Yu et al. Blood Cancer Journal (2020) 10:110 https://doi.org/10.1038/s41408-020-00378-z Blood Cancer Journal

ARTICLE Open Access VIS832, a novel CD138-targeting , potently induces killing of human and further synergizes with IMiDs or bortezomib in vitro and in vivo Tengteng Yu1,2, Bharat Chaganty 3,LiangLin1, Lijie Xing1, Boopathy Ramakrishnan3, Kenneth Wen1, Phillip A. Hsieh1, Andrew Wollacott3, Karthik Viswanathan3, Hedy Adari3, Shih-Feng Cho1,4,5,YuyinLi1,6, Hailin Chen1,7,WenjuanYang1, Yan Xu 1,2,GangAn2,LuguiQiu2,NikhilMunshi1, Gregory Babcock3, Zachary Shriver3, James R. Myette3, Kenneth C. Anderson 1 and Yu-Tzu Tai 1

Abstract Therapeutically targeting CD138, a define multiple myeloma (MM) antigen, is not yet approved for patients. We here developed and determined the preclinical efficacy of VIS832, a novel therapeutic monoclonal antibody (MoAb) with differentiated CD138 target binding to BB4 that is anti-CD138 MoAb scaffold for indatuximab ravtansine (BT062). VIS832 demonstrated enhanced CD138-binding avidity and significantly improved potency to kill MM cell lines and autologous patient MM cells regardless of resistance to current standard-of-care therapies, via robust antibody- dependent cellular cytotoxicity and phagocytosis mediated by NK and macrophage effector cells, respectively. Specifically, CD38-targeting -resistant MM cells were highly susceptible to VIS832 which, unlike 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; daratumumab, spares NK cells. Superior maximal cytolysis of VIS832 vs. daratumumab corresponded to higher CD138 vs. CD38 levels in MM cells. Furthermore, VIS832 acted synergistically with or bortezomib to deplete MM cells. Importantly, VIS832 at a sub-optimal dose inhibited disseminated MM1S tumors in vivo as monotherapy (P < 0.0001), and rapidly eradicated myeloma burden in all mice concomitantly receiving bortezomib, with 100% host survival. Taken together, these data strongly support clinical development of VIS832, alone and in combination, for the therapeutic treatment of MM in relapsed and refractory patients while pointing to its potential therapeutic use earlier in disease intervention.

Introduction Multiple myeloma (MM) is characterized by excess monoclonal plasma cells (PCs) in the bone marrow (BM) producing monoclonal immunoglobulins, and is asso- Correspondence: Kenneth C. Anderson ([email protected]) ciated with hypercalcemia, renal dysfunction, anemia, and or Yu-Tzu Tai ([email protected]) osteolytic bone disease1,2. Despite the recent use of novel 1 Jerome Lipper Multiple Myeloma Center, LeBow Institute for Myeloma therapies including proteasome inhibitors, i.e., bortezo- Therapeutics, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA mib (btz), immunomodulatory drugs (IMiDs), i.e., lenali- 2State Key Laboratory of Experimental Hematology, National Clinical Research domide (len) and pomalidomide (pom), and monoclonal Center for Hematological Disorders, Institute of Hematology and Blood antibodies (mAbs), i.e., daratumumab (dara) and isatux- Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin 300020, China imab targeting CD38, alone or in combination, MM Full list of author information is available at the end of the article

© The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a linktotheCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 2 of 13

remains an incurable disease in most patients due to the Materials and methods development of drug resistance underlying relapse of CD138 antibodies disease3,4. Thus, it is urgent to develop well tolerated VIS832 was produced by transient vector transfection in novel targeted to treat relapsed and Chinese hamster ovary cells, as described in Supplemen- refractory (RR) MM, which can then be used to treat tary Methods. The secreted mAb was purified from cell earlier stages of disease and further improve patient culture media using protein A affinity capture on Fast outcome. Protein Liquid chromatography (FPLC), in accordance CD138 (Syndecan-1, SDC1), a member of the integral with established protocols. membrane family of heparan sulfate proteoglycans, is highly expressed on differentiated PCs. It is overexpressed Antibody-dependent cellular cytotoxicity (ADCC) assay in patient MM cells compared with normal PCs, and a ADCC activity was mainly measured based on the validated diagnostic biomarker of MM. Importantly, calcein-AM release from calcein-AM labeled target MM CD138 is a co-receptor for MM cell growth, adhesion, and cells co-incubated with human NK or PBMC effector cells – survival5 8, as well as other critical aspects of myeloma isolated from multiple healthy donors or MM patients, in – biology9 13. Its expression further correlates with disease the presence of test Abs. Unless otherwise noted, target – progression and prognosis14 16. Thus, CD138 is a pro- MM cells and NK effector cells were co-cultured at an mising antigen for mAb-based of all effector: target (E:T) ratio of 10:1. ADCC data following stages of MM. Although CD138 targeted agents have and antibody titration were plotted using a nonlinear regres- will be developed first to treat RRMM, CD138 is also sion analysis and a 4-parameter curve fit (GraphPad), expressed on smoldering myeloma (SMM), prior to the from which EC50 values were calculated. development of MM without evidence of end-organ Bioluminescence (BLI)-based assays were also used – damage17 19. Indeed, several clinical trials of MM thera- when target MM cell transfectants expressing a luciferase pies are now under evaluation to delay the progression of gene were used28. When determining VIS832-induced – SMM to active disease20 24. Such therapies must not only ADCC against autologous patient MM cells in BMMCs delay progression of disease, but also be well tolerated, from patient BM samples, flow cytometry-based analysis since these patients are asymptomatic. gated on viable BCMA+ fraction was used. To date, targeting CD138 for treatment of MM was clinically evaluated using indatuximab ravtansine Antibody-dependent cellular phagocytosis (ADCP) assay (BT062), an anti-CD138 mAb (BB4) drug conjugate Indicated target MM cell lines (n ≥ 5) sensitive or (ADC) specifically delivering a cytotoxic maytansine resistant to (dex), IMiDs, or dara were derivative to MM cells25. Indatuximab ravtansine showed labeled with a fluorescence cell membrane dye Cell encouraging anti-MM activity with a reasonable safety Trace™ Far Red, washed, and incubated for 4 h with profile in a recent Phase Ib/IIa open label, multi-dose macrophages labeled green with CFSE in the presence of escalation, clinical trial in RRMM26. In a preclinical VIS832 or antibody isotype control in triplicate at indi- study, CD138-based chimeric antigen receptor T cells cated E:T ratios (1:1, 2:1, or 4:1). Cells were fixed with 1% (CAR-T) also demonstrated significant killing of MM paraformaldehyde followed by flow cytometry to detect cells, without off tumor cytotoxicity against normal epi- double-positive (DP) fluorescence, an indication of pha- thelial or endothelial cells27.Ananti-CD138mAb-based gocytic activity29. therapy with potent immune-mediated cytotoxicity could represent an effective and safer modality compared to an Murine model of disseminated human MM disease ADC or CAR-T. Both ADC and CAR-T approaches have The in vivo efficacy of VIS832 and btz, alone or in potential liabilities or limitations, e.g., dose-limiting combination, were evaluated using a murine model of the toxicities in the case of ADCs or a less durable clinical dissemination of MM1S-luc cells intravenously (i.v.) response, cytokine release syndrome, prohibitive cost of injected into CB-17 SCID mice. After 14 days, whole body manufacture, and limited point of patient access in the tumor burden was monitored by BLI. Mice (n = 9 per case of CAR-T. group) were randomized and treatments started. Vehicle We here developed a novel anti-CD138 mAb VIS832 control (PBS) was dosed once daily by intraperitoneal and characterized its mode of actions in multiple pre- (i.p.) injection (0.2 ml/20 g per mouse); VIS832 (4 mg/kg) clinical models of MM. In concert with its effective and btz (1 mg/kg) were dosed twice weekly by i.p. injec- immune-mediated killing of MM cells as a single agent, tion for 52 days. Treatment in all groups was discontinued in vitro and in vivo evaluations of VIS832 combined at Day 53; surviving animals were monitored for an with len or btz demonstrated augmented efficiency and additional 3 weeks till Day 73. tolerability, providing the rationale for its clinical See Supplementary “Materials and Methods” for addi- evaluation in MM. tional details.

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 3 of 13

Results minimal ADCC (Fig. 1c). Afucosylation of the Fc N glycan Robust binding of VIS832 to CD138 on MM cell lines and on VIS832 further optimized its efficiency and potency, patient MM cells with decreased EC50 and >7-fold increased maximal U266 CD138 protein and mRNA levels were first assayed by MM cell lysis (28.67 ng/ml (0.2 nM), Supplemental Fig. S2). flow cytometry analysis using an anti-CD138 DL-101 Using calcein-AM release assay, VIS832 ADCC activity clone and qRT-PCR, respectively, in MM cell lines (n = was next quantified in the co-cultures of calcein-AM-pre- 12) including paired sensitive and resistant to dex labeled MM cell lines with NK cells from multiple healthy (MM1S/MM1R), IMiDs (both len and pom) (MM1S/ donors (n ≥ 3). VIS832, dose-dependently, induced ADCC MM1S(R), H929/H929(R)), and btz (ANBL6/ANBL6- against MM cell lines (n = 12) (Supplementary Fig. S3A). 28,30 BR) (Supplementary Fig. S1A). All MM cell lines EC50 values of VIS832 ranged from 2.22 ± 0.37 to 15.3 ± highly expressed CD138, as shown by geometric mean 2.71 ng/ml, and % maximal lysis ranged from 37.06 ± fluorescence intensity (MFI) values, with some variability 1.45% to 97.3 ± 3.34% across tested MM cell lines (Fig. 2a in relative expression levels. MM cell lines (n = 29) and Supplementary Table S2). VIS832-induced ADCC expressed highest CD138 levels, compared with 10 other against MM cells generally correlated with CD138 surface hematological cancer cell line cohorts (Supplementary expression levels (Supplementary Fig. S3B). VIS832- Fig. S1B). Soluble CD138 generally correlated with protein induced lower ADCC activity against target JJN3 (Fig. and transcript levels in most of the corresponding MM 2a, b) and ANBL6 (Fig. 2a, c) MM cells, presumably due cell lines (Supplementary Fig. S1C) and was elevated in to relatively lower CD138 expression (e.g., relative to patient serum samples vs. healthy donors (P < 0.001) levels in H929 or MM1S). VIS832-induced ADCC against (Supplementary Fig. S1D). MM cell lines was also consistent with ADCC activity VIS832 is a humanized IgG1κ mAb targeting human induced by the parental mAb 2810, the non-humanized CD138 and has been optimized for its binding avidity to predecessor to VIS832 (Supplementary Fig. S3C). These cell membrane CD138. The epitope of VIS832 includes a comparable data (Supplementary Table S2, EC50 values membrane proximal region of CD138 predicted to confer ranging from 2.39 to 28.16 ng/ml across 10 MM cell lines) productive immune synapse formation for improved Fc confirmed the successful antibody humanization of effector-dependent, immune cell-mediated cytotoxicity. VIS832 with full retention of biological activity. The relative binding affinity of VIS832 to MM cells was The dependency of CD138 expression on VIS832- compared with anti-CD138 antibody BB4, whose VH and induced NK cell-mediated MM cell lysis was validated VL variable regions comprise the paratope for indatux- using two CRISPR/CAS9-generated CD138 (SDC1) gene imab ravtansine25,26. VIS832 showed robust and dose- knockout (KO) H929 and JJN3 MM cell lines (Supple- dependent target engagement, with potent binding mentary Fig. S3D). VIS832 did not induce ADCC against avidity (sub-to low-nanomolar) across MM cell lines (n CD138 KO transfectants vs. the parental (target control, = 8), as indicated by MFI values (Fig. 1a). Significantly, NT) cells with high (in H929) and low (in JJN3) CD138 binding of VIS832 was more intense (P <0.005)thanBB4 levels (Fig. 2b). in all MM cell lines, with multi-fold increases in apparent VIS832 induced equivalent ADCC against paired MM binding affinity (Supplemental Table S1). Importantly, cell lines sensitive and resistant to dex (MM1S/MM1R), VIS832 showed dose- and CD138 target-dependent len/pom-(H929/H929(R)), and btz (ANBL6/ANBL6- binding of PCs derived from BM of patients (4 RR and BR) (Fig. 2a, c). In the presence of BM stromal cells – 1SMM)(Fig.1b). Thus, VIS832 is a potential therapeutic (BMSCs) or osteoclasts (OCs)37 40,twokeyMM- candidate targeting CD138 for further investigation in supporting accessory cells in the BM milieu, VIS832- human MM. induced ADCC against MM cells was minimally affected (Fig. 2d). These data indicated that VIS832 could gen- VIS832 directs target-specific ADCC against MM cell lines erate potent cytotoxicity in MM cells resistant to cur- with acquired resistance to current therapies and in the rent therapies and in the tumorigenic (and generally presence of MM growth promoting BM cells immunosuppressive) BM microenvironment. Antibody-dependent cellular cytotoxicity (ADCC) is an effective and therapeutically validated immune-based VIS832 effectively targets autologous patient MM cells – mechanism for targeted MM cell killing29,31 36.ADCCof VIS832-dependent ADCC against autologous patient VIS832 was first determined using a reporter bioassay CD138 + cells was next evaluated using BMMCs freshly comprised of an engineered Jurkat cell line overexpressing isolated from patient BM aspirates (n = 7, including 6 the human FcγRIIIa receptor as effector cell surrogate co- RRMM and 1 SMM (MM12)). BMMCs that included cultured with target U266 MM cell line. EC50 value for CD138 + MM cells, immune cells, and other BM VIS832 was ~99.31 ng/ml (0.66 nM), whereas the BB4 accessory cells, were directly incubated with various possessing an isotype matched human IgG1 induced only dilutions of VIS832 for 1 day. The fraction of viable

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 4 of 13

MM1S MM1R H929 U266 a 6000 VIS832 10000 VIS832 20000 VIS832 10000 VIS832 **** **** BB4 BB4 **** **** **** ** BB4 **** BB4 8000 **** 8000 15000 4000 **** 6000 6000 10000 MFI MFI

**** MFI

MFI **** **** 4000 4000 2000 5000 2000 *** 2000 ** µ µ µ g/ml 0 0.01 0.1 1 10 g/ml 00.010.11 10 g/ml 00.010.11 10 µg/ml 00.010.11 10 OPM2 RPMI8226 JJN3 MOLP8 20000 2500 1500 VIS832 6000 VIS832 VIS832 VIS832 **** BB4 BB4 **** BB4 **** *** **** BB4 2000 **** **** 15000 **** **** **** 1000 4000 1500 I **** 10000 F M MFI MFI MFI 1000 500 **** 2000 5000 500 * * **

µg/ml 00.010.11 10 µg/ml 00.010.11 10 µg/ml 00.010.11 10 µg/ml 00.010.11 10

b 12000 MM1 c MM2 30 VIS832 MM3 BB4 9000 MM4 MM5 20 6000 MFI induction d (ADCC) 10

3000 Fol

0 0 VIS832 µg/ml 00.010.11 10 µg/ml 0.0001 0.001 0.01 0.1 1 10 100

Fig. 1 Binding of VIS832 was significantly more intense than BB4 to CD138 on MM cell membrane, potentiating FcR-mediated MM cell cytotoxicity. a Indicated MM cell lines (n = 8) were incubated with a serial dilution of VIS832 or BB4 (0–10 μg/ml) followed by flow cytometry (FC) analysis using an APC-conjugated secondary anti-human IgG1 antibody for detection. Binding intensity was reported as geometric mean fluorescence intensity (MFI) normalized to isotype control. Isotype control for VIS832 and BB4 (10 μg/ml) were used as background controls. Three experiments were done in triplicates for each concentration. Data were presented as means (MFIs) ± standard errors of means (SEMs, error bars). **P < 0.01, ***P < 0.005, ****P < 0.001. b Bone marrow mononuclear cells (BMMCs) from five patients (MM1-5) were incubated with VIS832 (0–10 µg/ ml). Cells were subsequently co-stained with PE-BCMA antibody (to identify patient myeloma cells) and APC-anti-human IgG1 followed by FC analysis on PE + APC + cells. All data is presented as means (MFIs) ± SEMs (error bars) of triplicates at each VIS832 concentration. c Relative antibody- dependent cellular cytotoxicity (ADCC) of VIS832 vs. BB4 possessing an isotype matched human IgG1 were measured using an ADCC reporter cell assay where Jurkat cells stably transfected with human FcRIIIa as surrogate for effector cells were co-cultured with target U266 MM cells. ADCC is reported as a fold induction over no antibody background control (effector + target cells only).

patient MM cells was independently measured by results also replicated ADCC data in another RRMM quantitative flow cytometry analysis after staining with patient cohort (n = 6, MM13-18) treated with the par- an mAb against BCMA, another PC marker41,42. ental 2810 mAb, including 90% killing in 3 of 6 MM VIS832-induced autologous cytolysis of patient MM patient samples (Fig. 2e, right). EC50 values ranged from cells was determined by the elimination of the viable 8.58 to 86.04 and 17.28–179.3 ng/ml for VIS832 and BCMA+ cell population relative to control (Supple- 2810 mAb, respectively. mentary Fig. S4). VIS832 depleted patient MM cells in a Paired CD138-positive and CD138-negative patient cell dose-dependent manner (Fig. 2e, left), inducing >60% populations were next separated and pre-labeled with maximal lysis of autologous patient MM cells, with calcein-AM, followed by co-culture with autologous >90% killing in 4 out of 7 MM patient samples. These PBMCs (n = 3) in the presence of VIS832 (0–1 µg/ml).

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 5 of 13

H929 NT MM1S abH929 CD138 KO c MM1R JJN3 NT H929 8000 20 100 100 MFI EC50 JJN3 CD138 KO **** H929(R) **** ANBL6 80

E 80 6000 15 ANBL6-BR

C **** 50 is 60 60 s n 4000 10 ysis

g/ **** m %ly 40 %l **** 40

l ** CD138 MFI 2000 5 20 ** 20 * 0 0 0 0 S 9 2 ) ) 6 R 1R 8 R L -5 -4 -3 -2 -1 0 1 2 (R B -5 -4 -3 -2 -1 0 1 H92 U266 8226 LP 9 µ MM1MM OPM I JJN3O log VIS832 [µ g/ml] log VIS832 [ g/ml] M AN BL6-B MM1S(H92 N RPM A d e 120 100 EC (ng/ml) MM6 MM13 50 MM7 MM14 MM1S 7.37 100 MM8 80 MM9 MM15 MM1S+BMSC 7.57 80 s MM10 MM16 60 MM1S+OC 7.94 MM11 MM17 ysi H929 2.36 l 60 MM12 MM18

40 H929+BMSC 2.81 % %lysis 40 H929+OC 2.83 20 20

0 0 -5 -4 -3 -2 -1 0 1 2 -5 -4 -3 -2 -1 0 1 -5 -4 -3 -2 -1 0 1 logVIS832[µg/ml] log VIS832 [µ g/ml] log 2810 [µ g/ml]

f 10 CD138-positive Patients EC50 (ng/ml) Patients EC50 (ng/ml) CD138-negative # of VIS832 # of 2810

DCC) 8 **** MM6 12.21 MM13 103 **** MM7 8.58 MM14 54.64 n(A 6

tio MM8 13.35 MM15 48.06 4 MM9 20.17 MM16 65.23 duc ** MM10 25.15 MM17 179.3 2

ld in MM11 27.66 MM18 17.28 o

F 0 MM12 86.04 -5 -4 -3 -2 -1 0 1 log VIS832 [µ g/ml]

Fig. 2 VIS832 specifically induced ADCC against MM cell lines and patient MM cells sensitive or resistant to current therapies in the presence or absence of BMSCs or OCs. a Human NK effector cells from 3 healthy donors were incubated with indicated MM cell lines (n = 12) (E:T = 10:1) in a serial concentration of VIS832 (0–10 μg/ml). VIS832-induced ADCC to lyse MM cells was determined using the calcein-AM release

assay. Derived EC50 values (triangle, right label) were plotted in parallel with CD138 MFI (bar, left label). b VIS832-induced ADCC was done using target cells including parental (target control, NT) and CD138-knockout (KO) H929 and JJN3 cell lines. Parental H929 cells express higher CD138 than JJN3 cells. c Additional NK effector cells (n = 3) were co-incubated with target cells including paired MM1S and dex-resistant MM1R, H929, and len/ pom-resistant H929(R), as well as ANBL6 and btz-resistant ANBL6-BR cells. d VIS832-induced ADCC against target MM1S or H929 MM cell lines was performed in the presence or absence of BMSCs or OCs. All data are presented as means (% lysis) ± SEMs (error bars) of three independent experiments using NK cells from three healthy donors with each performed in triplicate at each concentration. e BMMCs freshly derived from BM aspirates from patients (MM6-12, left; MM13-18, right) were treated with VIS832 (left) or mAb 2810 (right, a mouse-human chimeric antibody of VIS832) (0–10 μg/ml) for 24 h, followed by FC analysis to determine % lysis of patient MM cells. The % lysis indicated % depletion of viable BCMA+ patient PCs in BMMCs, and data was normalized to minus isotype control human IgG1 (set as 100% BCMA+ cells). BMMCs included target patient PCs, immune effector cells, and other non-MM BM accessory cells. f Calcein-AM-based ADCC assays were done using target cells including paired CD138-positive and CD138-negative cells separated from BMMC samples from three RRMM patients. Each target cell fraction was co-cultured with PBMCs from the same individual in the presence of VIS832 (0–1 µg/ml). Each experiment was performed in triplicate at each concentration and results are shown as means (% lysis) ± SEMs (error bars). *P < 0.05, **P < 0.01, ****P < 0.001.

Specific ADCC activity was both VIS832-dependent and VIS832 showed increased ADCC potency than dara and required CD138 + target patient MM cells (Fig. 2f). The directs significant target-specific killing of dara-resistant correlation of MM cell binding and ADCC using VIS832 MM cells without depleting NK cells and 2810 mAb, therefore, corroborated effective CD138 We next evaluated relative potencies of VIS832 vs. target engagement, leading to productive and specific daratumumab (dara), to induce effector cell-dependent autologous patient MM cell killing. killing of target MM cell lines using luminescence-based

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 6 of 13

100 MM1S-luc a b 80 dara VIS832 MM1S(R)-luc VIS832 80 *** H929(R)-luc *** 60 **** 60 MM1S-luc sis MM1S(R)-luc dara 40

%ly 40 H929(R)-luc ** %lysis 20 * 20 MM1S

0 0 -5 -4 -3 -2 -1 0 1 2 0 0.001 0.01 0.1 1 10 log mAb [µg/ml] mAb [µg/ml] c d 100 100 H929 NT RPMI8226 VIS832 80 H929 CD38 KO-1 VIS832 80 RPMI8226-DR H929 CD38 KO-2 RPMI8226 s dara 60 60 RPMI8226-DR si

H929 NT y H929 CD38KO-1 %lysis 40 dara %l 40 H929 CD38KO-2 20 20

0 0 -5 -4 -3 -2 -1 0 1 2 -5 -4 -3 -2 -1 0 1 2 log mAb [µg/ml] log mAb [µg/ml] e f 60 **** VIS832 [µg/ml] 15 **** mAb [µg/ml]

**** ells 0 **** 0 **** 0.001 Kc 0.001 40 **** 0.01 10

+N *** 0.01 *** 0.1 V 0.1 in

1 x 20 5 1

ne 10 n % CD107a+ NK cells 0 %A 0 NK NK + MM1S dara VIS832

Fig. 3 VIS832, unlike daratumumab, induced comparative ADCC against MM cells resistant to Dara- and IMiDs, without adverse effects on NK effector cells. a, b VIS832 showed a higher ADCC potency than dara, as measured by two orthogonal assays including BLI measurement (a) and quantitative FC analysis (b). a Target MM1S, MM1S(R), and H929(R) cells expressing firefly-luciferase (luc) were co-cultured with NK cells from three healthy donors in the presence of VIS832 or daratumumab (dara) (0–10 µg/ml). Luciferase activity in the cell culture media was quantified using the Dual-Luciferase® Reporter Assay Kit (Promega % cell lysis was normalized to 100% lysis of target cells set following mild detergent solubilization using assay compatible lysis buffer). b MM1S targeted cells were treated with VIS832 or dara in the presence of NK cells (n = 3 healthy donors) for 24 h followed by quantitative FC analysis to determine the % lysis (% of depleting viable BCMA+ cells). c, d VIS832- vs. dara-induced ADCC were determined in the 4-h calcein-AM release assay in co-cultures of indicated paired MM cell lines and NK cells from three healthy donors. Target MM cells included parental and two CD38 CRISPR-cas9 knockout (CD38KO-1, CD38KO-2) H929 (c) or dara-resistant RPMI8226 (PMIR8226-DR) (d) MM cells. Dara-resistant RPMI8226 cells were generated through extended exposure to a lower dose of dara with NK cells in the long-term culture. e NK cell activation was measured by double marker FC analysis (CD56 + and CD107a + ). CD107a cell surface expression was used as a surrogate for quantifying cellular degranulation of NK cells. NK cell activation was both VIS832- and target MM cell-dependent. f Effects of targeting CD38 vs. CD138 on NK cell viability. Fresh NK cells purified from healthy donors (n = 3) were incubated with serially diluted dara or VIS832 (0–10 µg/ml) for 24 h. NK cell apoptosis was assessed by Annexin V/PI staining followed by FC analysis. All assays included three independent experiments using NK cells from three healthy donors. Data are presented as means (% lysis) ± SDs (standard deviations, error bars). *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001.

(Fig. 3a) and flow cytometry analysis (Fig. 3b). Dara tar- vs. CD38 levels in all MM cell lines (P < 0.0001, Supple- geting CD38 has recently been approved to treat newly mentary Fig. S5A). Importantly, patient MM cells also diagnosed MM43. VIS832, more potently than dara, expressed significantly higher CD138 vs. CD38 (P < increased maximal lysis by ~3-fold of all target MM cell 0.0001), and CD138 levels were further increased with lines, regardless of resistance to both IMiDs. Higher progression of disease (Supplementary Fig. S5A, B). MM1S cell lysis induced by VIS832 vs. dara was also We next asked whether VIS832 remained active against confirmed in BCMA-based flow cytometry analysis (P < MM cells resistant to dara. CD38 knock-out (CD38 KO) 0.05). These results confirmed significantly higher CD138 through CRISPR/CAS9 gene editing was done in H929

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 7 of 13

cells to derive 2 CD38 KO transfectants (Supplementary Furthermore, VIS832-induced ADCC against target Fig. S5C). VIS832 induced equivalent ADCC to lyse MM cell lines H929 and RPMI8226 was also synergisti- CD38 KO transfectants and the parental control H929 cally enhanced when VIS832 combined with btz, at even (H929 NT) cells (Fig. 3c). CD138 levels were unchanged lower E:T ratio of 1:1, with CI values of <1 (Fig. 4c, d and in CD38 KO and parental control H929 cells. The dose Supplementary Fig. S6C, D). response curves for three cell lines overlapped, with similar % maximal lysis and EC50 values of VIS832. VIS832 induced antibody-dependent cellular phagocytosis Compared to dara, VIS832 induced significantly higher (ADCP) against MM cells sensitive and resistant to dex, ADCC, with increased maximal lysis (~2.7-fold higher) of IMiDs, or dara the parental H929 target cells. In contrast, dara failed to ADCP activity was next measured by dual colored flow induce ADCC against CD38 KO H929 transfectants. We cytometry analysis using in vitro culture-differentiated also developed dara-resistant RPMI8226 (RPMI8226-DR) macrophages co-incubated with target MM1R cells cells via extended exposure of dara in ex vivo NK-MM resistant to dex. Percentage of phagocytosis as the co-cultures. Unlike RPMI8226 cells, dara did not induce double-positive (DP) fraction was quantitated in the ADCC against RPMI8226-DR cells, confirming resistance presence of VIS832, isotype control Ab, or no Ab control, to dara (Fig. 3d). VIS832 still induced lysis of RPMI8226- at various macrophage effector to MM cell ratios. Dara DR to a similar extent as its parental RPMI8226 cells. served as a positive control for its reported ADCP Again, VIS832, more effectively than dara, induced activity45. Phagocytosis was dependent on both VIS832 greater % maximal lysis of the parental RPMI8226 cells. (e.g., in comparison to the isotype control antibody) and The effect of VIS832 on MM-specific NK cell activation the E:T ratio (P <0.001,Fig. 5a). The ADCP activity of was next assayed by quantitative flow cytometry analysis VIS832 was significantly higher than that observed for for % CD107a surface expression on CD56 + CD3- cells in dara (P < 0.05, Fig. 5a, b). Furthermore, VIS832-induced the NK-MM co-culture. VIS832 induced CD107a degra- ADCP was comparable in MM cell lines sensitive or nulation in NK cells in a MM cell- and concentration- resistant to IMiDs and dara (Fig. 5b). While dara failed to dependent manners (Fig. 3e). induce ADCP against RPMI822-DR cells, VIS832 still NK cells express CD38 at the highest level among other induced significant ADCP (P < 0.0001). normal hematological lineage cells. In fact, dara depletes patient NK cells within the first week after treatment, VIS832 efficacy in a murine model of disseminated human affecting its efficacy to target MM cells and thereby limiting MM durability of responses36,44. Using annexin V staining fol- In vivo VIS832 efficacy was assessed using a murine lowed by flow cytometry analysis to measure apoptotic cell model xenografted with the dissemination of MM1S-luc fraction, dara, but not VIS832, induced NK cell death in a cells by i.v. injection into CB-17 SCID mice, which pre- concentration-dependent manner after 1-day treatment serve aspects of innate immunity relevant to antibody (Fig. 3f). Thus, unlike dara, VIS832 even selectively induced mechanisms of action29,31,32, including NK cell- and MM cell lysis via NK activation, without depleting NK cells. macrophage-mediated cytotoxicity of antibody-targeted CD138-expressing MM cells. Dosing intervals and route Enhancement of targeted cellular killing activity of VIS832 of administration (n = 9 mice per group) were as descri- in combination with lenalidomide or bortezomib bed in methods and in Fig. 6a. Whole BLI of MM1S-luc IMiDs (len and pom), which upregulate ADCC ability cells was used to non-invasively quantify site-specific of therapeutic mAbs29,32,34, are routinely used in com- disease burden (Supplementary Fig. S7). bination with dex in mAb-based immunotherapies in In the vehicle control (VC, PBS) group, the median time MM. Using calcein-AM release assays in MM-NK cell co- to evaluation size was ~28 days, and the median tumor cultures at lower E:T ratio of 4:1, we next tested effects of burden doubling time was 2 days (range of 1.5–2.2 days), len on VIS832-induced cytotoxicity against MM cells. consistent with tumor implantation in all animals and an When VIS832 and len were added concomitantly, aggressive progression of tumor growth. The median synergistic MM cell lysis was seen, with combination survival of animals in the VC group was 30 days, and all index (CI) values of <1 at all concentrations of both drugs animals were deceased by day 51 (Table 1 and Supple- (Fig. 4a and Supplementary Fig. S6A). Pretreatment with mentary Tables S3 and 4). In contrast, VIS832 mono- len alone for 3 days did not induce significant killing of therapy at 4 mg/kg effectively reduced tumor burden with MM1S and MM1R target cell lines (<10%) (Fig. 4b). Also, a prolonged time to evaluation (Supplementary Table S4, VIS832 significantly induced higher ADCC against median survival of >60 days), and reduction of dis- MM1S and MM1R target cells pretreated with len when seminated tumor growth (Fig. 6b and Supplementary Fig. compared to control medium groups (P <0.05),withCI S7). Five of 9 animals (56%) remained on study for its values <1 (Supplementary Fig. S6B). duration (day 73). Significant survival benefit of VIS832

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 8 of 13

acMM1S MM1R RPMI8226 H929

80 **** ctrl 80 ctrl **** *** **** **** **** *** *** µ ** **** 60 *** ** **** ** len 1 M 60 *** **** *** btz 1 nM * **** *** ** ** **** **** *** ** len 5µM **** **** **** btz 2 nM 40 **** ** 40 ** **** *** ** µ * *** *

% lysis len 10 M *** ** btz 4 nM ** % lysis * ** * 20 * 20 * ** * 0 0 0 0.001 0.01 0.1 1 0 0.001 0.01 0.1 1 0 0.001 0.01 0.1 1 00.0010.010.11 VIS832 [µg/ml] VIS832 [µg/ml] VIS832 [µg/ml] VIS832 [µg/ml] 2 2 2 2

CI 1 CI 1 CI 1 CI 1

0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 Fa Fa Fa Fa

bdTarget cell pre-treatment with len Target cell pre-treatment with btz MM1S MM1R RPMI8226 H929 **** **** 80 **** **** ctrl **** **** **** **** 100 100 **** ** **** **** ctrl **** **** 80 btz 1 nM * **** **** µ ****60 **** **** **** 80 **** **** len 1 M **** **** s **** 60 * *** btz 2 nM **** **** µ **** ** 60 *** **** len 5 M **** **** 40 ** * **** 40 **** * * * btz 4 nM 40 **** len 10µM %lysis ** * **** %lysi * * 20 20 20 0 0 0 0 0.001 0.01 0.1 1 0 0.001 0.01 0.1 1 0 0.001 0.01 0.1 1 0 0.001 0.01 0.1 1 VIS832 [µg/ml] VIS832 [µg/ml] VIS832 [µg/ml] VIS832 [µg/ml] 2 2 2 2

CI 1 CI 1 CI 1 CI 1

0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 Fa Fa Fa Fa

Fig. 4 Lenalidomide and bortezomib enhanced VIS832-induced MM cell lysis in a synergistic manner. Indicated MM cell lines (a, b MM1S, MM1R; c, d RPMI8226, H929) were pretreated without (a, c) or with lenalidomide (len, b) or bortezomib (btz, d), and then drugs were washed out. Target MM cells were co-incubated with NK cells (E:T = 4:1) in the presence of varying concentrations of VIS832 or control media (ctrl) in the 2-h calcein-AM release assay. a, c VIS832 was added together with Len (a) or btz (c) in the co-cultures of target MM cells with NK cells. Percentages of MM cell lysis were used to calculate the combination index (CI) (Fa-CI plot) shown below each graph. CI < 1 indicated a synergistic interaction. Fa, fraction affected. All assays included three independent experiments using NK cells from 3 healthy donors. Data are presented as means (% lysis) ± SDs (error bars). *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.001.

was noted at day 53 (P < 0.0001, Fig. 6c), 3 weeks fol- combination vs. VC; P < 0.02 for combination vs. VIS832; lowing treatment discontinuation. VIS832 treatment also P < 0.0001 for combination vs. btz). This efficacy was led to disease regression in a subset of mice evaluated up sustained even after discontinuation of treatment on Day to day 53 (the last day of dosing). 53, 3 weeks prior to study termination. The apparent lack In the combination group, a sub-optimal dose of btz was of any discernible tumor burden suggested an achievement chosen (1 mg/kg) that achieved a tumor growth delay of of minimal residual disease in animals receiving combi- ~10 days and a 40% increase in lifespan, but no animals nation therapy. Loss of body weight during the first week exhibited either partial or complete regression or survived of treatment and mean body weight change was minimal (median survival of 42 days) to the end of study at Day 73 (<5%). In fact, animals treated with 4 mg/kg VIS832 (Fig. 6 and Supplementary Table S4). VIS832 alone sig- experienced an upward trend in mean body weight. In nificantly extended host survival when compared with btz contrast, treatment-related toxicity was observed in the (P = 0.02). Combined VIS832 and btz treatments sig- two groups initially receiving btz at the 2 mg/kg dose level, nificantly enhanced efficacy, evidenced by complete tumor but reduction in dose from 2 mg/kg to 1 mg/kg led to regression and 100% survival to end of study (Fig. 6c). recovery of body weight, particularly in mice receiving Clear survival advantages were confirmed (P < 0.0001 for 1 mg/kg of btz and 4 mg/kg VIS832 combination.

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 9 of 13

a VIS832 dara **** **** DP MM **** **** 40 ** * VIS832 **** ** **** dara **** **** 30 **** isotype MP no Ab **** no Ab isotype 20 ns ns ns 10 % phagocytosis Cell trace Far red Cell

0 1:1 2:1 4:1 MP:MM1R (E:T) ratio CFSE (green)

25 **** b *** * VIS832 *** **** **** **** **** dara 20 **** ns isotype **** no Ab 15 **** ns

10 ns hagocytosis ns %p 5 ns

0 target cell: MM1S MM1S(R) RPMI8226 RPMI8226-DR MP:MM (E:T=1:1)

Fig. 5 VIS832 induced antibody-dependent cellular phagocytosis (ADCP) against MM cells. a Target MM1R (MM) cells were pre-labeled with a fluorescence cell membrane dye Cell Trace™ Far Red and co-cultured for 4 h with macrophage (MP) effector cells pre-labeled with CFSE-FITC (green) in the presence of VIS832, daratumumab (dara), or isotype control (4.5 µg/ml each) at various E:T ratios of 1:1 to 4:1. Cells were fixed with 2% paraformaldehyde, followed by FC analysis to determine double-positive (DP) fluorescence fraction. Percent phagocytosis was determined as the number of DP cells (upper right panel) divided by the total number of target MM cells. Shown are representative dot plots at an E:T ratio of 2:1 (left) and summarized % phagocytosis at 3 E:T ratios (right). b Paired IMiDs- or dara-resistant target MM cell lines (n = 4) were co-incubated with MP cells (E:T = 1:1). All experiments were done in triplicates at each condition using MP cells derived from three healthy donors. ns, not significant; *P < 0.05, **P < 0.01, ****P < 0.001.

Thus, VIS832 induced significant in vivo anti-MM modeling of ADCC for well characterized antibodies such activity as monotherapy, and further completely eradi- as with potent Fc-mediated ADCC indicates the cated MM in mice treated in combination with btz. critical importance of epitope location and topological proximity to the cell surface as a key determinant of Discussion immune synapse formation46. The epitope recognized by Given its enhanced and prevalent presentation on the VIS832 includes a membrane proximal region of CD138 MM cell surface, CD138 is a compelling target for whose targeted binding is predicted to confer productive immunotherapy in MM. Many of the previously char- immune synapse formation for improved Fc effector acterized anti-CD138 mAbs, such as BB4 in indatuximab dependent, immune cell-mediated cytotoxicity. As such, ravtansine, target the integrin binding domain (IBD), an indatuximab ravtansine, which targets the IBD of CD138, immunodominant region of CD138, which is distally is predicted to exhibit poor immune cell-based killing25 located from the cell membrane. VIS832 evaluation has when compared with VIS832. Indeed, VIS832 induces been sufficiently characterized with respect to its physi- superior MM cell lysis than the BB4 possessing an isotype cochemical and biophysical attributes, as measured by matched human IgG1. several different drug “developability” assays and metrics VIS832-dependent NK cell-mediated cytotoxicity, while indicating favorable pharmaceutical properties. It engages target dependent, did not precisely correlate with a differentiated epitope on CD138 comprising two, non- CD138 surface expression levels on target cells. This sug- contiguous regions that are required for optimal target gests a sufficiency of CD138 target abundance (threshold) to binding. Importantly, when measured in the same assay, drive a potent antibody-mediated biological response. Other we showed stronger and more intense binding of VIS832 aspects include VIS832 target epitope engagement, target vs. BB4 to all MM cell lines, including those with relatively binding avidity, and other yet identified mechanisms of lower CD138 density. In parallel, further structural action. Significantly, our data indicate that the differentiated

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 10 of 13

a 0 1453 73 Days Treatment duration Tumor Treatment Observation % Increased lifespan inoculation VC (PBS) end termination i.v. VIS832 btz VIS832+btz

11 b 10 VC (PBS) 1010 VIS832 c 109 btz Median lifespan (days)

VIS832+btz r disease or treatment-related causes. SEM 108 ± 7

10 f BLI

Photons/sec 106

105 %TFS 14 28 42 56 70

Days post tumor inoculation e

c 100 VC (PBS) %CR l VIS832 C

iva 75 /

btz T rv Δ u 50 VIS832+btz S

% 25

0

14 28 42 56 70 (day 36) Days post tumor inoculation

Fig. 6 VIS832 treatment efficacy as monotherapy and in combination with btz in a murine model of disseminated human MM. a In vivo work plan. CB-17 SCID mice injected with MM1S-luc cells by i.v. route at day 0 were randomized on day 14 as the first day of dosing twice weekly with vehicle control (VC, PBS), VIS832 (4 mg/ kg), btz (1 mg/kg), or combination of VIS832 and btz (n = 9 mice per group). Tumor burden was assessed by whole body BLI and quantified

by IVIS imaging. Treatment was discontinued at day 53 (dosing phase), cacy Survival followed by a three-week washout period with end of study at day 73. fi Ef Tumor growth delay (days) Median % b Group efficacy data. Disseminated tumor burden was reported as BLI ± SEM at each time point. Tumor growth was significantly inhibited starting at day 36 in VIS832 or btz vs. VC groups; the

combination-treated vs. VC groups. (at Day 36, VIS832 vs. VC, P < cacy study. 0.0001; btz vs. VC, P = 0.0001; VIS832 + btz vs. VC, P < 0.0001; VIS832 vs. fi VIS832 + btz, P < 0.0001; btz vs. VIS832 + btz, P < 0.0001; VIS832 vs. btz, % Treatment- related deaths P = 0.0004, by Tukey’s multiple comparison test). c Kaplan–Meier 0.0 13.9 1.1 22.2 >60 >100

survival curves. The percentage of survival was defined as mice d). 6 surviving to a pre-determined euthanasia criterion related to disease- related morbidity such as weight loss >20%, severely impaired central nervous system function, severely impaired movement, or loss of righting reflexes. VIS832 significantly increased the median overall d 4.3 13.6 ------30 -- 13.69.8 11.1 0.0 10.6 >45 8.2 0.0 0 100 >90 42 >60 >100 40 Max % weight change – – survival of animals, and VIS832 with btz completely eradicated MM – – tumors, with all mice tumor-free (VC, 30d; VIS832 > 60d; btz, 42d; VIS832 + btz, >73d) (VIS832 vs. VC, P < 0.0001; btz vs. VC, P = 0.0004; VIS832 + btz vs. VC, P < 0.0001; VIS832 vs. VIS832 + btz, P < 0.02; btz vs. a VIS832 + btz, P < 0.0001; VIS832 vs. btz, P = 0.02). schedule 2QW 2QW epitope of VIS832 in comparison with BB4 within inda- tuximab ravtansine confers a more productive target bind- rst treatment in the study (not the day of tumor implant) for each animal. It captures the day of death for all animals that either die or are euthanized fo fi b b

ing and augmented Fc effector function, leading to 1.0 improved immune synapse formation, and ultimately to its more potent MM targeting and superior cellular cytotoxicity activity, including ADCC and ADCP. +

Besides ADCC that presumptively represents a primary btz 4.0 mechanism for the biological potency of VIS832, VIS832 Dose level of bortezomib adjusted down from 2 mg/kg (2QW) toNegative 1 change mg/kg in on weight day observed 18 during due initial to dosing toxicity only including (see the Fig. death of one animal in treatment group (Gp). Complete regression (CR) (see Supplementary Tables S3 and 4). Vehicle (PBS) administered by oral gavage. Measured from the day of Tumor-free survivor (TFS) (see Supplementary Tables S3 and 4). Animals euthanized for sampling or other reasons unrelated to disease or therapy (e.g., end of study) are excluded from this calculation. Table 1 Group (nine miceGp each group) summary for Treatment in vivo ef Dose (mg/kg) Dose a b c d e f 2 VIS832 4.0 2QW 4 VIS832 3 btz 1.0 delivered additional Fc-mediated ADCP to eliminate MM 1 Vehicle 0.2 ml/20 g 5QW

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 11 of 13

cells. ADCP activity of therapeutic mAbs including dara Combined treatment of tumor-bearing mice with both and GSK2857916 () is an important VIS832 and dose-adjusted btz resulted in complete tumor mechanism for their in vivo potency29,35,41. Significantly, regression and 100% survival of all animals to the end of VIS832 exhibited increased ADCP activity than dara in study at day 73. Complete tumor elimination in all mice multiple MM cell lines, regardless of resistance to dex and was sustained even after discontinuation of treatment on IMiDs. VIS832 induced comparable ADCP to deplete Day 53, 3 weeks prior to study termination. The apparent RPMI8226-DR cells resistant to dara, indicating ADCP as lack of any discernible tumor burden suggests an achieve- a crucial mechanism of action of VIS832 to overcome ment of minimal residual disease in animals receiving the dara resistance. This supplementary cytotoxic function of two therapies at sub-optimal doses in combination. More- VIS832 confirms its multi-faceted immune-mediated anti- over, the superior efficacyofcombinedVIS832andbtz MM activity, which could aid in overcoming multi-drug when compared to either agent alone suggests synergistic resistance and prolonging response. in vivo activity, consistent with synergistic in vitro cyto- As in the case for dara, , , and toxicity of this combination (CI < 1). Recently, CD138 – belantamab mafodotin29,34,47 49, the activating effect of len downregulation sensitizes MM cells to btz treatment in on immune effector cells also augments VIS832-induced another animal model50. The augmentation of activity of cytotoxicity against MM cells. Such enhanced efficacy of low-dose btz by VIS832 shown here suggests enhanced combination vs. monotherapy likewise would be antici- efficacy and tolerability at lower doses in the clinic. pated with other IMiDs including pom. Higher maximal Taken together, the significant in vivo efficacy of VIS832 killing of MM cells of VIS832 than dara also correlated in the MM1S xenograft murine model, coupled with its with higher CD138 than CD38 target expression. mechanisms of action and in vitro MM cytotoxicity, both VIS832 still potently induced ADCC against MM cells confirm CD138 as a promising MM target and provide resistant to dara, either due to the loss of CD38, or the basis for clinical development of VIS832 as a poten- acquired drug resistance through long-term culture selec- tially effective mAb-based immunotherapy. Given the tion with dara in ex vivo NK-MM co-cultures. Moreover, current therapeutic landscape and clinical approaches to dara, but not VIS832, induced apoptosis in NK cells. Thus, MM treatment, the use of immune targeted therapies VIS832 would not deplete NK cells, thereby avoiding any including VIS832 will likely be best suited to augment negative impact on its efficacy and therapeutic window. standard-of-care through an appropriate use of drug VIS832, like its predecessor mAb 2810, induced robust combinations, and/or as additional lines of therapies in immune cell-mediated cellular killing of patients MM MM. Such an approach will be of particular therapeutic cells resistant to btz, IMiDs, and/or dara. Its selective value in the treatment of patients with relapsed disease cytotoxicity against autologous patient MM cells was and/or refractory to prior treatment (e.g., dara). Indeed, confirmed using whole BMMCs, as well as purified our current data indicate that VIS832 can overcome CD138 + cells freshly harvested from patients. Since resistance to dex, btz, and IMiDs. Once its efficacy is various BM accessory cells confer immunosuppression, established in RRMM, its favorable therapeutic index these data suggest that VIS832 would be active in the should allow for moving rapidly to earlier stages of dis- patient BM milieu. Its selective ADCC against autologous ease, newly diagnosed MM and even SMM. patient MM cells, but not CD138-negative BM cells, further supporting a favorable therapeutic index. Acknowledgements fi We acknowledge the technical assistance from the flow cytometry facility at The in vivo ef cacy of VIS832 was convincingly Dana-Farber Cancer Institute (DFCI). We especially thank Drs. Jiye Liu and demonstrated in an aggressive disseminated MM1S Eugenio Morelli at DFCI, Dr. Ronald Taylor at University of Virginia for helpful xenograft model of human MM in CB-17 SCID mice, technical assistance and discussion for in vitro experiments. We thank Hamid Tissire and Ketan Deotale (Visterra, Inc.) for valuable technical assistance in the both as monotherapy and in combination with btz. When biochemical characterization of anti-CD138 antibodies. We thank all other lab used alone, VIS832 reduced median tumor burden and members as well as the clinical research coordinators at the Jerome Lipper improved overall survival. This efficacy was sustained Multiple Myeloma Center and the LeBow Institute for Myeloma Therapeutics of the DFCI for all support and help in providing primary tumor specimens for after discontinuation of treatment 3 weeks prior to study this study. This work was supported in part by grants from the National termination. An evaluation of VIS832 PK/PD relation- Institutes of Health Specialized Programs of Research Excellence (SPORE) P50 ships in a prior dose range finding study in the same CA100707, P01CA155258, RO1 CA 207237, R01 CA050947. This work was fi supported in part by Dr. Miriam and Sheldon G. Adelson Medical Research in vivo model indicated suf cient exposure of VIS832 at Foundation and the Riney Family Multiple Myeloma Initiative. Dr. Kenneth this dose level (4 mg/kg) to achieve a biological response C. Anderson is an American Cancer Society Clinical Research Professor. (data not shown). Even lower dose levels, while not evaluated, are predicted to be sufficient to achieve such a Author details 1 response. Importantly, administration of VIS832 was well Jerome Lipper Multiple Myeloma Center, LeBow Institute for Myeloma Therapeutics, Dana-Farber Cancer Institute, Harvard Medical School, Boston, tolerated, with no overt treatment-related morbidities or MA, USA. 2State Key Laboratory of Experimental Hematology, National Clinical toxicities, especially seen in btz-treated group. Research Center for Hematological Disorders, Institute of Hematology and

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 12 of 13

Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking 11. Sanderson,R.D.&Yang,Y.Syndecan-1:adynamicregulatorofthemyeloma Union Medical College, Tianjin 300020, China. 3Visterra Inc., Waltham, MA, USA. microenvironment. Clin.Exp.Metastasis25,149–159 (2008). 4Division of Hematology & Oncology, Department of Internal Medicine, 12. Yang, Y. et al. The syndecan-1 heparan sulfate proteoglycan is a viable target Kaohsiung Medical University Hospital, Kaohsiung Medical University, for myeloma therapy. Blood 110, 2041–2048 (2007). Kaohsiung 80708, Taiwan. 5Faculty of Medicine, College of Medicine, 13. Seidel, C. et al. High levels of soluble syndecan-1 in myeloma-derived bone Kaohsiung Medical University, Kaohsiung 80708, Taiwan. 6School of marrow: modulation of hepatocyte growth factor activity. Blood 96, Biotechnology, Tianjin University of Science and Technology, Key Lab of 3139–3146 (2000). Industrial Fermentation Microbiology of the Ministry of Education, State Key 14. Aref, S., Goda, T. & El-Sherbiny, M. Syndecan-1 in multiple myeloma: rela- Laboratory of Food Nutrition and Safety, Tianjin 300457, China. 7Hematology tionship to conventional prognostic factors. Hematology 8,221–228 (2003). Department, Yueyang Hospital of Integrated Traditional Chinese and Western 15. Andersen, N. F. et al. Syndecan-1 and angiogenic cytokines in multiple Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, myeloma: correlation with bone marrow angiogenesis and survival. Br. J. China Haematol. 128,210–217 (2005). 16. Seidel, C. et al. Serum syndecan-1: a new independent prognostic marker in Author contributions multiple myeloma. Blood 95,388–392 (2000). Conception and design: J.R.M., K.C.A., Y.-T.T. Development of methodology: T.Y., 17. Rajkumar,S.V.,Landgren,O.&Mateos,M.V.Smolderingmultiplemyeloma. B.C., L.L., B.R., L.X., Y.T.T., Z.S., J.R.M. Acquisition of data: T.Y., L.L., B.C., B.R., L.X., S.-F. Blood 125,3069–3075 (2015). C., Y.L., K.W., P.A.H., H.C., W.Y., Y.X., G.A., L.Q., A.W., K.V., H.A. Analysis and 18. Lakshman, A. et al. Risk stratification of smoldering multiple myeloma incor- interpretation of data (statistical analysis, biostatistics analysis): T.Y., B.C., L.L., L. porating revised IMWG diagnostic criteria. Blood Cancer J. 8,59(2018). X., B.R., S.-F.C., Y.L., K.W., G.B., J.R.M., Y.-T.T. Reagents provided: Z.S., J.R.M. Animal 19. Bolli, N. et al. Genomic patterns of progression in smoldering multiple mye- work plan and data requisition and analysis: Z.S., J.R.M. Provided and managed loma. Nat. Commun. 9, 3363 (2018). patient samples: N.M., K.C.A. Writing, review, and/or revision of the manuscript: 20. Detweiler-Short, K. et al. Long-term results of single-agent thalidomide as T.Y., J.R.M., K.C.A., Y.-T.T. Study supervision: K.C.A., Y.-T.T. initial therapy for asymptomatic (smoldering or indolent) myeloma. Am. J. Hematol. 85,737–740 (2010). Conflict of interest 21. Sundararajan, S., Kumar, A., Korde, N.&Agarwal,A.Smolderingmultiple B.C., B.R., A.W., K.V., H.A, G.B., Z.S., and J.R.M. are employees of Visterra Inc. N.M. myeloma: emerging concepts and therapeutics. Curr. Hematol. Malig. Rep. 11, – serves on advisory boards to Millennium-Takeda, Celgene, and Novartis. K.C.A. 102 110 (2016). fi serves on advisory boards Celgene, Millennium-Takeda, Bristol-Myers Squibb, 22. Bustoros, M. et al. Genomic pro ling of smoldering multiple myeloma iden- fi Gilead Sciences, Janssen, and Sanofi-Aventis and is a Scientific founder of ti es patients at a high risk of disease progression. J. Clin. Oncol. 0, OncoPep and C4 Therapeutics. All other authors declare no potential conflicts JCO.20.00437 (2020). of interest. 23.Landgren,C.O.etal.Daratumumabmonotherapyforpatientswith intermediate-risk or high-risk smoldering multiple myeloma: a randomized, – Publisher’s note open-label, multicenter, phase 2 study (CENTAURUS). Leukemia 34,1840 1852 Springer Nature remains neutral with regard to jurisdictional claims in (2020). published maps and institutional affiliations. 24. Lonial, S. et al. Randomized trial of lenalidomide versus observation in smol- dering multiple myeloma. J. Clin. Oncol. 38, 1126–1137 (2020). Supplementary Information accompanies this paper at (https://doi.org/ 25. Ikeda, H. et al. The monoclonal antibody nBT062 conjugated to cytotoxic 10.1038/s41408-020-00378-z). Maytansinoids has selective cytotoxicity against CD138-positive multiple myeloma cells in vitro and in vivo. Clin. Cancer Res. 15,4028–4037 (2009). 26. Jagannath, S. et al. Indatuximab ravtansine (BT062) monotherapy in patients Received: 3 August 2020 Revised: 16 September 2020 Accepted: 2 October with relapsed and/or refractory multiple myeloma. Clin. Lymphoma Myeloma 2020 Leuk. 19,372–380 (2019). 27. Sun, C. et al. Safety and efficacy of targeting CD138 with a chimeric antigen receptor for the treatment of multiple myeloma. Oncotarget 10,2369–2383 (2019). 28. Xing, L. et al. A novel BCMA PBD-ADC with ATM/ATR/WEE1 inhibitors or References bortezomib induce synergistic lethality in multiple myeloma. Leukemia 34, 1. Kumar, S. K. et al. Multiple myeloma. NatRev.Dis.Primers3, 17046 (2017). 2150–2162 (2020). ‐ fi 2. Rajkumar,S.V.Multiplemyeloma:2018updateondiagnosis,riskstrati cation, 29. Tai,Y.T.etal.Novelanti-B-cellmaturation antigen antibody-drug conjugate – and management. Am.J.Hematol.93,981 1114 (2018). (GSK2857916) selectively induces killing of multiple myeloma. Blood 123, 3. Chim,C.S.etal.Managementofrelapsed and refractory multiple myeloma: 3128–3138 (2014). novel agents, antibodies, immunotherapies and beyond. Leukemia 32, 30. Lin, L. et al. Preclinical evaluation of CD8+ anti-BCMA mRNA CAR T cells for – 252 262 (2018). treatment of multiple myeloma. Leukemia https://doi.org/10.1038/s41375-020- 4. Pawlyn,C.&Morgan,G.J.Evolutionary biology of high-risk multiple myeloma. 0951-5 (2020). – Nat. Rev. Cancer 17, 543 556 (2017). 31. Tai, Y. T. et al. Anti-CS1 humanized monoclonal antibody HuLuc63 inhibits 5. Hassan, H. et al. Syndecan-1 modulates beta-integrin-dependent and myeloma cell adhesion and induces antibody-dependent cellular cytotoxicity interleukin-6-dependent functions in breast cancer cell adhesion, migration, in the bone marrow milieu. Blood 112, 1329–1337 (2008). – and resistance to irradiation. FEBS J. 280,2216 2227 (2013). 32. Tai, Y. T. et al. Potent in vitro and in vivo activity of an Fc-engineered 6. Barbouri, D. et al. Syndecans as modulators and potential pharmacological humanized anti-HM1.24 antibody against multiple myeloma via augmented targets in cancer progression. Front. Oncol. 4,4(2014). effector function. Blood 119,2074–2082 (2012). 7. Beauvais,D.M.etal.Syndecan-1(CD138) suppresses apoptosis in multiple 33. de Weers, M. et al. Daratumumab, a novel therapeutic human CD38 mono- myeloma by activating IGF1 receptor: prevention by synstatinIGF1R inhibits clonal antibody, induces killing of multiple myeloma and other hematological – tumor growth. Cancer Res. 76,4981 4993 (2016). tumors. J. Immunol. 186,1840–1848 (2011). 8. McCarron, M. J., Park, P. W. & Fooksman, D. R. CD138 mediates selection of 34. Jiang, H. et al. SAR650984 directly induces multiple myeloma cell death via – mature plasma cells by regulating their survival. Blood 129,27492759 lysosomal-associated and apoptotic pathways, which is further enhanced by (2017). pomalidomide. Leukemia 30,399–408 (2016). fi 9. Akl, M. R. et al. Molecular and clinical pro les of syndecan-1 in solid and 35. Nijhof, I. S. et al. Preclinical evidenceforthetherapeuticpotentialofCD38- hematological cancer for prognosis and precision medicine. Oncotarget 6, targeted immuno- in multiple myeloma patients refractory to – 28693 28715 (2015). lenalidomide and bortezomib. Clin. Cancer Res. 21,2802–2810 (2015). 10. Khotskaya, Y. B. et al. Syndecan-1 is required for robust growth, vascularization, 36. Krejcik, J. et al. Daratumumab depletes CD38+ immune regulatory cells, – and metastasis of myeloma tumors in vivo. J. Biol. Chem. 284, 26085 26095 promotes T-cell expansion, and skews T-cell repertoire in multiple myeloma. (2009). Blood 128,384–394 (2016).

Blood Cancer Journal Yu et al. Blood Cancer Journal (2020) 10:110 Page 13 of 13

37. An, G. et al. Osteoclasts promote immune suppressive microenvironment in 44. Wang, Y. et al. Fratricide of NK cells in daratumumab therapy for multiple multiple myeloma: therapeutic implication. Blood 128,1590–1603 (2016). myeloma overcome by ex vivo-expanded autologous NK cells. Clin. Cancer 38. Tai, Y. T., Cho, S. F. & Anderson, K. C. Osteoclast immunosuppressive effects in Res. 24, 4006–4017 (2018). multiple myeloma: role of programmed cell death ligand 1. Front. Immunol. 9, 45. Overdijk, M. B. et al. Antibody-mediated phagocytosis contributes to the anti- 1822 (2018). tumor activity of the therapeutic antibody daratumumab in lymphoma and 39. Tai, Y. T. et al. Bruton tyrosine kinase inhibition is a novel therapeutic strategy multiple myeloma. MAbs 7,311–321 (2015). targeting tumor in the bone marrow microenvironment in multiple myeloma. 46. Cleary, K. L. S. et al. Antibody distance from the cell membrane regulates Blood 120,1877–1887 (2012). antibody effector mechanisms. J. Immunol. 198,3999–4011 (2017). 40. Tai, Y. T. et al. CRM1 inhibition induces tumor cell cytotoxicity and impairs 47. Feng, X. et al. Targeting CD38 suppresses induction and function of T reg- osteoclastogenesis in multiple myeloma: molecular mechanisms and ther- ulatory cells to mitigate immunosuppression in multiple myeloma. Clin. Cancer apeutic implications. Leukemia 28,155–165 (2014). Res. 23, 4290–4300 (2017). 41. Cho,S.F.etal.BCMA-targetingtherapy: driving a new era of immunotherapy 48. van der Veer, M. S. et al. Towards effective immunotherapy of myeloma: in multiple myeloma. Cancers (Basel) 12, 1473. https://doi.org/10.3390/ enhanced elimination of myeloma cells by combination of lenalidomide with cancers12061473 (2020). the human CD38 monoclonal antibody daratumumab. Haematologica 96, 42. Tai,Y.T.&Anderson,K.C.Bcellmaturation antigen (BCMA)-based immu- 284–290 (2011). notherapy for multiple myeloma. Expert Opin. Biol. Ther. 19, 1143–1156 (2019). 49. Zhang, L. et al. Regulatory B cell-myeloma cell interaction confers immuno- 43. Mateos, M. V. et al. Overall survival with daratumumab, bortezomib, mel- suppression and promotes their survival in the bone marrow milieu. Blood phalan, and prednisone in newly diagnosed multiple myeloma (ALCYONE): a Cancer J. 7, e547 (2017). randomised, open-label, phase 3 trial. Lancet 395, 132–141 (2020). 50. Akhmetzyanova, I. et al. Dynamic CD138 surface expression regulates switch between myeloma growth and dissemination. Leukemia 34,245–256 (2020).

Blood Cancer Journal