An Antibody–Drug Conjugate Targeting the Endothelin B Receptor for the Treatment of Melanoma

Published OnlineFirst January 18, 2011; DOI: 10.1158/1078-0432.CCR-10-2340

Clinical Cancer Cancer Therapy: Preclinical Research

Jyoti Asundi, Chae Reed, Jennifer Arca, Krista McCutcheon, Ronald Ferrando, Suzanna Clark, Elizabeth Luis, Janet Tien, Ron Firestein, and Paul Polakis

Abstract Purpose: To identify and evaluate targets amenable to antibody therapy in melanoma. Experimental Design: We searched for mRNA transcripts coding for cell-surface proteins with expression patterns similar to that of the melanoma oncogene MITF. One such candidate, the endothelin B receptor (EDNBR), was first analyzed for a functional contribution to tumor growth by conditional induction of shRNA. Second, antibodies were raised to the receptor, conjugated with monomethyl auristatin E, and tested for efficacy against melanoma tumor models generated from cell lines. Results: Conditional knockdown of the receptor in tumor xenograft models resulted in only a modest impact on tumor growth. A monoclonal antibody reactive with the N-terminal tail of EDNBR was found to internalize rapidly into melanoma cells. When conjugated with monomethyl auristatin E, the antibody– drug conjugate (ADC) showed remarkable efficacy against human melanoma cell lines and xenograft tumor models that was commensurate with levels of receptor expression. Comparative immunohistochemistry revealed a range of EDNBR expression across a panel of human melanomas, with the majority expressing levels equivalent to or greater than that in the models responsive to the ADC. Conclusion: An ADC targeting the EDNBR is highly efficacious in preclinical models of melanoma. Clin Cancer Res; 17(5); 1–11. 2011 AACR.

Introduction various cell types, including melanocytes and melanoma cells, pointed to functions independent of their role in Melanoma is an aggressive form of skin cancer that has endothelium. It is now well recognized that the endothelin recently undergone an alarming increase in incidence (1). B receptor (EDNBR) is critical for the faithful derivation of Although cures can be achieved with surgical resection of melanocytic cells emanating from the neural crest during localized lesions, the advanced stages of melanoma are embryonic development (6, 7). Melanocyte precursors rely only poorly responsive to currently approved therapies. on EDNBR activity to proliferate and migrate from the The 5-year survival rate for stage IV metastatic melanoma is neural tube to their final destinations (8, 9). Mice with approximately 10% (1). New therapeutic approaches, defective genes coding for either EDNBR or endothelin-3 including antisense to Bcl2, antibodies to CTLA4, small (ET-3) exhibit a pigmentation deficit in their coats and a molecule RAF kinase inhibitors, and adoptive immu- shortage of enteric ganglion cells, also derived from the notherapy, are currently in clinical testing for metastatic neural crest. These characteristics strongly resemble those melanoma (2). The results from some of these recent associated with the WS4 variant of Waardenburg syndrome studies seem to be encouraging, but a durable impact on in humans, which has been attributed to germline muta- overall survival will likely require therapeutic combina- tions in either ET-3 or EDNBR (10–12). An additional tions including additional new agents. variant of this syndrome, WS2, has been mapped to heri- More than 20 years ago, endothelin-1 (ET-1) was isolated table mutations in the microphthalmia-associated tran- from aortic endothelial cells and found to have potent scription factor (MITF), a key regulator of melanocyte vasoconstrictive activity (3). The receptors for endothelins development and a melanoma proto-oncogene (13–15). were cloned shortly thereafter (4, 5) and their expression in The strong genetic evidence linking EDNBR activity to the fate of melanoblasts underscores a potential role for Authors' Affiliation: Genentech, Inc., South San Francisco, California this receptor in the progression of melanoma. The expression of EDNBR mRNA and protein was reported to increase Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). during disease progression from dysplastic nevi to meta- Corresponding Author: Paul Polakis, Genentech, Inc., 1 DNA Way, South static melanoma (16). Blockade of EDNBR activity by 2 San Francisco, CA 94080. Phone: 650-225-5327; Fax: 650-225-6127; independent small molecule inhibitors interfered with E-mail: [email protected] growth and survival of melanoma cells and tumor xeno- doi: 10.1158/1078-0432.CCR-10-2340 grafts (17–19). These preclinical studies implicate EDNBR 2011 American Association for Cancer Research. as a potential driver of melanoma progression. However,

www.aacrjournals.org OF1

Asundi et al.

Translational Relevance transfected with full-length EDNBR clone or purified EDNBR protein. Immunization, hybridoma selection, Targeted cancer therapies exploit features unique to and antibody purification were done as described pre- the cancer cell and thereby minimize damage to normal viously (22). tissues. Accordingly, we have targeted the endothelin B For fluorescence-activated cell sorting (FACS), cells were receptor (EDNBR) on the basis of its overexpression on harvested in PBS with 2.5 mmol/L EDTA and washed in PBS buffer containing 1% FBS. All subsequent steps were human melanoma cells. This receptor is particularly attractive as a target for antibody–drug conjugate carried out at 4 C. Cells were incubated for 1 hour each m (ADC) therapy because of its minimal expression on with 2 to 5 g/mL primary antibodies, followed by the normal tissue, its cell-surface localization, and rapid appropriate secondary antibodies. Cells were then analyzed endocytosis. The receptor was expressed on the vast with a FACS Calibur flow cytometer (BD Biosciences). majority of melanoma tissue specimens examined, Primary antibodies, monoclonal or chimeric in-house gen- and the ADC showed preclinical efficacy even in a erated anti-EDNBR antibodies (5E9 and 24C7) for EDNBR model representing the low end of the expression range. cell surface detection, FLAG M2 mAb (F-3165; Sigma Life As a platform, ADCs such as SGN-35 and TDM1 have Science) for N-Term FLAG tag detection were used. Alexa recently shown very encouraging results in clinical trials. 488–conjugated anti-mouse or anti-human IgG fluorescent Our preclinical results suggest that the ADC targeting the detection reagent (A11017, A11013; Invitrogen) was used. EDNBR may provide an effective new therapy for mel- For Western blotting, the following reagents were used anoma patients. for detection of proteins: anti-EDNBR in-house generated mAb 1H1.8.5, rabbit polyclonal anti-GAPDH (glyceralde- hyde-3-phosphate dehydrogenase) antibody (PA1-987; translation of EDNBR inhibition by small molecules into Affinity Bioreagents), p44/42 mitogen-activated protein the melanoma clinic has been somewhat disappointing. kinase (MAPK) and phospho-p44/42 MAPK [(Thr202/ No objective responses were observed in phase II trials in Tyr 204) 9102, 9101S; Cell Signaling Technology], and which bosentan, a dual endothelin A/B-receptor antago- b-tubulin (556321; BD Pharmingen/BD Biosciences). nist, was administered either as monotherapy or in com- Detection reagents Alexa 680–conjugated anti-mouse or bination with dacarbazine in first-line therapy (20, 21). anti-rabbit IgG (A21057, A21076; Molecular Probes/ To identify targets suitable for antibody therapy in mel- Invitrogen) and IRDye800-conjugated anti-mouse or anoma, we carried out a search for transcripts encoding anti-rabbit IgG (610-132-121, 611-132-122; Rockland cell-surface proteins with expression correlative to that of Immunochemicals for Research) were used. MITF, the melanoma oncogene. Accordingly, EDNBR was For immunofluorescence, 10,000 cells were plated on 4- readily identified in a panel of metastatic melanomas and well chamber slides and incubated for 72 hours at 37Cin melanoma cell lines. Herein, we describe the application of 5% CO2. Cells were fixed with 4% Ultra Pure Formalde- antibodies to EDNBR and, in particular, report that EDNBR hyde (18814; Polysciences Inc.) and permeabilized with is an extremely efficient target for antibody–drug conjugate 0.05% saponin (S-4521; Sigma Life Science). Nonspecific (ADC) therapy. signal was blocked using PBS containing 5% goat serum (G-9023; Sigma Life Science) and primary and secondary Materials and Methods antibodies were applied to cells in PBS containing 2% goat serum. Live cells were incubated with 5 mg/mL anti-EDNBR EDNRA and EDNBR cDNAs antibodies for 1 hour on ice for surface expression detec- Human EDNRA cDNA from OriGene, human EDNBR tion. Internalization of antibodies was evaluated by apply- cDNA from Incyte, mouse EDNBR cDNA from Image ing 2 mg/mL anti-EDNBR chimeric antibody 5E9 for 2 (Invitrogen), cynomologus monkey EDNBR cDNA from hours at 37C to live cells in complete medium containing liver and lung cDNA libraries, and rat EDNBR cDNA from 25 mg/mL leupeptin and 10 mg/mL pepstatin (Roche neonatal cDNA library were cloned into a CMV promoter- Applied Science). Localization of internalized EDNBR driven N-terminal FLAG-tagged vector. These constructs was further determined on these cells by incubating fixed were used to generate D64 EDNBR lacking the N-terminal and permeabilized cells for 1 hour at room temperature 64 amino acids and the EDNRA/EDNBR chimeras in which with antibodies for organelle markers; 4 mg/mL antibody one of each of the 4 extracellular domains of EDNRA was for lysosomal-associated membrane protein 1 [(LAMP1) substituted into EDNBR. The N-His-DSP-EDNBR cDNA L1418; Sigma Life Science] or 5 mg/mL early endosomal coding for N-terminal 8xHis-tagged EDNBR protein with- antigen 1 (EEA1, 324610; EMD Biosciences). Cy3-conju- out native signal peptide was subcloned into Baculovirus gated anti-mouse or anti-human detection reagents (115- vector pAcGP67-B (BD Biosciences) and employed for the 165-166, 109-165-098; Jackson ImmunoResearch) and production of immunogen used to raise the 5E9 antibody. Alexa 488–conjugated anti-rabbit detection reagent (A11008; Invitrogen) were used. Cells were mounted with Immunologic procedures VECTASHIELD (H-1200; Vector Labs, Inc) and sealed with For production of monoclonal antibodies (mAb), BALB/ nail polish. Images were taken at 60 magnification, using c mice were immunized with either HEK293 cells stably Nikon microscope.

OF2 Clin Cancer Res; 17(5) March 1, 2011 Clinical Cancer Research

An Antibody–Drug Conjugate Targeting the Endothelin B Receptor

EDNBR immunohistochemistry on tissue sections was noma cell lines. The experiments were controlled using carried out on the Ventana Discovery XT Platform, using shRNA targeting luciferase gene. Mutant EDNBR gene, CC1 standard antigen retrieval as a pretreatment. Sections impervious to knockdown by this shRNA, was generated were incubated with goat anti-human EDNBR primary by changing target nucleotide sequence to 50-AGCTTAAA- antibody (sc-21196; Santa Cruz Biotechnologies) at 0.05 CAGTTGTATCA-30 while maintaining the protein identity. mg/mL (1:4,000) for 60 minutes at 37C, followed by incubation with an unconjugated rabbit anti-goat IgG MAPK phosphorylation and calcium influx assays secondary antibody (Vector Labs) at 5 mg/mL (1:200) for For MAPK phosphorylation, 2 106 UACC-257 cells 32 minutes at room temperature. Sections were subse- were serum starved for 24 hours on a 60-mm dish. Anti- quently incubated for 16 minutes with anti-rabbit Omni- body 24C7 (5 mg/mL) or BQ788 sodium salt (100 nmol/L; Map HRP reagent (Ventana Medical Systems), followed by PED-3788-PI; Peptide International Inc.) was added and Ventana DAB colorimetric peroxidase substrate. Sections the cells were incubated for 1 hour at 37C prior to the were counterstained with Ventana Hematoxylin II reagent addition of 100 nmol/L human ET-1 (PED-4198-v; Peptide and subsequently dehydrated, cleared with xylenes, and International Inc.) for 5 minutes at 37C. Cell lysates were coverslipped. ChromPure goat IgG (005-000-003; Jackson probed for total or phosphorylated p44/42 MAPK and for ImmunoResearch) was used as the isotype control. The b-tubulin. Protein bands were quantitated using Odyssey human melanoma tissue samples were purchased from US infrared imaging system from LI-COR Biosciences. Phos- Biomax, Inc. The overall staining intensity of tissue samples phorylated p44/42 (Thr202/Tyr204) MAPK bands were was scored as none (0), weak (1þ), moderate (2þ), or normalized to total p44/42 MAPK and the b-tubulin load- strong (3þ). ing control. Anti-EDNBR binding sites per cell was determined by For calcium influx, 25,000 HEK293 cells transfected with Scatchard analysis essentially as described previously (22). EDNBR were plated on a 96-well black-sided plate and incubated at 37 C, 5% CO2. After 48 hours, medium was Cell culture aspirated and cells were loaded with dye for 2 hours at The cell lines A2058, A375, C32, COLO 829, G-361, Hs- 37C, using FLIPR Calcium 3 Assay Kit (catalogue no. 294-T, Hs-695-T, LOX-IMVI, Malme-3M, MDA-MB-435, R8090; Molecular Devices). Loading dye was spiked with MeWo, RPMI-7951, SK23, SK-MEL-23, SK-MEL-28, 5 mg/mL 24C7 Ab or 100 nmol/L BQ788 sodium salt as UACC-257, and WM-266-4 were obtained from the Amer- appropriate, prior to addition to cells. The plate was loaded ican Type Culture Collection or NCI-60 [National Cancer onto the FLIPR High-throughput Cellular Screening System Institute (NCI)] and the cell lines 526 mel, 537 mel, 624 (Molecular Devices) to obtain instant readout of the cal- mel, 888 mel, 928 mel, and 1,300 mel were a generous gift cium mobilization on addition of 100 nmol/L ET-1. from Paul Robbins (Center for Cancer Research, Tumor Immunology Section, NCI) and were grown in appropriate Xenograft models media at 37 C and 5% CO2. Short hairpin RNA (shRNA)- All studies were conducted in accordance with the Guide mediated knockdown was induced in clonal populations for the Care and Use of Laboratory Animals (Ref: Institute of by culturing these cells for 72 hours in complete medium Laboratory Animal Resources (NIH publication no. 85-23), containing 2 mg/mL doxycycline (631311; Clontech Washington, DC: National Academies Press; 1996). Laboratories). shRNA-bearing clonal populations were Antibodies were conjugated with monomethyl auristatin selected by the addition of 1 to 5 mg/mL puromycin E (MMAE) as described previously (26). The drug to anti- (631306; Clontech Laboratories) as appropriate. Stable body ratio for 5E9-vc-MMAE ranged from 2.95 to 3.55, and HEK293-transfected populations were selected using 500 the drug to antibody ratio for control IgG-vc-MMAE ranged mg/mL Geneticin (10131-027; Gibco/Invitrogen). Cells from 2.5 to 3.2. Average tumor volumes with SDs were were transfected with DNA, using Lipofectamine 2000 determined from 10 animals per group. For efficacy studies (11668-019; Invitrogen) as per the recommended proto- with ADCs, 6- to 8-week-old female CRL Nu/Nu mice from col. To assess the effects of ADCs on cell viability, cells were Charles River Laboratories were inoculated subcutaneously plated at 1,500 per well in 50 mL of normal growth medium in the dorsal right flank with 10 106 UACC-257 or with in 96-well clear-bottom plates. Twenty-four hours later, an 5 106 A2058 cells in HBSS with Matrigel. When tumor additional 50 mL of culture medium with serial dilutions of volumes reached approximately 200 mm3 (day 0), animals antibodies was added to triplicate wells. Five days later, cell were randomized into groups of 10 each and administered survival was determined using CellTiter-Glo Luminescent a single intravenous (IV) injection of anti-EDNBR chimeric Cell Viability Reagent (G7572; Promega Corporation) and or mAb conjugated to MMAE through the valine-citrulline with an EnVision 2101 Mutilabel Reader (Perkin-Elmer). linker (26). MMAE conjugated to anti-gD antibody Cell lines with inducible shRNA-bearing lentivirus con- (human isotype) or to anti-ragweed antibody (mouse iso- structs were generated essentially as described previously type) was used as a control. Tumor volumes were measured (23–25). shRNA targeting EDNBR at nucleotides 1,363 to twice per week until study end. The same procedure was 1,381 of NM_000115, target sequence 50-TCACT- followed for growth studies involving A2058 and UACC- GAATTCCTGCATTA-30 (D-003657-02; Dharmacon RNAi 257 cell lines expressing inducible shRNA, except that Technologies), was used to knockdown EDNBR in mela- when tumor volumes reached approximately 200 mm3

www.aacrjournals.org Clin Cancer Res; 17(5) March 1, 2011 OF3

Asundi et al.

(day 0), animals were randomized and treated with either Affymetrix data were obtained from Gene Logic Inc. The 5% sucrose water or 1 mg/mL doxycycline in 5% sucrose analysis shown for probe set ID 204271_s_at was carried water. All water bottles were changed thrice weekly. out using the HGU133 Plus v2 GeneChip on 3,879 normal human tissue samples (green symbols), 1,605 human Protein purification cancer tissue samples (red symbols: 1,291 primary and For the purification of ligand-bound His-tagged 314 metastatic), and 3,872 human noncancer disease tissue EDNBR protein, we followed a procedure involving samples (blue symbols). Microarray data were normalized expression in insect sf9 cells essentially as described using the Affymetrix MAS (Microarray Analysis Suite) ver- previously (27), with a modification to the extraction sion 5.0 software, with sample expression values scaled to a of cell membranes. Protein was solubilized by adding 1% trimmed mean of 500. The correlation between MITF n-dodecyl-b-D-maltopyranoside (DDM) and 15% glycerol (probe 226066_at) and EDNBR (204271_s_at) mRNA and nutating the mixture overnight at 4C. Supernatant expression was assessed on the 28 human metastatic mel- was collected after high-speed centrifugation and anoma samples contained in these data. For the correlation adjusted to 300 mmol/L sodium chloride concentration of MITF and EDNBR in the indicated melanoma cell lines, and pH 8.0. The sample was loaded on to Ni-NTA Super- data were generated at Genentech by using the same flow Column (30410; Qiagen). The column was washed Affymetrix probes, chip set, and scaling factors. with PBS buffer adjusted to 300 mmol/L sodium chloride, 10 mmol/L imidazole, 1% DDM, 15% glycerol, and Results protease inhibitors (pH 8.0). Protein was eluted with PBS buffer adjusted to 300 mmol/L sodium chloride To identify antibody targets for melanoma, we containing 500 mmol/L imidazole, 0.1% DDM, and searched for mRNA transcripts that coded both for protease inhibitors (pH 8.0). cell-surface proteins and exhibited expression levels highly correlated to that of the melanoma oncogene Microarray gene expression analysis MITF (14). In microarray data generated from a panel For the analysis of EDNBR mRNA expression in multiple of 28 metastatic melanomas, and 18 melanoma cell lines, human tumor and normal biopsy samples (Fig. 1B), the these criteria were satisfied by mRNA coding for the

Signal intensity (ave. diﬀ. x 10-33) A 0 510152025 30 35 1,6000 Cell lines B 1,4000 Adipose Adrenal 1,2000 Blood vessel 10,000 Bone 8,000 Bone marrow Brain 6,000 Breast Figure 1. 4,000 Cervix Expression of EDNBR mRNA. Measurements were 2,000 Colorectal Endometrium carried out on the Affymetrix 0 Esophagus U133P chip and are expressed as 0 5,000 10,000 15,000 20,000 25,000 Gallbladder scaled average difference. A,

EDNBR Head and neck relative expression of EDNBR Heart 40,000 y x Tumors Kidney mRNA ( -axis) and MITF ( -axis) 35,000 Liver for 18 melanoma cells lines (top) 30,000 Lung and 28 human metastatic 25,000 Lymphoid melanomas. B, relative expression 20,000 Muscle of EDNBR in human tissues. Each Myometrium 15,000 dot represents a normal (green), Nerve 10,000 Neurocrine tumor (red), or diseased nontumor 5,000 Ovary (blue) human tissue specimen. 0 Pancreas Rectangles encompass the 25 to 0 10,000 20,000 30,000 Placenta 75percentile range for each Prostate distribution. WBC, white blood Skin cells. Small intestine MITF Soft tissue Stomach Testis Thymus Thyroid Urinary WBC

OF4 Clin Cancer Res; 17(5) March 1, 2011 Clinical Cancer Research

An Antibody–Drug Conjugate Targeting the Endothelin B Receptor

EDNBR (Fig. 1A). An examination of microarray data generated from a large and diverse panel of human A tissues revealed that EDNBR mRNA is expressed more C 100 highly in melanoma than in the vast majority of all other NA 100 NA normal and cancer tissues analyzed (Fig. 1B). In the 80 80 category labeled bone marrow, a subset of samples exhi- 60 60 40 40 % of max biting high levels of EDNBR transcript was further clas- % of max sified as multiple myeloma. 20 20

A positive role for EDNBR as a driver of oncogenesis in 0 0 1,000100101 10,000 1 1,00010010 10,000 melanoma is consistent with previous studies of small FL1-H FL1-H molecule inhibitors on tumorigenicity of melanoma cell 100 1.5 nmol/L lines in vitro and in vivo (17–19). mAbs offer a comple- 100 100 nmol/L 80 ET-1 mentary therapeutic approach to cell-surface targets and 80 ET-1 60 typically operate with greater specificity relative to small 60 40 % of max molecule inhibitors. Therefore, we attempted to raise 40 therapeutically active antibodies specific to EDNBR. Mice % of max 20 20 0 were immunized with intact HEK293 cells overexpressing 1 1,00010010 10,000 0 FL1-H human EDNBR and the resulting hybridoma superna- 1,000100101 10,000 100 tants were screened against receptor-positive and -nega- FL1-H 3 nmol/L ET-1 tive cell lines. Antibody 24C7 was of particular interest B 80 because its binding to EDNBR was blocked by preincuba- 60 40 tion of cells with ET-1 (Fig. 2A and B). Moreover, the % of max 24C7 antibody did not cross-react with endothelin A 20 receptor (EDNRA), and by testing various EDNRA/B 0 1 1,00010010 10,000 chimeric receptors, we mapped the binding site to the 20 μM -ET-1 FL1-H second extracellular loop of EDNBR (Supplementary 100 6 nmol/L Fig. S1). The mutually exclusive binding of 24C7 and 80 ET-1

ET-1 suggested that the antibody might inhibit receptor– 60 ligand interactions. However, ligand-induced activation 40 of EDNBR signaling, as monitored by MAPK phosphor- % of max 20 ylation and calcium influx, was unaffected by the 24C7 + ET-1 0 antibody (Supplementary Fig. S2). 1 1,00010010 10,000 It was possible that the lack of 24C7 binding observed FL1-H following ET-1 incubation was due to receptor internalization. However, binding of additional distinct antibodies to Figure 2. ET-1 competes with antibody 24C7. A and B, UACC-257 EDNBR was not affected by preincubation with ET-1 (not cells were incubated with or without 100 nmol/L ET-1 for 45 minutes shown; Supplementary Fig. S3). Endothelins have been followed by 5 mg/mL of 24C7. Antibody binding was detected by flow reported to bind their receptors with particularly high cytometry (A) or immunoflourescence (B). C, HEK293 cells ectopically expressing EDNBR were incubated for 45 minutes with the indicated affinity (28), which could account for the inability of concentrations of ET-1 followed by 5 mg/mL 24C7. Flow cytometry 24C7 to inhibit receptor activation. Indeed, further analysis data represent binding by antibody (black line) or secondary antibody of competitive binding revealed that as little as 1.5 nmol/L only (gray line) NA, not applied. ET-1 partially blocked binding of 5 mg/mL (30 nmol/L) 24C7 whereas 6 nmol/L ET-1 completely blocked it antagonist in phase II melanoma clinical studies (20, 21), (Fig. 2C). Thus, the binding of 24C7 and ET-1 to EDNBR dissuaded us from further pursuing an approach to recep- is mutually exclusive, but the antibody lacks affinity suffi- tor inhibition. However, the exceptional specificity and cient to compete with the high-affinity binding of ET-1. abundant overexpression of EDNBR in melanoma make Before initiating an effort to obtain more suitable it an attractive target for an ADC. Therefore, we generated therapeutic antibodies, we sought to more directly assess additional mAbs to EDNBR that would bind with high theroleofEDNBRinmelanomagrowthin vivo.We affinity irrespective of endothelin ligand binding. Immu- engineered 2 melanoma cell lines, UACC-257 and nization of mice with purified recombinant EDNBR pro- A2058, in which EDNBR could be knocked down con- tein resulted in numerous EDNBR-specific mAbs, all of ditionally by induction of shRNA on the addition of which reacted strongly with the N-terminal extracellular doxycycline (Supplementary Fig. S4). Although adminis- sequence. tration of doxycycline to animals bearing tumor xeno- From this panel, an mAb designated 5E9 was chosen grafts derived from these cell lines enhanced progression- for further study on the basis of its cross-reactivity with free survival and inhibited tumor growth, the effects were rodent, non-human primate, and human EDNBR. The rather modest (Supplementary Fig. S4). These data, along epitope for 5E9 resides between amino acid positions 65 with the disappointing results of an endothelin receptor and 101 within the N-terminal extracellular sequence

www.aacrjournals.org Clin Cancer Res; 17(5) March 1, 2011 OF5

Asundi et al.

EDNBR LAMP1/EEA1 Merge

Figure 3. Cell internalization of EDNBR antibody 5E9. Live UACC- 257 cells were incubated with 2 mg/mL of 5E9 antibody for 2 hours at 37C and then fixed and costained with antibody to either LAMP1 (top, middle) or EEA1 20 μM (bottom middle). 5E9 was detected with Cy3-labeled secondary antibody (red) and antibodies to LAMP1 and EEA1 with Alexa 488–labeled secondary antibody (green). Merged images are presented in panels at right.

(Supplementary Fig. S1), thus precluding any loss of reac- 0.14, and 0.5 nmol/L respectively, for the 3 cell lines tivity that might result from proteolysis reported to occur at (Fig. 4B). Thus, response to free drug is probably not a amino acid R64 (29). Moreover, ET-1 does not compete significant factor contributing to the potency of the ADC on with binding of 5E9 to EDNBR (Supplementary Fig. S3). these cell lines. To further examine the relationship An important property of an ADC is the capacity to deliver between target density and cell killing, we established a the appended drug payload into cells following cell-surface stable derivative of the UACC-257 cell line in which binding. Accordingly, we carried out immunofluorescent EDNBR levels could be knocked down by conditional staining to examine the uptake of 5E9 into cells expressing expression of shRNA. The total levels EDNBR, and the EDNBR. The antibody was readily detected in vesicular amount of 5E9 antibody that undergoes internalization, structures in the fixed permeabilized cells and the staining were significantly reduced following induction of the colocalized with both the lysosomal marker LAMP1 and EDNBR-specific shRNA, whereas induction of the control the early endosomal marker EEA1 (Fig. 3). luciferase shRNA had no effect (Fig. 5A). Consistent with The rapid internalization of 5E9 into melanoma cells the reduction in target levels, cell killing by the ADC was prompted us to pursue an armed antibody approach. The impaired by induction of the EDNBR-specific shRNA but potent cytotoxic compound MMAE was conjugated to 5E9 not by induction of the luciferase control shRNA (Fig. 5A). by using a linker containing the protease sensitive valine- We also generated a derivative of the A2058 cell line that citrulline dipeptide (26). The ADC was first evaluated in overexpresses EDNBR. The increase in target expression vitro on cell lines expressing increasing levels of EDNBR. To enhanced cell killing by the ADC by more than 100-fold identify and rank EDNBR-positive cell lines, FACS was relative to the parental cell line (Fig. 5B). carried out on a panel of 18 melanoma cell lines (Supple- Two of the 3 cell lines, A2058 and UACC-257, were mentary Fig. S5). The majority of these cell lines were found to be competent to generate relatively uniform positive and, on the basis of this survey, we chose 3 cell tumor xenografts in the flanks of athymic mice and were lines, A2058, WM-266-4, and UACC-257, to represent low, thus chosen for in vivo efficacy studies. To gauge the level of medium, and high EDNBR expression, respectively. By EDNBR expression on these cell lines, relative to that Scatchard analysis, the number of 5E9 binding sites on present on human melanomas, a comparative immuno- A2058, WM-266-4, and UACC-257 was estimated at 1,582, histochemical analysis was carried out. Fixed, paraffin- 9,410, and 33,939 per cell, respectively, and the dissocia- embedded pellets of the A2058 and UACC-257 cell lines tion constants (Kd) were estimated at 1.4 to 2.4 nmol/L were sectioned and stained with antibody to EDNBR, along (Supplementary Fig. S6). Titration of these cell lines with with a tissue microarray containing sections from 40 the 5E9 ADC showed specific cell killing relative to control human melanomas. Melanoma sections were assigned ADC that was generally proportional to the level of EDNBR scores ranging from 0 to 3þ, where 36 of 40 sections were expression (Fig. 4A). positive with staining intensities of 1þ or greater (Fig. 6). Variables in addition to receptor density, such as The staining of the 2 cell lines is shown above the stained response to the released drug, might also contribute to melanoma sections representative of each of the 4 staining the relative efficacy of the ADC. To address this, we carried intensity levels. Both A2058 and UACC-257 expressed out titration of free MMAE with the A2058, WM-266-4, and levels of EDNBR comparable with 1þ and 3þ melanomas, UACC-257 cell lines and estimated EC50 values of 0.28, respectively. No anti-EDNBR staining was detected with the

OF6 Clin Cancer Res; 17(5) March 1, 2011 Clinical Cancer Research

An Antibody–Drug Conjugate Targeting the Endothelin B Receptor

A Receptors/cell 1,582 18 PBS 100 A2058

) 15 -6 80 12 IgG-vc -MMAE 60 9 40 % of max 6 5E9-vc-MMAE 20 3

0 units (x 10 Relative light 1 10 100 1,000 10,000 0 0.001 0.01 0.1 1 10 FL1-H ADC (µg/mL) Receptors/cell 9,410 Figure 4. PBS Relationship of receptor 100 WM-266-4 9 level to ADC cell killing in vitro.A, ) the relative amounts of cell- 80 -6 surface EDNBR on melanoma cell IgG-vc lines A2058 (top), WM-266-4 6 (middle), and UACC-257 (bottom) 60 -MMAE is shown by flow cytometry on live cells (green line). Secondary 40 % of max 5E9-vc-MMAE antibody alone was used as a 3 control (red line). The indicated 20 number of receptors/cell was estimated separately by 0 Relative light units (x 10 units Relative light 0 Scatchard analysis. Cell killing by 1 10 100 1,000 10,000 5E9 ADC titration is presented 0.001 0.01 0.1 1 10 FL1-H adjacent to each flow cytometry ADC (µg/mL) plot for the corresponding cell line. The indicated concentrations of Receptors/cell 33,939 12 5E9-vc-MMAE (red line), control 100 UACC 257 PBS

IgG-vc-MMAE (blue line), or ) equivalent amount of PBS vehicle 80 -6 9 (black line) were incubated with cells for 5 days and relative cell IgG-vc 60 viability (y-axis) assessed using 6 -MMAE CellTiter-Glo. B, the cell lines 40

A2058 (blue), UACC-257 (red), and % of max WM-266-4 (black) were treated 3 with increasing concentrations of 20 5E9-vc-MMAE free MMAE for 5 days and cell

viability was determined by the units (x 10 Relative light 0 0 CellTiter-Glo assay. 1 10 100 1,000 10,000 0.001 0.01 0.1 1 10 FL1-H ADC (µg/mL) B 100

50 UACC-257 WM-266-4 % survival A2058 25

0 0.001 0.01 0.1 1 10 Free MMAE (nmol/L)

www.aacrjournals.org Clin Cancer Res; 17(5) March 1, 2011 OF7

Asundi et al.

A shRNA: EDNBR Luciferase A2058 UACC-257 shRNAi: EDNBR Dox: _ + -DOX EDNBR GAPDH-

+ DOX 01+ 100

40 Luc - Dox 2+ 3+ % survival Luc +Dox EDNBR -Dox 20 EDNBR +Dox

0 00.001 0.01001 0.101 1 10 Ab concentration (µg/mL) B P OE EDNBR Score 01+2+3+ GAPDH- n 4 10 12 14 100

80 PBS (OE) Figure 6. Comparative EDNBR expression on human melanoma and IgG-vc-MMAE (OE) melanoma cell lines. Sections prepared from fixed, paraffin-embedded 5E9-vc-MMAE (P) pellets of UACC-257 and A2058 cells and from 40 fixed, paraffin- 60 5E9-vc-MMAE (OE) embedded human melanoma specimens were stained with antibody to EDNBR. A relative staining intensity (0–3þ) was assigned and a melanoma specimen representing each intensity level is presented. The number of

% survival 40 specimens (n) assigned to each intensity level is shown below.

20 renal cancer cell line 786-O, which lacks EDNBR expression, nor was detectable staining observed on EDNBR- 0 0.001 0.01 0.1 1 10 positive cell lines or tissues with isotype control antibody Ab concentration (µg/mL) (data not shown). For in vivo efficacy studies, tumor xenografts were grown to an average size of approximately 200 mm3, whereupon Figure 5. Altering EDNBR expression level affects ADC cell killing animals were randomized into groups of 10 each, with each in vitro . A, Western blot (top left) for EDNBR on total cell lysates from a group receiving a single IV injection of the 5E9 ADC or UACC-257 cell line expressing shRNA specific to EDNBR in response to doxycycline (Dox). These UACC-257 cells and a control cell line control ADC. Doses of 1, 3, and 6 mg/kg of 5E9 ADC or 6 expressing shRNA specific to luciferase were treated with or without mg/kg of control ADC were administered to animals bear- doxycycline and then fixed and permeabilized prior to immunofluorescent ing A2058 tumors. A sustained reduction in tumor volume detection of EDNBR (top right). A second set of live cells was was observed at the high dose of the 5E9 ADC relative to subjected to a cell killing assay with 5E9-vc-MMAE (bottom) as described in Figure 4. B, Western blot (top) of total cell lysate from the matching dose of control ADC or vehicle (Fig. 7A). the A2058 parental cell line (P) or a stable derivative overexpressing Efficacy was not apparent for the group dosed at 1 mg/kg, EDNBR (OE). The two cells lines were subjected to a cell killing and a trend toward growth retardation was noted for the 3 assay (bottom) with 5E9-vc-MMAE or control IgG-vc-MMAE as mg/kg dose group. Consistent with in vitro cell killing, the described in Figure 4. UACC-257 xenograft tumors were more responsive to the

OF8 Clin Cancer Res; 17(5) March 1, 2011 Clinical Cancer Research

An Antibody–Drug Conjugate Targeting the Endothelin B Receptor

mode of signaling in cells are consistent with a potential A2058 functional contribution to tumor progression (7). Germ- A line-inactivating mutations in either EDNBR or ET-3 are 1,800 5E9-vcE (1 mg/kg) ) responsible for the WS4 variant of Waardenburg syn- 3 Vehicle IgG-vcE (6 mg/kg) drome, which is characterized by melanocytic deficits 1,400 in skin and other tissues (12). The MITF is mutated in 5E9-vcE (3 mg/kg) another variant of Waardenburg syndrome, WS2, which 1,000 again is associated with melanocytic deficits (31). Thus, 600 5E9-vcE (6 mg/kg) loss of function of either EDNBR or MITF negatively affects the growth and survival of melanocyte progenitors. 200 Conversely, excess activity of these genes could reanimate 0 growth and migratory characteristics of these cells and 484032241680 Mean tumor volume (mm volume Mean tumor (d) thereby contribute to tumor progression. Indeed, muta- IV treatments tions in MITF, and amplification of the gene coding for it, have been identified in human melanoma (13, 14). Although the gene encoding EDNBR is neither amplified B UACC-257 nor sporadically mutated in melanoma, the transcript is 1,200 overexpressed and we find this correlates with MITF Vehicle transcript across tumor tissue samples and cancer cell ) 3 1,000 lines. Therefore, we sought to target EDNBR with ther- IgG-vcE (4 mg/kg) apeutic antibodies. 800 Our first attempt at generating antibodies to EDNBR employed HEK293 cells overexpressing the receptor as the 600 immunogen administered to BALB/c mice. This resulted in mAbs that required the second extracellular loop of the 400 receptor for binding that was blocked on preincubation 5E9-vcE (1 mg/kg) 200 with endothelin ligands. Nevertheless, these antibodies were incapable of inhibiting EDNBR activation by Mean tumor volume (mm 5E9-vcE (4 mg/kg) 0 endothelins. It is possible that the high affinity of ET-1 K 0 80604020 100 for EDNBR, with a reported d of 80 pmol/L (28), could IV treatments (d) preclude competition by our antibodies when used at therapeutically relevant concentrations. Therefore, the appropriate antibodies with very high affinity could con- Figure 7. In vivo efficacy of anti-EDNBR ADC. Subcutaneous tumors were established in mice inoculated with A2058 (A) or UACC-257 (B) cells. When ceivably interfere with ligand binding to EDNBR. However, tumor volumes reached approximately 200 mm3 (day 0), animals were before engaging an effort to generate such antibodies, we given a single IV injection of either control ADC (IgG-vcE) or anti-EDNBR conducted experiments to further assess the contribution of ADC (5E9-vcE) at the indicated doses. Average tumor volumes with EDNBR to tumorigenicity. standard deviations were determined from 10 animals per group. We carried out conditional knockdown of the EDNBR in 2 independent melanoma cell lines and observed positive effects on survival of tumor-bearing animals and retarda- ADC and regressed or remained static for the entire study tion of tumor growth. This was consistent with a previous period of more than 100 days in response to single dose of study in which inhibition of EDNBR impacted the growth 1 mg/kg 5E9 ADC (Fig. 7B). This result was indistinguish- of melanoma cell lines in vitro (19). Nevertheless, the able from that of the 4 mg/kg 5E9 ADC dose group, marginal impact on tumorigenicity in our studies suggested suggesting that a maximal effect had been achieved at 1 that more robust interference with EDNBR might be mg/kg. In animals treated with 4 mg/kg of control ADC, the required to significantly effect tumor growth. Indeed, tumors grew at the same rate as those treated with vehicle marked inhibition of melanoma xenografts by BQ788, a control. These results suggest that efficacy can be achieved potent EDNBR inhibitory compound, has been reported with the 5E9 ADC in tumors that correspond to the full (18). Despite these encouraging preclinical results, phase II expression range of EDNBR encountered in human mela- clinical trials with the endothelin receptor inhibitor bosen- nomas. tan have not produced any objective responses (20, 21). We therefore turned to an approach that exploits the over- Discussion expression of EDNBR in melanoma but does not rely on functional interference. The EDNBR is overexpressed in human melanoma and Advances in ADC technology have resulted in encoura- attempts to exploit it as a therapeutic target have been ging responses in recent clinical trials (32). The develop- reported (6, 30). It is an attractive target because the ment of more potent drugs and linkers with enhanced function of EDNBR in embryonic development and its stability has greatly improved the prospects for this

www.aacrjournals.org Clin Cancer Res; 17(5) March 1, 2011 OF9

Asundi et al.

approach. This has coincided with advances in high- individual cancers. The drug substance that is ultimately throughput technologies, enabling the identification of released from the ADC will also be subject to metabolism highly specific cell-surface antigens expressed on tumor and efflux that will vary with the genetic background of the cells. Our gene expression analysis, in which thousands individual cancer and, possibly, that of the patient. More- of human tumor and normal tissue samples are repre- over, acquired resistance to the drug might arise to varying sented, revealed overexpression of EDNBR mRNA in mel- degrees by selection during the course of treatment. Selec- anoma that greatly exceeded expression in any other tion against the antibody target represents an additional normal tissue. mode of acquired resistance. However, we have monitored Our selection of the 5E9 mAb for drug conjugation was EDNBR expression in our preclinical efficacy studies and based on high-affinity binding, cross-species reactivity, lack have not observed downregulation of the receptor follow- of interference by endothelins, and a high rate of cellular ing prolonged ADC exposure. internalization. When conjugated with MMAE through the Melanoma remains one of the most aggressive human valine-citrulline peptide linker, the 5E9 ADC exhibited cancers for which effective treatment options are extremely remarkable potency. Notably, the growth of the A2058 limited. Overall, our efficacy studies indicate that EDNBR is melanoma tumor xenograft was strongly retarded by a an excellent target for the application of ADCs in mela- single dose of 6 mg/kg of 5E9-vc-MMAE despite the esti- noma. mated expression of only 1,500 copies of EDNBR per cell. The UACC-257, which expresses 30,000 copies per cell, exhibited no signs of growth to 100 days following a single Disclosure of Potential Conflicts of Interest administration of 1 mg/kg 5E9-vc-MMAE. On the basis of No potential conflicts of interest were disclosed. EDNBR expression, the A2058 and UACC-257 represent the low and high expressers, respectively, relative to our Acknowledgments panel of 40 human melanomas. Although the density of EDNBR expression on mela- We thank Genentech employees Christina Ha, Andrew McGeehan, Anan noma cells is a primary factor that will influence efficacy, it Chuntharapai, Robert Kelley, and James Ernst for expert technical advice is likely that additional variables will also have an impact. and assistance in generating reagents used in this study. For example, the efficiency by which cancer cells generate Received August 31, 2010; revised December 8, 2010; accepted January the active drug from the internalized ADC may vary across 2, 2011; published OnlineFirst January 18, 2011.

References 1. Thompson JF, Scolyer RA, Kefford RF. Cutaneous melanoma in the 12. Pingault V, Ente D, Dastot-Le Moal F, Goossens M, Marlin S, Bondur- era of molecular profiling. Lancet 2009;374:362–5. and N, et al. Review and update of mutations causing Waardenburg 2. Ascierto PA, Streicher HZ, Sznol M. Melanoma: a model for syndrome. Hum Mutat 2010;31:391–406. testing new agents in combination therapies. J Transl Med 2010;8: 13. Garraway LA, Widlund HR, Rubin MA, Getz G, Berger AJ, Ramas- 38–45. wamy S, et al. Integrative genomic analyses identify MITF as a lineage 3. Yanagisawa M, Kurihara H, Kimura S, Tomobe Y, Kobayashi M, Mitsui survival oncogene amplified in malignant melanoma. Nature 2005; Y, et al. A novel potent vasoconstrictor peptide produced by vascular 436:117–22. endothelial cells. Nature 1988;332:411–5. 14. Levy C, Khaled M, Fisher DE. MITF: master regulator of melanocyte 4. Arai H, Hori S, Aramori I, Ohkubo H, Nakanishi S. Cloning and development and melanoma oncogene. Trends Mol Med 2006;12: expression of a cDNA encoding an endothelin receptor. Nature 406–14. 1990;348:730–2. 15. Cronin JC, Wunderlich J, Loftus SK, Prickett TD, Wei X, Ridd K, et al. 5. Sakurai T, Yanagisawa M, Takuwa Y, Miyazaki H, Kimura S, Goto K, Frequent mutations in the MITF pathway in melanoma. Pigment Cell et al. Cloning of a cDNA encoding a non-isopeptide-selective subtype Melanoma Res 2009;22:435–44. of the endothelin receptor. Nature 1990;348:732–5. 16. Demunter A, De Wolf-Peeters C, Degreef H, Stas M, Van Den Oord JJ. 6. Lahav R. Endothelin receptor B is required for the expansion of Expression of the endothelin-B receptor in pigment cell lesions of the melanocyte precursors and malignant melanoma. Int J Dev Biol skin. Evidence for its role as tumor progression marker in malignant 2005;49:173–80. melanoma. Virchows Arch 2001;438:485–91. 7. Saldana-Caboverde A, Kos L. Roles of endothelin signaling in mel- 17. Lahav R, Heffner G, Patterson PH. An endothelin receptor B antago- anocyte development and melanoma. Pigment Cell Melanoma Res nist inhibits growth and induces cell death in human melanoma cells in 2010;23:160–70. vitro and in vivo. Proc Natl Acad Sci U S A 1999;96:11496–500. 8. Reid K, Turnley AM, Maxwell GD, Kurihara Y, Kurihara H, Bartlett PF, 18. Bagnato A, Rosano L, Spinella F, Di Castro V, Tecce R, Natali PG, et al. et al. Multiple roles for endothelin in melanocyte development: reg- Endothelin B receptor blockade inhibits dynamics of cell interactions ulation of progenitor number and stimulation of differentiation. Devel- and communications in melanoma cell progression. Cancer Res opment 1996;122:3911–9. 2004;64:1436–43. 9. Lee HO, Levorse JM, Shin MK. The endothelin receptor-B is required 19. Lahav R, Suva ML, Rimoldi D, Patterson PH, Stamenkovic I. Endothe- for the migration of neural crest-derived melanocyte and enteric lin receptor B inhibition triggers apoptosis and enhances angiogen- neuron precursors. Dev Biol 2003;259:162–75. esis in melanomas. Cancer Res 2004;64:8945–53. 10. Puffenberger EG, Hosoda K, Washington SS, Nakao K, deWit D, 20. Kefford R, Beith JM, Van Hazel GA, Millward M, Trotter JM, Wyld DK, Yanagisawa M, et al. A missense mutation of the endothelin-B et al. A phase II study of bosentan, a dual endothelin receptor receptor gene in multigenic Hirschsprung's disease. Cell 1994;79: antagonist, as monotherapy in patients with stage IV metastatic 1257–66. melanoma. Invest New Drugs 2007;25:247–52. 11. Robertson K, Mason I, Hall S. Hirschsprung's disease: genetic muta- 21. Kefford RF, Clingan PR, Brady B, Ballmer A, Morganti A, Hersey P, tions in mice and men. Gut 1997;41:436–41. et al. A randomized, double-blind, placebo-controlled study of

OF10 Clin Cancer Res; 17(5) March 1, 2011 Clinical Cancer Research

An Antibody–Drug Conjugate Targeting the Endothelin B Receptor

high-dose bosentan in patients with stage IV metastatic melanoma 27. Yamaguchi T, Arimoto-Tahara I, Fujiyoshi Y, Doi T. Characterization receiving first-line dacarbazine chemotherapy. Mol Cancer 2010;9:69. and application of monoclonal antibodies against human endothelin 22. Chen Y, Clark S, Wong T, Chen Y, Chen Y, Dennis MS, et al. Armed B receptor expressed in insect cells. Biotechnol Lett 2004;26: antibodies targeting the mucin repeats of the ovarian cancer antigen, 293–9. MUC16, are highly efficacious in animal tumor models. Cancer Res 28. Chiou WJ, Magnuson SR, Dixon D, Sundy S, Opgenorth TJ, Wu-Wong 2007;67:4924–32. JR, et al. Dissociation characteristics of endothelin receptor agonists 23. Gray DC, Hoeflich KP, Peng L, Gu Z, Gogineni A, Murray LJ, et al. and antagonists in cloned human type-B endothelin receptor. pHUSH: a single vector system for conditional gene expression. BMC Endothelium 1997;5:179–89. Biotechnol 2007;7:61. 29. Grantcharova E, Furkert J, Reusch HP, Krell HW, Papsdorf G, Beyer- 24. Hoeflich KP, Jaiswal B, Davis DP, Seshagiri S. Inducible BRAF mann M, et al. The extracellular N terminus of the endothelin B (ETB) suppression models for melanoma tumorigenesis. Methods Enzymol receptor is cleaved by a metalloprotease in an agonist-dependent 2008;439:25–38. process. J Biol Chem 2002;277:43933–41. 25. Jaiswal BS, Janakiraman V, Kljavin NM, Chaudhuri S, Stern HM, Wang 30. Bagnato A, Rosano L. The endothelin axis in cancer. Int J Biochem W, et al. Somatic mutations in p85alpha promote tumorigenesis Cell Biol 2008;40:1443–51. through class IA PI3K activation. Cancer Cell 2009;16:463–74. 31. Tassabehji M, Newton VE, Read AP. Waardenburg syndrome type 2 26. Doronina SO, Toki BE, Torgov MY, Mendelsohn BA, Cerveny CG, caused by mutations in the human microphthalmia (MITF) gene. Nat Chace DF, et al. Development of potent monoclonal antibody Genet 1994;8:251–5. auristatin conjugates for cancer therapy. Nat Biotechnol 2003; 32. Senter PD. Potent antibody drug conjugates for cancer therapy. Curr 21:778–84. Opin Chem Biol 2009;13:235–44.

www.aacrjournals.org Clin Cancer Res; 17(5) March 1, 2011 OF11

An Antibody−Drug Conjugate Targeting the Endothelin B Receptor for the Treatment of Melanoma

Jyoti Asundi, Chae Reed, Jennifer Arca, et al.

Clin Cancer Res Published OnlineFirst January 18, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-10-2340

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2011/03/02/1078-0432.CCR-10-2340.DC1

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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2011/02/18/1078-0432.CCR-10-2340. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.