Functional Characterization of Ly49+CD8 T-Cells in Both Normal Condition and During Anti-Viral Response

The role of Ly49+CD8 T-cells in the immune system is not clear. Previously, several papers suggested Ly49+CD8 T-cells as immunosuppressors, while multiple studies also suggested their role as potent participants of the immune response. The mechanism of Ly49 expression on CD8 T-cells is also not clear. We investigated phenotype, functions, and regulation of Ly49 expression on murine CD8 T-cells in both normal state and during LCMV infection. CD8 T-cells express different Ly49 receptors compared with NK-cells. In intact mice, Ly49+CD8 T-cells have a phenotype similar to resting central memory CD8 T-cells and do not show impaired proliferation and cytokine production. Conventional CD8 T-cells upregulate Ly49 receptors during TCR-induced stimulation, and IL-2, as well as IL-15, affect it. At the same time, Ly49+CD8 T-cells change the Ly49 expression profile dramatically upon re-stimulation downregulating inhibitory and upregulating activating Ly49 receptors. We observed the expression of Ly49 receptors on the virus-specific CD8 T-cells during LCMV infection, especially marked in the early stages, and participation of Ly49+CD8 T-cells in the anti-viral response. Thus, CD8 T-cells acquire Ly49 receptors during the T-cell activation and show dynamic regulation of Ly49 receptors during stimulation.

Factors driving the expression of Ly49 receptors on CD8 Tcells are not defined yet. Numerous experiments describe the upregulation of Ly49 receptors on CD8 T-cells as a result of stimulation with the cognate antigen (either during the infection or during the exposure to self-antigens) (10,16,17), cytokines (18,19), or because of interactions with activated lymphocytes during the immune response (presumably, activated CD4 Tcells) (10)(11)(12). However, the exact cellular or molecular mechanisms that govern this process, as well as the stability of Ly49 expression with time, are unknown.
Considering the uncertain role and possible implications of the NKR + CD8 T-cells in the immune homeostasis, a better understanding of their functions and ways of generation will help to develop treatments against viral infections, cancer, or autoimmune pathologies and improve our knowledge about the immune system. In the present study, we investigated functions and ways of generation of murine Ly49 + CD8 T-cells in normal conditions and during virus infection using cell-based methods and LCMV infection model.
We observed the expression and dynamic regulation of Ly49 receptors on CD8 T-cells during the cell activation and assessed how IL-2 and IL-15 affect this process. We also demonstrate the positive effect of Ly49 + CD8 T-cells from LCMV-immune mice on the antiviral response upon adoptive transfer. Our data describe cellular mechanisms of Ly49 induction on CD8 T-cells as well as demonstrate the role of Ly49 + CD8 T-cells in the immune response.

Experimental Animals
C57BL/6 WT, CD45.1, RAG1 KO, Nur77 GFP+/− on the C57BL/6 background were obtained from the Model Animal Research Center of Nanjing University. Ly49 KO mice (C57BL/6 background) were a generous gift of Prof. Dong (Beijing). The mice were crossed with our C57BL/6 WT mice for two generations and then used for experiments. Unless indicated, experiments were performed using 6-to 12-wk-old C57BL/6 mice. Experiments were conducted under institutional guidelines for animal care and use.

LCMV Virus Production, Quantification, and Infection
LCMV Armstrong 53b strain, BHK-21, and Vero cells were a generous gift of Prof. Ye (Chongqing). The virus was grown on BHK-21 cells and quantitated by plaque assay on Vero cells as described elsewhere (20). For acute infections, mice were injected with 2•10 5 PFU of LCMV Armstrong i.p.
To determine LCMV titers in the spleen and liver, harvested organs were weighed, weight-adjusted to 1/10 with DMEM (Basal Media, Shanghai) with 2% FBS (Gibco) and 1% penicillin/streptomycin solution (Hangzhou Keyi Biotechnology Co., Ltd). Tissue samples were homogenized with metal beads at the Tissue Lyser 48 (Shanghai Jing Xin Industrial Development Co., Ltd) and clarified by centrifugation at 1,000 g for 10 min. Then the samples were frozen at −80°C and analyzed later. Organ homogenates were diluted in 10-fold increments and tested in the standard 4-day plaque assays.

Flow Cytometry
We used the standard protocols for flow cytometry analysis. Cells were stained with live/dead stain, incubated with Fc blocking antibodies, stained with antibodies (titrated prior experiments), and then either fixed in 2% paraformaldehyde. Cells used for staining of the intracellular antigens were stained according to the above-mentioned protocol, fixed in the IC fixation buffer (eBioscience) overnight and permeabilized with the help of 1× Permeabilization solution (eBioscience) according to the protocol and stained with the respective antibodies. Then the cells were washed once and fixed in 2% paraformaldehyde.
To analyze the Ly49 repertoire expressed on CD8 T-cells, CD8 T-cells were stained with a combination of monoclonal antibodies against Ly49 receptors and subdivided according to the expression of Ly49F in 2 populations: Ly49F + and Ly49F − . Then each of these cell populations was divided according to Ly49G expression and then the same for Ly49A and Ly49D +Ly49H. We identified Ly49 − CD8 T-cells as the cells, which express neither of the analyzed Ly49 receptors and excluded them from the subsequent analysis. All of CD8 T-cells that express any of Ly49 receptors (even if the cell population is very small) were summarized together and considered as 100%. The amount of CD8 T-cells that express the particular combination of Ly49 receptors were expressed as a fraction of total Ly49 + CD8 T-cells. Figure S2 describes our gating strategy.
Similar calculations were used in assessing Ly49C/I expression. The cells were subdivided according to Ly49C/I into Ly49C/I + and Ly49C/I − . Then each of these two populations was subdivided according to Ly49G expression. Ly49C/I + CD8 T-cells were used to study the co-expression pattern of the other Ly49 receptors studied (except for Ly49F) and Ly49C/I − G + CD8 T-cells were used as a control for consistency of staining and further calculations.

Cell Isolation
CD4 or CD8 T-cell enrichment was performed using magnetic enrichment kits (STEMCELL Technologies) according to the manufacturer instructions or with the help of biotinylated antibodies against the particular cell surface antigens (CD4, CD8a, CD45R, CD44, Biolegend). The cell suspension was adjusted to 100•10 6 /ml, incubated with rat serum (STEMCELL Technologies), antibodies at the desired concentration (titrated before experiments), washed three times, and incubated with MojoSort streptavidin nanobeads (Biolegend) followed by incubation on a magnet (STEMCELL Technologies). Enrichment efficacy of the desired cell fraction was assessed using flow cytometry. For fluorescence-associated cell sorting, cells were incubated with fluorochrome-labeled antibodies 20 min on ice as described and sorted (FACSAria, BD Bioscience).
Intrahepatic lymphocytes were isolated according to the commonly accepted method (21). The resulting cell suspensions were mixed with 40% isotonic Percoll and centrifuged on the gradient of 70% isotonic Percoll at 600 g RT for 30 min without break. Mononuclear cells were collected from the interphase between both layers. The cells were used for flow cytometry analysis (TCRb + CD8a + cells).
Peripheral blood lymphocytes were collected from the saphenous vein into the tubes with EDTA at the final concentration of 5 mM. The resulting suspension was additionally diluted with PBS and centrifuged on 70% isotonic Percoll for 30 min at 600 g RT without break. Mononuclear cells were collected from the interphase between the layers.

Intracellular Cytokine Staining
Cytokine production by FACS-purified CD8 T-cells from naïve mice was measured using polyclonal stimulation. Sorted cell subsets were stimulated with plate-bound anti-CD3ϵ and soluble anti-CD28 (1 mcg/ml coated overnight and 0.4 mcg/ml, respectively) or with IL-15 (50 ng/ml) for 3 days. Then the cells were re-stimulated with PMA and ionomycin (final concentration is 50 ng/ml and 500 ng/ml, respectively) or left untreated (only applicable for the cells cultured on IL-15, which were used as a control), following with the addition of Brefeldin A for 4 h.

Upregulation of Ly49 Markers on CD8 T-Cells During Homeostatic Expansion
Upregulation of Ly49 markers on CD8 T-cells during homeostatic expansion was measured by the adoptive transfer of 0.2•10 6 sorted CD8 T-cells alone or together with 0.5•10 6 sorted CD4 T-cells. FACS-purified cell subsets were injected i.v. into RAG1 KO mice. Blood collection was made monthly for 5 months from the saphenous vein. Erythrocyte-depleted PBLs were stained with fluorochrome-labeled antibodies against TCRb, CD4, CD8, CD44, and Ly49. The phenotype of T-cells was analyzed. 5 months later the mice were euthanized, splenocytes were collected, and the phenotype of T-cells was assessed.

Anti-Viral Protection
Protective functions of different subsets of CD8 T-cells were studied using the adoptive transfer of IL-15-expanded memory CD8 T-cells from LCMV-immune mice into naïve hosts. Cell subsets from LCMV-immune mice (2 months after infection) were FACS-purified and expanded on IL-15 (100 ng/ml) for 2 weeks. The cells were checked to confirm phenotype and viability and 2•10 6 cells of each type were injected i.v. into naïve WT hosts. Control mice received i.v. injection with HBSS. The same day those mice were infected with LCMV Armstrong and the mice were sacrificed 5 days later.

Upregulation of Ly49 Markers on CD8 T-Cells During the Antiviral Response
Upregulation of Ly49 markers on CD8 T-cells during the antiviral immune response was assessed by the adoptive transfer of naïve WT CD45.2 naïve CD8 T-cells alone or together with total CD45.1 CD8-depleted splenocytes into sublethally irradiated (2.5 Gy, RS-2000, Rad Source Technologies) congenic CD45.1 mice. The recipients were either left untreated or infected with LCMV Armstrong.

Statistical Analysis
Statistical analysis was conducted using Statistica (Statsoft), GraphPad (Prism) or MS Excel (Microsoft). The equality of variances between experimental groups was tested using F-test or Levene's test prior to further analysis. Each data set was subjected to two-sided Student t-test or ANOVA with subsequent Newman-Keuls post-hoc tests. We used the last observation carried forward imputations method to deal with dropouts measuring mouse body weight changes during the colitis prevention study. The data is presented as the mean ± the standard error of the mean. P-value <0.05 was considered significant.

Ly49 + CD8 T-Cells Have the Phenotype of Resting Central Memory CD8 T-Cells and Express Ly49 Receptors in a Combinatorial Way
We isolated and checked distribution and cell surface phenotype of Ly49 + CD8 T-cells from spleen, pooled peripheral lymph nodes, mesenteric lymph nodes, bone marrow, peripheral blood, and liver. Depending on the site analyzed ( Figure 1A), the amount of Ly49 + CD8 T-cells among total CD8 T-cells varies from 4 to 8% in lymph nodes and spleen with only a slightly higher percentage in the liver and significant enrichment within the bone marrow compared with the other organs analyzed ( Figure 1A). Ly49 + CD8 T-cells show quite a uniform cell surface phenotype among the organs tested, so we concentrated our attention mainly on the splenic population of Ly49 + CD8 T-cells. The cell surface phenotype of Ly49 + CD8 Tcell suggests that these cells are very close to conventional memory phenotype CD8 T-cells, not showing exhaustion or excessive cell activation ( Figure 1B). Additionally, Ly49 + CD8 Tcells do not express Nur77 expression in naïve Nur77 GFP+/− mice ( Figure 1C).
CD8 T-cells express a wide variety of combinations of Ly49 receptors ( Figure 1D, gating strategy is described in the relevant section of Materials and Methods and depicted in Figure S1) with the absolute predominance of Ly49F + cells. The next most expressed receptors are Ly49G and Ly49A, which are mostly expressed together with Ly49F (16.4% of Ly49 + CD8 T-cells are Ly49F + G + , 3.2% are Ly49F + A + , and~11.3% are Ly49F + G + A + ). Ly49C and Ly49I are the least expressed inhibitory receptors among tested. Cumulatively, 26% of total Ly49 + CD8 T-cell express activating Ly49D and Ly49H receptors mostly without other inhibitory Ly49 receptors. Only a minority of Ly49 + CD8 Tcell co-express inhibitory and activating receptors in naïve mice.
One-week-old mice show a percentage of Ly49 + CD8 T-cell comparable with adult mice, and this percentage is relatively stable throughout life ( Figure 2B). Only the advanced age mice accumulate Ly49 + CD8 T-cells significantly, but it is accompanied by a much more pronounced expansion of Ly49 − memory phenotype CD8 T-cells. The absolute number of Ly49 + CD8 T-cells grows with age gradually, but there is no selective accumulation of exactly this cell population ( Figure 2C).
To assess the involvement of autoantigens and/or cytokines on Ly49 induction, we injected lymphopenic RAG1 KO mice with naïve CD8 T-cells (nCD8) alone or together with different populations of CD4 T-cells (total peripheral CD4 T-cells (tCD4), or CD25 + CD4 T-cells, or CD25 − CD4 T-cells). RAG1 KO mice injected with total CD8 and total CD4 T-cells (tCD8/tCD4) were used as a control group. Injected T-cells repopulated RAG1 KO mice successfully and acquired CD44 (data not shown). But naïve CD8 T-cells from all of the experimental group upregulated Ly49 receptors only on a very small number of cells ( Figure 2D), while pre-existing Ly49 + CD8 T-cells (RAG1 KO mice injected with total CD8 and total CD4 T-cells) preserved Ly49 expression on their surface and mostly retained their quantity stably. Only a fraction of CD8 T-cells acquired Ly49 receptors by the end of the study period, and these were activating Ly49 receptors ( Figure 2E).

Ly49+CD8 T-Cell Do Not Have Proliferation Defects
Ly49 + CD8 + T-cells demonstrate proliferation comparable to conventional memory phenotype CD8 T-cells ( Figure 3A) and proper upregulation of CD25 and CD69 in response to anti-CD3ϵ/anti-CD28 antibodies or IL-15 in vitro ( Figure 3B). At the same time, we noticed a considerable reduction in expression of inhibitory Ly49 receptors (cells were stained with a mixture of antibodies against Ly49C/F/I/H, Ly49A, Ly49G) on the surface of Ly49 + CD8 T-cells upon TCR stimulation while stimulation with IL-15 did not have such an effect ( Figure 3B).
Upon activation, Ly49 + CD8 T-cells express proinflammatory cytokines (IL2, TNFa, and IFNg) and granzyme B at the level comparable to the conventional memory CD8 Tcells ( Figure 3C) and do not express IL10 or TGFb (data not shown).

Ly49+CD8 T-Cells Participate in the Antiviral Respons
We infected mice with LCMV Armstrong and checked T-cell phenotype and IFNg expression at different time points (3,5,8,12, and 24 days after infection). 5 days post-infection, CD8 Tcells demonstrate pronounced activation and expansion of memory CD8 T-cells and their differentiation towards the effector phenotype ( Figures 4A, C, and S2A). The relative amount of Ly49 + CD8 T-cells also increases, but modestly (from 4.25% to more than 7.5% among all CD8 T-cells, P (t) <0.05, Figures 4A, B, and S2A), and they also show effector differentiation ( Figure 4C). At this time point, the relative amount of IFNg producers among Ly49 + CD8 T-cells is higher than among conventional memory phenotype CD8 T-cells (P (t) <0.05, Figures 4D, E, and S2B). Overall, Ly49 + CD8 T-cells account for approximately 10% of total IFNg producers in response to viral peptides (data not shown) at this time point. 8 days post-infection effector CD8 + T-cells expand greatly ( Figure 4C), and Ly49 + CD8 T-cells are outnumbered by the conventional memory CD8 + T-cells ( Figure 4C and S2A). Ly49 + CD8 T-cells also contain a high proportion of IFNg producers ( Figures 4D, E, and S2B).
Ly49 + CD8 T-cells contract and reacquire CD62L expression faster than conventional memory CD8 T-cells ( Figures 4B, C). The percentage of Ly49 + CD8 T-cells approaches the one of noninfected mice only 24 days after infection ( Figure 4B). The amount of IFNg producers also decreases, but they are still detectable even 24 days after infection (1.43% of Ly49 + CD8 Tcells produce IFNg in response to gp33 comparing to 3.6% of conventional memory CD8 T-cells; Figures 4E and S2B).
To check whether the absence of Ly49 + CD8 T-cells as well as the possibility to acquire the Ly49 receptors, is beneficial for CD8 T-cells or not, we assessed the phenotype of mice completely lacking the whole cluster of Ly49 genes.
Ly49 KO mice do not show significant differences in the thymopoiesis or ratios of major peripheral T-cell subsets compared to WT ( Figure 5A). We found only marginal differences in the proliferation of naïve CD8 T-cells during the polyclonal stimulation in vitro ( Figure 5B). 8 days post-infection with LCMV Ly49 KO demonstrate the marginally less pronounced expansion of CD8 T-cells ( Figures 5C-E), acquisition of KLRG1 by CD8 T-cells ( Figure S3A), and slightly higher viral titer in the spleen on days 4 and 5 ( Figure  5F). Overall, cell dynamics and cytokine production are not perturbed in CD8 T-cells from Ly49 KO mice (Figures S3B-D).
We assessed the role of Ly49 + CD8 T-cells in the secondary immune response using the adoptive transfer of memory CD8 Tcells from LCMV-immune mice (Ly49 -CD44 + CD8 or Ly49 + CD8 cells from WT mice or total CD44 + CD8 cells from Ly49 KO mice) into naïve WT hosts, which were infected subsequently. As we intended to preserve the antigen specificity of Ly49 + CD8 T-cells avoiding the selection of particular clones, we expanded those populations of CD8 T-cell using IL-15. According to our ( Figure 3A) and previously published observations Ly49 + CD8 T-cells expand nicely in response to this cytokine (12).
Mice injected with WT Ly49 − memory phenotype CD8 Tcells and total memory phenotype CD8 T-cells from Ly49 KO mice showed robust expansion of effector CD8 T-cells ( Figures  6A-C), decreased viral titer in the spleen and liver (Figures 6D,  E) and an increased amount of IFNg producing cells in the spleen ( Figure 6F). Ly49 + CD8 T-cells from WT LCMV-immune mice did not affect the antiviral response that much strong, yet still evident. The relative increase in effector CD8 T-cells was small but evident (54.6% of effector cells among memory phenotype CD8 T-cells in HBSS-treated mice vs 61.2% in Ly49 + CD8-treated animals, P (t) <0.05; data not shown). Ly49 + CD8 T-cells of the mice injected with Ly49 + CD8 T-cells from WT LCMV-immune mice showed more active effector differentiation compared with the HBSS-treated mice (58.8% of effector cells among Ly49 + CD8 T-cells in HBSS-treated mice vs 70.8% in Ly49 + CD8-treated animals, P (t) <0.05; data not shown). Moreover, Ly49 + CD8 Tcells showed protective properties as far as their transfer reduced viral load in the spleens of infected animals compared with the control group ( Figure 6D) and increased the amount of IFNg producing cells among CD8 T-cells ( Figure 6F).

Activation-Induced Alterations of Ly49 Expression on CD8 T-Cells
Sorted populations of CD8 T-cells from naïve WT mice were stimulated with anti-CD3ϵ and anti-CD28 antibodies alone or in the presence of IL-15 or IL-2 or with cytokines only for 3 days in vitro to assess changes in Ly49 expression profile.
TCR stimulation of Ly49 + CD8 T-cells reduced the expression of inhibitory Ly49 receptors, especially Ly49F, significantly while upregulating activating receptors ( Figure 7A, Figure S4A, and Table S1). IL-15 supports the expression of Ly49F alone or together with other inhibitory Ly49 receptors (Ly49F + A + , Ly49F + G + , Ly49F + G + A + ) while keeping the expression of activating Ly49 receptors low. Ly49 + CD8 T-cells preserve the expression of Ly49F and show reduced expression of activating receptors if they are TCR stimulated in the presence of IL-15. Strikingly, Ly49 + CD8 T-cells cultured on IL-2 show the considerable accumulation of Ly49G + cells, usually in combination with Ly49F and Ly49A ( Figure 7A, Figure S4A, and Table S1). Simultaneous TCR and cytokine stimulation of Ly49 + CD8 T-cells led to the combined effect of both stimuli.
As conventional CD8 T-cells do not upregulate inhibitory Ly49 receptors massively ( Figure 3B) we grouped the cells according to the expression of any inhibitory (Ly49A, G, F, and C/I) or activating (Ly49D and H) receptors or both types together. While the conventional CD8 T-cells (naïve and Ly49memory phenotype) express low levels of inhibitory Ly49 receptors upon TCR stimulation, there is an evident increase in expression of activating Ly49 receptors ( Figures 7B, C, S4B, C). Major factors affecting the expression of activating Ly49 receptors on naïve CD8 T-cells included stimulation strength (data not shown) as well as the presence of IL-2 and IL-15 in the culture medium ( Figure 7B). Memory phenotype CD8 T-cells also upregulate activating Ly49 receptors upon TCR ligation, but we did not find strong dependence on the strength of TCR stimulation or the presence of cytokines ( Figure 7C).
To confirm our results with in vitro stimulation, we tested the impact of LCMV infection on the repertoire of Ly49 receptors expressed on CD8 T-cells. During LCMV infection, the Ly49 repertoire of CD8 T-cells changes significantly. We observed an increase in the percentage of Ly49D/H + and Ly49G + CD8 T-cells and a decrease for the other combinations of Ly49 receptors, especially those including Ly49F (Ly49F + , Ly49F + A + , Ly49F + G + , Ly49F + G + A + ; Figure 8A). IFNg producers among Ly49 + CD8 Tcells express both inhibitory and activating Ly49 receptors, but those cells expressing inhibitory receptors account for a significantly higher percentage ( Figure 8C; gating strategy is depicted in Figure 8B).
To confirm the antigen-induced upregulation of Ly49 receptors, we transferred naive CD8 T-cells alone or with the total splenocytes depleted of CD8 T-cells into sublethally irradiated congenic mice. Only a few of the transferred cells upregulated activating Ly49 receptors in non-infected mice ( Figures 9A-C) with no induction of inhibitory Ly49 receptors regardless of the presence of CD8-depleted splenocytes. Strikingly, transferred CD8 T-cells acquired activating Ly49 receptors under the infection conditions ( Figures 9A, B) and expanded actively ( Figure 9C). The presence of CD8-depleted splenocytes facilitated the expansion of Ly49DH + CD8 T-cells as well as the amount of the antigen-specific cells able to produce IFNg in response to the viral peptides ( Figure 9D) while not affecting the quantity and percentage of Ly49ACIFG + CD8 T-cells.

DISCUSSION
Our data suggest that murine CD8 T-cells upregulate Ly49 receptors upon activation. These Ly49 + CD8 T-cells lack immunosuppressive activity and, instead, arise during the antiviral response and contribute to it, although, are not critical participants. We observed stepwise acquisition of Ly49 receptors by CD8 T-cells. Overall, Ly49 expression on CD8 Tcells is not stable, but rather a dynamic process, which depends on TCR stimulation and factors secreted by the microenvironment. Common g-chain cytokines, IL-2 and IL-15, affect Ly49 expression. Organ distribution and cell surface phenotype of Ly49 + CD8 T-cells are similar to the conventional central memory CD8 Tcells, showing high expression of CD122, CD44, CD62L, and absence of markers showing activation (CD69, CD25, KLRG1, Nur77) or inhibitory receptors (Lag3, PD1, Tim3). Ly49 + CD8 Tcells mostly express Ly49F, Ly49A, and Ly49G along with a handful of expression of activating receptors, Ly49D and Ly49H. This expression profile contrasts with the Ly49 repertoire of NKcells described before, which highly express Ly49D+H alone or together with Ly49G or Ly49A and minutely express Ly49F (22,23). It is interesting as according to these data the majority of Ly49 receptors expressed on Ly49 + CD8 T-cells do not have strong MHC class I ligands among the MHC class I proteins expressed in H2 b hosts. It is a quite notable similarity between murine Ly49 + CD8 and human KIR + CD8 T-cells: receptors expressed by both cell types usually do not match with the host NK-cells (24,25). NK-cells upregulate NKRs and maturate during the contact with self-MHC class I molecules (26). Contrary to this, NKR + CD8 T-cells express the least responsive Ly49 proteins, which is similar to murine TCRb + CD8aa intestine intraepithelial lymphocytes and, to some extent, to NK T-cells [ (27,28) and our unpublished observations].
The role of Ly49 + CD8 T-cells in the immune system is uncertain. Several authors suggest that Ly49 + CD8 T-cells have immunosuppressive properties (11,12,29,30). At the same time, other studies suggest that CD8 T-cells upregulate Ly49 receptors as a response to TCR or cytokine stimulation under certain conditions, either because of exposure to autoantigens (10,16) or during the antiviral response (13,15,31,32). There is data claiming that NK1.1 + CD8 T-cells include a significant fraction of Ly49 + cells and play an important role during the early stage of the immune response, show some innate-like features (33,34) and impact the anticancer immunity significantly (35).
CD8 T-cells upregulate Ly49 receptors quite early during the response as 5 days after infection Ly49 + CD8 T-cells account for approximately 10% of total IFNg-producing CD8 T-cells (data not shown). Although the Ly49 + CD8 T-cell pool is incomparably smaller than of the conventional CD8 T-cells and contracts faster, the antigen-specific Ly49 + CD8 T-cells are detectable even 24 days post-infection. We also observed the broad antigen specificity of Ly49 + CD8 T-cells generated during the infection ( Figures 4E and S2B), which agrees with the polyclonal nature of Ly49 + CD8 T-cells (8,16). We did not perform more extensive experiments to assess the effector functions of Ly49 + CD8 T-cells due to the limited number of this cell population even during the infection. However, some reports demonstrate cytotoxicity against the virus-infected target cells (13,15).
To test whether Ly49 + CD8 T-cells support or repress the antiviral response, we isolated and expanded Ly49 + and Ly49 − memory phenotype CD8 T-cells from LCMV-immune mice, expanded them in vitro, and transferred them into the naïve syngeneic hosts. Although we were not able to distinguish the transferred cells from the host's cells, we can still judge the response and the virus load in tissues. Transferred Ly49 + CD8 Tcells not only did not suppress the response, but the mice that received them had an increased amount of effector CD8 T-cells, decreased LCMV load in the spleen, and an increased number of IFNg + CD8 T-cells upon the peptide re-stimulation ( Figures 6A,  D, F).
Whether the expression of MHC class I specific NKRs gives additional survival benefits to CD8 T-cells is a matter of debate. KIR receptors expressed on human CD8 T-cells protect the host cells from activation-induced cell death under certain conditions (4,6,7,36,37). Transgenic expression of human KIR2DL3 receptor and the corresponding MHC class I ligand HLA-Cw3 in mice facilitates the accumulation of memory phenotype CD122 + CD8 T-cells (38). Transgenic expression of Ly49 receptors on T-cell lymphoma cell lines protects them from the activation-induced cell death (39). However, CD8 T-cell-restricted Ly49 expression in mice shows conflicting results regarding the protection against activation-induced cell death (40).
We found no major differences in the phenotype peripheral T-cells in Ly49 KO mice ( Figure 5A) and observed just a slightly higher proliferation of naïve CD8 T-cells from Ly49 KO mice during the polyclonal stimulation in vitro ( Figure 5B). There were some defects in the primary antiviral response (Figures 5D, F, and S3A), but overall the response was impaired not critically. The antiviral response upon secondary re-encounter was also unimpaired ( Figure 6), which suggests that the formation of the memory pool is not impaired because of the absence of Ly49 receptors. Thus, we conclude Ly49 receptors do not grant additional major benefits to CD8 T-cells. Although, we must be careful with these results as the lack of the whole cluster of Ly49 genes might affect all the cell types, which express them (NK-cells and NKT-cells), and more experiments are required (41). The major factors affecting the upregulation of Ly49 receptors by CD8 T-cells include the immune response and STAT1 signaling (10,13,15,16,31). Immunization with MHC class II-restricted antigens also facilitates the upregulation of Ly49 receptors on CD8 T-cells, which suggests interactions with some other cells, presumably with activated T-cells (11,12,30). Further, some studies propose the involvement of CD4 T-cells as well as CD4 T-regs in the induction or survival of Ly49 + CD8 T-cell (10,42). Attenuation of IL-15 signaling in either IL-15 KO mice or mice deficient in IL-15 signaling components reduces the amount of Ly49 + CD8 T-cells in mice (18,19). Ly49 + CD8 T-cells accumulate during the culture of enriched CD8 T-cells with IL-2 in vitro (43).
The lymphopenic environment itself cannot induce Ly49 receptors on CD8 T-cells, which is comparable to the previous studies (10,15,44). We also did not observe the dramatic accumulation of Ly49 + CD8 T-cells in mice with age described by other researchers (2, 10), which might result from the differences in the animal housing conditions. These data also suggest that CD8 T-cells acquire Ly49 receptors post-thymically, and the most active accumulation of Ly49 + CD8 T-cells takes place during sexual maturation.
We found valuable clues about the regulation of Ly49 receptors in mice from the in vitro stimulation and infection studies. CD8 T-cells stimulated in vitro demonstrate dynamic regulation and reversible expression of Ly49 receptors. CD8 Tcells stimulated in vitro demonstrate dynamic regulation and reversible expression of Ly49 receptors. Ly49 + CD8 T-cells downregulate inhibitory Ly49 receptors (especially, Ly49F) and upregulate activating Ly49D+H receptors upon TCR engagement in vitro ( Figures 7A and S4A). Stimulation with cytokines (IL-2 and/or IL-15) itself or combined with anti-CD3ϵ promotes the expression of inhibitory Ly49 receptors (IL-2 promotes the expression of Ly49G and Ly49A together with Ly49F, while IL-15 mostly promotes the expression of Ly49F), but only on a fraction of CD8 T-cells, which already expressed Ly49 receptors before. We were not able to induce inhibitory Ly49 receptors on conventional CD8 T-cells under the tested in vitro condition, but we observed a measurable upregulation of activating Ly49 receptors during TCR stimulation what generally is very similar to the effects observed with Ly49 + CD8 T-cells ( Figures 7B, C, S4B, C). Importantly, the results from LCMV infection experiments support our stimulation in vitro results. A fraction of the virus-specific CD8 T-cells express both inhibitory and activating Ly49 receptors and the major fraction of these Ly49 + CD8 T-cells expresses inhibitory Ly49 receptors ( Figure 8C).
The possible regulatory mechanism of NKR expression on CD8 T-cells is still elusive. In humans, inhibitors of DNA methylation added to CD8 T-cells during the TCR stimulation could promote KIR upregulation (8). Another study demonstrated the presence of distinct transcription machinery in human NK-cells and CD8 T-cells, which supports the expression of different spectrums of KIR receptors (45). Our data suggest a stepwise acquisition of Ly49 receptors upon CD8 T-cell activation. According to this, TCR stimulation induces activating Ly49 receptors. Common g-chain cytokines (IL-2 and/ or IL-15) and interactions with microenvironment affect the Ly49 expression supporting cells bearing Ly49 receptors and promoting the acquisition of inhibitory Ly49 receptors. This process depends on CD8 T-cells itself, while also requires factors delivered from the microenvironment of CD8 T-cells as well as IL-2 and/or IL-15. Some similarities with human KIR + CD8 Tcells may be noticed (46). Later, Ly49 + CD8 T-cells can either downregulate Ly49 receptors during the following TCR CD8 T-cells that showed detectable production of IFNg in response to the combination of gp33 and np396 at 5 mcg/ml were selected and these IFNg + cells were divided according to the expression of Ly49A+C/I+F+G and Ly49D+H. Frequency of each population was determined and calculated. The representative plot is presented. (C) The relative amount of CD8 T-cells expressing the particular Ly49 receptors among CD8 Tcells producing IFNg in response to the mixture of gp33 and np396 peptides. *P (t) < 0.05 comparing to the cell that express inhibitory Ly49 receptors. The sample includes four mice.
stimulation, or the antigen-specific Ly49 + CD8 T-cells may just vanish over time. These results broaden our understanding of the mechanism regulating MHC class I NK-cell receptor expression on CD8 T-cells and may be used to additionally unveil this process.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Review Committee of Zhejiang University School of Medicine.

ACKNOWLEDGMENTS
We thank Professor Zhongjun Dong for providing us with Ly49 KO mice. We thank Professor Lilin Ye for providing us with cell and virus stocks as well as giving us valuable recommendations on establishing the LCMV experimental model.