NK Cells Kill Myeloma Cells By Increasing ER Stress and Decreasing Autophagy Levels. NKG2D and NKP30 Are Involved In These Processes

Multiple myeloma (MM) plasma cells are equipped with an extensive endoplasmic reticulum (ER) for antibody production. Misfolding of newly synthesized immunoglobulins leads to ER stress and cell death. Plasma cells evade this by enhancing protein degradation through the proteasome pathway and the autophagy machinery. Because of this, MM cells are sensitive to proteasome inhibitors (PIs). However, autophagy induced by PIs and by reactive oxygen species (ROS) can increase drug resistance and lead to eventual relapse.

We have previously demonstrated in vitro and in vivo sensitivity of MM cells to cord blood-derived natural killer (NK) cells, and have observed that this effect may involve NKG2D and NKP30.

In this study we used confocal microscopy, chromium cytotoxicity assays and flow cytometry and we aimed to determine if there is a unique mechanism of NK cell-mediated cytotoxicity on MM cells involving ER stress, ROS and its dependence on NK cell receptors NKG2D and NKp30.

We first demonstrated a difference in NK-mediated killing of MM cells vs K562 cells. As expected inhibition of GranzymeB (GrB) and Caspase3 pathway yielded a 13% reduction in NK-cytotoxicity against K562 cells; however, a 12% increase in killing was seen against ARP1 cells (p<0.001), indicating a GrB-Caspase3 independent mechanism for NK cell cytotoxicity against ARP1 cells. We observed a decrease in ROS levels after NK treatment for both cell lines (p<0.001), but inhibition of GrB and Caspase 3 abrogated the decrease in ROS only for K562 (p<0.001). By confocal microscopy we observed that NKG2D and NKP30 were transferred to tumor cells and entered the endocytic vesicle pathway. These receptors also appeared to play a more relevant role in the killing of MM cells (33%, 49% and 73% in ARP1 vs 4%, 3% and 19% in K562 cells of killing reduction after blocking NKG2D, NKP30 and both together, p<0.001). Regarding ER stress in ARP1 cells, we saw that after NK treatment there was a significant decrease in endocytic vesicle markers (EEA1 and Rab11, p<0.001) indicating a decrease in the transport of properly folded proteins. There was also a decrease in Rab7 and in Lysosomes (indicated by Lyso-Tracker probes) (p<0.001) and a dramatic increase for DNA-damage-inducible transcript 3 (CHOP) expression (p<0.001) indicating a decreased level of misfolded protein degradation and higher ER stress after exposure to NK cells. Blocking NKG2D abrogated the NK effect on ER stress (indicated by CHOP levels) and ROS levels in ARP1 cells. Blocking of NKP30 abrogated the NK effect on protein degradation activity (Rab7 levels), and ROS levels. To confirm that these receptors mediate killing of ARP1 cells by increasing ER stress, we performed cytotoxicity assays blocking these receptors and adding the PI bortezomib. Bortezomib did not improve NK cell killing at baseline in MM cells. However, bortezomib was able to partially reverse the deleterious effect of blocking NKG2D or NKp30 (33% enhanced killing for NKG2D-blocked and 57% for NKP30-blocked). Finally, we observed that NKG2D and NKP30 were also transferred secondarily to neighboring ARP1 from ARP1 cells which had been primarily synapsed to NK cells (p<0.001). In addition, these secondary ARP1 cells demonstrated decreased ROS levels after this contact. In a cytotoxicity assay, those ARP1 cells which had previously synapsed with NK cells demonstrated 7% of killing against new ARP1 cells.

In summary, we describe a new mechanism of killing of MM cells by NK cells mediated by increased ER stress, decreased autophagy and dependent on NKG2D and NKP30. Furthermore, this cytotoxicity can be transmitted between MM cells. These findings could lead to new cellular therapeutic strategies for MM treatment.