Antigen-specific expansion and differentiation of natural killer cells by alloantigen stimulation

The Ly49D and Ly49H receptors associate with common adapter molecules for their activating signaling (Smith et al., 1998; Lanier, 2009); hence, we speculated that activation through Ly49D, which specifically recognizes H-2D^d, might induce the expansion and differentiation of Ly49D⁺ NK cells in response to alloantigen stimulation. To investigate the fate of Ly49D⁺ NK cells that respond to challenge with allogeneic cells, we adoptively transferred WT C57BL/6 NK cells (∼50% of NK cells are positive for Ly49D) into syngeneic DAP10 and DAP12 double-deficient (DAP10+12 KO) mice that lack expression of Ly49D and Ly49H, treated the mice with a depleting anti-CD8 mAb to ablate alloreactive CD8⁺ T cells, and then injected allogeneic BALB/c splenocytes (deleted from NK cells and T cells to avoid graft-versus-host reactions). Although prior studies have shown that Ly49D⁺ NK cells in C57BL/6 mice reject H-2D^d+ splenocytes in vivo (George et al., 1999a,b), we observed no expansion of the adoptively transferred Ly49D⁺ NK cells nor did we observe alterations in the phenotype of the Ly49D⁺ NK cells that would indicate their differentiation. Specifically, we did not detect the up-regulation of KLRG1 and Ly6C and down-regulation of CD27 and DNAM-1 that accompany the transition of Ly49H⁺ NK cells from naive to memory cells after MCMV infection (unpublished data).

The expansion and generation of MCMV-specific memory Ly49H⁺ NK cells requires IL-12, which is efficiently induced during viral infection (Sun et al., 2012). Therefore, we speculated that the expansion of alloantigen-specific Ly49D⁺ NK cells might similarly require robust production of cytokines, such as IL-12, that are not induced by challenge with alloantigen alone. Therefore, as schematically outlined in Fig. 1 A, after adoptive transfer of WT NK cells into DAP10+12 KO recipient mice treated with a depleting anti-CD8 mAb on day −1, the mice were infected with a sublethal dose of a MCMV mutant strain lacking m157 (Δm157 MCMV) to elicit an inflammatory response. The m157-deleted MCMV strain was used to prevent expansion of MCMV-specific Ly49H⁺ NK cells, which might complicate interpretation of the results given that a subset of the Ly49D⁺ NK cells coexpresses Ly49H. To stimulate the adoptively transferred Ly49D⁺ NK cells, BALB/c splenocytes (depleted from T and NK cells to prevent graft-versus-host reactions) were injected intravenously into recipient mice on days 0.5, 1.5, and 3 postinfection (pi; hereafter referred to as Ly49D alloantigen stimulation). In these conditions, donor Ly49D⁺ NK cells in the blood preferentially proliferated after Ly49D alloantigen stimulation, whereas Ly49D⁻ NK cells did not expand (Fig. 1, B and C). Subsequently, Ly49D⁺ NK cells underwent contraction, but were detectable in the blood and the spleen up to a month after Ly49D alloantigen stimulation (Fig. 1, C and D). The same trend was observed in the differentiation of long-lived, alloantigen-primed NK cells on days 28–30 pi in the spleen and blood (unpublished data). The kinetics of the alloantigen-specific Ly49D⁺ NK cell response strikingly resemble those of Ly49H⁺ NK cells during MCMV infection (Sun et al., 2009a), although the magnitude of the Ly49D⁺ NK cell response is less than the MCMV-induced response of Ly49H⁺ NK cells. Long-lived, alloantigen-primed Ly49D⁺ NK cells displayed the same phenotype (KLRG1^high, CD27⁻, CD11b⁺, Ly6C⁺, and DNAM-1⁻) as MCMV-specific memory Ly49H⁺ NK cells (Fig. 1 E; Sun et al., 2009a; Nabekura et al., 2014). These findings demonstrate that Ly49D is capable of providing the activating signal for expansion and differentiation of Ly49D⁺ NK cells, if accompanied by inflammation or infection.

We addressed the factors necessary for the expansion and differentiation of Ly49D⁺ NK cells after alloantigen stimulation. As compared with the response of donor Ly49D⁺ NK cells after stimulation with BALB/c splenocytes and Δm157 MCMV infection, donor Ly49D⁺ NK cells exhibited neither expansion nor differentiation when stimulated with syngeneic splenocytes accompanying Δm157 MCMV infection (Fig. 2, A and B). Similarly, injections of BALB/c splenocytes without Δm157 MCMV infection did not support expansion and differentiation of Ly49D⁺ NK cells (Fig. 2, C and D). Neutralization of IL-12 on the day before Δm157 MCMV infection and day 3 pi impaired the expansion and differentiation of Ly49D⁺ NK cells induced by alloantigen stimulation (Fig. 2, E and F), consistent with our prior studies documenting the role of IL-12 in the differentiation of memory Ly49H⁺ NK cells during MCMV infection (Sun et al., 2009b, 2012). These results indicate that both the interaction of Ly49D with its ligand and an optimal inflammatory condition to provide IL-12 are required for the acquisition of adaptive immune features by Ly49D⁺ NK cells.

Ly49D and Ly49H associate with DAP10 in addition to DAP12 to form activating receptor complexes (Smith et al., 1998; Lanier, 2009). To examine the role of DAP10 in expansion and differentiation of Ly49D⁺ NK cells after alloantigen stimulation, CD45.1⁺ WT Ly49D⁺ NK cells and CD45.2⁺ DAP10-deficient (Hcst^−/−) Ly49D⁺ NK cells were admixed and adoptively transferred into DAP10+12 KO mice, and then stimulated with BALB/c splenocytes after Δm157 MCMV infection (Fig. 3 A). DAP10-deficient Ly49D⁺ NK cells showed a significant impairment in the expansion of Ly49D⁺ NK cells (Fig. 3 B), consistent with our prior study analyzing the response of Ly49H⁺ NK cells to MCMV infection (Orr et al., 2009). However, long-lived Ly49D⁺ NK cells were still detected a month after the primary alloantigen stimulation (Fig. 3, B and C), and these DAP10-deficient alloantigen-primed Ly49D⁺ NK cells displayed the prototypic memory phenotype (KLRG1^high CD11b⁺ CD27⁻ Ly6C⁺ DNAM-1^low) equivalently to WT alloantigen-primed Ly49D⁺ NK cells (unpublished data). Naive, alloantigen-stimulated, and long-lived alloantigen-primed DAP10-deficient NK cells and WT NK cells equivalently expressed Ly49D (unpublished data). When DAP10-deficient alloantigen-primed Ly49D⁺ NK cells isolated from the primary recipient mice 28 d after the original Ly49D alloantigen stimulation were transferred into naive DAP10+12 KO recipient mice (Fig. 3 A), they again proliferated poorly in response to a secondary alloantigen stimulation (Fig. 3 D). These results demonstrate that DAP10 is required for the optimal expansion of naive and alloantigen-primed Ly49D⁺ NK cells, but dispensable for differentiation of alloantigen-stimulated Ly49D⁺ NK cells.

Both the activating Ly49D receptor and the inhibitory Ly49A recognize H-2D^d (Karlhofer et al., 1992; George et al., 1999a). We examined the impact of the inhibitory signal through Ly49A on the alloantigen-induced expansion and differentiation of Ly49D⁺ NK cells expressing Ly49A (Ly49D⁺A⁺) compared with Ly49D⁺ NK cells lacking Ly49A (Ly49D⁺A⁻). Ly49D⁺A⁻ NK cells were more proliferative than Ly49D⁺A⁺ NK cells at the peak of the initial expansion after alloantigen stimulation (Fig. 4 A). Although both Ly49D⁺A⁺ NK cells and Ly49D⁺A⁻ NK cells highly expressed KLRG1 after alloantigen stimulation as compared with naive Ly49D⁺ NK cells, alloantigen-stimulated Ly49D⁺A⁺ NK cells showed a less activated phenotype than alloantigen-stimulated Ly49D⁺A⁻ NK cells 7 and 28 d after the alloantigen stimulation, as indicated by lower expression of KLRG1 (Fig. 4 B). When the alloantigen-stimulated Ly49D⁺ NK cells were examined a month after the alloantigen challenge, the skewing toward the Ly49D⁺A⁻ NK cell subset was no longer evident (Fig. 4, A and D). Interestingly, Ly49D⁺A⁺ NK cells expressed higher amounts of Bcl-2 compared with Ly49D⁺A⁻ NK cells in the contraction phase of the response (Fig. 4 C). These results demonstrate that expression of the inhibitory Ly49A receptor suppresses activation of Ly49D⁺ NK cells and confers a selective survival advantage during the contraction phase after alloantigen stimulation. When these alloantigen-induced, long-lived Ly49D⁺ NK cells were isolated from the primary recipient mice a month after initial stimulation and transferred into naive recipient mice, followed by a secondary alloantigen challenge, Ly49D⁺A⁻ NK cells again showed better secondary expansion than Ly49D⁺A⁺ NK cells, similar to the primary expansion of naive Ly49D⁺ NK cells (Fig. 4 E). Collectively, these findings demonstrate that Ly49A suppresses activation and expansion of both naive and long-lived, alloantigen-stimulated Ly49D⁺ NK cells, but provides a survival advantage during contraction to the long-lived alloantigen-stimulated Ly49D⁺ NK cell subset.

To determine whether Ly49D⁺ NK cells generated by alloantigen stimulation exhibit enhanced effector functions as compared with naive Ly49D⁺ NK cells, naive and long-lived, alloantigen-stimulated Ly49D⁺ NK cells were co-cultured with RMA (H-2^b) transfectants expressing both m157 and H-2D^d or m157 alone (Fig. 5 A). Alloantigen-primed Ly49D⁺ NK cells bearing Ly49H (Ly49D⁺H⁺) and not expressing Ly49H (Ly49D⁺H⁻) produced more IFN-γ and degranulated more efficiently than naive Ly49D⁺H⁺ and Ly49D⁺H⁻ NK cells after co-culture with target cells expressing m157 and H-2D^d, respectively (Fig. 5 B). Alloantigen-primed Ly49D⁺H⁺ also showed stronger effector functions than naive Ly49D⁺H⁺ NK cells when they were co-cultured with m157-expressing target cells in the absence of H-2D^d (Fig. 5 B). Ly49A suppressed effector functions of naive and alloantigen-primed Ly49D⁺H⁺ NK cells only when Ly49D⁺H⁺ NK cells were co-cultured with target cells expressing H-2D^d (Fig. 5 C). These results show that the Ly49D alloantigen stimulation generates Ly49D⁺ NK cells with enhanced effector functions and Ly49A modulates effector functions of alloantigen-primed Ly49D⁺ NK cells against H-2D^d–expressing target cells.

Because a subset of the alloantigen-stimulated Ly49D⁺ NK cells coexpresses Ly49H, we assessed the antigen specificity of the secondary responses of the alloantigen-primed Ly49D⁺ NK cells (Fig. 6 A). Ly49D⁺H⁻ NK cells and Ly49D⁺H⁺ NK cells, but not Ly49D⁻H⁺ NK cells, underwent expansion, contraction, and differentiation after the primary Ly49D alloantigen stimulation (Fig. 6, B and C). Ly49D⁻ NK cells and long-lived Ly49D⁺ NK cells were isolated from the primary recipients 28 d after the primary alloantigen stimulation and were transferred into naive DAP10+12 KO mice, followed by infection with MCMV (Fig. 6 A). Alloantigen-primed Ly49D⁺H⁺ NK cells showed a robust expansion after the MCMV infection (Fig. 6 D).

We investigated the specificity of recall responses of memory Ly49H⁺ NK cells that had been generated after MCMV infection (Fig. 6 E). When WT NK cells were transferred into DAP10+12 KO mice followed by infection with MCMV, donor NK cell subsets bearing Ly49H efficiently expanded and differentiated into memory NK cells, consistent with our prior studies (Fig. 6, F and G). Donor Ly49H⁺ and Ly49H⁻ NK cells were isolated from these mice 28 d after infection and were transferred into recipient DAP10+12 KO mice followed by alloantigen stimulation (Fig. 6 E). MCMV-induced memory Ly49D⁺H⁺ NK cells, but not MCMV-induced memory Ly49D⁻H⁺ NK cells, proliferated in response to the Ly49D alloantigen stimulation (Fig. 6 H). Moreover, when alloantigen-primed Ly49D⁺H⁺ NK cells generated by the primary alloantigen stimulation were adoptively transferred into naive Ly49H-deficient or DAP12-deficient recipient mice, which lack functionally competent Ly49H⁺ NK cells and are unable to control early replication of MCMV (Sjölin et al., 2002; Sun et al., 2009a), they showed an improved protective effect against MCMV challenge on day 3 pi (Fig. 6 I). These results indicate that NK cells coexpressing both Ly49D and Ly49H can be primed to mount a more effective antiviral response when initially stimulated either by alloantigen or viral antigen.

Prior studies have established that NK cells can mount an immune response with many of the features of adaptive immunity when challenged with viruses or chemical haptens (O’Leary et al., 2006; Sun et al., 2009a, 2010; Paust et al., 2010; Min-Oo et al., 2013). Here, we have examined whether alloantigens also elicit the expansion of alloantigen-specific NK cells with enhanced responses upon rechallenge. Because the Ly49D receptor specific for H-2D^d and the Ly49H receptor specific for MCMV m157 are both expressed in C57BL/6 mice and use the same signaling subunits (e.g., DAP10 and DAP12; Karlhofer et al., 1992; Smith K.M. et al., 1998; Arase et al., 2002; Smith H.R. et al., 2002; Lanier, 2009), we compared the response of Ly49D⁺ NK cells to alloantigen stimulation with the well characterized response of Ly49H⁺ NK cells to MCMV infection. Although prior studies have demonstrated that Ly49D⁺ NK cells in C57BL/6 mice can reject transplanted allogeneic H-2D^d splenocytes (George et al., 1999a,b), we have determined that alloantigen alone is insufficient to induce the expansion and differentiation of these alloantigen-specific Ly49D⁺ NK cells. However, Ly49D⁺ NK cells can be induced to expand and differentiate, and mount a more potent recall response, as determined by ex vivo stimulation, when the alloantigen stimulation is accompanied by inflammation, in this case provided by viral infection at the time of alloantigen challenge. As with the generation of MCMV-specific Ly49H⁺ memory NK cells (Sun et al., 2012), the alloantigen-specific expansion of Ly49D⁺ NK cells was dependent on IL-12 produced in response to the accompanying viral infection. As with allogeneic challenge, when we have injected C57BL/6 mice with syngeneic tumor cells transduced to express m157 these m157-bearing tumors are eliminated by Ly49H⁺ NK cells, but m157 alone did not induce expansion and generation of long-lived Ly49H⁺ memory NK cells (unpublished data). Rather, the optimal expansion and differentiation of NK cells requires not only a specific ligand for the activating Ly49 receptor but also an appropriate inflammatory environment to provide the necessary growth and differentiation factors.

H-2D^d–driven expansion of Ly49D⁺ NK cells after the Ly49D alloantigen stimulation was less efficient than m157-driven expansion of Ly49H⁺ NK cells during MCMV infection (Sun et al., 2010). This is likely explained by the differences in binding affinities of Ly49D and Ly49H to H-2D^d and m157, respectively. We previously reported than Ly49H has a high affinity for m157 (K_d = 0.936 µM; Adams et al., 2007). In contrast, Ly49D has a much lower binding affinity to H-2D^d than Ly49A and Ly49G, which are inhibitory receptors recognizing H-2D^d (Ly49A:H-2D^d K_d = 4.4 µM; and Ly49G:H-2D^d K_d = 46.1 µM; Deng et al., 2008; Ma et al., 2014), which was directly quantified by using H-2D^d-tetramers (Hanke et al., 1999). Therefore, the less efficient expansion of Ly49D⁺ NK cells is likely attributed to the lower binding affinity of Ly49D to H-2D^d and the subsequently weaker activating signaling. Our results demonstrate that DAP10 is required for optimal expansion of alloantigen-specific Ly49D⁺ NK cells. The cytoplasmic domain of DAP10 is only 19 aa, and the only motif that is evolutionarily conserved is the YINM sequence that we and others have shown binds to PI3 kinase and Vav1 (Lanier, 2009). Therefore, these signaling molecules would not be recruited to the Ly49D receptor in DAP10-deficient NK cells when it is activated. We also have previously shown that in the absence of DAP10 there is less downstream activation of ERK1/2 when Ly49H is engaged (Orr et al., 2009). Therefore, the diminished expansion of DAP10-deficient alloantigen-stimulated Ly49D⁺ NK cells is almost certainly caused by the lack of recruitment of Vav1 and PI3 kinase to the receptor complex, which augments ERK1/2 signaling.

The antigen-specific expansion of both Ly49H⁺ and Ly49D⁺ NK cells is influenced by coexpression of inhibitory Ly49 receptors on the responding cells. Both the expansion and effector functions of the H-2D^d-specific response of Ly49D⁺ NK cells are restrained in Ly49D⁺ NK cells that coexpress Ly49A, an inhibitory receptor for H-2D^d. However, the long-lived alloantigen-stimulated Ly49D⁺ NK cells that persisted for more than a month after the initial challenge were not skewed to a Ly49A-negative subset, compared with the impairment of expansion of Ly49D⁺ NK coexpressing Ly49A during the peak of expansion. Similarly, we previously observed that the Ly49H⁺ NK cells expanding during MCMV infection preferentially lacked the inhibitory Ly49 receptors recognizing self H-2^b ligands at the peak of expansion (Orr et al., 2010), but at later times this skewing was no longer evident within the memory NK cell pool (unpublished data). The role of inhibitory receptors is to dampen the activating signals by recruitment of phosphatases, often SHP-1, and the dephosphorylation of key substrates such as Vav1 (Lanier, 2005). The recruitment of phosphatases would impair proliferation initiated through an activating receptor, but also likely suppress activation-induced cell death induced by strong signaling, accounting for the preferential survival of the Ly49A-bearing Ly49D⁺ NK cells during the contraction phase of the response. We observed high levels of Bcl-2 in the alloantigen-stimulated Ly49D⁺ NK cells coexpressing Ly49A during the contraction phase of the response, likely accounting for their preferential survival. Thus, we speculate that the inhibitory Ly49 receptors are selected against during the acute phase of an NK cell response, when the inhibitory receptors engaging their self-ligands would restrain expansion and effector functions, but during the contraction phase of the response these inhibitory receptors may provide a survival advantage by dampening signaling to prevent apoptosis.

Because the Ly49 receptors are expressed on overlapping subsets of NK cells, this provided us with the opportunity to determine whether NK cells coexpressing both Ly49H and Ly49D primed by stimulation through one of the receptors would demonstrate an enhanced immune response when subsequently activated in an antigen-specific manner through the other receptor. This was clearly the case. Ly49D⁺H⁺ NK cells that had been generated weeks earlier by alloantigen stimulation via Ly49D in vivo exerted enhanced effector functions on a per cell basis in response ex vivo to target cells expressing either m157 and H-2D^d or m157 alone. More importantly, these alloantigen-primed Ly49D⁺H⁺ NK cells provided more potent antiviral activity when challenged with MCMV infection, when compared with naive Ly49D⁺H⁺ NK cells. Reciprocally, we previously observed that MCMV-specific Ly49H⁺ memory NK cells showed enhanced responses when stimulated ex vivo through other activating receptors, for example NK1.1 (Sun et al., 2009a).

Our studies in the Ly49D model system have important implications for allogeneic transplantation in humans. As with Ly49D, some of the activating KIR in humans have specificity for polymorphic HLA class I ligands, for example KIR2DS1 specifically recognizes HLA-C alleles with a lysine at residue 80 (referred to as group 2 HLA-C; Colonna et al., 1993; Chewning et al., 2007). Our findings would predict that grafted tissues or cells from a donor expressing group 2 HLA-C proteins into a recipient lacking group 2 HLA-C alleles, but expressing KIR2DS1⁺ NK cells, likely would be insufficient to cause the expansion and enhanced effector function of these alloantigen-specific NK cells–unless the transplant recipient experienced a viral infection at the time of the transplantation. Viral infection, in particular HCMV reactivation, is common in transplant patients, particularly in patients receiving hematopoietic stem cell transplants. Remarkably, subjects who undergo allogeneic hematopoietic stem cell transplantation for the treatment of acute myeloid leukemia (AML) have a reduced risk of relapse and an advantage of overall survival when the donor and/or recipient are HCMV-seropositive before the transplantation (Behrendt et al., 2009; Elmaagacli et al., 2011). Although the cellular and molecular mechanisms of this anti-leukemia effect in HCMV-seropositive patients are not understood, these observations raise the intriguing possibility that virus-induced memory NK cells bearing activating receptors capable of recognizing AML cells might contribute to the prevention of relapse. Previous studies have reported that AML patients who receive an allogeneic hematopoietic stem cell transplant from a donor with polymorphic inhibitory KIR that are mismatched with the recipients HLA haplotype have a significantly reduced frequency of relapse and a survival advantage (Miller et al., 2005; Ruggeri et al., 2002, 2007). Moreover, AML patients that receive allogeneic hematopoietic stem cell grafts with a gene encoding the activating KIR2DS1 have a low rate of relapse, and patients receiving allogeneic hematopoietic stem cell grafts with a gene encoding the activating KIR3DS1 exhibit decreased mortality (Venstrom et al., 2012). Collectively, these results suggest that alloantigen-specific or virus-specific stimulation of NK cells may have a beneficial effect on the host response to transplantation, pathogens, and cancer.

C57BL/6, congenic CD45.1⁺ C57BL/6, and BALB/c mice were purchased from the National Cancer Institute. WT, DAP10, and DAP12 double-deficient (DAP10+12 KO; Inui et al., 2009; provided by T. Takai, Tohoku University, Sendai, Japan), DAP10-deficient (Hcst^−/−; Hyka-Nouspikel and Phillips, 2006), Ly49H-deficient (Klra8^−/−; Fodil-Cornu et al., 2008; provided by S. Vidal, McGill University, Montreal, Quebec, Canada), and DAP12-deficient (Tyrobp^−/−) mice (Bakker et al., 2000) on a C57BL/6 background were maintained at the University of California, San Francisco, in accordance with the guidelines of the Institutional Animal Care and Use Committee. Smith strain WT MCMV and Δm157 MCMV (Bubić et al., 2004; generously provided by U. Koszinowski, Max von Pettenkofer-Institut, Munich, Germany) were prepared by using C57BL/6 3T3 cells as described previously (Bubić et al., 2004). Mice were infected by intraperitoneal injection of 5–10 × 10⁴ plaque-forming units (pfu) WT MCMV or 5 × 10⁴ pfu Δm157 MCMV.

NK cells were enriched by incubating splenocytes with purified rat mAbs against mouse CD4, CD5, CD8, CD19, Gr-1, and Ter119, followed by anti–rat IgG antibodies conjugated to magnetic beads (QIAGEN) as described previously (Nabekura et al., 2014). In some experiments, NK cells were stained with antibodies against CD45.1 or CD45.2 and purified by using a FACS Aria III (BD).

Splenocytes from BALB/c mice were depleted of T cells and NK cells by incubating with purified rat mAbs against mouse CD4, CD8, and NKp46, followed by anti–rat IgG antibodies conjugated to magnetic beads. For Ly49D alloantigen stimulation, DAP10+12 KO recipient mice were inoculated intraperitoneally with 200 µg of a depletion mAb against mouse CD8 (clones 2.43 or YTS169.4) to prevent cytotoxic T lymphocyte–mediated rejection of BALB/c splenocytes, and 3 × 10⁵ WT NK cells or 1.5 × 10⁵ WT Ly49D⁺ NK cells were injected intravenously into DAP10+12 KO mice on the day before Δm157 MCMV infection. 4,000,000–5,000,000 T cell– and NK cell–depleted BALB/c splenocytes were injected intravenously on days 0.5, 1.5, and 3 pi. In some experiments, mice were inoculated intraperitoneally with 200 µg of a neutralizing mAb against mouse IL-12 p40 (clone C17.8) or an isotype-matched rat IgG2a on the day before Δm157 MCMV infection and on day 3 pi. For WT MCMV infection, 6 × 10⁵ WT NK cells or 10⁵ WT Ly49H⁺ NK cells were injected intravenously into Ly49H-deficient or DAP10+12 KO mice on the day before infection with MCMV.

Fc receptors (CD16 and CD32) were blocked with 2.4G2 mAb before surface or intracellular staining with the indicated fluorochrome-conjugated mAbs or isotype-matched control antibodies (BD, eBioscience, BioLegend, or TONBO Biosciences). Samples were acquired on an LSRII or a FACSCalibur (BD) and analyzed with FlowJo software (Tree Star).

10,000 naive or alloantigen-primed Ly49D⁺ NK cells expressing Ly49H were transferred separately into Ly49H-deficient or DAP12-deficient mice and infected with WT MCMV. The copy number of MCMV IE1 gene in DNA prepared from peripheral blood on day 3 pi was determined by quantitative PCR analysis using a SYBR green master mix reagent (Roche) as described previously (Nabekura et al., 2014).

100,000–150,000 NK cells were co-cultured with 1.5 × 10⁵ RMA or RMA transfectants expressing m157 alone or both m157 and H-2D^d for 6 h at 37°C in the presence of PE-conjugated anti-CD107a mAb and GolgiStop (BD), followed by staining for surface molecules and intracellular IFN-γ.

The Student’s t test was used to compare results. The Mann–Whitney U test was used to compare MCMV viral titers. P < 0.05 was considered statistically significant.

We thank the Lanier laboratory for comments and discussions.

The work was supported by National Institutes of Health grants AI068129 to L.L. Lanier. L.L. Lanier. is an American Cancer Society Professor and T. Nabekura is supported by Japan Society for the Promotion of Science.

The authors declare no competing financial interests.

Author contributions: T. Nabekura contributed to project planning, experimental work, data analysis, and writing the manuscript. L.L. Lanier contributed to project planning, data analysis, and writing the manuscript.