Homocysteine Activates B Cells via Regulating PKM2-Dependent Metabolic Reprogramming

Homocysteine (Hcy) is a sulfur-containing amino acid formed during the metabolism of the essential amino acid methionine. Accumulating evidence suggests that hyperhomocysteinemia (HHcy) is an independent risk factor for cardiovascular diseases in which inflammation plays a key role (1, 2). Our previous studies have shown that HHcy accelerates early atherosclerotic lesion formation in apolipoprotein E–deficient (ApoE^−/−) mice and that Hcy stimulation in vitro and ex vivo can induce B cell proliferation and IgG Ab secretion (3–5). However, the direct effects of HHcy on B cell function in vivo, the underlying mechanisms, and the potential pathophysiological significance remain to be elucidated.

Recent studies have revealed the interaction of multiple pathways in the regulation of immune and metabolic systems (6). Alterations in metabolism at both the cellular and tissue level affect specific lymphocyte functions (6). The Warburg effect, or aerobic glycolysis, was first discovered in highly proliferating tumor cells (7). Recently, similar metabolic changes have also been observed in immune cells. Activated dendritic cells, M1 macrophages, and effector T cells can switch their metabolic program from oxidative phosphorylation to aerobic glycolysis to meet the bioenergetic and biosynthetic demands of cell growth or effector functions (6, 8, 9). Although B cells share several features with T cells, it has recently been reported that B cells increase their rate of both glycolysis and oxidative phosphorylation in a relatively balanced fashion upon BCR or LPS stimulation (10). Moreover, in the intestinal immune system, IgA⁺ plasma cells in the intestinal lamina propria use both glycolytic and oxidative metabolism, whereas naive B cells in Peyer’s patches preferentially use oxidative metabolism (11). These investigations have revealed an important role of metabolic reprogramming in B cell activation.

Glucose metabolism is important for B cell activation (12). Pyruvate kinase is one of the key enzymes in the glycolytic pathway. There are four mammalian pyruvate kinase isoforms. Pyruvate kinase muscle isozyme 2 (PKM2) is mainly expressed in embryonic cells and tumor cells, whereas pyruvate kinase muscle isozyme 1 (PKM1) is found in highly differentiated tissues, such as muscles and the brain. The pyruvate kinase RBC isozyme and pyruvate kinase liver isozyme are tissue-specific isoforms and are found in RBCs (pyruvate kinase RBC isozyme) or in liver and kidney cells (pyruvate kinase liver isozyme) (13). Of all these isoforms, PKM2 has been the most extensively studied in tumor cells and has been found to be critical for tumor cell growth (14–16). The expression of PKM2 in tumor cells allows for an increase in both glycolytic and anabolic metabolic rates to support cell growth and proliferation (14). There have been a few recent reports showing that PKM2 is also required for normal cells (17–20). M1 macrophages upregulate PKM2 expression to increase glycolytic flux in support of cell activation (18, 19). Upon activation, B cells increase their cellular metabolism and proliferate rapidly. However, whether cellular metabolism is changed during HHcy-induced B cell activation is unclear, and if it is changed, the underlying mechanism is unknown.

In this study, we demonstrate that HHcy induces B cell proliferation and Ab secretion both in vivo and in vitro. PKM2 expression and enzyme activity were increased in HHcy-induced B cells to promote metabolic reprogramming, with an increase in both oxidative phosphorylation and glycolysis. The inhibition of PKM2 effectively reversed HHcy-induced B cell proliferation, Ab secretion, and the early stage of atherogenesis in ApoE^−/− mice. Therefore, our results suggest that PKM2 is a critical metabolic regulator of HHcy-induced B cell activation and may serve as a potential therapeutic target in treating HHcy-related atherosclerosis and B cell–associated inflammatory diseases.

Six-week-old C57BL/6J mice and ApoE^−/− mice were purchased from the Animal Center of Peking University Health Science Center (Beijing, China) and were maintained under specific pathogen-free conditions. For the induction of HHcy in mice, C57BL/6J mice were fed a normal mouse chow diet and were provided drinking water supplemented with or without 1.8 g/l dl-Hcy (Sigma-Aldrich, St. Louis, MO) for 2 wk as previously described (5). In the treatment study, ApoE^−/− mice were i.p. injected with 1.2 mg/kg shikonin (SKN; Sigma-Aldrich) or solvent control every 3 d. Three days after the first injection, ApoE^−/− mice were fed a normal mouse chow diet and provided water supplemented with or without 1.8 g/l Hcy for 2 wk. All studies were performed according to protocols approved by the Committee on the Ethics of Animal Experiments of Peking University Health Science Center. The investigation conformed to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health.

Splenic B cells were isolated from mice and purified via positive selection using magnetic microbeads against CD19 (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s protocol. Purified B cells were cultured in RPMI 1640 medium (Life Technologies, Gaithersburg, MD) supplemented with 10% FBS (Gemini Bio-Products, West Sacramento, CA) and maintained with 0.1 μg/ml LPS (Sigma-Aldrich). B cells were incubated with or without 100 μM Hcy or 5 μg/ml LPS for the indicated times. For some experiments, B cells were pretreated with 0.5 mM 2-deoxy-d-glucose (2-DG; Sigma-Aldrich), 0.25 μM SKN, or 10 nM rapamycin (Sigma-Aldrich) for 30 min before Hcy stimulation.

Scramble small interfering RNA (siRNA) and siRNA targeting Pkm2 (5′-CTTGCAGCTATTCGAGGAA-3′) were designed by and purchased from RiboBio (Guangzhou, China). Scramble and PKM2 siRNA were transfected into B cells using an Amaxa mouse B cell Nucleofector kit (Lonza Cologne, Cologne, Germany) according to a modified version of the manufacturer’s protocol.

Cell extracts containing equal amounts of total protein were resolved via 10% SDS-PAGE and were then transferred to a nitrocellulose membrane. After blocking with 5% BSA for 1 h, the membrane was incubated with different primary Abs at 4°C overnight, including anti-PKM2, anti-PKM1, anti–IFN regulatory factor 4 (IRF4), anti–phospho-mechanistic target of rapamycin (mTOR; Ser²⁴⁴⁸), anti-mTOR, anti–phospho-AKT (Ser⁴⁷³), anti-AKT, anti–phospho-S6 ribosomal protein (S6RP; Ser^235/236), anti–β-actin, anti-GAPDH (1:1000; all from Cell Signaling Technology, Danvers, MA), and anti–eukaryotic translation initiation factor 5 (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA). The membrane was then incubated with IRDye 700– or IRDye 800–conjugated secondary Abs (1:20,000; Rockland, Gilbertsville, PA) for 1 h at room temperature. The fluorescence signal was detected and analyzed using an Odyssey infrared imaging system (LI-COR Biosciences, Lincoln, NE).

Total RNA was extracted using TRIzol reagent (Life Technologies, Grand Island, NY) and was then reverse transcribed to cDNA and amplified using the AMV reverse transcription system (Promega, Madison, WI). For quantitative PCR, all amplification reactions involved the use of the Mx3000 multiplex quantitative PCR system (Stratagene, La Jolla, CA) and SYBR Green I reagent with results normalized to β-actin mRNA levels. Detection of postswitch transcripts was performed as described in a previous study (21). The primer sequences used for PCR analyses based on mouse genes are as follows: Aicda (forward, 5′-AAGGGACGGCATGAGACCTA-3′; reverse, 5′-GCCGAAGTTGTCTGGTTAGC-3′), Iμ-Cγ3 (forward, 5′-CTCTGGCCCTGCTTATTGTTG-3′; reverse, 5′-CTCAGGGAAGTAGCCTTTGACA-3′), Iμ-Cγ1 (forward, 5′-CTCTGGCCCTGCTTATTGTTG-3′; reverse, 5′-GGATCCAGAGTTCCAGGTCACT-3′), Iμ-Cγ2b (forward, 5′-CTCTGGCCCTGCTTATTGTTG-3′; reverse, 5′-CACTGAGCTGCTCATAGTGTAGAGTC-3′), Pkm (forward, 5′-TCGAGAACCATGAAGGCGTC-3′; reverse, 5′-ACTTGGTGAGCACTCCTG-3′), Pkm2 (forward, 5′-TCGAGAACCATGAAGGCGTC-3′; reverse, 5′-CGGCGGAGTTCCTCGAATAG-3′), Icam-1 (forward, 5′-AGCTCGGAGGATCACAAA-3′; reverse, 5′-TCTGCTGAGACCCCTCTTG-3′), Vcam-1 (forward, 5′-CTGTTCCAGCGAGGGTCTA-3′; reverse, 5′-CACAGCCAATAGCAGCACA-3′), Tnf-α (forward, 5′-ACAGAAAGCATGATCCGCGAC-3′; reverse, 5′-CCGATCACCCCGAAGTTCAGTA-3′), Ifn-γ (forward, 5′-TGGCTGTTTCTGGCTGTTAC-3′; reverse, 5′-TTCGCCTTGCTGTTGCTGAAG-3′), Mcp-1 (forward, 5′-CAGATGCAGTTAACGCCC-3′; reverse, 5′-ATTCCTTCTTGGGGTCAGC-3′), Il-2 (forward, 5′-CAGGAACCTGAAACTCCCCA-3′; reverse, 5′-AGAAAGTCCACCACAGTTGC-3′), β-actin (forward, 5′-GTGACGTTGACATCCGTAAAGA-3′; reverse, 5′-GCCGGACTCATCGTACTCC-3′). The Pkm splicing experiment was performed based on modified methods from a previous report (22). PCR products were digested with or without NcoI and PstI (both from Takara Bio, Shiga, Japan) to distinguish Pkm1 and Pkm2 isoforms. The products were then resolved on an agarose gel and analyzed using the GeneGenius bio-imaging system (Syngene, Nuffield Road, Cambridge, U.K.).

For the analysis of cell culture supernatants, B cells (1 × 10⁶ cells per well) were seeded in a 48-well plate for the indicated times and supernatants from the cell cultures were harvested. The IgM and IgG levels of the culture supernatants (1:10 for IgM and 1:1 for IgG) or blood plasma (1:5,000 for IgM and 1:50,000 for IgG) were analyzed and quantified using mouse-specific IgM and IgG ELISA kits (Bethyl Laboratories, Montgomery, TX) according to the manufacturer’s protocol.

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were analyzed using an XF24 extracellular flux analyzer (Seahorse Bioscience, Billerica, MA). B cells (1 × 10⁶ cells per well) were seeded in an XF24 microplate coated with polylysine (Sigma-Aldrich). Mitochondrial stress tests were performed under basal conditions and in response to metabolic reagents, including 1 μM oligomycin, 1 μM carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), 1 μM rotenone, and 1 μM antimycin A (all from Sigma-Aldrich). Glycolysis stress tests were performed under basal conditions and in response to 10 mM glucose, 1 μM oligomycin, and 100 mM 2-DG (all from Sigma-Aldrich). Wave software and the XF Mito/Glycolysis stress test report generator (Seahorse Bioscience) were used to analyze the OCR in mitochondrial stress tests and the ECAR in glycolysis stress tests. Additionally, for the analysis of ECAR in mitochondrial stress tests, basal and maximal ECARs were calculated as previously described (23). Briefly, basal ECAR was the basal rate before the addition of metabolic reagents, and maximal ECAR was assessed after the addition of rotenone and antimycin A. Glycolytic reserve was calculated by subtracting basal ECAR from maximal ECAR.

For metabolite extraction, cultured cells were washed with PBS twice and lysed in 80% aqueous methanol (v/v) equilibrated at −80°C. [5-¹³C]glutamine (Cambridge Isotope Laboratories, Tewksbury, MA) was added as an internal standard. Cell supernatants of metabolite extracts were collected, dried, and stored at −80°C until used for metabolomic analyses. For liquid chromatography–tandem mass spectrometry (LC-MS/MS), samples were reconstituted in water and analyzed using a QTRAP 6500 LC-MS/MS System (AB SCIEX, Concord, ON, Canada). MultiQuant v3.0 software (AB SCIEX) was used to process all raw liquid chromatography–mass spectrometry data and integrate chromatographic peaks. Integrated peak areas corresponding to metabolite concentrations were further analyzed using MetaboAnalyst software (24). Metabolite abundance was expressed relative to the internal standard and was normalized to the number of cells.

Pyruvate kinase activity was measured using a lactate dehydrogenase (LDH)-linked assay as previously described (25). B cells were lysed in a pyruvate kinase lysis buffer (50 mM Tris-HCl [pH 7.5], 1 mM EDTA, 150 mM NaCl, 1% IGEPAL CA-630) supplemented with protease inhibitors and PMSF. Protein concentrations of the cell lysates were determined, and 2 μg of total protein was mixed with the reaction mixture (50 mM Tris-HCl [pH 7.5], 100 mM KCl, 5 mM MgCl₂, 0.6 mM ADP, 0.5 mM phosphoenolpyruvate, 450 μM NADH, and 8 U LDH). The final reaction volume was 200 μl, and reactions were performed in 96-well plates. The decrease in absorbance by NADH at 340 nm was monitored and measured to determine pyruvate kinase activity using a Varioskan flash multimode reader (Thermo Fisher Scientific, Waltham, MA).

To assess glucose uptake, B cells were cultured and treated as indicated, and 2-NBDG (Life Technologies) was then added into the cell culture medium, and cultures were incubated for 1 h. For CD138 staining, cells were collected and stained with a PE-conjugated CD138 Ab (BD Pharmingen, San Diego, CA). Plasma inflammatory cytokine levels were measured using a cytometric bead array mouse inflammation kit (BD Biosciences, San Jose, CA). For cell proliferation, B cells were stained with CFSE (Molecular Probes, Eugene, OR) prior to culture and analyzed at certain times. Flow cytometry was performed using a FACSCalibur flow cytometer (BD Biosciences) and the data were analyzed using FlowJo software (Tree Star, Ashland, OR). Cells labeled with 2-NBDG were also analyzed using a FlowSight imaging flow cytometer (Amnis, Seattle, WA). Bright-field and fluorescent images were acquired and data analyses were completed using IDEAS software (Amnis).

Plasma Hcy levels were quantified via an enzymatic cycling assay as previously described (26). Total plasma triglycerides and cholesterol levels were analyzed using kits from BioSino Biotechnology & Science (Beijing, China) according to the manufacturer’s protocols.

Hearts were collected and prepared as previously described (26). Briefly, hearts were collected and embedded in OCT compound (Tissue-Tek; Sakura Finetek, Torrance, CA) and frozen. Aortic roots were serially cross-sectioned at 7-μm intervals, collected on glass slides, and then stained with Oil Red O to visualize lesion areas.

All data are presented as the mean ± SEM unless otherwise stated. Comparisons of data sets were performed using unpaired Student t tests for comparing two groups and one-way ANOVAs followed by Newman–Keuls post hoc tests for comparing multiple groups. Statistical analyses were performed using GraphPad Prism (GraphPad Software, La Jolla, CA). A p value <0.05 was considered statistically significant for all experiments.

We have previously reported that B cells from ApoE^−/− mice with HHcy show an increase in both cell proliferation and IgG secretion in response to LPS stimulation ex vivo compared with those from control ApoE^−/− mice (3, 4). To investigate whether HHcy could regulate normal B cell function in vivo, we fed C57BL/6J mice a normal chow diet and provided them drinking water supplemented with or without 1.8 g/l Hcy for 2 wk. The results show that mice given Hcy-supplemented water developed moderate HHcy compared with control mice (33.58 ± 1.28 versus 8.14 ± 0.42 μM; p < 0.05). In the treated mice, HHcy significantly increased B cell numbers in the spleen (Fig. 1A). Both plasma IgM (from 0.22 ± 0.01 to 0.41 ± 0.06 mg/ml) and IgG (from 0.95 ± 0.04 to 2.21 ± 0.40 mg/ml) levels were upregulated in HHcy mice (Fig. 1B, 1C). These data indicate that HHcy activates B cells in vivo.

The terminal differentiation of B cells into Ab-secreting plasma cells depends on the expression of several transcription factors (27), whereas the activation of Ig class switch recombination allows for genetic alterations of Ig genes and the generation of different Ab classes (28). We further tested whether Hcy regulated these processes during B cell activation. Protein expression of IRF4, which is the key regulator of B cell differentiation, was increased by ∼70% in HHcy-treated B cells (Fig. 1D). Quantitative PCR analysis also showed that HHcy-treated B cells had a significantly elevated mRNA expression of Aicda (Fig. 1E, left), which is critical for the initiation of class switch recombination (21). Moreover, the expression of postswitch transcripts, including Iμ-Cγ2b and Iμ-Cγ1, was increased in HHcy-treated B cells, whereas Iμ-Cγ3 had a slight tendency for increased expression (Fig. 1E, right). Thus, HHcy activated B cells through the regulation of terminal differentiation-related transcription factors and class switch recombination in vivo.

Our in vitro studies further confirmed the results of the in vivo experiments. Consistent with our previous findings, B cells proliferated at 48 h (Fig. 1F) and secreted IgM and IgG Abs at 72 h (Fig. 1G, 1H) when treated with Hcy (100 μM). Intriguingly, Hcy markedly enhanced B cell differentiation into CD138⁺ cells (from 2.37 ± 0.09% in control cells to 6.94 ± 0.39% in Hcy treatment cells; Fig. 1I), which was accompanied by increased protein expression of IRF4 (Fig. 1J). Additionally, Hcy significantly upregulated the mRNA levels of Aicda, Iμ-Cγ2b, and Iμ-Cγ3 in B cells, whereas Iμ-Cγ1 showed a tendency for increased expression (Fig. 1K). Collectively, these results show that HHcy can induce B cell proliferation and Ab secretion both in vivo and in vitro.

Upon activation, lymphocytes proliferate, differentiate, and exert effector functions. Activated lymphocytes need to alter their metabolic program to produce enough energy and biosynthetic substrates to meet the demands of cell growth and function (29). To identify whether metabolic reprogramming is required in HHcy-activated B cells, we first measured the metabolic parameters of B cells purified from control and HHcy mice using an extracellular flux analyzer. Analysis revealed that OCR and ECAR were significantly increased in HHcy-treated B cells (Fig. 2A, 2B, upper panels). The increase in basal OCR (from 124.2 ± 5.4 pmol/min in control cells to 181.2 ± 11.7 pmol/min in HHcy-treated cells) and ECAR (from 3.1 ± 0.4 mpH/min in control cells to 6.1 ± 0.3 mpH/min in HHcy-treated cells) revealed the activation of both glycolytic and oxidative metabolism in HHcy-stimulated B cells, whereas the increases in maximal OCR (from 208.2 ± 3.4 pmol/min in control cells to 253.4 ± 26.9 pmol/min in HHcy-treated cells) and ECAR (from 5.5 ± 0.4 mpH/min in control cells to 10.0 ± 0.5 mpH/min in HHcy-treated cells) reflected a higher maximal metabolic potential in HHcy-treated B cells (Fig. 2A, 2B, lower panels). Spare respiratory capacity (SRC) and glycolytic reserve are measurements that reflect the metabolic capacity in oxidative and glycolytic metabolism, respectively. HHcy-treated cells had similar spare respiratory capacity but significant higher glycolytic reserve (Fig. 2A, 2B, lower panels), demonstrating an elevated metabolic capacity in glycolytic metabolism rather than oxidative metabolism. In comparing the ratio of OCR to ECAR, an indicator for metabolic switching, we found that HHcy-activated B cells depended more on glycolytic metabolism than on oxidative metabolism, as evidenced by a 35% reduction in the OCR/ECAR ratio in HHcy-treated cells compared with that of the control cells (Fig. 2C, lower panel). These results indicate dramatic metabolic reprogramming in HHcy-treated B cells.

Similar results were obtained when we performed in vitro experiments. Both OCR (141.6 ± 0.4 versus 91.4 ± 8.0 pmol/min for basal OCR, 274.7 ± 6.5 versus 139.7 ± 16.1 pmol/min for maximal OCR) and ECAR (2.8 ± 0.1 versus 1.0 ± 0.4 mpH/min for glycolysis, 3.6 ± 0.2 versus 0.9 ± 0.3 mpH/min for glycolytic capacity) were significantly increased in B cells treated with Hcy (100 μM) for 24 h relative to the control cells (Fig. 2D, 2E). Hcy-stimulated B cells showed higher metabolic capacity in oxidative and glycolytic metabolism, with both increased SRC and glycolytic reserve (Fig. 2D, 2E). The OCR/ECAR ratio tended to be lower in Hcy-treated B cells, although the difference was not statistically significant (29.6 ± 3.9 versus 41.4 ± 7.8; Fig. 2F, lower panel). As glucose is a major source for both glycolytic and oxidative metabolism, we next assessed glucose uptake using a fluorescent glucose analog (2-NBDG). Hcy increased glucose uptake in B cells, as evidenced by fluorescent images and quantification of fluorescence intensity from flow cytometry analysis (Fig. 2G, 2H). Taken together, these results suggest that both glycolysis and oxidative phosphorylation are increased in Hcy-treated B cells in vivo and in vitro, and that these cells show a tendency for switching to glycolysis.

Furthermore, we incubated splenic B cells from control and HHcy mice with LPS for an additional 24 h to test the metabolic changes caused by stress ex vivo. HHcy-treated B cells displayed a higher OCR (174.6 ± 7.1 versus 139.3 ± 8.3 pmol/min for basal OCR, 304.8 ± 19.5 versus 220.8 ± 19.1 pmol/min for maximal OCR, 130.2 ± 16.0 versus 81.5 ± 13.1 pmol/min for SRC) and ECAR (8.7 ± 0.8 versus 6.2 ± 0.4 mpH/min for basal ECAR, 16.09 ± 2.3 versus 11.2 ± 0.8 mpH/min for max ECAR, 9.4 ± 1.2 versus 5.0 ± 0.5 mpH/min for glycolytic reserve) than did control B cells (Supplemental Fig. 1A, 1B). The OCR/ECAR ratio was not changed in HHcy-treated B cells in response to LPS (Supplemental Fig. 1C). Flow cytometric analysis showed that glucose uptake was increased by ∼30% in HHcy-treated B cells (Supplemental Fig. 1D). Thus, HHcy can increase metabolic potential in B cells under both basal and stress conditions.

The extracellular flux analysis identified metabolic changes in Hcy-activated B cells. To further explore the detailed changes in metabolic pathways and the potential underlying mechanisms involved, metabolites were extracted from B cells stimulated with or without Hcy for 24 h and analyzed. All samples were profiled using LC-MS/MS. The heat map generated from hierarchical clustering and a principal component analysis of metabolites revealed a distinct metabolic profile in B cells upon Hcy stimulation relative to the control group (Fig. 3A, 3B). The heat map also showed an enrichment of metabolites in Hcy-treated B cells. Further analysis demonstrated that Hcy-induced B cells had significantly higher levels of several glycolytic intermediates (Fig. 3C), including GADP and 3-phosphoglycerate (3PG). The levels of many metabolites in the TCA cycle (aconitate, isocitrate, α-ketoglutarate, and succinyl-CoA) were enriched in B cells following Hcy stimulation (Fig. 3C), which is in line with our observations of metabolic changes in the extracellular flux analysis. Moreover, the pentose phosphate pathway, a shunt from the glycolytic pathway, was activated, as determined by the accumulation of xylulose 5-phosphate, sedoheptulose 7-phosphate, and NADPH (Fig. 3C). The pentose phosphate pathway, a well-known biosynthetic pathway, is assumed to provide biosynthetic substrates to support B cell growth and activation (12, 29). Thus, our metabolomic profiling analysis not only confirms a metabolic upregulation in both glycolytic and oxidative metabolism, but it also suggests the activation of the biosynthesis-related pentose phosphate pathway in Hcy-treated B cells.

As revealed by the metabolic flux analysis and metabolomic profiles, glucose metabolism is activated in Hcy-stimulated B cells and may be critical for Hcy-induced B cell activation. To investigate the causal relationship between glucose metabolism and Hcy-induced B cell activation, we used 2-DG to inhibit the first step of glucose metabolism in the glycolytic pathway. As expected, 2-DG (0.5 mM) sharply suppressed Hcy-induced B cell proliferation and differentiation to plasma cells (Fig. 4A, 4B). Moreover, 2-DG (0.5–5 mM) inhibited the IgM and IgG secretion promoted by Hcy in a dose-dependent manner (Fig. 4C, 4D). Therefore, glucose metabolism is critical for Hcy-induced B cell proliferation and Ab secretion.

The glycolytic pathway, which has many enzyme-catalyzed reactions, is the initial and major pathway in metabolizing glucose. Pyruvate kinase is a rate-limiting glycolytic enzyme and controls the last step of this pathway (13). Although PKM2 has been reported to be mainly expressed in tumor cells and embryonic cells (13, 14), recent reports showed that PKM2 is the main pyruvate kinase isoform expressed in the mouse spleen (20, 22). To test whether B cells, one of the major cell types in the spleen, express PKM2, we first measured the mRNA and protein levels of the pyruvate kinase isoforms expressed in B cells. The results showed that B cells expressed mainly PKM2 but not PKM1, as detected by Western blotting (Fig. 5A), and by RT-PCR with amplicons digested with restriction enzymes to distinguish Pkm1 and Pkm2 (Fig. 5B). We then sought to explore the role of PKM2 in B cells activated by HHcy. PKM2 protein expression was upregulated in B cells from HHcy mice by 50% relative to control B cells (Fig. 5C). Moreover, the glycolytic enzyme activity of PKM2 was also increased in HHcy-treated B cells (Fig. 5D), suggesting a possible role of PKM2 in regulating metabolic changes in B cells. In agreement with this, our in vitro experiments revealed that Hcy stimulation for 24 h directly upregulated PKM2 protein expression (Fig. 5E) as well as enzyme activity (Fig. 5F) in B cells, although gene expression of Pkm2 remained unchanged (Fig. 5G). Therefore, PKM2 protein expression and enzyme activity in B cells are positively regulated by HHcy both in vivo and in vitro.

The Akt-mTOR signaling pathway is a critical pathway that influences cellular metabolism in immune cells through regulating gene transcription and the posttranslational modification of glycolytic enzymes (6). Therefore, we determined whether PKM2 was modulated by the Akt-mTOR pathway in B cells upon Hcy stimulation. Our results show that Hcy activated the Akt-mTOR pathway by increasing the phosphorylation of Akt (Fig. 6A) and mTOR (Fig. 6B) in B cells. We then used the mTOR inhibitor rapamycin to explore the role of the mTOR pathway in the regulation of PKM2. Treatment with a low dose of rapamycin (10 nM) strongly suppressed phosphorylation of mTOR and its downstream target S6RP in Hcy-treated B cells (Fig. 6C). Although the protein expression of PKM2 was not affected (Fig. 6D), the PKM2 enzyme activity was greatly diminished by rapamycin in Hcy-induced B cells (Fig. 6E). PKM2, therefore, can be one of the potential targets regulated by the mTOR pathway. Because previous reports have established that B cell growth and proliferation require the activation of the Akt-mTOR pathway (12), we next determine whether the Akt-mTOR pathway is necessary for Hcy-induced B cell activation. Our results showed that Hcy-induced B cell proliferation was partially repressed by rapamycin (Fig. 6F). Furthermore, rapamycin also reversed Hcy-induced IgM and IgG secretion (Fig. 6G, 6H). Collectively, these results point to a critical role for the Akt-mTOR pathway in regulating Hcy-induced B cell activation. That PKM2 could be regulated by the mTOR pathway indicates that PKM2 might serve as a possible regulator in Hcy-induced B cell activation.

Pyruvate kinase is a key enzyme in the glycolytic pathway that converts phosphoenolpyruvate and ADP to pyruvate and ATP (Fig. 7A). To verify the potential role of PKM2 in regulating Hcy-induced metabolic changes and B cell activation, the potent PKM2 inhibitor SKN was used. SKN can selectively inhibit the glycolytic enzyme activity of PKM2 without affecting other pyruvate isoforms and may limit the subsequent lactate production and oxidative metabolism in mitochondria (30). We first performed a dose titration of SKN to determine the optimal dose for our experiments. Although lower doses of SKN (0.1, 0.15, 0.2 μM) only slightly impaired PKM2 activity in Hcy-treated B cells, 0.25 μM SKN showed a significant inhibitory effect. Although a higher dose of SKN (0.5 μM) also effectively reduced PKM2 activity, the enzyme activity was lower than that of the control group (Fig. 7B). To determine the contribution of Hcy-induced PKM2 enzyme activity, a dose of 0.25 μM SKN was used in our further experiments. Although Hcy-enhanced PKM2 enzyme activity was effectively reduced, SKN did not affect PKM2 protein expression (Fig. 7B, 7C). Next, we measured the metabolic state in B cells in response to PKM2 inhibition. Notably, both OCR and ECAR, which were upregulated by Hcy, were significantly reduced by SKN (49.9% for basal OCR, 50.2% for maximal OCR, and 50.4% for SRC, 26.9% for glycolysis, and 25.0% for glycolytic capacity; Fig. 7D, 7E), suggesting a critical role of PKM2 in Hcy-induced glycolytic and oxidative metabolism in B cells. Through the inhibition of metabolic changes, SKN eventually reversed Hcy-induced B cell proliferation (Fig. 7F), differentiation to plasma cells (Fig. 7G), and Ab production (Fig. 7H, 7I). To further confirm the role of PKM2 in Hcy-induced B cell function, we knocked down PKM2 in B cells using siRNA. Similar to SKN, titration of PKM2 siRNA was performed and 50 nM was chosen as the optimal dose, as this dose of PKM2 siRNA effectively restricted Hcy-induced PKM2 activity without affecting the basal PKM2 activity (Fig. 7J). PKM2 was successfully knocked down as shown by the Western blot analysis (Fig. 7K). Similar results were obtained, as Hcy-induced B cell differentiation and Ab secretion were reduced by PKM2 silencing (Fig. 7L–N). Taken together, these results indicate that PKM2 plays an important role in Hcy-induced B cell activation.

In addition to Hcy stimulation, we also tested the role of PKM2 in B cells upon LPS stimulation to compare similarities and differences between Hcy and this classical stimulus of B cell activation. LPS (5 μg/ml) induced both the mRNA and protein expression of PKM2 in B cells, whereas Hcy (100 μM) only increased PKM2 protein expression (Supplemental Fig. 2A, 2B). B cells responded to LPS with increased PKM2 enzyme activity, the induced level of which was relatively lower than that of the Hcy-induced one (Supplemental Fig. 2C). SKN (0.25 μM) effectively diminished both Hcy- or LPS-induced PKM2 activity without affecting PKM2 protein expression (Supplemental Fig. 2B, 2C). Note that the same dose of SKN reduced LPS-induced PKM2 activity to a much lower level compared with the Hcy plus SKN group. We then measured oxidative and glycolytic metabolism in B cells, and similar changing patterns as those in PKM2 activity were seen in both OCR and ECAR analysis (Supplemental Fig. 2D, 2E). Similar results were also obtained when we examined B cell activation, as LPS induced B cell proliferation, differentiation to plasma cells, and Ab secretion, the induced levels of which were relatively lower compared with Hcy group, whereas most of the LPS-induced effects were also reduced by SKN, to a greater extent than Hcy-induced effects (Supplemental Fig. 2F–I). Collectively, these data indicate a similar role of PKM2-dependent metabolic changes in B cells responded to LPS and Hcy. Although the underlying mechanisms regulating PKM2 expression may differ, these results further confirm a critical role of PKM2 in B cell activation.

We have previously reported that HHcy accelerates early atherosclerotic lesions in ApoE^−/− mice (5). Recent studies have also reported that B cells can be tightly associated with the development of atherosclerosis (31). Given the capability of the PKM2 inhibitor SKN to prevent Hcy-induced B cell proliferation and Ab production in vitro, we examined whether SKN could affect B cell metabolism and function in vivo and further tested the effects of PKM2 inhibition in HHcy-accelerated atherosclerosis in ApoE^−/− mice. ApoE^−/− mice were i.p. injected with 1.2 mg/kg SKN or solvent control every 3 d. Three days after the first SKN injection, ApoE^−/− mice were fed a normal mouse chow diet and given drinking water supplemented with or without 1.8 g/l Hcy for 2 wk (Fig. 8A). Two weeks later, a significant increase in plasma Hcy level was detected both in HHcy mice and HHcy plus SKN mice (Supplemental Fig. 3A), indicating that SKN treatment does not directly affect plasma Hcy levels. Additionally, no differences were found in body weight or plasma total triglycerides and total cholesterol levels among the four groups (Supplemental Fig. 3B, 3C).

However, treatment with SKN greatly restored the number of B cells in the spleen, which was increased by HHcy in ApoE^−/− mice (Fig. 8B). Notably, whereas plasma IgM levels were only slightly reduced by SKN (Fig. 8C), SKN treatment significantly repressed plasma IgG production in HHcy plus SKN mice relative to production in HHcy mice (Fig. 8D). These results indicate that SKN inhibits HHcy-induced B cell function in vivo. Additionally, B cells from ApoE^−/− mice were collected to perform metabolic-related experiments. Consistent with our in vitro results, PKM2 enzyme activity was enhanced in HHcy-treated B cells and was significantly inhibited by SKN treatment (Fig. 8E). Extracellular flux analysis revealed that the Hcy-induced elevations in both OCR and ECAR were reversed in HHcy plus SKN mice (Fig. 8F, 8G). Furthermore, SKN treatment in vivo ameliorated HHcy-accelerated atherosclerotic lesion formation in aortic roots (Fig. 8H). Gene expression of local inflammatory cytokines in isolated aortas, including Icam-1, Tnf-α, Ifn-γ, and Il-2, was significantly increased in HHcy ApoE^−/− mice, and this increase was reversed by SKN treatment (Fig. 8I). Plasma inflammatory cytokines were also examined via a cytometric bead array analysis, and no significant changes were found among the groups (Supplemental Table I). Taken together, these results provide evidence that SKN can directly diminish HHcy-activated B cell function in vivo via the inhibition of PKM2 enzyme activity and metabolic changes. Moreover, SKN treatment can ameliorate HHcy-accelerated atherosclerosis.

Accumulating evidence demonstrates that HHcy is an independent risk factor for cardiovascular diseases (1). As described in our previous studies, Hcy stimulation in vitro and ex vivo can lead to B cell proliferation and IgG secretion (3, 4). In this study, we further verified the effects of HHcy on B cells in vivo and to our knowledge performed the first study to explore the metabolic reprogramming and potential mechanisms involved in this process.

The Warburg effect, or aerobic glycolysis, which is characterized by increased lactate production rather than oxidative phosphorylation, promotes the accumulation of intermediate metabolites for biosynthetic pathways responsible for cell growth and function (32). Because previous reports have demonstrated that aerobic glycolysis is critical for the proliferation and growth of tumor cells, numerous studies have been carried out to identify whether aerobic glycolysis is essential for innate and adaptive immunity. Surprisingly, activated dendritic cells, M1 macrophages, and effector T cells share some of the same properties as tumor cells, switching their metabolic program from oxidative phosphorylation to aerobic glycolysis upon activation, whereas resting dendritic cells, M2 macrophages, and naive and memory T cells maintain oxidative phosphorylation for energy generation (6, 8, 9). A series of studies have revealed that B cells can be either proatherogenic or atheroprotective, which may be dependent on the specific context these cells reside in and on different subsets of B cells (31). However, B cell metabolism is much less studied.

Until recently, two reports have suggested that BCR- or LPS-activated B cells and IgA⁺ plasma cells adopt both glycolysis and oxidative phosphorylation to maintain cell growth and function (10, 11). Consistent with these findings, our present study shows that Hcy stimulation causes an increase in both glycolytic and oxidative metabolic rates in B cells. Interestingly, our in vivo study reveals that HHcy-treated B cells, an in vivo model of chronic Hcy stimulation, preferentially use glycolysis over oxidative phosphorylation, whereas acute (24 h) Hcy stimulation in vitro leads to a slight decrease in the OCR/ECAR ratio without significant changes, reflecting a more balanced use of glycolytic and oxidative metabolism in response to an acute stimulation. Moreover, both OCR and ECAR are more dramatically upregulated in HHcy-treated B cells after 24 h of LPS stimulation relative to the control cells, implying a higher metabolic potential in HHcy-treated B cells under metabolic stress.

Our metabolomic profiling analysis further confirms the increased metabolic flux in both glycolysis and the TCA cycle. Although the extracellular lactate production detected using the extracellular flux analyzer was significantly increased upon Hcy stimulation, the intracellular level of lactate was not significantly altered. We speculate that Hcy-activated B cells may swiftly export lactate produced by glycolysis to maintain intracellular pH homeostasis. Moreover, the pentose phosphate pathway is activated in B cells treated with Hcy, as evidenced by the accumulation of the immediate metabolites in this pathway. The activation of all these pathways generates substantial energy and substrate supplies for biosynthesis to support Hcy-induced B cell growth, proliferation, and Ab secretion.

Glucose metabolism is critical for B cell growth and activation (12). There are many glycolytic enzymes in the glycolytic pathway; however, whether these enzymes regulate B cell function remains largely unknown. Pyruvate kinase is a rate-limiting enzyme that converts phosphoenolpyruvate and ADP to pyruvate and ATP. Previous studies of PKM2 mainly focused on tumor cells and have established an important role of PKM2 in regulating tumor cell growth through aerobic glycolysis (14). Recent reports have indicated that PKM2 is critical for modulating macrophage activation and function (18, 19), revealing a potential role of PKM2 in immune cells. In this study, we focused on B cells and provide evidence that PKM2 is essential for B cell proliferation and Ab secretion both in vivo and in vitro. B cells activated by Hcy increase their expression and enzyme activity of PKM2, as well as their glucose uptake. Our findings that both extracellular lactate production and oxygen consumption are markedly increased in B cells treated with Hcy indicate that PKM2, a common upstream glycolytic enzyme of lactate production and the TCA cycle, may be responsible for the metabolic changes in B cells.

SKN, a potent PKM2 inhibitor, which has previously been reported to limit glycolysis in tumor cells and protect mice from sepsis through the inhibition of PKM2 in macrophages (18, 30), was used in this study. Our in vitro and in vivo experiments demonstrate that SKN inhibits the Hcy-induced upregulation of PKM2 enzyme activity and metabolic reprogramming, leading to the prevention of Hcy-induced B cell proliferation and Ab secretion. Although SKN shows potent inhibitory effects on PKM2, SKN also possesses additional pharmacological effects (33). As an additional approach, we used siRNA to knock down PKM2 and found that Hcy-induced B cell functions are effectively inhibited, further demonstrating a critical role of PKM2 in B cell activation. The enhanced glycolytic function of PKM2 facilitates more glucose conversion into pyruvate and the subsequent activation of lactate production and mitochondrial oxidative phosphorylation. This altered metabolic phenotype allows B cells to quickly accumulate metabolites and to produce ATP simultaneously for B cell growth and function. However, in addition to acting as a traditional glycolytic enzyme, PKM2 also might have alternative functions independent of enzyme activity (34). Whether PKM2 exerts nonglycolytic functions in B cells and whether these functions are involved in B cell activation remain to be further investigated. In addition to Hcy-activated B cells, B cells that responded to LPS stimulation were also studied. Our data reveal a similar role of LPS in regulating PKM2-dependent metabolic reprogramming and B cell activation. However, differences were seen in the underlying mechanisms regulating PKM2 expression. LPS may increase PKM2 protein expression directly through gene transcription as evidenced by the increased mRNA level of Pkm2, whereas Hcy, which does not affect Pkm2 gene expression, may likely regulate this process through posttranslational modification.

The Akt-mTOR signaling pathway is a key regulator of glucose metabolism (35). Activation of this pathway can promote glycolysis through altered gene transcription as well as posttranslational modifications (6). It has been reported that the increases in glucose metabolism and cell growth in B cells activated by anti-IgM can be effectively reversed by the inhibition of this pathway (12). In agreement with previous findings, we demonstrate in the present study that the Akt-mTOR pathway is critical for Hcy-induced B cell proliferation and Ab secretion. Our findings further reveal that PKM2 could be one of the targets regulated by mTOR pathway, as PKM2 activity was limited by rapamycin treatment. However, whether PKM2 is directly regulated by mTOR, or by the downstream targets of the mTOR pathway, still needs further investigation.

Although our present study mainly focused on glucose metabolism in B cells, lipid metabolism may also be involved. Recent reports have shown that fatty acid oxidation is markedly downregulated (10), whereas glucose-dependent de novo lipogenesis is increased in LPS-activated B cells (36). This increased lipogenesis may support membrane expansion of the endoplasmic reticulum and Golgi network that can influence Ab synthesis and secretion in B cells (36). These observations highlight the importance of lipid and cholesterol homeostasis in lymphocyte activation and function. Our previous study has shown that liver X receptor activation inhibits Hcy-induced IgG secretion in B cells (4). Liver X receptor activation may influence many target genes involved in cholesterol and fatty acid metabolism, indicating a possible role of lipid metabolism in Hcy-induced B cell function. These possibilities should be thoroughly explored in future investigations.

The role of B cells and Igs in atherosclerosis remains controversial. Earlier studies have suggested an atheroprotective role for B cells (37, 38). However, two recent studies have revealed that the development and progression of atherosclerosis are suppressed using a CD20-targeted B cell depletion therapy in ApoE^−/− and low-density lipoprotein receptor–deficient mice, respectively (39, 40). These findings suggest a more complex role for B cells in atherosclerosis, which may be dependent on the specific context the cells reside in and on different subsets of B cells. Activated B cells secrete Abs to maintain humoral immunity. Although IgM is currently suggested to act as a protective role in atherosclerosis, IgG Abs have been suggested to be either proatherogenic or atheroprotective (31). However, there is still a lack of direct evidence for the role of Igs in atherosclerosis.

Despite these controversial findings, our results presented in this study show that HHcy increases the number of splenic B cells, elevates the level of plasma IgG, and increases atherosclerotic lesion formation, which can be all restored by SKN treatment, indicating a tight association between B cells and the development of atherosclerosis in ApoE^−/− mice. Although blood plasma levels of inflammatory cytokines remain unchanged, local aortic inflammatory cytokine levels are upregulated in HHcy ApoE^−/− mice and are restored by SKN treatment. Changes in local aortic inflammatory cytokines, including Tnf-α, Il-2, and Ifn-γ, secreted by macrophages and T cells, which are in line with our previous reports (26), suggest a role of macrophages and T cells in HHcy-accelerated atherosclerosis, and that macrophages and T cells may also be targeted by SKN in vivo. Collectively, these results indicate that many different cell types, including B cells, T cells, and macrophages, might together contribute to the development of HHcy-accelerated atherosclerosis. Further investigations are needed to confirm the exact role of B cells and Igs in this context.

Although it has been extensively established that metabolic reprogramming is indispensable for immune cell activation and effector function (6, 8–10), the direct relevance of cellular metabolism in relationship to atherosclerosis remains largely unknown. Aerobic glycolysis is critical for M1 macrophage activation (9) and has been suggested to have a proatherogenic effect by driving proinflammatory cytokine production (41). However, the overexpression of glucose transporter 1 in macrophages does not aggravate atherosclerosis in low-density lipoprotein receptor–deficient mice (41). This unexpected result has called into question the role of cellular metabolism in atherosclerosis. A recent investigation focusing on hematopoietic stem and multipotential progenitor cells has shown that glucose transporter 1 deficiency prevents hematopoietic stem and multipotential progenitor cell proliferation, myelopoiesis and the recruitment of Ly6C^hi monocytes in atherosclerotic lesions, and eventually reduces atherosclerosis in ApoE^−/− mice (42). Actually, the most recent report on this topic has demonstrated an overexpression of PKM2 in ex vivo–generated macrophages collected from patients with coronary artery disease compared with expression in those from healthy controls and has shown that PKM2 is required for M1 polarization and proinflammatory cytokine secretion from these coronary artery disease macrophages (43). Our present study shows that the PKM2 inhibitor SKN ameliorates HHcy-accelerated atherosclerosis in ApoE^−/− mice. Our observation that the inhibition of PKM2 with SKN reverses HHcy-induced local aortic proinflammatory cytokine secretion by macrophages may further support the important role of PKM2 in macrophages and atherosclerosis. Despite that the systemic effects of SKN in vivo have not been fully defined, our present study and this most recent report consistently highlight an important role of cellular metabolism in the regulation of atherosclerosis and support the notion that limiting the rate of glycolytic and oxidative metabolism by targeting PKM2 can be a possible therapeutic approach to prevent atherosclerosis.

Collectively, our present study reveals an essential role for PKM2-dependent metabolic reprogramming in Hcy-induced B cell activation and function. Targeting PKM2 may provide a possible therapeutic approach for HHcy-related atherosclerosis and B cell–related inflammatory diseases.

The authors have no financial conflicts of interest.