Human ascites does not affect NK-92 cytotoxicity and intraperitoneal compared with intravenous administration of NK-92 shows superior survival in human ovarian cancer xenografted mice

Ovarian cancer (OC) is the most lethal gynecological malignancy. Frontline treatment involves cytoreductive surgery and chemotherapy, with alternative therapies reserved for those who fail to respond. Cellular immunotherapy is an option to treat relapsed disease and while cytotoxic natural killer (NK) cells are involved in tumor surveillance and clearance, NK function in many cancers is impaired, making allogeneic NK cell lines such as NK-92 an attractive alternative. Here, we sought to determine if there is an advantage of intraperitoneal (IP) over intravenous (IV) delivery of NK-92 cells in an ovarian cancer xenograft model and also investigated the effect of primary human ascites fluid on NK-92 function. Primary tumor and ascites fluid samples were obtained from patients undergoing standard of care treatment. Cytotoxicity of NK-92 cells was evaluated by chromium release assay. NOD-SCID gamma null mice were injected IP with 2 × 106 SKOV-3 to generate a murine xenograft model. For in vivo trafficking, NK cells were labelled with DiR and monitored by IVIS up to 72 hours, after which organs were harvested and imaged. NK-92 was highly cytotoxic against primary OC cells (37.4 ± 2.1% at 10:1 effector:target ratio n = 9). Pre-treatment of NK-92 in ovarian cancer ascites fluid for 4 or 18 hours had no significant effect on NK-92 viability compared to culturing with XVIVO media without IL-2, and cytotoxicity against SKOV-3 was not impaired. In a murine xenograft model, treatment with six doses of NK-92 (total of 15 × 10e6 cells per dose) significantly increased overall survival in animals that received NK cells via IP (p = 0.009) or a combination of IP and IV (p = 0.05) injections. Treatment with NK-92 cells delivered IV did not improve survival compared with untreated controls (p = 0.665). Tracking studies showed that IV delivery of NK-92 resulted in rapid migration of the NK cells to lungs and spleen, but after 72 hours the cells were detected in the kidney, lung and bone. IP delivered NK-92 cells remained within the peritoneal cavity, localised to areas with tumor cells and were only detected in the liver and spleen by 72 hours. These studies suggest that NK-92 delivered IP may improve survival in OC due to enhanced localisation within the peritoneal cavity where a higher effector to target ratio is present and cytotoxicity is maintained.