Lactate dehydrogenase A-dependent aerobic glycolysis promotes natural killer cell anti-viral and anti-tumor function.

Natural killer (NK) cells are cytotoxic lymphocytes capable of rapid cytotoxicity, cytokine secretion, and clonal expansion. To sustain such energetically demanding processes, NK cells must increase their metabolic capacity upon activation. However, little is known about the metabolic requirements specific to NK cells in vivo. To gain greater insight, we investigated the role of aerobic glycolysis in NK cell function and demonstrate that their glycolytic rate increases rapidly following viral infection and inflammation, prior to that of CD8+ T cells. NK cell-specific deletion of lactate dehydrogenase A (LDHA) reveals that activated NK cells rely on this enzyme for both effector function and clonal proliferation, with the latter being shared with T cells. As a result, LDHA-deficient NK cells are defective in their anti-viral and anti-tumor protection. These findings suggest that aerobic glycolysis is a hallmark of NK cell activation that is key to their function.

INTRODUCTION

During their activation, natural killer (NK) cells require metabolic reprogramming in order to support cytokine production and rapid proliferation (O’Brien and Finlay, 2019). In vitro metabolic studies of mouse and human NK cells have demonstrated that exposure to proinflammatory cytokines (e.g., interleukin [IL]-2, IL-12, and IL-15) increases both oxidative phosphorylation and glycolysis (Keating et al., 2016; Keppel et al., 2015; Marçais et al., 2014). These findings suggest that upregulation of metabolic pathways is one of the mechanisms by which cytokines activate NK cells. Key metabolic regulators that cytokines have been shown to activate include mTOR (Donnelly et al., 2014; Keating et al., 2016; Marçais et al., 2014) and MYC (Dong et al., 2019; Loftus et al., 2018). During in vivo immune challenges, NK cells are exposed to inflammatory cytokines and activating receptor stimuli; however, the changes these signals induce in the metabolic profile of NK cells are poorly characterized. Furthermore, although NK cells undoubtedly have metabolic demands that must be met for optimal activation, the precise metabolic pathways that are essential during in vivo immune challenges are also not well understood.

Blocking glucose metabolism using 2-deoxy-d-glucose (2-DG), a competitive inhibitor of glucose-6-phosphate production, or through culture in glucose-free media has suggested this pathway to be critical to optimal NK cell-activating receptor-induced interferon (IFN)-γ production (Keppel et al., 2015) and control of mouse cytomegalovirus (MCMV) (Mah et al., 2017) in vitro and in vivo, respectively. However, because of potential pleiotropic and off-target effects of 2-DG in vivo, it is currently unknown which aspects of glucose metabolism are critical to NK cell function. Because glycolysis produces both ATP and substrates for biosynthetic pathways (e.g., nucleotide production) (Lane and Fan, 2015), we investigated the dependence of NK cell functional responses on aerobic glycolysis using a NK cell-specific genetic ablation of lactate dehydrogenase A (LDHA), as well as in vivo models of viral infection and tumor challenge. The conversion of pyruvate to lactate by LDHA helps maintain nicotinamide adenine dinucleotide (NAD+) levels in the cell, sustaining glycolysis and the pentose phosphate pathways (PPPs). A better understanding of the role glycolytic regulation plays in NK cell functional responses will allow us to identify disease states where limited glycolysis is the barrier to optimal NK cell function, and aid in the development of NK cell-based therapies.

RESULTS AND DISCUSSION

NK cells rapidly induce glycolysis as part of their response to MCMV infection

During MCMV infection, NK cells mount a robust effector response, with Ly49H+ cells undergoing clonal expansion and generating long-lived memory (Sun and Lanier, 2011). Thus, MCMV provides an excellent well-defined model in which to study the metabolic pathways regulating anti-viral NK cell effector functions. To investigate the metabolic changes occurring in NK cells after viral infection, we performed metabolite profiling using liquid chromatography-mass spectrometry on NK cells from the spleens of naive and infected mice. Analysis of more than 100 metabolites revealed that the levels of many glycolytic and pentose phosphate pathway intermediates are increased in NK cells within the first two days of MCMV infection (Figures 1A, 1B, and S1A). Phosphofructokinase uses ATP to convert fructose 6-phosphate to fructose 1,6-bisphosphate (FBP) and is the most important regulator of glycolytic flux (Sola-Penna et al., 2010). This process is inhibited by high concentrations of ATP and citrate, allowing it to react to the energy charge of a cell (Usenik and Legiša, 2010). In activated NK cells, we observed increased levels of FBP, which, in addition to dihydroxyacetone phosphate (DAP) (Figure 1A), have been shown to correlate with a higher rate of glycolysis (Tanner et al., 2018). Thus, we measured the glycolytic rate of naive and MCMV-activated Ly49H+ NK cells ex vivo by assessing their extracellular acidification rate (ECAR) as a proxy for lactic acid production rate (Wu et al., 2007).We found that the ECAR of Ly49H+ NK cells dramatically increases following MCMV infection, peaking at day 2 post-infection (PI) before returning to baseline in day 7 PI effector cells and day 30+ PI memory cells (Figure 1C). From the same MCMV-infected mice, we compared naive (day 0) and activated bulk (day 4 PI) or virus-specific (anti-M45 day 7 PI and day 30+ PI) CD8+ T cells. In contrast to NK cells, as well as data from transgenic OT-1 T cells in Listeria monocytogenes-ovalbumin (OVA) infection, where the ECAR peaked at day 3 PI (Ma et al., 2019), the ECAR does not peak until day 7 PI in endogenous MCMV-reactive CD8+ T cells; however, the maximum glycolytic rate reached similar levels to that of NK cells (Figures 1C and 1D). Thus, activation by viral infection in vivo rapidly increases the rate of aerobic glycolysis in NK cells prior to CD8+ T cells. These observations agree with the widely accepted function of NK cells as the innate counterpart of the CD8+ T cell, able to rapidly respond while providing time for the adaptive immune responses to develop.

Figure 1.

NK cells rapidly upregulate glycolysis during MCMV infection

(A and B) NK cells were harvested from naive and MCMV-infected mice at 2 days post-infection (PI), and mass spectrometry was used to assess metabolite relative abundance.

(A) Glycolytic and pentose phosphate pathway metabolites that passed significance testing (unpaired t test, p ≤ 0.05, fold change [FC] ≥ 2) between day 0 and day 2 PI are shown in the heatmap. Each sample represents an average of two technical replicates. Raw peak area counts for each metabolite were baselined to the mean of all samples prior to log2 transformation.

(B) Diagram of the glycolysis and pentose phosphate pathways; metabolites in bold are upregulated in NK cells on day 2 PI. P, phosphate; DAP, dihydroxyacetone phosphate; 2-PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate.

(C and D) Representative and summary plots of the extracellular acidification rate(ECAR) of 5 × 105splenic Ly49H+ NK cells (C) or CD8+ T cells (D) ex vivo, pooled from naive (day 0) or MCMV-infected mice at indicated number of days following MCMV infection under basal conditions (in summary plots), and in response to indicated metabolic inhibitors.

(E) Heatmap of KEGG glycolysis/gluconeogenesis pathway gene expression measured by RNA-seq. Genes shown are those that are differentially expressed in splenic Ly49H+ NK cells at day 0 versus day 2 PI. mRNA expression levels are row-scaled and log-transformed.

(F) Normalized counts of Ldha mRNA expression in Ly49H+ NK cells at various time points during MCMV infection measured by RNA-seq.

(G) Normalized counts of Ldha mRNA expression measured by RNA-seq in Ly49H− NK cells from mice at day 0 and day 2 PI.

(H) Expression level of Ldha mRNA relative to Actb mRNA measured by qRT-PCR in purified splenic NK cells stimulated for 3 h with IL-2 + IL-12 + IL-15 + IL-18. Means are plotted ± SD. *p < 0.05, **p < 0.01, ****p < 0.0001 (Student’s t test). Data are representative of two or more independent experiments; n = 2–8.

To better understand what drives the early increase in the glycolytic rate of NK cells during MCMV infection, we utilized transcriptional profiling of anti-viral NK cells (Lau et al., 2018) to examine which glycolytic pathway genes are differentially regulated in activated compared to naive NK cells. Using Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis at day 2 PI, we observed that many genes required for glycolysis and gluconeogenesis were transcriptionally upregulated compared to day 0 (Figure 1E). Among the upregulated genes were the glycolytic rate-limiting enzyme hexokinase (Hk2) and the enzyme key to aerobic glycolysis LDHA (Ldha) (Peng et al., 2016; Tanner et al., 2018). Expression of Ldha transcripts in NK cells during the course of MCMV infection showed the same pattern as their ECAR, with a rapid increase to peak Ldha expression at day 2 PI before returning to basal levels by days 7 and 30+ PI (Figures 1C and 1F).

T cells have been shown to rapidly upregulate aerobic glycolysis in response to T cell receptor (TCR) and coreceptor ligation (Frauwirth et al., 2002; Menk et al., 2018; Xu et al., 2021). Although early NK cell responses to MCMV infection are dependent on activation of pro-inflammatory cytokines (e.g., IL-12, IL-18, and type I IFNs; Madera and Sun, 2015; Sun et al., 2012), NK cells can also sense MCMV-infected cells via ligation of their activating receptor Ly49H to the virally encoded glycoprotein m157 (Arase et al., 2002; Brown et al., 2001; Daniels et al., 2001; Smith et al., 2002; Sun et al., 2009). Because the ECAR and transcriptomic data were generated using antigen-specific Ly49H+ NK cells, the question remained whether induction of Ldha expression in NK cells was dependent on receptor ligation in addition to the inflammatory environment in vivo. Interestingly, we observed similar Ldha upregulation in Ly49H− NK cells after MCMV infection (Figure 1G), and Ldha mRNA was upregulated in NK cells following in vitro stimulation in the absence of Ly49H signaling with a cocktail of proinflammatory cytokines (IL-2 + IL-12 + IL +15 + IL-18) known to be produced during MCMV infection (Figure 1H). Taken together, these findings suggest that stimulation through proinflammatory cytokines alone can promote increases in the glycolytic capacity of NK cells. Thus, during viral infection, the glycolytic rate of NK cells is upregulated by the initial cytokine surge, but then returns to basal levels after 7 days, whereas glycolysis in anti-viral CD8+ T cells peaks later, likely due to control of Ldha expression by TCR ligation rather than cytokines (Xu et al., 2021).

LDHA is not essential for NK cell development and homeostasis

LDHA converts pyruvate to l-lactate, and in doing so converts reduced NAD (NADH) into NAD+, thus maintaining the supply of NAD+ required to sustain rapid rates of glycolysis. As such, inhibition or ablation of LDHA can prevent cells from performing aerobic glycolysis (Oshima et al., 2020; Peng et al., 2016). To investigate whether aerobic glycolysis is essential for NK cell development, homeostasis, or anti-viral function in vivo, we generated mice with a NK cell-specific deletion of LDHA (Ncr1 × Ldha mice, hereafter designated as NK cell-Ldha). As expected, when cultured with IL-15, LDHA-deficient NK cells produced significantly less lactate and consumed less glucose in vitro (Figures S2A and S2B). However, LDHA-deficient NK cells did not demonstrate significant differences in development, with similar numbers of NK cells in the spleens of NK cell-Ldha mice and their littermate wild-type (WT) control mice (Figure S2C). Expression of the receptors KLRG1, NKG2A/C/E, Ly49G2, Ly49H, and Ly49I were also similar in LDHA-deficient NK cells when compared to NK cells from littermate controls (Figure S2D). The proportion of mature NK cells, as defined by the loss of CD27 and gain of CD11b, was also similar between the two groups (Figure S2E–S2G). Next, we compared the basal ECAR and oxygen consumption rate (OCR) of LDHA-deficient NK cells to those of WT counterparts from mixed bone marrow chimeras. The basal ECAR and OCR of WT and LDHA-deficient NK cells were similar (Figure S2H), suggesting that aerobic glycolysis is not required for the development and maintenance of naive NK cells.

LDHA-mediated induction of aerobic glycolysis is essential for optimal anti-viral NK cell function

Because we did not observe any NK cell defects in the absence of LDHA at steady state, we investigated the importance of aerobic glycolysis during the response against MCMV infection. First, we measured the ECAR of ex vivo NK cells at day 2 PI. We observed a significant reduction in basal ECAR in LDHA-deficient NK cells (Figure 2A), suggesting that LDHA is driving aerobic glycolysis in anti-viral NK cells. This phenotype became even more pronounced after the addition of oligomycin (which prevents ATP synthesis, hyperpolarizes the mitochondria, and blocks the tricarboxylic acid [TCA] cycle), upon which the WT NK cells further increase their ECAR to compensate for the inhibition of oxidative phosphorylation while the LDHA-deficient NK cells failed to do so (Figure 2A). Comparison of oxidative respiration between WT and LDHA-deficient NK cells through measurement of OCR revealed a lack of difference (Figure 2A), suggesting that LDHA supports the production of critical biochemical substrates and/or maintenance of NAD+, as opposed to the generation of ATP. Our data also suggests that LDHA deficiency prevents anti-viral NK cells from increasing their capacity to perform aerobic glycolysis.

Figure 2.

LDHA is essential for optimal anti-viral NK cell function

(A) Representative ex vivo ECAR of 5 × 105 splenic NK cells from mixed Ncr1 × Ldha (CD45.2) and WT (CD45.1) bone marrow chimeric mice at day 2 PI. Means are plotted ± SD and are representative of three independent experiments. ***p < 0.001.

(B) Kaplan-Meier representation of the survival of mice containing LDHA-deficient NK cells (Ncr1 × Ldha) or WT NK cells (Ldha) following MCMV challenge, with significant differences determined by a log-rank test (p = 0.0007). Plot combines the results of three independent experiments.

(C–E) RNA-seq was generated from Ly49H+ NK cells sorted from the spleens of mixed Ncr1 × Ldha (CD45.2) and WT (CD45.1) bone marrow chimeric mice at day 0 versus day 2 PI (n = 3). (C) PCA of RNA-seq data, showing the impact of Ldha deletion on gene expression. (D) Heatmap of 2,878 genes differentially expressed in WT or LDHA-deficient NK cells at day 2 PI (padj < 0.05) and (E) pathways predicted to be impacted by the data in (D) defined using the PANTHER database.

To determine the functional importance of aerobic glycolysis induction in NK cells during MCMV infection, we challenged NK cell-Ldha mice and their WT littermates with a high dose of virus. NK cell-Ldha mice more rapidly succumbed to MCMV infection compared to WT controls (Figure 2B), suggesting that the induction of aerobic glycolysis in NK cells is vital for an effective anti-viral host immune response in vivo.

To better understand the impact of LDHA deletion in activated NK cells, we sorted and performed transcriptome analysis on WT and LDHA-deficient NK cells at steady state and following MCMV infection. Principal-component analysis (PCA) showed little variance between naive WT and LDHA-deficient NK cells (Figure 2C), corroborating our findings that LDHA ablation did not impact NK cell maturation or homeostasis (Figures S2C–S2G). However, the transcriptome of LDHA-deficient NK cells was clearly distinct from that of WT NK cells at day 2 PI (Figure 2C), with 2,580 genes differentially expressed (adjusted p [padj] < 0.05) between WT and LDHA-deficient NK cells on day 2 PI (Figure 2D). Analysis of genes differentially expressed on day 2 PI highlighted several pathways that were affected by the absence of LDHA, including the p53 and Ras pathways, as well as oxidative stress responses (Figure 2E). Genes regulating metabolism (Hif1a, Mtor, and Vhl), NK receptor and cytokine signaling (Zap70, Stat5b, and Mapk8/9), and cellular proliferation (Rras, Kras, and Cdk9/19) were prominent among dysregulated pathways (Figure 2E). Thus, the activity of LDHA may be required for many aspects of anti-viral NK cell function.

LDHA is critical for NK cell clonal proliferation during MCMV infection

Given that many genes involved in cell survival and proliferation are dysregulated in LDHA-deficient NK cells during MCMV infection, we investigated the impact of LDHA deletion on anti-viral NK cell expansion. To do this, we transferred an equal number of NK cells from congenically distinct WT and NK cell-Ldha mice into Ly49H-deficient hosts and infected them with MCMV. In this system, only adoptively transferred NK cells that express Ly49H are capable of antigen-specific clonal proliferation and memory formation in response to MCMV infection (Sun et al., 2009). We observed that the WT NK cells responded robustly to MCMV infection by rapidly expanding; however, the co-transferred LDHA-deficient NK cells were unable to mount an effective response and were quickly outcompeted by the WT cells at day 7 PI and beyond (Figures 3A and S3A). Furthermore, LDHA-deficient Ly49H+ NK cells in spleen and liver showed diminished percentages of memory cells compared to WT controls (Figures 3B and S3B). Interestingly, LDHA-deficient Ly49H+ NK cells showed a reduced expression level of Ly49H compared to WT Ly49H+ NK cells at day 7 PI (Figure 3C), suggesting that LDHA is also required for the recently described avidity selection of NK cells (Adams et al., 2019; Grassmann et al., 2019). To confirm that the difference in effector and memory cell expansion was due to a proliferative defect, we labeled the NK cell populations with CellTrace Violet (CTV) priorto transfer and MCMV infection. By day 3.5 PI, a clear defect in the proliferative rate (measured by CTV dilution) of LDHA-deficient NK cells was observed when compared to their WT counterparts (Figure 3D). We also demonstrated that LDHA is critical for cytokine-driven proliferation in vitro, where cocultured WT but not LDHA-deficient NK cells underwent a significant expansion (Figure 3E). Thus, our data highlight an essential role for LDHA in the proliferative expansion of Ly49H+ NK cells and host protection during MCMV infection, suggesting that the rapid upregulation of glycolytic metabolism is required for an effective NK cell-mediated anti-viral response.

Figure 3.

LDHA promotes NK and CD8+ T cell proliferation in response to MCMV infection

(A and B) Percentage of transferred Ly49H+ NK cells that are WT (CD45.1) versus Ncr1 × Ldha (CD45.2) in the blood of Ly49H-deficient hosts during the course of MCMV infection (A), and in the spleen and liver harvested at >30 days PI (B).

(C) Ly49H expression level of Ly49H+ WT and LDHA-deficient NK cells at day 7 PI.

(D) Prior to transfer into Ly49H-deficient recipient mice, the two NK cell populations were labeled with CellTrace Violet (CTV). Mice were sacrificed 3.5days PI, and the representative flow plot and graph show cell division assessed by dilution of CTV.

(E) Ex vivo WT (CD45.1) and Ncr1 × Ldha (CD45.2) NK cells were co-cultured in IL-15, and cell numbers were determined by flow cytometry at indicated time points.

(F) Percentage of M45- and M38-specific populations within total CD8+ T cells from the peripheral blood of CD4 × Ldha and Ldha littermate controls (WT) at various time points following MCMV infection.

(G) Percentage of M45- and M38-specific memory CD8+ T cells from CD4 × Ldha (CD45.2) or WT (CD45.1) populations in the spleen and liver of mixed bone marrow chimeras at >30 days PI.

Means are plotted ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Student’s t test). Data are representative of three to four independent experiments with three to five mice per time point.

We next investigated whether LDHA-dependent aerobic glycolysis was a shared requirement for clonal expansion between NK cells and T cells responding to virus infection. LDHA-deficient CD4+ T cells had been previously shown to proliferate normally in vitro after stimulation with anti-CD3 and anti-CD28 in the presence of IL-2 (Peng et al., 2016). However, when we measured the proportion of CD8+ T cells specific for MCMV epitopes derived from M45 and M38, representing canonical and inflationary T cell responses (Munks et al., 2006), respectively, we observed a higher prevalence of virus-specific CD8+ T cells in the peripheral blood of WT mice compared to littermates with LDHA-deficient T cells (CD4 × Ldha mice, hereafter designated as T cell-Ldha) at all time points during MCMV infection (Figure 3F). These findings were corroborated with mixed bone marrow chimeric mice, where WT memory CD8+ T cells were found in greater abundance than LDHA-deficient counterparts in both spleen and liver >30 days PI (Figure 3G). Thus, CD8+ T cells require LDHA to mount a robust effector and memory response against viral infection similar to Ly49H+ NK cells, suggesting an important shared role for aerobic glycolysis in driving both innate and adaptive cytotoxic lymphocyte responses.

LDHA is required for optimal anti-tumor NK cell effector function

In addition to LDHA expression being important for driving proliferation, our transcriptional analysis of WT versus LDHA-deficient NK cells following MCMV infection also uncovered genes involved in NK cell activation (Figure 2E). To investigate the functional relevance of this finding, we stimulated naive NK cells from NK cell-Ldha mice and their WT littermates with multiple concentrations of IL-12 + IL-18 in vitro and observed a similar proportion of cells producing IFN-γ between the two groups (Figure 4A). WT and LDHA-deficient NK cells similarly produced comparable levels of IFN-γ and granzyme B early after MCMV infection (Figure 4B). Thus, LDHA appears dispensable for cytokine-induced production of IFN-γ by NK cells. In contrast, following in vitro stimulation with phorbol 12-myristate 13-acetate (PMA) + ionomycin, a significantly reduced fraction of LDHA-deficient NK cells produced IFN-γ compared to WT controls (Figure 4C). Because ionomycin induces Ca2+ release from the endoplasmic reticulum (ER), which synergizes with PMA to activate calcium-sensitive kinases that are part of activating receptor signaling cascades (Chatila et al., 1989; Matthews and Cantrell, 2009), crosslinking of the NK cell-activating receptor NK1.1 also resulted in lower IFN-γ production in LDHA-deficient NK cells (Figure 4D).

Figure 4.

LDHA is required for optimal NK cell effector and anti-tumor function

(A) Percentage of naive Ldha and WT NK cells producing IFN-γ upon in vitro stimulation with IL-12 + IL-18 for 5 h (A, 10 ng/mL IL-12 + 20 ng/mL IL-18; B, 1 ng/mL IL-12 + 2 ng/mL IL-18; C, 100 pg/mL IL-12 + 200 pg/mL IL-18).

(B) Representative histograms showing IFN-γ and granzyme B expression in WT (dashed line) and Ncr1 × Ldha (solid line) NK cells harvested from mice at day 2 PI.

(C) Graph shows percentage of naive Ldha and WT NK cells producing IFN-γ following in vitro stimulation with PMA + ionomycin for 5 h.

(D) Graph shows percentage of naive Ldha and WT NK cells producing IFN-γ following in vitro stimulation with plate-bound anti-NK1.1 (α-NK1.1) or isotype control antibodies for 5 h.

(E) Lytic killing activity of Ldha and WT NK cells incubated at a 20:1 effector/target ratio with calcein-acetoxymethyl-labeled Yac-1 target cells in vitro for 5 h.

(F and G) In vivo clearance of m157 transgenic (Ly49H ligand-expressing) (F) or β2m KO (MHC class I-deficient) (G) splenocytes 3 and 24 h after intravenous (i.v.) injection, respectively, in NK cell-Ldha or WT littermate control mice.

(H) Quantitation of the lung metastases in NK cell Ldha or WT littermate control mice challenged i.v. with the B16-F10 melanoma cell line. Results are combined from three independent experiments. *p < 0.05 (Mann-Whitney test).

Means are plotted ± SD. (A–G) Data are representative of two to five independent experiments with three to five mice per time point. *p < 0.05, **p < 0.01, ***p < 0.001 (Student’s t test).

We therefore tested whether LDHA deficiency had an impact on the cytotoxic function of NK cells by incubating calcein acetoxymethyl (calcein-AM)-labeled Yac-1 cells with NK cells isolated from either NK cell-Ldha mice or their WT littermates. We observed a reduced lysis of Yac-1 cells by the LDHA-deficient NK cells (Figure 4E). Next, to dissect the importance of LDHA expression in NK cells participating in both “missing self”-induced and activating receptor-induced cytotoxic responses in vivo, we injected splenocytes from β2-microglobulin β2m) knockout (KO) (major histocompatibility complex [MHC] class I-deficient) or m157 (Ly49H ligand) transgenic mice, respectively. Compared with labeled and co-injected WT control splenocytes, m157 transgenic or β2m KO splenocytes were less efficiently killed in NK cell-Ldha mice compared to their WT littermates (Figures 4F and 4G). Altogether, these in vitro and in vivo data suggest that LDHA expression in NK cells is important for both activating receptor- and missing self-driven cytotoxicity.

Finally, we tested whether LDHA is vital for NK cell-mediated tumor surveillance in vivo. Using the NK cell-sensitive B16-F10 melanoma model of tumor metastasis, we challenged NK cell-Ldha mice and their WT littermates and measured tumor burden 14 days later. The number of tumor nodules in the lung was significantly higher in NK cell-Ldha mice compared to their WT littermates, suggesting that aerobic glycolysis is not only essential for optimal anti-viral but also anti-tumor function of NK cells (Figure 4H). Our results support recent findings that LDHA deficiency reduced the efficacy of CD8+ T cells transferred as a therapeutic treatment into B16 tumor-bearing mice (Hermans et al., 2020).

In this study, we directly assess the glycolytic rate of NK cells ex vivo from virally infected mice. Previous in vitro studies of the metabolic regulation of NK cells have rapidly advanced our understanding of lymphocyte metabolism. However, artificial levels of metabolite availability in a tissue culture dish can produce misleading results (Ma et al., 2019), and blockade of metabolic pathways with synthetic inhibitors such as 2-DG can have potential off-target side effects (Andresen et al., 2012; Kurtoglu et al., 2007; Zhong et al., 2008). Furthermore, use of inhibitors in vivo would likely target more than just NK cells, and pleiotropic effects would complicate interpretation of any phenotypes. In the present study, we engineered a NK cell-specific deletion of LDHA to investigate the role of aerobic glycolysis in the NK cell response against virus infection and tumor challenge. We discovered that during viral infection, NK cells rapidly induce aerobic glycolysis to levels similar to antigen-specific CD8+ T cells, but with greater kinetics. Inhibition of aerobic glycolysis through cell-specific deletion of Ldha prevented NK cells and CD8+ T cells from mounting a robust effector and memory response against MCMV infection.

Genetic ablation of LDHA prevents the inter-conversion of pyruvate to lactate. The prevailing view has been that LDHA converts pyruvate to lactate in rapidly dividing immune cells (Pearce et al., 2013) and that lactic acid inhibits the effector function of NK cells and T cells (Brand et al., 2016; Dodard et al., 2020; Fischer et al., 2007; Haas et al., 2015; Husain et al., 2013; Mendler et al., 2012). However, recent in vivo tracing studies have questioned these assertions. Mice infused with 13C-lactate demonstrated significant lactate uptake by the spleen (Hui et al., 2017). Furthermore, ex vivo 13C-glucose tracing in CD8+ T cells from Listeria-infected mice demonstrated a significantly lower rate of lactate production than in vitro cultured T cells (Ma et al., 2019). Future studies will be required to demonstrate that activated NK cells are not utilizing lactate as a fuel source.

Proliferation of anti-viral NK cells is reliant on both aerobic glycolysis and oxidative phosphorylation (Mah-Som et al., 2021, this issue of Cell Reports This suggests that metabolic fitness is a key regulator of NK cell proliferation during an antiviral response. How these metabolic pathways regulate proliferation requires further investigation. It could simply be that a reduced concentration of metabolites becomes a rate-limiting factor for the biosynthesis of components required for proliferation. However, metabolites might also directly influence proliferation, as FBP was recently shown to activate Ras (Peeters et al., 2017). Thus, should stimulated NK cells suffer a reduction in levels of FBP or other metabolites, this may reduce proliferative signals and therefore the rate of cell division, possibly to preserve the cells and avoid apoptosis due to a lack of nutrient availability. This would be in line with previous findings from our lab that preventing the buildup of metabolism-induced stress, in the form of dysfunctional mitochondria, is key to robust NK cell memory formation (O’Sullivan et al., 2015).

It has previously been suggested that glucose metabolism promotes the production of IFN-γ in response to activating receptor ligation (Keppel et al., 2015). In this study, we demonstrate that LDHA tunes the responsiveness of NK cells to activating receptor stimulation potentially through aerobic glycolysis. In agreement with this finding, deletion of COX10, a critical component of the mitochondrial electron transport chain, had no negative impact on NK cell IFN-γ production in response to NK1.1 cross-linking or in vitro killing of Yac-1 cells (see accompanying article by Mah-Som et al., 2021). Thus, glucose metabolism likely promotes IFN-γ production and NK cell cytotoxicity through the glycolytic pathway rather than through oxidative phosphorylation.

Glycolysis has been suggested to regulate IFN-γ production in T cells at both the transcriptional and post-transcriptional levels. Transcriptionally, glycolysis has been shown to maintain the high levels of acetyl-coenzyme A required to enhance histone acetylation and RNA polymerase II (RNA pol II) recruitment to the Ifng locus (Peng et al., 2016). Post-transcriptionally, glycolysis has been shown to prevent binding of substrate-free glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to the 3′ UTR of Ifng mRNA, thereby relieving inhibition of translation (Chang et al., 2013). In NK cells, however, IFN-γ production in response to cytokine stimulation remained normal in the absence of LDHA, suggesting that Ifng expression is not impaired by either of these mechanisms. Moreover, LDHA-deficient NK cells exhibit defects in cytotoxicity, suggesting that glycolysis more broadly regulates their effector processes, likely through other mechanisms. A potential candidate is the reduced expression of members of the Ras and p38 mitogen-activated protein kinase (MAPK) pathways highlighted by our RNA sequencing (RNA-seq) analysis. This would seem to be in line with data suggesting that NK cell-activating receptor- but not cytokine-induced production of IFN-γ is dependent on the IκB kinase (IKK) complex and TPL2 (Map3k8) signaling (Piersma et al., 2019). A better understanding of how specific metabolic fitness regulates NK cells will be essential toward optimizing NK cell therapies.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Joseph Sun (sunj@mskcc.org).

Materials availability

This study did not generate new unique reagents.

Data and code availability

Data generated in this study have been deposited in the Gene Expression Omnibus (GEO accession number GSE149447). Previously generated data referenced in this study can be found in GEO using accession number GSE106138.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice

All mice used in this study were housed and bred under specific pathogen free conditions at Memorial Sloan Kettering Cancer Center, and handled in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC). The following mouse strains on a C57BL/6 background were used in this study: C57BL/6 (CD45.2), B6.SJL (CD45.1), B6 CD45.1xCD45.2, Ncr1 (Narni-Mancinelli et al., 2011), Ldha (Peng et al., 2016), Ncr1 x Ldha, CD4 x Ldha, Klra8 (Ly49H-deficient) (Fodil-Cornu et al., 2008), β2 m KO and m157 transgenic (Tripathy et al., 2008). Experiments were conducted using age- and sex-matched mice in accordance with approved institutional protocols.

Mixed bone marrow chimeras and adoptive transfers

Wild-type (WT) host CD45.1+ CD45.2+ mice were lethally irradiated with 900 Gy and reconstituted with a 1:1 mixture of bone marrow cells from WT (CD45.1+) and either Ncr1 × Ldha, CD4 × Ldha (CD45.2+) mice. Intraperitoneal injection of mixed bone marrow chimeric mice with 200 μg of anti-NK1.1 (clone PK136) ensured depletion of any residual donor or host mature NK cells.

Adoptive co-transfer studies were performed by intravenously injecting an equal number of Ly49H+ NK cells from WT (CD45.1) and gene-deficient (CD45.2) mice, or from mixed bone marrow chimeras, into Klra8 recipients before infecting recipient mice with MCMV.

METHOD DETAILS

Viral infection

Mixed bone marrow chimeric mice, Klra8 mice in adoptive transfer experiments, and mice in survival experiments received 7.5 × 103, 7.5 × 102, or 4 × 104 plaque-forming units (PFU) of MCMV (salivary gland passaged Smith strain) by i.p. injection, respectively.

Isolation of lymphocytes

Spleens and liver were dissociated through a 100-μm strainer. Dissociated liver was resuspended in a 40% Percoll solution and centrifuged at 700 × g (with reduced break speed) for 12 min to separate lymphocytes from hepatocytes. To isolate bone marrow cells, femur and tibia were ground with mortar and pestle and the resulting solution was filtered through a 100-μm strainer. Red blood cells in spleen, liver, blood, and bone marrow were lysed using ACK lysis buffer. In experiments involving sorted cells, splenocytes were incubated with rat anti-mouse CD3ε and CD8α (except when CD8+ T cells were required) along with Ly6G, CD4, CD19, and Ter-119 antibodies (Bio X Cell, clones 17A2, 2.43, 1A8, GK1.5, 1D3, and TER-119, respectively) followed by incubation with BioMag goat anti-rat IgG beads (QIAGEN) to remove Ab-bound cells without touching populations of interest.

Flow cytometry and cell sorting

Surface staining of single-cell suspensions from bone marrow and spleen was performed using fluorophore-conjugated antibodies in the presence of Purified Rat Anti-Mouse CD16/CD32 (BD Bioscience) to block non-specific Fc binding and Fixable Viability Dye eFluor 506 (eBioscience/Thermo Fisher Scientific). Intracellular staining was performed by fixing and permeabilizing cells with the BD Cytofix/Cytoperm Fixation and Permeabilization kit (BD Bioscience). The following antibodies were used for flow cytometry: CD3ε (17A2, Tonbo, #25-0032), CD8α (53-6.7, BioLegend, #100730), CD11b (M1/70, eBioscience/Thermo Fisher Scientific, #467108), CD27 (LG.7F9, eBioscience/Thermo Fisher Scientific, #465001) CD45.1 (A20, BioLegend, #110729), CD45.2 (104, BioLegend, #109821), CD49b (DX5, BioLegend, #108918), CD69 (H1.2F3, BioLegend, #104524), Granzyme B (GB11, Biolegend, #2114575), IFN-γ (XMG1.2, eBioscience/Thermo Fisher Scientific, #1834366), Klrg1 (2F1, eBioscience/Thermo Fisher Scientific, #1518768) Ly49G2 (4D11, eBioscience/Thermo Fisher Scientific, #763604), Ly49H (3D10, eBioscience/Thermo Fisher Scientific, #11-5886-81) Ly49I (YLI-90, BD Bioscience, #2534427), NK1.1 (PK136, Tonbo, #65-5941), NKG2A/C/E (20d5, BD Bioscience, #465305) and TCRβ (H57-597, BioLegend, #109220). MHC class I tetramers were generated by conjugating Db/HGIRNASFI (M45) or Kb/SSPPMFRV (M38) monomers (NIH Tetramer Facility) to streptavidin-PE or streptavidin-APC (BioLegend, #405204 and #405207, respectively). NK cell proliferation was analyzed by labeling cells with 5 μM CellTrace Violet (CTV, Thermo Fisher) according to manufacturer’s protocol prior to transfer into mice.

Flow cytometry and cell sorting were performed on LSR II and Aria II cytometers (BD Biosciences), respectively. NK cell activation on d2 PI was confirmed by expression of CD69. Virus-specific CD8+ T cells were sorted using M45 tetramer staining at d7 PI and d30+ PI. Cells were sorted into 100 μL of FBS for RNA sequencing and mass spectrometry, or 20% FBS, 1mM EDTA, 25 mM HEPES in PBS for Seahorse. Flow cytometry data were analyzed with FlowJo software (Tree Star).

In vitro NK cell functional assays

For receptor crosslinking, purified antibodies against NK1.1 (PK136) or isotype control were dispensed into high-binding ELISA plates at the concentration of 25 ug/ml in PBS, and allowed to adhere overnight at 4°C. For stimulation with cytokines, 20 ng/ml IL-18 plus 10 ng/ml of IL12 were added to the media. For PMA and ionomycin, 10 ng/ml and 1 μg/ml, respectively were added to the media. For all ex vivo stimulation, 2×105 bead enriched NK cells were plated in RPMI, and Brefeldin A (10 μg/ml; Sigma) and Golgi-Stop (0.7 μl/106 cells; BD) were added 2 hours after the culture started. Cells were analyzed by flow cytometry at 5h post-stimulation.

In vitro killing assay

Yac-1 cells at a concentration of 1×106 cells per ml in phenol red free complete RPMI were stained with calcein-acetoxymethyl (calcein-AM) dye at a concentration of 15 μM for 30 minutes at 37°C. Excess dye was washed off prior to incubation with NK cells at a 20:1 effector:target ratio (3×103 targets per well) total volume of 200 μL in phenol read free complete RPMI in a 96 well plate. A maximum lysis control was generated by adding 100 μL of 2% Triton X-100 instead of effector cells, spontaneous lysis controls were created by adding 100 μL of media in place of effector cells. Cells were incubated together for 4 h at 37°C. Following incubation, the top 100 μL of cell free media was carefully transferred to a flat bottomed white 96 well plate and calcein AM release was read in a plate reader set to an excitation of 485 nm ± 9 nm and an emission of 525 nm ± 15nm. A percentage lysis was calculated by subtracting spontaneous lysis from all values then assuming the maximum lysis to be 100% lysis (Neri et al., 2001).

In vivo killing assay

25 × 106 β2 m KO or m157 transgenic splenocytes were labeled with CTV or CTB and WT control splenocytes were stained with the other. (Labels was alternated in target cells between experimental repeats to control for any effect of the different dyes). Splenocytes were injected i.v. and mice were bled 24 h (β2M) or 3h (m157 transgenic) later. Killing was measured relative to co-transferred WT control splenocytes.

B16 metastasis

For the tumor metastasis in vivo, 2×105 B16-F10 cells were injected i.v. and mice were euthanized 14 days later. Lungs were harvested, and metastatic plaques counted.

Cytokine stimulation and qPCR

Sorted NK cells were cultured in complete RPMI with 10% FBS, containing IL-2 (20 ng/mL) + IL-12 (20 ng/mL) + IL-15 (20 ng/mL) + IL-18 (10 ng/mL) at 37°C for 3 hours. RNA was isolated using the PicoPure RNA Isolation Kit following the manufacturer’s instructions (Thermo Fisher Scientific). RNA was reverse transcribed into cDNA with a High capacity cDNA RT kit (Thermo Fisher Scientific). Quantitative real time PCR was carried out using the TaqMan system (Ldha Mm01612132_g1, Applied Biosystems), and values were normalized to Actb (Mm02619580_g1, Applied Biosystems) expression.

Cell culture

Sorted NK cells were cultured in complete IMDM with 10% FBS containing IL-15 (50 ng/ml) at 1.25-2.5 × 105 NK cells per mL. Cell counts were performed using 123count eBeads (Thermo Fisher Scientific) on a flow cytometer. Media was clarified by centrifugation at 800 xg for 3 mins and analyzed on a YSI biochemistry analyzer (YSI).

Seahorse bioanalyzer

5 × 105 NK cells or CD8+ T cells were plated in buffer-free, glucose-free media (Seahorse Biosciences/ Agilent Technologies) with glutamine (2 mM), sodium pyruvate (0.5 mM), and glucose (10mM) for OCR and ECAR measurements, which were made under basal conditions and following addition of oligomycin (1 μM), FCCP (1 μM), Rotenone (1 μM) and Antimycin A (1 μM) at indicated time points and recorded on a Seahorse XFe96. All compounds added during the Seahorse run are from Seahorse Biosciences/Agilent Technologies.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) sample preparation

Sorted NK cells were pelleted at 800 × g for 3 mins 4°C, supernatant was aspirated, and pellets were frozen in liquid nitrogen. The intracellular metabolites were extracted by adding 1 mL ice-cold 80% methanol directly to the cell pellet and incubated overnight at −80°C to aid protein precipitation. The following day, the methanol extracts were centrifuged at 20,000 × g for 20 min at 4°C and 900 μL of the supernatant was transferred to a new tube and evaporated to dryness in a vacuum concentrator (GeneVac). The samples were reconstituted in 40 μL of 97:3 water:methanol containing 10 mM tributylamine and 15 mM acetic acid (mobile phase A) and incubated on ice for 20 min, vortexing every 5 min to ensure adequate re-suspension. All samples underwent one final centrifugation step (20,000 × g for 20 min at 4°C) to remove any residual particulate.

LC-MS/MS analysis

The reconstituted samples were subjected to MS/MS acquisition using an Agilent 1290 Infinity LC system equipped with a quaternary pump, multisampler, and thermostatted column compartment coupled to an Agilent 6470 series triple quadrupole system using a dual Agilent Jet Stream source for sample introduction. Data were acquired in dynamic MRM mode using electrospray ionization (ESI) in negative ion mode. The capillary voltage was 2000 V, nebulizer gas pressure of 45 psi, drying gas temperature of 150°C and drying gas flow rate of 13 L/min. A volume of 5 μL of sample was injected on to an Agilent Zorbax RRHD Extend-C18 (1.8 μm, 2.1 × 150 mm) column operating at 35°C. The 24 min chromatographic gradient was performed using 10 mM tributylamine and 15 mM acetic acid in 97:3 water:methanol (mobile phase A) and 10 mM tributylamine and 15 mM acetic acid in methanol (mobile phase B), at a 0.25 mL/minute flow rate. At the end of the 24 min, the gradient included a backflush of the analytical column for 6 min with 99% acetonitrile at a 0.8 mL/min flow rate, followed by a 5 min re-equilibration step at 100% A (MassHunter Metabolomics dMRM Database and Method, Agilent Technologies).

Each batch included method blanks and pooled samples. The pooled samples were prepared by mixing 5 μL of each sample in the batch. This mixture was injected multiple times as a control for changes during the run. Four technical replicates of the pooled sample were injected at the start of the batch to condition the system, followed by samples in randomized order. Analysis of the pooled samples was used to monitor the reproducibility of the system and the stability of the run over time. Initial data analysis was performed using MassHunter Quantitative Analysis (v. B.09.00). Following quantitative analysis, the files were exported to Mass Profiler Professional (v. 15.0) software for differential metabolite analysis. Statistical analysis was performed by applying a t test between the d0 and d2 groups and Benjamini-Hochberg correction (p < 0.05).

Sample preparation for RNA-seq

We sorted 2–6 × 104 splenic NK cells from mixed bone marrow chimeras for RNA-sequencing at d0 and d2 PI. Ly49H+ NK cells were defined asTCRβ−NK1.1+ DX5+ Ly49H+ on d0, and also CD69+ on d2 PI, along with either CD45.1+ (WT) or CD45.2+ (KO). Each sample represents an individual mouse. RNA was isolated from sorted cell populations using the PicoPure RNA Isolation Kit following the manufacturer’s instructions (Thermo Fisher Scientific). After RiboGreen quantification and quality control by Agilent BioAnalyzer, 2 ng total RNA with RNA integrity numbers ranging from 9.4 to 10 underwent amplification using the SMART-Seq v4 Ultra Low Input RNA Kit (Clonetech catalog # 63488), with 12 cycles of amplification. Subsequently, 10 ng of amplified cDNA was used to prepare libraries with the KAPA Hyper Prep Kit (Kapa Biosystems KK8504) using 8 cycles of PCR. Samples were barcoded and run on a HiSeq 4000 in a 50bp/50bp paired end run, using the HiSeq 3000/4000 SBS Kit (Illumina). An average of 34 million paired reads were generated per sample and the percent of mRNA bases per sample ranged from 59% to 71%.

RNA-seq analysis

RNA-seq analysis on Ly49H+ NK cells harvested at various time points during MCMV infection has been previously described (Lau et al., 2018). For RNA-seq generated in this study, paired-end reads were trimmed for adaptors and removal of low-quality reads using Trimmomatic (v0.36) (Bolger et al., 2014). Transcript quantification was based on the mm10 UCSC Known Gene models and performed using the quasi-mapping-based mode of Salmon (v.0.13.1) (Patro et al., 2017) correcting for potential GC bias. Transcripts were summarized to gene level using tximport (v1.10.1) (Soneson et al., 2015). Differential analyses were executed with DE-Seq2 (v1.22.2) (Love et al., 2014) using the UCSC Known Gene models as reference annotations. Genes were considered differential if they showed a P value (padj) < 0.05, adjusted for multiple hypothesis correction. Principal component analysis was performed using log2 values calculated by the rlog function from the DESeq2 package.

For gene set analysis, metabolism gene sets were retrieved from KEGG (https://www.kegg.jp/kegg/) through the Bioconductor KEGGREST package with the following identifiers: mmu00010 - Glycolysis/Gluconeogenesis; mmu00190 - Oxidative phosphorylation; mmu01212 - Fatty acid metabolism). Gene set enrichment was performed on downregulated differentially expressed genes (LDHA-deficient versus WT) with goseq (v1.34.1) (Young et al., 2010), using pathways from the PANTHER database (http://www.pantherdb.org/) retrieved through the Bioconductor Panther.db package. P values were corrected for multiple-hypothesis testing using the Benjamini & Hochberg method.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical analyses

For graphs, data are shown as mean ± SD, and unless otherwise indicated, statistical differences were evaluated using a two-tailed unpaired Student’s t test. Equal sample variance was assumed; unless an F test demonstrated significant variance then Welch’s correction was applied. Statistical differences in survival were determined by log-rank (Mantel-Cox) Test analysis. p < 0.05 was considered significant. Graphs were produced and statistical analyses were performed using GraphPad Prism.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-Mouse CD3ε (clone 17A2)	Tonbo Biosciences	Cat#25-0032; RRID:AB_2621619
Anti-Mouse TCRβ (clone H57-597)	BioLegend	Cat#109220; RRID:AB_893624
CD8α (clone 53-6.7)	BioLegend	Cat#100734; RRID:AB_2075238
NKG2A/C/E (clone 20d5)	BD Biosciences	Cat#740549; RRID:AB_465305
Anti-Mouse NK1.1 (clone PK136)	Tonbo Biosciences	Cat#65-5941; RRID:AB_2621910
Anti-Mouse NKp46 (clone 29A1.4)	BioLegend	Cat#137604; RRID:AB_2235755
Anti-Mouse Ly49H (clone 3D10)	eBioscience/Thermo Fisher	Cat#11-5886-81; RRID:AB_1257160
Anti-Mouse CD45.1 (clone A20)	BioLegend	Cat#110729; RRID:AB_1134170
Anti-Mouse CD45.2 (clone 104)	BioLegend	Cat#109821; RRID:AB_493730
Anti-Mouse CD49b (clone Dx5)	BioLegend	Cat#108918; RRID: AB_2265144
Anti-Mouse/Human CD11b (clone M1/70)	BioLegend	Cat#101223; RRID:AB_755985
Anti-CD27 (clone LG.7F9)	eBioscience/Thermo Fisher	Cat#14-0271-81; RRID:AB_467182
Anti-Mouse KLRG1 (clone 2F1)	Tonbo Biosciences	Cat# 25-5893-82; RRID: AB_1518768
Anti-Mouse Ly49D (clone 4E5)	BioLegend	Cat#138308; RRID:AB_10639939
Ly49G2 (clone 4D11)	eBioscience/Thermo Fisher	Cat #46-5781-82; RRID: AB_1834437
Anti-Mouse Ly49I (clone YLI-90)	BD Biosciences	Cat# 550595; RRID:AB_393770
Anti-Mouse CD69 (clone H1.2F3)	BioLegend	Cat#104524; RRID:AB_2074979
Anti-Human/Mouse Granzyme B (clone GB11)	BioLegend	Cat#515403; RRID:AB_2114575
Anti-Mouse IFN gamma (clone XMG1.2)	Tonbo Biosciences	Cat#20-7311; RRID:AB_2621616
Anti-Mouse CD107a (clone 1D4B)	BioLegend	Cat#121611; RRID:AB_1732051
streptavidin–PE	BioLegend	Cat #405204
streptavidin-APC	BioLegend	Cat #405207
MHC class I monomer Db/HGIRNASFI (M45)	NIH Tetramer Facility	N/A
MHC class I monomer Kb/SSPPMFRV (M38)	NIH Tetramer Facility	N/A
InVivoMab Anti-Mouse CD8α (NK cell enrichment, clone 2.43)	Bio X Cell	Cat#BE0061; RRID:AB_1125541
InVivoMab Anti-Mouse CD4 (NK cell enrichment, clone GK1.5)	Bio X Cell	Cat#BE0003-1; RRID:AB_1107636
InVivoMab Anti-Mouse CD19 (NK cell enrichment, clone 1D3)	Bio X Cell	Cat#BE0150; RRID:AB_10949187
InVivoMab Anti-Mouse Ter-119 (NK cell enrichment, clone TER-119)	Bio X Cell	Cat#BE0183; RRID:AB_10949625
InVivoMab Anti-Mouse Ly6G (NK cell enrichment, clone 1A8)	Bio X Cell	Cat#BE0075-1; RRID:AB_AB_1107721
InVivoMAb anti-mouse CD3 (NK cell enrichment, clone 17A2)	Bio X Cell	Cat#BE0002; RRID:AB_1107630
Anti-Mouse NK1.1 (receptor crosslinking clone PK136)	BioLegend	Cat#108758 RRID:AB_2800568
Mouse IgG2a, κ Isotype Ctrl (receptor crosslinking clone MOPC-173)	BioLegend	Cat#400264; RRID:AB_11148947
Purified Rat Anti-Mouse CD16/CD32 (Fc Block clone 2.4G2)	BD Biosciences	Cat#553142; RRID:AB_394657
Bacterial and virus strains
Murine Cytomegalovirus (MCMV)	J. Sun (PI)	Smith Strain
Chemicals, peptides, and recombinant proteins
Recombinant Mouse IL-12 Protein	R&D Systems	Cat#419-ML
Recombinant Mouse IL-18	MBL	Cat#B002-5
Recombinant Mouse IL-2 Protein	R&D Systems	Cat#402-ML
Recombinant human IL-15	miltenyi	Cat#130-095-766
Phorbol 12-myristate 13-acetate (PMA)	Sigma-Aldrich	Cat#P8139
Ionomycin calcium salt from Streptomyces conglobatus (Ionomycin)	Sigma-Aldrich	Cat#I0634
Critical commercial assays
BD Cytofix/Cytoperm	BD Biosciences	Cat#554714
Calcein, AM, cell-permeant dye	Thermo Fisher Scientific	Cat#C1430
Seahorse XF Cell Mito Stress Test Kit	Agilent	Cat#103010-100
BioMag Goat Anti-Rat IgG (NK cell enrichment)	QIAGEN	Cat#310107
CellTrace Violet Cell Proliferation Kit	Thermo Fisher Scientific	Cat#C34557
Fixable Viability Dye eFluor 506	eBioscience	Cat#65-0866-18
PicoPure RNA Isolation Kit	Thermo Fisher Scientific	Cat#KIT0214
Deposited data
Raw Data Files for RNA Sequencing	NCBI Gene Expression Omnibus	GEO: GSE149447
Experimental models: Cell lines
Yac-1	L. Lanier (PI)	N/A
B16-F10	ATCC	CRL-6475
Experimental models: Organisms/strains
Mouse: WT or CD45.2: C57BL/6J	The Jackson Laboratory	Stock#000644; RRID:IMSR_JAX:000664
Mouse: WT or CD45.1: B6.SJL-Ptprca Pepcb/BoyJ	The Jackson Laboratory	Stock#002014; RRID:IMSR_JAX:002014
Mouse: CD45.1xCD45.2	J. Sun (PI)	N/A
Mouse: Ldhaflox	M. Li (PI)	N/A
Mouse: β2 m KO	The Jackson Laboratory	Stock#002087; RRID:IMSR_JAX:002087
Mouse: Ncr1iCre	E. Vivier (PI) (Narni-Mancinelli et al., 2011)	N/A
Mouse: m157 transgenic	W. Yokoyama (PI) (Tripathy et al., 2008)	N/A
Mouse: CD4Cre	The Jackson Laboratory	Stock#022071; RRID:IMSR_JAX:022071
Mouse: VavCre	The Jackson Laboratory	Stock#008610; RRID:IMSR_JAX:008610;
Mouse: Klra8−/− or Ly49H-deficient	S. Vidal (PI) (Fodil-Cornu et al., 2008)	N/A
Oligonucleotides
TaqMan Ldha Mm01612132_g1	Thermo Fisher Scientific	Cat#4331182
TaqMan Actb Mm02619580_g1	Thermo Fisher Scientific	Cat#4331182
Software and algorithms
DESeq2 (v.1.22.2)	Love et al., 2014	http://bioconductor.org/packages/release/bioc/html/DESeq2.html
Trimmomatic (v.0.36)	Bolger et al., 2014	http://www.usadellab.org/cms/?page=trimmomatic
Bowtie2 (v.2.3.4)	Langmead and Salzberg, 2012	http://bowtie-bio.sourceforge.net/bowtie2/index.shtml
Salmon (v.0.13.1)	Patro et al., 2017	https://salmon.readthedocs.io/en/latest/salmon.html
tximport (v.1.10.1)	Soneson et al., 2015	https://bioconductor.org/packages/release/bioc/html/tximport.html
goseq (v.1.34.1)	Young et al., 2010	https://bioconductor.org/packages/release/bioc/html/goseq.html
KEGG.db (v.3.2.3)	Bioconductor	https://bioconductor.org/packages/release/data/annotation/html/KEGG.db.html
PANTHER.db (v.1.0.4)	https://bioconductor.org/packages/release/data/annotation/html/PANTHER.db.html	https://bioconductor.org/packages/release/data/annotation/html/PANTHER.db.html
R (v.3.5.1)	https://www.r-project.org/	https://www.r-project.org/
UCSC mm10 Known Gene Annotation Package (v.9)	Bioconductor	https://bioconductor.org/packages/release/data/annotation/html/TxDb.Mmusculus.UCSC.mm10.knownGene.html