CD28-stimulated ERK2 phosphorylation is required for polarization of the microtubule organizing center and granules in YTS NK cells

Abstract

Activation of natural killer (NK) cell cytotoxicity requires adhesion and formation of a conjugate with a susceptible target cell, followed by actin polymerization, and polarization of the microtubule organizing center (MTOC) and cytolytic granules to the NK cell immune synapse. Here, by using the YTS NK cell line as a model, CD28 is shown to be an activating receptor. It signals cytotoxicity in a process dependent on phosphoinositide-3 kinase activation, leading to sustained extracellular signal-regulated kinase 2 (ERK2) phosphorylation. ERK and phospho-ERK localize to microtubule filaments. Neither conjugation with targets nor actin polymerization is affected by blocking ERK2 activation. However, both polarization of the MTOC and cytolytic granules to the synaptic region and NK cell cytotoxicity are strongly reduced by blocking ERK2 activation. A role for the CD28/CD80 interaction in cytotoxicity of human peripheral NK cells also was established. By contrast, lymphocyte function-associated antigen 1 (LFA-1) ligation transduces only a transient ERK2 activation and fails to induce killing in YTS cells. Thus, in YTS cells, a CD28 signal is used to polarize the MTOC and cytolytic granules to the NK cell immune synapse by stimulating sustained ERK2 activation.

Fig. 1.

Surface receptor phenotype and cytotoxicity of YTS NK and target cells. (A and B) YTS tumor cells (A) and P815, K562, and 721.221 target cells (B) were stained with FITC- or phycoerythrin-conjugated or unlabeled mAbs and analyzed by flow cytometry. Staining of unlabeled mAb was revealed with FITC-conjugated goat anti-mouse secondary Ab. The mean fluorescent intensities (MFI) of YTS staining (A) and target cell staining (B) after three independent experiments are shown. (C) YTS cells were used in a ⁵¹Cr-release cytotoxicity assay with the indicated targets. The average percentage of target cell lysis was calculated after three independent experiments at effector/target cell ratio of 10:1.

Fig. 2.

Cytotoxicity of YTS cells mediated by CD28 but not LFA-1. (A) YTS and 721.221 cells were allowed to form cell–cell conjugates, fixed, and stained with anti-CD28 and anti-LFA-1 mAb, followed with Alexa Fluor 568-conjugated secondary Abs (represented by red). The cells then were stained intracellularly with phalloidin-Alexa Fluor 488 to visualize F-actin (green) and Alexa Fluor 647-conjugated anti-GB mAb (blue) at the same time. Representative images are shown. (B) Stable transfectants of P815 and K562 target cells with the listed molecules were generated and tested with YTS cells in ⁵¹Cr-release cytotoxicity assays. Symbols at the left are in the same order as lines at the right. (C) After preincubation of YTS effector cells with anti-CD28 mAb or control mAb, CD86-transfected K562 cells were used as targets for ⁵¹Cr-release cytotoxicity assays in the presence of mAbs shown.

Fig. 3.

Requirement of ERK2 signals for NK cell cytotoxicity. (A) Effector YTS cells were preincubated with 10 μM LY294002/200 nM wortmannin/50 μM PD98059/20 μM TAT-MEK₁₃ at 37°C for 1 h. 721.221 target cells were used for ⁵¹Cr-release cytotoxicity assays. (B) Lentivirus-based RNAi was used to knock down ERK1 or ERK2 in YTS cells. Positively infected cells were examined for ERK1, ERK2, and WASp protein levels by Western blot. The parallel experiment with PBS instead of virus was regarded as mock, and the empty lentivirus without shRNA was used as the control vector. (C) The ERK knockdown effector YTS cells and target 721.221 cells were used for ⁵¹Cr-release cytotoxicity assays. The average percentage of target cell lysis at different E/T ratios in three independent experiments is shown.

Fig. 4.

Requirement of ERK2 signals for MTOC polarization. The drug-treated YTS cells, together with the lentivirus-based RNAi treated YTS cells, were conjugated with 721.221 targets and stained with Alexa Fluor 488-conjugated Phalloidin, Cy3-conjugated anti-β-tubulin mAb, and Alexa Fluor 647-conjugated anti-GB mAb. The percentages of MTOC and GB polarization to the NKIS in three independent experiments were analyzed. Two representative images of empty lentivirus-treated YTS cells (A) and ERK2 knockdown YTS cells (B) are shown. (C) MTOC and granule polarization. More than 50 conjugates were observed in each experiment (P < 0.005). (D) YTS cells were electrotransfected with both the GB-mRFP (red) construct and the ensconsin-3XGFP construct that selectively associates with MT (19) (green). The cells then were stained with anti-pan ERK mAb followed by Alexa Fluor 647-conjugated secondary Ab (blue). (E) YTS cells were stained with FITC-conjugated anti-phospho-ERK mAb (green), Cy3-conjugated anti-β-tubulin mAb (red), and anti-GB-Alexa Fluor 647 mAb (blue). In the merged image, the yellow dots (Inset) indicate the colocalization of phospho-ERK with MT.

Fig. 5.

Different patterns of ERK2 phosphorylation induced by CD28 and CD18. (A) CD28 and/or LFA-1 on YTS cells were crosslinked by using the indicated biotinylated mAb and streptavidin for the time indicated at 37°C. The cells were stained with Alexa Fluor 647-conjugated anti-phospho-ERK mAb and then used for FACS analysis. The MFI in five independent experiments was averaged to show the increased ratio of phosphorylated ERK after crosslinking (see Materials and Methods for details). (B) The lentivirus-based ERK1/2 RNAi knockdown YTS cells were used for the same crosslinking experiment with both anti-CD28 and anti-LFA-1 mAb. (C) The YTS cells were crosslinked with control mIgG, anti-LFA-1, or anti-CD28 mAb for the time indicated, lysed, run on a Bis-Tris gel, Western blotted with anti-phospho-ERK mAb, washed and reblotted with anti-pan ERK polyclonal Ab, and then detected by chemiluminescence. The exposure time was 1 min, and the result was identical to a 5-min overexposure. X signifies crosslinking.

Fig. 6.

Enhancement of killing by human pNK cells after transfection of P815 cells with CD80 or CD86. (A) NK cells were isolated from the peripheral blood of six independent donors (see Materials and Methods). FACS analysis was performed to confirm that the CD56⁺CD3⁻ NK cells were obtained, and the expression of CD28 (with mAb YTH293.2), LFA-1, NKG2D, and NKp46 was verified on the gated population. Although the arbitrary cut-off suggested that 18% of pNK cells were CD28⁺ (Right Upper), the histogram indicated that nearly all of these cells were CD28⁺ (Left Lower). (B) Anti-CD28 (clone L293) and anti-CD80 mAbs were used to block in the killing assays to confirm that the killing was specific for the CD28 and CD80 interaction. This result is representative of three independent experiments with pNK isolated from three donors.