Antigen receptor–mediated changes in glucose metabolism in B lymphocytes: role of phosphatidylinositol 3-kinase signaling in the glycolytic control of growth

In response to antigen challenge, resting B lymphocytes exit the G₀ phase of the cell cycle and undergo a period of growth before committing to genome replication.^1,2  Growth corresponds to an accumulation of cell mass that is accompanied by increased size and is linked to increased de novo macromolecular synthesis.^3-5  That mammalian cell growth may be necessary for genome replication underscores its importance in adaptive immunity in that the clonal expansion of antigen-specific B lymphocytes is a prerequisite for humoral immune responses. Most investigations in B cells have focused on the role of genes whose function are important for B-cell antigen receptor (BCR)–induced protein synthesis and increased cell size.^4,5  It is recognized, however, that antigen receptor–triggered macromolecular synthesis and gene expression places enormous bioenergetic demands on lymphocytes.^5,6  Therefore, one of the fundamental aspects of B-cell responses to antigen challenge that may be critical in vivo is the provision of metabolic substrates to provide ATP and anabolic precursors for cellular growth.

Early studies in lectin-stimulated thymocytes highlighted the importance of glucose uptake and catabolism in providing energy and carbon for macromolecular synthesis.^7,8  Further, proliferating thymocytes meet their ATP demand mainly by glycolytic catabolism when sufficient glucose is available.⁹  It is widely viewed that glucose metabolism is regulated by homeostatic mechanisms wherein mammalian cells respond to a decreased ATP/ADP ratio by adjusting nutrient uptake and catabolism to meet the increased demand for ATP and macromolecular precursors. Recent work has brought new insights into the regulation of energy metabolism by suggesting that the bioenergetic activity of cells is not merely controlled by increased ATP demand but may also be coordinated by signal transduction pathways that act to directly modulate nutrient uptake and metabolism.^10-12 

Little is currently known about the intracellular signaling and metabolic pathways involved in the acquisition and utilization of glucose to support the bioenergetic demands encountered following BCR engagement.^13,14  Upon engagement of the BCR, quiescent B cells undergo a period of growth characterized by de novo RNA and protein synthesis and subsequently enter S phase of the cell cycle within 31 to 48 hours.^1,2  Treatment of B cells with inhibitors of phosphatidylinositol 3-kinase (PI-3K) activity in early G₁ phase of the cell cycle blunts the ongoing increase in cell size, suggesting that failure of anti-Ig–stimulated B cells to commit to genome replication in the absence of PI-3K activity results from a block at a critical growth checkpoint.¹⁵  It has been suggested that in this situation, impaired growth may result from an inability to enact Rel/NF-κB– and c-myc–dependent gene expression programs.^4,16,17  Importantly, genomewide analyses have found that Myc increases the expression of specific glycolytic enzyme genes.^18-20  It is therefore plausible that impaired growth of B cells in the absence of PI-3K activity may reflect an inability to supply glucose-derived ATP and/or anabolic precursors via glycolysis. With this in mind, we sought to determine if B lymphocytes increase glucose utilization in response to BCR engagement and to identify the signaling and metabolic pathways linking antigen receptors to glucose energy metabolism.

Antiphospho(Ser473)Akt and antiphospho(Thr389)p70S6 kinase antibody (Ab) were from Cell Signaling Technology (Beverly, MA). The anti-Glut1 Ab was from Research Diagnostics (Concord, MA). PE-conjugated F(ab′)₂ fragments of goat anti–rabbit IgG were obtained from CalTag Laboratories (Burlingame, CA). Anti-CD16/CD32 (FcγIII/II receptor) 2.4G2 monoclonal antibody (mAb) and the Cytofix/Cytoperm Kit were purchased from BD Biosciences (San Jose, CA). F(ab′)₂ fragments of goat anti–mouse IgM were obtained from Jackson ImmunoResearch Labs (West Grove, PA). F(ab′)₂ fragments of rabbit anti–mouse IgG and rabbit anti–mouse IgG were obtained from ICN Cappel (Aurora, OH). Wortmannin, rapamycin, and LY294002, were obtained from CalBiochem-NovaBiochem (San Diego, CA). Cholera toxin subunit B (CT-B)–Alexa Fluor 488 was obtained from Molecular Probes, Invitrogen (Carlsbad, CA). Protein determinations of whole cell extracts were carried out with a Bio-Rad Protein Assay Dye Reagent (Bio-Rad Laboratories, Hercules, CA).²¹ 

BALB/c mice and p85α-deficient mice (BALB/cAnNTac-Pik3r1 N12) were obtained from Taconic Farms (Germantown, NY) and housed at Boston College. Mice were cared for and handled at all times in accordance with National Institutes of Health and Boston College guidelines. Splenic B cells from mice at 8 to 12 weeks were purified and cultured in RPMI 1640 medium plus 10% FCS as previously described.²¹  Small dense B cells were then isolated following centrifugation through a discontinuous Percoll gradient as described by Hodgkin et al.²²  The A20 cell line expressing a myristoylated Akt–estrogen receptor fusion (Akt-mER) was provided by Dr Michael Gold (Department of Microbiology and Immunology, University of British Columbia).²³ 

B cells were resuspended at 3 × 10⁷/mL in uptake buffer (10 mM HEPES [pH 7.4], 136 mM NaCl, 4.7 mM KCl, 1.25 mM CaCl₂, and 1.25 mM MgSO₄). Transport was initiated by mixing 200 μL B cells with 200 μL uptake buffer containing 200 μM 2-deoxy-d-glucose (2-DOG) and 2 μCi (7.4 × 10⁴ Bq)/mL [³H]2-DOG (Amersham Biosciences, Piscataway, NJ). At the indicated times, 50 μL of the transport mixture was loaded onto a discontinuous gradient of 200 μL bromododecane per 40 μL 20% perchloric acid, and the gradient was then centrifuged at 14 000g (90 seconds).²⁴  Following centrifugation, the upper aqueous and bromododecane layers were decanted and the perchloric acid phase collected and placed into a vial containing 5 mL scintillation cocktail. The radioactivity was then quantitated by liquid scintillation spectrometry. To assess the contribution of glucose uptake mediated by glucose transporters, uptake of [³H]2-DOG was measured in the presence of 10 μM cytochalasin B, a potent inhibitor of glucose transporter activity.

B cells were washed twice in staining buffer (1 mL PBS containing 1% FCS/0.1% NaN₃) and then incubated for 20 minutes (4°C) with the anti-CD16/CD32 (2.4G2) mAb Fc block reagent (1:500 vol/vol). Cells were washed twice in 0.1 mL staining buffer and resuspended in 250 μL Cytofix/Cytoperm solution (4°C). After 20 minutes, cells were washed twice in 0.5 mL Perm/Wash solution, resuspended in 0.1 mL Perm/Wash solution, and incubated with 1:500 dilution of anti-Glut1 Ab or isotype control Ab (4°C). After 60 minutes, cells were washed 3 times in 0.1 mL Perm/Wash solution and then incubated with PE-conjugated F(ab′)₂ fragments of goat anti–rabbit IgG (1:800) for 1 hour (4°C). Cells were then washed in 0.1 mL Perm/Wash solution, and Glut1 staining was measured by flow cytometry on a BD FACSCanto cytometer and the data analyzed by FACSDiva software (BD Biosciences).

Splenic B cells were centrifuged onto glass slides using a Cytospin (Thermo Electron, Pittsburgh, PA) and incubated with 5 μg/mL CT-B–Alexa Fluor 488 for 30 minutes at 22°C. Slides were washed once with PBS and incubated in 3.7% (vol/vol) formaldehyde for 20 minutes (22°C). Permeabilization was performed in 0.5% (vol/vol) Triton X-100 (5 minutes), and slides were then blocked for 30 minutes in 2% (wt/vol) BSA (22°C). Slides were incubated for 1 hour (22°C) with anti-Glut1 rabbit polyclonal Ab, followed by a TRITC-conjugated goat anti–rabbit IgG (Jackson ImmunoResearch Labs). After 4 washes with PBS, slides were mounted with Aqua Polymount and analyzed by confocal laser scanning microscopy at 488 nm excitation (Leica TCS SP2 confocal microscope; Leica Microsystems, Wetzlar, Germany). A 100 ×/0.53 numeric aperture objective was used to visualize the images. Leica Confocal Software version 2.61 was used to acquire the images.

B cells (10⁶ cells/0.5 mL) were incubated with [5-³H]glucose (Amersham Biosciences) for 90 minutes. The initial rate of anti-Ig–induced glycolysis remained linear for 180 minutes. At the indicated times, 100 μL of cells was removed and placed in 1.5 mL microcentrifuge tubes containing 50 μL 0.2 N HCl. The microcentrifuge tubes were then placed in 20 mL scintillation vials containing 0.5 mL water and the vials capped and sealed. ³H₂O was separated from unmetabolized [³H]glucose by evaporation diffusion for 48 hours (25°C). The amount of diffused and nondiffused tritium was quantitated by liquid scintillation spectrometry and compared with parallel vials containing [5-³H]glucose only and ³H₂O alone.¹⁴  Tricarboxylic acid (TCA) cycle flux from glucose catabolism was measured by the rate of ¹⁴CO₂ production from [6-¹⁴C]glucose (Amersham Biosciences).¹⁴  Glucose and lactate were measured using assays obtained from Sigma Diagnostics (St Louis, MO) according to the manufacturer's instructions.

B cells (3 × 10⁷) were cultured in RPMI 1640 containing 10 mM [1-¹³C]glucose or [2-¹³C]glucose (Cambridge Isotope Laboratories, Andover, MA). B cells were extracted 3 times with 1 mL of 70% (vol/vol) ethanol. Washing of the pelleted cells was avoided because it caused efflux of a large portion of the lactate. The soluble ethanol extracts were pooled, frozen in liquid nitrogen, and lyophilized. The dry material was then resuspended in 0.5 mL D₂O for ¹H NMR analyses at 28°C. ¹H- and ¹³C-coupled ¹H spectra were acquired using a Varian INOVA 500 spectro-photometer (Varian, Palo Alto, CA) equipped with a 5-mm indirect detection probe as described previously.²⁵  In brief, conventional ¹H NMR spectra were acquired using a 5006.6 Hz sweep width, 9984 datum points, 90-degree pulse width, 3.0-second recycle delay time, and 128 transients. The 1D heteronuclear multiple quantum coherence (HMQC) spectra (which detect only those protons coupled to ¹³C) were acquired with a sweep width of 7509.6 Hz, 2048 datum points, 90-degree pulse width, a 1.0-second recycle delay time, and ¹³C decoupled. For ¹³C decoupling, the average one-bond ¹H-¹³C J-coupling constant was set to 135 Hz. Spectra were obtained using 4000 transients. Under these acquisition conditions, the lactate methyl doublet was not resolved, but integrated intensities of small molecules at early time points could easily be quantified.

We sought to determine if BCR signaling modulates glucose uptake in B cells. To this end, small dense (quiescent) splenic B cells exhibited a relatively low rate of 2-DOG uptake, whereas stimulation with 10 μg/mL F(ab′)₂ fragments of anti–mouse IgM (anti-Ig) for 15 hours resulted in an approximately 10-fold increase in 2-DOG uptake (Figure 1A). The initial rate of anti-Ig–stimulated 2-DOG uptake was linear for approximately 90 seconds and judged specific as cytochalasin B, an inhibitor of facilitated glucose transporters, blocked the increased 2-DOG uptake (Figure 1A, inset). BCR crosslinking induces PI-3K activation within minutes, leading to activation of Akt and p70S6 kinase.¹⁵  To determine if PI-3K activity was necessary for the increased glucose uptake following BCR ligation, LY294002 and wortmannin were used to inhibit PI-3K. In the presence of LY294002, anti-Ig–stimulated glucose uptake was significantly reduced in comparison with nontreated B cells (Figure 1B). Pretreatment of B cells with wortmannin also resulted in inhibition of anti-Ig–induced glucose uptake, although not to the same extent as that observed with LY294002 (Figure 1B). The efficacy of the PI-3K inhibitors was demonstrated insofar as pretreatment of B cells with LY294002 or wortmannin blocked BCR-induced Akt phosphorylation on Ser473 (Figure 1C). To determine the contribution of the mammalian homolog mTOR, we tested the effects of the mTOR inhibitor rapamycin on anti-Ig–induced glucose uptake. In several experiments, rapamycin only had a moderate inhibitory effect on anti-Ig–induced glucose uptake (Figure 1B). Importantly, rapamycin inhibited anti-Ig–induced p70S6 kinase phosphorylation on Thr389, which is dependent on mTOR, indicating that the concentration of rapamycin used was sufficient to inhibit mTOR (Figure 1C). These data suggest that BCR ligation acts to increase glucose transport and does so via a pathway dependent on PI-3K activity.

We next determined if the increased glucose uptake in response to anti-Ig might be attributed to changes in glucose transporter expression. Expression of Glut1 was evaluated, given reports that Glut1 is the primary facilitated glucose transporter expressed on hematopoietic cells.²⁶  Although quiescent B cells expressed Glut1, the expression of Glut1 increased in response to anti-Ig stimulation as determined by Western blot analysis and flow cytometry (Figure 2A-B, respectively). Of note, we also measured Glut1 staining of nonpermeabilized B cells cultured in medium alone or stimulated with anti-Ig for 12 hours; the mean fluorescence intensity of surface Glut1 expression increased 5-fold following anti-Ig stimulation (data not shown). Anti-Ig–induced Glut1 expression in B cells pretreated with LY294002 mirrored that of quiescent B cells, suggesting that increased Glut1 expression requires PI-3K activity (Figure 2B). Consistent with these findings, anti-Ig–induced Glut1 expression was impaired in splenic B cells from p85α-null mice, which exhibit markedly reduced BCR-induced PI-3K activity (Figure 2B).^27-29  BCR-induced Glut1 expression in wild-type B cells was not sensitive to rapamycin (data not shown). We also evaluated B cells for expression of Glut1 by immunofluorescence. To monitor surface staining, B cells were incubated with Alexa Fluor 488–conjugated CT-B prior to fixation (Figure 2C, left panels). CT-B binds to glycosphingolipids, with a strong affinity for GM1, and therefore can be used as a plasma membrane marker.³⁰  Quiescent B cells exhibited a relatively low level of anti-Glut1 Ab staining, which was increased in response to BCR engagement (Figure 2C).

To examine the significance of increased glucose uptake, we evaluated whether BCR signaling increased glucose flux into several metabolic pathways involved in glucose catabolism. Generation of [³H]H₂O resulting from the dehydration of [5-³H]glucose catalyzed by enolase, the penultimate step in the glycolytic pathway, was monitored as a readout for glycolysis.¹²  Quiescent B lymphocytes exhibited a measurable, albeit relatively low, rate of glycolysis (Figure 3A). Anti-Ig–stimulated B cells increased glycolysis when measured at 12 hours, which continued to increase for 31 hours, remaining elevated for at least 48 hours (Figure 3A). BCR-induced glycolysis was inhibited more than 90% in the presence of 1 mM 2-DOG (Figure 3A), a potent inhibitor of glycolysis.³¹  The increase in glycolysis following BCR engagement was blocked by LY294002 or wortmannin and, to a much lesser extent, by rapamycin (Figure 3B). To corroborate these findings, we examined ex vivo splenic B cells isolated from mice deficient in the p85α subunit of PI-3K.^21,29  We observed a striking impairment in BCR-induced glycolysis in p85α-null B cells in comparison with wild-type B cells (Figure 3C). To understand the biologic significance of increased glycolysis, we incubated quiescent B cells with 1 mM 2-DOG along with anti-Ig. 2-DOG blocked increases in cell size in response to anti-Ig, as assessed by mean forward scatter (Figure 3D). In addition, 2-DOG blocked increases in protein content in response to anti-Ig (Figure 3D). Notably, 2-DOG reduced cell number by 11% when measured at 18 hours (data not shown). Taken together, these results suggest that glycolytic flux is necessary for BCR-induced B-cell growth.

During the course of these experiments we observed that the increase in glycolysis following BCR crosslinking was accompanied by production of lactate (Figure 3E). We estimate that more than 80% of the glucose consumed in anti-Ig–stimulated B cells, as measured by glucose depletion from the culture medium, was converted to lactate when assayed at 24 hours and 48 hours. An important caveat to this interpretation is that carbon sources other than glucose can contribute to the lactate pool (eg, from incomplete oxidation of glutamine via malate and PEP-carboxykinase enzymes). To more fully assess glucose metabolism, we compared the flux of radiolabeled glucose through the glycolytic pathway and the TCA cycle in anti-Ig–stimulated B cells. Glucose flux through the glycolytic pathway was substantial in comparison with glucose oxidation through TCA cycle flux (Table 1). Because the ¹⁴CO₂ generated in these assays derives from the complete oxidation of [6-¹⁴C]glucose, we cannot entirely rule out the possibility that some glucose may be incompletely oxidized via a “truncated” TCA cycle.

We next sought to discern further the role of BCR signaling in controlling glucose carbon flow. Our approach was to use a ¹H NMR experiment, which specifically monitors ¹³C fixation from [1-¹³C]glucose into metabolites by selecting for only those ¹H's coupled to ¹³C nuclei.²⁵  We favored ¹H NMR because it allowed glucose metabolism to be followed immediately in response to BCR engagement, whereas the biochemical assays used in Figure 3A-C did not exhibit the sensitivity necessary to reproducibly monitor changes in glucose utilization early in response to anti-Ig stimulation (ie, 0 to 8 hours). There was little detectable difference in the normal ¹H spectrum between quiescent and anti-Ig–stimulated B cells during an initial 8-hour period, quantified by integrating the resonances for the lactate methyl group and the glutamate methylenes as a function of time after anti-Ig stimulation (data not shown). However, use of a 1D-HMQC sequence to select out protons attached to ¹³C adequately could follow glucose metabolism in quiescent and anti-Ig–stimulated B cells (Figure 4A-B, respectively). Integration of the ¹³C-coupled ¹H resonances revealed that both the lactate methyl group and the glutamate methylenes exhibited increased ¹³C content as the [1-¹³C]glucose was metabolized during the initial 8 hours of anti-Ig stimulation (Figure 4C). The specific increase in ¹³C could be monitored by normalizing the ¹³C-coupled ¹H resonance integral to the total ¹H resonance integral (Figure 4D). These data indicate that enhanced glucose utilization, as evidenced by increased ¹³C incorporation into the methyl group of lactate from [1-¹³C]glucose, was apparent within 2 hours following BCR engagement. Glutamate pools also increased in ¹³C content in response to BCR ligation. Notably, incorporation of ¹³C label into glutamate C-4 from transamination of the TCA cycle intermediate α-ketoglutarate reflects metabolism of [1-¹³C]glucose by the TCA cycle. C-3 of glutamate also becomes ¹³C labeled as material repetitively goes through the TCA cycle and the symmetric intermediate succinate. In data not shown, we observed that ¹³C label into glutamate moderately increased until 12 hours and then declined by 24 hours to a level similar to quiescent B cells.

We carried out a similar 1D-HMQC experiment to evaluate the contribution of PI-3K activity in BCR-induced ¹³C incorporation from [1-¹³C]glucose into the methyl group of lactate. In B cells stimulated via the BCR for 8 hours, incorporation of ¹³C label increased approximately 3.5-fold above quiescent B cells (Figure 4E, control). Incorporation of ¹³C label into the glutamate carbons also increased approximately 2.7- to 2.8-fold. The citrate pool, much of which derived from the tissue culture medium, was not appreciably labeled during the activation period. Pretreatment of B cells with wortmannin or LY294002 blocked ¹³C incorporation into both lactate and glutamate (Figure 4E), whereas rapamycin had little effect on ¹³C fixation into either lactate or glutamate (Figure 4E, +rap).

Because several glucose-derived metabolites from the pentose phosphate pathway can enter the glycolytic pathway upstream of enolase, the dehydration reaction of 2-phosphoglycerate used herein may reflect glucose catabolism via both glycolysis and the pentose phosphate pathway. To monitor diversion of glucose into the pentose phosphate pathway, a 1D-HMQC experiment was carried out with [2-¹³C]glucose. If [2-¹³C]glucose is used instead of [1-¹³C]glucose, there will be no ¹³C incorporation into the lactate methyl group by the glycolytic pathway; however, if the pentose phosphate pathway is operational, ¹³C will be incorporated into the methyl group of lactate. [1-¹³C]glucose rapidly labeled the lactate methyl group following BCR engagement (Figure 4F, •). In parallel B cells cultured with [2-¹³C]glucose, no significant ¹³C incorporation was detected in the lactate methyl group prior to 24 hours after BCR crosslinking (Figure 4F, ▪). As ¹³C label from [2-¹³C]glucose appeared in the lactate methyl group, it was depleted from the methine group of lactate (Figure 4F, ○), a finding consistent with decreased glucose catabolism by the glycolytic pathway and increased catabolism via the pentose phosphate pathway. By 48 hours, [2-¹³C]glucose flux through the pentose phosphate pathway represented approximately 60% of the total lactate production (Figure 4G). These results suggest that as stimulated B cells progress through the G₁/S transition, a shift occurs in the metabolism of glucose from primarily glycolytic to include the pentose phosphate pathway.

Murine A20 B cells expressing a myristoylated Akt-ER fusion protein (Myr Akt-ER) were used to determine if glucose utilization is contingent upon Akt activation, a downstream target of PI-3K. The Myr Akt-ER is constitutively expressed in A20 cells but remains inactive in the absence of 4-hydroxytamoxifen (4-HT).²³  Myr Akt-ER–expressing A20 B cells cultured in medium alone or stimulated with anti-Ig were devoid of detectable Myr Akt-ER Ser473 phosphorylation; addition of 4-HT led to increased Ser473 phosphorylation, a modification that correlates with catalytic activity (Figure 5A). The 4-HT–treated A20 B cells exhibited an increase in glucose uptake in comparison with nontreated A20 B cells (Figure 5B). Activation of the Myr Akt-ER fusion protein also resulted in higher glycolytic flux than nontreated A20 B cells (Figure 5C). In data not shown, A20 B cells expressing an empty vector did not exhibit changes in glucose utilization in response to 4-HT. Exposure to 4-HT did not alter the proliferation of Myr Akt-ER–expressing A20 B cells in comparison with nontreated A20 B cells, indicating that the sum macromolecular synthesis was equivalent (data not shown). These findings suggest that Akt activation is sufficient to increase glucose utilization in A20 B cells.

The FcγRIIB is a low-affinity receptor for the Fc portion of Ig, with IgG immune complexes as its physiologic ligand.³²  Coclustering of the BCR with FcγRIIB provokes a dominant-inhibitory signal that blocks B-cell growth induced by BCR engagement and ultimately leads to apoptosis.^33-35  FcγRIIB inhibits BCR signaling by reducing the accumulation of phosphoinositide 3-kinase products and recruiting SH2-containing inositol-5-phosphatase to FcγRIIB, hydrolyzing PIP₃, and inhibiting distal Akt activation.^36,37  We therefore tested if coengagement of FcγRIIB inhibited glucose utilization induced by BCR crosslinking. The level of Akt phosphorylation on Ser473 was significantly reduced by intact anti-Ig, which coligates the BCR and FcγRIIB, in comparison with F(ab′)₂ fragments of anti-Ig (Figure 6A). Inhibition of Akt phosphorylation required FcγRIIB engagement, because B cells incubated with intact anti-Ig plus mAb 2.4G2 exhibited levels of Ser473 phosphorylation comparable to BCR crosslinking alone. Coclustering of the BCR and FcγRIIB resulted in an approximately 82% reduction in glucose uptake as compared with B cells stimulated via the BCR alone (Figure 6B). Intact anti-Ig also blocked BCR-induced Glut1 expression and glycolysis (Figure 6C-D, respectively). The 1D-HMQC experiments described in Figure 4E also revealed that B cells treated with intact anti-Ig exhibited decreased ¹³C incorporation into both lactate and glutamate (Figure 6E). These results suggest that limiting glucose utilization may represent an important mechanism underlying the inhibitory action of FcγRIIB on B-cell growth.

The results herein suggest that the BCR regulates glucose utilization in small dense B cells by increasing glucose uptake and expression of the facilitated Glut1 transporter. The importance of this signaling-induced response is underscored in that quiescent B cells do not contain significant glycogen reserves. Therefore, the capacity to signal increased glucose transport in response to antigen challenge provides a mechanism to rapidly acquire glucose-derived energy and biosynthetic precursors necessary to support growth. Our results also indicate that BCR engagement leads to a rapid and sustained increase in glucose metabolism. The results from 1D-HMQC reveal that glycolysis is increased within 2 hours of BCR ligation, preceding the increase in cell size and entry into the cell cycle.^1,2,15  Our results also indicate that incubation of B cells with 2-DOG, a potent inhibitor of glycolysis resulting from the competitive inhibition of phosphoglucoisomerase and depletion of hexokinase II,³¹  prevents BCR-triggered increases in cell size and protein synthesis, suggesting that heightened glycolytic flux is a requirement for B-cell growth.

These results also suggest that prior to S-phase entry, glucose catabolism shifts from primarily glycolysis to include the pentose phosphate pathway. The biologic significance of the shift toward pentose phosphate pathway–mediated glucose catabolism is currently unknown. This pathway has several purposes, including providing reduced coenzyme NADPH for reductive biosynthesis and ribose 5-phosphate for nucleotide biosynthesis. That elevated pentose phosphate pathway activity coincides with the onset of S-phase entry may reflect a demand for ribose 5-phosphate to support nucleic acid synthesis. Notably, however, growth factor stimulation is accompanied by a rise in intracellular H₂O₂^38-40  Recent reports provide compelling support for the requirement of NADPH, not ribose 5-phosphate, for cellular proliferation, presumably to maintain the optimal redox status for cellular proliferation.^41,42 

The rapid increase in glycolysis following BCR engagement is accompanied by an increase in lactate production, with more than 80% of the glucose consumed by B cells being converted to lactate. An important caveat to our calculations is the possibility that nonglucose sources may contribute to lactate production, which might be significant given that glutamine oxidation accounts for 33% of lactate production in activated T cells.⁴³  At this juncture we cannot eliminate the contribution of nonglucose sources to lactate production; however, in a preliminary 1D-HMQC experiment, [¹³C]glutamine metabolism in anti-Ig–stimulated B cells (15 hours) was not accompanied by a measurable increase in ¹³C incorporation into lactate (data not shown). These results suggest that anti-Ig–stimulated B cells carry out aerobic glycolysis characterized by increased lactate production from glucose despite the presence of an intact TCA cycle.

The high rate of aerobic glycolysis in response to anti-Ig stimulation of primary B lymphocytes observed herein mirrors the situation in tumor cells. This might reflect a situation in which the capacity to carry out oxidative degradation in the mitochondria is limited in normal B cells. Of interest, Fitzpatrick and coworkers¹⁴  reported that B-cell hybridomas exhibit low levels of pyruvate dehydrogenase, which in turn might represent a barrier for glycolytic intermediates entering the TCA cycle. We note that in response to limiting extracellular glucose, B cells compensate by increasing glycolysis but not oxygen consumption (data not shown). Alternatively, B cells that have encountered antigen may rapidly transition to a high rate of glycolysis to allow for precise control over the time during which ATP and anabolic precursors are needed to support growth, clonal expansion, and Ig synthesis.^44,45  This would provide B cells the capacity to redirect glucose carbon flow into the pentose phosphate pathway to provide NADPH and ribose 5-phosphate or redirect the glycolytic end product, pyruvate, into a truncated TCA cycle. It has been demonstrated that citrate can be exported to the cytosol, thereby serving as a substrate for ATP citrate lyase, to produce acetyl-CoA and oxaloacetate⁴⁶ ; the former is a precursor for fatty acid and cholesterol synthesis and functions, at least in part, to promote membrane biogenesis.⁴⁷ 

Our results also suggest that increased glucose utilization results from a primary PI-3K signaling pathway originating from the BCR. This conclusion is supported by the findings that LY294002 and wortmannin block BCR-induced glucose uptake and glycolysis. In addition, splenic B cells from mice deficient in the class 1A PI-3K regulatory subunit p85α (Pik3r1), which exhibit impaired BCR-coupled PI-3K activation,^28,29  fail to increase Glut1 expression and glycolysis. We cannot rule out that a homeostatic response to decreased ATP and/or glycolytic metabolites associated with growth contributes, along with PI-3K signaling input, to increased glycolysis; however, we find that conditional activation of Akt is sufficient to stimulate glucose utilization in A20 B cells in the absence of increased growth. Taken together, these findings suggest that in B lymphocytes PI-3K/Akt signaling functions to regulate glucose utilization, thereby providing B cells the capability to meet the bioenergetic demands associated with growth.

In terms of downstream Akt effector signaling components, our results fail to conclusively implicate mTOR in BCR-induced glucose metabolism, because rapamycin treatment does not significantly abate the increase in glucose transport and glycolysis in B cells. This is surprising given the important role of Akt and mTOR signaling in regulating nutrient uptake in hematopoietic cells.^12,24  Interestingly, the growth of splenic B cells was reported to be only partially inhibited by rapamycin, whereas LY294002 completely blocked B-cell growth.¹⁶  In this study NF-κB–dependent c-myc transcription was shown to be critical for anti-Ig–stimulated B-cell growth. The potential significance of this pathway to the findings herein is underscored by the observation that many glycolytic genes are directly regulated by c-Myc.^18-20  Further, recent gene expression profiling in ex vivo splenic B cells revealed that BCR engagement up-regulates several genes encoding enzymes of the glycolytic pathway.⁴⁸  These findings taken together raise the possibility that PI-3K may link the BCR to increased glycolysis via c-Myc–dependent transcription of genes encoding glycolytic enzymes. In addition, increased glucose utilization may be regulated via posttranslational mechanisms. The key rate-limiting step in glycolysis is the phosphorylation of fructose-6-phosphate catalyzed by 1-phosphofructokinase (PFK).^49,50  Akt has been shown to phosphorylate PFK-2, an enzyme that generates fructose-2,6 bisphosphate, an allosteric regulator of PFK.⁴⁹  Along these lines, the atypical protein kinase C (PKC) isoforms, PKCλ and PKCζ, together with Akt, have been implicated in insulin-stimulated Glut4 translocation to the plasma membrane.^50,51 

In summary, our results provide evidence for the first time that both growth-promoting BCR signaling and growth-inhibitory FcγRIIB signaling directly modulate glucose energy metabolism. These results also raise the possibility that manipulation of the BCR-linked pathways involved in glucose acquisition and catabolism may prove effective in the treatment of lymphoproliferative disorders that arise from the clonal expansion of normal B lymphocytes.

We thank Dr David Fruman (Center for Immunology, Department of Molecular Biology and Biochemistry, University of California–Irvine) for helpful suggestions.