Transcription of the activating receptor NKG2D in natural killer cells is regulated by STAT3 tyrosine phosphorylation

Abstract

Signal transducer and activator of transcription 3 (STAT3) is considered a negative regulator of inflammation, as inhibition of STAT3 signaling enhances antitumor immunity. However, STAT3 activation is a key oncogenic pathway in natural killer (NK)-lineage large granular lymphomas, and we recently reported enhanced proliferation and function of human NK cells activated with IL-21, which signals primarily through STAT3. These IL-21-expanded NK cells also have increased NKG2D expression, which led us to focus our investigation on whether STAT3 regulates NKG2D. In this study, we show that modulation of STAT3 phosphorylation with cytokines and small-molecule inhibitors correlates with NKG2D expression on human NK cells, leading to altered NK-cell degranulation. Moreover, NKG2D expression on murine NK cells having conditional STAT3 ablation is lower than on NK cells from wild-type mice, and human NK cells carrying dominant-negative STAT3 mutations have decreased baseline NKG2D expression and blunted responses to IL-10 and IL-21. Lastly, we show binding of STAT3 to a predicted STAT3 binding site upstream of the NKG2D gene, which is enhanced by IL-10 and IL-21 and decreased by STAT3 inhibition. Taken together, these data show that NKG2D expression in NK cells is regulated at the transcriptional level by STAT3, resulting in a functional NK cell defect in patients with STAT3 mutations.

Figure 1

NK cells expanded ex vivo with aAPCs expressing mbIL21 have higher expression of NKG2D than those expanded with mbIL15. (A) Surface imunophenotyping by flow cytometry was performed on NK cells from each donor after expansion for 3 weeks with the indicated aAPCs. MFI of NKG2D expression is shown for the CD3⁻CD56⁺ population. Four donors in this experiment are representative of over 20 donors tested. P values indicated are for 2-tailed Student paired t test. (B) mRNA was purified from freshly-purified NK cells, or NK cells expanded for 3 weeks as above, and assessed for gene expression copy number of DAP10, DAP12, and NKG2D transcripts on the NanoString platform. Five donors in this experiment are representative of 2 independent experiments. P values indicated are for Student t tests of comparisons to fresh NK cells, with Holm-Sídák correction for multiple comparisons. *P < .05; **P < .01; ***P < .001. NS, no statistical significance.

Figure 2

STAT3 blockade decreases baseline NKG2D expression of primary human NK cells. (A) To assess the potential direct toxicity of STAT3 inhibitors, freshly-purified human NK cells were cultured in the presence of the STAT3 inhibitors, JSI-124 and S3I-201. NK-cell viability was evaluated at 24, 48, and 72 hours by flow-cytometric analysis of 7-AAD staining. (B) To assess the effect of STAT3 inhibition on NKG2D expression, NK cells were treated with the nontoxic concentrations of STAT3 inhibitors (JSI-124 and S3I-201) for 24 hours, after which NKG2D expression was evaluated by flow cytometry. Replicates are from 3 different donors. P values indicated are for 2-tailed Student t tests of comparisons to untreated NK cells with Bonferroni correction. *P < .05; ***P < .001. NS, no statistical significance.

Figure 3

Enhanced NKG2D expression on primary human NK cells by IL-10 and IL-21 is STAT3 dependent. (A) To assess the activation of STAT1, STAT3, and STAT5, NK cells were treated with 20 ng/mL of IL10 or IL21, with or without 0.1 µM of JSI-124 for 24 hours. Total and tyrosine-phosphorylated STAT were evaluated by western blot of cell lysates using STAT- or phospho–STAT-specific antibodies as indicated. (B) To assess cytokine-mediated induction of NKG2D expression, primary NK cells were treated with IL-10 and IL-21 at the indicated concentrations. After treatment of 24, 48, and 72 hours, NKG2D expression was evaluated by flow cytometry. (C) To assess the contribution of STAT3 signaling to cytokine-mediated induction of NKG2D, primary NK cells were treated with 20 ng/mL IL-10 or IL-21 with or without 0.1 µM of JSI-124 for 24 hours. NKG2D surface expression on NK cells was then determined by flow cytometry. (A,C) Representative of at least 3 independent experiments. (B) Pooled data from 4 donors.

Figure 4

STAT3 activation enhances NK-cell degranulation and cytotoxicity. Primary human NK cells were treated with 20 ng/mL of IL10 or IL21 with or without 0.1 µM of JSI-124 for 24 hours, and followed by coincubation with K562 cells for 4 hours. Target-mediated NK-cell degranulation was assessed by flow cytometry for CD107a. (A) Representative dot plot; (B) Pooled data from 3 donors. P values are shown for 2-tailed unpaired Student t tests.

Figure 5

Conditional ablation of STAT3 in murine hematopoietic cells or congenital STAT3 dysfunction in humans results in decreased basal NKG2D surface expression and decreased upregulation by STAT3-activating cytokines. (A) NK cells were isolated from the spleens of STAT3 conditional knockout and WT mice. The expression of NKG2D on NK1.1⁺ CD3⁻ splenocytes was determined by flow cytometry. PBMCs were isolated from NDs and patients with HIE and NK cells (gated on CD56⁺ CD3⁻ lymphocytes), and were assessed for baseline expression of NKG2D by flow cytometry and expressed in panel (B) as MFI, or in panel (C) as % of positive cells compared with isotype controls. (D) PBMCs were then cultured with 20 ng/mL of IL10 or IL21, and NKG2D expression determined at 24 and 48 hours. Data were obtained in 2 separate experiments by different investigators, the results of which were pooled after normalization of the 2 experiments to mean expression at 48 hours. P values for 1-tailed unpaired Student t tests are shown or designated as *P < .05; **P < .01. NS, no statistical significance. HIE sample numbers from Table 1 are indicated at the far right.

Figure 6

Phosphorylated STAT3 binds upstream of the human NKG2D gene translational start site. (A) The genomic sequence upstream of the NKG2D start site was analyzed using TFSEARCH software to identify putative STAT3 binding elements. Three potential regions were identified, with STAT3 binding elements indicated in underlined bold font. Sequences flanking the binding elements (identified in underlined font) were used as PCR primers. (B) Constitutive binding of pSTAT3 to the putative STAT3 binding sites was determined by PCR for each predicted region following co-ChIP with phosphorylated STAT3 from nonactivated primary human NK cells. (C) Enhanced binding of pSTAT3 to site 1 was assessed in response to treatment of NK cells with 20 ng/mL of IL10 or IL21, with or without JSI-124 for 24 hours. Data are representative of 2 replicate experiments.