CS1, a potential new therapeutic antibody target for the treatment of multiple myeloma

Abstract

Purpose: We generated a humanized antibody, HuLuc63, which specifically targets CS1 (CCND3 subset 1, CRACC, and SLAMF7), a cell surface glycoprotein not previously associated with multiple myeloma. To explore the therapeutic potential of HuLuc63 in multiple myeloma, we examined in detail the expression profile of CS1, the binding properties of HuLuc63 to normal and malignant cells, and the antimyeloma activity of HuLuc63 in preclinical models.

Experimental design: CS1 was analyzed by gene expression profiling and immunohistochemistry of multiple myeloma samples and numerous normal tissues. HuLuc63-mediated antimyeloma activity was tested in vitro in antibody-dependent cellular cytotoxicity (ADCC) assays and in vivo using the human OPM2 xenograft model in mice.

Results: CS1 mRNA was expressed in >90% of 532 multiple myeloma cases, regardless of cytogenetic abnormalities. Anti-CS1 antibody staining of tissues showed strong staining of myeloma cells in all plasmacytomas and bone marrow biopsies. Flow cytometric analysis of patient samples using HuLuc63 showed specific staining of CD138+ myeloma cells, natural killer (NK), NK-like T cells, and CD8+ T cells, with no binding detected on hematopoietic CD34+ stem cells. HuLuc63 exhibited significant in vitro ADCC using primary myeloma cells as targets and both allogeneic and autologous NK cells as effectors. HuLuc63 exerted significant in vivo antitumor activity, which depended on efficient Fc-CD16 interaction as well as the presence of NK cells in the mice.

Conclusions: These results suggest that HuLuc63 eliminates myeloma cells, at least in part, via NK-mediated ADCC and shows the therapeutic potential of targeting CS1 with HuLuc63 for the treatment of multiple myeloma.

Fig. 1

Expression of CS1 in normal, benign, and malignant plasma cells (PC). A, CS1 gene expression was quantified using probeset 219159 from the Affymetrix gene chip U133 in plasma cell samples from 24 healthy donors, 14 monoclonal gammopathies of undetermined significance (MGSUS), 35 smoldering myeloma (SMM), 532 multiple myeloma [MM;TotalTherapy (TT) 2/3 baseline], and 45 human myeloma cell lines (HMCL). The expression values ranged from 900-19,548 AIU for normal plasma cells, 1,771-24,632 AIU for monoclonal gammopathies of undetermined significance, 426-2,3697AIU for smoldering myeloma, 258-2,3727 AIU for multiple myeloma, and 28-11,698 AIU for human myeloma cell lines. B, CS1 gene expression in 256 multiple myeloma patients at baseline before Total Therapy 2 and ordered by molecular classification based on Zhan et al. (24). Inset, median AIU values, showing the low-bone disease (LB) group with the highest median expression level. C, paired gene expression analysis at time of diagnosis (black) versus at time of relapse after Total Therapy 2 (gray) shows that CS1expression is retained at significant levels at relapse. PR, higher gene expression associated with proliferation;MS, MMSET/FGFR3; HY, hyperdiploidy; CD1, CCND1; CD2, CCND3; MF, c-MAF/MAFB.

Fig. 2

Western blot analysis using anti-CS1mAbs 1G9, HuLuc63, andMuLuc90 show specificity for CS1. Cell extracts were made from OPM2 multiple myeloma cells (lanes1, 4, and 7), parental 293 cells (lanes 2, 5, and 8), and 293 cells transfected with a CS1 expression vector (lanes 3, 6, and 9). Immunoblots probed with anti-CS1mAbs 1G9, HuLuc63, andMuLuc90 show a broad reactive protein band(s) of ~60 kDa in the OPM2 and the CS-transfected 293 cells. M, markers in kDa.

Fig. 3

CS1 protein expression is restricted to normal and malignant plasma cells in tissues. Frozen section IHC stained with MuLuc63: (A) plasma cell staining in the lamina propria of cecum (×1,000); B, malignant plasma cell staining in extramedullary plasmacytoma (×100; inset, ×400). Paraffin section IHC stained with 1G9; C, a case of lymphoplasmacytic lymphoma (H&E, ×200) showing CS1expression (inset, ×200) in neoplastic cells with plasmacytic maturation; D, two examples of bone marrow with multiple myeloma show typical strong plasma cell staining (×400).

Fig. 4

HuLuc63 mediates ADCC in vitro and inhibits myeloma tumor growth in vivo. A, HuLuc63-mediated ADCC of Cr-51 – labeled primary myeloma cells by purified allogeneic or autologous NK cells. Shown is the average specific lysis from multiple chromium release assays using as effectors purified NK cells from healthy allogeneic donors (5 assays) or autologous NK cells (6 assays) in the presence of (1) control isotype antibody (10 μg/mL), (2) HuLuc63 (10 μg/mL), or (3) HuLuc63 (10 μg/mL) and anti-CD16 Ab (10 μg/mL). The effector to target ratio was 30:1. B, significant ADCC of Cr-51 – labeled OPM2 cells by human PBMCs occurred in the presence of increasing concentrations of HuLuc63 (closed squares, solid line) compared with isotype control Ab (solid circles, solid line). Depletion of NK cells from the same PBMC preparation significantly reduced HuLuc63-mediated lysis (open squares, dashed line). Control Ab exhibited no lysis activity with NK-depleted PBMCs (open circles, dashed line). C and D, mice with established OPM2 xenograft tumors (average of ~100 mm³) were randomized into groups 2 to 3 wk after inoculation and were then treated with either humanized IgG1control antibody or HuLuc63. In all experiments, the average tumor volumes at necropsy (on study days 40-45) were plotted as percent control for comparisons between the different animalmodels. C, antitumor activity of HuLuc63 was compared with HuLuc63-LF (both dosed at 5 mg/kg) and then to the CD16-binding mutant HuLuc63-AA (both dosed at 10 mg/kg). Compared with isotype control (1), HuLuc63 (2), and HuLuc63-LF (3) caused significant tumor growth inhibition (P < 0.01 and P < 0.003, respectively) but were not significantly different from each other. In the second experiment, HuLuc63 (5) exhibited significant antitumor activity compared with control (4; P < 0.00001), whereas HuLuc63-AA (6) did not inhibit tumor growth (P = 0.4). D, compared with isotype control (1), HuLuc63 (2) at 10 mg/kg per dose did not exhibit significant antitumor activity in ICR-SCID mice that were depleted of NK cells with an anti – asialo-GM1antibody (P = 0.1) or in NOD-SCID/IL-2Rγ knockout mice (P > 0.05), which lack NK cells.