Glucose-dependent de Novo Lipogenesis in B Lymphocytes

Background: The metabolic requirements underlying B lymphocyte differentiation are poorly understood. Results: Differentiation is accompanied by glucose metabolism into fatty acid and cholesterol synthesis, mediated by ATP-citrate lyase (ACLY). Conclusion: ACLY-dependent lipogenesis is required for several phenotypic changes defining plasma cell differentiation. Significance: This study proposes a critical role for ACLY coupled glucose-dependent de novo lipogenesis in LPS-induced B lymphocyte differentiation. Bacterially derived lipopolysaccharide (LPS) stimulates naive B lymphocytes to differentiate into immunoglobulin (Ig)-secreting plasma cells. Differentiation of B lymphocytes is characterized by a proliferative phase followed by expansion of the intracellular membrane secretory network to support Ig production. A key question in lymphocyte biology is how naive B cells reprogram metabolism to support de novo lipogenesis necessary for proliferation and expansion of the endomembrane network in response to LPS. We report that extracellularly acquired glucose is metabolized, in part, to support de novo lipogenesis in response to LPS stimulation of splenic B lymphocytes. LPS stimulation leads to increased levels of endogenous ATP-citrate lyase (ACLY), and this is accompanied by increased ACLY enzymatic activity. ACLY produces cytosolic acetyl-CoA from mitochondrially derived citrate. Inhibition of ACLY activity in LPS-stimulated B cells with the selective inhibitor 2-hydroxy-N-arylbenzenesulfonamide (compound-9; C-9) blocks glucose incorporation into de novo lipid biosynthesis, including cholesterol, free fatty acids, and neutral and acidic phospholipids. Moreover, inhibition of ACLY activity in splenic B cells results in inhibition of proliferation and defective endomembrane expansion and reduced expression of CD138 and Blimp-1, markers for plasma-like B cell differentiation. ACLY activity is also required for LPS-induced IgM production in CH12 B lymphoma cells. These data demonstrate that ACLY mediates glucose-dependent de novo lipogenesis in response to LPS signaling and identify a role for ACLY in several phenotypic changes that define plasma cell differentiation.

The molecular mechanisms underlying how membrane phospholipid and cholesterol content increases during B lymphocyte differentiation have received little attention (18,19,21,22). Overall rates of fatty acid synthesis in most mammalian tissues are relatively low, in part, because lipids are acquired from the circulation. By contrast, cancer cells and proliferating mammalian cells exhibit high rates of de novo lipogenesis in order to support membrane biogenesis associated with proliferation as well as fundamental cellular processes, such as signal transduction (23)(24)(25)(26). Along these lines, one possible biosynthetic fate of glucose is metabolism by fatty acid and mevalonate synthesis pathways, whereby glucose-derived carbon in the form of pyruvate is imported into the mitochondria, decarboxylated to acetyl-CoA, and then condensed with oxaloacetate to form citrate (11,12). Citrate can be exported from the mitochondria via the malate-citrate shuttle system and used as a substrate for ATP-citrate lyase (ACLY) (13,27,28). ACLY catalyzes the formation of acetyl-CoA and oxaloacetate from cytosolic citrate and coenzyme A in the presence of ATP (27)(28)(29). ACLY therefore serves as a cross-link between glucose metabolism and the fatty acid and mevalonate synthetic pathways.
The acly gene is expressed in a wide variety of tissues and deletion of the gene in mice results in embryonic lethality (30 -33). ACLY is abundantly expressed in the liver and white adipose tissue and is also expressed and catalytically active in pancreatic beta cells (13,31,34). Receptor signaling via the PI3K/ Akt pathway stimulates ACLY activity predominantly through phosphorylation of ACLY also on serine 454; phosphorylation of ACLY contributes to its protein stabilization (27,34,35). Several additional protein kinases have been shown to phosphorylate ACLY, including nucleoside diphosphate kinase and cyclic AMP-dependent protein kinase (34). Alterations in expression or activity of ACLY have been observed in a variety of different metabolic and pathological conditions (13). Notably, ACLY is up-regulated and activated in several types of human cancers, and its inhibition induces proliferation arrest in cancer cells in vitro and in vivo (11, 26, 34 -38). We sought herein to investigate whether B lymphocytes stimulated with LPS acquire the capacity to increase cellular lipid content via a pathway that involves ACLY-mediated acetyl-CoA production from extracellularly acquired glucose. Our studies also investigate whether ACLY plays a role in the differentiation of mature B lymphocytes in response to LPS stimulation.
B Lymphocyte Isolation and Culture-BALB/cJ mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Mice were cared for and handled at all times in accordance with National Institutes of Health and Boston College Institutional Animal Care and Use Committee guidelines. Splenic B cells from mice at 8 -12 weeks were isolated and purified by negative selection using a B cell isolation kit, MidiMACS TM separator, and LS column (Miltenyi Biotec Inc., Auburn, CA). Small dense B cells were isolated following centrifugation through a discontinuous 72:65:50% Percoll gradient. B cells were cultured in RPMI 1640 medium (Mediatech, Inc., Manassas, VA) plus 10% fetal calf serum (FCS), 2 mM L-glutamine, 10 mM HEPES, pH 7.4, 50 M ␤-mercaptoethanol, 50 units/ml penicillin, and 50 g/ml streptomycin as described (6).
ELISA-CH12 B lymphoma cells were cultured in the presence or absence of LPS for the indicated times. Tissue culture supernatant was collected (250 l, corresponding to 10 4 CH12 B cells), and the amount of IgM was measured by ELISA using a mouse IgM quantitation set (E90-101) according to the manufacturer's instructions (Bethyl Laboratories, Inc., Montgomery, TX).
Flow Cytometry and Cell Cycle Measurements-B cells (10 5 cells) were collected, washed in PBS, and resuspended in PBS containing 0.1% Triton X-100, 50 g/ml propidium iodide, and 50 g/ml RNase A for 30 min (37°C). DNA content was measured by flow cytometry using a FACSCanto cytometer (BD Biosciences). Acquired data were analyzed with Modfit LT version 3.0 (Verity Software House, Topsham, ME). To assess viability, B cells (10 4 cells) were incubated in PBS containing 0.5% FCS and 1.0 g/ml propidium iodide followed by flow cytometry as described (6). Proliferation of B lymphocytes (10 4 cells) was measured by bromodeoxyuridine (BrdU) incorporation and flow cytometry using a Pharmingen FITC-BrdU flow kit according to the manufacturer's instructions. For CD138 surface staining, 10 6 splenic B cells were washed with ice-cold staining buffer (PBS containing 1% FCS, 0.1% NaN 3 ) and then incubated with 100 l of staining buffer containing 1.5 l of rat anti-mouse CD15/CD32 Ab (BD Biosciences) for 20 min. Cells were then stained with PE-conjugated rat anti-mouse CD138 Ab or PE-conjugated rat IgG2a, isotype control (BD Biosciences) for 45 min (4°C). Cells were washed three times with staining buffer and analyzed by flow cytometry. B lymphocyte-induced maturation protein-1 (Blimp-1) staining was carried out with 10 6 splenic B cells that were washed with ice-cold staining buffer and then fixed and permeabilized with the transcription factor buffer set according to the manufacturer's instructions (BD Pharmingen); cells were stained with Alexa Fluor 647-conjugated rat anti-mouse Blimp-1 (BD Biosciences) and analyzed by flow cytometry. For ER-Tracker TM staining, 10 6 splenic B cells were washed with ice-cold PBS. Cells were then incubated with 1 M ER-Tracker TM dye for 30 min (37°C), washed three times with PBS, and analyzed by flow cytometry. In all experiments, analysis was carried out on the gated viable cell population. Flow cytometry was carried with a BD FACSCanto cytometer, and the data were analyzed by FACSDiva software (BD Biosciences).
Lipid Extraction and High Performance Thin Layer Chromatography (HPTLC)-B cells (7.5 ϫ 10 7 ) were collected and washed in PBS, and total cellular lipid was extracted in methanol/chloroform (1:1) overnight (25°C). Neutral and acidic lipids were separated by DEAE-Sephadex (A-25, GE Healthcare) column chromatography as described (40). For each condition and/or time point, an equal number of B cells were extracted. To ensure that the same amount of lipid was spotted on the chromatography plate, 10 g of olealyl alcohol was added to the initial B cell extract. As an additional control for extraction efficiency, B cell extracts were spiked with 10 g of lysophosphocholine and cerebrosides because these lipids were not detected in the bulk lipid isolation of B cells. Lipids were then separated by HPTLC; the chromatography plates were developed first to 4.5 cm (of 10-cm total length) in chloroform/methanol/acetate/formate/water (35:15:6:2:1, v/v/v/v/v) and subsequently to 10 cm with hexane/diisopropyl ether/acetate (65: 35:2, v/v/v) as described (41,42). Lipids were visualized by charring in the presence of a solution of 3% Cu(OAc) 2 and 8% H 3 PO 4 . To monitor glucose incorporation into lipid, B cells (7.5 ϫ 10 7 ) were cultured in medium containing 1 Ci/ml D-[U-14 C]glucose. Neutral and acidic lipids were then isolated and separated by HPTLC, and the 14 C-labeled lipids were visualized by autoradiography. Densitometry of individual bands was carried out with a CamagTLC scanner equipped with ImageQuant software. The data represent -fold increase in individual lipids induced in response to LPS stimulation of B cells in comparison with unstimulated B cells and are normalized for HPTLC loading using the olealyl alcohol loading control.
ACLY Enzyme Activity Measurements-B cells (5 ϫ 10 6 ) were collected by centrifugation and lysed by the addition of 150 l of lysis buffer (10 mM Tris, pH 8.0, 150 mM NaCl, 20 mM deoxycholate, 1% Nonidet P-40, and 0.1% SDS) supplemented with 1 mM PMSF and protease inhibitor mixture. The lysate was centrifuged at 14,000 ϫ g for 10 min, the supernatant was collected, and ACLY activity was measured by the method of Ma et al. (43). The amount of [ 14 C]acetyl-CoA produced was then measured by liquid scintillation spectrophotometry.
Gas Chromatography-Mass Spectroscopy Analysis-B cells (6.5 ϫ 10 6 ) were cultured in modified RPMI 1640 containing 10% dialyzed FCS and 50% [U-13 C]glucose (1.0 g/liter) or 50% [U-13 C]glutamine (0.3 g/liter). At the indicated times, B cells were centrifuged at 500 ϫ g, and the supernatant was collected and stored at Ϫ80°C. The resulting cell pellet was washed several times in PBS and then stored at Ϫ80°C. The cell pellets were extracted with a 1:1:1 mix of water/methanol/chloroform. The chloroform extracts were methylated to produce fatty acid methyl esters and analyzed by gas chromatography-mass spectroscopy (GC-MS) as described (44). C17:0 fatty acid was added to the samples to serve as an internal standard. Relative amounts of fatty acids were calculated based on the intensity of methyl ester ions and adjusted for internal standard as needed. Mass spectra of fatty acid methyl esters were corrected for heavy isotopes at all positions in the molecules except for the carbon atoms in the fatty acids (44); these corrected spectra were used to deduce the percentage of fatty acid synthesis and 13 C labeling of constituent acetyl units in fatty acids using methods from Lee et al. (45), modified as described (46).

RESULTS
Distribution of Neutral and Acidic Lipids in Naive and LPSstimulated Splenic B Lymphocytes-We initially evaluated bulk neutral and acidic lipid distribution in control and LPS-stimulated ex vivo splenic B lymphocytes (41). Individual neutral lipids were identified based on their migration distance on HPTLC in comparison with neutral lipid standards and a purified murine brain neutral lipid fraction (Fig. 1A). In the absence of LPS stimulation, B lymphocytes contained several predominant neutral lipids, including PE, phosphatidylcholine (PC), cholesterol, cholesterol esters, and a relatively lower abundance of ceramide (CM) and sphingomyelin. Similarly, acidic lipids were identified based on migration with acidic lipid standards and a purified murine brain acidic lipid fraction (Fig. 1A); the major acidic lipids detected in unstimulated B cells corresponded to phosphatidylinositol (PI), phosphatidylserine (PS), and cardiolipin (CL) (Fig. 1A). In an experiment not shown, we identified by two-dimensional thin layer chromatography the nature of an LPS-induced acidic lipid detected in the region of the sulfatide doublet on HPTLC as corresponding to phosphatidylglycerol (PG) (Fig. 1A). Stimulation of B cells with LPS for 48 and 72 h resulted in elevated levels of several neutral lipids, including PC, PE, CM, and cholesterol ( Fig. 1, A and B); this time frame corresponds to the initial endomembrane expansion and induction of IgM production in ex vivo splenic B cells stimulated with mitogenic antigens (15). LPS stimulation also resulted in the production of triglycerides (Fig. 1A) as well as increased levels of the acidic lipids, PI, PS, and CL in comparison with unstimulated B cells (Fig. 1, A and B). The HPTLC also identified free fatty acids in the acidic lipid fraction; however, we observed that the migration distance of the free fatty acid band relative to the origin on HPTLC varied over several experiments, as did the relative -fold increase in free fatty acids in response to LPS.
Stimulation of B Lymphocytes with LPS Leads to Glucose-dependent de Novo Fatty Acid and Lipid Synthesis-We next determined if extracellular glucose was directly utilized to support de novo lipid synthesis in B lymphocytes. For these experiments, B cells were cultured in the absence (24 h) and presence of LPS (24,48, and 72 h) in medium supplemented with [ 14 C]glucose. B cells were then collected, neutral and acidic lipid fractions were isolated, and the amount of [ 14 C]glucose incorporation into individual lipids was determined by autoradiography following HPTLC. As shown in Fig. 2A, under the experimental conditions used, unstimulated B cells exhibited no detectable incorporation of 14 C label into the predominant neutral and acidic lipids previously identified by bulk lipid analysis (see Fig. 1A). Unstimulated B cells exhibited incorporation of 14 C label into two slower migrating neutral lipid species ( Fig.  2A), which was increased at 24 h and then declined at 48 and 72 h following LPS stimulation ( Fig. 2A). The identity of these lipid species is currently unknown. In contrast to unstimulated B cells, LPS treatment led to increased 14  It should be noted that under these experimental conditions, the presence of 14 C label incorporation into long chain polyunsaturated fatty acids, such as arachidonic acid (C20:4) and C22:6, was not detected. The level of arachidonic acid (C20:4) in the total unlabeled lipid fraction was low and similar in naive B cells and in the LPS-stimulated B cells after 72 h (ϳ5.9 mol %).
We next sought to determine the contribution of glucose metabolism to de novo fatty acid synthesis by GC-MS. B cells were cultured in the presence of [ 13 C]glucose or [ 13 C]glutamine during LPS stimulation for 24 and 48 h. Mass spectra for 13 Clabeled fatty acids were assumed to be dependent on the random incorporation by fatty acid synthase of 13 C-labeled or unlabeled acetyl units from acetyl-CoA into the fatty acid molecule (up to C16:0 palmitate, incorporating 8 acetyl units) and could therefore be fitted to a binomial model. This fitting process (described in detail in Refs. 45   extension of palmitate to C18:0 stearate (versus salvage) by elongation was calculated based on the extra ϩ2 mass labeling in stearate, corresponding to the incorporation of an extra labeled acetyl unit, compared with the expected mass labeling based on the mass distribution in palmitate (46). Upon LPS stimulation for 24 h, incorporation of 13 C label from glucose to acetyl-CoA increased to 36% (Fig. 3A), whereas the incorporation of 13 C label from glutamine to acetyl-CoA was not affected with LPS stimulation. At 48 h after LPS stimulation, the percentage contribution of [ 13 C]glucose (36%) and [ 13 C]glutamine (10%) to the acetyl-CoA pool combined to be equivalent to 92% of the total acetyl-CoA content (because input of 13 C label is 50% of either substrate). These results suggest that in response to LPS, the majority of acetyl-CoA used for fatty acid synthesis is derived from glucose; most of the remainder comes from glutamine, and only a small fraction (the remaining 8%) is derived from unlabeled sources, such as turnover of (unlabeled) fatty acids or other metabolites. The labeling of free fatty acids was also evaluated in control and LPS stimulated B cells. The total relative amount of free fatty acids incorporating 13 C label increased with LPS stimulation at 24 h and to a greater extent at 48 h (Fig. 3B). In each case, the amount of 13 C label converted to fatty acids was normalized to extracts of unstimulated B cells cultured with [ 13 C]glutamine for 24 h. Without stimulation, the percentage of de novo synthesis of free cellular palmitate (C16:0) was less than 4% but increased to 60% at 48 h after LPS stimulation (Fig. 3C). A similar increase was observed in sterate (C18:0) synthesis, whereas myristate (C14:0) synthesis could not be measured in unstimulated B cells because of low amounts and low labeling but was also ϳ60% newly synthesized after 48 h in stimulated B cells (Fig. 3C). We also observed varying increases in the relative amounts of free fatty acids (C14:0, C16:0, C16:1, C18:0, and C18:1 (two species corresponding to a major and minor peak eluting from the GC), C18:2, C20:0) in LPS-stimulated B cells (Fig. 3B).
Expression and Activity of ACLY in B Lymphocytes-ACLY catalyzes a key regulatory step in de novo lipogenesis in mammalian cells by cleaving mitochondrially derived citrate to generate cytosolic acetyl CoA plus oxaloacetate (28,29). To investigate the role of ACLY in B lymphocytes, endogenous ACLY protein expression was measured by Western blot in ex vivo splenic B cells culture in medium alone or stimulated with LPS (Fig. 4). ACLY was expressed in unstimulated B cells, and the level of ACLY protein increased over time following LPS stimulation (Fig. 4). ACLY undergoes phosphorylation on serine 454 in response to growth factor stimuli, and this modification has been reported to increase its enzymatic activity (27,47). We found a relatively low level of phosphorylated ACLY in unstimulated B lymphocytes (Fig. 4). In response to LPS stimulation, ACLY phosphorylation on serine 454 was increased at 6 h following LPS stimulation but more substantially between 12 and 72 h (Fig. 4). To determine if ACLY enzymatic activity was modulated in response to LPS stimulation of B cells, we measured ACLY activity in whole cell extracts prepared from unstimulated and LPS-stimulated B cells. ACLY enzymatic activity was measured as a function of the amount of [ 14 C]acetyl-CoA produced from the conversion of [ 14 C]citrate by endogenous ACLY (43). Naive B cells exhibited a basal level of ACLY activity, which was further increased 2-and 6-fold in response to LPS when measured at 12 and 48 h, respectively (Fig. 5A). Similar results were obtained by a second independent enzyme assay based on coupling the ACLY enzymatic reaction to malate dehydrogenase (data not shown) (48).
ACLY Activity Is Required for Glucose-dependent de Novo Lipogenesis in B Lymphocytes-We next determined the contribution of ACLY to de novo lipogenesis by inhibiting ACLY with the pharmacological inhibitor compound-9 (C-9) (39). C-9 is a cell-permeable 2-hydroxy-N-arylbenzene-sulfonamide analog that specifically inhibits ACLY activity in mammalian cells with an IC 50 ϭ 0.13 M and demonstrates efficacy in vivo mouse models (39). C-9 inhibited endogenous B cell ACLY activity (data not shown), in keeping with previous reports in mammalian cells (39). In control experiments, treatment of splenic B , and 0.047 (C20:0). The two C18:1 species are shown in the order they elute from the GC, with the major peak first and the minor peak second. In C, the p values comparing B cells cultured in the absence of LPS with B cells stimulated with LPS at 24 and 48 h correspond to 6.6 ϫ 10 Ϫ5 and 6.0 ϫ 10 Ϫ5 for C16:0, respectively, and 5.2 ϫ 10 Ϫ3 and 5.8 ϫ 10 Ϫ4 for C18:0, respectively. Cells were then collected, cellular lysates were prepared in Triton X-100 buffer, and equivalent amounts of total protein were examined by Western blot for total ACLY (ACLY), ACLY phosphorylated in serine 454 (pACLY-(Ser454)), and ␤-actin as a loading control for electrophoresis, as described previously (6). Individual bands were subjected to densitometric scanning using ImageQuant software. Numbers represent -fold increase in pACLY(Ser-454) and ACLY bands at each time point following LPS stimulation in comparison with B cells cultured in medium alone. The data are representative of two independent experiments. cells with the ACLY inhibitor C-9 did not block LPS-induced glycolysis when measured at 12 h (Fig. 5B). We next determined if culture of LPS-stimulated B cells with [ 14 C]glucose in the presence or absence of C-9 blocked the incorporation of 13 C label into the major neutral lipids at 24 and 48 h, as detectable by autoradiography and HPTLC (Fig. 6, A and B). Incorporation of 13 C label into the acidic lipids was also blocked by C-9 in LPS-stimulated B cells treated at 24 and 48 h (Fig. 6, A and B). Treatment of B cells with C-9 also eliminated the increase in de novo synthesis incorporation of palmitate or stearate fatty acid species in LPS-stimulated B cells when measured at 36 h by GC-MS (Fig. 7A) and completely reversed the increase in the amounts of various free fatty acids observed with LPS stimulation (Fig. 7B). As mentioned above, 13 C label arachidonic acid (C20:4) was not detectable by HPTLC and autoradiography following incubation of B cells with [ 14 C]glucose; however, GC-MS revealed the presence of 13 C-labeled arachidonic acid (C20:4) in unstimulated splenic B cells cultured with [ 13 C]glucose (Fig. 7C). LPS stimulation of B cells led to a statistically significant, albeit relatively small, increase in 13 C-labeled arachidonic acid (C20:4) of ϳ0.7-fold above that of unstimulated B cells (Fig. 7C). The increase in 13 C label incorporation into arachidonic acid (C20:4) following LPS stimulation was blocked by C-9 (Fig. 7C). Collectively, these data suggest that ACLY activity is required for glucose-dependent de novo lipid synthesis in B lymphocytes.
We also determined if activation of naive B lymphocytes by other surface receptor-linked pathways led to glucose-dependent de novo lipogenesis. Activation of B cells via cross-linking the B cell antigen receptor with anti-Ig together with IL-4 led to increased 13 C label incorporation into several neutral and acidic lipids (Fig. 8, A and B). Qualitatively, a similar profile of neutral and acidic lipids incorporated [ 14 C]glucose; however, the relative amounts of [ 14 C]glucose incorporation into individual lipids differed in B cells stimulated with anti-Ig plus IL-4 in comparison with LPS. This may reflect that anti-Ig plus IL-4 induces proliferation of naive B cells, whereas LPS induces both proliferation and differentiation, the latter requiring additional lipid production (to support endomembrane expansion for Ig secretion) beyond that needed for cell growth and division. The increase in [ 14 C]glucose label incorporation into neutral and acidic lipids following stimulation of B cells with anti-Ig plus IL-4 was also reduced following inhibition of ACLY activity with C-9 (Fig. 8, A and B). Whereas treatment of LPS-stimulated B cells with C-9 reduced 13 C label incorporation into neutral and acidic lipids to levels not detectable by HPTLC and autoradiography (Fig. 6), treatment of anti-Ig plus IL-4-stimulated B cells with C-9 led to reduced 13 C label incorporation that varied with respect to individual neutral and acidic lipids (Fig. 8, A and B). Our studies also revealed that culture of B lymphocytes with the combination of anti-Ig plus IL-4 resulted in increased ACLY phosphorylation on serine 454 (Fig. 8C). Moreover, inhibitors of PI3K blocked the increase in ACLY serine 454 phosphorylation, suggesting the involvement of PI3K in regulating ACLY activity (Fig. 8C). (20). We therefore determined whether ACLY activity was necessary for B cell proliferation in response to LPS. Inhibition of ACLY activity with C-9 was accompanied by reduced BrdU incorporation into DNA in LPS-stimulated B cells, corresponding to 0.9% (48 h) and 0.4% (72 h) in comparison with BrdU incorporation in parallel B cells cultured in the absence of C-9 (8.2% at 48 h and 15.3% at 72 h) ( Table 1). B lymphocytes stimulated with LPS in the presence of C-9 also exhibited a reduction in B cell viability corresponding to 31.2 and 13.5% at 48 and 72 h, respectively ( Table 1). The decrease in cell viability was accompanied by an increased percentage of hypodiploid DNA content, corresponding to 28.3 to 35.0% at 48 and 72 h, respectively (Table 1).

ACLY Activity Is Required to Maintain Survival of ex Vivo B Lymphocytes in Response to LPS-Proliferation of mammalian cells requires a doubling of membrane phospholipid content
ACLY Activity Is Required for Several Phenotypic Changes Defining Plasma Cell-like Differentiation-The results above suggest that inhibition of ACLY activity blocks glucose-dependent synthesis of neutral and acidic lipids, including PC and CM (Fig. 6A); both of these lipids are required for ER expansion in B lymphocytes (17)(18)(19). It therefore seemed likely that ACLY activity might contribute to expansion of the endomembrane secretory network in response to LPS stimulation. To assess the expansion of the ER compartment, we stained ex vivo B lymphocytes with Dapoxyl ER-Tracker TM , an ER-specific probe. Intracellular staining of B cells cultured in the absence of LPS revealed a small area of staining that expanded substantially following stimulation with LPS for 24 h (Fig. 9A). Treatment of LPS-stimulated B cells with C-9 resulted in decreased intracellular staining (Fig. 8A). Flow cytometry of ER-Tracker TM stained B cells also revealed a significant reduction in mean fluorescence intensity in LPS-stimulated B cells treated with C-9 (Fig. 9B). For these experiments, the viable B cell population was gated and subsequently analyzed for fluorescence intensity by flow cytometry.
To investigate further the involvement of ACLY activity in plasma-like B cell differentiation, we evaluated whether ACLY activity was required for LPS-induced surface CD138 (Syndecan-1) expression in splenic B lymphocytes; CD138 is expressed upon B cell differentiation into plasmablasts and plasma cells (15). LPS stimulation of splenic B cells increased surface CD138 expression (Fig. 10A). To minimize the confounding effects of decreased viability at later time points observed with C-9, B cells were instead treated with Medica 16, a small molecule inhibitor of ACLY that is structurally unrelated to C-9 and has been used both in vitro and in vivo (33). Treatment of LPSstimulated B cells with Medica 16 resulted in a significant reduction in CD138 expression in viable B cells analyzed by flow cytometry (Fig. 10A). We also evaluated the expression of Blimp-1, which is required for development of immunoglobulin-secreting cells and for maintenance of long lived plasma cells (16). We found that LPS-induced Blimp-1 expression was reduced in splenic B cell cultures incubated with Medica 16 (Fig. 10B).
We also sought to determine the effects of ACLY inhibition on IgM production. For these studies, the CH12 B lymphoma cell line was used as a model system to evaluate the role of ACLY in the differentiation of B cells into IgM secreting cells  (17,18,49). CH12 B lymphoma cells express surface IgM and MHC II and undergo morphological differentiation and IgM secretion in response to LPS, similar to splenic B cells (17,49). CH12 B lymphoma cells stimulated with LPS for 24 h exhibited an increase of [ 14 C]glucose incorporation into neutral and acidic lipid species that was blocked following inhibition of ACLY activity with C-9 (data not shown). To test whether ACLY activity was necessary for IgM production, CH12 B lymphoma cells were stimulated with LPS in the presence and absence of C-9 and IgM levels in the tissue culture supernatant measured using enzyme-linked immunosorbent assay. The production of IgM at 48 and 72 h in LPS-stimulated CH12 B lymphoma cells was significantly reduced in the presence of C-9 (Fig.  10C); the decrease in IgM production was not due to reduced cell viability following exposure to C-9 (data not shown).

DISCUSSION
LPS induces the proliferation and differentiation of B lymphocytes into Ig-secreting plasma cells. A doubling of membrane phospholipid content precedes proliferation, whereas differentiation involves a major morphological change of the cell interior, including a substantial expansion of the rough ER and Golgi network (17)(18)(19). Distinct changes in the ER network are evident as early as 24 h after exposure of naive B lymphocytes to LPS (14,15). B cells are capable of anticipating their secretory role via enacting a sequential and multistep program of protein expression to accommodate the massive Ig production that follows (4,5). A recent metabolomic study identified several bioenergetic pathways that discriminate between the proliferation and Ig-secreting phases (50). By contrast, the metabolic reprogramming necessary to support de novo phospholipid synthesis has not been described in depth. Lymphocytes and hematopoietic cells must access extracellular nutrients, such as glucose in order to support growth and effector function (1, 2). PI3K/Akt signaling represents a mechanism that links antigen receptor engagement to glucose uptake and glycolytic metabolism in B and T lymphocytes (2,6,9,10). Germane to the studies herein, one of the biosynthetic fates of glucose metabolism in highly proliferating cancer cells is its conversion to fatty acids and cholesterol (11,13,51,52). We reasoned therefore that LPS signaling might promote de novo lipogenesis, at least in part, by reprogramming glucose utilization.
Our results indicate that the levels of neutral and acidic lipids detected by HPTLC increase after LPS stimulation of B lymphocytes, but the bulk composition does not change substantially, consistent with a recent report in CH12 B lymphoma cells  (C18:0), and 1.5 ϫ 10 Ϫ2 (C20:0), respectively. The two C18:1 species are shown in the order they elute from the GC, with the major peak first and the minor peak second. C, the relative amount of arachidonic acid (C20:4) was also calculated based on peak total ion counts by GC-MS. The data are expressed as -fold change over unstimulated B cells cultured for 5 h. Error bars, S.D. of duplicate measurements. The p value comparing B cells stimulated with LPS for 36 h (*) with B cells cultured in medium alone (U) for 36 h corresponds to p Ͻ 0.05. (18). LPS leads to an increase in neutral and acidic lipid biosynthesis that is dependent on the acquisition of extracellular glucose. The predominant neutral lipids whose de novo synthesis derives carbon, in part, from glucose, measured by 13 C label incorporation from [ 14 C]glucose, include PE, PC, cholesterol, and CM, whereas the acidic lipids include PS, PI, PG, and cardiolipin. Subsequent stable isotope tracking identified 13 C-labeled fatty acids whose synthesis relies on [ 13 C]glucose in response to LPS stimulation of B cells. The predominant fatty acids identified by GC-MS include eicosanoic acid (20:0), palmitate (16:0), and palmitoleate (16:1) as well as stearate (18:0) and oleate (18:1), which are derived from the desaturation and/or elongation of palmitate. Collectively, these findings support a role for glucose in de novo lipogenesis in splenic B cells.
In addition to de novo synthesis, we cannot rule out the possibility that the cellular levels of lipids in B lymphocytes may also result from the uptake of fatty acids from external sources. On this point, polyunsaturated fatty acid species (C20:4 or C22:6) were present in the total unlabeled lipid fraction;   however, we were unable to detect (by HPTLC and autoradiography) de novo synthesis of long chain polyunsaturated fatty acids (C20:4 or C22:6) from [ 14 C]glucose. By contrast, GC-MS analysis revealed the presence of detectable 13 C-labeled arachidonic acid (C20:4) in naive B cells cultured with [ 13 C]glucose; LPS stimulation led to a statistically significant, albeit modest, increase in 13 C-labeled arachidonic acid (C20:4). It is possible that acquisition of polyunsaturated fatty acids under the experimental conditions used herein is derived from an exogenous source (e.g. serum in the culture medium). A similar finding was reported in a previous study of concanavalin A-stimulated lymphocytes (53). GC-MS analysis revealed a time-dependent decrease in several C16:0 and C18:0 fatty acids in control B lymphocytes devoid of LPS stimulation. Because these analyses were carried out with viable B cells, we do not believe the findings reflect cell death. Lymphocytes require the continuous presence of extrinsic signals (e.g. pro-survival cytokines) from their in vivo microenvironment to maintain cell viability. Ex vivo T cells cultured in medium devoid of these signals undergo atrophy and eventually died. Germane to the studies herein, atrophy is characterized by reduced cell size and decreased glycolytic rate (54). It is possible that the time-dependent decrease in glucose-dependent C16:0 and C18:0 de novo synthesis observed in ex vivo B lymphocytes devoid of LPS reflects decreased glycolysis over time in culture and thus reduced glucose-derived metabolites available to support de novo fatty acid synthesis.
We cannot rule out the possibility that other metabolites, in addition to glucose, may be used to support de novo fatty acid synthesis. Fatty acid synthesis precursors can be produced via other metabolic pathways, such as glutamine (55,56). For example, in tumor cells with defective mitochondria, reductive carboxylation of glutamine-derived ␣-ketoglutarate produces citrate to support de novo lipogenesis (55). The findings herein indicate that the majority of acetyl-CoA used to support de novo fatty acid synthesis is derived from glucose and not glutamine when measured at 24 h following LPS stimulation, whereas at 48 h, a relatively smaller fraction of acetyl-CoA is derived from glutamine. Interestingly, analysis of metabolites consumed and released during B lymphocyte differentiation in ex vivo cultures revealed recruitment of distinct metabolic pathways during the proliferation phase (i.e. lactate production increased), whereas the onset of Ig secretion was parallel with glutamine being used primarily as carbon and energy sources (50).
Our findings suggest that ACLY activity is required for glucose-dependent de novo lipogenesis, as evidenced by inhibition of glucose incorporation into fatty acid, cholesterol, and neutral and acidic lipids following treatment of B cells with the ACLY inhibitor C-9. The results also indicate that ACLY is regulated at several levels by LPS. LPS stimulation of B lymphocytes promotes an increase in ACLY enzymatic activity. This is probably attributable, in part, to increased endogenous ACLY protein levels following LPS stimulation; however, we also found that phosphorylation of ACLY on serine 454 is increased; phosphor-  ylation of ACLY on 454 has been linked to increased catalytic activity (27,47). Taken together, these results point to a key role for ACLY activity in regulating glucose-dependent de novo lipogenesis in LPS-stimulated B lymphocytes.
Our studies also revealed that stimulation of naive B cells with the combination of anti-Ig plus IL-4 resulted in increased ACLY phosphorylation on serine 454. Moreover, inhibitors of PI3K blocked the increase in ACLY serine 454 phosphorylation, suggesting the involvement of PI3K in regulating ACLY activity. By contrast, we were unable to demonstrate a role for PI3K activity in LPS-induced ACLY phosphorylation on serine 454 (data not shown). Whereas treatment of LPS-stimulated B cells with the ACLY inhibitor, C-9 reduced glucose-dependent de novo synthesis of acidic and neutral lipids to levels not detected by HPTLC and autoradiography, C-9 did not completely block de novo neutral and acidic lipid synthesis in anti-Ig plus IL-4stimulated B cells. The molecular basis for this discrepancy remains unclear at present.
It is possible that additional regulatory mechanisms may direct glucose utilization to support de novo lipogenesis in response to LPS. For example, fatty acid synthase has emerged as a key regulatory enzyme, given its role in the synthesis of long chain fatty acids by using acetyl-CoA as a primer, malonyl-CoA as a two-carbon donor, and NADPH (25). Moreover, a recent global analysis of lipogenic genes expressed in LPS-stimulated CH12 B lymphoma cells revealed the existence of a genetic program underlying the coordinate regulation of fatty acid synthesis genes, fatty acid modification (acyl chain elongation and desaturation), and PC biosynthesis (18).
Several lipids identified herein whose de novo synthesis is increased in response to LPS, including PI and cholesterol, play critical roles in B lymphocyte signaling and/or differentiation (3,57). Along these lines, PC synthesis is increased following LPS stimulation. PC is the most abundant membrane phospholipid and the primary phospholipid in the ER compartment; it is also a precursor to two other abundant phospholipids in B cells, sphingomyelin and PE (reviewed in Ref. 20). PC represents the major membrane lipid in CH12 B lymphoma cells, and its levels increase in response to LPS stimulation (18). Both proliferation and endomembrane expansion require an increased supply of phospholipids to support membrane and organelle biogenesis, a requirement met in part by increased PC synthesis (17,18). A recent report demonstrated that CTP:phosphocholine cytidylyltransferase ␣, a rate-controlling enzyme in the biosynthesis of PC, was necessary for B lymphocyte proliferation in vivo and in vitro in response to LPS (21). We show herein that inhibition of de novo lipogenesis following treatment of B cells with the ACLY inhibitor C-9 results in proliferation arrest and decreased survival of LPS-stimulated splenic B cells. Interestingly, B cells treated with C-9 at 48 h exhibit a significant decrease in cell viability (31.2%), whereas at 24 h, viability remains relatively high (89.2%). It is possible that B cells possess the capacity to maintain viability short term (i.e. for ϳ24 h) following inhibition of de novo lipogenesis, perhaps by acquiring lipids from the tissue culture medium; however, this may not be sufficient to maintain viability long term in the absence of de novo lipogenesis. It is noteworthy that cell death caused by ACLY blockage in tumor cells has been attributed to a shortage of cholesterol and isoprenoids (36,58). Apoptosis caused by inhibiting genes downstream of ACLY, such as fatty acid synthase, can be rescued by fatty acid supplementation in some cancer cell types (36).
Our results also reveal that LPS-induced CM synthesis is dependent on glucose metabolism via an ACLY-dependent pathway. CM synthesis is required for ER expansion and protein glycosylation in plasma cells (19). CM synthesis is up-regulated in an X-box binding protein 1 (Xbp1)-dependent manner and in turn results in increased levels of sphingolipids (19,20). XBP-1 is a key transcription factor that drives the unfolded protein response (20). B cells deficient in xbp-1 reveal markedly defective Ig secretion in vivo in response to immunization and in vitro following LPS stimulation of ex vivo splenic B cells (22). Exogenous expression of XBP-1 is sufficient to induce many of the phenotypic changes that define plasma cell differentiation, including increased cell size, increased protein synthesis, and expanded organelle biogenesis (16). Moreover, Xbp1-deficient B cells exhibit impaired LPS-induced accumulation of several lipids, including PI and PC, and this correlates with reduced expansion of the endomembrane secretory pathway (22). We show that impaired de novo lipogenesis, resulting from ACLY inhibition, correlates with an inability to physically expand the ER compartment in B lymphocytes stimulated with LPS.
The decrease in splenic B cell viability after 31 h of treatment with C-9 precluded evaluation of the role of ACLY activity in IgM production in LPS-stimulated B cells. Stimulation of B cells with lower doses of LPS (e.g. 1.0 and 10 g/ml) in the presence of C-9 did not allow the recovery of a sufficient number of B cells to reproducibly measure IgM production (data not shown). We were, however, able to monitor the expression of several B cell differentiation markers by flow cytometry in viable lymphocyte populations treated with the ACLY inhibitor Medica 16. LPS-induced surface expression of CD138, a marker for B cell differentiation, was reduced following inhibition of ACLY activity. The increased expression of Blimp-1 in LPSstimulated splenic B cells was also reduced in the presence of Medica 16. Collectively, these findings support a role for ACLY activity in plasma-like splenic B cell differentiation.
The role of ACLY activity was also evaluated in CH12 B lymphoma cells, which represents a model for studying LPS-induced IgM production in B cells (17,18,49). We found that LPS-stimulated CH12 B lymphoma cells undergo glucose-dependent de novo lipogenesis and that inhibition of ACLY activity blocks glucose incorporation into neutral and acidic lipids (data not shown). Our results revealed that ACLY activity is required for LPS-induced IgM production in CH12 B lymphoma cells, as evidenced by a significant reduction in IgM secretion following treatment with the ACLY inhibitor C-9. In summary, our study provides many new insights into the molecular pathways underlying the reprogramming of glucose metabolism that accompanies B lymphocyte differentiation. The results suggest a previously unknown role for glucose utilization and ACLY activity in the de novo lipogenesis during LPS-induced B lymphocyte differentiation. Taken together, the results from primary B lymphocytes and CH12 B lymphoma cells suggest that ACLY plays a role in several phenotypic changes that define B lymphocyte differentiation. This pathway may provide new targets for therapeutic exploration in treating plasma cell dyscrasias.