Ex vivo generation of highly purified and activated natural killer cells from human peripheral blood

Abstract

Adoptive immunotherapy using natural killer (NK) cells has been a promising treatment for intractable malignancies; however, there remain a number of difficulties with respect to the shortage and limited anticancer potency of the effector cells. We here established a simple feeder-free method to generate purified (>90%) and highly activated NK cells from human peripheral blood-derived mononuclear cells (PBMCs). Among the several parameters, we found that CD3 depletion, high-dose interleukin (IL)-2, and use of a specific culture medium were sufficient to obtain highly purified, expanded (∼200-fold) and activated CD3(-)/CD56(+) NK cells from PBMCs, which we designated zenithal-NK (Z-NK) cells. Almost all Z-NK cells expressed the lymphocyte-activated marker CD69 and showed dramatically high expression of activation receptors (i.e., NKG2D), interferon-γ, perforin, and granzyme B. Importantly, only 2 hours of reaction at an effector/target ratio of 1:1 was sufficient to kill almost all K562 cells, and the antitumor activity was also replicated in tumor-bearing mice in vivo. Cytolysis was specific for various tumor cells, but not for normal cells, irrespective of MHC class I expression. These findings strongly indicate that Z-NK cells are purified, expanded, and near-fully activated human NK cells and warrant further investigation in a clinical setting.

FIG. 1.

Depletion of CD3⁺ cells is essential to natural killer (NK) cell–specific expansion. Peripheral blood mononuclear cells (PBMCs) were isolated by gradient centrifugation, using Ficoll-Paque PLUS. The purity of primary NK cells was ≥96%. The data are expressed as the means±SEM. (A) Growth curve of the total and selected primary CD3⁻/CD56⁺ NK cells (left graph), and final yield of CD3⁻/CD56⁺ NK cells at day 14 (right graph). Note the log scale on the left graph. (B) Growth curve of the total cells (left graph) and the time course of the CD3⁻/CD56⁺ cell percentage (right graph) during the culture periods of various materials (CD3⁻: CD3-depleted PBMCs; CD14⁻: CD14-depleted PBMCs; and CD19⁻: CD19-depleted PBMCs). Note the log scale on the left graph. (C) Mean ratio of cell components at 14 days of cultivation from each material shown in (B).

FIG. 2.

Impact of culture medium on the production of NK cells from CD3-depleted PBMCs. An NK cell culture was established as shown in Fig. 3A using various culture media with high dose IL-2 (2500 IU/mL). Fourteen days later, the total cell number, purity, CD3-/CD56 NK cell yield, and fold expansion were assessed. Successful expansion was seen only when using Stemline II, CellGro SCGM, and KBM501, and the latter tended to show consistently better results.

FIG. 3.

Phenotype of zenithal-NK (Z-NK) cells. (A) Schematic diagram of feeder-free culture sequences for the production of Z-NK. (B) Microscopic morphology of primary NK and Z-NK at the end of 14 days of cultivation (left panels) and the expression of a typical activation inducer molecule, CD69 (right graph). (C) FACS analyses assessing the positive cell ratio of various surface markers (CD19, CD14, CD94, CD16, NKG2D), natural cytotoxicity receptors (NKp30, NKp46), and killer cell immunoglobulin-like receptors (KIR2DL5, KIR2DL3, KIR3DL1/DL/2). (D) Quantitative expression level of some typical activation receptors (NKp30, NKp46, and NKG2D) assessed by mean fluorescent intensity (MFI). Results are shown as the means±SEM. *p<0.01. Color images available online at www.liebertpub.com/hgtb

FIG. 4.

Antitumor potency of Z-NK cells. (A) Interferon (IFN)-γ expression of Z-NK and primary NK cells. Each line of NK cells was co-cultured with K562 target cells at a 2:1 effector-to-target ratio for 2 hr, and the expression of IFN-γ was assessed by intracellular FACS. Left panels show representative FACS profiles, and the right graph is the quantitative analysis based on MFI. Note that Z-NK cells contain a large CD56^high population. Results are shown as the means±SEM. *p<0.01. (B) Expression of perforin and granzyme B as assessed by intracellular FACS. The left panels show representative FACS profiles, and the right graphs show the results of the quantitative analyses. Results are shown as the means±SEM. *p<0.01.

FIG. 5.

Antitumor activity of Z-NK cells. (A) Cytolytic activity of Z-NK and primary NK cells against K562. In the assay of NK cell–mediated cytotoxicity (left panel), nearly 100% cytotoxicity was achieved only after 2 hr at E:T=1:1. The corresponding NK cell degranulation, labeled by CD107a (right panel), was assessed by FACS analyses. Results are shown as the means±SEM. *p<0.01. (B) The effect of primary CD3⁺ T cells on the cytotoxicity of Z-NK cells. After Z-NK cells were co-cultured with primary CD3⁺ T cells at a ratio of 1:5 for 16 hr, the number of NK cells in the cell culture was assessed by FACS and was taken as the number of effector cells. Then, NK cell–mediated cytotoxicity was assessed at a 1:5 effector-to-target ratio for 2 hr. The mean purity of primary T cells was ≥93%. No inhibitory effect of CD3⁺ T cells was evident. Results are shown as the means±SEM. *p<0.01. (C) Time-dependent Z-NK–mediated cytotoxicity. The %lysis values against K562 at effector-to-target ratios of 1:5 and 1:10 were assessed at various time points (2, 6, and 12 hr). A time-dependent increase of cytolysis was apparent, suggesting that Z-NK cells mediated the killing of multiple cells. (D) Direct cytolysis of Z-NK cells against normal cells (human umbilical vein endothelial cells [HUVECs] and human embryonic lung fibroblast line [MRC5]) and various tumor cells with varied MHC class I expression patterns. NK cell–mediated cytotoxicity was assessed 2 or 12 hr later at E:T=1:1. Note that (1) no apparent cytolysis was found in normal cells; (2) incubation time-dependent cytolysis was seen when using Z-NK cells, not primary NK cells; and 3) the cytolysis could not be predicted by the expression patterns of HLA-ABC. Results are shown as the means±SEM. *p<0.01.

FIG. 6.

Cytolytic function of Z-NK cells after cryopreservation and thawing and correlation with the preservation temperature. (A) The cell viability and cytolytic activity of thawed Z-NK cells. Z-NK cells were cryopreserved over 7 days, then thawed and analyzed in comparison with thawed cells cultured for 1 day in KBM501 with 5% human serum. The cell viability was assessed by trypan blue staining (left panel), and NK cell–mediated cytotoxicity was assessed 2 hr later at E:T=1:1 (right panel). The cryopreservation appeared not to have seriously affected the viability of Z-NK cells. However, the cytolytic activity of thawed Z-NK cells was seriously impaired in comparison with that seen by the pre-preserved Z-NK cells; the impaired activity could be largely recovered by 1 day of culture under KBM501. Results are shown as the means±SEM. (B) The effect of preservation temperature on cytolytic activity. Z-NK cells at the end of 14 days of cultivation were preserved for 3 days at various temperatures (4°C, 15°C, 23°C, 30°C, 37°C, 39°C, and 42°C). NK cell–mediated cytotoxicity was assessed 2 hr later at E:T=1:1. The preservation temperature was found to be a critical factor which dramatically affected the killing activity of Z-NK cells. Results are shown as the means±SEM. *p<0.01.

FIG. 7.

Antitumor effect of Z-NK cells in vivo. The antitumor effects of primary NK and Z-NK cells were assessed by using a mouse tumor xenograft model. Four days after tumor cell inoculation on the abdominal wall of NOD/SCID mice, when the tumor was established as ∼3 mm in diameter, NK or Z-NK cells were administered by bolus injection via a tail vein. The data show the results of two independent experiments. (A) Typical and representative gross observation of mice with K562 tumors treated with or without primary NK or Z-NK cells 19 days after tumor cell inoculation. Note the dramatic inhibition of tumor growth in mice treated with 1×10⁶ Z-NK cells by bolus injection. (B) Time course of the tumor volume treated. Results are shown as the means±SEM. *p<0.01. (C) Survival Kaplan–Meier curve. Note that the bolus injection of 1×10⁶ Z-NK cells significantly delayed the death of the animals, and two of seven animals showed complete remission of the tumors and long-term survival. The data were analyzed by log-rank test. *p<0.01.