Targeting Mechanisms for Natural Killer Cell Dysfunction in Patients with Multiple Myeloma

Patients with multiple myeloma (MM) have a multi-faceted immune dysfunction with poor NK cell cytotoxic function. We show that NK cells in MM have a disturbed homeostasis, featured by an increased frequency of terminally differentiated CD57⁺ NK cells and almost complete absence of the less mature CD56^hi NK cell subset. MM patient NK cells have reduced levels of killer activating receptors (KAR) [Nkp46, Nkp44 and NKG2D] and increased levels of killer inhibitory receptor (KIR) [KIR2DL1 and KIR2DL3]. Despite expressing normal levels of the NK effector transcription factors Tbet and eomesdermin, and the effector protein perforin, MM patient NK cells demonstrate poor cytotoxic function against myeloma cell lines and lower levels of the cytotoxic proteins granzyme B (gzmB).

Elotuzumab (Elo, Bristol Myers Squibb) is a human IgG1 antibody which targets SLAM-F7, a protein highly expressed on myeloma plasma cells and on specific subsets of immune cells. Elo mediates dual activation of NK cells and has shown promising clinical responses when combined with lenalidomide; however the complete mechanism(s) whereby Elo mediates these effects have not been fully described. Thus, this project investigates the Elo mechanisms of action in myeloma. To explore whether MM patient NK cells were capable of ADCC, we examined NK cell subset degranulation (CD107a) and IFN-γ secretion when co-cultured with the SLAM-F7 expressing OPM-2 myeloma cells and Elo. Healthy donor NK cells degranulated in response to Elo-labeled OPM-2 cells, which correlated with loss of the CD56^loCD16⁺ NK cell subset. In contrast, newly diagnosed (NEW) and refractory/relapsed MM (RRMM) patient NK cells had a decreased response to Elo-labeled OPM-2 cells as the CD56^loCD16⁺ NK subset was retained, and they degranulated at a significantly lower level than that observed in healthy donor NK cells. Therefore, all MM patient NK subsets (CD56^loCD16+, CD161+ and CD57+) assessed were low responders to Elo-labeled OPM-2 cells.

We then visually examined NK cell killing kinetics of Elo-labeled myeloma target cells using time lapse live video microscopy (TLLVM). NK cells used ADCC to kill OPM-2 cells in the presence of Elo, the killing kinetics of was equivalent to direct killing induced in presence of the human IgG isotype control, the delivery of the lethal hit occurring in under three minutes. Elo also mediated serial killing of OPM-2 cells. Interestingly, Elo did delay the NK cell detachment from dying myeloma plasma cell targets, which results in prolonged signaling to the NK cell and consequently higher levels of secreted cytokines and pro-inflammatory chemokines. Taken together, this data indicates Elo-induces a profound change in the NK-mediated killing kinetics of myeloma cells and that likely recruits other immune effectors to the myeloma killing site.

Given that NK cells can also express SLAM-F7, we assessed whether they were likely to be a target for other NK cells when bound to Elo, and induce fratricide. We used killing assays, flow cytometry and TLLVM to show that in the presence of Elo, healthy donor NK cells were capable of NK fratricide under the assay conditions tested. Interestingly, fratricide induced death saw the depletion of the CD56^loCD16⁺ NK subset. This data may partially explain why there is a rapid but transient reduction of NK cells from the peripheral blood of MM patients who received single agent Elo.

 In summary, MM patient NK dysfunction is due to aberrant NK cell homeostasis featured by terminal differentiation of the NK cells to a subset with poor natural cytotoxic function and high KIR expression. In vitro, Elo is less effective at inducing ADCC responses by NK cells from MM patients compared to normal donors. Nonetheless, Elo induces efficient normal donor NK cell killing of myeloma target cells, the killing kinetics is featured by a protracted immune synapse time and secretion of high levels of pro-inflammatory cytokines/chemokines by 24 hours, plus serial killing of myeloma target cells. These data likely explain the single agent efficacy of Elo in RRMM patients. The data also underlines the exciting potential of Elo to induce NK mediated ADCC of myeloma plasma cells, and recruit multiple arms of the immune system to the killing site. To harness its full potential, combination of Elo with other drugs (eg. Lenalidomide and anti-KIR) maybe required to normalize MM patient NK cell function and initiate the immune attack.