Novel Strategy to Expand Super-Charged NK Cells with Significant Potential to Lyse and Differentiate Cancer Stem Cells: Differences in NK Expansion and Function between Healthy and Cancer Patients

Natural killer (NK) cells are known to target cancer stem cells and undifferentiated tumors. In this paper, we provide a novel strategy for expanding large numbers of super-charged NK cells with significant potential to lyse and differentiate cancer stem cells and demonstrate the differences in the dynamics of NK cell expansion between healthy donors and cancer patients. Decline in cytotoxicity and lower interferon (IFN)-γ secretion by osteoclast (OC)-expanded NK cells from cancer patients correlates with faster expansion of residual contaminating T cells within purified NK cells, whereas healthy donors’ OCs continue expanding super-charged NK cells while limiting T cell expansion for up to 60 days. Similar to patient NK cells, NK cells from tumor-bearing BLT-humanized mice promote faster expansion of residual T cells resulting in decreased numbers and function of NK cells, whereas NK cells from mice with no tumor continue expanding NK cells and retain their cytotoxicity. In addition, dendritic cells (DCs) in contrast to OCs are found to promote faster expansion of residual T cells within purified NK cells resulting in the decline in NK cell numbers from healthy individuals. Addition of anti-CD3 mAb inhibits T cell proliferation while enhancing NK cell expansion; however, expanding NK cells have lower cytotoxicity but higher secretion of IFN-γ. Expansion and functional activation of super-charged NK cells by OCs is dependent on interleukin (IL)-12 and IL-15. Thus, in this report, we not only provide a novel strategy to expand super-charged NK cells, but also demonstrate that rapid and sustained expansion of residual T cells within the purified NK cells during expansion with DCs or OCs could be a potential mechanism by which the numbers and function of NK cells decline in cancer patients and in BLT-humanized mice.


CD3T-positive cells and CD3T-negative cells (CD16 positive cells) were treated with rh-IL-2
(1000 U/ml) for 18-20 hours before they were tested for cytotoxicity using a standard 4-hour 51 Cr release assay against the OSCSCs (C) and K562 (D) cell lines. The lytic units 30/10 6 cells were determined using the method described in Materials and Methods, for OSCSCs and K562 respectively. The supernatant was harvested from the culture and IFN-γ secretion was determined using single ELISA (E). NK cells, CD3T, CD4T, CD8T, and γδT cells were purified from PBMC as described in materials and methods, and were activated with rh-IL-2 for 18-20 hours, before they were tested for cytotoxicity using a standard 4-hour 51 Cr release assay against the OSCSCs (F). The lytic units 30/10 6 cells were determined using the inverse number of lymphocytes required to lyse 30% of OSCSCs x 100 (F). NK and T cells were purified from PBMCs as described in materials and methods, NK cells were treated as described in Materials and Methods. T cells were activated with anti-CD3 (1 ug/ml) and anti-CD28 (3 ug/ml) 18-20 hours before they were cultured with autologous OCs, and expanded lymphocytes were counted manually using microscope day 4 after the culture (G). NK and T cells were purified and cultured with OCs, and counted on day 4 as described in Fig. S2G, fold expansion of lymphocytes expanded by the OCs were divided by fold expansion of lymphocytes without the OCs (H). To generate osteoclasts, monocytes were cultured in medium containing M-CSF (25 ng/ml) and RANKL (25 ng/ml) for 21 days, K562 tumor cell lines were cultured as described in materials and methods. Highly purified NK cells (1x10 6 cells/ml) were treated with the combination of IL-2 (1000 U/ml) and anti-CD16mAb (3ug/ml) for 18 hours before they were co-cultured with K562 and autologous OCs in the presence of sAJ2 bacteria at 1:2:4 ratios (OCs:NK:sAJ2), respectively. Cells from the cultured were counted manually using microscope on day 6, 10 and 13 (A). The osteoclasts generated as described in Fig. S1A and K562 tumor cells lines were irradiated at 40 grays (Gy) as described in the materials and methods. NK cells were purified and treated as described in Materials and Methods, before they were co-cultured with irradiated K562 and irradiated autologous OCs in the presence of sAJ2 bacteria at 1:2:4 ratios (OCs:NK:sAJ2), respectively. Cells from the cultured were counted manually using microscope on day 6, 10 and 13 (B). NK cells were purified and cultured with OCs and K562 as described in Fig. S1A, cytotoxicity of lymphocytes co-cultured for 6 days was determined using a standard 4-hour 51 Cr release assay against OSCSCs. The lytic units 30/10 6 cells were determined using method described in Fig. S2F (C). NK cells were purified and cultured with OCs and K562 as described in Fig. S3B, cytotoxicity of lymphocytes co-cultured for 6 days was determined using a standard 4-hour 51 Cr release assay against OSCSCs. The lytic units 30/10 6 cells were determined using method described in Fig. S2F (D). NK cells were purified and cultured with OCs and OSCSCs as described in Fig. S3A, cytotoxicity of lymphocytes cocultured for 6 days was determined using a standard 4-hour 51 Cr release assay against OSCSCs. The lytic units 30/10 6 cells were determined using method described in Fig. S2F (E). NK cells were purified and cultured with OCs and K562 as described in Fig. S2A, supernatant was harvested on day 3, 6, 7, 10 and 13, and IFN-γ secretion was determined using single ELISA (F). NK cells were purified and cultured with irradiated OCs and irradiated K562, as described in Fig. S3B, supernatant was harvested on day 3, 6, 7, 10 and 13, and IFN-γ secretion was determined using single ELISA (G). NK cells were purified and cultured with OCs and OSCSCs as described in Fig. S3A, supernatant was harvested on day 1, 3, 6 and 8, and IFN-γ secretion was determined using single ELISA (H).  Fig. S4C   Fig, S4D  Fig. S4E   Fig. S4F   Fig. S4G   Fig. S4H   Fig. S4I  Fig. S4J  Fig. S4K   Fig. S4.

Small fraction of contaminating T cells within purified NK cells from cancer patient expand faster and crowd out NK cells likely due to decreased NK cell function.
Freshly purified NK cells from a healthy donor and a pancreatic cancer patient were treated and co-cultured with monocyte-derived allogeneic (from different healthy donor) osteoclasts as described in Fig. S1A.  Freshly purified NK cells were treated and co-cultured with monocyte-derived autologous osteoclasts as described in Fig. S1A, day 9 after the cultures, expanded NK cells were frozen.
NK cells were thawed and treated with rh-IL-2 (1000 U/ml), day 6 and day 9 after the culture, the supernatant was harvested and IFN-γ secretion was determined using single ELISAs (A). NK cells were cultured as described in Fig. S1A, and the cytotoxicity of lymphocytes day 6 and 9 after the culture was determined using a standard 4-hour 51 Cr release assay against OSCSCs. The lytic units 30/10 6 cells were determined using method described in Fig. S2F  antibodies. Isotype control antibodies were used as a control (Fig. S6).

Fig. S7. Cytokines, chemokines and growth factors and ligands secreted by primary and osteoclast-expanded NK cells
Highly purified NK cells and monocytes were obtained from peripheral blood mononuclear cells (PBMCs) of healthy donors and NK cells were treated (1x10 6 cells/ml) with IL-2 (1000 U/ml) for 18 hours before the supernatant was harvested. To generate osteoclasts, monocytes were cultured in alpha-MEM media containing M-CSF (25 ng/ml) and RANKL (25 ng/ml) for 21 days. For expansion, purified NK cells (1x10 6 cells/ml) were treated with the combination of IL-2 (1000 U/ml) and anti-CD16mAb (3ug/ml) for 18 hours before they were co-cultured with autologous osteoclasts in the presence of sAJ2 bacteria at 1:2:4 ratios (OC:NK:sAJ2), respectively.
The supernatant was harvested after 6 days of co-culture and multiplex assay was used to determine cytokines (A), chemokines (B) and growth factors (C) levels secreted by the primary and expanded NK cells. Highly purified T cells and monocytes were obtained from peripheral blood mononuclear cells (PBMCs) of healthy donors and T cells were treated (1x10 6 cells/ml) with IL-2 (100 U/ml) and anti-CD3 (1ug/ml) for 18 hours before they were co-cultured with autologous osteoclasts in the presence of sAJ2 bacteria at 1:2:4 ratios (OC:T cells:sAJ2), respectively. The cells were analyzed for CD3, CD16 and CD56 on day 9 after the culture. Anti-CD3 treated T cells did not lose Forward and side scatter in the absence of NK cells (Fig.   S8).