VIS832, a novel CD138-targeting monoclonal antibody, potently induces killing of human multiple myeloma and further synergizes with IMiDs or bortezomib in vitro and in vivo

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 daratumumab-resistant MM cells were highly susceptible to VIS832 which, unlike 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 lenalidomide 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.


Production of VIS832
VIS832 was produced by the transient co-transfection of two mammalian expression vectors separately encoding heavy chain (HC) and light chains (LC) for production as full length, human IgG1kappa antibodies in Chinese hamster ovary cells (ExpiCHO, Thermo Fisher Scientific). DNA transfection was performed using lipid-based transfection reagents (e.g., ExpiFectamine CHO Transfection Kit, Thermo Fisher Scientific) and standard protocols as recommended by the manufacturer. VIS832 was produced under cell culture conditions designed to reduce core fucose in the respective N-glycans at asparagine 297 and ultimately for the purpose of further enhancing Fc-mediated ADCC activity. Highly-reduced core fucosylation was achieved metabolically through the use of the decoy substrate 2-deoxy-2-fluoro-l-fucose (2FF), which was added concurrently to the culture media at 0.15 mM at four hours following transient transfection. Cell culture was carried out in shake flasks at 600 ml cell culture scale for a period of 7 days without feed supplementation. Secreted monoclonal antibodies were subsequently purified from cell culture media using protein A affinity capture on Fast Protein Liquid chromatography (FPLC). Purified VIS832 was formulated in phosphate buffered saline (PBS), pH 7.4, and quantified by UV absorbance at 280 nm using a theoretically derived extinction coefficient of 1.41 (ε 0.1%). Purity and molecular mass were confirmed by both SDS-PAGE and SEC-HPLC, validating predicted molecular weight and >95% purity.

Reagents (Test antibodies)
BB4 -The mouse anti-CD138 antibody clone BB4 was cloned as a chimeric molecule with human IgG1 heavy chain constant region (Fc), recombinantly produced by recombinant transfection in EXPI 293 as an afucosylatated antibody and purified by Protein A affinity capture chromatography, as otherwise described for VIS832.
Daratumumab -Darzalex® (Daratumumab), 5 mg/20 ml solution, was purchased through Pharmaceutical Buyers International, a fully licensed third-party purchaser of pharmaceutical products for use in research.

MM cell culture and patient derived cells
Unless otherwise stated, all human MM cell lines express CD138 and BCMA, and they were obtained from American Type Culture Collection (ATCC), as previously described [1][2][3] . Drug resistant cell lines including H929 cells resistant to len/pom H929(R) and MM1S resistant to len/pom MMIS(R) were generated by culturing these cell lines in medium containing gradually increasing concentrations of len and pom (up to 100 µM) 1 . CRISPR/CAS9 technique was used to knock out CD138 expression in H929 and JJN3 cells. Similar technique was used to knock out CD38 expression in H929 cells to generate H929 CD38 KO-1 and H929 CD38 KO-2. All MM cell lines were cultured in media containing RPMI-1640 (Invitrogen, Carlsbad, CA), 10% FBS (Hyclone, Logan, UT), HEPES buffer (ThermoFisher Scientific), L-glutamine (ThermoFisher Scientific), penicillin, and streptomycin (ThermoFisher Scientific). They were routinely checked by Human short tandem repeat (STR) DNA Profiling Cell Authentication for their authenticity and for mycoplasma contamination by PCR and the MycoSEQ Mycoplasma detection assay kit (ThermoFisher Scientific), as previously described 1,2 . MM1S, MM1S(R), and H929(R) cell lines were subject to the generation of stable luciferase reporter lines (MM1S-luc, MM1S(R)-luc, H929(R)-luc) for the use in indicated ADCC assays. MM1S-luc cells were also used for in vivo work.
Healthy donors and MM patient samples were obtained with informed consent, in accordance with the Declaration of Helsinki and under the auspices of a Dana-Farber Cancer Institute Institutional Review Board (IRB) approved protocol. Primary CD138+ plasma cells were purified from BM aspirates using anti-CD138 microbeads (Miltenyi Biotech, Auburn, CA). Residual CD138-negative BM mononuclear cells (BMMCs) were further cultured for 3 to 6 weeks to derive stromal cells (BMSCs).
Anti-CD138 monoclonal antibodies VIS832 and BB4 were also assessed by flow cytometry for their relative capacity to bind to CD138 on the surface of CD138+ human myeloma cell lines (n>8). MM cells were cultured with recombinant antibody at varying concentrations as noted for up to 4 hours followed by washing in PBS and detection using secondary anti-human IgG1 antibody conjugated with APC or related fluorophore. Antibody cell binding was reported as geometric mean fluorescence intensity (MFI).

Real-time quantitative RT-PCR (qRT-PCR)
Total RNA was isolated using the TRIzol reagent (Life Technologies), according to manufacturer's instructions. cDNA was obtained from 0.5 μg of RNA using the cDNA Reverse Transcription kit (Life Technologies). Real-time PCR was performed using TaqMan gene expression assay primer sets for SDC1 (CD138) (Hs00896423_m1), CD38 (Hs01120071_m1) and internal controls (18S, HS03003631_g1; GAPDH, Hs02758991_g1) on an ABI Prism 7700 sequence detection system (Applied Biosystems).

ADCC Reporter Assay
VIS832 and other anti-CD138 antibodies were independently evaluated for antibody dependent cellular cytotoxicity (ADCC) activity against U266 target MM cell line using an orthogonal luciferase based ADCC reporter bioassay (Catalog number G7010, Promega). In this assay, engineered Jurkat cells were stably transfected with the human FcγRIIIa receptor and served as surrogate effector cell line. U266 target cells were incubated with mAbs that were 3fold serially diluted from a starting concentration of 3 µg/ml and subsequently added to effector cells at a cell ratio of effector: target (E:T) of 10:1. After 16-hour incubation at 37 o C, luciferase activity was assessed as a measure of ADCC activity and reported as relative luminescence units (RLU). ADCC data were plotted using nonlinear regression analysis and a 4-parameter curve fit.
Controls include effector cell only to determine non-specific background. Data are typically plotted as fold-induction of ADCC activity relative to no antibody control.

Syndecan-1 ELISA
Soluble syndecan-1 levels in cell culture media were quantified by a sandwich enzymelinked immunosorbent assay (ELISA) using Human Syndecan-1 DuoSet ELISA Kit (R&D Systems, Inc., MN). The syndecan-1 concentrations in the samples were determined by a standard protein curve, according to the manufacturer's protocol. The reported assay range is listed by the vendor as 125.0 -8,000 pg/ml. The same assay was also used to quantify syndecan-1 in cell culture supernatants.

Luciferase-based ADCC assay
In experiments comparing the ADCC activities of VIS832 and daratumumab, a modified luciferase-based assay was also used. In brief, MM1S-, MM1S(R)-, and H929(R)-luc cell lines were used as target cells to co-culture with human NK cells at an E:T ratio of 10:1. Both target and effector cells were also cultured in the presence of varying concentrations of either VIS832 or daratumumab for 4 hours. Cells were collected by centrifugation and lysed with 20 µl passive lysis buffer (to establish 100% lysis); firefly luciferase value was quantified in accordance with instructions as provided in the Dual-Luciferase® Reporter Assay System Technical Manual.

VIS832-induced cytotoxicity of patient derived myeloma cells
The potency of VIS832 in mediating ADCC against patient derived cells was evaluated in two separate methods. In the first method, BMMCs derived from RRMM patients were used as both the source of target and effector cells. These cells were directly treated with varying concentrations of VIS832 (0-10 µg/ml) for 24 hours. The population of patient derived plasma cells was detected and quantified by flow cytometry using BCMA as a surrogate plasma cell surface marker. The % of viable BCMA + patient MM cells remaining was used to determine the extent of cell killing ability of VIS832.
In the second method, MM cells from refractory/relapsed patients were sorted by CD138 microbeads. CD138+ or CD138-cells were co-cultured for 1.5 hours with PBMCs from the matched individual (E:T = 10:1), in the presence of varying concentrations of VIS832 (0-1 µg/ml). Target cells (CD138+ or the matched CD138-cells) were pre-labeled with Calcein AM, followed by the Calcein AM release assay to determine CD138-specific cytotoxicity (% lysis) of VIS832.

Combined use of VIS832 and lenalidomide or bortezomib
ADCC activity was measured as described above using NK cells and calcein AM-prelabeled MM cell lines (n>4), with some modifications. Target MM cell lines were co-treated with sub-optimal doses of len or btz (Selleck Chemicals) and varying concentrations of VIS832 in the presence of effector cells (lower E:T ratio of 4:1) for a period of 2 hours, prior to measuring released calcein-AM. In some experiments, various MM cell lines were pre-treated with len or btz for 10 hours prior to VIS832-induced ADCC assays (E:T ratio of 1:1) for 1.5 hours prior to analysis of cellular lysis.

Murine model of disseminated human MM disease
Five-to-seven-week-old female CB-17 SCID mice (C.B-17/IcrHsd-Prkdc scid ) were injected intravenously (i.v.) with 5.0x10 6 live (trypan-excluding) MM1S-luc cells on day 0 and staged for therapeutic dosing on day 14, after normalizing the baseline tumor burden between the groups set by bioluminescence (BLI) as 2.90x10 6 photons/second (p/s, range of group means, 2.81-3.07x10 6 p/s). Whole body tumor burden was monitored by BLI using an IVIS Lumina S5 (PerkinElmer, MA). At 14 days after tumor implant, mice (n=9 per group) were randomized to ensure that the mean whole-body tumor burden for all groups was within 10% of the overall mean whole-body tumor burden for the study population before treatments started. Animals were imaged five at a time under 1-2% isoflurane gas anesthesia. Each mouse was injected IP with 150 mg/kg (15 mg/ml) D-luciferin (Promega) formulated in saline for injection and imaged in the prone, then supine positions 10 minutes after the injection. Large binning of the CCD chip was used, and the exposure time was adjusted (5 seconds to 2 minutes) to obtain at least several hundred counts per image and to avoid saturation of the CCD chip.
Animals were imaged once weekly at staging and on days 14, 21, 29, 36, 43, 53, and 73 (end of study) for all animals remaining on study. Images were analyzed using Living Image 4.7.1 (PerkinElmer, MA) software. Whole body fixed-volume regions of interest (ROIs) were placed on prone and supine images for each individual animal and labeled based on animal identification. Total flux (photons/sec) was calculated and exported for all ROIs to facilitate analyses between groups. The prone and supine ROIs were summed together to estimate whole body tumor burden. Subject ROIs were also placed over each mouse, and region-specific ROIs were placed over spine and hindlimbs for each animal. These ROIs were then quantified independent of the whole-body ROIs in order to calculate signal from only hind limbs or spine.
Primary and secondary efficacy were generated using in vivo whole body BLI and included tumor burden as defined. Endpoints included whole body BLI ratio T/C (where T = treatment group and C = vehicle control) measured on each imaging day and reported on day 36, and tumor growth delay, as measured as time to tumor burden as pre-defined as a BLI signal of 1.0x10 9 p/s. Additional efficacy endpoints included tumor regression, tumor free survival (TFS), and median life span normalized to vehicle control (% increased life span); this latter endpoint was calculated based on the first day of treatment and not the day of implant. Overall survival was calculated based on animals remaining on study.
Additional measurements included body weights (3x weekly), daily clinical observations, and necropsy following euthanasia. Pre-defined euthanasia criteria were related to disease progression and included >20% loss in body weight or disease-related morbidity such as hind limb paralysis, paresthesia, or other signs of severe CNS or musculoskeletal impairment.

Statistical analysis
Multiple groups (≥3) were analyzed by one-way ANOVA, and paired groups were analyzed by two-way ANOVA or Student t test. All data were graphed and analyzed using GraphPad Prism 8.3.1 (San Diego, CA) with P<.05 indicating statistical significance. Tumor growth curves in vivo were analyzed by Tukey's multiple comparison test and 2-way ANOVA, after log-transformation of data. The survival curve in the murine model was plotted by using the Kaplan-Meier method, and compared by the log-rank test.
Drug interactions between VIS832 and len or btz were assessed by CompuSyn software using the Chou-Talalay method to determine combination index (CI) 4 . CI<1 indicates synergism whereas CI>1 antagonism and CI=1 additive effect. Table S1. Fold increases in binding affinity of VIS832 vs. BB4 to CD138 on MM cell membrane. Target binding of VIS832 vs BB4 mAb in MM cell lines was evaluated by quantitative flow cytometry (FC) analysis using an APC-conjugated anti-human IgG1 antibody for detection (Fig.  1A). 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 (n=3 replicates; ± SEM). Fold changes were calculated by the MFI values for VIS832 divided by those for BB4. Table S2. EC50 of VIS832 and its chimeric predecessor 2810 mAb. VIS832 (left)-and 2810 mAb (right)-induced ADCC to lyse indicated MM cell lines (n=12, left; n=10, right) using the calcein-AM release assay ( Fig. 2A, Supplementary Figure S3A, C). Human NK effector cells from 3 healthy donors were incubated with indicated MM cell lines (E:T=10:1) in a serial concentration of VIS832. The percent maximum lysis and EC50 were determined from 3 independent experiments, with triplicate measurements for each assay. CD138 MFI values was determined by flow cytometry analysis using anti-CD138 mAb DL101 clone (Supplementary Figure S1A). Data are shown as means (% lysis) + SEMs of three independent experiments.  Table S4. Group (9 mice each group) summary for in vivo efficacy study

Definitions for animal efficacy endpoints.
Median T/C -Is a group endpoint. It is calculated for each day of treatment as: Time to Evaluation Size (TES) -TES is an individual mouse endpoint, and it is expressed in days from tumor implant. It is the time it takes the tumor burden to reach a specified value, for this study, tumor burden is evaluated by BLI. It is calculated by log-linear interpolation between the two closest data points that bracket the chosen tumor burden.
= ℎ − ((log ℎ − log ) * ( ℎ − )) (log ℎ − log ) where: = -the day evaluation size is reached ℎ -the day of the first measurement greater than the was reached -the day of the last measurement before the was reached ℎ -The tumor burden on day ℎ -the tumor burden on -the evaluation size Tumor growth delay (TGD) -TGD is a group endpoint. Tumor growth delay is expressed in units of days and is calculated from the median times it takes the mice in a group to reach a specified tumor burden (time to evaluation size, TES as defined above). It can be calculated as: TGD = median TES treatedmedian TES control Complete regression (CR)-An animal is credited with a complete regression if its tumor burden decreases to less than a declared background BLI signal level. In this particular study, this background signal level was measured as 7.20E+05p/s Tumor-free Survivor (TFS) -A TFS is any animal that (1) survives until termination of the study, and (2) has no reliably measurable evidence of disease at study termination. Mice that are tumor-free at some point during the study but are then euthanized for sampling or other purposes prior to the end of the study are not considered in this assessment. They are excluded from calculation of the %TFS.

Figure S2. Optimization of VIS832
Fc afucosylation further increased maximal ADCC in the Promega reporter cell assay, where Jurkat cells stably transfected with human FcRIIIa as surrogate for effector cells were cocultured with target U266 MM cell line (as in Figure 1C). . Example of flow cytometry gating strategy on VIS832 autologous killing in samples of patients with MM. VIS832 was incubated for 24 hours with autologous patient MM cells and effector immune cells isolated from bone marrow aspirates. Cell killing was assessed by flow cytometry to monitor antibody-dependent loss of BCMA+ plasma cells, as described for VIS832 in methods. More than 90% lysis of patient (BCMA+) MM cells was achieved in selected patients. . In vivo treatment effects followed by changes in BLI in animals A, Representative images of individual mice from each group (Fig. 6) in the efficacy study, chosen based on their proximity to median BLI at day 36. Image intensity was adjusted for purposes of normalization (min of 1.0e106; max of 3.0x10 8 ) and does not imply lack of tumor burden at day 14 (day of staging and prior to first dose). Note that all animals in VC (PBS) group were deceased by day 51. B, BLI follow-up in individual mouse (n=9 per group).