Anti-Endosialin Antibody–Drug Conjugate: Potential in Sarcoma and Other Malignancies Cecile Rouleau, Diego A

Published OnlineFirst July 16, 2015; DOI: 10.1158/1535-7163.MCT-15-0312

Large Molecule Therapeutics Molecular Cancer Therapeutics Anti-Endosialin Antibody–Drug Conjugate: Potential in Sarcoma and Other Malignancies Cecile Rouleau, Diego A. Gianolio, Robert Smale, Stephanie D. Roth, Roy Krumbholz, Jay Harper, Kenneth J. Munroe, Tessa L. Green, Bruce C. Horten, Steven M. Schmid, and Beverly A. Teicher

Abstract

Endosialin/TEM1/CD248 is a cell surface protein expressed at were grown as xenograft tumors in nude mice. The SK-N-AS high levels by the malignant cells of about 50% of sarcomas and neuroblastoma and the A-673 Ewing sarcoma lines were select- neuroblastomas. The antibody–drug conjugate (ADC) anti-endo- ed for in vivo efficacy testing of the anti-endosialin-MC-VC- sialin-MC-VC-PABC-MMAE was selectively cytotoxic to endosia- PABC-MMAE conjugate. The treatment groups included a vehi- lin-positive cells in vitro and achieved profound and durable cle control, unconjugated anti-endosialin, an admix control antitumor efficacy in preclinical human tumor xenograft models consisting of anti-endosialinandadoseoffreeMMAEequiv- of endosialin-positive disease. MC-VC-PABC-MMAE was conju- alent to the dose administered as the ADC, and the anti- gated with anti-endosialin with 3–4 MMAE molecules per ADC. endosialin-MC-VC-PABC-MMAE conjugate. The unconjugated The anti-endosialin-MC-VC-PABC-MMAE conjugate was tested anti-endosialin had no antitumor activity and resulted in for activity in four human cell lines with varied endosialin levels. similar tumor growth as the vehicle control. The admix control The HT-1080 fibrosarcoma cells do not express endosialin, A-673 produced a modest tumor growth delay. Administration of the Ewing sarcoma cells and SK-N-AS neuroblastoma cells are mod- anti-endosialin-MC-VC-PABC-MMAE conjugate resulted in a erate expressers of endosialin, and SJSA-1 osteosarcoma cells marked prolonged tumor response of both xenograts. These express very high levels of endosialin. To determine whether proof-of-concept results break new ground and open a prom- endosialin expression was maintained in vivo, A-673 Ewing sar- ising drug discovery approach to these rare and neglected coma, SK-N-AS neuroblastoma, and SJSA-1 osteosarcoma cells tumors. Mol Cancer Ther; 14(9); 2081–9. 2015 AACR.

Introduction types, is associated with tumor neovascularization and inflam- mation and has emerged as a molecular marker and therapeutic Endosialin/CD248/TEM1, a transmembrane glycoprotein target for sarcoma (1–12). First recognized as the antigen of an expressed on pericytes and fibroblasts during tissue develop- antibody raised in mice against human fetal fibroblasts (FB5), ment and present in the adult in several mesenchymal cell endosialin was found to be expressed by human solid tumor vasculature (13) and was detected in a subset of cells enriched Genzyme Corporation, Framingham, Massachusetts. for endothelium via selection with P1H12, an anti-CD146 Note: Supplementary data for this article are available at Molecular Cancer antibody, from a colorectal tumor specimen (14). Endosialin Therapeutics Online (http://mct.aacrjournals.org/). expression in tumor vasculature occurs mainly in pericytes and stromal fibroblasts (15–17). C. Rouleau and D.A. Gianolio contributed equally to this article. In mouse embryos, endosialin/TEM1-lacZ colocalizes with Current address for Cecile Rouleau: Tufts University, Jaharis 601, 136 Harrison most vimentin-positive cells and a large portion of CD31- or Avenue, Boston, MA; current address for Diego A. Gianolio, Sanofi Oncology, desmin-positive cells. In the mouse, endosialin is expressed 500 Kendall Street, Cambridge, MA; current address for Robert Smale, Covance, 2440 S. Sepulveda Blvd., Suite 220, Los Angeles, CA; current address for throughout embryonic and adult development in mesenchymal / Stephanie D. Roth, Cherokee Nation Businesses (CNT), Research Directorate, cells related to blood vessels (18). Endosialin mice have no US Army Institute of Surgical Research, 3698 Chambers Pass, Suite B, JBSA Fort defect in pericyte recruitment, suggesting a role for endosialin in Sam Houston, TX; current address for Roy Krumbholz, 15443 Grosbeak Pass, San pericyte/endothelial cell cooperation during vascular patterning Antonio, TX; current address for Jay Harper, MedImmune Oncology Research, (3). Endosialin / mice have higher than normal bone mass due One MedImmune Way, Gaithersburg, MD; current address for Tessa L. Green, MD to increased osteoblast-mediated bone formation (1). Growth of Anderson Cancer Center, UT MD Anderson Cancer Center, Houston, TX; current CyD/CyD address for Bruce C. Horten, Integrated Oncology-New York, 521 West 57th syngeneic tumors was reduced in CD248 mice, which lack CyD/CyD Street, New York, NY; current address for Steven Schmid, Vivo Biosciences, Inc., the endosialin cytoplasmic domain. CD248 fibroblasts 3714 Hunters Point Street, San Antonio, TX; and current address for Beverly A. have impaired PDGF-BB–induced migration, decreased matrix Teicher, National Cancer Institute, 9609 Medical Center Drive, RM 4-W602, MSC metalloproteinase secretion, and higher transcript levels of the 9735, Bethesda, MD. tumor suppressors transgelin (SM22a), Hes and Hey1 (6). þ Corresponding Author: Beverly A. Teicher, National Cancer Institute, RM Endosialin is expressed by human, but not mouse, CD8 naive þ þ 4-W062, MSC9735, 9609 Medical Center Drive, Bethesda, MD 20892. Phone: T cells, specifically CD8 CCR7 CD11a low naive T cells, and on þ þ 240-276-5972; Fax: 240-276-7895; E-mail: [email protected] CD8 T cells in the thymus. Endosialin knockdown in naive CD8 doi: 10.1158/1535-7163.MCT-15-0312 T cells increased cell proliferation; thus, endosialin has opposing þ 2015 American Association for Cancer Research. functions on hematopoietic (CD8 ) and stromal cells (5).

www.aacrjournals.org 2081

Rouleau et al.

The function and mechanisms of regulation of endosialin are described previously (12). Synthetic procedures for the conjuga- still incompletely understood. Recently, several new anti-endo- tion of the linker-functionalized monomethylauristatin E sialin monoclonal antibodies became available, two recognize (MMAE) to anti-endosialin were performed as detailed in the C-type lectin-like domain-Sushi/SCR/CCP and four recognize the following published procedure reports (35, 36). Briefly, the sialomucin domain. In addition, a yeast-derived anti-endo- anti-endosialin-MC-VC-PABC-MMAE was prepared by partial sialin biobody-78 was developed (4, 19). reduction of the antibody interchain disulfides with 3.3 molar The earliest indication that endosialin may be expressed by equivalents of tris-(2-carboxyethyl)-phosphine hydrochloride in malignant cells was in the 1992 publication by Rettig and col- sodium borate buffer pH ¼ 8.0 for 2 hours at 37C. After cooling leagues who reported immunoreactivity of FB5 in several neuro- on ice, 4.8 molar equivalents of maleimide-linker-MMAE deriv- blastoma cell lines and mentioned FB5þ malignant cells in a ative in DMSO were added and the reaction was allowed to subset of sarcomas (13). Further evidence for endosialin expres- proceed for 30 minutes at 4C. Excess small molecule was sion by tumor cells came in 2005 with immunostaining of removed using QuadraPure DET (Sigma-Aldrich Co.) polystyrene malignant fibrous histiocytoma and liposarcoma showing tumor scavenging beads. Yields were 85% to 95% based upon protein cell immunoreactivity (20). Furthermore, endosialin expression recovery. was assessed in 86 formalin-fixed, paraffin-embedded human The drug/antibody ratio (DAR) was determined by C8 reversed- clinical sarcoma specimens. Immunoreactive tissue components phase LC/MS following published procedures (35, 36). The ADC were malignant cells, stromal cells, and vasculature. Seventy was deglycosylated with PNGase F overnight at 37C, dialyzed to (81%) were positive for endosialin, with 44 (51%) reaching at remove salt, and reduced to smaller fragments (light and heavy least 50% coverage of immunoreactive tissue components. Stain- chains) using 20 mmol/L dithiothreitol (DTT) for 30 minutes at ing intensity was scored on the scale 0, 1þ,2þ,3þ. All nine 37C. Aliquots were injected in the LC/MS instrument and elec- sarcoma subtypes tested included specimens with at least 50% trospray ionization mass spectra of light and heavy chains were immunoreactive tissue components positive with a minimum of recorded and deconvoluted, revealing conjugation profiles that 2þ staining intensity, indicating the high prevalence of endosialin included free, mono-, di-, and tri-conjugated species. The inte- in sarcomas (12). A retrospective analysis of diagnostic reports grated peak areas were then used to determine the weighted showed that endosialin can be detected in high-grade disease and average ratios for light and heavy chains. The DAR was determined metastasis. In disseminated human sarcoma xenografts, endosia- by doubling and adding the light and heavy chains ratios. The lin protein expression was maintained at different anatomic sites DAR for the ADC was in the range of 3 to 5. Aggregation was (7–9). An anti-endosialin MORAb-004, which is a humanized determined by size exclusion HPLC and the conjugate was >98% FB5 antibody, has completed phase I clinical trial and is currently monomeric. Purity was determined by C18 RP-HPLC and there in phase II clinical trials in endometrial cancer, ovarian cancer, was <0.5% unconjugated linker-functionalized MMAE derivative melanoma, neuroectodermal tumors, and sarcoma (21, 22). in each sample. The requirements for a cell surface molecule to be suitable as an antibody–drug conjugate (ADC) target are well established Cells (23–30). The optimal ADC has antigen recognition that is not The human sarcoma cell lines HT-1080 (fibrosarcoma), A-673 different from the unconjugated antibody. ADCs usually include (Ewing sarcoma), SJSA-1 (osteosarcoma), and the SK-N-AS neu- two to four highly potent anticancer agent small-molecule drugs. roblastoma cell line were obtained from the ATCC and used over The covalent linker that tethers the antibody to the small-mole- the next 6 months. The SK-N-AS cells underwent cytogenetic cule drug must be stable in plasma and labile when internalized analysis and MYCN FISH analysis and the A673 cells underwent by the target cell. cytogenetic analysis before use in in vivo studies (12). All cells were The drugs often used in ADCs, maytansines and dolastatins, propagated in RPMI medium supplemented with 10% heat- target microtubules. The dynamic flux of microtubules is a key inactivated FBS (Invitrogen). target of anticancer therapies. Although principally recognized in mitotic function for their role in separating the duplicate set of Flow cytometry chromosomes during cell division, microtubules are an essential Analysis of endosialin expression in live cells by flow cytometry cytoskeleton component and are critical in directional transport was conducted as described previously using a fully human of proteins and organelles, maintenance of cell motility, cell shape monoclonal antibody raised against human endosialin, anti- and scaffolding, intracellular transport, secretion, neurotransmis- endosialin-MC-VC-PABC-MMAE, and a fully human isotype con- sion and relay of signaling between cell surface receptors and the trol antibody raised against dinitrophenol (DNP; ref. 12). Sample nucleus (31, 32). The biologic function of microtubules relies on acquisition was conducted on a FACS Calibur instrument (Becton the assembly and disassembly dynamics of tubulin polymeriza- Dickinson Labware) and analysis was conducted with Flow Jo tion (33, 34). (Tree Star Inc.). An anti-endosialin-MC-VC-PABC-monomethyl auristatin E ADC was prepared and assessed in cell culture and in two human Growth inhibition tumor xenograft models, demonstrating high specificity and The cells were detached using tryspin-EDTA (Invitrogen) and profound, durable antitumor efficacy. washed once with RPMI-1640 medium supplemented with 5% FBS. The cells (2 103) were plated in a 96-well plate in RPMI supplemented with 5% FBS. After 24 hours, the cells were exposed Materials and Methods to MMAE-conjugated or unconjugated anti-endosialin for 96 Anti-endosialin-MC-VC-PABC-monomethylauristatin E hours at 37C. After 96 hours, cell number was determined with The fully human anti-endosialin antibody was generated the CellTiter-Glo reagent (Promega) using a calibration curve. through a partnership with Kyowa Hakko Kirin Co., Ltd, as Luminescence was measured with a Bio-Tek (Highland Park)

2082 Mol Cancer Ther; 14(9) September 2015 Molecular Cancer Therapeutics

Anti-Endosialin ADC

Synergy HT plate reader utilizing the associated Kineticalc soft- dimensions were measured twice weekly. The data are presented ware, Version #3.4. Luminescence data were converted to growth as mean tumor volume SD. The antitumor activity of the fraction by comparison with the luminescence for the untreated compounds was determined by calculating tumor growth delay 3 control for each cell line, and IC50 values were determined from (T-C) in days at a tumor volume of 1,500 mm . Increase-in- the graphical data. Each cell line was tested in at least two lifespan was determined as a secondary endpoint with removal independent experiments. from study due to tumor size. Fold increase-in-lifespan was calculated from median survival in days for the treated versus IHC control groups. Kaplan–Meier survival curves were prepared using IHC was performed as described previously (12). the GraphPad Prism software to determine the median survival times for each treatment group. In vivo xenografts All procedures were conducted according to a protocol Results approved by the Institutional Animal Care and Use Committee in accordance with the Federal Animal Welfare Act (9 CFR, 1992) Monomethylauristatin E (MMAE) was reacted with maleimi- in an AAALAC-accredited facility. For subcutaneous models, SK- docaproyl-valine-citrulline with a p-aminobenzylcarbamate spac- N-AS neuroblastoma cells and A-673 Ewing sarcoma cells grown er to produce MC-VC-PABC-MMAE, which was ready for conju- in culture were implanted subcutaneously (1 106) in the flanks gation with anti-endosialin. The interchain sulfhydryl groups of of nude mice (Harlan Laboratories, Inc.). The gender of the mice the antibody were reduced to allow reaction with MC-VC-PABC- matched that of the cells. Animals were euthanized by CO2 MMAE to produce the ADC in a manner analogous to the asphyxiation when weight loss reached 10% of body weight or preparation of brentuximab vedotin (37–40). The reaction con- when experiencing any sign of pain or distress. The efficacy of anti- ditions were optimized to produce ADCs with a mean number of endosialin-MC-VC-PABC-MMAE was compared with that of drug molecules in the range 3 to 5 (Fig. 1). High-performance unconjugated anti-endosialin, to a mixture of anti-endosialin liquid chromatography was used to determine the antibody:drug and free MMAE providing the same dose of MMAE as in the ratio and measure the percentage of protein that was unconju- conjugate (admix), and to the vehicle (PBS). Treatments were gated or highly conjugated. Liquid chromatography coupled with initiated when tumors reached 200 mm3. Animals were random- mass spectroscopy allowed determination of the number of drug ized by tumor size and assigned to treatment or control groups molecules present on the antibody light chain and heavy chain (n ¼ 9–10). Anti-endosialin-MC-VC-PABC-MMAE was tested on (Fig. 1). two schedules: alternate days (M, W, F) for 2 weeks (total 6 doses) In a group of 42 human sarcoma cell lines grown in monolayer and once weekly for 4 weeks. Tumor volumes were calculated culture, SJSA-1 osteosarcoma cells expressed the highest endosia- using the formula (w2 l) 0.52. Mouse weight and tumor lin levels by flow cytometry (7–9). Live SJSA-1 cells were stained

Figure 1. Chemical structure of anti-endosialin- MC-VC-PABC-MMAE. A, high- performance liquid chromatography proﬁle of anti-endosialin-MC-VC- PABC-MMAE with a drug:antibody ratio (DAR) equal to 3.9. B, deconvoluted light chain LC/MS. C, deconvoluted heavy chain LC/MS.

www.aacrjournals.org Mol Cancer Ther; 14(9) September 2015 2083

Rouleau et al.

Flow cytometry of live human SJSA-1 osteosarcoma cells 100

Unconjugated anti-endosialin 80 Anti-endosialin-MC-VC-PABC-MMAE (3.8 MMAE/mAb) Isotype control-MC-VC-PABC-MMAE (4.9 MMAE/mAb)

% of Max 40

0 100 101 102 103 FL2-H

Figure 2. Endosialin protein expression by live SJSA-1 osteosarcoma cells in culture by ﬂow cytometry as determined by unconjugated anti-endosialin (red), anti-endosialin- MC-VC-PABC-MMAE (green), and an antibody isotype control-MC-VC-PABC-MMAE (gray).

with naked anti-endosialin, anti-endosialin-MC-VC-PABC- endosialin-MC-VC-PABC-MMAE conjugate. The anti-endosia- MMAE (3.8 MMAE/mAb), or an isotype control-MC-VC-PABC- lin antibody does not cross-react with the mouse homolog of MMAE (4.9 MMAE/mAb; Fig. 2). The flow cytometry histograms endosialin, therefore the only endosialin-expressing cell in show that there is no binding of the isotype control antibody to the study were the human tumor cells. For the SK-N-AS the SJSA-1 cells, while the binding of the unconjugated endosialin experiment, treatment was initiated when the tumors reached and that of anti-endosialin-MC-VC-PABC-MMAE to the SJSA-1 200 mm3 in volume. All treatments were administered by cells is similar, indicating that conjugating the drug plus linker to intravenous injection into the tail vein on alternate days for the antibody protein did not alter the ability of the antibody to 2 weeks for a total of 6 injections. The treatment groups bind to the cell surface antigen. included a vehicle control, unconjugated anti-endosialin The anti-endosialin-MC-VC-PABC-MMAE conjugate was (20 mg/kg), an admix control consisting of anti-endosialin, tested for activity in a 96-hour growth inhibition assay in and a dose of free MMAE equivalent to the dose administered four human cell lines expressing varied endosialin levels as the ADC (20 mg/kg), and the anti-endosialin-MC-VC-PABC- (Fig. 3). The HT-1080 fibrosarcoma cells do not express endo- MMAEconjugate(20mg/kg;Fig.5A).Noneofthetreatments sialin, A-673 Ewing sarcoma cells and SK-N-AS neuroblastoma produced body weight loss in the mice (Supplementary Fig. cells are moderate expressers of endosialin, and SJSA-1 osteosar- S1).The unconjugated anti-endosialin had no antitumor activ- coma cells express very high levels of endosialin. HT-1080 cells are ity and resulted in similar tumor growth as the vehicle control. not sensitive to the anti-endosialin-MC-VC-PABC-MMAE conju- The admix control produced a tumor growth delay of 10 days. gate even at the highest concentration tested. A-673 and SK-N-AS Administration of the anti-endosialin-MC-VC-PABC-MMAE cells are similarly sensitive to the anti-endosialin-MC-VC-PABC- conjugate resulted in a marked prolonged tumor response. MMAE conjugate with IC50s of 0.5 and 0.3 mg/mL mAb, respec- Two of 9 mice were lost to tumor growth, one on day 54 and tively (Fig. 3B and C), and an IC50 of 1.5 mg/mL mAb was reached one on day 81. The experiment was terminated on day 97. in SJSA-1 cells (Fig. 3D). Survival for the mice is shown in Fig. 5B. Administration of the To determine whether endosialin expression was maintained in unconjugated anti-endosialin did not alter the survival of the vivo, A-673 Ewing sarcoma, SK-N-AS neuroblastoma, and SJSA-1 mice compared with the vehicle control. Treatment with the osteosarcoma cells were grown as xenograft tumors in nude mice. unconjugated anti-endosialin along with the MMAE-free small Tumors were collected when they reached approximately 400 molecule (admix control) produced an increased median mm3 in volume, then formalin fixed and paraffin embedded. The survival of 11 days, whereas treatment with the anti-endosia- tumor specimens were analyzed for endosialin expression by IHC lin-MC-VC-PABC-MMAE conjugate resulted in an extended and scored for staining intensity by two pathologists (Fig. 4). survival that did not reach the median by day 97 when the Although A-673 cells and SK-N-AS cells in culture expressed very experiment was terminated. similar endosialin levels, when grown in vivo, A-673 tumors The in vivo efficacy experiment with A-673 Ewing sarcoma expressed endosialin with 1þ intensity, whereas SK-N-AS tumors examined the effects of a once weekly schedule for four doses expressed endosialin with 2þ intensity. SJSA-1 tumors had a and a dose-response range of ADC doses. The treatment groups markedly heterogeneous endosialin expression pattern. Some included a vehicle control, unconjugated anti-endosialin regions of the tumor expressed endosialin with 3þ intensity and (15 mg/kg), an admix control in which anti-endosialin was some regions expressed no endosialin. administered with MMAE-free small molecule (15 mg/kg), and The SK-N-AS neuroblastoma and the A-673 Ewing sarcoma three dose levels of the anti-endosialin-MC-VC-PABC-MMAE lines were selected for in vivo efficacy testing of the anti- conjugate (1, 5, and 15 mg/kg; Fig. 6A). As in the SK-N-AS

2084 Mol Cancer Ther; 14(9) September 2015 Molecular Cancer Therapeutics

Anti-Endosialin ADC

A HT-1080 fibrosarcoma B A-673 Ewing sarcoma

µ µ IC50 >3 g/mL IC50 0.5 g/mL

µg/mL antibody µg/mL antibody

C SK-N-AS neuroblastoma D SJSA-1 osteosarcoma

µ IC50 0.3 g/mL µ IC50 1.5 g/mL

µg/mL antibody µg/mL antibody

Anti-endosialin-VC-MC-PABC-MMAE (3.8 MMAE/mAb) Isotype control-VC-MC-PABC-MMAE (4.9 MMAE/mAb)

Figure 3. Determination of endosialin protein expression by flow cytometry and cellular growth inhibition by anti-endosialin-MC-VC-PABC-MMAE and by an antibody isotype control-MC-VC-PABC-MMAE upon 96 hours of exposure to the agent for: A, human HT-1080 fibrosarcoma cells; B, human A-673 Ewing sarcoma cells; C, human SK-N-AS neuroblastoma cells; D, human SJSA-1 osteosarcoma cells. Experiments were repeated three times independently; bars, SEM. efficacy study, the unconjugated anti-endosialin had no effect prolonged antitumor effect such that 80% of the mice survived on the growth of A-673 Ewing sarcoma tumors. The admix until the experiment was terminated on day 150. control produced a modest tumor growth delay. A clear dose response was observed with the anti-endosialin-MC-VC-PABC- MMAEconjugate.Thedoseof1mg/kghadasmalleffect, Discussion whereas the dose of 5 mg/kg had a moderate effect on 50% of Bone and soft tissue sarcoma are currently treated with conven- the mice and a marked antitumor effect on 50% of the mice tional cytotoxic agents, including vincristine, dacarbazine, doxoru- (Fig. 6B). The dose of 15 mg/kg produced a marked and bicin, cyclophosphamide, cisplatin, gemcitabine, and docetaxel

Figure 4. Expression of endosialin by IHC in three human tumor xenograft specimens. Scoring of IHC endosialin staining intensity is shown.

A-673 human Ewing sarcoma SK-N-AS human neuroblastoma SJSA-1 human osteosarcoma 1+ endosialin IHC staining 2+ endosialin IHC staining 3+ endosialin IHC staining

www.aacrjournals.org Mol Cancer Ther; 14(9) September 2015 2085

Rouleau et al.

A 2,500 SK-N-AS human neuroblastoma 2,250

2,000

) Vehicle 3 1,750 Unconjugated anti-endosialin (20 mg/kg, i.v. QODx3, 2 cycles) Admix control:anti-endosialin+MMAE (20 mg/kg, i.v. QODx3, 2 cycles) 1,500 Anti-endosialin-MC-VC-PABC-MMAE (20 mg/kg, i.v. QODx3, 2 cycles)

1,250

1,000 Figure 5. A, growth of subcutaneously implanted human SK-N-AS neuroblastoma xenografts 750

Mean tumor volume (mm Mean tumor volume in female nude mice treated with vehicle (Control, &), unconjugated anti-endosialin 500 (20 mg/kg, i.v., QODx3, 2 cycles; &), anti- endosialin þ MMAE (20 mg/kg, i.v., QODx3, 250 2 cycles; ADMIX,^), or anti-endosialin-MC- VC-PABC-MMAE (20 mg/kg, i.v., QODx3, 2 0 cycles; ~). Treatments were administered 4 6 81012141618202224262830323436384042444648505254565860626466687072747678808284 by intravenous injection 3-times per week Days after tumor cell implantation for 2 weeks. The data are the mean SD for groups of 9 mice. B, survival of female nude Rx :4, 6, 8, 11, 13, 15 mice bearing subcutaneously implanted human SK-N-AS neuroblastoma xenografts after treatment with vehicle (Control, &), B SK-N-AS human neuroblastoma unconjugated anti-endosialin (20 mg/kg, i. v., QODx3, 2 cycles; &), anti-endosialin þ 100 MMAE (20 mg/kg, i.v., QODx3, 2 cycles; 90 ADMIX, ^), or anti-endosialin-MC-VC- PABC-MMAE (20 mg/kg, i.v., QODx3, 2 80 cycles; ~). Treatments were administered by intravenous injection three times per 70 week for 2 weeks. The data are the mean 60 SD for groups of 9 mice. Vehicle 50 Unconjugated anti-endosialin (20 mg/kg, i.v. QODx3, 2 cycles) Admix control:anti-endosialin+MMAE (20 mg/kg, i.v. 40 QODx3, 2 cycles) 30 Anti-endosialin-MC-VC-PABC-MMAE (20 mg/kg, i.v. QODx3, 2 cycles) 20 Percent of mice remaining (%) Percent 10

0 4 6 81012141618202224262830323436384042444648505254565860626466687072747678 Days after tumor cell implantation

(41). High-risk neuroblastoma is currently treated with anthracy- subset of malignant cells. Ontuxizumab is being investigated as a clines, alkylators, platinum compounds, topoisomerase II inhibi- monoclonal antibody for the treatment of several types of cancer in tors, radiotherapy, and myeloablative therapy (42). New therapies adults and children and has received orphan drug designation for for these rare and neglected tumors are urgently needed. ADCs are sarcoma. A phase I study in pediatric patients with recurrent or now a clinically established therapeutic modality, most recently refractory solid tumors or lymphoma was reported recently (43). A exempliﬁed in the approval of trastuzumab emtansine (T-DM1; phase I study of ontuxizumab in Japanese adults with solid tumors Kadcyla), in which the monoclonal anti-HER2 antibody trastuzu- has been completed (44). A phase II sarcoma study, which utilizes mab is covalently linked to the maytansine derivative DM1, for adaptive design to identify subgroup populations during study, is HER2-positive metastatic breast cancer. The accessibility of endo- underway (45). sialin, as a cell surface antigen, to antibodies and its high level of The function of endosialin remains largely unknown and it expression in sarcomas and neuroblastomas makes it a potentially appears to be expressed at relatively high levels by about 50% of suitable ADC target (10). A humanized monoclonal anti-endosialin sarcomas. The implication is that endosialin is not essential for antibody, ontuxizumab (MORAb-004), is currently being tested in cell viability or proliferation. The requirements for an ADC target phase I and II clinical trials. Ontuxizumab binds to endosialin on include being expressed on the cell surface and that an antibody to tumor vascular pericytes, tumor stromal cells, and directly on a the extracellular domain of the target protein be internalized into

2086 Mol Cancer Ther; 14(9) September 2015 Molecular Cancer Therapeutics

Anti-Endosialin ADC

A A673 human Ewing sarcoma 2,400

2,200

2,000 Control ) 1,800 3

1,600 Naked Ab (15 mg/kg)

Figure 6. 1,400 Admix (15 mg/kg) A, growth of subcutaneously implanted human A-673 Ewing 1,200 Endo-MMAE (1 mg/kg) sarcoma xenografts in female nude 1,000 mice treated with vehicle (Control, Endo-MMAE (5 mg/kg) & ), unconjugated (naked) anti- 800 endosialin (15 mg/kg, i.v., once Endo-MMAE (15 mg/kg) weekly 4; *), anti-endosialin þ

Mean tumor volume (mm Mean tumor volume 600 MMAE (15 mg/kg, i.v., once weekly 4; ADMIX, ^), or anti-endosialin- 400 MC-VC-PABC-MMAE (1 mg/kg, i.v., once weekly 4; ), anti- 200 endosialin-MC-VC-PABC-MMAE (5 mg/kg, i.v., once weekly 4; ~), 0 or anti-endosialin-MC-VC-PABC- 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 MMAE (15 mg/kg, i.v., once weekly 4; !). The data are the Days after tumor cell implantation mean SD for groups of 10 mice. B, survival of female nude mice bearing subcutaneously implanted A673 human Ewing sarcoma human A-673 Ewing sarcoma B xenografts after treatment with 100 vehicle (Control, ), unconjugated (naked) anti-endosialin (15 mg/kg, 90 i.v., once weekly 4; ), anti- endosialin þ MMAE (15 mg/kg, i.v., 80 once weekly 4; ADMIX, ), or anti-endosialin-MC-VC-PABC- 70 MMAE (1 mg/kg, i.v., once Control weekly 4; ), anti-endosialin- Naked Ab (15 mg/kg) MC-VC-PABC-MMAE (5 mg/kg, i.v., 60 Admix (15 mg/kg ) Endo-MMAE (1 mg/kg) once weekly 4; ), or anti- 50 Endo-MMAE (5 mg/kg) endosialin-MC-VC-PABC-MMAE Endo-MMAE (15 mg/kg) (15 mg/kg, i.v., once weekly 4; ). The data are the mean SD 40 for groups of 10 mice. 30

Percent of mice remaining (%) Percent 20

0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150

Days after tumor cell implantation the cell along with the target protein to enable delivery of the detection of endosialin protein using the IHC assay previously cytotoxic drug pay-load. Some ADC targets such as CD30 occur described (12) and used herein should be a key component of any at high levels on the cell surface of specific diseases, anaplastic endosialin-directed therapy as a companion diagnostic to select large cell lymphoma, and Hodgkin lymphoma with sufficient potential responders. frequency that diagnostic testing for the target is not essential. This present proof-of-concept study demonstrates that Brentuximab vedotin (SGN-35) reached FDA approval in 2011 endosialin-positive tumors can be specifically and effectively for treatment of refractory Hodgkin lymphoma and systemic targeted by a monoclonal anti-endosialin antibody conjugat- anaplastic large cell lymphoma without the requirement for ed to the potent cytotoxic small-molecule MMAE. The testing tumor cells for the CD30 (46). Given the variability of response is complete and durable, warranting extensive pre- endosialin expression across tumors and within tumors (7–9, 12), clinical evaluation of endosialin-directed ADCs in endosialin-

www.aacrjournals.org Mol Cancer Ther; 14(9) September 2015 2087

Rouleau et al.

positive disease. Anti-endosialin-MMAE brings the promise of Analysis and interpretation of data (e.g., statistical analysis, biostatistics, personalized medicine to sarcoma and neuroblastoma computational analysis): C. Rouleau, D.A. Gianolio, J. Harper, K.J. Munroe, patients. B.C. Horten, S.M. Schmid, B.A. Teicher Writing, review, and/or revision of the manuscript: C. Rouleau, D.A. Gianolio, J. Harper, B.C. Horten, B.A. Teicher Disclosure of Potential Conﬂicts of Interest Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Rouleau, R. Smale, R.D. Krumbholz, J. Harper No potential conﬂicts of interest were disclosed. Study supervision: D.A. Gianolio, S.D. Roth, J. Harper, S.M. Schmid, B.A. Teicher Authors' Contributions Conception and design: C. Rouleau, D.A. Gianolio, J. Harper, B.A. Teicher Grant Support Development of methodology: C. Rouleau, D.A. Gianolio, J. Harper, This work was funded in whole or in part with funds from Genzyme S.M. Schmid, B.A. Teicher Corporation. Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Rouleau, D.A. Gianolio, R. Smale, R.D. Krumbholz, Received April 20, 2015; revised June 23, 2015; accepted July 7, 2015; J. Harper, K.J. Munroe, T.L. Green published OnlineFirst July 16, 2015.

References 1. Naylor AJ, Azzam E, Smith S, Croft A, Poyser C, Duffield JS, et al. The 17. Lax S, Hardie DL, Wilson A, Douglas MR, Anderson G, Huso D, et al. The mesenchymal stem cell marker CD248 (Endosialin) is a negative regulator pericyte and stromal cell marker CD248 (endosialin) is required for of bone formation in mice. Arthritis Rheum 2012;64:3334–43. efficient lymph node expansion. Eur J Immunol 2010;40:1884–9. 2. Naylor AJ, McGettrick HM, Maynard WD, May P, Barone F, Croft AP, et al. A 18. Huang HP, Hong CL, Kao CY, Lin SW, Lin SR, Wu HL, et al. Gene taregting differential role for CD248 (endosialin) in PDGF-mediated skeletal muscle and expression analysis of mouse Tem1/endosialin using a lacZ reporter. angiogenesis. PLos One 2014;9:e107146. Gene Express Patterns 2011;11:316–26. 3. Simonavicius N, Ashenden M, van Weverwijk A, Lax S, Husoi DL, Buckley 19. Zhao A, Nunez-Cruz S, Li C, Coukos G, Siegel DL, Scholler N. Rapid CD, et al. Pericytes promote selective vessel regression to regulate vascular isolation of high-affinity human antibodies against the tumor vascular patterning. Blood 2012;120:1516–27. marker endosialin/TEM1, using a paired east-display/secretory scFv library 4. Kontesekova S, Repic AO, Polcicova K, Tuomaala P, Pastorek J, Pastorekova platform. J Immunol Methods 2011;363:221–32. S, et al. Novel monoclonal antibodies specific for CTLD-SSC and sialo- 20. Dolznig H, Schweifer N, Puri C, Kraut N, Rettig WJ, Kerjaschki D, et al. mucin domains of endosialin, a mural cell marker of tumor vasculature. Characterization of cancer stroma markers: in silico analysis of an mRNA Int J Oncol 2012;41:1365–72. expression database for fibroblast activation protein and endosialin. 5. HardieDL,BaldwinMJ,NaylorA,HaworthOJ,ZhengT,LaxS,etal.The Cancer Immun 2005;5:1–8. stromal cell antigen CD248 (endosialin) is expressed on naive CD8þ 21. Carvajal RD, Gounder MM, Coughlin CM, Fishel J, Heyburn J, O'Shannessy human T cells and regulates proliferation. Immunology 2011;133: D, et al. A first-in-human phase I study of MORAb-004 (MOR4), a 288–95. humanized monoclonal antibody recognizing TEM-1 (endosialin), in 6. Maia M, DeVriese A, Janssens T, Moons M, Lories RJ, Tavernier J, et al. patients with solid tumors. J Clin Oncol 2011;29:Abstr 3086. CD248 facilitates tumor growth via its cytoplasmic domain. BMC Cancer 22. Diaz LA, Coughlin CM, Weil SC, Fishel J, Gounder MM, Lawrence S, et al. A 2011;11:162–73. first-in-human phase I study of MORAb-004, a monoclonal antibody to 7. Rouleau C, Sancho J, Campos-Rivera J, Teicher BA. Endosialin expression in endosialin in patients with advanced solid tumors. Clin Cancer Res side populations in human sarcoma cell lines. Oncol Lett 2011;2:325–9. 2015;21:1281–8. 8. Rouleau C, Smale R, Sancho J, Fu YS, Kurtzberg L, Weber W, et al. 23. Teicher BA. Antibody-drug conjugate targets. Curr Cancer Drug Targets Endosialin: a novel malignant cell therapeutic target for neuroblastoma. 2009;9:982–1004. Int J Oncol 2011;39:841–51. 24. Miller ML, Roller EE, Wu X, Leece BA, Goldmacher VS, Chari RV, et al. 9. Rouleau C, Smale R, Fu YS, Hui G, Wang F, Hutto E, et al. Endosialin is Synthesis of potent taxoids for tumor-specific delivery using monoclonal expressed in high grade and advanced sarcomas: evidence from clinical antibodies. Bioorg Med Chem Lett 2004;14:4079–82. specimens and preclinical modeling. Int J Oncol 2011;39:73–89. 25. Hellstrom I, Hellstrom KE, Senter PD. Development and activities of the 10. Valdez Y, Maia M, Conway EM. CD248: Reviewing its role in health and BR96-doxorubicin immunoconjugate. Methods Mol Biol 2001;166:3–16. disease. Curr Drug Targ 2012;13:432–9. 26. Chari RV, Jackel KA, Bourret LA, Derr SM, Tadayoni BM, Mattocks KM, et al. 11. Bagley RG. Endosialin: from vascular target to biomarker for human Enhancement of the selectivity and antitumor efficacy of a CC-1065 sarcomas. Biomark Med 2009;3:589–604. analogue through immunoconjugate formation. Cancer Res 1995;55: 12. Rouleau C, Curiel M, Weber W, Smale R, Kurtzberg L, Mascarello J, et al. 4079–84. Endosialin protein expression and therapeutic target potential in human 27. Chari RV, Martell BA, Gross JL, Cook SB, Shah SA, Bl€attler WA, et al. solid tumors: sarcoma versus carcinoma. Clin Cancer Res 2008;14: Immunoconjugates containing novel maytansinoids: promising antican- 7223–36. cer drugs. Cancer Res 1992;52:127–31. 13. Rettig WJ, Garin-Chesa P, Healey JH, Su SL, Jaffe EA, Old LJ. Identification 28. Scott CF Jr, Goldmacher VS, Lambert JM, Chari RV, Bolender S, Gauthier of endosialin, a cell surface glycoprotein of vascular endothelial cells in MN, et al. The antileukemic efficacy of an immunotoxin composed of a human cancer. Proc Natl Acad Sci U S A 1992;89:10832–6. monoclonal anti-Thy-1 antibody disulfide linked to the ribosome-inacti- 14. St. Croix B, Rago C, Velculescu V, Traverso G, Romans KE, Montgomery E, vating protein gelonin. Cancer Immunol Immunother 1987;25:31–40. et al. Genes expressed in human tumor endothelium. Science 2000;289: 29. Coombes RC, Buckman R, Forrester JA, Shepherd V, O'Hare MJ, Vincent M, 1197–202. et al. In vitro and in vivo effects of a monoclonal antibody-toxin conjugate 15. Carson-Walter EB, Winans BN, Whiteman MC, Liu Y, Jarvela S, Haapasalo for use in autologous bone marrow transplantation for patients with breast H, et al. Characterization of TEM1/endosialin in human and murine brain cancer. Cancer Res 1986;46:4217–20. tumors. BMC Cancer 2009;9:417. 30. Lambert JM, Senter PD, Yau-Young A, Blattler WA, Goldmacher VS. 16. O'Shannessy DJ, Somers EB, Chandrasekaran LK, Nicolaides NC, Bordeaux Purified immunotoxins that are reactive with human lymphoid cells. J, Gustavson MD. Influence of tumor microenvironment on prognosis in Monoclonal antibodies conjugated to the ribosome-inactivating proteins colorectal cancer: tissue architecture-dependent signature of endosialin gelonin and the pokeweed antiviral proteins. J Biol Chem 1985;260: (TEM-1)and associated proteins. Oncotarget 2014;5:3983–95. 12035–41.

2088 Mol Cancer Ther; 14(9) September 2015 Molecular Cancer Therapeutics

Anti-Endosialin ADC

31. Rowinsky EK, Donehower RC. Antimicrotubule agents. In:Chabner BA, specific cytotoxicity by peptide-linked anti-CD30-auristatin conjugates. Longo DL, editors. Cancer Chemotherapy and Biotherapy. New York: J Biol Chem 2006;281:10540–7. Lippincott-Raven; 1996. p. 263–96. 40. Francisco JA, Cerveny CG, Meyer DL, Mixan BJ, Klussman K, Chace DF, et al. 32. Bhat KM, Setaluri V. Microtubule-associated proteins as targets in cancer cAC10-vcMMAE, an anti-CD30-monomethyl auristatin E conjugate with chemotherapy. Clin Cancer Res 2007;13:2849–54. potent and selective antitumor activity. Blood 2003;102:1458–65. 33. Jordan MA, Thrower D, Wilson L. Mechanism of inhibition of cell prolif- 41. Teicher BA. Perspective: opportunities in recalcitrant, rare and neglected eration by Vinca alkaloids. Cancer Res 1991;51:2212–22. tumors. Oncol Rep 2013;30:1030–4. 34. Jordan MA, Wilson L. Microtubules as a target for anticancer drugs. Nat Rev 42. Irwin MS, Park JR. Neuroblastoma: paradigm for precision medicine. Cancer 2004;4:253–65. Pediatr Clin North Am 2015;62:225–56. 35. Gianolio DA, Rouleau C, Bauta WE, Lovett D, Cantrell WR, Recio A III, 43. Norris RE, Fox E, Reid JM, Ralya AT, Liu X, Ahern CH, et al. Phase I trial of et al. Targeting HER2-positive cancer with dolastatin 15 derivatives ontuxizumab (MORAb-004) in children with relaspsed or refractory solid conjugated to trastuzumab, novel antibody-drug conjugates. Cancer tumors: A Children's Oncology Group Study. J Clin Oncol 2015;33(suppl): Chemother Pharmacol 2012;70:439–49. Abstr 10036. 36. Stefano JE, Busch M, Hou L, Park A, Gianolio DA. Micro- and mid-scale 44. Ikeda M, Aramaki T, Inaba Y, Nakai K, Doi T, Oksuaka T, et al. Phase I study maleimide-based conjugation of cytotoxic drugs to antibody hinge region of ontuxizumab, a humanized monoclonal antibody (mAb) recognizing thiols for tumor targeting. Methods Mol Biol 2013;1045:145–71. endosialin in Japanese patients (pts) with hepatocellular carcinoma 37. Deng C, Pan b, O'Connor OA. Brentuximab vedotin. Clin Cancer Res (HCC). J Clin Oncol 2015;33(suppl): Abstr 2521. 2013;19:22–7. 45. Chawla SP, Schoffski P, Blay JY, Jones RL, Gelderblom H, Attia S, et al. A 38. Alley SC, Okeley NM, Senter PD. Antibody-drug conjugates: targeted drug study of the safety and efficacy of the combination of gemcitabine and delivery for cancer. Curr Opin Chem Biol 2010;14:529–37. docetaxel with on tuxizumab (MORAb-004) in metastatic soft tissue 39. Sutherland MSK, Sanderson RJ, Gordon KA, Andreyka J, Cerveny CG, Yu C, sarcoma. J Clin Oncol 33, 2015 (suppl): Abstr TPS10577. et al. Lysosomal trafficking and cysteine protease metabolism confer target- 46. Teicher B. Antibody drug conjugates. Curr Opin Oncol 26, 2014:476–83.

www.aacrjournals.org Mol Cancer Ther; 14(9) September 2015 2089

Anti-Endosialin Antibody−Drug Conjugate: Potential in Sarcoma and Other Malignancies

Cecile Rouleau, Diego A. Gianolio, Robert Smale, et al.

Mol Cancer Ther 2015;14:2081-2089. Published OnlineFirst July 16, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-15-0312

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2015/07/17/1535-7163.MCT-15-0312.DC1

Cited articles This article cites 42 articles, 15 of which you can access for free at: http://mct.aacrjournals.org/content/14/9/2081.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/14/9/2081.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/14/9/2081. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.