An efficient feeder-free and chemically-defined expansion strategy for highly purified natural killer cells derived from human cord blood

Abstract

Introduction: Natural killer cells (NKCs) are immune cells that can attack cancer cells through the direct recognition of ligands without prior sensitization. Cord blood-derived NKCs (CBNKCs) represent a promising tool for allogenic NKC-based cancer immunotherapy. Efficient NKC expansion and decreased T cell inclusion are crucial for the success of allogeneic NKC-based immunotherapy without inducing graft-versus-host reactions. We previously established an efficient ex vivo expansion system consisting of highly purified-NKCs derived from human peripheral blood. Herein, we evaluated the performance of the NKC expansion system using CB and characterized the expanded populations.

Methods: Frozen CB mononuclear cells (CBMCs), with T cells removed, were cultured with recombinant human interleukin (rhIL)-18 and rhIL-2 under conditions where anti-NKp46 and anti-CD16 antibodies were immobilized. Following 7, 14, and 21 days of expansion, the purity, fold-expansion rates of NKCs, and the expression levels of NK activating and inhibitory receptors were assessed. The ability of these NKCs to inhibit the growth of T98G, a glioblastoma (GBM) cell line sensitive to NK activity, was also examined.

Results: All expanded T cell-depleted CBMCs were included in over 80%, 98%, and 99% of CD3^-CD56⁺ NKCs at 7, 14, and 21 days of expansion, respectively. The NK activating receptors LFA-1, NKG2D, DNAM-1, NKp30, NKp44, NKp46, FcγRIII and NK inhibitory receptors TIM-3, TIGIT, TACTILE, NKG2A were expressed on the expanded-CBNKCs. Two out of three of the expanded-CBNKCs weakly expressed PD-1, yet gradually expressed PD-1 according to expansion period. One of the three expanded CBNKCs almost lacked PD-1 expression during the expansion period. LAG-3 expression was variable among donors, and no consistent changes were identified during the expansion period. All of the expanded CBNKCs elicited distinct cytotoxicity-mediated growth inhibition on T98G cells. The level of cytotoxicity was gradually decreased based on the prolonged expansion period.

Conclusions: Our established feeder-free expansion system yielded large scale highly purified and cytotoxic NKCs derived from human CB. The system provides a stable supply of clinical grade off-the-shelf NKCs and may be feasible for allogeneic NKC-based immunotherapy for cancers, including GBM.

Keywords: Allogeneic NKC; Cancer immunotherapy; Cell-based immunotherapy; Cord blood; Glioblastoma.

Fig. 1

Ex vivo expansion of human CBNKCs stimulated by NK activating receptor antibodies and defined cytokines. (a) Representative morphological patterns of the expanded cells on days 0, 7, 14, and 21. (b) The cellular proportion (CD3⁻CD56-: non-T, non-NKC, CD3⁻CD56⁺: NKC, CD3⁺CD56-: T cells; and CD3⁺CD56⁺: NKT cells) was analyzed by flow cytometry. The dot plots on days 0, 7, 14, and 21 of three donors (CBMC#1, #2, and #3) are depicted. The x-axis and y-axis show the fluorescence intensity of cells stained with FITC-conjugated anti-CD3 and APC-conjugated anti-CD56 antibodies, respectively. (c) The kinetics of the total cell viability (left), NKC purity (center), and NKC expansion ratio (right) during the expansion period. Blue lines and circles, red lines and squares, green lines and triangles indicate CBMC#1, #2, and #3, respectively.

Fig. 2

NK activating and inhibitory receptor expression on the expanded CBNKCs. The NK activation receptors, LFA-1, NKG2D, DNAM-1, NKp30, NKp44, NKp46, FcγRIII, and NK inhibitory receptors, PD-1, LAG-3, TIM-3, TIGIT, TACTILE, NKG2A, and KIR were evaluated by flow cytometry. Histogram data of each receptor on the expanded CBNKCs gated on the CD56 positive population. Days 7, 14, and 21 are arranged to upper, middle, and lower panels in the NK activating (tops) and inhibitory receptor (bottoms) regions. (a), (b), (c) indicate CBMC#1, #2, and #3, respectively.

Fig. 3

Kinetics of NK activating and inhibitory receptor expression on the expanded CBNKCs in the expansion process. The frequency and mean fluorescent intensity (MFI) of NK activating and inhibitory receptors on day 7, 14, and 21 are shown. Blue lines and circles, red lines and squares, green lines and triangles indicate CBMC#1, #2, and #3, respectively. Data are expressed as the mean ± SE of triplicate experiments. Statistical differences were determined by a two-way ANOVA followed by a Tukey's test. ∗∗∗∗P < 0.0001; ∗∗∗P < 0.001; ∗∗P < 0.01; ∗P < 0.05, ns: not significant.

Fig. 4

Cytotoxicity-mediated growth inhibition effects of T98G GBM cells by the expanded CBNKC-including populations. (a) Images depict the real-time growth inhibition of a NK activity-sensitive GBM cell line, T98G, by the expanded CBNKC-included population. The left, middle, and right panel lines depict the 7-, 14-, 21-day cultures, respectively. The top, intermediate, and bottom lines depict CBMC#1, #2, and #3, respectively. The X- and Y-axes respectively depict the co-culture time and relative normalized cell index of each time point divided by the cell index of the co-culture starting point. Data represent the mean ± SD of triplicate experiments. The red, light blue, and blue lines indicate target (T98G) only; the effector (the expanded CBNKC-including population) to target (E:T) ratio was 0.5:1 and 1:1, respectively. (b) The cytotoxicity of the expanded CBNKC-including population on T98G cells. The X- and Y-axes respectively depict culture days for the CBNKC-including population and cytotoxicity. Blue lines and circles, red lines and squares, and green lines and triangles indicate CBMC#1, #2, and #3, respectively. Data represent the mean ± SE of triplicate experiments. Statistical differences of the integrated data of three donors were determined by a two-way ANOVA followed by a Tukey's test. ∗∗∗∗P < 0.0001.