Natural Killer Cells Promote The Generation Of Memory CD8~+T Cells In Perforin-depedent Manner

Antigen-specific immunological memory is a cardinal feature of immunity, which depends on generation and maintenance of memory T lymphocytes,B lymphocytes and plasma cells with antigen specificity.Mechanisms for generation and maintenance of memory CD8T(CD8Tm) cells are of special interests as they play a crucial role in immune responses against infections and tumors.Much attention has been paid to key molecules regulating the generation and functions of CD8Tm cells,whereas little is known about the role of populations of cells in regulation of CD8Tm cell generation and functions.Recent data demonstrated that dendritic cells(DCs) promote homeostatic proliferation of memory CD8T cells via IL-15 and IL-15Rα.However,it is unknown whether other immune cells,such as natural killer(NK) cells which are closely related to CD8T cell mediated immune responses,play a role in the regulation of CD8Tm cells.NK cells mediate fast cytotoxicity against target cells and play crucial roles in immune defense early in an infection.It has been clearly shown in recent years that activated NK cells mediate equally important immuno-modulatory functions. In innate immune response early in an infection,NK cells polarize the ensuing T cell response directly via secretion of soluble factors such as perforin,granzyme and IFN-γ.Meanwhile,NK cells influence T cell response indirectly through promoting maturation of DCs via a both cell-to-cell contact and soluble dependent manner.These data suggest that NK cells may have regulatory functions in CD8Tm generation.In our current study,we generated antigen specific CD8 T cell immune response and memory using a murine model of Listeria Monocytogenes(LM) infection.We investigated the possible roles of NK cells in the regulation of CD8Tm cell generation and the underlying cellular and molecular mechanisms.1.Natural killer cells promote CD8 Tm cell generation by regulating the extent of contraction.To generate CD8Tm cells,we adoptively transferred Thy-1.1 OT-I CD8 T cells in to host C57 mice(Thy-1.2) followed by injection of a virulance attenuated stain of Listeria Monocytogenes(ΔactA LM-OVA)via the tail veil 24h later.On day7 after infection,we detected robust expansion in number of Thy-1.1⁺ CD8T cells. On day40 after infection,donor derived Thy-1.1⁺ CD8Tm cells were detectable in various organs including peripheral blood,spleen,lymph nodes,and liver.These cells were CD44^hi,CD62L^hi/CD62L^lo in phenotype and more importantly,these cells were readily detectable for intracellular IFN-γsecretion several hours after ex vivo restimulation with OVA_257-264 peptide.Moreover,mice carrying Thy-1.1⁺ CD8Tm cells were significantly more resistant to challenge with virulant strain of OVA expressing LM(LM-OVA).Thy-1.1⁺ CD8Tm cells were detectable with similar phenotype and function 100 days after infection.These data suggest that we have generated long-lived and functional CD8Tm cells with antigen specificity.We next investigated whether NK cells are required in primary CD8 T cell response and/or CD8Tm generation by in vivo depletion of NK cells using PK136 monoclonal antibody which is specific to murine NK1.1.We gave one dose intraperitoneally of PK136 Ab or isotype control Ab on day-2 and day0, respectively.We observed a NK cell depletion of over 90%within 3 weeks since the last dose of Ab.Normal NK cell number was observed at the end of week 4. We speculated that excess dose of PK136 Ab existed in mice which depleted newly generated NK cells.We found that after NK cell depletion,the percentage and the absolute number of donor derived Thy-1.1⁺ CD8 T cells were comparable in PK136 Ab and isotype Ab treated mice on day 7 after infection,with similar phenotype(CD27,CD44,CD62L,CD122,CD127,KLRG-1)and cytokine secretion profile(IFN-γ,TNF-α,IL-2).Thus,our data demonstrated that NK cells are not required in primary CD8T cell response afterΔactA LM-OVA infection.Unexpectedly,we observed significantly lowered percentage and number of Thy-1.1⁺ CD8 Tm cells 40 days after infection.We first exclude the possibility that Thy-1.1⁺CD8T cells were depleted by PK136 Ab by showing that only marginal percentage of Thy-1.1⁺ CD8T effector cells were positive for NK1.1 staining on days 0,3,7,14,21 and 40 after infection,and PK136 Ab did not deplete Thy-1.1⁺ CD8T effector cells when injected on day 7 after infection and analyzed 24 hours later.By blood sample collection at different time points,we monitored the kinetics of the percentage of Thy-1.1⁺ CD8T cells in blood CD8⁺ T cells.Though similar percentage of Thy-1.1⁺ effector CD8T cells were generated on day 7 after infection,the percentage of Thy-1.1⁺ CD8T cells in total blood CD8⁺ T cells in NK cell depleted group was about 1/5 that in control group on day 21,and this proportion was essentially fixed in the ensuing days and weeks which resulted in reduced generation of CD8Tm cells.Intracellular detection of Bcl-2 and BrdU incorporation suggested that the survival of Thy-1.1⁺ CD8T cells was impaired in NK cell depleted mice,though proliferation of Thy-1.1⁺CD8T cells in two groups was indistinguishable.We analyzed surface marker expression as well as cytokine secretion by Thy-1.1⁺ CD8Tm cells.Though small in number,Thy-1.1⁺ CD8Tm cells generated in NK cell depleted mice were similar in surface marker expression and were equally able in IFN-γ,TNF-α,and IL-2 secretion when restimulated ex vivo with OVA_257-264peptide.After challenging withLM-OVA on day 40,we observed reduced protection of CD8Tm cells as a population in NK cell depleted mice as CFUs of LM-OVA in both the spleen and the liver were significantly higher in NK cell depleted mice.Note that NK cells recovered to a normal level on day 40,our data suggested that NK cells promote the survival of activated CD8 T cells in the contraction phase and thus help to generate a protective CD8Tm cell pool.We did further experiments to confirm that the role of NK cells on activated CD8 T cells was mediated in the contraction phase rather than in the expansion phase,where NK cells may presumably influence the intrinsic quality of CD8 T effector cells which is not detected in our experiments.Thy-1.1⁺ OT-I CD8 T cells were adoptively transferred into Thy-1.2⁺ host C57 mice treated with either isotype or PK136 Ab as mentioned above followed by injection i.v.withΔactA LM-OVA.On day 7 after infection,equal number of Thy-1.1⁺ CD8 T effector cells from either hosts were sorted and adoptively transferred into second host mice(Thy-1.2) which were also on day 7 after infection and treated with isotype or PK136 Ab,respectively.Three weeks after transfer of CD8 effector T cells, absolute number of spleen Thy-1.1⁺ CD8 Tm cells and percentages of Thy-1.1⁺ cells in CD8 T cells in peripheral blood,spleen and lymph nodes were analyzed. CD8 Tm cell generation was impaired in NK cell depleted second hosts irrespective of whether effector CD8 T cells were generated in NK cell depleted or control mice.In contrast,both effector Thy-1.1⁺ CD8T cells isolated from control or NK cell depleted mice generated normal number of CD8Tm cells after transfer into control second hosts.Thus,we provided convincing data showing that the role of NK cells in promoting CD8Tm cell generation were mediated during the contraction phase rather than in the expansion phase after primary infection.2.NK cells promote the survival of contracting CD8 T cells via a perforin dependent mechanism.We went on to ask which molecule(s) is required for NK cells in the promotion of the survival of contracting CD8T cells.Perforin is constitutively expressed by NK cells and is a key molecule in NK mediated cytotoxicity.In heat shock protein gp96-peptide complexes induced tumor rejection model,antigen specific CD8T cells(perforin competent) adoptively transffered into perforin knock out mice failed to proliferate and differentiate into CTLs whereas co-transfer of wide type NK cells rescued the clonal expansion and CTL differentiation of transferred CDST cells.These results suggested important immuno-modulatory roles of perforin and presumably different roles of perforin secreted by NK cells versus by CD8⁺ CTLs.Using the same method mentioned above,we adoptively transferred Thy-1.1 OT-I CD8 T cells into wide type B6 mice(W-F) and perforin knock out(PFR)mice on a B6 background followed byΔactA LM-OVA infection.On day 7 after infection, comparable numbers of effector Thy-1.1⁺ CD8 T cells with similar phenotype and cytokine secretion profile were generated in WT and PFR mice.However,a more extensive contraction of Thy-1.1⁺ CD8 T cells was observed in PFR mice compared with those in WT mice.Consistent with those in NK cell depleted mice, the percentage of effector Thy-1.1⁺ CD8 T cells in PFR mice was significantly lower than that in WT control group and number of Thy-1.1⁺ CD8Tm cells was significantly decreased compared with that in WT mice 40 days after infection.In accordance with those in NK cell depleted mice,Thy-1.1⁺ CD8Tm cells generated in PFR mice were similar in phenotype and cytokine secretion profile on a single cell basis.To confirm that reduced CD8Tm cell generation in PFR mice was due to perforin incompetence within NK cell population rather than in other cells that could also express perforin after infection(e.g.CTLs),we adoptively transferred WT NK cells or PFR NK cells isolated from simultaneously infected WT or PFR mice on day 7 after infection into PFR mice transferred with Thy-1.1⁺ CD8 T cells followed by infection.WT NK cells almost fully rescued generation of CD8Tm cells in PFR mice whereas PFR NK cells failed to do that.Activated CD8 CTLs isolated from infected WT mice could not rescue the generation of Thy-1.1⁺ CD8Tm cells in PFR mice,either.Thus,our data clearly showed that perforin is required in NK cell mediated promotion of CD8 Tm cell generation.This conclusion was further confirmed when effector CD8 T cells were generated in WT mice and adoptively transferred into WT or PFR second hosts on day 7 after infection.Co-injection of WT NK cells almost fully rescued CD8Tm cell generation in PFR second host mice whereas PFR NK cells did not.We analyzed the proliferation and the survival of Thy-1.1⁺ effector CD8 T cells in the contraction phase.The proliferation of effector Thy-1.1 ⁺CD8 T cells was not different as shown by BrdU incorporation.However,analysis of Bcl-2 expression showed that effector Thy-1.1 ⁺CD8 T cells were more resistant to apoptosis in VVT mice than PFR mice.And adoptive transfer of WT NK cells restored Bcl-2 level in effector Thy-1.1 ⁺CD8 T cells in PFR mice.Thus,our data demonstrate that perforin is required for NK cells in promotion of CD8Tm generation via promoting survival of contracting CD8T cells.3.NK cells promote IL-15 and IL-15Rαexpression by dendritic cells in a perforin dependent manner,outlining one of possible mechamisns for the promotion of memory CD8 T cell generation by NK cellsWe have shown in the first two parts that NK cells promote CD8Tm cell generation via promoting survival of CD8T cells in a perforin dependent manner.It is reported that a)NK cells form conjugates with DCs and promote IL-15 expression by DCs;b)IL-15,especially membrane bound IL-15 on DCs promotes the survival of CD8T cells in the contraction phase.Based on our experimental data and these reports,we speculate that during the contraction phase of CD8T cell immune response,NK cells may promote IL-15 expression by DCs in a perforin dependent manner and thus indirectly promote the generation of CD8Tm cells via DCs.We first analyzed IL-15Rαand membrane bound IL-15 expression by murine splenic DCs during the contraction phase.Using FACS analysis of splenocytes,we observed reduced IL-15Rαand membrane bound IL-15 expression by CD11c^hila^b+ splenic DCs in both NK cell depleted WT mice and PFR mice when compared with WT control mice.More importantly,adoptive transferred activated WT NK cells restored IL-15Rαand membrane bound IL-15 expression by CD11c^hila^b+ splenic DCs in PFR mice,whereas PFR NK cells failed to restore IL-15Rαand IL-15 on DCs.By using fluorenscent immunohistochemical analysis,we observed that during the contraction phase of primary CD8T cell immune response,NK cells and DCs were localized closely in the area between T cell zones and B cell zones.NK-DC juxtaposition could be visualized which indicated direct cell-cell contact between NK cells and DCs in the contraction phase.We observed that after stimulation with IL-2 and IL-15 in vitro,perforin protein was evenly distributed within cytoplasma of NK cells derived from wild type mice but not PFR mice.When cocultured with bone marrow-derived matured DCs (mDCs)for 1 hour,these in vitro activated NK cells form conjugates with mDCs and more importantly,perforin was centered to the site of NK-DC synapse.These data indicate that NK cells could have certain impact on mDCs via perforin secretion.Activated wild type NK cells,but not PFR NK cells could promote mDCs to express IL-15Rαand membrane bound IL-15 after coculture in vitro.When wild type NK cells and PFR NK cells were mixed in a 1:1 ratio,they could promote IL-15Rαand membrane bound IL-15 expression by DCs to the same level as wild type NK cells.Considering that IL-15 is crucial in the generation of CD8Tm cells, these data indicate that enhancement of IL-15 and IL-15Rαexpression on DCs by NK cell-drerived perforin might be one of mechanisms by which NK cells promote CD8Tm cell generation.In conclusion,our data demonstrate that NK cells promote the generation of CD8Tm cells via promoting the survival of activated CD8T cells in the contraction phase after primary immune responses.NK cell-derived perforin is required in the promotion of CD8Tm cell generation by NK cells.We have also shown that NK cells form conjugates with DCs and promote IL-15 and IL-15Rαexpression by DCs in a perforin dependent manner both in vitro and in vivo.Thus,NK cells promote CD8Tm cell generation presumbly via promoting the expression of IL-15Rαand membrane bound IL-15 on DCs.In CD8Tm cell-mediated immune responses against infections,the quantity and quality of a pathogen/antigen specific CD8 Tm cell pool are key factors that determine whether the host could survive or die of uncontrolled infection.The generation of protective CD8 T cell immunity is also crucial in inoculation with inactivated or virulance attenuated vaccines to fight against possible infections. Thus,exploring the cellular as well as molecular mechanisms that regulate virulance attenuated pathogen-induced CD8Tm cell generation quantitively and qualitively will be valuable both theoretically and practically.In our current study, we provide with a novel perspective in respect to a promoting role of NK cells in the generation of CD8Tm cells.And our data may provide theoretical basis for better immunotherapy as well as vaccination design in the future.