Virology 490 (2016) 75–82

Contents lists available at ScienceDirect

Virology journal homepage: www.elsevier.com/locate/yviro

Brief Communication

Mcl-1 regulates effector and memory CD8 T-cell differentiation during acute viral infection Eui Ho Kim a, Brandon Neldner a, Jingang Gui b,1, Ruth W. Craig b, M. Suresh a,n a b

Department of Pathobiological Sciences, University of Wisconsin-Madison, 2015 Linden Drive, Madison, WI 53706, USA Department of Pharmacology and Toxicology, Dartmouth Medical School, Hanover, NH 03755, USA

art ic l e i nf o

a b s t r a c t

Article history: Received 9 December 2015 Returned to author for revisions 11 January 2016 Accepted 12 January 2016

Mcl-1, an anti-apoptotic member of Bcl-2 family maintains cell viability during clonal expansion of CD8 T cells, but the cell intrinsic role of Mcl-1 in contraction of effectors or the number of memory CD8 T cells is unknown. Mcl-1 levels decline during the contraction phase but rebound to high levels in memory CD8 T cells. Therefore, by overexpressing Mcl-1 in CD8 T cells we asked whether limiting levels of Mcl-1 promote contraction of effectors and constrain CD8 T-cell memory. Mcl-1 overexpression failed to affect CD8 T-cell expansion, contraction or the magnitude of CD8 T-cell memory. Strikingly, high Mcl-1 levels enhanced mTOR phosphorylation and augmented the differentiation of terminal effector cells and effector memory CD8 T cells to the detriment of poly-cytokine-producing central memory CD8 T cells. Taken together, these findings provided unexpected insights into the role of Mcl-1 in the differentiation of effector and memory CD8 T cells. & 2016 Elsevier Inc. All rights reserved.

Keywords: Myeloid cell leukemia 1 Lymphocytic choriomeningitis virus Effector Memory Central memory Effector memory Differentiation Apoptosis Mammalian target of rapamycin

Introduction Following vaccination or infections, antigen receptor signaling in concert with appropriate inflammatory and costimulatory signals trigger antigen-specific naïve CD8 T cells to expand and differentiate into effector cells in secondary lymphoid organs (Jameson and Masopust, 2009; Kaech and Cui, 2012; Zhang and Bevan, 2011). Subsequently, effector CD8 T cells traffic to the peripheral tissues and control the infection by MHC I-restricted cell-mediated cytotoxicity and/or by producing cytokines such as IFN-γ and TNF-α (Sprent and Surh, 2002; Zhang and Bevan, 2011). During the contraction phase,
90% of the effector CD8 T cells undergo apoptosis and the remainder of the expanded CD8 T cells differentiate into effector or central memory CD8 T cells. The self-renewing population of memory CD8 T cells provides rapid immunity to re-infection but such protective immunity depends upon the number, anatomical localization and functional attributes of memory CD8 T cells (Jameson and Masopust, 2009; Kaech and Cui, 2012; Zhang and Bevan, 2011). The number of memory CD8 T cells is a function of the extent of n

Corresponding author. E-mail address: [email protected] (M. Suresh). 1 Current Address: Laboratory of Immunology, Beijing Pediatric Research Institute, Beijing Children's Hospital and Capital Medical University, Beijing 100045, PR China. http://dx.doi.org/10.1016/j.virol.2016.01.008 0042-6822/& 2016 Elsevier Inc. All rights reserved.

expansion and contraction of effector CD8 T cells. Hence, there is high level of interest in determining the mechanisms that govern expansion and contraction of CD8 T cells during an immune response. The number of antigen-specific CD8 T cells at a given time during the T cell response is controlled by the relative rates of cellular proliferation and/or apoptosis. While it is clear that death receptormediated apoptosis can govern the death of activated T cells in vitro (Brunner et al., 1995; Dhein et al., 1995; Ju et al., 1995), cell-intrinsic pathway of mitochondria-mediated apoptosis is considered as the major mechanism of the death of effector T cells in vivo (Hildeman et al., 2002; Prlic and Bevan, 2008; Weant et al., 2008). The Bcl-2 family proteins are critical regulators of mitochondriamediated apoptosis, and consist of pro-apoptotic members (Bim, Bad, Puma, Noxa, Bid, Bax, Bak and Bok) and anti-apoptotic members (Bcl-2, Bcl-xL, Mcl-1 and A1) (Chipuk et al., 2010; Czabotar et al., 2014). These Bcl-2 family proteins are known to interact with each other with varying affinities (Chen et al., 2005), and the balance of the pro and anti-apoptotic proteins in the cell determines the cell fate. For example, Bim binds to all anti-apoptotic members with similar affinity, while Noxa binds to Mcl-1 and A1 selectively (Chen et al., 2005; Kurtulus et al., 2010). It is well established that the pro-apoptotic Bim is required for contraction of effector CD8 T cells and Bcl-2 is known as an antagonist to Bim (Prlic and Bevan, 2008; Wojciechowski et al., 2006, 2007). However, unlike Bim deficiency, overexpression of Bcl-2 fails to rescue effector CD8 T

76

E.H. Kim et al. / Virology 490 (2016) 75–82

cells from apoptosis in vivo (Petschner et al., 1998). These data suggested that high level of Bcl-2 alone is insufficient to overcome the pro-apoptotic effects of Bim in effector CD8 T cells during the contraction phase of the CD8 T-cell response. Unlike Bcl-2, which is highly expressed in naïve and memory T cells and rapidly downregulated after T cell activation, Mcl-1 expression is strongly induced after TCR stimulation (Dunkle et al., 2011; Opferman et al., 2003). Therefore, Mcl-1 is believed to maintain T cell viability during antigen-driven proliferation, especially when Bcl-2 levels are extremely low. Consistent with this idea, Hildeman's group recently reported that Mcl-1 deficiency impaired the accumulation of effector T cells following acute viral infection, a likely sequel to increased mitochondria-mediated apoptosis (Tripathi et al., 2013). Further, Gui et al. (2015) reported that transgenic expression of Mcl-1 in bone marrow-derived cells enhanced the number of memory CD8 T cells following infection of mice with vaccinia virus. Reports also indicate that Mcl-1 expression is induced by IL-7 and IL15 in T cells, and it might share the function of antagonizing Bim with Bcl-2 (Dunkle et al., 2011). However, the CD8 T cell-intrinsic role of Mcl-1 in regulating contraction and development of memory and whether Mcl-1 overexpression can override Bim-dependent apoptosis of effector CD8 T cells remain unknown. T cell activation leads to profound alterations in cellular energy metabolism characterized by increased glucose uptake and glycolysis to support the enhanced bioenergetics needs of the proliferating cell (Pearce et al., 2013; Pollizzi and Powell, 2014). Exquisite control of cellular energy metabolism is not only essential for optimal expansion and effector functions it is also linked to the differentiation of memory CD8 T cells (Finlay and Cantrell, 2011; Finlay, 2012; Pearce et al., 2013; Prlic and Bevan, 2009; van der Windt et al., 2012). In this context, Mcl1 is a unique member of the Bcl-2 family because, independent of its canonical function to enhance cell survival, it regulates mitochondrial physiology and cellular energy metabolism (Perciavalle and Opferman, 2013). Notably, loss of Mcl-1 leads to reduced oxidative phosphorylation and cellular ATP levels (Perciavalle et al., 2012) in cells such as fibroblasts and hepatocytes. However, it remains to be determined whether Mcl-1 regulates mitochondrial function, energy homeostasis and the differentiation of effector and memory CD8 T cells by cellintrinsic mechanisms. Mcl-1 deletion invariably results in apoptosis of naïve T cells (Opferman et al., 2003), which precludes experiments to assess the cell survival-independent roles of Mcl-1 in T cells. To overcome this constraint in studying the role of Mcl-1 in CD8 T-cell immunity, we have assessed the effects of Mcl-1 overexpression on the dynamics of the CD8 T-cell response in vivo (Zhou et al., 1998). This approach allowed us to delineate the cell-intrinsic role of Mcl-1 in regulating homeostasis and differentiation of effector and memory CD8 T cells. Surprisingly, we find that overexpression of Mcl-1 had a minimal impact on the expansion or the contraction of effector CD8 T cells and the magnitude of CD8 T-cell memory. However, elevated levels of Mcl-1 promoted the differentiation of short-lived effector cells (SLECs) at the expense of the memory precursor effector cells (MPECs) and skewed the development of effector memory CD8 T cells to the detriment of central memory CD8 T cells and their polyfunctionality. These results ascribe a previously unidentified role for Mcl-1 in regulating the differentiation of effector and memory CD8 T cells, which might have implications in vaccine development.

Results Mcl-1 expression in naïve, activated, effector and memory CD8 T cells First, we examined the expression of Mcl-1 in naïve (CD44LO) and activated/memory (CD44HI) CD8 T cells from uninfected B6 mice by flow cytometry. Interestingly, CD44HI CD8 T cells showed higher levels

of Mcl-1 protein, as compared to those in naïve CD8 T cells (Supplementary Fig. 1A). We also investigated whether activation of CD8 T cells induced the expression of Mcl-1. Consistent with a previous report (Dzhagalov et al., 2008), in vitro stimulation with anti-CD3 and anti-CD28 antibodies rapidly increased (3-fold) Mcl-1 protein expression within 24 h, as compared to un-stimulated CD8 T cells; Mcl-1 protein levels further increased in the next 24 h. (Supplementary Fig. 1B). Together, these data demonstrated that Mcl-1 is robustly expressed in naïve CD8 T cells, but their activation or differentiation into memory further enhanced Mcl-1 levels in CD8 T cells. During a CD8 T cell response, naïve CD8 T cells undergo clonal expansion and differentiation into effector cells. Subsequently,
90% of the effector cells are lost during the contraction phase. The net number of CD8 T cells during expansion or contraction are linked to the rate of cellular proliferation and/or apoptosis (Hildeman et al., 2007; Kaech and Cui, 2012). Since Bcl-2 family proteins are important regulators of this process (Hildeman et al., 2002; Kaech et al., 2003b; Wojciechowski et al., 2006, 2007), we assessed the expression dynamics of anti-apoptotic Bcl-2 family proteins Bcl-2 and Mcl-1 in virus-specific CD8 T cells during an acute LCMV infection. Consistent with previous reports (Hildeman et al., 2007, 2002), Bcl-2 levels in CD8 T cells plummeted during the activation and expansion phase, but levels rapidly rebounded during the contraction phase (Supplementary Fig. 1C). By contrast, Mcl-1 levels showed only a transient drop at day 5 post infection (PI), rose dramatically between days 5 and 8 PI, and then decreased during the early contraction phase (Supplementary Fig. 1C). During the late contraction and memory phase however, Mcl-1 expression rebounded to high levels. Thus, Bcl-2 and Mcl-1 may exert their anti-apoptotic functions in CD8 T cells at different time points during the T cell response in vivo. Effect of elevated Mcl-1 expression on clonal expansion of CD8 T cells Recently, Hildeman's group reported that Mcl-1 is required for the survival and accumulation of effector T cells during the expansion phase of the T cell response to LCMV (Tripathi et al., 2013). Further, it is worth noting that both Bcl-2 and Mcl-1 levels were substantially reduced at day 5 PI (Supplementary Fig. 1C). Therefore, it was of interest to determine whether lower Mcl-1 levels limit the clonal expansion and the magnitude of CD8 T cell memory during LCMV infection. To address this issue, we utilized mice that express human Mcl-1 transgene in the bone marrow compartment. Uninfected Mcl-1 transgenic (tg) mice did not exhibit significant alterations in the frequencies of B cells, CD4 T cells or CD8 T cells in spleen (Supplementary Fig. 2A). Additionally, the relative proportions of various T-cell subsets were largely unaltered in spleen of uninfected Mcl-1 tg mice (Supplementary Fig. 2B). Next, we determined whether forced expression of Mcl-1 in the bone marrow compartment would enhance clonal expansion and/or CD8 T cell memory during an acute LCMV infection. Data in Fig. 1A show that in the Mcl-1 tg mice, LCMV-specific CD8 T cells contained 1.5-2 fold more Mcl-1 protein, as compared to their non-transgenic counterparts during the course of LCMV infection. At day 8 PI, the frequencies of LCMVspecific Mcl-1 tg effector CD8 T cells were comparable to those in WT mice (Fig. 1B). By virtue of increased spleen size and the resulting augmentation of CD8 T cell numbers, the absolute numbers of LCMVspecific effector CD8 T cells were increased by
2 fold in Mcl-1 tg mice (Fig. 1C). Interestingly, there was a modest but statistically significant (po0.05) reduction in the apoptosis of Mcl-1 tg CD8 T cells at day 8 PI when the expression of Mcl-1 was highly induced (Fig. 1D). Previously, Wensveen et al. (2010) showed that Mcl-1/Noxa axis affected the TCR diversity of CD8 T cells by regulating the apoptosis of low avidity CD8 T cells. Therefore, we investigated whether forced expression of Mcl-1 affected the immunodominance hierarchy of LCMV-specific CD8 T cells. Mcl-1 transgene did not significantly affect the responses to either immunodominant (NP396, GP33 and GP276) or subdominant (NP205, GP118 and L2062) epitopes (Fig. 1E). Taken

E.H. Kim et al. / Virology 490 (2016) 75–82

77

Fig. 1. Effect of elevated Mcl-1 on expansion and differentiation of virus-specific effector CD8 T cells. WT and Mcl-1 tg mice were infected with LCMV. (A) Expression of hMcl-1 transgene was measured in NP396- and GP33-specific CD8 T cells over the course of infection. The median fluorescence intensities (MFIs) of hMcl-1 in Mcl-1 tg LCMV-specific CD8 T cells were normalized to those of WT group. (B, C) At day 8 PI, splenocytes from WT and Mcl-1 tg mice were stained with NP396- or GP33-specific class I MHC tetramers along with anti-CD8 for quantifying the frequencies and numbers of total CD8 T cells and LCMV-specific CD8 T cells. (D) Splenocytes were incubated in vitro for 4 h, and the levels of active caspase-3 in NP396-specific CD8 T cells were measured by flow cytometry. (E) WT and Mcl-1 tg splenocytes from day 8 PI were cultured for 5 h in the presence or absence of LCMV epitope peptides, followed by intracellular staining of IFN-γ. * indicates po0.05. Data are representative of two independent experiments with 3 mice per group.

Fig. 2. Effects of Mcl-1 overexpression on the contraction of LCMV-specific CD8 T cells. WT and Mcl-1 tg mice were infected with LCMV. After an acute LCMV infection, the percentages (A) and numbers (B) of virus-specific CD8 T cells were examined at day 8, 15 and 42 PI. (C) Fold contraction was calculated based on the reduction in the number of LCMV-specific CD8 T cells from day 8 to days 15 and 42. * indicates p o 0.05. Data are representative of two independent experiments with 3 mice per group/time point.

together, these data suggested that the elevated level of Mcl-1 in the bone marrow compartment increased the accumulation of virusspecific CD8 T cells during the expansion phase without affecting the immune-dominance hierarchy of the LCMV-specific CD8 T-cell response. Effect of elevated Mcl-1 protein on the contraction phase of the CD8 T cell response To investigate whether elevated Mcl-1 levels and reduced apoptosis (Fig. 1D) exerted beneficial effects on the survival of

effector CD8 T cells during the contraction phase, we tracked the frequencies and numbers of LCMV-specific CD8 T cells. Frequencies of both NP396- and GP33-specific CD8 T cells in WT and Mcl-1 tg mice were not significantly different during the contraction phase (Fig. 2A). However, the absolute numbers of LCMV-specific CD8 T cells were always slightly but reproducibly higher in the Mcl-1 tg mice (Fig. 2B). However, there was no significant difference in the contraction rate between WT and Mcl-1 tg CD8 T cells (Fig. 2C). These data suggested that overexpression of anti-apoptotic Mcl-1 in the bone marrow compartment did not augment the survival of effector CD8 T cells during the contraction phase.

78

E.H. Kim et al. / Virology 490 (2016) 75–82

Fig. 3. Cell-intrinsic effects of increased Mcl-1 levels on CD8 T cell responses to an acute viral infection. Ly5.1/P14 CD8 T cells from P14/WT or P14/Mcl-1 mice were purified and adoptively transferred into congenic Ly5.2/WT B6 mice, and 24 hours later the recipient mice were infected with LCMV. (A) Contour plots show percentages of donor P14 CD8 T cells (Ly5.1 þ ) in spleen at indicated time points. (B) Kinetic changes of P14 CD8 T cell numbers in P14/WT or P14/Mcl-1 group during the course of infection. (C, D) The levels of intracellular active caspase-3 (C) and nuclear Ki-67 (D) in P14 CD8 T cells were quantified by flow cytometry at indicated time points. * indicates p o 0.05. Data are representative of three independent experiments with 4 to 5 mice per group.

CD8 T cell-intrinsic role of Mcl-1 during an acute LCMV infection To determine the cell-intrinsic effects of elevated Mcl-1 in CD8 T cells, Mcl-1 tg mice were crossed with P14 tg mice to generate P14/ Mcl-1 tg (P14/Mcl-1) mice. Naive CD8 T cells from P14/WT or P14/Mcl1 mice were adoptively transferred into B6 mice, and those recipient mice were subsequently infected with LCMV. The frequencies of P14/ Mcl-1 CD8 T cells were not significantly different from their P14/WT counterparts at all times after LCMV infection (Fig. 3A). Further, the numbers of P14/Mcl-1 CD8 T cells were identical to their P14/WT counterparts (Fig. 3B). Notably, Mcl-1 overexpression only modestly reduced apoptosis and did not alter the proliferation of effector P14 CD8 T cells (Fig. 3C and D). These data suggested that elevated Mcl-1 expression did not alter the overall homeostasis of virus-specific CD8 T cells during expansion, contraction or memory maintenance. Cell-intrinsic effects of elevated Mcl-1 expression on the differentiation of effector and memory CD8 T cells To reiterate, Mcl-1 overexpression did not have detectable effects either on the frequencies or the total number of P14 CD8 T cells at all phases of the CD8 T-cell response (Fig. 3). Since energy homeostasis regulates CD8 T-cell differentiation and Mcl-1 is known to regulate mitochondrial energy production (Perciavalle et al., 2012), we questioned whether increased Mcl-1 expression affected the differentiation of effector and memory CD8 T cells. The heterogeneous population of effector CD8 T cells can be divided into at least two subsets: the short lived effector cells (SLECs; KLRG1 þ CD127  ) and memory precursor effector cells (MPECs; KLRG1  CD127 þ ) (Rutishauser et al., 2009). The percentages of SLECs among P14/Mcl-1 effector cells were substantially greater, as compared to those among P14/WT effector CD8 T cells (Fig. 4A and B). The increase in the representation of SLECs among P14/Mcl-1 effectors occurred at the expense of MPECs. This point is illustrated in Fig. 4A; in comparison to P14/WT CD8 T cells, the MPEC: SLEC ratios for the P14/Mcl-1 CD8 T cells were reduced by 42-fold at the outset on day 8 PI and similar or higher magnitude of augmentation was maintained for at least day until day 36 PI (Fig. 4A). Thus, overexpression of Mcl-1 promoted the differentiation of SLEC CD8 T cells and limited the development of MPEC

CD8 T cells. Increase in the ratio of the transcription factors T-bet and Eomes can enhance the representation of SLECs (Kaech and Cui, 2012). However, we find that Mcl-1 transgene did not significantly affect the Tbet:Eomes ratio in LCMV-specific effector CD8 T cells at day 8 PI (Supplementary Fig. 3). Next, we investigated the effect of Mcl-1 transgene on the differentiation of central (TCM) and effector (TEM) memory CD8 T cells. As shown in Fig. 4C, the percentage of P14/WT TCM CD8 T cells gradually increased after day 8 PI. In comparison to the P14/WT CD8 T cells, the emergence of TCM CD8 T cells was substantially blunted by overexpression of Mcl-1. Thus, high levels of Mcl-1 promoted the differentiation of TEM CD8 T cells during an acute LCMV infection. Since Mcl-1 overexpression regulated the differentiation of central and effector memory CD8 T cells, it was of interest to examine whether it also affected the functional attributes of memory CD8 T cells. This is because multi-cytokine-producing ability (ability to produce IFN-γ, TNF-α and IL-2) is typically associated with central memory CD8 T cells (Kaech and Cui, 2012; Wherry et al., 2003). As shown in Fig. 4D, most of the P14/WT and P14/Mcl-1 memory CD8 T cells produced IFN-γ upon ex vivo antigenic stimulation at all times after LCMV infection. At day 8 PI, comparable percentages of P14/WT and P14/Mcl-1 CD8 T cells produced IFN-γ and TNF-α or IL-2 (Fig. 4E and F). As effector CD8 T cells differentiated into memory, there was a gradual increase in the fraction of P14/WT cells that produced IFN-γ and TNF-α or IL-2. By contrast, the percentage of multiple cytokine-producing CD8 T cells was reduced in P14/Mcl-1 CD8 T cells. Consequently, at day 36 PI, a considerably lower percentage of P14/Mcl-1 CD8 T cells produced IFN-γ and TNF-α or IL-2 as compared to P14/WT CD8 T cells (Fig. 4E and F). These data indicated that high levels of Mcl-1 inhibit the development of multi-cytokine-producing memory CD8 T cells. Elevated Mcl-1 expression enhances mTOR activation in effector CD8 T cells As mentioned before, apart from preventing apoptosis, Mcl-1 regulates the physiology of mitochondria by promoting optimal oxidative phosphorylation and inhibition of superoxide production in cells (Perciavalle et al., 2012). Since mitochondrial respiratory

E.H. Kim et al. / Virology 490 (2016) 75–82

79

Fig. 4. Cell-intrinsic regulation of CD8 T-cell differentiation by elevated levels of Mcl-1 during an acute LCMV infection. Purified naive Ly5.1 þ P14/WT or P14/Mcl-1 CD8 T cells were adoptively transferred into congenic Ly5.2/B6 mice. Subsequently, mice were infected with LCMV and the responses of donor P14 CD8 T cells was assessed at different days postinfection. (A) The differentiation of effector P14 CD8 T cells was analyzed by co-staining for KLRG1 and CD127 at day 8, 15 and 36 PI. Numbers represent percentages of each subset. (B) Total numbers of KLRG1 þ CD127  and KLRG1  CD127 þ population were plotted during an acute LCMV infection. (C) Kinetic changes of central memory (CD62LHI; TCM) and effector memory (CD62LLO; TEM) frequencies among P14 CD8 T cells. (D–F) Cytokine production by P14 CD8 T cells was examined by intracellular cytokine staining at day 8, 15 and 36 PI after an acute LCMV infection. Plots display frequency of IFN-γ (D), IFN-γ and TNF (E) and IFN-γ and IL-2 (F) producers among P14 CD8 T cells. (G) Levels of dihydroethidium in KLRG1  or KLRG1 þ P14 CD8 T cells were assessed by flow cytometry at day 8 PI. Data are representative of two or three independent experiments with 4 to 5 mice per group. (H) Phospho-mTOR levels in P14/WT and P14/Mcl-1 CD8 T cells were quantified by flow cytometry at day 8 PI; data are the median fluorescence intensities. (I) Ly5.1/P14/Mcl-1 transgenic CD8 T cells were transferred into congenic Ly5.2/B6 mice and infected with LCMV. These mice were treated daily with PBS or rapamycin (Rapa) beginning day 7 PI. On day 22 PI, SLECs, MPECs and TCM in the blood were quantified by flow cytometry (n¼5). * indicates po0.05.

capacity might regulate memory versus effector differentiation of CD8 T cells (van der Windt et al., 2012), we assessed whether regulation of CD8 T-cell differentiation by Mcl-1 included alterations in mitochondrial membrane potential and cellular superoxide levels by staining with DiOC6 and dihydroethidium (DHE) respectively. Data in Fig. 4G show that Mcl-1 overexpression led to a slight reduction of the cellular superoxide levels, especially in the KLRG1 þ subset of effector cells but no detectable alteration in the mitochondrial membrane potential (data not shown). Thus Mcl-1 overexpression minimally affected cellular superoxide levels or the mitochondrial potential in LCMV-specific effector CD8 T cells. It is well established that mammalian target of rapamycin (mTOR) activation promotes the differentiation of SLECs and effector memory CD8 T cells (Araki et al., 2009; Rao et al., 2010). We investigated whether Mcl-1 transgene-driven differentiation of SLECs and effector memory CD8 T cells was associated with enhanced activation of mTOR. Naïve P14/WT and P14/Mcl-1 CD8 T cells were adoptively transferred into congenic B6 mice, which were subsequently infected with LCMV. At day 8 after LCMV infection, phosphorylation of mTOR was readily detected in P14/WT effector CD8 T cells directly ex vivo. Strikingly, phosphorylation of mTOR in P14/Mcl-1 effector CD8 T cells was significantly higher than in P14/WT CD8 T cells (Fig. 4H). Thus, Mcl-1 overexpression leads to enhanced activation of mTOR in antigen-specific effector CD8 T cells in vivo. Next, we determined whether mTOR inhibition by rapamycin treatment can reverse the effects of Mcl-1 overexpression on the differentiation of SLECs and TEM. Naïve P14/Mcl-1

CD8 T cells were adoptively transferred into congenic mice, which were subsequently infected with LCMV. Cohorts of LCMV-infected mice were treated daily with PBS or Rapamycin beginning day 7 PI. At day 22 PI, we compared the percentages of SLECs, MPECs and TCM in the peripheral blood of PBS and Rapamycin-treated groups (Fig. 4I). Data in Fig. 4I illustrate that the percentages of MPECs and TCM among P14 CD8 T cells in the rapamycin group were significantly greater than in the PBS group. Thus, mTOR inhibition can overcome some of the effects of Mcl-1 overexpression on differentiation of effector and memory CD8 T cells. Based on data presented in Fig. 4H–I, we propose that Mcl-1 promotes differentiation of SLECs and TEM at least in part by enhancing mTOR activation.

Discussion CD8 T cells play a crucial role in defense against viral, intracellular bacterial and protozoon infections. Vaccine-induced protective immunity to these infections depends upon the quantity and quality of memory CD8 T cells (Jameson and Masopust, 2009; Kaech and Cui, 2012; Zhang and Bevan, 2011). Deciphering the molecular and cellular mechanisms that regulate protective memory CD8 T cells is an area of active investigation. The apoptotic rate is one mechanism, which limits the accumulation of effector cells during the expansion phase and the number of memory CD8 T cells that survive the contraction phase. Therefore, modulating the apoptotic rate following vaccination can augment

80

E.H. Kim et al. / Virology 490 (2016) 75–82

the number of memory CD8 T cells (Hand and Kaech, 2009; Schluns and Lefrancois, 2003). Bcl-2 family proteins play seminal roles in regulating mitochondria-mediated apoptosis in different types of mammalian cells (Chipuk et al., 2010; Czabotar et al., 2014). More specifically, the balance between pro-apoptotic Bcl-2 members and anti-apoptotic members dictates cellular fate. In this context, mutually antagonistic proteins Bim and Bcl-2 have been investigated extensively for their roles in regulating the homeostasis of T cells (Wojciechowski et al., 2006, 2007). Unlike most members of the Bcl-2 family, Mcl-1 not only promotes cellular viability, it is implicated in the maintenance of mitochondrial structure, function and energy homeostasis (Perciavalle et al., 2012). Since energy homeostasis regulates differentiation of effector and memory CD8 T cells (Prlic and Bevan, 2009; van der Windt et al., 2012), Mcl-1 can potentially control the number and differentiation of memory CD8 T cells. However, the cell-intrinsic role of Mcl-1 in the differentiation of effector CD8 T cells or the generation of CD8 T-cell memory remains unknown. The main point of this manuscript is that Mcl-1 levels regulate the differentiation of effector and memory CD8 T cells. We find that the Mcl-1 protein levels undergo dynamic alterations during a T cell response in vivo. Mcl-1 levels are rapidly elevated during the expansion phase, but expression is downregulated during the contraction phase of the CD8 T-cell response. Interestingly, following the contraction phase during effector-memory transition, Mcl-1 expression raises again to new heights. On the other hand Bcl-2 expression pattern is reciprocal to that of Mcl-1, at least in the expansion phase. How are Bcl-2 and Mcl-1 levels regulated in a dynamic fashion? Antigen receptor signaling during clonal expansion likely drives down Bcl-2 levels and hyper-induce Mcl-1 in responding CD8 T cells (Dunkle et al., 2011; Opferman et al., 2003). Bcl-2 has been shown to be anti-proliferative and therefore it will be imperative for cells to overcome this effect to promote antigen-driven proliferation of CD8 T cells. At the same time, low Bcl-2 levels might render proliferating T cells susceptible to apoptosis (Grayson et al., 2000; Kaech et al., 2003a) but it is believed that concurrently augmented expression of Mcl-1 maintains cellular viability during antigen-driven expansion. This idea is consistent with a recent report that loss of Mcl-1 abrogates the accumulation of effector CD8 T cells during the T cell response to LCMV (Tripathi et al., 2013). Despite elevated Mcl-1 levels, CD8 T cell proliferation during the expansion phase is associated with significant apoptosis (Sullivan et al., 2012). This begged the question whether Mcl-1 levels constitute a limiting factor for the accumulation of effector CD8 T cells during the expansion phase. Data presented in this manuscript suggest that this is not likely the case because overexpression of Mcl-1 in P14 CD8 T cells failed to augment the number of effector CD8 T cells at the peak of the T cell response to LCMV. However, overexpression of Mcl-1 in all bone marrow-derived cells increased the number of vaccinia virus-specific MPECs (Gui et al., 2015) and LCMV-specific memory CD8 T cells (Fig. 1). Since transgenic expression of Mcl-1 in the bone marrow compartment doubles the splenic cellularity (Zhou et al., 1998), we suspect that the increase in the number of MPECs and memory CD8 T cells in these mice are linked to increase in the precursor frequencies of antigen-specific T cells in the naïve T cell repertoire. This theory is consistent with the report that increased precursor frequency enhances the differentiation of MPECs and memory CD8 T cells (Badovinac et al., 2007; Wirth and Harty, 2009). It is interesting that the levels of both Bcl-2 and Mcl-1 are transiently reduced at day 5 PI (Supplementary Fig. 1) during clonal expansion. How is T-cell viability maintained during this transient dip in Bcl-2/ Mcl-1 levels? T cell vaiability can be maintained by several plausible mechanisms that are not mutually exclusive. One possibility is that the reduced levels of Mcl-1 are sufficient to maintain T-cell viability. Alternatively, the high proliferation rate might exceed the apoptotic rate during this phase, leading to accumulation of activated CD8 T

cells. Thirdly, it is possible that anti-apoptotic molecules such as BclxL (Song et al., 2005) might promote T cell survival transiently when Bcl-2 and Mcl-1 levels are low. Loss of antigen receptor signaling and gain of responsiveness to cytokines such as IL-7 and/or IL-15 might induce Bcl-2 and Mcl-1 expression in CD8 T cells (Kaech et al., 2003a), following the expansion phase of the T cell response. The massive contraction of effector CD8 T cells that ensues the peak of clonal expansion has been attributed to Bim-dependent apoptosis (Wojciechowski et al., 2006, 2007). However, overexpression of the Bim-antagonist Bcl-2 fails to mitigate the loss of effector CD8 T cells and Bcl-xL does not appear to play a role in this process (Petschner et al., 1998). These reports led to the hypothesis that overexpression of Mcl-1 can override Bim-dependent apoptosis of effector cells. Surprisingly, data presented in this manuscript suggest that this hypothesis needs reexamination or that the elevated Mcl-1 levels in Mcl-1 tg CD8 T cells are insufficient to mitigate Bim-driven apoptotic program of effector CD8 T cells. The most surprising finding in this manuscript is that Mcl-1 regulates the differentiation of effector and memory CD8 T cells. We find that high levels of Mcl-1 enhanced the number of SLECs at the expense of the MPECs. The increased representation of SLECs among P14/Mcl-1 effectors could be due to increased proliferation and/or reduced apoptosis and/or augmented differentiation of this subset. The increased proportion of SLECs among P14/Mcl-1 effector cells is less likely to be consequential to enhanced survival or proliferation because the increase in the numbers of SLECs was clearly associated with a corresponding decrease in the number of MPECs. If the increase in the number of SLECs is primarily driven by selective enhancement in proliferation and/or reduced apoptosis of this subset, the total number of MPECs should have remained unaltered. Although we cannot exclude the contribution of altered proliferation/apoptosis to the augmentation of SLEC:MPEC ratio for Mcl-1 tg effectors, we favor the inference that high levels of Mcl-1 promotes differentiation of SLECs at the expense of MPECs. Increased representation of SLECs among P14/Mcl-1 effectors suggests that high levels of Mcl-1 might have accentuated the process of terminal differentiation. As noted above, alteration in cellular energy metabolism can skew the differentiation of effector cells. The truncated form of Mcl-1 is known to traffic to the mitochondrial matrix, augment oxidative phosphorylation and elevate the cellular ATP levels (Perciavalle et al., 2012). We hypothesize that high Mcl-1 levels induce dysregulated mitochondrial metabolism, excessive oxidative phosphorylation and high cellular ATP levels, which drive terminal differentiation of effector cells. Additionally, higher phosphorylation of mTOR in P14/Mcl-1 P14 CD8 T cells might have contributed to the excessive differentiation of SLECs and TEM. Notably, this effect is potentially reversible because mTOR inhibition by Rapamycin treatment restored at least in part the differentiation of MPECs and TCM from P14/Mcl-1 effector CD8 T cells. Nevertheless, the finding that Mcl-1 promotes phosphorylation of mTOR is extremely intriguing and surprising, but the underlying mechanism(s) is obscure. The increased representation of effector memory cells among P14/Mcl-1 memory CD8 T cells and the loss of IL-2-producing ability also support the idea that high Mcl-1 levels drive CD8 T cells towards terminal differentiation. Notably, deficiency for Noxa, an antagonist of Mcl-1 leads to increased numbers of effector memory phenotype T cells (Wensveen et al., 2010). It is possible that in the absence of Noxa, elevated Mcl-1 activity might drive the differentiation of effector memory CD8 T cells. Thus the balance between Mcl-1 and Noxa activities might regulate the development of TCM and TEM. Conclusions Data presented in this manuscript show that overexpression of Mcl-1 minimally affected the apoptosis or proliferation of CD8 T

E.H. Kim et al. / Virology 490 (2016) 75–82

cells during a physiological T cell response to an acute viral infection. Unexpectedly, we find that high levels of Mcl-1 drive terminal differentiation of effector cells into SLECs and promote the development of effector memory CD8 T cells at the expense of central memory CD8 T cells. Thus, modulation of Mcl-1 levels during vaccinations might be a strategy to enhance CD8 memory T cell-dependent protective immunity.

81

In vitro CD8 T cell activation Splenic CD8 T cells were enriched by CD8 þ T Cell Purification Kit (Miltenyi Biotec). Subsequently, CD8 T cells were cultured at a density of 2  105 cells per well in flat-bottom 96-well plate pre-coated with anti-CD3 (10 μg/ml) and anti-CD28 (2 μg/ml). Cells were cultured for 48 h with IL-2 (100 U/ml). Expression of Mcl-1 and Bcl-2 and the level of apoptosis were assessed by flow cytometry.

Materials and methods

Statistical analysis

Mice, viral infection and rapamycin treatment

Control and experimental groups were statistically compared by Student t test using StatPlus (AnalystSoft), and significance was defined at po0.05. Error bars represent Standard error of the mean (SEM).

C57BL/6 (B6) mice were purchased from the National Cancer Institute. The generation of Mcl-1 tg mice that express human Mcl-1 transgene has been described elsewhere (Zhou et al., 1998). To generate P14/Mcl-1 double tg mice, P14 mice, expressing the transgenic TCR specific for the Db-restricted GP33 epitope of LCMV, were crossed with Mcl-1 tg mice. Mice were infected with LCMV Armstrong strain (2  105 PFU, intra-peritoneal injection) to induce an acute infection. LCMV-infected mice were treated daily with rapamycin between day 7 and day 22 PI, as before (Kim et al., 2012). Experiments were conducted in accordance with protocols approved by the institutional animal care committee. Adoptive transfer of P14 CD8 T cells Splenic P14 (Ly5.1 þ ) CD8 T cells from P14/WT or P14/Mcl-1 mice were purified using a CD8 þ T-Cell Purification Kit (Miltenyi Biotec). Approximately 103 P14 cells from P14/WT or P14/Mcl-1 mice were adoptively transferred into congenic Ly5.2 þ B6 mice, and the recipient mice were subsequently infected with LCMV. Flow cytometry Mononuclear cells from spleens were harvested using standard techniques. For staining surface molecules, 106 cells were incubated with fluorochrome-labeled antibodies such as anti-CD8, anti-CD44, anti-CD62L, anti-KLRG1 and anti-CD127. To detect virus-specific CD8 T cells, MHC class I tetramers specific for LCMV epitopes NP396 and GP33 were used. In the P14 adoptive transfer experiment, antiLy5.1 antibody was used along with anti-CD8 antibody to track donor P14 CD8 T cells. Mitochondrial membrane potential and superoxide production were measured by DiOC6 (40 nM) and dihydroethidium (DHE), respectively, using the surface-staining protocol. For intracellular cytokine staining, splenocytes were restimulated ex vivo with LCMV-epitope peptides in the presence of Golgi inhibitor (Brefeldin A) and IL-2 for 5 h. After surface staining, cells were fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences), then stained with anti-IFN-γ, TNF-α and IL-2. For measurement of intracellular Bcl-2-family proteins and granzyme B, splenocytes were stained with anti-Bcl-2 and anti-granzyme B antibodies following the surface staining and permeabilization steps. Anti-Mcl-1 antibody (sc-819, Santa Cruz) and corresponding blocking peptides (sc-819P, Santa Cruz) were used to measure Mcl1-specific binding. Staining with anti-Mcl-1 antibodies preincubated with the blocking peptide was used as a specific negative control for flow cytometry. Active caspase-3 was assessed after 4 h culture in RPMI medium containing 10% FBS. Nuclear proteins such as transcription factors and Ki-67 were stained using the Foxp3 Staining Kit (eBioscience). Phospho-staining protocol was used to quantify mTOR phosphorylation in P14 CD8 T cells, as before (Kim et al., 2012). All samples were acquired with FACSCalibur, LSR II or LSR Fortessa flow cytometer (BD Bioscience). Data were analyzed with FlowJo software (TreeStar, Ashland, OR).

Acknowledgments This work was supported by PHS grants from the National Institutes of Health (AI48785 and AI101976) to Dr. M. Suresh. We thank Dr. David Gasper for editing the manuscript.

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.virol.2016.01.008.

References Araki, K., Turner, A.P., Shaffer, V.O., Gangappa, S., Keller, S.A., Bachmann, M.F., Larsen, C.P., Ahmed, R., 2009. mTOR regulates memory CD8 T-cell differentiation. Nature 460, 108–112. Badovinac, V.P., Haring, J.S., Harty, J.T., 2007. Initial T cell receptor transgenic cell precursor frequency dictates critical aspects of the CD8( þ ) T cell response to infection. Immunity 26, 827–841. Brunner, T., Mogil, R.J., LaFace, D., Yoo, N.J., Mahboubi, A., Echeverri, F., Martin, S.J., Force, W.R., Lynch, D.H., Ware, C.F., et al., 1995. Cell-autonomous Fas (CD95)/ Fas-ligand interaction mediates activation-induced apoptosis in T-cell hybridomas. Nature 373, 441–444. Chen, L., Willis, S.N., Wei, A., Smith, B.J., Fletcher, J.I., Hinds, M.G., Colman, P.M., Day, C.L., Adams, J.M., Huang, D.C., 2005. Differential targeting of prosurvival Bcl-2 proteins by their BH3-only ligands allows complementary apoptotic function. Mol. Cell 17, 393–403. Chipuk, J.E., Moldoveanu, T., Llambi, F., Parsons, M.J., Green, D.R., 2010. The BCL-2 family reunion. Mol. Cell 37, 299–310. Czabotar, P.E., Lessene, G., Strasser, A., Adams, J.M., 2014. Control of apoptosis by the BCL-2 protein family: implications for physiology and therapy. Nat. Rev. Mol. Cell Biol. 15, 49–63. Dhein, J., Walczak, H., Baumler, C., Debatin, K.M., Krammer, P.H., 1995. Autocrine Tcell suicide mediated by APO-1/(Fas/CD95). Nature 373, 438–441. Dunkle, A., Dzhagalov, I., He, Y.W., 2011. Cytokine-dependent and cytokineindependent roles for Mcl-1: genetic evidence for multiple mechanisms by which Mcl-1 promotes survival in primary T lymphocytes. Cell Death Dis. 2, e214. Dzhagalov, I., Dunkle, A., He, Y.W., 2008. The anti-apoptotic Bcl-2 family member Mcl-1 promotes T lymphocyte survival at multiple stages. J. Immunol. 181, 521–528. Finlay, D., Cantrell, D.A., 2011. Metabolism, migration and memory in cytotoxic T cells. Nat. Rev. Immunol. 11, 109–117. Finlay, D.K., 2012. Regulation of glucose metabolism in T cells; new insight into the role of phosphoinositide 3-kinases. Front. Immunol. 3. Grayson, J.M., Zajac, A.J., Altman, J.D., Ahmed, R., 2000. Cutting edge: increased expression of Bcl-2 in antigen-specific memory CD8 þ T cells. J. Immunol. 164, 3950–3954. Gui, J., Hu, Z., Tsai, C.Y., Ma, T., Song, Y., Morales, A., Huang, L.H., Dmitrovsky, E., Craig, R.W., Usherwood, E.J., 2015. MCL1 enhances the survival of CD8þ memory T Cells after viral infection. J. Virol. 89, 2405–2414. Hand, T.W., Kaech, S.M., 2009. Intrinsic and extrinsic control of effector T cell survival and memory T cell development. Immunol. Res. 45, 46–61. Hildeman, D., Jorgensen, T., Kappler, J., Marrack, P., 2007. Apoptosis and the homeostatic control of immune responses. Curr. Opin. Immunol. 19, 516–521. Hildeman, D.A., Zhu, Y., Mitchell, T.C., Bouillet, P., Strasser, A., Kappler, J., Marrack, P., 2002. Activated T cell death in vivo mediated by proapoptotic Bcl-2 family member Bim. Immunity 16, 759–767.

82

E.H. Kim et al. / Virology 490 (2016) 75–82

Jameson, S.C., Masopust, D., 2009. Diversity in T cell memory: an embarrassment of riches. Immunity 31, 859–871. Ju, S.T., Panka, D.J., Cui, H., Ettinger, R., El-Khatib, M., Sherr, D.H., Stanger, B.Z., Marshak-Rothstein, A., 1995. Fas(CD95)/FasL interactions required for programmed cell death after T-cell activation. Nature 373, 444–448. Kaech, S.M., Cui, W., 2012. Transcriptional control of effector and memory CD8 þ T cell differentiation. Nat. Rev. Immunol. 12, 749–761. Kaech, S.M., Tan, J.T., Wherry, E.J., Konieczny, B.T., Surh, C.D., Ahmed, R., 2003a. Selective expression of the interleukin 7 receptor identifies effector CD8 T cells that give rise to long-lived memory cells. Nat. Immunol. 4, 1191–1198. Kaech, S.M., Tan, J.T., Wherry, E.J., Konieczny, B.T., Surh, C.D., Ahmed, R., 2003b. Selective expression of the interleukin 7 receptor identifies effector CD8 T cells that give rise to long-lived memory cells. Nat. Immunol. 4, 1191–1198. Kim, E.H., Sullivan, J.A., Plisch, E.H., Tejera, M.M., Jatzek, A., Choi, K.Y., Suresh, M., 2012. Signal integration by Akt regulates CD8 T cell effector and memory differentiation. J. Immunol. 188, 4305–4314. Kurtulus, S., Tripathi, P., Opferman, J.T., Hildeman, D.A., 2010. Contracting the ‘mus cells’–does down-sizing suit us for diving into the memory pool? Immunol. Rev. 236, 54–67. Opferman, J.T., Letai, A., Beard, C., Sorcinelli, M.D., Ong, C.C., Korsmeyer, S.J., 2003. Development and maintenance of B and T lymphocytes requires antiapoptotic MCL-1. Nature 426, 671–676. Pearce, E.L., Poffenberger, M.C., Chang, C.H., Jones, R.G., 2013. Fueling immunity: insights into metabolism and lymphocyte function. Science 342, 1242454. Perciavalle, R.M., Opferman, J.T., 2013. Delving deeper: MCL-1's contributions to normal and cancer biology. Trends Cell Biol. 23, 22–29. Perciavalle, R.M., Stewart, D.P., Koss, B., Lynch, J., Milasta, S., Bathina, M., Temirov, J., Cleland, M.M., Pelletier, S., Schuetz, J.D., Youle, R.J., Green, D.R., Opferman, J.T., 2012. Anti-apoptotic MCL-1 localizes to the mitochondrial matrix and couples mitochondrial fusion to respiration. Nat. Cell Biol. 14, 575–583. Petschner, F., Zimmerman, C., Strasser, A., Grillot, D., Nunez, G., Pircher, H., 1998. Constitutive expression of Bcl-xL or Bcl-2 prevents peptide antigen-induced T cell deletion but does not influence T cell homeostasis after a viral infection. Eur. J. Immunol. 28, 560–569. Pollizzi, K.N., Powell, J.D., 2014. Integrating canonical and metabolic signalling programmes in the regulation of T cell responses. Nat. Rev. Immunol. 14, 435–446. Prlic, M., Bevan, M.J., 2008. Exploring regulatory mechanisms of CD8þ T cell contraction. Proc. Natl. Acad. Sci. USA 105, 16689–16694. Prlic, M., Bevan, M.J., 2009. Immunology: a metabolic switch to memory. Nature 460, 41–42. Rao, R.R., Li, Q., Odunsi, K., Shrikant, P.A., 2010. The mTOR kinase determines effector versus memory CD8þ T cell fate by regulating the expression of transcription factors T-bet and Eomesodermin. Immunity 32, 67–78.

Rutishauser, R.L., Martins, G.A., Kalachikov, S., Chandele, A., Parish, I.A., Meffre, E., Jacob, J., Calame, K., Kaech, S.M., 2009. Transcriptional repressor Blimp-1 promotes CD8( þ ) T cell terminal differentiation and represses the acquisition of central memory T cell properties. Immunity 31, 296–308. Schluns, K.S., Lefrancois, L., 2003. Cytokine control of memory T-cell development and survival. Nat. Rev. Immunol. 3, 269–279. Song, A., Tang, X., Harms, K.M., Croft, M., 2005. OX40 and Bcl-xL promote the persistence of CD8 T cells to recall tumor-associated antigen. J. Immunol. 175, 3534–3541. Sprent, J., Surh, C.D., 2002. T cell memory. Annu. Rev. Immunol. 20, 551–579. Sullivan, J.A., Kim, E.H., Plisch, E.H., Peng, S.L., Suresh, M., 2012. FOXO3 regulates CD8 T cell memory by T cell-intrinsic mechanisms. PLoS Pathog. 8, e1002533. Tripathi, P., Koss, B., Opferman, J.T., Hildeman, D.A., 2013. Mcl-1 antagonizes Bax/ Bak to promote effector CD4( þ ) and CD8( þ) T-cell responses. Cell Death Differ. 20, 998–1007. van der Windt, G.J., Everts, B., Chang, C.H., Curtis, J.D., Freitas, T.C., Amiel, E., Pearce, E.J., Pearce, E.L., 2012. Mitochondrial respiratory capacity is a critical regulator of CD8 þ T cell memory development. Immunity 36, 68–78. Weant, A.E., Michalek, R.D., Khan, I.U., Holbrook, B.C., Willingham, M.C., Grayson, J. M., 2008. Apoptosis regulators Bim and Fas function concurrently to control autoimmunity and CD8þ T cell contraction. Immunity 28, 218–230. Wensveen, F.M., van Gisbergen, K.P., Derks, I.A., Gerlach, C., Schumacher, T.N., van Lier, R.A., Eldering, E., 2010. Apoptosis threshold set by Noxa and Mcl-1 after T cell activation regulates competitive selection of high-affinity clones. Immunity 32, 754–765. Wherry, E.J., Teichgraber, V., Becker, T.C., Masopust, D., Kaech, S.M., Antia, R., von Andrian, U.H., Ahmed, R., 2003. Lineage relationship and protective immunity of memory CD8 T cell subsets. Nat. Immunol. 4, 225–234. Wirth, T., Harty, J.T., 2009. Initial TCR transgenic precursor frequency alters functional behaviour of CD8 T cells responding to acute infection. Adv. Exp. Med. Biol. 633, 71–80. Wojciechowski, S., Jordan, M.B., Zhu, Y., White, J., Zajac, A.J., Hildeman, D.A., 2006. Bim mediates apoptosis of CD127(lo) effector T cells and limits T cell memory. Eur. J. Immunol. 36, 1694–1706. Wojciechowski, S., Tripathi, P., Bourdeau, T., Acero, L., Grimes, H.L., Katz, J.D., Finkelman, F.D., Hildeman, D.A., 2007. Bim/Bcl-2 balance is critical for maintaining naive and memory T cell homeostasis. J. Exp. Med. 204, 1665–1675. Zhang, N., Bevan, M., 2011. CD8þ T Cells: foot soldiers of the immune system. Immunity 35, 161–168. Zhou, P., Qian, L., Bieszczad, C.K., Noelle, R., Binder, M., Levy, N.B., Craig, R.W., 1998. Mcl-1 in transgenic mice promotes survival in a spectrum of hematopoietic cell types and immortalization in the myeloid lineage. Blood 92, 3226–3239.