SAR650984, a Novel Humanized CD38-Targeting Antibody, Demonstrates Potent Antitumor Activity in Models of Multiple Myeloma and Other Cd38þ Hematologic Malignancies

Published OnlineFirst July 1, 2014; DOI: 10.1158/1078-0432.CCR-14-0695

Clinical Cancer Cancer Therapy: Preclinical Research

SAR650984, A Novel Humanized CD38-Targeting Antibody, Demonstrates Potent Antitumor Activity in Models of Multiple Myeloma and Other CD38þ Hematologic Malignancies

Jutta Deckert1, Marie-Cecile Wetzel2, Laura M. Bartle1, Anna Skaletskaya1, Victor S. Goldmacher1, Francois¸ Vallee 3, Qing Zhou-Liu4, Paul Ferrari4,Stephanie Pouzieux3, Charlotte Lahoute2, Charles Dumontet5, Adriana Plesa5, Marielle Chiron2, Pascale Lejeune2, Thomas Chittenden1, Peter U. Park1, and Veronique Blanc2

Abstract Purpose: The CD38 cell surface antigen is expressed in diverse hematologic malignancies including multiple myeloma, B-cell non-Hodgkin lymphoma (NHL), B-cell chronic lymphocytic leukemia, B-cell acute lymphoblastic leukemia (ALL), and T-cell ALL. Here, we assessed the antitumor activity of the anti- CD38 antibody SAR650984. Experimental Design: Activity of SAR650984 was examined on lymphoma, leukemia and multiple myeloma cell lines, primary multiple myeloma samples, and multiple myeloma xenograft models in immunodeﬁcient mice. Results: We identiﬁed a humanized anti-CD38 antibody with strong proapoptotic activity independent of cross-linking agents, and potent effector functions including complement-dependent cytotoxicity, antibody-dependent cell-mediated cytotoxicity, and antibody-dependent cellular phagocytosis (ADCP), þ þ equivalent in vitro to rituximab in CD20 and CD38 models. This unique antibody, termed SAR650984, inhibited the ADP-ribosyl cyclase activity of CD38, likely through an allosteric antagonism as suggested by 3D structure analysis of the complex. In vivo, SAR650984 was active in diverse NHL, ALL, and multiple þ myeloma CD38 tumor xenograft models. SAR650984 demonstrated single-agent activity comparable with rituximab or cyclophosphamide in Daudi or SU-DHL-8 lymphoma xenograft models with induction of the proapoptotic marker cleaved capase-7. In addition, SAR650984 had more potent antitumor activity than bortezomib in NCI-H929 and Molp-8 multiple myeloma xenograft studies. Consistent with its mode of þ action, SAR650984 demonstrated potent proapoptotic activity against CD38 human primary multiple myeloma cells. Conclusion: These results validate CD38 as a therapeutic target and support the current evaluation of this þ unique CD38-targeting functional antibody in phase I clinical trials in patients with CD38 B-cell malignancies. Clin Cancer Res; 20(17); 4574–83. 2014 AACR.

Introduction aggressive non-Hodgkin lymphoma (NHL) through its use Rituximab, a chimeric anti-CD20 antibody, dramatically as a single agent or in combination with chemotherapy (1, improved the treatment outcome of both indolent and 2). Although rituximab ultimately depletes B cells through binding to CD20, its mechanism of action is attributed to a combination of direct apoptosis induction and activation of 1Research and Development, ImmunoGen, Inc., Waltham, Massachusetts. immune effector mechanisms, including antibody-depen- 2Oncology Business Division, Sanofi R&D, Vitry-sur-Seine, France. 3SCP dent cell-mediated cytotoxicity (ADCC) and phagocytosis fi LGCR/Structure Design & Informatics/Structural Biology, Sano R&D, (ADCP) and complement-dependent cytotoxicity (CDC; Vitry-sur-Seine, France. 4Biorealization/Operations and Protein Tools, Sanofi R&D, Vitry-sur-Seine, France. 5Hematology Laboratory, Hospices ref. 3). Despite the success of rituximab, incomplete treat- Civils de Lyon, Pierre Benite, France and CRCL, INSERM 1052-CNRS ment responses and emergence of resistance represent 5286, Lyon, France. important limitations (1). In an attempt to overcome these Current address for P. Lejeune: Oncology Department, Global Drug Dis- limitations, a plethora of second-generation anti-CD20 covery, Bayer Pharma AG, Berlin, Germany; current address for P.U. Park, Biology, Mersana Therapeutics, Inc., Cambridge, MA; and current address antibodies has been introduced into clinical development, for C. Lahoute, Pharmacology Department, Invectys, Paris, France. most notably ofatumumab and obinutuzumab which have Corresponding Author: Jutta Deckert, ImmunoGen, Inc., 830 Winter gained regulatory approval in B-cell chronic lymphocytic Street, Waltham, MA 02451. Phone: 781-895-0600; Fax: 781-895-0611; leukemia (CLL; ref. 4). CD20 expression is confined to a B- E-mail: [email protected] cell subset, restricting the utility of these agents to patients doi: 10.1158/1078-0432.CCR-14-0695 with CLL and B-cell NHL. Therefore, identification of new 2014 American Association for Cancer Research. targets with a broader expression spectrum and potential for

4574 Clin Cancer Res; 20(17) September 1, 2014

SAR650984: A Novel Antibody for CD38þ Malignancies

malignant cells and cell lines have been described in the Translational Relevance literature (10, 22–27). A recently developed antibody, dar- The CD38 antigen is a promising target for an anti- atumumab, is currently undergoing clinical trials in patients body-based therapeutic as it is expressed in a number of with multiple myeloma and reported to have a manageable hematologic malignancies including multiple myeloma, safety profile and early signs of efficacy (28–30). Another B-cell non-Hodgkin lymphoma, B-cell chronic lympho- anti-CD38 antibody, MOR202, is being evaluated in a cytic leukemia, B-cell acute lymphoblastic leukemia phase I/IIa trial in patients with relapsed/refractory multiple (ALL), and T-cell ALL. On the basis of the success of þ myeloma (31, 32). rituximab in treating CD20 lymphoma, targeting CD38 Here, we describe the generation of a new humanized with a functional antibody is a promising therapeutic anti-CD38 antibody, SAR650984, selected for its multiple approach for these diverse indications. SAR650984, a mechanisms of action. SAR650984 demonstrated potent novel anti-CD38 antibody described here, induces proapoptotic activity in the absence of cross-linking agents, a unique combination of effects on cancer cells including elicited immune effector mechanisms such as CDC, ADCC, þ complement-dependent cytotoxicity, antibody-depen- and ADCP against CD38 malignant cells, and in addition dent cell-mediated cytotoxicity, antibody-dependent cel- strongly inhibited CD38 enzymatic activity. This unique lular phagocytosis, and apoptosis, and demonstrates combination of activities translated into strong in vivo þ potent in vitro and in vivo activity against cells derived efficacy of SAR650984 against diverse CD38 tumor xeno- from these malignancies. Consequently, clinical þ graft models. trials with this antibody in patients with CD38 hematologic malignancies are on-going (NCT01084252, Materials and Methods NCT01749969), with initial signs of promising efficacy. Cell culture Ramos, Raji, Daudi, and Namalwa lymphoma cells were obtained from the American Type Culture Collection (ATCC). All other cell lines were obtained from Deutsche antibody mechanisms similar to rituximab could yield Sammlung von Mikroorganismen und Zellkulturen novel antibody therapeutics for a wider range of hemato- (DSMZ). Cells were cultured in RPMI-1640 medium sup- logic malignancies. plemented with 10% FBS, 2 mmol/L glutamine, and 1% The cell surface molecule CD38 represents such an oppor- penicillin-streptomycin (Life Technologies) at 37Cina tunity because it is expressed in multiple hematologic humidified 5% CO2 incubator. malignancies derived from both lymphoid and myeloid lineages, including multiple myeloma, a CD20 negative Antibody generation B-cell malignancy (5). CD38 is a 45 kDa type II transmem- CD38-binding murine monoclonal antibodies were gen- brane glycoprotein that has been described as both a recep- erated by standard hybridoma technology following immu- tor and a multifunctional enzyme involved in the produc- nizations with murine 300-19 cells transfected to express the tion of nucleotide-metabolites (6–8). It has been proposed full-length human CD38 antigen. Antibodies were chimer- to associate with cell surface proteins in lipid rafts, regulate ized (ch) as huIgG1-k and humanized (hu) by variable þ cytoplasmic Ca2 flux, and mediate signal transduction domain resurfacing (33). Chimeric antibodies, chHB7 and in lymphoid and myeloid cells (9–12). During hemato- chOKT10, were produced on the basis of sequences derived þ poiesis, CD38 is expressed on CD34 stem cells and line- from their respective murine hybridomas (ATCC HB-136, age-committed progenitors of lymphoid, erythroid, and CRL-8022). Ab005 (a surrogate antibody for daratumu- myeloid cells, while pluripotent early stem cells are defined mab) and huAT13/5 were produced on the basis of pub- þ as CD34 CD38 . CD38 expression is absent during the lished sequences (Genbank: ADS96869 and ADS96865, intermediate stages of lymphocyte maturation, then reap- corresponding to sequences 17 and 12 from US7829673 pears during the final stages of maturation, and continues for Ab005, CAA03366 and CAA03368 for huAT13/5). IB4 through the plasma cell stage (13–16). and SUN-4B7 were a gift from F. Malavasi, University of CD38 expression is an important prognostic factor in B- Turin, Italy. Rituximab, cyclophosphamide (CPA), and CLL that identifies patients with aggressive disease and poor bortezomib were obtained from commercial sources. survival (17–19). It has been proposed that CD38 plays a role in the proliferation and survival of CLL cells (20). High Apoptosis induction CD38 expression has been reported on plasma cells as well Cells were incubated at 2 105 cells/mL in complete as their malignant counterparts, multiple myeloma cells, medium (RPMI-1640, 10% FBS, 2 mmol/L L-glutamine) leading to its widespread use as a surface marker, combined with 1.5 mg/mL (10 nmol/L) of indicated antibodies for with CD138, to enumerate malignant myeloma cells in 20 hours at 37 C with 5% CO2. Cells were stained with patient bone marrow (5, 21). Overall, CD38 expression in Annexin V-FITC in accordance with the manufacturer’s hematologic malignancies and its correlation with disease instructions (Life Technologies). Samples were analyzed by progression make CD38 an attractive target for antibody- flow cytometry on a FACSCalibur flow cytometer with based therapeutics. Several anti-human CD38 antibodies CellQuest Pro (v5.2) for acquisition control and data anal- differing in their functional properties against various ysis (both BD Biosciences).

www.aacrjournals.org Clin Cancer Res; 20(17) September 1, 2014 4575

Deckert et al.

Binding assay Mononuclear cells were prepared by standard Ficoll– Antibodies were incubated with 5 104 cells per assay in Paque centrifugation. Cells were incubated at 106 cells RPMI-1640 with 2% normal goat serum and 0.1% sodium per sample with 100 mg/mL SAR650984 for 18 hours at azide (both Sigma-Aldrich) for 3 hours on ice followed by 37C in culture media containing 50% autologous serum incubation with FITC-conjugated goat-anti-mouse or anti- (all patient samples contained > 18 mg/mL of IgG; see human IgG antibody (Jackson ImmunoResearch) for 1 Supplementary Table S1). Samples were stained with anti- hour on ice. Samples were analyzed on a FACSCalibur flow CD38-APC (HB7, BD Biosciences) and anti-CD138-PC5 cytometer as above. Sigmoidal dose–response curves were (BA38, Beckman Coulter) antibodies to identify multiple generated and EC50 values calculated using GraphPad Prism myeloma cells and the degree of apoptosis were measured (GraphPad Software). by Annexin-V-FITC staining. Samples were analyzed by flow cytometry on a FACSCalibur or a FACS Canto II Expression analysis (both BD Biosciences). Cells were incubated in PBS with 0.1% sodium azide and 0.5% BSA (Sigma-Aldrich) at 2 105 cells per assay with 5 Enzymatic activity assay mg/mL anti-CD38-PE or 16 mg/mL anti-CD20-PE (BD Bios- A recombinant soluble human CD38-Fc fusion protein ciences) for 2 hours on ice. Samples were washed, fixed, and was produced, and enzymatic activity assays carried out as acquired in conjunction with QuantiBRITE beads (BD described in the Supplementary Data section (36). Biosciences) by flow cytometry according to the manufacturer’s instructions. Values for antibodies bound per cell 3D structure analysis (ABC) were extrapolated from the standard curve. Crystals of the non-glycosylated soluble human CD38 antigen (extracellular domain R45-I300) and the SAR650984 ADCC, ADCP, and CDC assays Fab fragment were obtained and analyzed as described in ADCC activity with isolated natural killer (NK) cells the Supplementary Data section. High-resolution data for 4 and 5 10 target cells at an effector:target (E:T) ratio crystals up to 1.53 A were collected at beamline ID-29 at the of 3:1 was measured by lactate dehydrogenase release European Synchrotron Radiation Facility (ESRF). (Cytotoxicity Detection Kit, Roche) after 4-hour incubation using standard protocols (34). The degree of specific In vivo xenograft studies lysis was calculated as follows: %lysis ¼ (experimental Female CB.17 SCID mice (Charles River Laboratories) release spontaneous release)/(maximum release - spon- were inoculated subcutaneously with 1 107 cells. Treat- taneous release) 100. ment was initiated as indicated and tumor volumes were Monocyte-derived macrophages were prepared by isolat- measured twice weekly for the duration of the study. %T/C þ ing monocytes from peripheral blood using CD14 beads was calculated as the median tumor volume of each treated and culturing them in the presence of 50 ng/mL hGM-CSF (T) group divided by the median tumor volume of the (both Miltenyi Biotec) for 7 days. ADCP was evaluated using vehicle control (C) group. Activity was assessed according 2 105 Ramos target cells labeled with PKH26 (Sigma- to National Cancer Institute (NCI, Bethesda, MD) standards Aldrich), then incubated with antibodies and isolated (%T/C 42% ¼ active, %T/C value 12% ¼ highly active). human monocyte-derived macrophages in round-bottom For survival studies, SCID mice were inoculated intrave- 96-well plates at 37C for 3 hours at an E:T ratio of 3:1. nously with 5 106 cells per mouse through lateral tail vein Subsequently, samples were stained with anti-CD11b-Alexa injection. Animals were sacrificed when hind leg paralysis Fluor-488 (AF; Life Technologies) to identify macrophages, was present or body weight decreased by >20% of pretreat- washed and fixed with 4% paraformaldehyde. Cell-associ- ment weight. Kaplan–Meier analysis was performed for ated fluorescence of PKH26 (FL2, Target cells) and AF (FL1, survival studies and %increase of lifespan (%ILS) was Macrophages) was measured by flow cytometry using a calculated as %ILS ¼ 100 (median survival of treated FC500 cytometer. Dual-labeled cells were considered as group median survival of control group)/median survival phagocytic events, and the degree of phagocytosis was of control group. In accordance with NCI standards, %ILS > determined as: %phagocytosis ¼ 100 [# of FL2þFL1þ 25% was considered active. Induction of cleaved caspase-7 events/(# of FL2þFL1þ events þ # of FL2þFL1- events)]. in tumors was evaluated by immunoblot analysis as In CDC assays, antibodies were incubated with 5 104 described in the Supplementary Data section. target cells per well in the presence of 5% human complement (Sigma-Aldrich) for 2 hours as previously described Results (35). Viability of remaining cells was determined by Ala- Antibody generation and selection marBlue assays (Life Technologies). We generated a large panel of murine monoclonal antibodies that specifically bound to human CD38 (huCD38). Primary multiple myeloma samples To identify antibodies with intrinsic anticancer cell activity, Bone marrow aspirates from patients with multiple antibodies were screened for their ability to induce apopto- myeloma were obtained from Hospital Edouard Herriot sis in lymphoma cells in the absence of any cross-linking (Lyon, France) after written informed consent under agents. One of the antibodies identified had notably strong protocols approved by an Institutional Review Board. proapoptotic activity, inducing 35% Annexin V-positive

4576 Clin Cancer Res; 20(17) September 1, 2014 Clinical Cancer Research

SAR650984: A Novel Antibody for CD38þ Malignancies

Daudi cells compared with 6% Annexin V-positive cells in Immune effector functions mediated by rituximab were the absence of antibody (Fig. 1A). In contrast, other anti- reported to contribute to its clinical activity. We therefore CD38 antibodies generated in our screen, such as clones 7 compared immune effector activities of SAR650984 in þ þ and 23, did not induce a significant increase in Annexin V CD38 CD20 cell lines with those of rituximab. staining of cells. Previously generated anti-CD38 antibodies SAR650984 induced strong CDC in the presence of huAT13/5, chOKT10, IB4, and SUN-4B7 also resulted at human serum as well as strong ADCC via human NK best in a minor (14% or less) increase in the percentage of cells in Raji and Daudi cells, with similar activity as Annexin V-positive Daudi cells. This murine lead clone was rituximab (Fig. 2C and D). In addition, SAR650984 therefore chosen for further analysis based on its unique mediated ADCP with human macrophages against Ramos proapoptotic activity and humanized by variable domain cells to a similar extent as rituximab at 10 mg/mL (Fig. 2E). resurfacing. The chimeric and humanized versions of the lead clone retained specific, high-affinity binding to human SAR650984 activity against a panel of diverse cell lines CD38 with an apparent Kd of 0.2 nmol/L (Fig. 1B). Its CD38 is expressed on a wider range of malignant humanized IgG1 version was designated as SAR650984. cell types than CD20, providing the potential for an expanded utility of a therapeutic anti-CD38 antibody (5). We therefore characterized a larger panel of cell lines for their Functional activity of the humanized anti-CD38 CD38 expression level and SAR650984 in vitro activity. antibody SAR650984 CD38 levels were determined by quantitative flow cyto- Superior proapoptotic activity by SAR650984 over ritux- metry and ranged from 790,000 to 59,000 ABC in the 15 cell imab in the absence of cross-linking agents was seen in Raji, lines tested representing B-cell NHL, multiple myeloma, B- Daudi, and Ramos lymphoma cells, which express CD20 CLL, T-cell acute lymphoblastic leukemia (ALL), and B-cell and CD38 at similar levels (Fig. 2A and B). In addition, þ ALL (Fig. 3A). Most notably, the three multiple myeloma SAR650984 induced apoptosis in CD38 CD20 SU-DHL- cell lines evaluated ranked among the six highest expressors 8 diffuse large B-cell lymphoma (DLBCL) cells, Molp-8 with 790,000 to 233,000 ABC. SAR650984 induced apo- multiple myeloma cells, and DND-41 T-ALL cells (Fig. 2A). ptosis in eight of 15 evaluated cell lines with up to 90% Annexin V-positive cells and EC50 values from 2 to 16 ng/ mL (Fig. 3B). Strong CDC activity was seen in seven of 15 cell lines evaluated with up to 90% maximum lysis and EC50 values varying widely from 8 to 230 ng/mL (Fig. 3C). SAR650984-mediated potent ADCC against all 15 cell lines evaluated with maximum lysis of 30% to 90% of target cells and EC50 values of the dose–response ranged from approximately 0.2 to 8 ng/mL (Fig. 3D). SAR650984 mediated ADCP with 60% phagocytosed Ramos cells, compared with 25% in untreated samples, with an EC50 value of 5 ng/mL (Fig. 3E). Although ADCC activity was seen against all þ CD38 cell lines tested regardless of CD38 expression level, apoptosis induction and CDC activity were most often seen against cell lines with the highest CD38 levels (Fig. 3A). Binding of SAR650984 to normal human peripheral blood cells was assessed by flow cytometry. SAR650984 specifically bound to monocytes and T cells (Supplemen- tary Fig. S1A and S1B) with the same affinity seen on tumor cells. We observed low levels of CD38 expression in normal human blood cells with 3,000 to 14,000 ABC, much lower than levels of CD20 expression on B cells (Supplementary Fig. S1C). SAR650984 treatment did not deplete peripheral blood cells cultured in vitro (Supplementary Fig. S1D), whereas rituximab resulted in B-cell depletion and alemtu- zumab in both B- and T-cell depletion.

Apoptosis induction in primary multiple myeloma patient cells The ability of SAR650984 to induce apoptosis without the addition of external cross-linking agent was also tested Figure 1. Identiﬁcation of the lead anti-CD38 antibody. A, induction of apoptosis by indicated antibodies in Daudi cells measured by Annexin V- in primary cells isolated from bone marrow aspirates of 7 FITC staining. B, binding of the humanized anti-CD38 antibody patients with multiple myeloma. SAR650984 noticeably SAR650984 and its chimeric variant to Daudi cells by ﬂow cytometry. increased the percentages of Annexin V-positive cells over

www.aacrjournals.org Clin Cancer Res; 20(17) September 1, 2014 4577

Deckert et al.

Figure 2. In vitro activity of SAR650984 and rituximab. A, induction of apoptosis by SAR650984 or rituximab in various cell lines. B, CD20 and CD38 expression in lymphoma cells by ﬂow cytometry. C, CDC activity against Raji (left) or Daudi (right) cells. D, ADCC activity against Raji (left) or Daudi (right) target cells. E, ADCP activity against Ramos target cells.

background levels in all multiple myeloma samples tested epitope bound by SAR650984-Fab are highlighted with a mean increase of 25% Annexin V-positive cells in Fig. 4D and notably include a continuous segment from (Fig. 3F). Met110 to Cys119. Mutations introduced to silence the four glycosylation sites of CD38 are not part of the huCD38/ Enzymatic activity SAR650984-Fab interface. Inhibition of the ADP ribosyl-cyclase enzymatic activ- A close-up view of the epitope shows that the interface ity of human CD38 by SAR650984 and other anti-CD38 between huCD38 and SAR650984-Fab involves only part antibodies was evaluated in vitro. SAR650984 inhibited the of the residues in the CDR loops (Fig. 4E). Heavy chain enzymatic function of 5 nmol/L recombinant CD38 almost CDR Loop H3 protrudes out of the structure of completely at concentrations of 200 and 20 nmol/L with SAR650984-Fab and is particularly critical for complex stabi- partial inhibition at 2 nmol/L concentration (Fig. 4A). A lization, whereas Loop H1, H2, L1, L2, and L3 are located at human IgG1 isotype control antibody had no measurable the periphery of the interface. Four salt-bridges form the inhibitory effect in these experiments. Other anti-CD38 interactions between H1 and H2 with Arg194 and Lys111 antibodies tested, Ab005 (a surrogate antibody for daratu- of huCD38 (Supplementary Fig. S2). Binding of SAR650984- mumab), chHB7, huAT13/5, and chOKT10, resulted in Fab induces local conformational changes in the huCD38 much lower or no inhibitory effect (Fig. 4B). structure. The most significant conformational change affects the segment Gln115-Thr116-Val117 of huCD38 that is 3D structure of the huCD38/SAR650984-Fab complex shifted by approximately 2.5 Atoward the light chain, creating Mapping of the epitope and identification of the paratope enough space to accommodate the H3 loop of SAR650984- were achieved by determining the crystal structure of sol- Fab (Fig. 4E). However, the configuration of key residues uble human CD38 (huCD38) in complex with the Fab involved in the ADP-ribosyl cyclase activity of CD38 is main- fragment from SAR650984 at 1.53 A resolution (PDB ID tained and the catalytic site remains accessible (Fig. 4F), code: 4 CMH; Fig. 4C). Residues that are part of the huCD38 suggesting that SAR650984 is likely an allosteric antagonist.

4578 Clin Cancer Res; 20(17) September 1, 2014 Clinical Cancer Research

SAR650984: A Novel Antibody for CD38þ Malignancies

þ Figure 3. In vitro activity of SAR650984 on CD38 cells. A, CD38 expression by quantitative ﬂow cytometry. Representative dose–response curve examples þ þ for (B) apoptosis induction, (C) CDC activity, (D) ADCC activity, and (E) ADCP activity of SAR650984. F, apoptosis induction by SAR650984 in CD38 CD138 primary multiple myeloma cells.

In vivo activity against lymphoma and leukemia models In vivo activity against multiple myeloma models We evaluated the in vivo efficacy of SAR650984 in com- Next, we examined the utility of SAR650984 in subcu- þ parison with established lymphoma treatments such as taneous xenograft models derived from CD20 CD38 rituximab and chemotherapy. SAR650984 showed similar multiple myeloma cell lines, Molp-8 and NCI-H929, in antitumor activity as rituximab in a disseminated xenograft comparison with bortezomib, an established multiple þ þ model established from CD20 CD38 Daudi lymphoma myeloma treatment. In a study with palpable Molp-8 cells (Fig. 5A). Treatment with 40 mg/kg SAR650984 or tumors, SAR650984 was well tolerated and active at rituximab twice weekly for 3 weeks markedly increased 40, 25, and 15 mg/kg administered twice weekly for survival with a median increase in lifespan (%ILS) of 3 weeks with T/C values of 8%, 10%, and 12%, respec- 68% for SAR650984 and 52% for rituximab as compared tively (Fig. 6A). Bortezomib treatment at 0.7 or 0.4 mg/kg with untreated control animals. SAR650984 was also highly administered twice weekly for 2 weeks was inactive, efficacious in a disseminated NALM-6 B-ALL xenograft whereas higher doses of 1.2 or 1.9 mg/kg were toxic. model, with a median%ILS of >200% at doses of 2.5 Similarly, in a study with established NCI-H929 tumors, mg/kg (Fig. 5B). In a subcutaneous xenograft model derived SAR650984 treatment was well tolerated and highly active from SU-DHL-8, a diffuse large B-cell lymphoma cell line, at 40, 20, and 10 mg/kg 6 with 10 of 10 tumor-free SAR650984 was active at all doses tested with T/C values of survivors and at 5 mg/kg 6 with 9 of 10 tumor-free 8%, 14%, 14%, or 26% at the 40, 20, 10, or 5 mg/kg dose, survivors at study end (Fig. 6B). Bortezomib administered respectively (Fig. 5C). No significant weight loss was twice weekly for 2 weeks was highly active at 0.6 mg/kg (T/C observed as a result of SAR650984 treatment in any of these 7.8%; 4/10 tumor-free survivors) and at 1 mg/kg (10/10 studies. CPA, an established lymphoma treatment compo- tumor-free survivors). A higher bortezomib dose of 1.6 mg/ nent, was inactive at the 56 mg/kg dose level, active at 145 kg was toxic, whereas a dose of 0.4 mg/kg was inactive. and 90 mg/kg (T/C values of 3% or 17%, respectively), and was not tolerated at a higher dose of 234 mg/kg. SU-DHL-8 Discussion tumors excised after a single SAR650984 treatment were The introduction of therapeutic monoclonal anti-CD20 analyzed for the levels of cleaved caspase-7, a marker for antibodies, such as rituximab, ofatumumab, and very apoptotic pathway activation, in comparison with untreat- recently obinutuzumab, is improving outcomes for patients ed tumors. SAR650984 induced a 64% or 169% increase in with B-cell malignancies, although unmet medical need cleaved caspase-7 levels after 14 or 24 hours, respectively remains. It would be desirable to develop antibody thera- (Fig. 5D). peutics for other hematologic diseases, in particular

www.aacrjournals.org Clin Cancer Res; 20(17) September 1, 2014 4579

Deckert et al.

Figure 4. CD38 enzymatic activity and 3D-structure of the huCD38/SAR650984-Fab complex. A, inhibition of CD38 enzymatic activity by SAR650984 þ measured in vitro by the production of cyclic GDPRibose (cGDPR) from NGD . B, inhibition of CD38 enzymatic activity at 90 minutes in the presence of 20 nmol/L of various anti-CD38 antibodies. C, overall structure of the complex in surface and ribbon representation with SAR650984-Fab heavy chains (yellow), light chains (green), and human CD38 (gray). D, amino-acid sequence of the R45-I300 huCD38 extracellular domain with SAR650984-Fab epitope residues in blue, mutations introduced to silence the four glycosylation sites of CD38 in light gray, CD38 active site residues in red. Epitope residues were deﬁned as residues which contain atoms within 4A from any atom of the SAR650984-Fab CDRs. E, detailed view of the epitope highlighting the contributions of loops H3 (yellow) and L1, L2, and L3 (green). F, overlay of CD38/SAR650984–FabonfreeCD38(PBD2EF1,red),CD38 in complex with GTP (PBD 3DZH, magenta), and with NAD (PBD 2I65, cyan). Essential residues of CD38 involved in the enzymatic activity are shown in stick representation.

multiple myeloma, that could also combine high efﬁcacy rituximab, SAR650984 mediated ADCC and ADCP effector with low systemic toxicity. Our goal was to build upon the activities with potent ADCC activity seen against cell lines þ success story of CD20 antibody therapy by developing a derived from diverse CD38 malignancies. Unlike type II novel multifunctional antibody directed against a different anti-CD20 antibodies, SAR650984 also induced CDC in target with expanded utility in a range of hematologic lymphoma cell lines with activity comparable with ritux- malignancies. Here, we describe the generation and char- imab. SAR650984 demonstrated potent in vivo antitumor acterization of SAR650984, a CD38-targeting antibody with activities similar to rituximab in both disseminated and þ þ a unique combination of antitumor activities. subcutaneous CD20 CD38 lymphoma models. Against The ability of SAR650984 to induce apoptosis in vitro was DLBCL SU-DHL-8 tumors SAR650984 activity compared superior to rituximab in B-cell lymphoma cell lines with favorably with CPA, a component of the standard treatment similar CD38 and CD20 expression, with potent activity regimen CHOP, whose in vivo activity was limited by tox- even in the absence of cross-linking agents. This type of icity. In this model, SAR650984 also induced cleaved cas- enhanced proapoptotic activity has been described for type pase-7, a marker of apoptotic cell death. Notably, II anti-CD20 antibodies such as tositumumab or obinutu- SAR650984 was active against Molp-8 and NCI-H929 mul- zumab (37). In addition, SAR650984 induced apoptosis in tiple myeloma xenograft tumors, whereas bortezomib, an multiple myeloma and T-ALL cell lines, which are typically agent commonly used in multiple myeloma treatment, was CD20 , as well as a CD20 DLBCL cell line. Similarly to only active against NCI-H929 tumors.

4580 Clin Cancer Res; 20(17) September 1, 2014 Clinical Cancer Research

SAR650984: A Novel Antibody for CD38þ Malignancies

Figure 5. In vivo efﬁcacy of SAR650984 in lymphoma and leukemia xenograft models. SCID mice (n ¼ 10 per group) bearing xenografts were treated with intraperitoneal administration of SAR650984 or reference compounds as indicated, (A) disseminated Daudi xenografts, (B) disseminated Nalm-6 xenografts, (C) subcutaneous SU-DHL-8 lymphoma xenografts treated with SAR650984 or CPA. D, induction of cleaved caspase-7 after a single dose of 40 mg/kg SAR650984 in SU-DHL-8 xenograft in SCID mice evaluated by immunoblot analysis.

Several anti-CD38 antibodies with various functional orthosteric antagonist of CD38. This is different from HB7, properties have been identified previously. Although an antibody that does not compete with SAR650984 (Sup- þ some of these can mediate in vitro killing of CD38 cell plementary Fig. S3), binds at a different location in the lines via ADCC and/or CDC effector function, their direct C-terminal portion of the CD38 extracellular domain (Sup- impact on cell proliferation or apoptosis remains unclear plementary Fig. S4), and does not inhibit CD38 enzymatic (10, 22–24). For example, the agonistic antibody IB4 was activity (Fig. 4B). The SAR650984 binding mode appears reported to prevent apoptosis or induce proliferation in unique as other anti-CD38 antibodies tested here or some contexts (10, 38), and induce apoptosis in another described elsewhere show no or much less potent inhibition (9). The more recently developed antibody daratumumab of the enzymatic activity. hasbeenshowntomediateADCCaswellasADCPbutto CD38 enzymatic activities produce nucleotide-metabo- þ induce apoptosis only in the presence of stromal cells or lites including cADPR, which induce Ca2 mobilization cross-linking agents (28, 39). It was selected on the basis and signaling (7, 8, 43). However, the physiologic role of of its CDC activity, which may be linked to its unique CD38 is still unknown, as functional studies have mainly epitope. Another recent addition, MOR202 primarily used cell lines ectopically overexpressing CD38 and anti- possesses ADCC activity with CDC and ADCP also being CD38 antibodies, which are nonphysiologic ligands (44, observed with no reported intrinsic proapoptotic activity 45). Therefore, the exact interplay between enzymatic activ- (31, 32). ity, receptor function and consequences on signaling path- Analysis of the 3D structure of SAR650984-Fab with ways and cellular response remains to be further documen- human CD38 revealed that SAR650984 binds to a discon- ted. As a consequence, the physiologic impact of tinuous epitope which includes amino acids located oppo- SAR650984 inhibition of CD38 enzymatic activity remains site to the catalytic site of CD38. Although significant to be deciphered and how the functional differences conformational changes are observed in CD38 upon bind- observed between various anti-CD38 antibodies affect their ing by SAR650984, the overall configuration of key residues therapeutic utility requires further investigation. The con- involved in the CD38 enzymatic activity and access to the tribution of distinct mechanisms of action of SAR650984 catalytic site appear conserved (40–42). Nevertheless, to clinical efficacy may also differ depending on the type SAR650984 strongly inhibits CD38 enzymatic activity sug- of hematologic malignancy, tumor biology, and tumor gesting that SAR650984 is an allosteric rather than an microenvironment.

www.aacrjournals.org Clin Cancer Res; 20(17) September 1, 2014 4581

Deckert et al.

antibody could offer the prospect of a highly effective therapy in multiple myeloma that might also be used readily in combination with other agents to improve clinical outcomes. On the basis of the results described here, SAR650984 has shown potent in vitro and in vivo activity þ against diverse CD38 lymphoma, leukemia as well as multiple myeloma-derived cell line models and can induce apoptosis in primary multiple myeloma cells. Therefore, it presents an exciting opportunity for a novel antibody ther- þ apeutic in CD38 hematologic malignancies and is currently in clinical testing in patients with multiple myeloma with early signs of efﬁcacy (48).

Disclosure of Potential Conﬂicts of Interest T. Chittenden holds ownership interest (including patents) in Immuno- Gen. No potential conﬂicts of interest were disclosed by the other authors.

Authors' Contributions Conception and design: M.-C. Wetzel, V.S. Goldmacher, C. Lahoute, C. Dumontet, P. Lejeune, P.U. Park, V. Blanc Development of methodology: M.-C. Wetzel, Q. Zhou-Liu, P. Ferrari, P.U. Park Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Deckert, M.-C. Wetzel, L.M. Bartle, A. Skalets- kaya, F. Vallee, S. Pouzieux, C. Dumontet, A. Plesa, M. Chiron, P.U. Park Analysis and interpretation of data (e.g., statistical analysis, biosta- tistics, computational analysis): J. Deckert, L.M. Bartle, F. Vallee, C. Dumontet, A. Plesa, M. Chiron, P. Lejeune, P.U. Park, V. Blanc Figure 6. In vivo efficacy in multiple myeloma xenograft models. SCID Writing, review, and/or revision of the manuscript: J. Deckert, M.-C. Wetzel, F. Vallee, Q. Zhou-Liu, C. Dumontet, M. Chiron, P. Lejeune, mice bearing subcutaneous xenografts of (A) Molp-8 or (B) NCI-H929 T. Chittenden, P.U. Park, V. Blanc cells treated as indicated with SAR650984 or bortezomib. Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): Q. Zhou-Liu, P.U. Park Use of chemotherapy and autologous stem cell trans- Study supervision: M.-C. Wetzel, P. Lejeune, T. Chittenden, P.U. Park plantation and the availability of new therapeutic agents including proteasome, kinase, and histone deacetylase inhi- Acknowledgments bitors, as well as immune-modulatory drugs have improved The authors thank Daniel Tavares, Michele Mayo, Helene Bastien, Josiane Le Parc, Francois¸ Michoux, Anne Pommeret, Dominique Quarteronet, survival in patients with myeloma (46, 47). However, the Florence Attenot, and Catherine Geslin for technical support, Ti Cai for majority of patients will eventually relapse underscoring the critical reading of the manuscript, and Walter Bl€attler, Thierry Hercend, need for additional therapeutic options. Moreover, many of Michel Streuli, and John Lambert for guidance and helpful discussions. The costs of publication of this article were defrayed in part by the these small-molecule agents have significant toxicities, such payment of page charges. This article must therefore be hereby marked as neuropathy, that limit their dosing and their potential for advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate use in combination therapy. Drawing on the example of this fact. rituximab, with its favorable safety profile and success in Received March 21, 2014; revised June 5, 2014; accepted June 23, 2014; combination therapy for lymphoma, a functional CD38 published OnlineFirst July 1, 2014.

References 1. Bello C, Sotomayor EM. Monoclonal antibodies for B-cell lymphomas: 7. Lee HC, Munshi C, Graeff R. Structures and activities of cyclic ADP- rituximab and beyond. Hematology Am Soc Hematol Educ Program ribose, NAADP and their metabolic enzymes. Mol Cell Biochem 2007:233–42. 1999;193:89–98. 2. Dotan E, Aggarwal C, Smith MR. Impact of rituximab (rituxan) on the 8. Berthelier V, Tixier JM, Muller-Steffner H, Schuber F, Deterre P. Human treatment of B-cell non-Hodgkin's mymphoma. P T 2010;35:148–57. CD38 is an authentic NAD(P)þ glycohydrolase. Biochem J 1998;330: 3. Alduaij W, Illidge TM. The future of anti-CD20 monoclonal antibodies: 1383–90. are we making progress? Blood 2011;117:2993–3001. 9. Morra M, Zubiaur M, Terhorst C, Sancho J, Malavasi F. CD38 is 4. Cang S, Mukhi N, Wang K, Liu D. Novel CD20 monoclonal antibodies functionally dependent on the TCR/CD3 complex in human T cells. for lymphoma therapy. J Hematol Oncol 2012;5:64. FASEB J 1998;12:581–92. 5. Konopleva M, Rissling I, Andreeff M. CD38 in hematopoietic malig- 10. Konopleva M, Estrov Z, Zhao S, Andreeff M, Mehta K. Ligation of cell nancies. Chem Immunol 2000;75:189–206. surface CD38 protein with agonistic monoclonal antibody induces a 6. Quarona V, Zaccarello G, Chillemi A, Brunetti E, Singh VK, Ferrero E, cell growth signal in myeloid leukemia cells. J Immunol 1998;161: et al. CD38 and CD157: a long journey from activation markers 4702–8. to multifunctional molecules. Cytometry B Clin Cytom 2013;84: 11. Zilber MT, Gregory S, Mallone R, Deaglio S, Malavasi F, Charron D, 207–17. et al. CD38 expressed on human monocytes: a coaccessory molecule

4582 Clin Cancer Res; 20(17) September 1, 2014 Clinical Cancer Research

SAR650984: A Novel Antibody for CD38þ Malignancies

in the superantigen-induced proliferation. Proc Natl Acad Sci U S A 31. Tesar M, Springer M, Swoboda B, Brocks B, Popp A, Broders O, et al. 2000;97:2840–5. Fully human antibody MOR202 against CD38 for the treatment of 12. Deaglio S, Vaisitti T, Billington R, Bergui L, Omede P, Genazzani AA, multiple myeloma and other blood-borne malignancies. J Clin Oncol et al. CD38/CD19: a lipid raft-dependent signaling complex in human B 2007;25:8106. cells. Blood 2007;109:5390–8. 32. Endell J, Boxhammer R, Wurzenberger C, Ness D, Steidl S. The Activity 13. Reinherz EL, Kung PC, Goldstein G, Levey RH, Schlossman SF. of MOR202, a Fully Human Anti-CD38 Antibody, Is Complemented by Discrete stages of human intrathymic differentiation: analysis of nor- ADCP and Is Synergistically Enhanced by Lenalidomide in Vitro and in mal thymocytes and leukemic lymphoblasts of T-cell lineage. Proc Natl Vivo. ASH Annual Meeting Abstracts 2012;120:4018. Acad Sci U S A 1980;77:1588–92. 33. Roguska MA, Pedersen JT, Keddy CA, Henry AH, Searle SJ, Lambert 14. Tedder TF, Clement LT, Cooper MD. Discontinuous expression of a JM, et al. Humanization of murine monoclonal antibodies through membrane antigen (HB-7) during B lymphocyte differentiation. Tissue variable domain resurfacing. Proc Natl Acad Sci U S A 1994;91: Antigens 1984;24:140–9. 969–73. 15. Terstappen LW, Huang S, Safford M, Lansdorp PM, Loken MR. 34. Shields RL, Namenuk AK, Hong K, Meng YG, Rae J, Briggs J, et al. High Sequential generations of hematopoietic colonies derived from single resolution mapping of the binding site on human IgG1 for Fc gamma RI, nonlineage-committed CD34þCD38- progenitor cells. Blood 1991;77: Fc gamma RII, Fc gamma RIII, and FcRn and design of IgG1 variants 1218–27. with improved binding to the Fc gamma R. J Biol Chem 2001; 16. Ferrero E, Malavasi F. The metamorphosis of a molecule: from soluble 276:6591–604. enzyme to the leukocyte receptor CD38. J Leukoc Biol 1999;65: 35. Gazzano-Santoro H, Ralph P, Ryskamp TC, Chen AB, Mukku VR. 151–61. A non-radioactive complement-dependent cytotoxicity assay for 17. Damle RN, Wasil T, Fais F, Ghiotto F, Valetto A, Allen SL, et al. Ig V gene anti-CD20 monoclonal antibody. J Immunol Methods 1997;202: mutation status and CD38 expression as novel prognostic indicators in 163–71. chronic lymphocytic leukemia. Blood 1999;94:1840–7. 36. Graeff RM, Walseth TF, Fryxell K, Branton WD, Lee HC. Enzymatic 18. Ibrahim S, Keating M, Do KA, O'Brien S, Huh YO, Jilani I, et al. CD38 synthesis and characterizations of cyclic GDP-ribose. A procedure for expression as an important prognostic factor in B-cell chronic lym- distinguishing enzymes with ADP-ribosyl cyclase activity. J Biol Chem phocytic leukemia. Blood 2001;98:181–6. 1994;269:30260–7. 19. Hamblin TJ, Orchard JA, Ibbotson RE, Davis Z, Thomas PW, Steven- 37. Mossner E, Brunker P, Moser S, Puntener U, Schmidt C, Herter S, et al. son FK, et al. CD38 expression and immunoglobulin variable region Increasing the efﬁcacy of CD20 antibody therapy through the engi- mutations are independent prognostic variables in chronic lympho- neering of a new type II anti-CD20 antibody with enhanced direct and cytic leukemia, but CD38 expression may vary during the course of the immune effector cell-mediated B-cell cytotoxicity. Blood 2010;115: disease. Blood 2002;99:1023–9. 4393–402. 20. Deaglio S, Vaisitti T, Aydin S, Ferrero E, Malavasi F. In-tandem insight 38. Zupo S, Rugari E, Dono M, Taborelli G, Malavasi F, Ferrarini M. from basic science combined with clinical research: CD38 as both CD38 signaling by agonistic monoclonal antibody prevents apo- marker and key component of the pathogenetic network underlying ptosis of human germinal center B cells. Eur J Immunol 1994; chronic lymphocytic leukemia. Blood. 2006;108:1135–44. 24:1218–22. 21. Lin P, Owens R, Tricot G, Wilson CS. Flow cytometric immunophe- 39. Overdijk MB, Verploegen S, Marijn B, van Egmond M, Groen RWJ, notypic analysis of 306 cases of multiple myeloma. Am J Clin Pathol Martens ACM, et al.Phagocytosis Is A Mechanism of Action for 2004;121:482–8. Daratumumab. ASH Annual Meeting Abstracts 2012;120:4504. 22. Stevenson FK, Bell AJ, Cusack R, Hamblin TJ, Slade CJ, Spellerberg 40. Liu Q, Kriksunov IA, Graeff R, Munshi C, Lee HC, Hao Q. Crystal MB, et al. Preliminary studies for an immunotherapeutic approach to structure of human CD38 extracellular domain. Structure 2005;13: the treatment of human myeloma using chimeric anti-CD38 antibody. 1331–9. Blood 1991;77:1071–9. 41. Liu Q, Kriksunov IA, Graeff R, Munshi C, Lee HC, Hao Q. Structural 23. Ausiello CM, Urbani F, Lande R, la Sala A, Di Carlo B, Baj G, et al. basis for the mechanistic understanding of human CD38-controlled Functional topography of discrete domains of human CD38. Tissue multiple catalysis. J Biol Chem 2006;281:32861–9. Antigens 2000;56:539–47. 42. Liu Q, Graeff R, Kriksunov IA, Lam CM, Lee HC, Hao Q. Conformational 24. Goldmacher VS, Bourret LA, Levine BA, Rasmussen RA, Pourshadi M, Closure of the Catalytic Site of Human CD38 Induced by Calcium. Lambert JM, et al. Anti-CD38-blocked ricin: an immunotoxin for the Biochemistry. 2008;47:13966–73. treatment of multiple myeloma. Blood 1994;84:3017–25. 43. Lee HC, Graeff RM, Walseth TF. ADP-ribosyl cyclase and CD38. Multi- 25. Ellis JH, Barber KA, Tutt A, Hale C, Lewis AP, Glennie MJ, et al. functional enzymes in Caþ2 signaling. Adv Exp Med Biol 1997;419: Engineered anti-CD38 monoclonal antibodies for immunotherapy of 411–9. multiple myeloma. J Immunol 1995;155:925–37. 44. Kitanaka A, Suzuki T, Ito C, Nishigaki H, Coustan-Smith E, Tanaka T, 26. Maloney DG, Donovan K, Hamblin TJ. Antibody therapy for treatment et al. CD38-mediated signaling events in murine pro-B cells expressing of multiple myeloma. Semin Hematol 1999;36:30–3. human CD38 with or without its cytoplasmic domain. J Immunol 27. Stevenson GT, Bell AJ, Evans TR, French RR, Glennie MJ, Hamblin TJ, 1999;162:1952–8. et al. Mechanisms in removal of tumor by antibody. Cell Biophys 45. Hoshino S, Kukimoto I, Kontani K, Inoue S, Kanda Y, Malavasi F, et al. 1994;24–25:45–50. Mapping of the catalytic and epitopic sites of human CD38/NADþ 28. de Weers M, Tai YT, van der Veer MS, Bakker JM, Vink T, Jacobs DC, glycohydrolase to a functional domain in the carboxyl terminus. et al. Daratumumab, a novel therapeutic human CD38 monoclonal J Immunol 1997;158:741–7. antibody, induces killing of multiple myeloma and other hematological 46. Munshi NC, Anderson KC. New strategies in the treatment of multiple tumors. J Immunol 2011;186:1840–8. myeloma. Clin Cancer Res 2013;19:3337–44. 29. Gimsing P, Plesner T, Nahi H, Lokhorst H, Valentin M-L, Lisby S, et al. A 47. Gentile M, Recchia AG, Mazzone C, Lucia E, Vigna E, Morabito F. Phase I/II, Dose-Escalation Study of Daratumumab, A CD38 Mab in Perspectives in the treatment of multiple myeloma. Expert Opin Biol Patients with Multiple Myeloma - Preliminary Safety Data. ASH Annual Ther 2013;13:S1–22. Meeting Abstracts 2011;118:1873. 48. Martin TG, Strickland SA, Glenn M, Zheng W, Daskalakis N, Mikhael 30. Plesner T, Lokhorst H, Gimsing P, Nahi H, Lisby S, Richardson PG. JR. SAR650984, a CD38 Monoclonal Antibody In Patients With Daratumumab, a CD38 Monoclonal Antibody in Patients with Multiple Selected CD38 þHematological Malignancies- Data From a Dose- Myeloma - Data From a Dose-Escalation Phase I/II Study. ASH Annual Escalation Phase I Study. ASH Annual Meeting Abstracts 2013; Meeting Abstracts 2012;120:73. 122:284.

www.aacrjournals.org Clin Cancer Res; 20(17) September 1, 2014 4583

SAR650984, A Novel Humanized CD38-Targeting Antibody, Demonstrates Potent Antitumor Activity in Models of Multiple Myeloma and Other CD38+ Hematologic Malignancies

Jutta Deckert, Marie-Cécile Wetzel, Laura M. Bartle, et al.

Clin Cancer Res 2014;20:4574-4583. Published OnlineFirst July 1, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-14-0695

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2014/07/14/1078-0432.CCR-14-0695.DC1

Cited articles This article cites 43 articles, 23 of which you can access for free at: http://clincancerres.aacrjournals.org/content/20/17/4574.full#ref-list-1

Citing articles This article has been cited by 27 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/20/17/4574.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://clincancerres.aacrjournals.org/content/20/17/4574. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.