vinci-Biochem).

Protein G Sepharose 4 fast flow (17-0618; GE Healthcare) waswashed 4Xwith ice-cold PBS, preloadedwith anti-EGFR antibody(Santa Cruz Biotechnology sc-03) in PBS/TritonX-100 and incu-bated overnight with gentle rocking at 4�C.

Cell lysates were precleared by rocking at 4�C for 20 minuteswith Protein G Sepharose previously washed 4Xwith ice-cold PBSand 1X with CLB buffer.

Preloaded resin/antibodies were incubated at 4�C rocking for 3hours with the cell lysate. Samples were transferred on cytospinfilter (CytoSignal cat. C00-100), washed 3X with CLB buffer, andeluted at 95�C in Laemmli-SDS buffer 2X.

ResultsWe assembled a panel of chordoma cell lines which included

cells of sacral origin like the widely used U-CH1 (28) and U-CH2(26), the recently established MUG-Chor1 (30) and JHC7 (8),and the first available clival cell line UM-Chor1 (31). In addition,we tested Chor-IN-1, a new cell line derived in house from a sacralchordoma surgical sample shown to meet validation criteria forchordoma cell lines (27). All cell lines were first authenticated bySTR fingerprinting. The STR profile of U-CH2 cells obtained fromtwo sources differed at a single locus, suggesting the acquisition oflittle differences during their propagation (Supplementary TableS1). Both clones were therefore tested in parallel.

Immunoblot analysis showed that all seven cell lines expressedthe "T" gene product brachyury, considered a defining marker ofchordoma (Fig. 1).

Brachyury

P-STAT3 (Thr705)

GAPDH

EGFR

PDGFRb

STAT3

P-AKT (Ser473)

AKT

P-MAPK

MAPK

MET

HER2

Figure 1.

Immunoblot analysis of tyrosine kinase receptors and other signaling moleculesin chordoma cell lines. Immunoblot analysis of U-CH1, U-CH2, UM-Chor1,Chor-IN-1, MUG-Chor1, and JHC7protein cell extracts resolved onSDS-PAGEgel,transferred onto nitrocellulose membrane, and probed with the indicatedantibodies.

Afatinib Targets EGFR and Brachyury in Chordoma

www.aacrjournals.org Mol Cancer Ther; 17(3) March 2018 605

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

MET and EGFR were shown to display similar expression levelsacross the different cell lines, and HER2 was barely detectable.PDGFRb expression was more heterogeneous, ranging from verylow (U-CH1) to high expression (U-CH2 and UM-Chor1). Thedownstream signaling molecules AKT (PKB), MAPK (MAPK1/MAPK3), andSTAT3weredifferentially activatedacross thedifferentcell lines, with no specific correlation to receptor expression (Fig. 1).

To evaluate the importance of these tyrosine kinase receptorsfor the growth of chordoma cells, we assembled a broad panel ofpotent MET, PDGFRb, and EGFR inhibitors, including approveddrugs as well as other reference inhibitors. The antiproliferativeactivity of compounds was determined after 144-hour incuba-tion, to allow two population doublings, considering the slowgrowth rate of chordoma cells (>60-hour doubling time). Inparallel, the inhibitors were profiled on reference cell lines: gastricadenocarcinoma MKN-45 with MET amplification (32), NSCLCNCI-H1703 with PDGFRA amplification (33), and epidermoidcarcinoma A431 with EGFR amplification were used as positivecontrols, whereas ovarian carcinoma A2780 cell line was used asgeneral negative control, for which no target-dependent activitywas expected.

Crizotinib (34, 35) and cabozantinib (36), two approved METinhibitor drugs, and the MET-selective inhibitor PHA-665752(35, 37) did not display considerable antiproliferative activityagainst chordoma cell lines, with IC50s higher than those mea-sured for the A2780 control cell line (Supplementary Table S2A).

A similar result was observed with several PDGFR inhibitors,including the approved drugs sunitinib and imatinib (35, 38, 39)as well as crenolanib (ref. 40; Supplementary Table S2B).

We then tested the clinically approved EGFR inhibitor drugserlotinib, gefitinib, afatinib, and lapatinib, which possess differ-ent potency and selectivitywithin theEGFR/HER2 family (35, 41–44). All compounds were active on U-CH1 and UM-Chor1 celllines, although with different potency (Table 1A). In particular,afatinibwas very potent against both cell lines, with IC50s of 0.014and 0.023 mmol/L, respectively, whereas the other three drugs

exhibited IC50s in the 0.1–0.8 mmol/L range. Moreover, afatinibwas the only drug displaying activity, although with differentpotency, against all cell lines (IC50 < 0.7 mmol/L), with theexception of JHC7. The other EGFR inhibitors erlotinib, lapatinib,and gefitinib, despite being active on U-CH1 and UM-Chor1 inthe submicromolar range, did not generally show considerableantiproliferative activity against the other chordoma cell lines.

To investigate whether these results were correlated with thepotency and/or the covalent mechanism of EGFR inhibition ofafatinib, we also tested the covalent EGFR inhibitors neratinib(35, 45) and dacomitinib (46). Both inhibited U-CH1 and UM-Chor1 in the nanomolar range, but had poor activity on the othercell lines,with comparable or higher IC50s than those observed forthe A2780 control cell line (Table 1B).

Interestingly, dacomitinib displayed IC50 values comparablewith afatinib in U-CH1 and UM-Chor1, but was not generallyactive in the other cell lines, therefore indicating that the activity ofafatinib across the chordoma panel is not just related to itsbiochemical potency against EGFR. Of note, the T790M EGFRmutant-selective covalent inhibitors osimertinib (AZD-9291;ref. 47) and rociletinib (48) were inactive or poorly active in allchordoma cell lines, as expected in the absence of EGFRmutations(13, 27, www.chordomafoundation.org).

These data suggest that the antiproliferative effect observedwith afatinib against chordoma is linked to EGFR inhibition, butis not just related to its potency or covalent mechanism of action.

To obtain a dynamic view of the effects induced by treatmentwith afatinib, we performed a kinetic live-cell analysis (see Sup-plementary Methods) using a multi-well automated microscope(IncuCyte ZOOM). U-CH1 cells were treated with afatinib ordoxorubicin standard, and a time-lapse movie was generatedintegrating the sequence of live-cell images acquired through144-hour treatment. Cell confluence was progressively decreasedover time with increasing concentrations of afatinib or doxoru-bicin, while increasing in the untreated cells (SupplementaryFig. S1A). At active compound concentrations, graphs bottomed

Table 1. Antiproliferative activity of EGFR inhibitors in chordoma cell lines

IC50 (mmol/L; in brackets StdDev)

Cell line U-CH1 UM-Chor1 MUG-Chor1 U-CH2 (Ch. F.) U-CH2 (ATCC) Chor-IN-1 JHC7 A431 ctrl A2780 ctrl

A. Approved EGFR inhibitorsAfatinib 0.014 0.023 0.258 0.494 0.531 0.668 1.346 0.026 1.915

(0.005) (0.007) (0.072) (0.409) (0.203) (0.351) (0.394) (0.009) (0.594)

Erlotinib 0.144 0.617 3.006 8.042 7.776 2.329 2.281 0.346 3.919(0.049) (0.069) (0.977) (1.714) (1.953) (0.774) (0.848) (0.033) (0.898)

Lapatinib 0.656 0.516 >10 >10 >10 >10 >10 0.562 3.578(0.257) (0.080) (—) (—) (—) (—) (—) (0.061) (0.771)

Gefitinib 0.791 0.751 6.241 6.259 5.936 9.040 7.010 0.333 4.762(0.446) (0.055) (1.390) (2.502) (2.115) (1.389) (0.856) (0.069) (0.911)

B. Other advanced EGFR inhibitorsNeratinib 0.038 0.110 1.946 1.926 3.253 2.009 2.893 0.282 0.603

(0.031) (0.023) (0.233) (0.189) (0.470) (0.569) (0.505) (0.059) (0.132)

Dacomitinib <0.019 <0.019 2.075 2.400 2.145 0.418 1.230 0.043 2.715(—) (—) (0.191) (0.042) (0.276) (0.054) (0.156) (0.014) (0.106)

Osimertinib 0.459 0.569 1.129 2.250 1.165 0.384 0.897 0.646 3.655(T790M) (0.120) (0.034) (0.228) (0.085) (0.134) (0.021) (0.113) (0.199) (0.049)

Rociletinib 2.124 2.108 3.623 3.631 3.688 1.973 3.236 1.599 1.026(T790M) (0.484) (0.586) (1.520) (1.152) (0.684) (1.250) (0.683) (1.527) (0.152)

Doxorubicin 0.166 0.067 0.455 0.152 0.348 0.340 0.881 0.083 0.010(0.043) (0.040) (0.185) (0.038) (0.191) (0.119) (0.164) (0.004) (0.005)

NOTE: In bold, registered EGFR inhibitors. The average IC50 (mmol/L) were determined as described under the Materials and Methods.All values were derived from technical duplicates and confirmed in multiple (n > 6) independent biological experiments.A431 and A2780 cell lines are shown for comparison. Doxorubicin is used as reference standard.

Magnaghi et al.

Mol Cancer Ther; 17(3) March 2018 Molecular Cancer Therapeutics606

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

to a background signal, representing the image analysis of celldebris. Conversely, caspase 3 (CASP-3) and caspase 7 (CASP-7)induction (Supplementary Methods), measured with a specificfluorescent dye, was evident upon doxorubicin treatment, butwas not detectable upon treatment with afatinib (SupplementaryFig. S1B and S1C). This experiment demonstrated that afatinibinhibits U-CH1 cell proliferation and induces cell death in a time-and dose-dependent manner, and indicated that the main mech-anism of cell death is not apoptosis.

We then studied the dose- and time-dependent biomarkermodulation following treatment with afatinib, erlotinib, or

lapatinib in U-CH1 cells, responsive to all EGFR inhibitors(Fig. 2). AKT and MAPK pathways were modulated in a dose-dependent manner by the three inhibitors, both after short(2-8-24 hours) and longer (48 hours) incubation times, atdoses consistent with antiproliferative IC50s. STAT3 activationinstead was not dependent upon EGFR signaling in these cells.Caspase 3 cleavage was not observed up to 48 hours, againsuggesting the involvement of mechanisms other than apopto-sis in the antiproliferative effect of these inhibitors (Fig. 2;Supplementary Fig. S1B). Interestingly, treatment with afatiniband erlotinib induced a decrease in the total level of both the

2

48

0.01 0.1 1 5 0.01 0.1 1 5 0.01 0.1 1 5 mmol/LC Afatinib Erlotinib Lapatinib C

p-AKT(Ser473)

Treatment (hours)

GAPDH48

2

48

p-STAT3 (Thr705)

2

48

p-MAPK(Tyr202/Tyr204)

LC3B8

48

Casp. 32

48

8

48Brachyury

24

48EGFR

IC50: 0.014 IC50: 0.144 IC50: 0.656 mmol/L

Figure 2.

Biomarker modulation upon treatment of U-CH1 cellline with different EGFR inhibitors. Immunoblot analysisof U-CH1 cells treated with the indicated doses ofinhibitors for 2-8-24 hours (top) or 48 hours (bottom).Protein cell extracts were resolved on SDS-PAGE gel andmembranes probed with the indicated antibodies.IC50s of the different inhibitors are reported.

Afatinib Targets EGFR and Brachyury in Chordoma

www.aacrjournals.org Mol Cancer Ther; 17(3) March 2018 607

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

cytoplasmic and membrane-bound forms of LC3B (MAP1LC3B),hinting a potential involvement of autophagic pathways.

A striking difference in the mechanism of action of the threecompounds was the effect on brachyury and EGFR protein levelsat the latest time points. After 48-hour incubation, afatinibinduced not only a strong decrease in the total level of EGFR,but also a dose-dependent decrease of brachyury protein. EGFRdepletion was observed but only to a lesser extent upon treatmentwith erlotinib, and was absent after treatment with lapatinib.Brachyury expression was not affected by treatment with erlotinibor lapatinib.

To assess whether the afatinib-induced EGFR and brachyurydownmodulation was linked to stronger potency or covalentmechanism of inhibition, we analyzed the effect of dacomitinib,which displayed similar IC50s on U-CH1 andUM-Chor1, but wasgenerally inactive on the other cell lines (Table 1). Treatment ofU-CH1 cells with dacomitinib induced a dose-dependent down-modulation of P-AKT levels, but, differently fromafatinib, did notaffect total levels of EGFR and brachyury up to 5 mmol/L (Sup-plementary Fig. S2). These data support the hypothesis that theactivity observed with afatinib against chordoma cell lines couldbe linked to its ability to ablate two proteins relevant for chor-doma cells growth, and not merely to its potency or covalentmechanism of EGFR inhibition.

To investigate whether this particularmechanism occurred alsoin the other afatinib-responsive chordoma cell lines, MUG-Chor1and Chor-IN-1 were treated with the compound and analyzed byimmunoblot. Figure 3 shows that downmodulation of EGFR andbrachyury occurred in the same dose-response manner within

each cell line, which correlated with the measured IC50s. Afatinibtherefore turned out to be the only EGFR inhibitor displayingactivity, althoughwith different potency, across the chordoma cellline panel and the only inhibitor inducing a downmodulation oftwo biomarkers known to play a major role in chordoma cellgrowth.

To explore potential correlations between EGFR andbrachyury,their reciprocal impact upon siRNAwas analyzed (SupplementaryMethods). Transfection of U-CH1 cells with specific siRNA oli-gonucleotides induced a complete ablation of the correspondingtarget protein (EGFR or brachyury), with no effect on the otherprotein. In both cases, a strong impairment of cell viability wasobserved, therefore confirming that brachyury and EGFR play acrucial role in chordoma cell growth (Supplementary Fig. S3A andS3B). We ruled out that this downregulation occurred transcrip-tionally, because treatment of U-CH1 cells with 0.1 and 1 mmol/Lafatinib for 24 or 48 hours did not affect brachyury mRNA andonly slightly decreased EGFR mRNA (Supplementary Fig. S4A).

Instead, treatment of U-CH1 cells with afatinib in the presenceof MG-132 (proteasome inhibitor) or bafilomycin (autophagyinhibitor) prevented afatinib-induced EGFR and brachyury pro-tein ablation (Supplementary Fig. S4B), therefore showing thatthe downmodulation of these proteins involves pathways ofprotein degradation.

We then assessed the activity of afatinib in vivo in U-CH1xenografts and SF8894, CF322, and CF365 PDX (29) mousemodels, generated as described in the Materials and Methodssection. Daily oral treatment of mice with 20 mg/kg afatinib forthe indicated times induced potent tumor growth inhibition in all

U-CH1

2

48

C Afatinib

p-AKT (Ser473)

Treatment (hours)

8

48Brachyury

24

48EGFR

GAPDH 48

MUG-Chor1

C Afatinib

IC50: 0.014 mmol/L IC50: 0.258 mmol/L

Chor-IN-1

C Afatinib

IC50: 0.668 mmol/L

0.01 0.1 1 5 mmol/L 0.01 0.1 1 5 mmol/L 0.01 0.1 1 5 mmol/L

Figure 3.

Biomarker modulation upon treatment of U-CH1, MUG-Chor1, and Chor-IN-1 cell lines with afatinib. Immunoblot analysis of U-CH1, MUG-Chor1, and Chor-IN-1 cellstreated with afatinib for 2-8-24 or 48 hours. Protein cell extracts were resolved on SDS-PAGE gel and membranes probed with the indicated antibodies.IC50s for each cell line are reported.

Magnaghi et al.

Mol Cancer Ther; 17(3) March 2018 Molecular Cancer Therapeutics608

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

the analyzed models, with no apparent signs of toxicity (Fig. 4).These data confirm that the activity observed in vitro with afatinibtranslates into impressive in vivo efficacy in four out of fourmodels,providing a strong rationale for testing this drug in the clinic.

Finally, we exploited the cell lines with different levels ofresponse to afatinib to start exploring potential biomarkers ofsensitivity to this drug. We investigated whether the differencesin the IC50s among these cell lines could be related to thetotal EGFR levels, to the extent of receptor activation or to adifferential response to the inhibitor. Total and membrane-bound EGFR levels were found to be comparable among thedifferent chordoma cell lines, as detected by flow cytometryanalysis with cetuximab (Supplementary Methods; Supplemen-tary Fig. S5A).

The level of EGFR phosphorylation was analyzed, both in basalconditions and after treatment with afatinib, by immunoprecip-itation of the receptor from total cell lysates followed by immu-noblot with anti–P-Tyr antibody. Figure 5A shows that EGFR isefficiently immunoprecipitated from all cell lysates (bottom) andthat sensitive cell lines harbor phosphorylated EGFR (top),where-as JHC7 has a minimal level of basal phosphorylation. Interest-ingly, treatment with afatinib induced complete EGFR dephos-phorylation in U-CH1 cell line, but did not decrease the weak P-Tyr band detected in JHC7 (Fig. 5A, top). Because treatment withafatinib completely abolished EGFR phosphorylation also in

MUG-Chor1, Chor-IN-1, and U-CH2 cell lines (SupplementaryFig. S5B), the band observed in JHC7might be due to recognitionof a different phosphorylation site, not linked to receptor activa-tion. Therefore, the sensitivity of chordoma cell lines to afatinibseems to generally correlate with EGFR phosphorylation, and thelack of EGFR activation in JHC7 cell line likely accounts for themeasured higher IC50.

To identify other potential determinants of differential sen-sitivity to afatinib, we analyzed the expression of AXL, reportedto mediate resistance to anti-EGFR therapies in different tumortypes (49). Figure 5B shows that AXL expression inverselycorrelates with sensitivity to afatinib, with a weak expressionin U-CH1 and UM-Chor1 cell lines, and higher levels in theless sensitive MUG-Chor1, U-CH2, and Chor-IN-1 cell lines. Noexpression of AXL was observed in JHC7, where the EGFRpathway is not activated.

Interestingly, in a whole kinome-targeted sequencing, evaluat-ing in parallel the differential expression of approximately 500kinases in U-CH1 versus the other sacral chordoma cell lines,STK33 emerged as the only kinasewith undetectable expression inU-CH1 and higher expression in the other cell lines (27). We thenevaluated STK33 expression in all the cell lines by RT-qPCR andfound no expression in both afatinib highly responsive U-CH1and UM-Chor1 cell lines, while confirming relevant expression inthe others.

Tum

or v

olum

e (m

m3 )

20 mg/kg; po; qdx28

20 mg/kg; po; qd to end

Tum

or v

olum

e (m

m3 )

2,000

1,800

1,600

1,400

1,200

1,000

800

600

400

200

00 10 20 30 40 50 0 10 20 30 40 50

1,200

1,000

800

600

400

200

0

900

800

700

600

500

400

300

200

100

00 10 20 30 40 50

Days

Days

Days

Days

2,500

2,000

1,500

1,000

500

00 20 40 60 80

Control

Afatinib

Control

Afatinib

UCH1 SF8894

CF322 CF365

Figure 4.

In vivo efficacy of afatinib on U-CH1xenograft and different PDX models.Nude mice bearing U-CH1 or SF8894,CF322, and CF365 tumors were doseddaily with afatinib 20 mg/kg po for 28days (bar indicates the length oftreatment) or to the end of treatment,as indicated.

Afatinib Targets EGFR and Brachyury in Chordoma

www.aacrjournals.org Mol Cancer Ther; 17(3) March 2018 609

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

Finally, we performed a pharmacodynamics study evaluatingEGFR signaling aswell as AXL and STK33 levels after 1 and 2hoursof treatment of U-CH1, MUG-Chor1, and Chor-IN-1 cell lineswith afatinib. No relevant alterations of AXL and STK33mRNA ortotal protein levels were detected (Supplementary Fig. S6A andS6B). Interestingly, although EGFR phosphorylation was abro-gated in all cell lines at 0.1 mmol/L afatinib, P-AKT was shutdown inU-CH1 cell line, but displayed a slight increasing trend inMUG-Chor1 and Chor-IN-1 (Supplementary Fig. S6B and S6C),highlighting that the activity of afatinib in these cell lines iscontributed by cellular mechanisms that go beyond EGFR kinaseinhibition.

DiscussionOnly a few clinical trials have been conducted so far to assess

the efficacy of targeted therapeutic agents in chordoma (20–24),

certainly limited by the exiguous number of preclinical models ofthis indication available so far. We recently analyzed in depth atthe genomic level a panel of chordoma cell lines of sacral originthat included also the novel cell line Chor-IN-1, established andcharacterized in our laboratories (27).With the goal of identifyingnew drugs that might have an immediate therapeutic applicationfor chordoma patients, we evaluated the in vitro antiproliferativeactivity of inhibitors ofMET, PDGFRb, andEGFR tyrosine kinases,reported to be frequently expressed in chordomas.

MET inhibitors did not display significant activity and were notfurther investigated. PDGFR inhibitors were also not active, evenin the cell lines with strong expression of PDGFRb such as U-CH2and UM-Chor1. Considering that a phase II clinical trial withimatinib demonstrated some clinical benefit for chordomapatients, PDGFRb inhibition in the tumor microenvironmentmight result in effects that go beyond mere antiproliferativeactivity on tumor cells in vitro, or PDGFRb in tumors might be

A

250

150

250

150I.P. W.B.

+ +

Afatinib sensitivity - +++ + + +++ + ++

+ : Afatinib 1 mmol/L

B

100

150AXL

GAPDH

Rel

ativ

e qu

antif

icat

ion

Afatinib sensitivity

STK33 RT-qPCR

I.P. a-EGFR tot.

a-EGFR tot.a-EGFR tot.

W.B. a-P-Tyr

4

3.5

3

2.5

2

1.5

1

0.5

0

U-CH1

U-CH2 (

CF)

U-CH2 (

ATCC)

UM-Chor1

Chor-IN-1

JHC7

Pos contro

l

Neg co

ntrol

Ref co

ntrol

MUG-Chor1

Figure 5.

Identification of potential determinants of sensitivity to afatinib. A, Immunoprecipitation analysis of EGFR phosphorylation in the different chordoma cell lines. Thereceptor was immunoprecipitated with anti-EGFR antibody from untreated or afatinib-treated cell lines, as indicated. Immunocomplexes were resolved onSDS-PAGE gel. Membranes were probed using an anti-EGFR or anti–P-Tyr antibody. B, Expression of AXL and STK33 in the different chordoma cell lines. Top:immunoblot analysis of AXL expression. Protein cell extracts were resolved on SDS-PAGE gel and membranes probed with anti-AXL antibody. Bottom: RT-qPCRanalysis of STK33 mRNA expression normalized to reference controls as described. Cell lines are ordered based on afatinib sensitivity ranking.

Magnaghi et al.

Mol Cancer Ther; 17(3) March 2018 Molecular Cancer Therapeutics610

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

activated by paracrine mechanisms that are lost in vitro. Interest-ingly, responses to imatinib observed in the clinic are primarilyrelated to changes in tumor features rather than dimension (17).

Multiple reports in the literature highlight the activity of dif-ferent EGFR inhibitors in chordoma cell lines, in animal models,and also sporadically in patients (reviewed in refs. 24, 50),providing evidence of a potential clinical relevance of EGFRinhibitors in chordoma, but also raising questions about themost suitable agent.

We focused our studies on clinically approved EGFR inhibitorsand observed that all of them displayed some activity, althoughwith different potencies, against U-CH1 andUM-Chor1 cell lines.Afatinib was the most potent inhibitor in these cell lines, whereaslapatinib was the least active in addition to being inactive on theremaining cell lines, whichmight have predicted the poor activityobserved for lapatinib in the clinic. Afatinibwas also the only drugwith activity across the chordoma cell lines panel, with theexception of JHC7, that despite expressing significant levels ofEGFR is not driven by EGFR signaling.

The activity of afatinib in chordoma cell lines is particularlyrelevant because this drug is very selective biochemically andits cellular activity is dependent upon EGFR pathway activation(41, 42). The activity of afatinib in U-CH1 and UM-Chor1chordoma cell lines is equivalent to that reported for mutantEGFR-dependent lung tumors in vitro (42, 48), suggesting thatchordomamight also represent a potential indication for this drug.

These cellular data were also corroborated by impressive in vivoefficacy,with tumor growth inhibition inone chordomaxenograftand in three different PDX models.

Scheipl and colleagues (51) published a focused compoundscreening describing the activity of EGFR inhibitors in a subset ofchordoma cell lines, reporting, in line with our data, that U-CH1and UM-Chor1 cell lines are sensitive to EGFR inhibitors, butunexpectedly they found afatinib active on UM-Chor1 but not inU-CH1 cell line. Considering the potent activity we repeatedlyobserved for afatinib on U-CH1 both in vitro, with modulation ofdownstream pathways, and particularly in vivo, the lack of activitythey observed in U-CH1 might be related to different screeningconditions or compound handling.

Differently from erlotinib, gefitinib, and lapatinib that arereversible inhibitors, afatinib contains an electrophilic groupcapable of Michael addition to conserved cysteines within thecatalytic domains of EGFR (Cys797), HER2 (Cys805), and HER4(Cys803), that induces complete and long-lasting inhibitionof receptor phosphorylation (42). However, the activity of afati-nib in chordoma cell lines appears to be related not only to itshigher potency and covalent mechanism, as neratinib and daco-mitinib, sharing the samemechanism, were active only in U-CH1and UM-Chor1 cell lines.

This activity appears to be strongly contributed by its uniqueability to downmodulate the total level of EGFR and brachyuryproteins, a feature not shared by the other inhibitors. Interesting-ly, afatinib was recently shown to induce a downmodulation ofthe total level both of HER2 and EGFR in tumors harvested fromNCI-N87 xenografts treated with afatinib (52).

Brachyury is a distinctive marker of sporadic chordomas,duplication of the T gene locus is typical of familial chordomas(6), and a DNA polymorphism in the T gene is associated withchordoma predisposition in the general population (53). Accord-ingly, silencing of brachyury in chordoma cell lines was widelyshown to impair tumor growth both in vitro and in vivo (7, 8;

Supplementary Fig. S3B). Brachyury would represent an obvioustarget for drug development in chordomas, but direct inhibitionof transcription factors with small molecules has thus far beenextremely challenging.

The identification of a small molecule that, in addition toinhibiting EGFR signaling, induces brachyury degradation repre-sents a new approach indirectly targeting this important tran-scription factor with a kinase inhibitor.

Overall, testing EGFR inhibitors across different cell lines andin vivomodels suggests that a subset of chordomas are driven bythe EGFR signaling pathway and are strikingly sensitive to EGFRinhibition. Preclinically, this sensitivity correlates with the levelof EGFR phosphorylation and it will be important to assess ifthis holds true in the clinical study. Interestingly, Mug-Chor1,U-CH2, and Chor-IN-1 chordoma cell lines display a strongexpression of AXL receptor, which deserves further investiga-tion, because it represents a mechanism of resistance in lungcancer (49). Should this occur also in chordoma, a combina-tion therapy with AXL inhibitors, which are already in theclinic, might be beneficial to these types of patients. STK33,which is absent in the cell lines most sensitive to afatinib, alsorepresents a candidate biomarker for afatinib sensitivity whichdeserves further investigation.

These data provide a strong rationale for formal evaluation ofafatinib in the treatment of chordoma patients. Accordingly, aEuropean phase II study on afatinib in advanced chordoma isabout to start patient recruitment.

Disclosure of Potential Conflicts of InterestS. Stacchiotti is a consultant/advisory board (Minor) member of Bayer. No

potential conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: P. Magnaghi, B. Salom, N. Amboldi, D. Ballinari,E. Tamborini, S. Pilotti, S. Stacchiotti, A. IsacchiDevelopment of methodology: P. Magnaghi, B. Salom, L. Cozzi, N. AmboldiAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): P. Magnaghi, B. Salom, N. Amboldi, F. Gasparri,L. Raddrizzani, C. Orrenius, F. Bozzi, J. SommerAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): P. Magnaghi, B. Salom, L. Cozzi, N. Amboldi,D. Ballinari, F. Gasparri, L. Raddrizzani, A. Somaschini, R. Bosotti, C. OrreniusWriting, review, and/or revision of the manuscript: P. Magnaghi, B. Salom,N. Amboldi, D. Ballinari, E. Tamborini, S. Pilotti, A. Galvani, J. Sommer,S. Stacchiotti, A. IsacchiAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): P.Magnaghi, B. Salom,N. Amboldi, C. OrreniusStudy supervision: P. Magnaghi, B. Salom, E. Tamborini, A. Montagnoli,A. Isacchi

AcknowledgmentsWe acknowledge the Chordoma Foundation for providing U-CH1, U-CH2,

JHC7, and UM-Chor1 cell lines. We acknowledge the Johns Hopkins Universityfor JHC7 and University of Michigan for UM-Chor1 cell lines.

We also acknowledge Clara Albanese, Stefano Camisasca, Antonella Land-onio, and Claudia Re from Nerviano Medical Sciences for technical support inflow cytometry analysis and cell culturing, and Michael Wick from START fortechnical support in the efficacy study.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received April 12, 2017; revised August 31, 2017; accepted November 30,2017; published OnlineFirst December 13, 2017.

Afatinib Targets EGFR and Brachyury in Chordoma

www.aacrjournals.org Mol Cancer Ther; 17(3) March 2018 611

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

References1. Stacchiotti S, Sommer J. Building a global consensus approach to chor-

doma: a position paper from the medical and patient community. LancetOncol 2015;16:e71–83.

2. Yamada Y, Gounder M, Laufer I. Multidisciplinary management of recur-rent chordomas. Curr Treat Options Oncol 2013;14:442–53.

3. Walcott BP, Nahed BV, Mohyeldin A, Coumans JV, Kahle KT, Ferreira MJ.Chordoma: current concepts, management, and future directions. LancetOncol 2012;13:e69–76.

4. Vujovic S, Henderson S, Presneau N, Odell E, Jacques TS, Tirabosco R, et al.Brachyury, a crucial regulator of notochordal development, is a novelbiomarker for chordomas. J Pathol 2006; 209:157–65.

5. Nibu Y, Jose-Edwards DS, Di Gregorio A. From notochord formation tohereditary chordoma: the many roles of Brachyury. Biomed Res Int2013;2013:826435.

6. Yang XR,NgD, AlcortaDA, LiebschNJ, Sheridan E, Li S, et al. T (brachyury)gene duplication confers major susceptibility to familial chordoma. NatGenet 2009;41:1176–8.

7. Presneau N, Shalaby A, Ye H, Pillay N, Halai D, Idowu B, et al. Role of thetranscription factor T (brachyury) in the pathogenesis of sporadic chor-doma: a genetic and functional-based study. J Pathol 2011;223:327–35.

8. HsuW,MohyeldinA, ShahSR, apRhysCM, Johnson LF, Sedora-RomanNI,et al. Generation of chordoma cell line JHC7 and the identification ofBrachyury as a novel molecular target. J Neurosurg 2011;115:760–9.

9. Heery CR, Singh BH, Rauckhorst M, Marte JL, Donahue RN, Grenga I, et al.Phase I trial of a yeast-based therapeutic cancer vaccine (GI-6301) targetingthe transcription factor brachyury. Cancer Immunol Res 2015;3:1248–56.

10. Tamborini E, Miselli F, Negri T, Lagonigro MS, Staurengo S, Dagrada GP,et al. Molecular and biochemical analyses of platelet-derived growth factorreceptor (PDGFR) B, PDGFRA, and KIT receptors in chordomas. ClinCancer Res 2006;12:6920–8.

11. Akhavan-Sigari R, Gaab MR, Rohde V, Abili M, Ostertag H. Expression ofPDGFR-alpha, EGFR and c-MET in spinal chordoma: a series of 52 patients.Anticancer Res 2014;34:623–30

12. Dewaele B,Maggiani F, Floris G, AmpeM, Vanspauwen V,Wozniak A, et al.Frequent activation of EGFR in advanced chordomas. Clin Sarcoma Res2012;1:4.

13. Tamborini E, Virdis E, Negri T, OrsenigoM, Brich S, Conca E, et al. Analysisof receptor tyrosine kinases (RTKs) and downstream pathways in chordo-mas. Neuro Oncol 2010;12:776–89.

14. de Castro CV, Guimaraes G, Aguiar S Jr, Lopes A, Baiocchi G, da Cunha IW,et al. Tyrosine kinase receptor expression in chordomas: phosphorylatedAKT correlates inversely with outcome. Hum Pathol 2013;44:1747–55.

15. Stacchiotti S, Longhi A, Ferraresi V, Grignani G, Comandone A, Stupp R,et al. Phase II study of imatinib in advanced chordoma. J Clin Oncol2012;30:914–20.

16. Hindi N, Casali PG, Morosi C, Messina A, Palassini E, Pilotti S, et al.Imatinib in advanced chordoma: a retrospective case series analysis. Eur JCancer 2015;51:2609–14.

17. Casali PG,Messina A, Stacchiotti S, Tamborini E, Crippa F, Gronchi A, et al.Imatinib mesylate in chordoma. Cancer 2004;101:2086–97.

18. Stacchiotti S, Marrari A, Tamborini E, Palassini E, Virdis E, Messina A, et al.Response to imatinib plus sirolimus in advanced chordoma. Ann Oncol2009;20:1886–94.

19. Stacchiotti S, Tamborini E, Lo Vullo S, Bozzi F, Messina A, Morosi C, et al.Phase II study on lapatinib in advanced EGFR-positive chordoma. AnnOncol 2013;24:1931–6.

20. Singhal N, Kotasek D, Parnis FX. Response to erlotinib in a patient withtreatment refractory chordoma. Anticancer Drugs 2009;20:953–5.

21. Houessinon A, BooneM, Constans JM, Toussaint P, Chauffert B. Sustainedresponse of a clivus chordoma to erlotinib after imatinib failure. Case RepOncol 2015;8:25–9.

22. Linden O, Stenberg L, Kjellen E. Regression of cervical spinal cord com-pression in a patient with chordoma following treatment with cetuximaband gefitinib. Acta Oncol 2009;48:158–9.

23. Launay SG, Chetaille B, Medina F, Perrot D, Nazarian S, Guiramand J, et al.Efficacy of epidermal growth factor receptor targeting in advanced chor-doma: case report and literature review. BMC Cancer 2011;11:423.

24. Di Maio S, Yip S, Al Zhrani GA, Alotaibi FE, Al Turki A, Kong E, et al. Noveltargeted therapies in chordoma: an update. Ther Clin Risk Manag 2015;11:873–83.

25. Bompas E, Le Cesne A, Tresch-Bruneel E, Lebellec L, Laurence V, Collard O,et al. Sorafenib in patients with locally advanced and metastatic chordo-mas: a phase II trial of the French SarcomaGroup (GSF/GETO). AnnOncol2015;26:2168–73.

26. Bruderlein S, Sommer JB, Meltzer PS, Li S, Osada T, Ng D, et al. Molecularcharacterization of putative chordoma cell lines. Sarcoma 2010;2010:630129.

27. Bosotti R, Magnaghi P, Di Bella S, Cozzi L, Cusi C, Bozzi Fet al Establish-ment and genomic characterization of the new chordoma cell line Chor-IN-1. Sci Rep 2017 Aug 23;7(1):9226.

28. Scheil S, Bruderlein S, Liehr T, Starke H, Herms J, SchulteM, et al. Genome-wide analysis of sixteen chordomas by comparative genomic hybridizationand cytogenetics of the first human chordoma cell line, U-CH1. GenesChromosomes Cancer 2001;32:203–11.

29. Davies JM, Robinson AE, Cowdrey C, Mummaneni PV, Ducker GS, ShokatKM, et al. Generation of a patient-derived chordoma xenograft andcharacterization of the phosphoproteome in a recurrent chordoma. JNeurosurg 2014;120:331–6.

30. Rinner B, Froehlich EV, Buerger K, Knausz H, Lohberger B, Scheipl S, et al.Establishment and detailed functional and molecular genetic characteri-sation of a novel sacral chordoma cell line, MUG-Chor1. Int J Oncol2011;40:443–51.

31. Owen JH,KomarckCM,WangAC, AbuzeidWM,KeepRF,McKeanEL, et al.UM-Chor1: establishment and characterization of the first validated clivalchordoma cell line J Neurosurg 2017 Apr 21:1–9. [Epub ahead of print]

32. Rege-Cambrin G, Scaravaglio P, Carozzi F, Giordano S, Ponzetto C,Comoglio PM, et al. Karyotypic analysis of gastric carcinoma cell linescarrying an amplified c-met oncogene. Cancer Genet Cytogenet 1992;64:170–3.

33. McDermott U, Ames RY, Iafrate AJ, Maheswaran S, Stubbs H, Greninger P,et al. Ligand-dependent platelet-derived growth factor receptor (PDGFR)-alpha activation sensitizes rare lung cancer and sarcoma cells to PDGFRkinase inhibitors. Cancer Res 2009;69:3937–46.

34. Zou HY, Li Q, Lee JH, Arango ME, McDonnell SR, Yamazaki S, et al. Anorally available small-molecule inhibitor of c-Met, PF-2341066, exhibitscytoreductive antitumor efficacy through antiproliferative and antiangio-genic mechanisms. Cancer Res 2007;67:4408–17.

35. Davis MI, Hunt JP, Herrgard S, Ciceri P, Wodicka LM, Pallares G, et al.Comprehensive analysis of kinase inhibitor selectivity. Nat Biotechnol2011;29:1046–51.

36. YouWK, Sennino B,WilliamsonCW, FalconB,HashizumeH, Yao LC, et al.VEGF and c-Met blockade amplify angiogenesis inhibition in pancreaticislet cancer. Cancer Res 2011;71:4758–68.

37. Christensen JG, Schreck R, Burrows J, Kuruganti P, Chan E, Le P, et al. Aselective small molecule inhibitor of c-Met kinase inhibits c-Met-depen-dent phenotypes in vitro and exhibits cytoreductive antitumor activity invivo. Cancer Res 2003;63:7345–55.

38. Sun L, Liang C, Shirazian S, Zhou Y, Miller T, Cui J, et al. Discovery of 5-[5-fluoro-2-oxo-1,2- dihydroindol-(3Z)-ylidenemethyl]-2,4- dimethyl-1H-pyrrole-3-carboxylic acid (2-diethylaminoethyl)amide, a novel tyrosinekinase inhibitor targeting vascular endothelial and platelet-derived growthfactor receptor tyrosine kinase. J Med Chem 2003;46:1116–9.

39. Buchdunger E, Cioffi CL, Law N, Stover D, Ohno-Jones S, Druker BJ, et al.Abl protein-tyrosine kinase inhibitor STI571 inhibits in vitro signal trans-duction mediated by c-kit and platelet-derived growth factor receptors.J Pharmacol Exp Ther 2000;295:139–45.

40. Mathias TJ, Natarajan K, Shukla S,Doshi KA, SinghZN, Ambudkar SV, et al.et al.The FLT3 and PDGFR inhibitor crenolanib is a substrate of themultidrug resistance protein ABCB1 but does not inhibit transport func-tion at pharmacologically relevant concentrations. Invest New Drugs2015;33:300–9.

41. Li D, Ambrogio L, Shimamura T, Kubo S, Takahashi M, Chirieac LR, et al.BIBW2992, an irreversible EGFR/HER2 inhibitor highly effective in pre-clinical lung cancer models. Oncogene 2008;27:4702–11.

42. Solca F, Dahl G, Zoephel A, Bader G, Sanderson M, Klein C, et al. Targetbinding properties and cellular activity of afatinib (BIBW 2992), anirreversible ErbB family blocker. J Pharmacol Exp Ther 2012;343:342–50.

43. ModjtahediH, Cho BC,MichelMC, Solca F. A comprehensive review of thepreclinical efficacy profile of the ErbB family blocker afatinib in cancer.Naunyn Schmiedebergs Arch Pharmacol 2014;387:505–21.

Magnaghi et al.

Mol Cancer Ther; 17(3) March 2018 Molecular Cancer Therapeutics612

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

44. Karaman MW, Herrgard S, Treiber DK, Gallant P, Atteridge CE, CampbellBT, et al. A quantitative analysis of kinase inhibitor selectivity. Nat Bio-technol 2008;26:127–32.

45. Rabindran SK, Discafani CM, Rosfjord EC, Baxter M, Floyd MB, Golas J,et al. Antitumor activity of HKI-272, an orally active, irreversible inhibitorof the HER-2 tyrosine kinase. Cancer Res 2004;64:3958–65.

46. Brzezniak C, Carter CA, Giaccone G. Dacomitinib, a new therapy for thetreatment of non-small cell lung cancer. Expert Opin Pharmacother2013;14:247–53.

47. Cross DA, Ashton SE, Ghiorghiu S, Eberlein C, Nebhan CA, Spitzler PJ,et al. AZD9291, an irreversible EGFR TKI, overcomes T790M-mediatedresistance to EGFR inhibitors in lung cancer. Cancer Discov 2014;4:1046–61.

48. Walter AO, Sjin RT,HaringsmaHJ,OhashiK, Sun J, LeeK, et al.Discovery ofa mutant-selective covalent inhibitor of EGFR that overcomes T790M-mediated resistance in NSCLC. Cancer Discov 2013;3:1404–15.

49. Scaltriti M, Elkabets M, Baselga J. Molecular pathways: AXL, a mem-brane receptor mediator of resistance to therapy. Clin Cancer Res2016;22:1313–7.

50. Bozzi F, Manenti G, Conca E, Stacchiotti S, Messina A, Dagrada G, et al.Development of transplantable human chordoma xenograft for preclinicalassessment of novel therapeutic strategies. Neuro Oncol 2014;16:72–80.

51. Scheipl S, Barnard M, Cottone L, Jorgensen M, Drewry DH, Zuercher WJ,et al. EGFR inhibitors identified as a potential treatment for chordoma in afocused compound screen. J Pathol 2016;239:320–34.

52. Janjigian YY, Viola-Villegas N, Holland JP, Divilov V, Carlin SD, Gomes-DaGama EM, et al. Monitoring afatinib treatment in HER2-positive gastriccancer with 18F-FDG and 89Zr-trastuzumab PET. J Nucl Med 2013;54:936–43.

53. Pillay N, Plagnol V, Tarpey PS, Lobo SB, Presneau N, Szuhai K, et al. Acommon single-nucleotide variant in T is strongly associated with chor-doma. Nat Genet 2012;44:1185–7.

www.aacrjournals.org Mol Cancer Ther; 17(3) March 2018 613

Afatinib Targets EGFR and Brachyury in Chordoma

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324

2018;17:603-613. Published OnlineFirst December 13, 2017.Mol Cancer Ther   Paola Magnaghi, Barbara Salom, Liviana Cozzi, et al.   Ability to Target EGFR and BrachyuryAfatinib Is a New Therapeutic Approach in Chordoma with a Unique

  Updated version

  10.1158/1535-7163.MCT-17-0324doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://mct.aacrjournals.org/content/suppl/2017/12/13/1535-7163.MCT-17-0324.DC1

Access the most recent supplemental material at:

   

   

  Cited articles

  http://mct.aacrjournals.org/content/17/3/603.full#ref-list-1

This article cites 51 articles, 14 of which you can access for free at:

  Citing articles

  http://mct.aacrjournals.org/content/17/3/603.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.org

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/17/3/603To request permission to re-use all or part of this article, use this link

on October 22, 2020. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst December 13, 2017; DOI: 10.1158/1535-7163.MCT-17-0324