Donor-dependent amplification capacity and phenotype of ex vivo–expanded iNKT cells.

e14526 Background: Invariant natural killer T (iNKT) cells combine rapid innate-like effector function with T-cell specificity, making them a compelling platform for CAR engineering. CAR-iNKT cells show potent antitumor activity, improved tumor infiltration, and a favorable safety profile with low GvHD risk, supporting their potential as off-the-shelf therapies, particularly for solid tumors. However, their clinical translation remains challenging due to low natural abundance, manufacturing complexity, and difficulties in achieving robust in vivo expansion and persistence. Methods: PBMCs from 14 healthy donors were used for iNKT cell expansion. iNKT cells were isolated using anti-iNKT microbeads (Miltenyi Biotec), stimulated with X-ray–irradiated, α-galactosylceramide–pulsed PBMCs, and cultured in RPMI-1640 with 10% FBS and IL-2. Differentiation status was assessed by CD45RA/CCR7 flow cytometry. Cells were lentivirally transduced with CD19 or GD2 CARs and further expanded. Antitumor activity was evaluated against lymphoma (Ramos, Jurkat, Raji) and neuroblastoma (SH-SY5Y) cell lines. Statistical analysis used Student's t test. Results: The proliferative capacity of iNKT cells from 14 donors was evaluated after ~3 weeks of in vitro culture, revealing marked interindividual variability ranging from 210-fold to 42,200-fold expansion under identical conditions. Six donors exhibited high amplification ( > 1,000-fold), while eight showed low amplification ( < 1,000-fold). Flow cytometric analysis demonstrated a significantly higher proportion of CD4⁺ cells in PBMCs from high-amplification donors compared with low-amplification donors (high/low = 1.43, P = 0.02). Although trends were observed for individual naïve- and central memory-like iNKT subsets, significant differences emerged when subsets were combined: naïve/central memory cells were enriched in the high group (high/low = 1.78, P = 0.02), whereas effector memory/effector cells were higher in the low group (high/low = 0.71, P = 0.02). No difference in CD4⁺/CD8⁺ ratio was observed. Expansion capacity was unrelated to age, sex, or BMI. Additionally, iNKT cells from selected donors were successfully transduced with lentiviral CAR constructs targeting CD19, GD2, MSLN, or other antigens and demonstrated potent antitumor activity against a range of hematologic and solid tumor cell lines, outperforming CAR-T cells in vitro. However, donor-dependent effects on CAR transduction efficiency and cytotoxicity require further validation in larger cohorts. Conclusions: iNKT cells exhibit marked donor-dependent variability in ex vivo expansion associated with immunophenotypic differences in donor PBMCs, highlighting donor selection as a critical factor for CAR-iNKT manufacturing. Further studies in larger cohorts are warranted to define predictive biomarkers of expansion efficiency and therapeutic potency.