Enhanced SLAMF7 Homotypic Interactions by Elotuzumab Improves NK Cell Killing of Multiple Myeloma

Abstract

Elotuzumab (Elo) is an IgG₁ monoclonal antibody targeting SLAMF7 (CS1, CRACC, and CD319), which is highly expressed on multiple myeloma (MM) cells, natural killer (NK) cells, and subsets of other leukocytes. By engaging with FcγRIIIA (CD16), Elo promotes potent NK cell-mediated antibody-dependent cellular cytotoxicity (ADCC) and macrophage-mediated antibody-dependent cellular phagocytosis (ADCP) toward SLAMF7⁺ MM tumor cells. Relapsed/refractory MM patients treated with the combination of Elo, lenalidomide, and dexamethasone have improved progression-free survival. We previously showed that Elo enhances NK cell activity via a costimulation mechanism, independent of CD16 binding. Here, we further studied the effect of Elo on cytotoxicity of CD16-negative NK-92 cells. Elo, but not other SLAMF7 antibodies, uniquely enhanced cytotoxicity mediated by CD16-negative NK-92 cells toward SLAMF7⁺ target cells. Furthermore, this CD16-independent enhancement of cytotoxicity required expression of SLAMF7 containing the full cytoplasmic domain in the NK cells, implicating costimulatory signaling. The CD16-independent costimulation by Elo was associated with increased expression of NKG2D, ICAM-1, and activated LFA-1 on NK cells, and enhanced cytotoxicity was partially reduced by NKG2D blocking antibodies. In addition, an Fc mutant form of Elo that cannot bind CD16 promoted cytotoxicity of SLAMF7⁺ target cells by NK cells from most healthy donors, especially if previously cultured in IL2. We conclude that in addition to promoting NK cell-mediated ADCC (CD16-dependent) responses, Elo promoted SLAMF7-SLAMF7 interactions in a CD16-independent manner to enhance NK cytotoxicity toward MM cells.

Figure 1.. Elotuzumab promoted maximal ADCC if NK cells expressed high affinity CD16 and required SLAMF7 expression on MM cells.

A) Comparison of CD16 expression (GRM1 antibody, Southern Biotech) on parental NK-92 cells (CD16⁻; grey) and NK-92 cells transduced to express 176V (solid line) or 176F (dashed) polymorphic variants of CD16, and mean GMFI ±S.D. from 5 experiments (bottom). B) SLAMF7 expression before and after the genetic modification of two human MM cell lines (Elo + anti-IgG1 secondary stain). MM.1R cells were transfected with SLAMF7 zinc finger nucleases (ZFN) to genetically deplete SLAMF7 (MM.1R SLAMF7KO; left), and RPMI8226 cells were stably transduced with SLAMF7 cDNA (RPMI8226+SLAMF7; right), and mean GMFI ±S.D. from 5 experiments (bottom). Cells were stained with biotinylated Elo plus secondary streptavidin-APC. C,D) Effect of SLAMF7 expression on NK cell mediated ADCC in the presence of various concentrations of Elo (0.001–300 ug/ml). Values are mean ±S.D. Asterisks indicate statistical significance comparing cytotoxicity by NK-92 cells expressing 176V or 176F variants of CD16 using Welch’s unequal variance t test. ** - P<0.01, * -P<0.05, n=5 independent experiments.

Figure 2.. Elotuzumab enhances natural cytotoxicity by NK cells in a CD16-independent manner.

A) Effect of Elo (0.001–300ug/ml) on natural cytotoxicity of parental NK-92 cells expressing SLAMF7 and lacking CD16 (SLAMF7+, CD16−), toward MM.1R (top) and RPMI8226 (bottom) cells. Grey squares are genetically-modified MM.1R SLAMF7KO (top panel) or RPMI83226+SLAMF7 (bottom panel) targets. Values are mean ±S.D. Overhead bars mark statistical significance between untreated and different Elo concentration groups using paired Student’s t test, ** - P<0.01, * -P<0.05, n=5 independent experiments. B) SLAMF7 expression on parental SKOV3 cells (solid line) and SKOV3 cells that were stably transduced with SLAMF7 cDNA (dashed line) and stained with the 162.1 SLAMF7 antibody (Biolegend), and mean GMFI ±S.D. from 5 experiments (bottom). C) Kinetics of parental SKOV3 (top graph) or SKOV3(+SLAMF7) (bottom graph) target cell death by parental NK-92 cells (SLAMF7+, CD16−) analyzed in xCELLigence assays., n = 4-6. D) Mean, S.D., and statistical analysis for percentage target cell death (compared to detergent lysis conditions as 100%) at 10 hours of xCELLigence assays from 4-6 independent experiments. Overhead bars mark statistical comparison between indicated groups using paired Student’s t test, ** - P<0.01, * -P<0.05. E) Schematic of Elo effect on cytotoxicity of NK cells against SLAMF7+ and SLAMF7− targets.

Figure 3.. Elotuzumab is unable to enhance natural cytotoxicity of NK cells lacking expression of both SLAMF7 and CD16.

A) SLAMF7 expression [162.1 SLAMF7 antibody (Biolegend)] on NK-92 control (GFP knockout, SLAMF7+, CD16−; solid line) and NK-92 SLAMF7 KO (CD16−; dashed line) cells generated using CRISPR/Cas9, and mean GMFI ±S.D. from 5 experiments (bottom). B) Cytotoxicity toward SKOV3(+SLAMF7) target cells by NK-92 (SLAMF7+, CD16−) and NK-92 (SLAMF7 KO, CD16−) cells studied over time in xCELLigence assays with or without Elo (10 ug/ml; top panel) or Elo Fc mutant (Elo Fc mut; 10 ug/ml; bottom panel). C) Mean ±S.D. target cell death at 10-hours from xCELLigence assays for Elo (top panel) and Elo Fc mut (bottom panel) toward SKOV3(+SLAMF7) target cells from 4-5 independent experiments. Overhead bars mark statistical comparison between indicated groups using paired Student’s t test, ** - P<0.01, * -P<0.05. D) Cytotoxicity toward SKOV3(+SLAMF7) target cells by NK-92 (SLAMF7+, CD16−) cells studied over time in xCELLigence assays (top panel) and mean ±S.D. target cell death at 10-hours (bottom panel) for Elo vs Elo F(ab’)₂ vs Fab (3 and 10 µg/ml). Overhead bars mark statistics as in panel C. n = 4 independent experiments. E) Schematic showing requirement for SLAMF7 expression on NK cells for CD16-independent NK cell-mediated cytotoxicity.

Figure 4.. Elotuzumab uniquely promotes SLAMF7-SLAMF7 interactions.

A) SLAMF7-TCRζ reporter cell line was cultured with four different plate-bound SLAMF7 antibodies (Elo, ChuLuc90, PDL241 and 162.1; coated overnight at 10 ug/ml) and mouse (m)IL-2 was assayed in culture supernatant. B) Indicated SLAMF7 antibodies or isotype controls (0.01 - 3 ug/ml) were added to 3A9 SLAMF7-TCRζ reporter cell line in plates coated with recombinant human(h) SLAMF7 overnight and mIL2 secretion was assayed. C) Soluble Elo (left panel) or hIgG1 isotype control (right panel) were added to SLAMF7-TCRζ reporter cell line on plate-bound (PB) recombinant hSLAMF7 or PB mesothelin control protein, or hSLAMF7 was added as soluble protein (+hSLAMF7; right panel). Plotted as mean ±SEM. Asterisks mark statistical comparisons for Elo Stimulated group using paired Student’s t test, ** - P<0.01, * -P<0.05.

Figure 5.. Elotuzumab promotes co-stimulation-mediated cytotoxicity.

Impact of various SLAMF7 antibodies (10 ug/ml each) on cytotoxicity by NK-92(SLAMF7+,CD16−) cells (A) or SLAMF7-deficient NK-92 (SLAMF7 KO,CD16−) cells (B) toward SKOV3(+SLAMF7) target cells in xCELLigence assays. Mean, S.D., and statistics for target cell death are shown at 10 hours from 3-4 independent xCELLigence assays. Overhead bars mark statistical significance between indicated groups using paired Student t test, ** P < 0.01, *P < 0.05. Elo = Elotuzumab, CL90 = ChLuc90, 162 = 162.1, P241 = PDL241. C), D) Test of the impact of blocking NKG2D (C) and DNAM1 (D) antibodies on Elo-induced cytotoxicity by NK-92(SLAMF7+, CD16−) cells toward SKOV3(+SLAMF7) target cells in xCELLigence assays. Mean, S.D., and statistics for target cell death at 10 hours from 5 independent assays using blocking antibodies at 10 µg/ml each. Overhead bars mark statistics as in panels A and B. Kinetic plots for all of these xCELLigence assays can be found in Supplemental Figure S5C-D.

Figure 6.. Elotuzumab enhances NK cell cytotoxicity only if SLAMF7 contains the full cytoplasmic domain.

A) Representative SLAMF7 expression [162.1 antibody (Biolegend)] on NK-92 SLAMF7 KO cells (grey) and reconstituted with SLAMF7-L (dashed) or SLAMF7-S (solid line) isoforms, and mean GMFI ±S.D. from 5 experiments (bottom). B) Impact of Elo (10 ug/ml) on cytotoxicity by NK-92(SLAMF7 KO,CD16−) cells reconstituted with either SLAMF7-L or SLAMF7-S toward SKOV3(+SLAMF7) target cells studied in xCELLigence assays. C) Mean, S.D., and statistics for target cell death at 10 hours from 5 independent assays. Overhead bars mark statistical comparisons between indicated groups using paired Student’s t test, ** - P<0.01, * -P<0.05.

Figure 7.. Co-stimulation of healthy donor NK cell cytotoxicity by Elo Fc mut depends on NK cell activation status.

A) Elo or Elo Fc mut (10 ug/ml) were added to SKOV3(+SLAMF7) target cells in the presence or absence of freshly purified NK cells from healthy donor (HD) #1 in xCELLigence assays. B) Compilation of cytotoxicity data for n=7 healthy donors. Percentage target cell death at the 10 hour time point from xCELLigence assays in the absence (−) or presence (+) of Elo Fc mut by purified NK cells toward SKOV3(+SLAMF7) target cells on the day of the purification (left plots) or after cultured in presence of recombinant human IL-2 (100U/ml; +) for 48 hrs (right plots). Each patient is designated by a unique icon and results with or without antibody addition are connected with lines. Overhead bars mark statistical comparison between indicated groups using Wilcoxon matched-pairs signed rank test. C,D) Representative time course data in which purified NK cells from HD#3 and HD#8 were assayed fresh (left) or after cultured in presence of recombinant human IL-2 (100U/ml) for 48 hrs (right).