Multiple signaling pathways promote B lymphocyte stimulator–dependent B-cell growth and survival

The majority of naive lymphocytes are small cells in the G_o stage of the cell cycle, a condition providing economy of space and efficient resource utilization. Lymphocytes at rest are not static but are actively maintained by the action of growth and survival factors necessary for nutrient uptake and suppression of apoptosis.¹  Two unique survival factors for B cells have been identified: B lymphocyte stimulator (BLyS; BAFF, TALL-1, THANK, TNFSF13B, and zTNF4) and APRIL, both members of the tumor necrosis factor superfamily of ligands.²  APRIL depletion has little effect on the homeostasis of naive B cells but compromises the survival of plasma cells.^3,4  In contrast, BLyS is essential for the survival of peripheral B lymphocytes.⁵  BLyS-deficient mice have striking deficits in marginal zone and follicular B-cell populations,^6,7  whereas ectopic expression of BLyS from a transgene markedly expands follicular and marginal peripheral zone B cells without affecting T cells, early (T1) transitional peripheral B cells, or developing B cells in the marrow.⁸  BLyS is also required for the maintenance of a variety of B cell tumors and dysregulated BLyS stimulation rescues autoantibody producing B cells from deletion.^9,10  Thus, BLyS has a critical role in the homeostasis of both normal and pathogenic B cells

BLyS acts through 3 receptors: BCMA (B-cell maturation antigen), TACI (transmembrane activator and CAML interactor), and BR3 (BLyS receptor 3) or BAFF receptor.⁸  Targeted mutation of BCMA has no striking effect on naive B-cell survival, but BCMA is up-regulated on and required for the survival of long-lived plasma cells.⁴  Elimination of TACI causes a marked increase of follicular and marginal zone B cells and the production of autoantibodies suggesting a primary regulatory role for this receptor.¹¹  BR3 is found on mature follicular and marginal zone B cells and on T2, T3 transitional B cells.¹²  Mice with a spontaneous mutation in the BR3 signaling motif (A/WySn) or mice in which the BR3 gene has been targeted lose these B-cell subpopulations.^13,14  The phenotype of BR3-deficient mice for the most part recapitulates that of BLyS-deficient mice, suggesting that naive B-cell survival depends primarily on BLyS acting through BR3.

The molecular mechanism by which BLyS affects naive B-cell survival has not been completely elucidated. BLyS stimulation mobilizes tumor necrosis factor receptor associated factors (TRAFs) promoting NF-κB activation through the induction of both canonical (p50, NF-kB1) and noncanonical (p52, NF-kB2) NF-kB pathways.^15,–17  NF-kB induction up-regulates genes for a variety of antiapoptotic proteins, including A1/Bfl-1, Bcl-x_L, and Bcl-2 any of which could account for BLyS-dependent apoptosis inhibition.^17,–19  Alternatively, inhibition of proapoptotic Bcl-2 family members is proposed as a mechanism of protection.^20,21 

We report here that BLyS attenuates atrophy and death of primary naive B lymphocytes by activating 2 independent growth and survival pathways, Akt/mTOR and Pim 2. In the absence of Pim 2, BLyS-mediated B-cell survival becomes rapamycin-sensitive in vitro and in vivo. A key effector molecule necessary for protection is Mcl-1, an antiapoptotic Bcl-2 family member; BLyS-dependent survival is not induced in B cells where Mcl-1 was conditionally deleted. The data suggest that 2 independent signaling pathways mediate BLyS-dependent survival and that Mcl-1 is a critical downstream mediator of BLyS action.

Pim-1^+/+2^+/+, Pim-1^−/−2^+/+, Pim-1^+/+2^−/−, and Pim-1^−/−2^−/− mice were generated from Pim-1^+/−2^+/− stock generously provided by Paul Rothman, Columbia University, New York, NY. C57BL/6 (B6) mice were obtained from Jackson Laboratory (Bar Harbor, ME). Mx-CRE transgenics were backcrossed with mice with floxed Mcl-1 to produce Mcl-1^f/f, Mcl-1^f/null, and Mcl-1^f/wt populations.²²  CBA/Ca × B6 mice expressing the human Bcl-2 transgene under the control of an Ig enhancer and promoter have been described.^23,24  Animals were maintained at the University of Pennsylvania, Harvard Medical School, or the University of Massachusetts Medical School under Institutional Animal Care and Use Committee guidelines.

Murine B cells were prepared by antithy1.2 and complement treatment of splenocytes followed by purification of resting B cells using a step Percoll gradient harvesting cells at the 60% to 70% interface.²³  CD23⁺ B cells were obtained by magnetic separation of murine splenocytes using biotinylated anti-CD23 antibody (BD Biosciences-Pharmingen, San Diego, CA) and streptavidin-coated microbeads (Miltenyi Biotec, Auburn CA). Splenic B cells were more than 90% B220⁺, whereas CD23⁺ B cells were more than 95% pure.

Purified B cells or CD23⁺ B cells were cultured in RPMI 1640 supplemented with 2-mercaptoethanol, MEM-nonessential amino acids, glutamine, penicillin, and streptomycin (CM) at 37°C in 5% CO₂ and air. For B-cell survival, enzyme and protein induction assays recombinant human (rhu) BLyS (Human Genome Sciences, Rockville, MD) was used at 50 to 100 ng/mL, the optimum dose for survival established by BlyS titrations between 1 and 500 ng/mL. Murine BLyS was purchased from Alexis Biochemicals (San Diego, CA) and interferon alpha (IFNα) from PBL Biomedical (Piscataway, NJ). Rapamycin at a final concentration of 50 nM, a concentration established in preliminary experiments to inhibit mitogen responses in vitro without affecting viability, was added to cultures from a stock dissolved in methanol; control B cells were treated with methanol only. For enzyme phosphorylation assays B cells cultured overnight at 37°C were activated with soluble antimurine IgM (5 μg/mL; Jackson ImmunoResearch, West Grove, PA), anti-CD40 (0.5 μg/mL; eBioscience, San Diego, CA), or 100 ng/mL of rhuBLyS added to assay buffer (Hanks balanced salt solution, 2% bovine serum albumin). To assess PI 3-kinase dependence of BLyS-induced Akt phosphorylation, B cells were treated for 60 minutes with LY294002 or the negative control LY 303511 (EMD Bioscience, La Jolla, CA) at 10 μM in DMSO before BLyS addition. Long-term Mcl-1 expression was determined using B cells from huBcl-2 tg⁺ donors (3 × 10⁶ cells /well in 12-well plates) cultured for 10 days with and without 50 ng/mL of rhuBLyS.

Anti-Pim 2 (1D12), Pim 1 (19F7), and actin (1-19) anti–mouse, – rabbit, and – goat Ig coupled to HRP were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit antiphospho-Akt (S473) and phospho-Akt (T308), phospho-p70 S6 kinase (T389), phosph-Fox01/Fox03a (T24/32), phospho-Gsk-3α/β (S21/9), phospho-4E-BP1 (T37/46), PI3 kinase (p85), Gsk-3β, 4E-BP1, and Akt were purchased from Cell Signaling (Beverly, MA). Rabbit anti–mouse Mcl-1 was purchased from Rockland (Wilmington, MA) and anti-Bcl-2 and anti-Bcl-X_L from eBioscience. Whole cell lysates were prepared by washing B cells in ice-cold phosphate-buffered saline and lysing pellets in RIPA (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris, pH 8.0) supplemented with protease inhibitors (Complete Mini, Roche Diagnostics, Indianapolis IN) and phosphatase inhibitor cocktails I and II (Sigma, St Louis, MO); 10 to 50 μg of protein was resolved on 4% to 12% NuPage bis-tris polyacrylamide gels (Invitrogen, Carlsbad, CA) and transferred to nitrocellulose. Blots were blocked with 5% bovine serum albumin, 0.2% Tween-20 in Tris-buffered saline and incubated with primary antibody in the same buffer overnight at 4°C. Blots were washed, incubated with secondary antibody conjugated with HRP, and developed using ECLplus (Amersham Bioscience, Piscataway, NJ). Blots were stripped for reprobing by incubation for 20 minutes at 65°C in Tris-buffered saline supplemented with 1% SDS and 100 μM β-mercaptoethanol. Quantitation of Western signals was performed with a Molecular Dynamics Densitometer (Amersham, Sunnyvale, CA) using BioRad (San Diego, CA) Multi-analyst software.

B cells at 5 × 10⁶/mL were cultured in 24-well tissue culture dishes in CM at 37°C with test stimulators and supplements. B cells were pretreated with 50 nM rapamycin or vehicle 1 hour before culture with the test stimulants with fresh rapamycin added after 48 hours. Survival was monitored daily by counting viable cells using trypan blue exclusion with each determination done in triplicate. Apoptosis and cell survival were monitored, respectively, by FACS using PI staining²⁵  or a Live/Dead Fixable Blue Stain kit (Molecular Probes, Eugene, OR). Cell size was determined with a Coulter Z2 particle size analyzer after removing dead cells by Percoll flotation.

Glucose use was determined for B cells cultured for 2 days with rhuBLyS, anti-Ig, or anti-CD40 as described.²⁶  Briefly, after culture, cells were harvested, washed, and pulsed with 10 μCi (370 kBq) of 5-³H-glucose in 24-well plates for 1 to 2 hours at 37°C. The reaction was stopped by the addition of 0.2N HCL. ³H₂O was separated from 5-³H-glucose by diffusion into a reservoir of H₂O in an airtight scintillation vial. Diffused and undiffused ³H₂O was measured using a 1450 Microbeta scintillation counter (Wallac, Turkua, Finland) to determine the rate of glycolysis.

Data were compared using a paired t test or one-way ANOVA test with Tukey posttest, InStat version 3 statistical package (Graphpad Software, San Diego, CA). P values are provided or represented as P is more than .05 (not significant), P is less than .05 (*), P is less than .01 (**), and P is less than .001 (***).

B lymphocytes removed from their normal microenvironment undergo death by neglect that can be blocked by inducers of cell growth and antiapoptotic molecules.²⁷  Positively selected CD23⁺ follicular B cells cultured without stimulation die rapidly ex vivo, demonstrated by direct observation using trypan blue dye exclusion (Figure 1A) or Flow analysis using blue-fluorescent reactive vital dye (Figure 1B). These 2 methods for determining viability yielded statistically indistinguishable results (P > .05). Survival was markedly enhanced (P < .01 for days 3, 4, and 5) for B cells cultured with recombinant huBLyS compared with unstimulated controls (Figure 1A,B). Cell loss by apoptosis was significantly inhibited by BLyS stimulation (Figure 1C). Maintenance of viability is not attributable to contaminating endotoxin in the BLyS preparation because BLyS-dependent survival of TLR4^−/− and wild-type CD23⁺ B cells in culture was equivalent and TLR4^−/− B cells were not rescued by LPS (Figure S1A, available on the Blood website; see the Supplemental Materials link at the top of the online article).

T lymphocytes cultured without growth factors lose cell volume as a consequence of nutrient deprivation.²⁸  This cell atrophy subsequent to growth factor withdrawal is seen in B lymphocytes. Freshly isolated B cells, averaging 129 plus or minus 8 fL in volume shrank to 100 plus or minus 3 fL, for viable cells recovered after 5 days in culture, whereas B cells cultured with BLyS increased in volume to 144 plus or minus 8 fL over the same period (Figure 1D). Contaminating endotoxin could not explain BLyS-dependent atrophy resistance, which was comparable in TLR4^−/− and wild-type CD23⁺ B cells, although TLR4^−/− B cells failed to enlarge when stimulated with LPS (Figure S1B). Maintenance of cell size reflects the preservation of essential metabolic pathways; B cells cultured with BLyS have significantly increased glucose metabolism (4.3 ± 1.9 nM/10⁶ B cells/hr) compared with B cells cultured without BLyS (1.4 ± 0.4 nM/10⁶B cells/h; Figure 1E). Stimulation with anti-Ig or anti-CD40 induced blastogenesis and a markedly increased glucose use relative to BLyS-treated B cells, which did not undergo blastogenesis (Figure 1E). For example, B cells stimulated through CD40 exhibited a greater cell volume (198 ± 3 fL), after 3 days of culture than BLyS-stimulated cells (140 ± 4 fL). BLyS stimulation induces a generalized metabolic enhancement; RNA (³H-uridine uptake) and amino acid (³H-leucine uptake) metabolism are increased, but ³H-thymidine incorporation remains at background levels (data not shown), and progression to the S phase of the cell cycle did not occur (Figure 1C). These data suggest that BLyS functions as both a growth and survival factor because BLyS-stimulated B cells retain viability, size, and metabolic activity ex vivo, whereas unstimulated B cells do not.

Maintenance of growth and survival are known functions of the Akt/mTOR pathway. To determine whether BLyS activated Akt, we stimulated primary murine B cells with recombinant human or murine BLyS or with anti-Ig, a known inducer of this pathway. Phosphorylation of the 2 target sites necessary for full Akt activity, T308 and S473, was readily apparent in B cells stimulated with either soluble human or murine rBLyS (Figure 2A). BlyS-induced Akt phosphorylation peaked at 60 minutes and was still detectable at 120 minutes (Figure 2A), whereas anti-Ig induced phosphorylation was more rapid. Akt phosphorylation was dependent on PI3-kinase as it was inhibited by LY294002 but not by the inactive analog LY303511 (Figure 2B). The activation of mTOR, a major Akt downstream effector associated with cell growth, was demonstrated by the phosphorylation of its substrate the translation inhibitor 4E-BP1 (Figure 2C). We found incubation of B cells without activator sometimes led to phosphorylation of 4E-BP1, but this was always more transient and less intense than activation with BLyS. Akt phosphorylation on S473 was not seen in extracts from cells cultured for 4 hours in the absence of BLyS (Figure 2D). We also found p70 S6 kinase, Gsk-3_β, FOXO1, and FOXO3a were phosphorylated by BLyS stimulation with kinetics similar to the phosphorylation of Akt (Figure S2).

Akt phosphorylation at S473 has recently been attributed to the normally rapamycin-insensitive rictor containing mTORC2 complex.²⁹  This complex can be disassembled, however, in some mammalian cell lines after prolonged rapamycin treatment making Akt phosphorylation at S473 rapamycin sensitive.³⁰  To assess the action of mTORC2 on Akt activation in primary B cells, B cells were cultured for 24 hours with 50 nM rapamycin before BLyS (Figure 2E) or anti-CD40 (Figure 2F) stimulation. We found Akt phosphorylation at S473 was significantly impaired for both activators in rapamycin-treated B cells.

Rapamycin inhibits growth-factor dependent maintenance of cell survival and size.^28,31,–33  To assess the effect of rapamycin on BLyS-mediated survival and growth, total B or CD23⁺ selected B cells were cultured with rapamycin or vehicle. Treatments with rapamycin neither exacerbated significantly spontaneous B-cell loss nor impeded BLyS-dependent survival (Figure 3A). In contrast, BLyS-dependent maintenance of cell size was inhibited by rapamycin with significant differences found for days 3 through 5 of culture (Figure 3B). This rapamycin-induced effect was not complete; B cells stimulated with both BLyS and rapamycin were still significantly larger than untreated or rapamycin only treated controls.

The failure of rapamycin to suppress BLyS-mediated survival in normal B cells suggested a BLyS-induced signaling pathway independent of Akt/mTOR. Pims are a family of 3 serine/threonine kinases induced in hematopoietic cells by numerous activators that can provide rapamycin-resistant protection against apoptosis and atrophy.^34,35  To assess Pim induction by BLyS, we analyzed extracts from purified B cells left unstimulated or stimulated with BLyS, anti-CD40, or anti-IgM antibody for 24 or 48 hours. Figure 4 shows that both isoforms of Pim 1 are induced by BLyS (2- to 4-fold over unstimulated controls) with similar kinetics but at a lower intensity than with the other activators. Likewise, all isoforms of Pim 2 were induced by BLyS (2- to 6-fold over unstimulated controls), with similar kinetics but a lower intensity than the other activators.

Can Pims account partially or fully for BLyS-dependent survival? Multiple signal pathways promote apoptosis resistance in hematopoietic cell lines and T cells.^21,36  To assess the potential contribution of simultaneously active protective pathways using Pims and/or Akt/mTOR, we determined the survival of normal (wild-type or Pim1^+/−2^+/− heterozygotes) and Pim 1^−/− 2^−/− doubly deficient CD23⁺ B cells treated with BLyS, and/or rapamycin or vehicle. BLyS-dependent survival of wild-type B cells was fully resistant to rapamycin (Figure 5A). Pim 1 and 2 doubly deficient B cells were also fully protected by culture with BLyS, showing that Pims were not obligatory for BLyS-mediated protection. Protection was completely abolished, however, in Pim 1– and 2–deficient B cells by rapamycin treatment (Figure 5B), showing that Akt/mTOR and Pims fully account for BLyS-dependent protection. Pim 1 and 2 mediate growth factor–induced apoptosis protection in B and T cells, respectively.^21,37  To determine whether one or both Pims were necessary to protect B cells in which mTOR was inhibited, we assessed the effect of rapamycin on B cells from mice deficient only in Pim 2. BLyS failed to rescue rapamycin–treated Pim 1^+/−2^−/− B cells (Figure 5C), showing that Pim 1 does not confer rapamycin resistance. These data, in toto, suggest that the Akt/mTOR– and Pim 2–dependent pathways independently mediate BLyS–dependent survival.

Because rapamycin induced only a partial inhibition of BLyS-dependent B-cell growth and Pims mediate both growth and survival in hematopoietic cells,^34,38  we asked if rapamycin-dependent growth inhibition was more effective in Pim 1 and 2 doubly deficient B cells. BLyS-dependent maintenance of cell size was unimpaired in Pim-deficient B cells and the magnitude of rapamycin–induced inhibition of BLyS-dependent atrophy protection was comparable in Pim-deficient and normal B cells (Figure 3B,C).

Mcl-1 plays a critical role in peripheral B- and T- cell homeostasis. Excision of Mcl-1 causes the loss of mature T and B cells in situ or of adoptively transferred lymphocytes in vivo.²²  Given these results, combined with our finding that neither wild-type nor Pim 1^−/−2^−/−–deficient B cells stimulated with BLyS consistently up-regulated Bcl-2 or Bcl-X_L proteins (Figure S3), we used B cells in which Mcl-1 could be conditionally deleted to test the hypothesis that BLyS-mediated survival required Mcl-1.B cells bearing a CRE transgene under the control of the interferon responsive MX promoter, and heterozygous for floxed and wild-type Mcl-1 alleles (Mcl-1^f/wt) or floxed and null Mcl-1 alleles (Mcl-1^f/null) were examined for in vitro viability when cultured with and without BLyS in the presence or absence of recombinant IFNα. BLyS-treated B cells from Mcl-1^f/wt mice were fully protected from death in culture and this BLyS-dependent protection was unaffected by IFNα in the cultures (P > .05; Figure 6A). In contrast, although B cells from Mcl-1^f/null mice were also protected from death by BLyS, protection was completely lost when Mcl-1 was excised on addition of IFNα to the cultures (P < .001; Figure 6A).

We also found that Mcl-1^f/wt B cells cocultured with BLyS in the absence of IFNα were better protected form death in vitro than B cells expressing Mcl-1^f/null alleles (P < .001, Figure 6A). We speculated that gene dose could diminish the amount of Mcl-1 protein found in the absence of interferon and account for the difference in BLyS-mediated protection. Indeed, the amount of Mcl-1 protein in untreated Mcl-1^f/null B cells was approximately half that found in either Mcl-1^f/f or Mcl-1^f/wt B cells (Figure 6B). To determine whether the amount of Mcl-1 protein influenced BLyS-dependent protection, we assessed B-cell viability of Mcl-1^f/wt, Mcl-1^f/f, and Mcl-1^f/null populations cultured with and without BLyS in the presence or absence of IFN. In the absence of IFN, BLyS-dependent survival was comparable for Mcl-1^f/wt and Mcl-1^f/f B cells (P, not significant) throughout the culture period, whereas Mcl-1^f/null B cells were protected significantly less well (P < .01-.001) compared with Mcl-1^f/wt and Mcl-1^f/f populations (Figure 6C). When Mcl-1 was excised after IFN-induced Cre expression, the protective effect of BLyS was lost in both Mcl-1^f/f and Mcl-1^f/null cells but unaffected in Mcl-1^f/wt B cells (comparing BLyS to BLyS plus IFNα, Figure 6C). These data suggest that Mcl-1 expression is required for BLyS-dependent survival, and the amount of available Mcl-1 protein dictates the efficiency of BLyS protection.

Mcl-1 is a short-lived protein needing constant translational replenishment.³⁹  Both mTOR and Pim 2 inactivate the translation inhibitor 4E-BP1,^34,40  suggesting in the absence of BLyS that both of these pathways would be dormant and the production of Mcl-1 protein reduced. To model this condition, we assayed Mcl-1 protein in cultured apoptosis–resistant Bcl-2-tg⁺ B cells. Consistent with the hypothesis, we found Mcl-1 is reduced in B cells cultured unstimulated for 1 to 10 days, whereas Mcl-1 was induced/maintained in cultures supplemented with BLyS (Figure S4).

We showed previously that Pim-deficient B cells become refractory to BLyS-mediated protection when cultured with rapamycin (Figure 5B). To determine if Mcl-1 expression were sensitive to combined rapamycin treatment and Pim deficiency, we assessed Mcl-1 protein in cell extracts from wild-type or Pim 1– and 2–deficient B cells cultured with and without BLyS and/or rapamycin. There was a decline in Mcl-1 protein in unstimulated normal B cells over 2 days in culture exacerbated somewhat by rapamycin (Figure 7A). Normal B cells cultured with BLyS, however, maintained a higher amount of Mcl-1 than vehicle or rapamycin-treated B cells, which was not significantly diminished by culture with BLyS and rapamycin combined (Figure 7A). BLyS stimulation also maintained Mcl-1 expression in Pim 1– and 2–deficient B cells, but rapamycin had a stronger inhibitory effect; Mcl-1 protein in B cells cocultured with BLyS and rapamycin was comparable with that seen in vehicle or rapamycin-treated populations (Figure 7A). Bcl-2 and Bcl-X_L proteins were unaffected by Pim deficiency, BLyS, or rapamycin treatment (Figure S3).

Because either Akt/mTOR or Pim 2 signaling can promote BLyS-dependent B-cell survival in vitro, then the administration of rapamycin in vivo should reduce mature peripheral B cell populations in Pim-deficient but not wild-type mice. To test this prediction, age-matched Pim 1 and 2 doubly deficient mice and wild-type littermates were injected intraperitoneally with rapamycin or diluent every other day for a week before death. Spleen cell populations were quantified and B-cell subpopulations identified by flow cytometry. There was no significant change in the total or follicular B-cell populations in normal mice treated with rapamycin and an increase in marginal zone B cells (7 ± 2 × 10⁻⁶ vs 11 ± 1 × 10⁻⁶, P < .01; Figure 7B). In marked contrast total, follicular and marginal zone B cells in Pim-deficient mice were all significantly reduced (P < .001), by more than 45%, after a single week of rapamycin treatment (Figure 7B).

We demonstrate in these studies that the growth and survival of B lymphocytes are processes regulated directly by BLyS stimulation acting through 2 pathways, Akt/mTOR and Pim 2. B cells, like T cells, removed from the protective niches of the in vivo microenvironment undergo atrophy. Although nutrients are abundant in culture, their entry and use is restricted by the lack of transporters and enzymes regulated by growth factors.^1,41  BLyS functions as a growth factor; it increases amino acid and nucleic acid uptake and the glycolytic rate, and it raises basal metabolic activity and blocks B-cell shrinkage. In so doing, BLyS confers a coincident heightened functional responsiveness marked by faster kinetics of proliferation induced with lower doses of activator than is seen in freshly isolated B cells or B cells undergoing atrophy (R.T.W., M.R.S., laboratory observations, December 2006).

The signaling pathways that connect BLyS stimulation to Akt activation and Pim 2 production are an active area of investigation. Akt activation requires PI3-kinase producing phosphoinositides that facilitate membrane localization of Akt and PDK1 with subsequent phosphorylation of Akt at T308; full activation of Akt requires phosphorylation at S473 by a second kinase recently suggested to be PKCβ,⁴²  or the rictor containing mTOR complex mTORC2.²⁹  We find that BLyS-induced Akt phosphorylation at S473 is fully inhibited in primary B cells cultured for 24 hours with rapamycin, revealing a major role for mTORC2 in Akt activation. BLyS stimulation also activates the noncanonical NF-κB2 pathway,^15,43  required for the production of Pim 2.¹⁷  Akt and Pim 2 can synergize to amplify the canonical NF-κB pathway by activating the serine/threonine kinase Cot promoting the survival of transformed lymphocytes.⁴⁴  This is consistent with the finding that the selective activation of the canonical NF-kB pathway, by expressing constitutively active IKKβ, replaces BR3-dependent survival signals for peripheral B cells.⁴⁵ 

B cells stimulated with BLyS in vitro are rescued from atrophy, a process by which lymphocytes shrink in size when cultured in the absence of growth factors necessary to facilitate nutrient uptake. BLyS-induced atrophy protection was fully active in normal as well as Pim 1– and 2–deficient B cells and was partially dependent on mTOR, as it could be significantly inhibited by rapamycin. It has been shown previously that atrophy resistance induced by growth factors results from increased glucose uptake and use and the up-regulation of nutrient transporters, and is Akt dependent and rapamycin sensitive.^28,46,47  In our studies, we found a rapamycin-insensitive component of BLyS-induced atrophy protection that cannot be explained by the action of Pim1 or 2; BLyS-stimulated Pim-deficient B cells were no more sensitive to atrophy when cultured with rapamycin than wild-type B cells. Pathways upstream of mTORC1 could provide this atrophy protection. Although we found that prolonged rapamycin treatment can block Akt S473 phosphorylation, many of the normal effectors of Akt, Gsk-3, and TSC2, and the mTORC1 effectors, S6 kinase and 4E-BP1, are activated even when mTORC2 is inactive.⁴⁸  It may be noteworthy that mitogen-induced B-cell growth during blastogenesis requires Akt and NF-κB but is rapamycin-insensitive.⁴⁹  Our findings with B cells support the notion that growth factor–dependent regulation of metabolism is a common attribute of the homeostasis of resting lymphocytes.

The action of Akt/mTOR and Pim 2 fully accounts for ex vivo B cell viability mediated by BLyS stimulation. This is shown by our finding that rapamycin, an acute inhibitor of mTORC1, and chronic regulator of Akt completely suppresses BLyS-dependent survival in Pim 2–deficient B cells but has no effect on Pim 2–sufficient wild-type B cells. Akt has been shown to act in parallel with other signaling pathways to effect growth factor–dependent survival in lymphocytes.^21,36  In these instances, inhibition of the overlapping pathway by mutation or drugs renders the cells susceptible to rapamycin-induced death. The survival and growth promoting properties of Akt are well known,^27,50,–52  and inhibitors of PI3 kinase provoke apoptosis in B cells.^42,53  The rapamycin sensitivity of BLyS-induced protection in the absence of Pim 2 is probably because of the translational enhancement necessary to maintain the short-lived protein Mcl-1⁵⁴  in primary B cells, and it is noteworthy that Pim 2 modulation of translation is rapamycin resistant.^21,33  Alternatively, Akt phosphorylates and inactivates FOXO3a, a transcription factor necessary for Bim expression.^55,56  Because Bim is a major regulator of B cells,⁵⁷  suppression of this key proapoptotic protein could also occur through BLyS stimulation. Furthermore, FOXO3a phosphorylation is dependent on the action of Akt 308/473⁴⁸  and would be sensitive to the destabilization of the mTORC2 complex by chronic rapamycin treatment of B cells. Of the 3 members of the Pim kinase family, only Pim 2 appears to be required for BLyS-dependent survival. BLyS did not protect B cells deficient in Pim 2 when mTOR was inhibited, although Pim 1 was readily induced by BLyS stimulation of Pim 2–deficient B cells. These results also suggest that Pim 3, if induced, is unnecessary for rapamycin resistant apoptosis protection.

The molecular basis for BLyS-dependent apoptosis attenuation has been controversial. It has been reported that BLyS stimulation: (1) up-regulates the Bcl-2 family of antiapoptotic proteins including, Bcl-2, Bcl-x_L, or Bfl-1/A1,^17,–19,58  (2) inhibits the induction or function of the proapoptotic proteins Bak, Bim, or Bad^17,20,59,60  and/or (3) blocks the translocation of catalytically active PKCδ to the nucleus preventing histone phosphorylation leading to apoptosis.^61,62  Recent studies using murine and human B cells and our own data have failed to show consistent up-regulation of Bcl-2, Bcl-x_L, or Bfl-1/A1,^16,63,–65  and B cells deficient in PKCβ were not rescued by BLyS, although PKCδ translocation to the nucleus was blocked.⁴² 

Mcl-1 is an antiapoptotic Bcl-2 family member necessary for mature B and T cell survival.²²  Mcl-1 is known to interact preferentially with and sequester the proapoptotic proteins Bak,⁶⁶  Bax,⁶⁷  and Bim,^22,68  the major negative regulators of B-cell homeostasis.⁶⁹  Our studies show that BLyS modulates Mcl-1 protein and that BLyS could not induce apoptosis protection in B cells when the Mcl-1 gene was deleted. Unlike other Bcl-2 family members Mcl-1 has a relatively short half-life and is dynamically regulated³⁹ ; suppression of the Akt pathway^70,71  or mTOR,⁵⁴  results in a decline of Mcl-1 protein and the induction of apoptosis in hematopoietic cells. As for Pim 2, it is known that Pims and Akt/mTOR can mediate overlapping functions,⁷²  and both can activate the canonical NF-κB pathway through the serine/threonine kinase Cot producing cyclic adenosine monophosphate response element binding protein (CREB) and NF-kB, transcription factors critical for Mcl-1 expression.^73,74  Pim 2 and mTOR both inactivate the translation inhibitor 4E-BP1, enhancing translation efficiency.^34,40  Akt and Pim 2 both inhibit Gsk-3β, which, when active, phosphorylates Mcl-1, leading to its ubiquitinylation and accelerated decay.⁷⁵  BLyS induces the activation of extracellular signal-regulated kinase (Erk),²⁰  which can phosphorylate Mcl-1 at T163 slowing turnover⁷⁶  and also reduces the formation of enzymatically active PKCδ, which phosphorylates and accelerates the catabolism of Mcl-1.⁷⁷  Accordingly, BLyS stimulation of Akt/mTOR and Pim 2 can regulate the expression and/or maintenance of Mcl-1 at the transcriptional, translational, and posttranslational levels.

BLyS supports the survival of malignant, autoreactive B cells and normal B cells. Multiple myelomas, lymphomas, and non-Hodgkin leukemias use BLyS for survival,^78,,,–82  and a subset of transformed B cells can produce BLyS, establishing a pathologic autocrine loop associated with poor disease prognosis.⁸⁰  Dysregulated BLyS expression is also seen in rheumatic diseases, such as arthritis, lupus erythematosis, and Sjogren.^83,,–86  In autoimmune responses, BLyS stimulation overcomes tolerance rescuing autoreactive B cells normally targeted for elimination.^60,87  These considerations have fostered a clinical approach for treating both autoimmune B cells and B-cell cancers, which consists of inactivating BLyS with monoclonal antibodies or recombinant decoy receptors.^88,–90  We suggest that pharmacologic inhibition of B-cell survival signaling may be a useful adjunct to this current approach. BLyS is expressed as a biologically active membrane bound or soluble ligand,^91,92  and inhibition of membrane-bound ligand may require higher concentrations of BLyS competitor than are feasible.

The identification of 2 and only 2 pathways, Akt/mTOR and Pim 2 mediating BLyS-induced survival, reduces the array of pharmacologic targets for clinical intervention. Indeed, we find that simultaneously targeting these pathways using a brief course of rapamycin treatment in Pim 1– and 2–deficient mice significantly reduced both marginal zone and follicular B cell populations without affecting Pim-sufficient littermate controls. Rapamycin is a well-tolerated and highly efficient inhibitor of mTOR, and genetically induced Pim 2 deficiencies are relatively benign,⁹³  as most biologic processes mediated by Pim 2, other than B-cell survival, can be effected by other members of the Pim family. A combination therapy targeting mTOR and Pim 2 may have clinical efficacy for the treatment pathogenic B cells.

The authors thank Sarah Kenward for expert technical assistance, the members of the Thompson and Woodland laboratories for helpful comments and support, and Dr Randolph Noelle (Dartmouth Medical School) for special reagents. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health (NIH).

This work was supported by NIH grants AI041054 and AI057463 (C.B.T., R.T.W.), an institutional Diabetes Endocrinology Research Center grant DK32520, and a Special Fellowship from the Leukemia Society (3277-05; C.J.F.).

National Institutes of Health

Contribution: R.T.W. conceived and designed the research plan, performed experiments, and wrote the manuscript; C.J.F. and P.S.H. contributed to the research plan and provided critical reagents; J.T.O., S.J.K., and D.M.H. provided critical reagents; M.R.S. performed experiments and contributed to the research plan; and C.B.T. contributed to the experimental design and interpretation.

Conflict-of-interest disclosure: D.M.H. owns stock in Human Genome Sciences. The other authors declare no competing financial interests.

Stanley J. Korsmeyer died on March 31, 2005.

Correspondence: Robert T. Woodland, University of Massachusetts Medical School, Department of Molecular Genetics and Microbiology, 55 Lake Ave N, Worcester, MA 01655; e-mail: Robert.Woodland@umassmed.edu.