An exploration of natural killer cell diversity and education as a determinant of responsiveness against pancreatic adenocarcinoma

Abstract

Natural Killer (NK) cells are critical mediators of cancer immunotherapy but understudied for the treatment of pancreatic cancer (PDAC). Driven by the activation or inhibition of specific receptors, NK cells can target tumours. NK cell receptor-ligand pairs are germline-encoded and highly polymorphic, but measurable, which may enable selection of specific populations for precision therapy. NK cell recognition is dependent on the integration of heterogeneous signals for activation and inhibition; defining key receptors contributing to this signaling could inform precise direction of NK cells against PDAC. I hypothesized that ligands expressed by PDAC determine NK cell responsiveness, and that NK cells can be selected with “inclusion” and “exclusion” criteria to enable effective and comprehensive treatment of PDAC. Using a 27-parameter flow cytometry panel, I assessed the expression of NK cell-associated receptors on circulating NK and T cells. NK cells with markers associated with maturation (i.e. KIR, NKG2C, and CD16) also exhibited increased expression of CD8, and greater degranulation (CD107a) against the missing-self target K562. CD8+ T cells acquired expression of several NK cell associated receptors as they enter a more terminally differentiated state, with effector memory and TEMRA subsets expressing NKG2A, KIR, and CD16. I also developed a humanized mouse model that both recapitulates the human HLA landscape and produces human IL-15 to help support primary human NK cells in vivo. Using bioinformatics, I found that primary PDAC tumours express ligands for NK cells, especially those known to bind the activating NKG2D receptor. After exposure of PDAC cells to immune cells or inflammation, I measured dynamic changes in expression of both activating and inhibitory ligands when compared to a resting phenotype. In vitro co-culture assays revealed a redundancy in the activating receptors engaged in NK:PDAC interactions, but that HLA-KIR signaling dominantly interrupted anti-PDAC activity. Using NK-competent humanized mice, I showed that adoptively transferred, unselected NK cells slowed tumour growth in a dose-dependent manner, but NK cells selected to avoid HLA I-driven inhibition were ideal. These results indicate that selection of allogeneic NK cells to remove inhibitory signals and support activating receptor engagement augments NK:PDAC killing.