Phenotypic characterization of chimeric ilt receptor natural killer (CIR-NK) cells targeting HLA-G+ tumors for allogeneic donor identification Free

Introduction

HLA-G is an immunosuppressing MHC-IB that acts through ILT2 and ILT4 receptors to block allogeneic immune responses at the placenta. HLA-G is derepressed in ~50% of tumors to evade immune responses. Chimeric ILT Receptors (CIR) were developed to better target HLA-G in tumors by using the HLA-G binding domains of ILT receptors to target the diverse isoforms of HLA-G with increased specificity than an antibody-derived scFv. In transduced allogeneic NK cells, CIR demonstrated remarkable anti-tumor activity against HLA-G⁺ AML, multiple myeloma and solid tumor models.

Our published results indicated that activation domains 4-1BB, DAP10 and MyD88 signaling domains significantly increased CIR-NK cell functional persistence for more than one month. Further, it was noted that functional persistence varied between donor cells with some being highly resistant to NK cell exhaustion. Identification of donors that generate engineered NK cells with high growth potential and persistent cytotoxicity is key to the success of allogeneic NK cells as an off-the-shelf cell therapy. To identify the signatures of highly productive blood donors, extensive phenotyping was performed with naive NK cells isolated from a collection of heathy blood donors, post-CIR transduction and after HLA-G⁺ tumor cell exposure.

Methods

CD56⁺ NK cells were purified from 11 healthy blood donors, cultured with IL-15 and IL-2 and activated with a feeder-free cocktail of an immobilized cytokine and activating ligand. Activated NK cells were transduced with γ-retroviruses directing multi-cistronic expression of CIR, soluble IL-15, and a DCD19 marker and co-transduced with RFP for fluorescent detection. Activation domains embedded in the CIR contained CD3ζ, 4-1BB.CD3ζ, 4-1BB.DAP10 or 4-1BB.DAP10.TLR2. Coculture assays were performed with GFPFluc-expressing HLA-G⁺ AML and solid tumor cell lines and monitored using an IncuCyte or CellInsight CX7 incubating microscope. Flow cytometric analysis was performed at days 0 and 5 before viral transduction, at day 13 preceding coculture with tumor cells, at day 21 following a one-week coculture with OE19, and at day 33 following four rounds of consecutive 5-day OE19 exposure.

Results

Cells from all donors expanded similarly following initial activation but marked differences were noted between donors during expansion post-transduction without respect to the CIR construct introduced. CIR-NK cells from all donors were transduced similarly and showed similar viability at day 8 and day 14. Consistently, the donors that showed the most persistent cytotoxicity after repeated stimulation with target cells also displayed superior expansion during coculture and prior to coculture (R²=0.60, p=0.0089). Interestingly, some donor's NK cells were cytotoxic during early cocultures but showed signs of NK cell exhaustion following repeated target exposure.

Flow cytometry analysis was performed with 5 panels that quantitated the expression of 26 proteins marking NK cell activation, maturation status, exhaustion, cytotoxicity, innate receptors, and inhibitors along with markers of CIR transduction. NK cells from each donor demonstrated a similar activation profile (ICAM-1, CD25, DNAM induction) prior to transduction with variation in the degree of induction of granzyme B, and CD57 and repression of NKp80, both are markers of NK cell maturation. Of note, 2 of the 11 donors that demonstrated superior long-term cytotoxicity and NK expansion expressed higher levels of CD25, NKG2D, CD16 (R²=0.60, p=0.0094), CD57 (R²=0.44, p=0.0362) and granzyme A (R²=0.63, p=0.0062) while the levels of NKG2A (R²=0.44, p=0.0375), ILT2, IFN-γ and TNF-α (R²=0.44, p=0.037) are lower than the other 8 donors.

When grouped between best 5 and worst 6 performing donors, donor NK cells that had the highest degree of CD16 (p=0.026) and CD161 (p=0.01) positivity prior to activation (day 0) led to CIR-NK cells that displayed poor functional persistence. However, during coculture, the transduced cells from the best performing donors maintained the highest levels of CD16 (p=0.0013 day 19, p=0.027 day 33).

Conclusions

These results indicate that ‘good’ donor NK cells can be identified by a combination of growth, cytotoxicity and expression signature possibly including CD161 and CD16 prior to manufacture at scale. Other markers including CD57, NKG2A and granzymes may be useful for phenotypic immune monitoring of NK cell persistence during clinical trial.