Cutting Edge: IL-2–Induced Expression of the Amino Acid Transporters SLC1A5 and CD98 Is a Prerequisite for NKG2D-Mediated Activation of Human NK Cells

The activating receptor NKG2D, which is expressed by NK cells, CD8⁺ T cells, subsets of CD4⁺ T cells, invariant NKT cells, and γδ T cells (1), recognizes a diverse array of ligands belonging to the MHC class I chain–related protein family (i.e., MICA and MICB) and the UL16-binding protein family (i.e., ULBP1–6) (1). NKG2D ligands are normally absent or expressed at low levels on healthy resting cells, but they are induced in many stressed, virally infected, or transformed cells (2). The importance of the NKG2D receptor in immune surveillance is emphasized by the evasion strategies cancers and viruses employ to prevent the surface expression of NKG2D ligands (1). Moreover, NKG2D is implicated in the development and/or progression of autoimmune diseases, for example, rheumatoid arthritis, diabetes, celiac disease, and Crohn’s disease (1). NKG2D associates with the DAP10 adapter molecule, which is required for cell surface expression of NKG2D and promotes signaling through the PI3K and Grb2-Vav pathways to regulate NK cell–mediated cytotoxicity and cytokine production (2). However, the functional outcome of NKG2D stimulation depends on the activation state of the cells. NKG2D stimulation of freshly isolated human peripheral blood NK cells alone is insufficient to induce cytotoxicity and cytokine production even though the surface expression of NKG2D and intracellular signaling machinery are present (3, 4). In contrast, NK cells primed by IL-2 efficiently kill NKG2D ligand–expressing target cells and produce cytokines after NKG2D stimulation (5, 6). The mechanisms responsible for the IL-2–mediated change in NKG2D responsiveness remain undetermined.

Human NK cells are subdivided into two major subsets. CD56^bright NK cells are immature, whereas CD56^dim NK cells comprise a mature NK cell subset. CD56^bright NK cells express the high-affinity IL-2 receptor, IL-2Rαβγ, whereas resting CD56^dim NK cells lack IL-2Rα (7). When IL-2 binds to its receptor, JAK1 and JAK3 are phosphorylated, which recruit and activate STAT1, STAT3, and STAT5. The IL-2R also signals through PI3K, AKT, and the MAPK pathways (8). Gene expression profiling of resting and IL-2–primed human NK cells has shown that IL-2 treatment upregulates expression of a wide range of genes (9). In this study, we found that IL-2 priming of human NK cells leads to a JAK3-dependent upregulation of the glutamine transporter solute carrier family 1 member 5 (SLC1A5) (ASCT2) and the amino acid transporter SLC3A2/SLC7A5 (CD98). Furthermore, we show that the activity of these transporters and mammalian target of rapamycin complex 1 (mTORC1) is essential for functional activation of both CD56^bright and CD56^dim NK cells in response to NKG2D stimulation.

PBMCs were isolated from blood obtained from the Blood Centers of the Pacific under an Institutional Review Board–approved protocol (no. 10-00265) by density gradient centrifugation using Ficoll-Paque Plus (GE Healthcare Bio-Sciences). NK cells were purified (>90–95%) using EasySep NK cell enrichment kits (Stemcell Technologies). NK cells were cultured in RPMI 1640 medium (Corning Cellgro; Mediatech) containing 10% FBS (Thermo Scientific), 1× MEM nonessential amino acids solution (Life Technologies), 1 mM sodium pyruvate (University of California San Francisco Cell Culture Facility), 2 mM l-glutamine (University of California San Francisco Cell Culture Facility), penicillin (100 IU/ml), streptomycin (100 μg/ml) (Corning Cellgro), and 200 U/ml human rIL-2 (provided by Prometheus Laboratories).

NK cells were cultured in medium with or without 200 U/ml rIL-2 for 5, 10, 15, 20, or 24 h at 37°C and 5% CO₂ prior to stimulation_. Where indicated, the following inhibitors were added to cells 1 h prior to stimulation: 10 mM l-glutamic acid γ-(p-nitroanilide) hydrochloride (GPNA) (Santa Cruz Biotechnology) (1 M GPNA solution was prepared in DMSO and kept at 37°C immediately prior to its addition), 100 mM d-phenylalanine (Sigma-Aldrich) (dissolved in medium for 1 h at 37°C by vortexing every 5–10 min prior to addition), 100 nM rapamycin (Calbiochem) (109.4 μM rapamycin in DMSO stock was stored at −20°C, thawed, and diluted in medium prior to addition), and 100 nM Torin-1 (ApexBio) (1.645 mM Torin-1 in DMSO stock was stored at −80°C, thawed, and diluted in medium before addition). Nunc MaxiSorp ELISA plates (Thermo Fisher Scientific) were washed twice with PBS and then coated with 5 μg/ml anti-NKG2D mAb (1D11; BioLegend) or 5 μg/ml control mouse IgG₁ (MOPC-21; University of California San Francisco mAb Core Facility) in PBS for 24 h at 4°C. After coating, the plates were washed twice with PBS and blocked in complete medium for 10 min at room temperature (RT). For IFN-γ measurements, GolgiSTOP was added to cells immediately prior to stimulation. Stimulation of 1.5 × 10⁵ cells per well proceeded for 5 h at 37°C and 5% CO₂.

NK cells were cultured in medium with or without 200 U/ml human rIL-2, 10 ng/ml human rIL-15 (R&D Systems), or 10 ng/ml human rIL-12 for 6 h. Where indicated, 0.05% DMSO or 0.5 μM Jak3i (provided by Dr. J. Taunton, Dr. A. Weiss, and Dr. G. Smith) (1 mM Jak3i in DMSO was stored at −20°C, thawed, and diluted in medium prior to addition) was added to the cells 2 h prior to treatment with cytokines. Total RNA was purified using an RNeasy mini kit (Qiagen), and cDNA was generated from 0.05 μg of RNA using the SuperScript III first-strand synthesis system (Invitrogen). Real-time PCR was performed with the SYBR Green master mix reagent (Invitrogen) using standard conditions (primer melting temperature used, 55°C) and the following primer pairs: SLC1A5 forward, 5′-GTGTCCTCACTCTGGCCATC-3′, reverse, 5′-TACAGGACCGGTCGACTAGC-3′; SLC7A5 forward, 5′-TCATCATCCGGCCTTCATCG-3′, reverse, 5′-GAGCAGCAGCACGCAGA-3′; SLC3A2 forward, 5′-AGCTGGAGTTTGTCTCAGGC-3′, reverse, 5′-GGCCAATCTCATCCCCGTAG-3′; and hypoxanthine phosphoribosyltransferase (HPRT) forward, 5′-GACCAGTCAACAGGGGACAT-3′, reverse, 5′-CTTGCGACCTTGACCATCTT-3′. Samples were normalized to HPRT and the relative expression of the glutamine transporters was determined using the comparative cycle threshold (CT) (i.e., 2^−ΔΔCT) method, using the media sample as the reference sample.

mAbs used for cell surface staining were: FITC–anti-CD98 (MEM-108; BioLegend), FITC–anti-CD56 (HCD56; BioLegend), PerCP-Cy5.5–anti-CD56 (HCD56; BioLegend), Alexa Fluor 700–anti-CD3 (HIT3a; BioLegend), and allophycocyanin–anti-NKG2D (149810; R&D Systems). mAbs used for intracellular staining were: PE–anti-pS6 (D57.2.2E; Cell Signaling Technology), Alexa Fluor 647–anti-pSTAT5 [47/Stat5(pY694); BD Biosciences], Alexa Fluor 647–anti-IFN-γ, anti-ASCT2 (SLC1A5) (D7C12; Cell Signaling Technology), and Alexa Fluor 647–conjugated goat anti-rabbit IgG (A-21245; Invitrogen). For surface staining, cells were incubated with the indicated mAbs or isotype-matched control Abs (BioLegend) for 20 min on ice. During the pSTAT5 and pS6 measurements, the Alexa Fluor 700–anti-CD3 and FITC–anti-CD56 were added to the cell culture 15 min prior to fixation in 1.5% paraformaldehyde in PBS. For intracellular IFN-γ staining, cells were fixed and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences) and subsequently stained with Abs for 20 min on ice. For intracellular staining of pS6 and pSTAT5, cells were fixed in 1.5% paraformaldehyde in PBS for 10 min at RT and permeabilized in cold 100% methanol for 20 min on ice. Samples were either stored overnight at −20°C in methanol before staining or stained immediately following permeabilization with the anti-pS6 and anti-pSTAT5 for 30 min at RT. For intracellular staining of SLC1A5, cells were fixed in 1.5% paraformaldehyde in PBS for 10 min at RT, permeabilized in cold 100% methanol for 20 min on ice, and then stained with anti-SLC1A5 for 45 min at RT, followed by staining with anti-rabbit IgG for 20 min at RT. Samples were acquired on an LSR II (BD Biosciences) and analyzed with FlowJo software (Tree Star). Live single cells were gated based on forward and side light scatter profiles. NK cells were gated as CD3⁻CD56^bright or CD3⁻CD56^dim.

IL-2 priming renders human NK cells able to produce IFN-γ in response to NKG2D stimulation (5, 6). To determine the kinetics of IL-2 priming needed to induce NKG2D responsiveness, we cultured freshly purified peripheral blood NK cells in medium with or without IL-2 for 5, 10, 15, 20, or 24 h. Cells were then stimulated with plate-bound anti-NKG2D for 5 h and IFN-γ production was measured by flow cytometry. Few CD56^bright NK cells and even fewer CD56^dim NK cells produced IFN-γ after only 10 h of IL-2 priming (Fig. 1A). After 15 h of IL-2 priming, intermediate levels of IFN-γ⁺ cells were detected from both subsets, whereas peak IFN-γ production required 20–24 h priming (Fig. 1A). Although the relationship between priming time and IFN-γ production was consistent between donors, donor variability was observed in the percentages of IFN-γ⁺ cells at 24 h, ranging from 21 to 35% in CD56^bright NK cells and from 8 to 33% in CD56^dim NK cells (Fig. 1B). These results indicate that extended IL-2 priming of human NK cells is required for robust IFN-γ production in response to NKG2D stimulation, suggesting that IL-2–induced synthesis of new proteins might be necessary.

Searching the National Center for Biotechnology Information Gene Expression Omnibus database (10) accession number GSE8059 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE8059) to identify genes differentially expressed in human NK cells after IL-2 priming (9), we found three genes of interest: SLC1A5, which encodes a high affinity l-glutamine transporter (also designated ASCT2) (11), and SLC7A5, which together with SLC3A2 encodes a heterodimeric bidirectional antiporter (collectively designated CD98) that regulates the exchange of intracellular l-glutamine with extracellular amino acids such as l-leucine, l-phenylalanine, and l-tryptophan (11, 12). Consistent with the published dataset (9), we observed an increase in the SLC1A5 mRNA levels in NK cells after IL-2 priming (Fig. 2A). We also observed an increase in expression level of SLC7A5 after IL-2 priming, whereas no change was observed for SLC3A2 (Fig. 2A). However, SLC3A2 was expressed at a high level in resting NK cells compared with the housekeeping gene HPRT, in contrast to the low basal levels of SLC1A5 and SLC7A5 (Fig. 2B). Treatment with IL-15, which signals through components of the IL-2 receptor (13), similarly increased the expression of SLC1A5 and SLC7A5, and to a lower extent SLC3A2 (Supplemental Fig. 1A). Treatment of NK cells with IL-12, which stimulates IFN-γ production by NK cells (14), also increased the expression of SLC1A5, SLC7A5, and SLC3A2 (Supplemental Fig. 1A), although to a lesser degree than IL-15 and IL-2. IL-2 treatment leads to phosphorylation of STAT5 (pSTAT5), and the level of phosphorylation remained high throughout the course of IL-2 treatment (Supplemental Fig. 1B). We used a selective covalent inhibitor of JAK3 (15), which completely blocks the phosphorylation of STAT5 in IL-2–primed NK cells (Supplemental Fig. 1C), to examine its role in the IL-2–induced expression of SLC1A5 and SLC7A5. The induction of SLC1A5 and SLC7A5 expression in the IL-2–primed NK cells was abrogated after treatment with the JAK3 inhibitor, whereas the high basal expression of SLC3A2 was unaffected (Fig. 2A). Expression of both SLC1A5 and CD98 protein was increased in CD56^bright and CD56^dim NK cells after IL-2 priming (Fig. 2C–F), with the highest increase being observed in CD56^bright NK cells. Thus, IL-2 priming of human NK cells induces expression of the amino acid transporters SLC1A5 and CD98 at the mRNA and protein level.

We treated IL-2–primed NK cells with inhibitors against SLC1A5 (i.e., GPNA) (12) or CD98 (i.e., d-phenylalanine) (12) for 1 h prior to NKG2D stimulation to determine their involvement in the NKG2D-induced production of IFN-γ. Pretreatment of IL-2–primed CD56^bright and CD56^dim NK cells with either inhibitor completely blocked the NK cells ability to produce IFN-γ after NKG2D stimulation (Fig. 3). Furthermore, blocking of SLC1A5 and CD98 activity decreased the NKG2D-induced degranulation by the IL-2–primed NK cells (Supplemental Fig. 2A). The lack of responsiveness was not due to a change in the surface level of NKG2D, as comparable levels were observed between the control-treated and inhibitor-treated cells (Supplemental Fig. 2B). These results indicate that the activities of SLC1A5 and CD98 play an important role in NKG2D-mediated IFN-γ production and degranulation.

Uptake of l-glutamine by SLC1A5 closely followed by export of that l-glutamine in exchange for essential amino acids by CD98 activates mTORC1 (12). To determine whether increased expression of SLC1A5 and CD98 activates mTORC1 in the IL-2–primed NK cells, we measured the phosphorylation of S6, which is a downstream target of the mTORC1 pathway (16), by flow cytometry. pS6 was detected in both CD56^bright and CD56^dim NK cells early after IL-2 treatment, and the pS6 levels continued to increase during the 24-h treatment period (Fig. 4). Treatment of the IL-2–primed (24 h) CD56^bright and CD56^dim NK cells with GPNA or d-phenylalanine almost completely abrogated phosphorylation of S6, whereas it only minimally affected phosphorylation of STAT5 (Supplemental Fig. 2C), confirming the essential roles of SLC1A5 and CD98 in S6 phosphorylation at this latter time point.

mTORC1 is a critical regulator of mRNA translation (16) and controls IFN-γ production by mouse NK cells in vivo after polyinosinic-polycytidylic acid injection (17) and in vitro by IL-15–primed mouse NK cells stimulated with IL-12 or plate-bound Abs against the Ly49H (18). To examine its role in NKG2D-mediated IFN-γ production, we treated IL-2–primed human NK cells with inhibitors against mTORC1 (i.e., Torin-1 and rapamycin) (19) for 1 h prior to stimulation. Treatment of the IL-2–primed CD56^bright and CD56^dim NK cells with either inhibitor significantly decreased IFN-γ production (Fig. 5). However, inhibition of mTORC1 did not affect NK cell degranulation after NKG2D stimulation (Supplemental Fig. 2A). The lack of responsiveness was again not due to a change in the surface level of NKG2D, as comparable levels were observed between the control-treated and inhibitor-treated cells (Supplemental Fig. 2B). Taken together, these results highlight the essential role of mTORC1 in the NKG2D-mediated IFN-γ production by IL-2–primed NK cells, but not the degranulation.

In this study we show that IL-2 priming of CD56^bright and CD56^dim NK cells leads to an upregulation of the amino acid transporters SLC1A5 and CD98 and an increase in mTORC1 activity. The rate-limiting step in mTORC1 activation by the bidirectional transport of l-glutamine has been shown to be the import of l-glutamine through SLC1A5 (12). The transport of l-glutamine out of the cells through CD98 appears to occur rapidly in the presence of essential amino acids (12), which makes it difficult to the measure the intracellular levels of l-glutamine in IL-2–primed NK cells under physiological conditions. Thus, it remains to be determined whether the import of l-glutamine by SLC1A5 or its export by CD98 prevails in regulating the mTORC1 activation in IL-2–primed NK cells.

In summary, our findings reveal a crucial role for SLC1A5 and CD98 in controlling NKG2D responsiveness in human NK cells. Blocking either of the transporters completely abrogated the NKG2D-mediated IFN-γ production by the IL-2–primed NK cells. Furthermore, activation of mTORC1 by the transporters was found to be critical for the ability of the IL-2–primed NK cells to produce IFN-γ after NKG2D stimulation. In contrast to IFN-γ production, mTORC1 activity was not essential for NKG2D-mediated degranulation. The presence of lytic granules in many human NK cells is independent of the activation status (20), likely accounting for the continued ability of IL-2–primed NK cells to degranulate in the absence of mTORC1-promoted protein synthesis. Higher intracellular concentrations of l-leucine or l-arginine has been shown to cause the exchange of GDP for GTP among Rag GTPases, thereby promoting translocation of mTORC1 to the lysosomal membranes and subsequent activation by the GTPase Rheb (21). Further characterization of the precise mechanism that regulates the mTORC1 activity in the IL-2–primed NK cells and whether additional glutamine-dependent pathways are involved will guide us toward a more comprehensive understanding of how amino acid transport regulates the NKG2D responsiveness in NK cells. A recent study by Nakaya et al. (22) reported that SLC1A5 and CD98 activity was required for activation of mTORC1 in mouse T cells following stimulation, and the absence of SLC1A5 resulted in fewer IFN-γ–producing T cells after Listeria monocytogenes infection (22). Our data suggest a similar linkage between l-glutamine transport, mTORC1 activation, and IFN-γ production in human NK cells and thus provide novel insight into the cellular changes caused by IL-2 priming that controls the NKG2D responsiveness in NK cells.

We thank the Lanier laboratory for comments and discussions, Prometheus Laboratories for providing human IL-2, and Dr. Jack Taunton, Dr. Art Weiss, and Dr. Geoff Smith for providing the Jak3 inhibitor.

L.L.L. and the University of California, San Francisco have licensed intellectual property rights regarding NKG2D for commercial applications. The other authors have no financial conflicts of interest.