Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Immunol. Author manuscript; available in PMC 2010 October 8.
Published in final edited form as:
PMCID: PMC2951739

Interleukin 15–mediated survival of natural killer cells is determined by interactions among Bim, Noxa and Mcl-1


Interleukin 15 (IL-15) promotes the survival of natural killer (NK) cells by preventing apoptosis through mechanisms unknown at present. Here we identify Bim, Noxa and Mcl-1 as key regulators of IL-15-dependent survival of NK cells. IL-15 suppressed apoptosis by limiting Bim expression through the kinases Erk1 and Erk2 and mechanisms dependent on the transcription factor Foxo3a, while promoting expression of Mcl-1, which was necessary and sufficient for the survival of NK cells. Withdrawal of IL-15 led to upregulation of Bim and, accordingly, both Bim-deficient and Foxo3a−/− NK cells were resistant to cytokine deprivation. Finally, IL-15-mediated inactivation of Foxo3a and cell survival were dependent on phosphotidylinositol-3-OH kinase. Thus, IL-15 regulates the survival of NK cells at multiple steps, with Bim and Noxa being key antagonists of Mcl-1, the critical survivor factor in this process.

Natural killer (NK) cells contribute to innate immune responses through the lysis of damaged cells and the production of cytokines and chemokines1,2. NK cells recognize and rapidly respond to stressed or transformed cells by means of receptors that detect a lack of expression of major histocompatibility complex class I (ref. 3) and/or the expression of stress-inducible ligands. These attributes make NK cells important effectors in the clearance of virus-infected cells and certain tumors2,4. Given the importance of NK cells to immunity in mammals, identifying mechanisms that mediate the survival of NK cells might assist in improving strategies for the immunotherapy of infectious diseases or cancer.

The lack of rearranged antigen receptors on NK cells means that unlike B cells and T cells, NK cells must rely on factors other than antigen for their development, activation and survival. One such factor is the pleiotropic cytokine interleukin 15 (IL-15). Mice deficient in IL-15 (Il15−/−) lack peripheral NK cells5, as do mice deficient in any one of the three IL-15 receptor (IL-15R) subunits (α, β or γ)68 or those lacking the IL-15R signaling molecule Jak3 (refs. 9,10). IL-15 is critical not only for the development of NK cells but also for their survival in vivo, something best demonstrated by the observation that mature NK cells fail to survive when transferred into Il15−/− mice11,12.

Cell survival is controlled to a large extent by the balance between pro- and antiapoptotic members of the Bcl-2 protein family. Apoptosis signaling is initiated by the transcriptional or post-translational activation of proapoptotic Bcl-2 homology domain (BH3)–only proteins, such as Bim, Noxa and Puma13. These proteins bind to antiapoptotic Bcl-2 family members, such as Bcl-2 itself, Bcl-xL and Mcl-1. This step is thought to ‘unleash’ the proapoptotic Bax and/or Bak proteins from their inactive state, leading to permeabilization of the outer mitochondrial membrane, release of apoptogenic molecules (such as cytochrome c) and activation of the caspase cascade14.

Bcl-2 has been proposed to be the critical factor by which IL-15 promotes the survival of NK cells, because IL-15 has been reported to maintain or increase Bcl-2 expression in human or mouse NK cells, respectively11,15,16, and overexpression of Bcl-2 inhibits apoptosis of NK cells after their transfer into Il15−/− mice. However, most resting NK cells have high constitutive expression of Bcl-2 (refs. 11,15) and experience only a slight decrease in Bcl-2 while undergoing apoptosis16, indicating that Bcl-2 may have only limited involvement in the survival of NK cells. Although so far none of the BH3-only proteins have been suggested as being critical initiators of NK cell apoptosis, Bim is a likely candidate because it is essential for the homeostasis of hematopoietic cells and cytokine deprivation–induced apoptosis of B lymphocytes and T lymphocytes17, osteoclasts18 and mast cells19. The phenotype of Bim-deficient lymphocytes has been reviewed20.

Here we report the molecular basis by which IL-15 mediates the survival of NK cells. We found that IL-15 stimulation maintained Mcl-1 expression in NK cells and that IL-15 was still able to promote the survival of NK cells even when Bcl-2 and Bcl-xL were blocked specifically by the BH3-mimetic compound ABT-737 but not when Mcl-1 was inhibited by the selective BH3 ligand Noxa. Moreover, Bim-deficient (Bcl2l11−/−; called ‘Bim−/−’ here) NK cells accumulated at a late maturation stage in vivo and were resistant to apoptosis caused by IL-15 deprivation. IL-15 inhibited Bim activation at many levels, including transcription initiation through phosphotidylinositol-3-OH kinase (PI(3)K)–dependent inactivation of transcription factor Foxo3a, and promoted proteasome-mediated degradation of Bim through activation of kinases Erk1 and Erk2 (called ‘Erk1/2’ here). Our results elucidate the pathway by which IL-15 regulates the survival of NK cells and identify Bim, Mcl-1 and, to a lesser extent, Noxa as the key factors in this process.


IL-15 blocks Bim accumulation in NK cells

To delineate the prosurvival versus mitotic effects of IL-15 on NK cells, we first established a concentration of IL-15 (5 ng/ml) that did not induce NK cell proliferation but was sufficient to maintain viability above 95% for NK cell populations previously expanded with IL-15. NK cells cultured in the absence of IL-15 rapidly underwent apoptosis, with over 80% dead by 24 h (Fig. 1a). Given the critical function of the BH3-only protein Bim in the cytokine deprivation–induced apoptosis of lymphocytes, myeloid cells, osteoclasts and mast cells20, we determined whether Bim was regulated by IL-15 in NK cells. Although we readily detected BimEL and BimL, the most highly expressed isoforms of Bim21, in IL-15-stimulated NK cells by immunoblot analysis, there was an increase in both isoforms after 2 h and 4 h of IL-15 deprivation, whereas Bcl-2 remained constant (Fig. 1b).

Figure 1
IL-15 deprivation causes accumulation of Bim in NK cells

We next investigated the signaling pathways activated in NK cells by IL-15 that may have been connected with the regulation of Bim. IL-15 stimulation activated the mitogen-activated protein kinase and PI(3)K pathways in NK cells, with increases in phosphorylated c-Akt (protein kinase B) and phosphorylated Erk1/2 (p44/p42) within 30 min of exposure to IL-15 (Fig. 1c). During this interval, the amount of Bim in the IL-15 stimulated NK cells did not vary much. The Erk inhibitor U0126 completely blocked phosphorylation of both Erk1/2 and Bim (BimEL as well as BimL), indicating that Bim phosphorylation probably occurs ‘downstream’ of Erk1/2 in IL-15-stimulated NK cells (Fig. 1d) and that in NK cells, as in osteoclasts18, Bim may be a substrate for Erk. Moreover, Bim increased in IL-15-stimulated NK cells treated with the proteasome inhibitor PS341 (Fig. 1e), indicating that IL-15 primes Bim for proteasomal degradation, possibly through Erk-mediated phosphorylation and subsequent ubiquitination, as has been described for lymphocytes22. NK cells cultured in the presence of IL-15 and the Erk1/2 inhibitor U0126 were much less viable than cells cultured with IL-15 alone, confirming that mitogen-activated protein kinase activity contributes to the IL-15-mediated survival of NK cells (Supplementary Fig. 1a online).

Bim is not essential for NK cell effector function

Next we investigated the consequences of loss of Bim on the development and function of NK cells. The percentage of NK cells in the spleens of Bim−/− mice was lower than that of control (wild-type) mice; however, given the greater cellularity of Bim−/− spleens17, the total number of NK cells was similar, as it was in the liver and bone marrow (Figs. 2a,b). Bim-deficient NK cells proliferated normally in response to IL-15; they also responded normally to IL-18 and IL-12, as assessed by production of interferon-γ (IFN-γ) and modulation of the expression of NK1.1 or KLRG1, respectively (Figs. 2ce). Moreover, cytolysis of major histocompatibility complex class I–deficient target cells and NKG2D ligand–positive target cells was unaffected by loss of Bim (Fig. 2f), indicating that Bim is not required for normal NK cell function.

Figure 2
Bim is not essential for development or function of NK cells

Altered NK cell maturation and IL-15-regulated

As NK cell numbers were normal in Bim−/− mice, we next investigated the maturation state of these NK cells. Mature (Mac-1+) NK cells can be categorized into subsets on the basis of their expression of CD27 and KLRG1, with KLRG1CD27+ NK cells considered precursors to KLRG1+CD27 NK cells2325. Notably, Bim−/− mice had a significantly higher percentage and total number of KLRG1+ NK cells in the spleen, liver and bone marrow than did wild-type mice (P < 0.02; Fig. 3a,b and data not shown). To determine whether the accumulation of KLRG1+ NK cells in Bim−/− mice correlated with impaired apoptosis, we cultured wild-type and Bim−/− NK cells in the presence or absence of IL-15. Bim−/− NK cell populations expanded with IL-15 survived cytokine withdrawal significantly better than wild-type NK cells did (P < 0.0008; Fig. 3c). We obtained similar results with freshly isolated Bim−/− and wild-type NK cells (Supplementary Fig. 1b).

Figure 3
Apoptosis-resistant NK cells accumulate in Bim−/− mice

To determine if such differential survival after IL-15 withdrawal occurred in vivo, we transferred equal numbers of Bim−/− and wild-type NK cell populations expanded in vitro into Il15−/− recipient mice and monitored their persistence over 72 h (Fig. 3d). During this time essentially all wild-type NK cells disappeared and thus Bim−/− NK cells comprised over 90% of the remaining population. The same ‘preferential’ survival of Bim−/− cells occurred when we used primary NK cells isolated from spleen, although the kinetics of loss were different in that a greater fraction of wild-type NK cells survived at 24 h.

NK cells are important in controlling many viral infections. Infection of mice with mouse cytomegalovirus (MCMV) has provided useful insights into NK cell biology; in MCMV-resistant mouse strains, such as C57BL/6J, MCMV infection is controlled by activities mediated by NK cells26. Notably, NK cell numbers expand in a MCMV-specific way27 before contracting to homeostatic numbers by day 14 after infection24. Thus, we next examined NK cell dynamics in the spleens of wild-type and Bim−/− mice after infection with MCMV. At both day 10 and day 14 after infection, Bim−/− mice retained significantly more NK cells than wild-type mice did (P < 0.01; Fig. 3e). Furthermore, the phenotype of the expanded NK cell population differed in wild-type versus Bim−/− mice, with the latter showing a greater frequency of KLRG1+ NK cells (Fig. 3e). These results indicate that Bim is a key regulator of NK cell homeostasis and that enhanced survival of NK cells in vivo resulting from loss of Bim leads to the accumulation of late-stage KLRG1+ NK cells.

PI(3)K facilitates IL-15-mediated NK survival

Given our finding that IL-15 activated the PI(3)K pathway in NK cells (Fig. 1c) and the importance of PI(3)K in lymphocyte signaling, homeostasis and survival28, we investigated the requirement for PI(3)K in the IL-15-mediated survival of NK cells. LY294002, a PI(3)K inhibitor29, reduced the survival of NK cells cultured in IL-15 in a dose-dependent way (Fig. 4a and Supplementary Fig. 1a) and required Bim, as Bim−/− NK cells were significantly more resistant to LY294002-mediated apoptosis (P < 0.003; Fig. 4b). As anticipated, LY294002 decreased phosphorylated Akt in IL-15-stimulated NK cells without affecting the phosphorylation of Erk1/2 (Fig. 4c). Notably, LY294002 treatment of IL-15-stimulated NK cells increased Bim in cultured NK cells (Fig. 4c), suggesting a link between PI(3)K activity and Bim stability. Notably, inhibition of PI(3)K activity in IL-15-stimulated NK cells did not affect the amount of Mcl-1 (Fig. 4c). To further delineate the requirement for PI(3)K in the IL-15-induced survival of NK cells, we analyzed mice with an inactivating mutation in the p110δ subunit of PI(3)K28. NK cells with this inactivating mutation developed normally, proliferated in the presence of IL-15 and died after IL-15 withdrawal exactly as wild-type NK cells did (Supplementary Fig. 2a–c online). Finally, treatment with LY294002 in the presence of IL-15 killed NK cells with this inactivation mutation as effectively as it killed wild-type NK cells (Supplementary Fig. 2d). These experiments collectively indicate that PI(3)K signaling ‘downstream’ of the IL-15R complex does not require the p110δ subunit.

Figure 4
PI(3)K is essential for IL-15-mediated survival of NK cells

IL-15 suppresses Bim in NK cells via PI(3)K-Foxo3a

Given that the inhibition of PI(3)K increased Bim protein concentrations and blocked IL-15-mediated survival of NK cells in a Bim-dependent way (Fig. 4), we next investigated whether PI(3)K-Akt signaling regulated Bim expression in IL-15-stimulated NK cells. Foxo3a (also called FKHR-L1), a member of the family of Forkhead box class O transcription factors, increases Bim transcription in certain hematopoietic cells when they are deprived of their requisite growth factors30. In cytokine-stimulated cells, Foxo3a is inactivated through cytosolic sequestration after Akt-mediated phosphorylation30,31. Indeed, treatment of NK cells with IL-15 increased Foxo3a phosphorylation, and this was blocked by LY294002 in a concentration-dependent way, indicating a relationship between PI(3)K and Foxo3a activity in NK cells (Fig. 5a). Mice deficient in Foxo3a (Foxo3a−/− mice) had normal numbers of NK cells in the spleen that seemed grossly normal in phenotype except for a small but significant increase in the fraction that were KLRG1+ (less than that in Bim−/− mice; P < 0.01; Fig. 5b). Notably, after IL-15 withdrawal in culture, Foxo3a−/− NK cells failed to upregulate Bim (Fig. 5c) and survived significantly better than wild-type NK cells (albeit less well than Bim−/− NK cells; P < 0.05; Fig. 5d). Moreover, LY294002 induced significantly less apoptosis of Foxo3a−/− NK cells than of wild-type NK cells (P < 0.01; Fig. 5e), indicating that PI(3)K contributed to the survival of NK cells in part through the inhibition of Foxo3a-mediated Bim transcription.

Figure 5
IL-15 suppresses Foxo3a-mediated Bim upregulation

IL-15-mediated maintenance of Mcl-1 and NK cell survival

Having identified Bim as the key proapoptotic protein after IL-15 withdrawal, we next sought to identify the prosurvival Bcl-2 family members that were responsible for the IL-15-mediated survival. It has been shown that Mcl-1 is critical for the survival of B cells and T cells in vivo and that IL-15 is a potent inducer of Mcl-1 in T cells32. Consistent with those findings, we noted that IL-15 was required for the maintenance of Mcl-1 in NK cells, as IL-15 withdrawal resulted in rapid loss of Mcl-1 (Fig. 6a). We assessed the importance of the various prosurvival Bcl-2 family members in IL-15-mediated survival of NK cells with two specific inhibitors: ABT-737, the BH3 mimetic that inhibits Bcl-2, Bcl-xL and Bcl-w but not Mcl-1 (ref. 33), and enforced expression of Noxa, which acts selectively on Mcl-1 but not on Bcl-2, Bcl-xL or Bcl-w34. Treatment of NK cells with ABT-737 had no effect on their survival in the presence of IL-15, although it accelerated NK cell death after IL-15 withdrawal (Fig. 6b). Conversely, enforced expression of Noxa by retroviral gene transduction led to efficient killing of primary NK cells in the presence of IL-15. This finding was typified by the much lower frequency of cells positive for green fluorescent protein (GFP), a marker of retroviral infection, in Noxa-transduced NK cell cultures than in cultures of NK cells transduced with retrovirus encoding only GFP or GFP plus Noxa-3E, the inactive mutant form of Noxa (P < 0.025; Fig. 6c). Furthermore, GFP expression in the remaining Noxa-transduced NK cells was much lower than that of cells infected with the mutant Noxa-3E construct, consistent with selection against high Noxa expression in these cells (Fig. 6d). Notably, control fibroblasts refractory to the proapoptotic effects of Noxa34 expressed similar amounts of retrovirally transduced Noxa or Noxa-3E (data not shown). These results indicate that Mcl-1 was critical for the survival of NK cells in the presence of IL-15 and that other prosurvival Bcl-2 family members had minor involvement, perhaps contributing to the survival of NK cells when IL-15 was scarce and Mcl-1 abundance diminished.

Figure 6
IL-15-induced upregulation of Mcl-1 promotes the survival of NK cells

Noxa is upregulated after IL-15 withdrawal

Involvement of Noxa itself in NK cell apoptosis was suggested both by the importance of Mcl-1, a known target of Noxa34,35, in the survival of NK cells and the proapoptotic effects of Noxa overexpression. Indeed, real-time PCR analysis showed that Noxa mRNA was significantly upregulated in NK cells after IL-15 withdrawal (P < 0.04; Fig. 7a). We therefore investigated the consequences of loss of Noxa on NK cell numbers in vivo and the survival of NK cells in vitro. Although Noxa-deficient mice (Pmaip1−/−; called ‘Noxa−/−’ here) had normal numbers of NK cells in all organs analyzed (data not shown), both freshly isolated Noxa−/− NK cells (Supplementary Fig. 3a online) and Noxa−/− NK cell populations expanded in vitro survived significantly better than did wild-type NK cells when cultured without IL-15 (P < 0.05; Fig. 7b). Furthermore, a combined loss of Bim plus Noxa protected NK cells against IL-15 deprivation in vitro significantly better than did loss of Bim alone (P < 0.015; Fig. 7b). This effect of Noxa loss was specific, as Bad−/− NK cells died normally in the absence of IL-15 and loss of Bad did not enhance the resistance of Bim-deficient NK cells to IL-15 deprivation (Supplementary Fig. 3b). These results demonstrate that Noxa acts together with Bim in inducing the apoptosis of NK cells after IL-15 withdrawal.

Figure 7
Noxa acts together with Bim in mediating NK cell apoptosis


Unlike B cells and T cells, the development and survival of NK cells is not governed by signals generated from classic antigen receptors but instead relies on the cytokine IL-15 (ref. 36). The proapoptotic BH3-only protein Bim is known to be essential for developmentally ‘programmed’ and cytokine deprivation–induced death of B cells and T cells20. Very little is known about the control of NK cell survival and the function that Bim might have in this process. Here we have identified Bim and the pathways that regulate its expression as the key molecular targets of the IL-15-mediated survival of NK cells.

The absolute dependence of NK cells on IL-15–IL-15R signaling and the ease with which they are grown in vitro have made these cells ideal for investigating pathways involved in cytokine-induced survival. The balance between prosurvival proteins, such as Bcl-2, Bcl-xL, Bcl-w and Mcl-1, and proapoptotic proteins, such as Bim, Bid, Noxa, Puma and Bad, controls lymphocyte survival20. Our main finding here was that stimulation with IL-15 limited the amount of Bim in NK cells by activating two distinct signaling pathways. First, Bim phosphorylated by an Erk1/2-dependent process was targeted for degradation by the proteasome. Second, IL-15 activated PI(3)K and its ‘downstream’ target Akt, which in turn phosphorylated Foxo3a, thereby preventing it from upregulating Bim transcription30,31. IL-15 therefore joins IL-2 and IL-3 as a cytokine that regulates leukocyte survival by inactivating Foxo3a by means of PI(3)K-Akt–mediated phosphorylation37,38. Notably, the p110δ subunit of PI(3)K was not required for the inhibition of Foxo3a activity, repression of Bim expression or survival of NK cells, although it is essential for normal survival of B cells and T cells28. This result indicates that p110δ is critical in survival signaling from antigen receptors but not that from cytokine receptors.

Noxa and Puma are both BH3-only proteins that can be activated by p53-dependent apoptotic stimuli (such as DNA damage) as well as p53-independent apoptotic stimuli. Notably, our study has demonstrated that Noxa was also induced after IL-15 withdrawal and paralleled the induction of Noxa and resulting apoptosis seen in human T cells after glucose deprivation39. Puma is critical for apoptosis of myeloid cells induced by IL-3 deprivation40 and it is upregulated in T cells in a Foxo3a-dependent way when IL-2, which uses the same β- and γ-receptor subunits as IL-15, is withdrawn41. This suggests that Puma may also be involved in NK cell apoptosis and demonstrates that Noxa and Puma do not function exclusively in response to DNA damage but also participate in apoptosis elicited by the withdrawal of cytokines or other growth factors.

Given the critical involvement of Bim in apoptosis induced by IL-15 deprivation and the importance of IL-15 in NK cell development, the finding of normal numbers of NK cells in Bim−/− mice was unexpected. In contrast, the development of B cells and T cells is greatly affected by the absence of Bim, with Bim−/− mice having two- to threefold more B cells and T cells than control mice have17. Similarly, mice overexpressing Bcl-2 in all hematopoietic cells accumulate B cells and T cells but have normal numbers of NK cells, although these are also very resistant to IL-15 deprivation (N.H. and A.S., unpublished data). There are many potential explanations for the lack of NK cell accumulation in Bim−/− and Bcl2-transgenic mice. First, it is possible that abnormal population expansion of B cells and T cells limits the availability of IL-15, thereby preventing abnormal accumulation of NK cells. Alternatively, it is possible that although NK cell numbers are normal in Bim−/− and Bcl2-transgenic mice, their lifespan is abnormally long and, as a compensatory process, their production in vivo is lower. Such a scenario has been noted for Bim−/− osteoclasts, which are resistant to cytokine deprivation but do not accumulate in vivo18. Notably, although Bim−/− mice have normal overall numbers of NK cells, they accumulate at a late developmental stage, characterized as KLRG1+CD27, and are mostly post-mitotic23,24. Notably, during the early stages of MCMV infection, proliferation of NK cells results in the accumulation of KLRG1+ cells, which ‘preferentially’ downregulate Bcl-2 and undergo apoptosis during the contraction phase of the response24. The accumulation of KLRG1+ NK cells in Bim−/− mice could therefore reflect the normal upregulation of Bim in KLRG1+ NK cells in response to diminished IL-15R signaling. It will be useful to compare the lifespans of terminally differentiated Bim−/− and wild-type NK cells and the production of such cells from their progenitors.

There was also abnormal accumulation of KLRG1+ NK cells in Foxo3a−/− mice, although to a lesser extent than that in Bim−/− mice. This finding suggests that Foxo3a is responsible for a proportion of the Bim activation required for normal NK cell turnover in vivo. Notably, although Bim still underwent dephosphorylation after IL-15 withdrawal in Foxo3a−/− NK cells, Bim did not increase appreciably. This result suggests that transcriptional induction has a dominant function in regulating Bim concentrations during IL-15 deprivation and posttranslational modifications have only a minor function. Our finding that Foxo3a−/− NK cells were not as resistant to IL-15 deprivation as Bim−/−NK cells were, however, demonstrates that the amount of Bim in Foxo3a−/− NK cells was sufficient to trigger apoptosis when the concentration of Mcl-1 decreased.

Although Mcl1−/− mice are not viable42, lineage-specific deletion of Mcl-1 has demonstrated a requirement for this molecule in the development and sustained survival of B lymphocytes and T lymphocytes32. There are no reports at present detailing a requirement for Mcl-1 in the survival or development of NK cells, presumably because mice expressing Cre recombinase under the control of an NK cell–specific promoter are not yet available. However, the requirement for IL-15 in maintaining Mcl-1 concentrations in NK cells and the upregulation of Mcl-1 in IL-15-stimulated thymocytes32 indicate that Mcl-1 may also be essential for the survival of NK cells. In support of that hypothesis, Noxa, which binds only Mcl-1 and Bcl-z-related protein A1 (refs. 34,35), was upregulated in IL-15-deprived NK cells, and enforced Noxa expression killed NK cells even in the presence of IL-15. Finally, Noxa-deficient NK cells were abnormally resistant to IL-15 withdrawal. These observations, along with the finding that combined inactivation of Bcl-2, Bcl-xL and Bcl-w with the BH3 mimetic ABT-737 had no effect on the survival of IL-15-stimulated NK cells, indicate that Mcl-1 is probably the crucial inhibitor of apoptosis in NK cells.

On the basis of the findings reported here for NK cells and elsewhere for osteoclasts17 and lymphocytes22, we suggest that IL-15 maintains small amounts of Bim in NK cells by two complementary mechanisms. First, IL-15 activates the Erk1/2-mediated phosphorylation, ubiquitination and proteasomal degradation of Bim. Second, IL-15 represses transcription of Bim through the PI(3)K-Akt–mediated phosphorylation of Foxo3a. The small amounts of Bim coupled with the high expression of Mcl-1 after IL-15 exposure shift the balance toward survival. In contrast, withdrawal of IL-15 results in both inactivation of Erk1/2 and PI(3)K and accumulation of Bim in the cytoplasm. Noxa is also upregulated and, together with Bim induction and the decreasing abundance of Mcl-1, shifts the balance toward cell death.

Given that NK cells are critical for host defense against many infectious pathogens and transformed cells, understanding the mechanisms that govern the survival of NK cells could improve immunotherapeutic approaches to such diseases. Our finding that Bim and Mcl-1 are critical proapoptotic and antiapoptotic proteins, respectively, will allow research to focus on targeting these molecules in efforts to manipulate NK cell numbers and activity in vivo. For example, our finding that ABT-737 does not kill NK cells (even at concentrations used in preclinical cancer studies in mice) might have implications for the immunotherapy of cancer, as NK cell–mediated cytotoxicity would not be expected to be affected by treatment with ABT-737, as long as the concentration of IL-15 in vivo was not compromised.



C57BL/6, Bim−/− mice17, Noxa−/− mice43, Bim−/−Noxa−/− mice (E.M.M. and A.S., unpublished data), Bad−/− mice44, Bim−/−Bad−/− mice45, Foxo3a−/− mice (N.M., unpublished data), Il15−/− mice5 and mice with an inactivating mutation in the p110δ subunit of PI(3)K28 were bred and maintained at the Walter and Eliza Hall of Medical Research or the Ludwig Institute for Cancer Research (Melbourne, Australia). Mice of all mutant strains were generated on the C57BL/6 background or were backcrossed to this background for at least eight generations, except Foxo3a−/− mice, which were backcrossed for five generations. The relevant Animal Ethics and Experimentation Committees approved animal experiments according to the guidelines of the National Health and Medical Research Council Australia.

Antibodies, flow cytometry and cell sorting

Antibodies specific for NK1.1 (PK136), Ly49C/I (SW5e6), CD11b (Mac-1; M1/70), TCR-β (H57-5921), IFN-γ (HB170 and XMG1.2), KLRG1 (2F1) and CD49b (DX5 and HMα2) were from BD PharMingen. Antibody to CD27 (anti-CD27; LG.7F9) was from eBioscience. Single-cell suspensions were prepared by forcing of organs through metal sieves. Lymphocytes from liver were isolated from a 40–80% Percoll gradient (Amersham Pharmacia Biotech) centrifuged for 20 min at 1,300g. For flow cytometry, single-cell suspensions were stained with the appropriate monoclonal antibody in PBS containing 2% (vol/vol) FCS. Biotinylated monoclonal antibodies were visualized with indodicarbocyanine- or phycorythrin-streptavidin (Southern Biotechnology Associates). LSR, FACSDiva (BD Biosciences) and MoFlo (Cytomation) apparatuses were used for cell sorting and analysis, with dead cells excluded by propidium iodide staining.

In vitro analysis of NK cell function

NK cell proliferation was assayed by culture for 72 h of 3 × 104 cells in 100 µl of Iscove’s modified Dulbecco’s medium supplemented with 10% (vol/vol) FCS plus gentamycin (50 ng/ml; Sigma) and a ‘titration’ of recombinant human IL-15 (0.5–50 ng/ml), mouse IL-21 (100 ng/ml), IL-12 (2 ng/ml) and IL-18 (10 ng/ml; R&D Systems), in 96-well, flat-bottomed plates. Plates were then pulsed with 1 µCi [3H]thymidine (NEN) and samples were collected after 8 h onto glass-fiber filters (Packard), followed by scintillation counting. Survival was assessed by culture for 72 h of freshly sorted or IL-15-expanded NK cell samples at a density of 3 × 104 cells per well or 1 × 106 cells per well in 100 µl of Iscove’s modified Dulbecco’s medium plus FCS, in 96-well, flat-bottomed plates. Live and dead cells were discriminated by staining with propidium iodide and annexin V–fluorescein isothiocyanate (BD PharMingen) followed by analysis on a FACScan (Becton-Dickinson). Induction of KLRG1 expression was analyzed by expansion of sorted NK cell populations in IL-15 (50 ng/ml) with cells at a density of 1 × 106 cells per ml in Nunclon six-well flat-bottomed tissue culture plates (Nunc). After 5 d, IL-12 (2 ng/ml) and IL-18 (10 ng/ml) were added alone or in combination and KLRG1 expression was monitored by flow cytometry after 24 and 48 h. Analysis of IFN-γ production and cytotoxicity assays were done as described46. The kinase and proteasome inhibitors U0126, LY294002 and PS341 were dissolved in dimethyl sulfoxide and were added for 6–24 h. The BH3 mimetic ABT-737 was from Abbot Laboratories.

Immunoblot analysis

Sorted NK cells were grown for 5–8 d in six-well tissue culture plates with IL-15 (50 ng/ml). For IL-15 withdrawal, 5 × 106 NK cells were washed twice and were cultured for various times in Iscove’s modified Dulbecco’s medium plus FCS. Cells were collected and were washed in PBS and total cell lysates were prepared in lysis buffer (20 mM Tris–HCl, pH 7.4, 135 mM NaCl, 1.5 mM MgCl2, 1% (vol/vol) Triton X-100 and 10% glycerol) supplemented with Pefabloc (0.5 mg/ml) and 1 mg/ml each of leupeptin, aprotinin, soya bean trypsin inhibitor and pepstatin (Sigma). Proteins were separated by electrophoresis through either 12% or 4–20% gradient Novex gels (Invitrogen) and were transferred to Hybond-C extra nitrocellulose membranes (Amersham) with an Excell II blot module (Novex, Invitrogen). Proteins were detected as described47. Nuclear proteins were obtained with the NucBuster Extraction kit (Novagen). Antibodies used included rat monoclonal anti-Bim (3C5; Alexis Biochemicals), rabbit anti-Mcl-1 (600-401-394; Rockland Immunochemicals; hamster monoclonal anti-Bcl-2 3F11; BD Pharmingen); mouse monoclonal anti-Hsp70 (N6); rabbit polyclonal antibodies to phosphorylated Akt (9271), total Akt (9272), phosphorylated Erk1/2 (9102) and total Erk1/2 (9101; all from Cell Signaling); antibodies to phosphorylated Foxo3a (06–952) and total Foxo3a (06–951; both from Upstate); and anti-β-actin (sc-7210; Santa Cruz Biotechnology). Band intensities were measured by densitometry (Molecular Dynamics) and were analyzed with ImageQuant v5 software. Densities are expressed as the ratio of staining relative to staining with anti-β-actin and the ‘fold change’ in this ratio relative to that of unstimulated samples.

Cell transfer

NK cell populations from control (wild-type) or Bim−/− bone marrow were expanded for 7 d with IL-15 and then were labeled with CFSE (carboxyfluorescein succinimidyl diester) at a concentration of 10 µM or 1 µM, respectively. Labeled wild-type and Bim−/− NK cells were mixed at near-equal numbers and were placed in culture with or without IL-15 (10 ng/ml) or were injected into Il15−/− recipient mice. Survival of NK cells was determined 24 and 72 h later by analysis of cultures and spleens of recipient mice to determine the ratio of Bim−/− to wild-type NK cells, identified as CFSE+CD49b+. The proportion of persisting wild-type NK cells was determined as follows: (percent wild-type NK) / (percent wild-type NK + percent Bim−/− NK). Ex vivo wild-type and Bim−/− NK cells were isolated from donor spleens with DX5 MACS beads (Miltenyi Biotec), were differentially labeled with CFSE and then were injected into Il15−/− recipient mice. Recipient mice were analyzed at 24 and 72 h after injection and NK cell persistence was determined as described above.

Retroviral transduction of primary NK cells

Retroviral expression vectors pMIG-Noxa and pMIG-Noxa-3E have been described34,35. Retroviral constructs were transiently transfected into Phoenix Ecotropic packaging cells48 with Lipofectamine (Invitrogen) and viral supernatants were used to infect cells with a combination of RetroNectin (Takara) and spin inoculation. Twelve-well non–tissue culture plates were coated for 12 h at 4 °C with RetroNectin (4 µg/cm2 in PBS) before being blocked with PBS containing 2% (wt/vol) BSA. Viral supernatant (2 ml) was added and plates were then centrifuged for 15 min at 4 °C and 1,200g. Bone marrow cells were cultured for 48 h in IL-15 (50 ng/ml) and 2 × 106 cells per ml were added to each well after the viral supernatant was removed. Cells were then cultured for 6 h at 37 °C and 2 ml fresh viral supernatant was added together with IL-15 (50 ng/ml) and two rounds of ‘spin infection’ 18 h apart were done as described49. Plates were returned to 37 °C and infection efficiency was determined by flow cytometry of GFP expression.

MCMV infection

Mice were infected by intraperitoneal administration of 5 × 103 plaque-forming units of MCMV (K181 Perth strain) diluted in PBS containing 0.05% (vol/vol) FCS. Control mice received PBS containing 0.05% (vol/vol) FCS only. Organs were collected at various times and were processed for further analysis.

Quantitative PCR

Total RNA from 1 × 106 cells was purified and treated with DNAse I with RNeasy mini columns (Qiagen). SuperScript II reverse transcriptase (Invitrogen) was used for first-strand cDNA synthesis according to the manufacturer’s instructions. Samples without reverse transcriptase were used as controls for contaminating DNA. One tenth of the reverse-transcription reaction product was amplified by quantitative PCR with Quantitect SYBR Green PCR Master Mix (Qiagen) and the ABI PRISM system (Applied Biosystems). Primers used included β-actin (5′-AGTGTGACGTTGACATCCGTA-3′ and 5′-CCAGAGCAGTAATCTCCTTC-3′) and Noxa (5′-GCAGAGCTACCACCTGAGTTC-3′ and 5′-ATCTGCGCCAGAACCACAG-3′). A 100–base pair DNA template (Geneworks) spanning each primer set was purified by high-performance liquid chromatography and was used as a standard. The amount of Noxa in each sample was normalized to the amount of Actb present.

Statistical analysis

A standard Student’s t-test with two-tailed distributions for two-samples with equal variance was used for statistical analysis. P values of 0.05 or less were considered significant; exact P values are provided.

Supplementary Material

Supplemental Data


We thank C. Scott, L. Lee, Y. Hayakawa, J. Brady, C. Vandenberg and the support staff of the Walter and Eliza Hall Institute of Medical Research for technical assistance and reagents; P. Bouillet, J. Adams and S. Cory (Walter and Eliza Hall Institute) and B. Vanhaesenbroek (Babraham Institute) for gene-targeted mice; R. Anderson (The Peter MacCallum Cancer Centre) for mouse monoclonal anti-Hsp70; and Abbot Laboratories for ABT-737. Supported by the Cancer Council Victoria (N.D.H. and S.N.W.), the National Health and Medical Research Council of Australia (257502, 251608 and 356202), the National Cancer Institute (US; CA 80188 and CA 43540), the Leukemia and Lymphoma Society of America (SCOR 7015), and the Juvenile Diabetes Research Foundation–National Health and Medical Research Council and the Walter and Eliza Hall of Medical Research Metcalf Fellowship (S.L.N.).


Note: Supplementary information is available on the Nature Immunology website.


N.D.H. did experiments, analyzed data and wrote the manuscript; H.P., P.G., E.N., H.T., G.D. and S.N.W. did experiments; E.M.M. and N.M. provided unpublished genetically modified mice; M.A.D.-E., D.C.S.H. and M.J.S. designed research; and S.L.N., D.M.T. and A.S. designed research, analyzed data and wrote the manuscript.


The authors declare no competing financial interests.

Reprints and permissions information is available online at


1. Trinchieri G. Biology of natural killer cells. Adv. Immunol. 1989;47:187–376. [PubMed]
2. Biron CA, Nguyen KB, Pien GC, Cousens LP, Salazar-Mather TP. Natural killer cells in antiviral defense: function and regulation by innate cytokines. Annu. Rev. Immunol. 1999;17:189–220. [PubMed]
3. Lanier LL. NK cell recognition. Annu. Rev. Immunol. 2005;23:225–274. [PubMed]
4. Smyth MJ, Hayakawa Y, Takeda K, Yagita H. New aspects of natural-killer-cell surveillance and therapy of cancer. Nat. Rev. Cancer. 2002;2:850–861. [PubMed]
5. Kennedy MK, et al. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. J. Exp. Med. 2000;191:771–780. [PMC free article] [PubMed]
6. DiSanto JP, Muller W, Guy-Grand D, Fischer A, Rajewsky K. Lymphoid development in mice with a targeted deletion of the interleukin 2 receptor gamma chain. Proc. Natl. Acad. Sci. USA. 1995;92:377–381. [PubMed]
7. Suzuki H, Duncan GS, Takimoto H, Mak TW. Abnormal development of intestinal intraepithelial lymphocytes and peripheral natural killer cells in mice lacking the IL-2 receptor β chain. J. Exp. Med. 1997;185:499–505. [PMC free article] [PubMed]
8. Lodolce JP, et al. IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation. Immunity. 1998;9:669–676. [PubMed]
9. Nosaka T, et al. Defective lymphoid development in mice lacking Jak3. Science. 1995;270:800–802. [PubMed]
10. Park SY, et al. Developmental defects of lymphoid cells in Jak3 kinase-deficient mice. Immunity. 1995;3:771–782. [PubMed]
11. Cooper MA, et al. In vivo evidence for a dependence on interleukin 15 for survival of natural killer cells. Blood. 2002;100:3633–3638. [PubMed]
12. Prlic M, Blazar BR, Farrar MA, Jameson SC. In vivo survival and homeostatic proliferation of natural killer cells. J. Exp. Med. 2003;197:967–976. [PMC free article] [PubMed]
13. Huang DC, Strasser A. BH3-only proteins—essential initiators of apoptotic cell death. Cell. 2000;103:839–842. [PubMed]
14. Strasser A, O’Connor L, Dixit VM. Apoptosis signaling. Annu. Rev. Biochem. 2000;69:217–245. [PubMed]
15. Ranson T, et al. IL-15 is an essential mediator of peripheral NK-cell homeostasis. Blood. 2003;101:4887–4893. [PubMed]
16. Carson WE, et al. A potential role for interleukin-15 in the regulation of human natural killer cell survival. J. Clin. Invest. 1997;99:937–943. [PMC free article] [PubMed]
17. Bouillet P, et al. Proapoptotic Bcl-2 relative Bim required for certain apoptotic responses, leukocyte homeostasis, and to preclude autoimmunity. Science. 1999;286:1735–1738. [PubMed]
18. Akiyama T, et al. Regulation of osteoclast apoptosis by ubiquitylation of proapoptotic BH3-only Bcl-2 family member Bim. EMBO J. 2003;22:6653–6664. [PubMed]
19. Alfredsson J, Puthalakath H, Martin H, Strasser A, Nilsson G. Proapoptotic Bcl-2 family member Bim is involved in the control of mast cell survival and is induced together with Bcl-XL upon IgE-receptor activation. Cell Death Differ. 2005;12:136–144. [PubMed]
20. Strasser A. The role of BH3-only proteins in the immune system. Nat. Rev. Immunol. 2005;5:189–200. [PubMed]
21. O’Connor L, et al. Bim: a novel member of the Bcl-2 family that promotes apoptosis. EMBO J. 1998;17:384–395. [PubMed]
22. Puthalakath H, et al. ER stress triggers apoptosis by activating BH3-only protein Bim via de-phosphorylation and transcriptional induction. Cell. (in the press) [PubMed]
23. Hayakawa Y, Smyth MJ. CD27 dissects mature NK cells into two subsets with distinct responsiveness and migratory capacity. J. Immunol. 2006;176:1517–1524. [PubMed]
24. Robbins SH, Tessmer MS, Mikayama T, Brossay L. Expansion and contraction of the NK cell compartment in response to murine cytomegalovirus infection. J. Immunol. 2004;173:259–266. [PubMed]
25. Huntington ND, et al. NK cell maturation and peripheral homeostasis is associated with KLRG1 up-regulation. J. Immunol. 2007;178:4764–4770. [PubMed]
26. Andoniou CE, Andrews DM, Degli-Esposti MA. Natural killer cells in viral infection: more than just killers. Immunol. Rev. 2006;214:239–250. [PubMed]
27. Dokun AO, et al. Specific and nonspecific NK cell activation during virus infection. Nat. Immunol. 2001;2:951–956. [PubMed]
28. Okkenhaug K, et al. Impaired B and T cell antigen receptor signaling in p110δ PI 3-kinase mutant mice. Science. 2002;297:1031–1034. [PubMed]
29. Vlahos CJ, Matter WF, Hui KY, Brown RF. A specific inhibitor of phosphatidylinositol 3-kinase, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002) J. Biol. Chem. 1994;269:5241–5248. [PubMed]
30. Dijkers PF, Medema RH, Lammers JW, Koenderman L, Coffer PJ. Expression of the pro-apoptotic Bcl-2 family member Bim is regulated by the forkhead transcription factor FKHR-L1. Curr. Biol. 2000;10:1201–1204. [PubMed]
31. Brunet A, et al. Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor. Cell. 1999;96:857–868. [PubMed]
32. Opferman JT, et al. Development and maintenance of B and T lymphocytes requires antiapoptotic MCL-1. Nature. 2003;426:671–676. [PubMed]
33. Oltersdorf T, et al. An inhibitor of Bcl-2 family proteins induces regression of solid tumours. Nature. 2005;435:677–681. [PubMed]
34. Chen L, et al. Differential targeting of prosurvival Bcl-2 proteins by their BH3-only ligands allows complementary apoptotic function. Mol. Cell. 2005;17:393–403. [PubMed]
35. Willis SN, et al. Proapoptotic Bak is sequestered by Mcl-1 and Bcl-xL, but not Bcl-2, until displaced by BH3-only proteins. Genes Dev. 2005;19:1294–1305. [PubMed]
36. Di Santo JP. Natural killer cell developmental pathways: a question of balance. Annu. Rev. Immunol. 2006;24:257–286. [PubMed]
37. Dijkers PF, et al. Forkhead transcription factor FKHR-L1 modulates cytokine-dependent transcriptional regulation of p27KIP1. Mol. Cell. Biol. 2000;20:9138–9148. [PMC free article] [PubMed]
38. Stahl M, et al. The forkhead transcription factor FoxO regulates transcription of p27Kip1 and Bim in response to IL-2. J. Immunol. 2002;168:5024–5031. [PubMed]
39. Alves NL, et al. The Noxa/Mcl-1 axis regulates susceptibility to apoptosis under glucose limitation in dividing T cells. Immunity. 2006;24:703–716. [PubMed]
40. Ekert PG, et al. Cell death provoked by loss of interleukin-3 signaling is independent of Bad, Bim, and PI3 kinase, but depends in part on Puma. Blood. 2006;108:1461–1468. [PubMed]
41. You H, et al. FOXO3a-dependent regulation of Puma in response to cytokine/growth factor withdrawal. J. Exp. Med. 2006;203:1657–1663. [PMC free article] [PubMed]
42. Rinkenberger JL, Horning S, Klocke B, Roth K, Korsmeyer SJ. Mcl-1 deficiency results in peri-implantation embryonic lethality. Genes Dev. 2000;14:23–27. [PubMed]
43. Villunger A, et al. p53- and drug-induced apoptotic responses mediated by BH3-only proteins puma and noxa. Science. 2003;302:1036–1038. [PubMed]
44. Ranger AM, et al. Bad-deficient mice develop diffuse large B cell lymphoma. Proc. Natl. Acad. Sci. USA. 2003;100:9324–9329. [PubMed]
45. Kuroda J, et al. Bim and Bad mediate imatinib-induced killing of Bcr/Abl+ leukemic cells, and resistance due to their loss is overcome by a BH3 mimetic. Proc. Natl. Acad. Sci. USA. 2006;103:14907–14912. [PubMed]
46. Huntington ND, Xu Y, Nutt SL, Tarlinton DM. A requirement for CD45 distinguishes Ly49D-mediated cytokine and chemokine production from killing in primary natural killer cells. J. Exp. Med. 2005;201:1421–1433. [PMC free article] [PubMed]
47. Puthalakath H, Huang DC, O’lReilly LA, King SM, Strasser A. The proapoptotic activity of the Bcl-2 family member Bim is regulated by interaction with the dynein motor complex. Mol. Cell. 1999;3:287–296. [PubMed]
48. Kinsella TM, Nolan GP. Episomal vectors rapidly and stably produce high-titer recombinant retrovirus. Hum. Gene Ther. 1996;7:1405–1413. [PubMed]
49. Huntington ND, et al. CD45 links the B cell receptor with cell survival and is required for the persistence of germinal centers. Nat. Immunol. 2006;7:190–198. [PubMed]