ALBERT

Description: Natural killer (NK) cells hold great promise as a source for allogeneic cell therapy against hematological malignancies, including acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS). NK cell recognition of allogeneic tumors is strictly regulated by inhibitory killer cell immunoglobulin-like receptors (KIR) that bind to groups of HLA class I alleles. However, KIR expression on NK cells is highly diverse due to variation in gene content, polymorphism and copy number in combination with stochastic expression of the protein in individual cells. As a consequence, the number of efficacious allogeneic NK cells within a product isolated and expanded from random donors can vary a great deal and potentially be negligible. Our group has defined a repertoire of NK cells that is uniquely found in individuals with prior exposure to cytomegalovirus (CMV). Interestingly, these cells were shown to share many attributes usually reserved for adaptive immune cells including increased longevity, memory, and serial killing. We have previously described a 14-day protocol to enrich for adaptive NKG2C+CD57+ NK cells from CMV sero-positive donors with a homogenous expression of one single self-HLA specific KIR (self KIR). Here, we present new data on the GMP-transfer and clinical scale-up of this protocol, providing a route to off-the-shelf adaptive NK cell therapy for refractory high-risk AML/MDS. By screening 〉250 healthy donors, we first established the prerequisites for robust expansion of adaptive NK cells from peripheral blood of CMV+ donors and found that donors with 〉15% pre-existing adaptive NK cells showed efficient expansion of adaptive NK cells (Figure 1A-B). Apheresis products from a pool of pre-screened third-party donors are currently being collected for GMP freezing and use in an off-the-shelf setting intended for HLA mismatched patients to maximize alloreactivity by "missing" self. The GMP compatible protocol led to a robust expansion of clinical doses of self-KIR+ adaptive NK cells, with an average frequency of 60% self-specific KIR+ cells in the end product (Figure 1C-D). Based on the expression of self-KIR the expanded cells were educated, showing large dense-core granules and high levels of granzyme B. Further characterization in CyTOF using 36 phenotypic and functional markers revealed a highly activated state with high expression of DNAM-1 and CD2, which are critical for NK cell adhesion and function (Figure 1E). Notably, the expanded adaptive NK cells were negative for the HLA-E binding inhibitory receptor NKG2A, which is a major check point for T- and NK-cell based therapies. A microchip single-cell imaging platform revealed high serial killing capacity of the expanded adaptive NK cells. In flow cytometry-based killing assays and long-term killing assays this enhanced capacity for serial killing correlated with highly efficient targeting of mismatched PHA blasts (Figure 1F), tumor cell lines (Figure 1G), and MDS blasts. These pre-clinical data demonstrate the feasibility of off-the-shelf therapy with a non-engineered and yet highly specific NK cell population, representing the first route to clinical testing of missing self-recognition as it was originally defined over thirty years ago. Disclosures Valamehr: Fate Therapeutics Inc.: Employment. Alici:Vycellix: Consultancy, Equity Ownership, Patents & Royalties, Research Funding; Intellia: Membership on an entity's Board of Directors or advisory committees. Ljunggren:Fate Therapeutics: Patents & Royalties; Vycellix: Equity Ownership, Membership on an entity's Board of Directors or advisory committees. Malmberg:Fate Therapeutics Inc.: Consultancy, Research Funding.

Description: Key Points Activated NK cells display heterogeneity in their cytotoxic responses that justifies grouping them into 5 distinct classes of NK cells. A subpopulation of particularly active “serial killer” NK cells deliver their lytic hits faster and release more perforin in each hit.

Description: Although adoptive transfer of NK cells with IL-2 can induce complete remissions in 30-50% of patients with refractory AML, efficacy is limited by IL-2 mediated induction of Tregs and by lack of antigen specificity. Thus we generated a 161533 trispecific killer engager (TriKE) molecule containing an anti-CD16 scFv to engage NK cells, an anti-CD33 scFv to engage myeloid targets (including myelodysplastic syndrome [MDS]), and a modified IL-15 linker. We have previously shown that this molecule is superior to a 1633 bispecific killer engager (BiKE) in killing of AML targets and that it promotes enhanced survival and in vivo expansion of NK cells. We questioned the mechanism for the increased potency of the 161533 TriKE and if the TriKE could activate the dysfunctional and suppressed NK cells found in patients with MDS. Cryopreserved mononuclear cells (obtained from the NMDP Sample Repository) collected from 8 patients with advanced MDS were tested to investigate how the TriKE might enhance functionality in this setting. Previously we reported that MDS patients have significantly decreased frequencies of NK cells due to increased CD33+ myeloid derived suppressor cells. Our 161533 TriKE enhanced the function of NK cells derived from MDS patients against acute promyelocytic CD33+ leukemia HL-60 tumor targets (Figure 1A), when compared to 1633 BiKE, in a flow cytometry assay measuring NK cell degranulation (% CD107a: 41.8±3.8 vs. 30.3±3.2, p=0.004), and inflammatory cytokine production (% IFNg: 40.7±5.0 vs. 30.0±4.9, p=0.009; % TNFa: 36.9±5.5 vs. 28.4±4.8, p=0.009). State of the art microchip-based live cell imaging was then employed to evaluate NK cell function and contact-to-target dynamics (Figure 1B). Briefly, resting NK cells and HL-60 target cells were stained with distinct dyes and then co-incubated in the presence of BiKE or TriKE in microwells at 37°C and 5% CO2. Cells were then imaged for 12 hours using a Zeiss 880 microscope and analyzed by Matlab. In contrast to those incubated with BiKE, NK cells cultured in the presence of TriKE had augmented cytotoxicity (37%±6% vs. 59%±6%, p=0.02) and killed their targets remarkably faster (time to first target kill = 148±30 min vs. 75±26 min, p〈 0.0001). In addition, NK cell serial killers were more common in the presence of TriKE compared to BiKE (number of killed targets ≥3: 18%±7% vs. 9%±1%, p=0.04). Having shown the robust killing dynamics of TriKE primed NK cells, we next designed an in vivo dose escalation study to evaluate HL-60 tumor control in our xenogeneic mouse model. NSG (NOD scid gamma) mice were conditioned with a sub-lethal dose of radiation (275 cGy) and engrafted with 750,000 HL-60luc cells, which allow for tracking of tumor growth using bioluminescent imaging (BLI). Mice were infused with 1 million fresh healthy donor human NK cells and treated daily with recombinant IL-15 (5 ug/injection), 20-200 ug of 161533 TriKE, or left untreated. Higher doses of TriKE provided better tumor control (6.3x109±2.9x109 [200 ug 1615133] vs. 1.6 x1010±2.7x109 [20 ug 161533] p/sec/cm2/sr) demonstrating dose responsiveness at days 14 and 21 This was not due to increased proliferation of effectors (NK cell numbers) induced by the IL-15 segment of the TriKE as the absolute PBNK cell counts in mice were not different across the TriKE concentrations. However, the day 14 and 21 absolute PBNK counts were higher in all of the TriKE groups when compared to the IL-15 treated mice, including in the 20 ug 161533 TriKE group which best matched the 5 ug IL-15 dose. These data indicated that the TriKE molecule mediates superior NK cell expansion or maintenance in vivo. Taken together the in vivo data suggests that 161533 TriKE not only mediates tumor control by inducing NK cell proliferation and survival in a methodology that exceeds signals provided by IL-15 alone, but also increases tumor killing through enhanced killing kinetics. In summary, we have shown that the 161533 TriKE can rescue dysfunctional NK cells suppressed in patients with MDS and can mediate potent in vivo tumor killing presumably through better NK cell maintenance and enhanced killing kinetics This is of translational importance and demonstrates that the cancer induced immune suppression is reversible by the activating signals induced by the TriKE molecule. The 161533 TriKE represents a promising modality to maximizing NK cell based immunotherapies against MDS and AML and will be in phase I clinical testing the first half of 2017. Disclosures Cooley: Fate Therapeutics: Research Funding. Vallera:Oxis Biotech: Consultancy, Membership on an entity's Board of Directors or advisory committees. Miller:Fate Therapeutics: Consultancy, Research Funding; Oxis Biotech: Consultancy, Other: SAB.