Large-scale ex vivo expansion and characterization of natural killer cells for clinical applications

Abstract

Background aims: Interest in natural killer (NK) cell-based immunotherapy has resurged since new protocols for the purification and expansion of large numbers of clinical-grade cells have become available.

Methods: We have successfully adapted a previously described NK expansion method that uses K562 cells expressing interleukin (IL)-15 and 4-1 BB Ligand (BBL) (K562-mb15-41BBL) to grow NK cells in novel gas-permeable static cell culture flasks (G-Rex).

Results: Using this system we produced up to 19 × 10(9) functional NK cells from unseparated apheresis products, starting with 15 × 10(7) CD3(-) CD56 (+) NK cells, within 8-10 days of culture. The G-Rex yielded a higher fold expansion of NK cells than conventional gas-permeable bags and required no cell manipulation or feeding during the culture period. We also showed that K562-mb15-41BBL cells up-regulated surface HLA class I antigen expression upon stimulation with the supernatants from NK cultures and stimulated alloreactive CD8 (+) T cells within the NK cultures. However, these CD3 (+) T cells could be removed successfully using the CliniMACS system. We describe our optimized NK cell cryopreservation method and show that the NK cells are viable and functional even after 12 months of cryopreservation.

Conclusions: We have successfully developed a static culture protocol for large-scale expansion of NK cells in the gas permeable G-Rex system under good manufacturing practice (GMP) conditions. This strategy is currently being used to produce NK cells for cancer immunotherapy.

Figures

Figure 1

Ex vivo expansion of peripheral blood NK cells. (A) G-Rex100s were seeded with PBMC containing 2 × 106, 4 × 106 or 8 × 106 NK cells with irradiated K562-mbIL15-41BBL feeders at a 1:10 ratio of NK cells to K562-mb15-41BBL cells. Cells were counted on days 6, 7 and 8 of culture. Aliquots of cells from each G-Rex100 were analyzed by flow cytometry using anti-CD56 and anti-CD3 monoclonal antibodies, and the numbers of CD56+ CD3− NK cells (solid black symbols) and CD3+ CD56−T cells (open gray symbols) were plotted. *P < 0.001 difference in NK cell numbers between 4 × 106 and 2 × 106 and 8 × 106 and 2 × 106 groups (n = 5). (B) Expansion of CD56+ CD3− NK cell on days 6, 7 and 8 of culture. (C) Glucose levels were analyzed on days 5–8 of culture. Media were changed when glucose levels dropped below 100 mg/dL. (D) Increasing frequency of CD56+ CD3− NK cells over 8 days of culture. (E) Expression of NK-specific receptors on day 8. (F) Cytotoxicity of day 8-expanded NK cells against U266 MM, Raji Burkitt's lymphoma and K562 erythroleukemia cells. Result from five different donors are shown.

Figure 2

G-Rex yields higher numbers of NK cells and requires fewer culture manipulations than cultures in bags. (A) Schematic view of NK cell expansion in bags and G-Rexes. Less cell processing is required to grow a similar number of cells in G-Rexes. (B) NK cells from three different donors were seeded in the 197-mL bags and G-Rex100s. Bags were seeded with PBMC containing 2 × 106 NK cells and 2 × 107 K562-mb15-41BBL cells in a 40-mL starting volume and cultures required feedings with 40 mL fresh media and cytokine every other day, i.e. days 2, 4, 6 and 8, resulting in 200 mL of total volume by day 8. G-Rex100s were seeded with PBMC containing 2 × 106 NK cells and 2 × 107 K562-mbIL15-41BBL in 400 mL medium and were not subsequently fed. Cells were counted and analyzed by flow cytometry for percentage CD56+ CD3− cells (n = 3). (C) Fold expansion of CD56+ CD3− cells in G-Rexes and bags (n = 3). *P < 0.5.

Figure 3

Proliferation and frequency of the K562-mb15-41BBL feeder cells in the final NK products. (A) Proliferation of K562-mbIL15-41BBL cells was analyzed in a Click-iT EdU Alexa Fluor 647 flow cytometry assay. Irradiated and non-irradiated positive control cells (1 × 106) were incubated for 1 h or overnight (16–24 h) with 10 mm EdU. EdU was detected with a Click-iT reaction cocktail containing Alexa Fluor 647 azide. The experiments were performed at least three times. (B) K562-mb15-41BBL cell gating in the final NK cell product was established by gating K562-mb15-41BBL cells (GFP+) alone and spiking 1000 and 5000 of those cells into 106 cells of the final NK product. The spiking experiments were performed three times and 12 NK products were analyzed. Representative results are shown.

Figure 4

Alloreactive CD8+ T cells with an activated phenotype dominated the T-cell component of NK cultures. (A) HLA class I expression on K562-mb15-41BBL was induced with IFN-γ/TNF-α and NK culture supernatants. MFI, mean fluorescent intensity. (B) NK cell cultures were gated on CD3+ cells and analyzed for expression of CD4, CD8, γθ-T cell receptor (TCR) and intracellular Foxp3. (C) CD8+ T cells were stained for the surface expression of CD69, HLA-DR and CD86, and intracellular expression of granzyme B (GrB). The percentage of double-positive cells is depicted. The results are representative of three independent experiments. (D) Alloreactivity of CD8+ CD56− T cells purified by magnetic bead separation and flow sorting from NK cultures of the donor A*02,03; B*07,41; C*07,17 (class I mismatched to K562-mb15-41BBL cells). The cytotoxic specificity of the CD8+ T cells was analyzed in a 6-h chromium51 release assay using (1) K562-mbIL15-41BBL cells stimulated with supernatants from PBMC cultured with K562-mb15-41BBL cells (Sup-PMBC-K562), (2) K562-mb15-41BBL cells, (3) the NK cell donor (autologous) LCL and (4) HLA Cw*05-matched LCL (n = 3). (E) A cytotoxicity blocking assay (n = 3) with anti-HLA-ABC (clone W6/32) and HLA-DP, -DQ, -DR (clone CR3/43) antibodies was performed at an E:T ratio of 20:1. Targets as described in (D). *P < 0.5.

Figure 5

Stability of cryopreserved TC-NK products. (A) Cryopreserved NK cells were thawed in a 37°C waterbath and washed in pre-warmed 1 × phosphate-buffered saline (PBS). Live and dead cells were evaluated by flow cytometry using 7-AAD. (B) NK cells from three donors (n = 3) were analyzed in a cytotoxicity assay using K562 cells as targets. Fresh or cryopreserved NK cells were used as effectors. Frozen NK cells were used immediately after thawing or were rested overnight in SCGM medium supplemented with 10% FBS and 10 IU/mL IL-2. *P < 0.5.