The role of metabolic checkpoint regulators in B cell survival and transformation

In response to mitogenic stimulation, B cells activate different pro‐anabolic signaling pathways such as c‐Myc‐ and mTORC1‐dependent networks to satisfy the energetic demands of biomass synthesis and proliferation. In order to preserve viability and function, cell growth cannot progress unchecked and must be adjusted according to the availability of nutrients. Nutrient‐sensing proteins such as AMPK antagonize mTORC1 activity in response to starvation. If pro‐anabolic signaling pathways are aberrantly activated, B cells may lack the metabolic capacity to accommodate their energetic needs, which can lead to cell death. On the other hand, metabolic hyperactivation is a salient feature of cancer cells, suggesting that mechanisms exist, which allow B cells to cope with metabolic stress. The aim of this review is to discuss how B cells respond to a mismatch between energy supply and demand and what the consequences are of metabolic dysregulation in normal and malignant B cells.


| INTRODUC TI ON
B cells contribute substantially to protection against pathogens.
Their ability to produce pathogen-specific antibodies, to present antigen to T cells, and to regulate other cells through the secretion of cytokines makes them major players of our adaptive immune response. 1,2 On the flip side, B cells are also known to play pathologic roles in a number of diseases including autoimmune disorders 3,4 and various malignancies. 5 Throughout their life, B cells assume different roles and are exposed to changing environments, which is reflected by dynamic adjustments of their gene expression profile and their metabolic signature. 6 To progress in their development, large pre-B cells need to arrest proliferation. 13 Consistent with their different rates of proliferation, pro-B cells, large and small pre-B cells, and immature B cells differ significantly in their metabolic profile. 14,15 To foster cell growth, IL-7 induces activation of signaling molecules such as mechanistic target of rapamycin complex 1 (mTORC1) and c-Myc, which drive an anabolic metabolic program 15,16 indispensable for B cell development ( Figure 1). Both mTORC1 and c-Myc are known metabolic master regulators, which are activated in response to mitogenic signals. mTORC1 phosphorylates downstream targets promoting protein synthesis, oxidative metabolism, and glycolysis. 17 Myc is a transcription factor and drives the expression of genes involved in nutrient uptake, mitochondrial function, glycolysis, and glutaminolysis. 18 While it may not be surprising that c-Myc/mTORC1-driven nutrient uptake and energy generation are essential for the proliferation of B cell precursors, the involvement of these pathways in regulating B cell metabolism prompts the question what effects would unrestrained mTORC1/c-Myc activity have on B cell development.
Would mTORC1-driven hypermetabolism promote B cell expansion and render the cells vulnerable to malignant transformation?
Or would unrestrained metabolic activity lead to an energetic crisis and a metabolic collapse? Recent studies suggest that developing B cells are particularly sensitive to a mismatch between energy consumption and nutrient availability. In normal cells, mTORC1 is inhibited upon starvation to restrict energy-consuming processes such as protein biosynthesis. An important inhibitor of mTORC1 activity in response to reduced energy levels is the enzyme AMP-activated protein kinase (AMPK; Figure 1). The deletion of the AMPK activator liver kinase B1 (LKB1) has been shown to result in a complete block in B cell development 19 (Figure 2). Similarly, mice with a B cell-specific ablation of the AMPK-interacting protein folliculin-interacting protein 1 (Fnip1) lack mature B cells 20 (Figure 2). The signaling components downstream of the pre-BCR are expressed and activated normally in Fnip1-/-pre-B cells and the expression of a rearranged BCR cannot rescue B cell development in Fnip1-deficient mice suggesting that B cell loss is not caused by a missing pre-BCR signal or a failure to induce the recombination of the light chain locus in this mouse model. 20 Rather than playing a role in pre-BCR signaling, Fnip1 seems to be required to inhibit cell growth and metabolic activity when energy and nutrient consumption exceed the availability of the necessary resources. Fnip1-deficient pre-B cells show increased AMPK activation, but fail to efficiently terminate mTORC1 signaling in response to the chemical AMPK activator 5-aminoimidazole-4-carboxyamide-1-beta-D-ribofuranoside (AICAR). 16 In comparison with wildtype cells, Fnip1-/-pre-B cells display enhanced glucose uptake, increased mitochondrial mass and a hypermetabolic phenotype but reduced ATP levels in response to IL-7 stimulation. In vitro, Fnip1-/-pre-B cells are able to progress in their development to IgM positive cells; however, the generation or survival of these cells is strongly dependent on high levels of amino acids. 20 Collectively, these studies suggest that the maintenance of metabolic homeostasis is crucial for B cell development and an inability to limit anabolic processes results in cell death rather than excessive proliferation.

| ME TABOLI C HOMEOS TA S IS IN B CELL-DERIVED LEUKEMIA
The dramatic loss of cell viability in response to aberrantly increased metabolic demands in B cell precursors suggests that metabolic checkpoints are essential for B cell development. Intriguingly, not all B cell subsets require LBK1/AMPK activity for their survival or proliferation. While a disruption of AMPK signaling results in increased apoptosis of developing B cells, mature stimulated B cells are capable of proliferation in the absence of LBK1 or AMPK activity. 21,22 This suggests that these metabolic regulators are not essential for B cell proliferation in general. This prompts the question: What is the evolutionary benefit of such a high sensitivity to metabolic dysregulation in B cell precursors? From the work of Markus Müschen, 23 a compelling model has emerged, in which metabolic checkpoints represent safeguard mechanisms against malignant transformation or autoreactivity. B cell precursors are at particular risk for malignant transformation due to their active DNA rearrangement and high levels of proliferation. Therefore, mechanisms mitigating the risk of malignant transformation are crucial at this stage.
Cancer cells are characterized by rapid proliferation associated with increased metabolic activity. Signaling pathways driving anabolic metabolism are frequently hyperactivated in cancer. 24 Enhanced mTORC1 activity, and consequently increased metabolism, may represent a warning sign for B cells that transforming events have occurred. Several studies suggest that an inability to curtail mTORC1-signaling does not only inhibit B cell development idence that metabolic control of Myc function is disrupted as well. 26 It is tempting to speculate that the inability of Myc transgenic cancer cells to grow in the absence of Fnip1 shows that unrestrained metabolic activity is a liability rather than an advantage for cancer cells. Foxo1 activity is essential for the expression of IL7-receptor and the Rag enzymes, which mediate VDJ recombination 30 ( Figure 1).
Consistently, the deletion of the negative PI3K regulator phosphatase and tensin homolog (PTEN) inhibits normal B cell development. 15,31 However, constitutively active Foxo1 cannot rescue B cell development in PTEN-deficient mice suggesting that PI3K signaling also plays Foxo1-independent roles in B cell precursors. 15 Recent studies demonstrate that Pten deletion compromises BCR-Abl1-or NRAS-driven leukemogenesis 32 (Figure 2). Upon PTEN loss, B cell acute lymphoblastic leukemia (B-ALL) cells show increased glucose consumption and lactate production but decreased ATP levels. 32 This need for PTEN expression in B-ALL cells is surprising considering that PTEN deletion in BCR-Abl1-driven chronic myeloid leukemia (CML) does not inhibit cell survival. 32 Moreover, PTEN is known to play a role as a tumor suppressor in many other types of cancer including B cell lymphoma. 33,34 In summary, these studies highlight the importance of metabolic To maintain adequate redox balance, B cells express PP2A, which redirects carbon from glycolysis toward the pentose phosphate pathway. As a consequence, PP2A limits glycolytic ATP production. The deletion of the PP2A subunit Ppp2r1a has been shown to increase glucose consumption, ATP levels, and lactate production in BCR-ABL1-transformed B-ALL but not CML. On the other hand, overall carbon flow through the PPP is reduced upon Ppp2r1a-deletion in B-ALL cells, the levels of NADPH/NADP are decreased, and production of reactive oxygen species (ROS) is increased demonstrating that PP2A plays an important role in the maintenance of redox balance. PP2A is needed for the survival of both normal and B-ALL cells demonstrating the importance of maintaining PPP activity through this pathway in B cells. 35 In summary, these studies show that B cell precursors need to walk the fine line of supporting proliferation of normal cells and limiting oncogene-driven growth. In addition to mTORC1, glycogen synthase kinase 3 (GSK3) also regulates cell metabolism and plays an important role in B cell maturation and/or maintenance 37 ( Figure 2). GSK3 is a constitutively active kinase that phosphorylates a wide range of proteins. 45 Many of GSK3 substrates are pro-survival or pro-proliferation factors such as Mcl1, CyclinD3, and c-Myc and are targeted for degradation upon GSK3 mediated phosphorylation ( Figure 1). GSK3 has been shown to also inhibit mTORC1 activity 46 (Figure 1), although this pathway does not appear to be strongly modulated by GSK3 in B cells. 37 GSK3 deficiency seems to interfere with normal B cell development or survival. Similar to Cd19cre × Tsc1flox mice, Cd19cre × Gsk3a/bflox mice show a reduction in mature B cell numbers. 37 Of note, GSK3deficient mature B cells display increased cell mass and glucose uptake. If GSK3 is inducibly deleted in mature B cells, these cells slowly accumulate cell mass and exhibit a reduced half-life. 37 In summary, these findings are indicative of metabolic quiescence playing an important role in B cell maturation or the maintenance of the mature B cell pool. However, this view turns out to be too simplistic. Studies exist that demonstrate that B cells are able to cope with increased metabolic activity and to progress in their development despite a metabolically active phenotype in certain situations.

| ME TABOLI C QU IE SCEN CE IN MATURE B CELL S
An example in which enhanced metabolic activity goes hand in hand with increased survival is aberrantly activated signaling of the B cell activating factor (BAFF) receptor in B cells. BAFF is a cytokine secreted by stromal cells and various cell types of myeloid origin. This pro-survival factor is absolutely essential for mature B cell maintenance and also shapes the selection into the mature B cell pool. Autoreactive B cells are thought to be less capable of competing for this limiting resource than normal B cells and as a result are excluded from the mature B cell pool. 47 If BAFF is available in excess, autoreactive B cells are able to persist. Increased BAFF production has been shown to promote autoantibody production and lupus-like autoimmunity in mice. 48 Serum BAFF levels are increased in human patients suffering from different autoimmune disorders. 49 The interaction of BAFF with the BAFF receptor (BAFF-R) triggers the activation of different signaling pathways such as the alternative NFkB pathway, PI3K/Akt signaling, mTORC1 activation, and GSK3 inhibition 50 (Figure 1). In addition to driving a pro-survival program, BAFF-R-mediated Akt activation also supports the basal metabolic needs of mature B cells and confers resistance to atrophy. 51 So how does chronic exposure to BAFF alter B cell metabolism? B cells isolated from BAFF transgenic mice show increased respiratory capacity in the resting state and are poised for rapid induction of glycolysis upon stimulation. 52 Similarly, TNF receptor-associated factor 3 (TRAF3)-deficient naive B cells show a phenotype of increased metabolic activity. 53 TRAF3 is a negative regulator of NIK, an essential kinase in the alternative NFKB pathway 50 ( Figure 1).

Thus, TRAF3 deficiency in B cells mimics chronic BAFF-R signaling
and leads to increased B cell survival. 54 Naive TRAF3-deficient B cells display increased Glut1 and hexokinase II expression as well as glucose uptake and oxygen consumption. 53 Mitochondrial mass or ROS production is however not affected by the loss of TRAF3. 54 In conclusion, excessive BAFF-R signaling achieved through increased BAFF availability or TRAF3 deletion results in increased metabolic activity, but instead of inhibiting B cell maturation boosts B cell survival. Considering that GSK3 inhibition and mTORC1 activation are downstream of BAFF-R-mediated signaling, it would be interesting to investigate why GSK3 and TSC1 deficiency interfere with B cell maturation or survival but chronic exposure to BAFF does not. Does BAFF-R activation or TRAF3 deficiency provide additional pro-survival signals that rescue metabolically stressed cells? Or do B cells from BAFF transgenic mice retain the ability to modulate GSK3 and mTORC1 activity and limit their effects on cell metabolism to a sustainable level? Answering these questions could help to identify metabolic vulnerabilities in autoreactive B cells to be exploited for treatment.

| ME TABOLI C BAL AN CE IN AC TIVATED AND G ERMINAL CENTER B CELL S
When naive B cells encounter antigen, they need to rapidly rewire their metabolic program to foster the proliferative burst that follows activation. Many studies to this date have focused on elucidating which intracellular signaling pathways support this metabolic reprograming. 7,8 Perhaps not surprisingly, many signaling pathways that are initiated during B cell activation simultaneously promote cell cycle progression, survival, and the acquisition of the appropriate metabolic profile. Consequently, the inability to activate these signaling pathways results in a failure to induce anabolic metabolism and thus prevents cell growth and proliferation.
Questions that are less well investigated to this date however are as follows: How is metabolic homeostasis maintained in activated  (Figure 2). In contrast, expression of mutant alleles of the gene RagC that drive partial but not complete uncoupling of mTORC1 activity from amino acid dependence has been shown to boost GC expansion 68 (Figure 2). Nutrient and energy levels are also sensed by the LKB1/AMPK pathway, which impinges on mTORC1 activity. The role of the LKB/AMPK signaling pathway in GC B cells is complex and extends beyond the role of metabolic regulation. LKB1 has been shown to be activated downstream of DNA double-strand breaks and to be needed for plasma cell differentiation. 22 Additionally, LKB1 inhibits NFκB signaling and IL6 secretion. 22 Due to altered cytokine secretion, B cell-specific LKB1 deletion results in T cell hyperactivation and spontaneous GC formation. Interestingly however, the generated GCs are mostly cells that have escaped Lkb1 deletion suggesting that LKB1-sufficient cells outcompete LKB1-deficient cells in the GC. 22 While it is not clear whether LKB1 is also important for GC B cells to respond to metabolic stress, these findings demonstrate that the ability to restrict cell growth in response to stress signaling in general is essential for an appropriate GC response. Moreover, since LKB1 drives plasma cell differentiation, it would be interesting to analyze whether the induction of differentiation is a general mechanism for B cells to escape metabolic stress. Among other targets, LKB1 regulates B cell metabolism by activating AMPK. However, AMPKα1 ablation does not affect B cell proliferation, nutrient uptake, or GC formation suggesting that other LKB1 targets play a more prominent role in GC B cells. 21 Lastly, a typical adaptive response to low oxygen levels is the stabilization of the transcription factor hypoxia-inducible factor 1α (Hif1α) and a metabolic shift toward glycolysis. 69 Consistent with the GC being a hypoxic environment, GC B cells have been found to accumulate Hif1α. 37,62,63 Interestingly, Hif1α hyperstabilization has been shown to be detrimental to GC B cell survival. 62

| MALI G NANT B CELL S AND CELLUL AR S ENE SCEN CE
Similarly to pre-B cells, GC B cells are at a particularly high risk of undergoing malignant transformation. GC B cells manifest phenotypic features such as massive proliferation, genome instability, and resistance to DNA damage that create an inherent risk for malignant outgrowth. 5  B cells navigate the decision between cell death, quiescence, and senescence when exposed to metabolic stress? Considering our current knowledge on the role of metabolic checkpoint regulators in B cells, these questions cannot be unequivocally answered at this time.
As discussed above, AMPKα1 deletion in Myc-driven lymphoma accelerates lymphoma development by boosting mTORC1 activity, biosynthesis, and Hif1α-dependent glycolytic activity. 28 Interestingly, while Hif1α is not essential for the survival of normal lymphoma cells, AMPKα1-deficient cells become addicted to this transcription factor. 28 The deletion of the AMPK-interacting protein Fnip1 in the same mouse model prevents lymphoma genesis. Fnip1deficient B cells have been shown to retain high mTORC1 signaling even when AMPK activity is induced. 20 Similar to AMPK, the loss of LKB1 has been suggested to contribute to lymphoma development.
Mice with a B cell-specific LKB1 deletion (Cd19cre × Lkb1flox mice) spontaneously develop tumors despite impaired survival of mature B cells 77 (Figure 2). In addition to AMPK, RagC/RagA GTPases control mTORC1 activity in response to nutrient deprivation. While RagA mutations are rare, RagC mutations leading to reduced dependence of mTORC1 signaling on amino acids are frequently found in follicular lymphoma cells. 68 Expression of these mutants in a mouse model in which the pro-survival factor Bcl2 is overexpressed accelerates lymphoma development 68 (Figure 2). These findings suggest that partial dysregulation of AMPK/mTORC1 activity benefits lymphoma cells, but should not exceed a certain threshold. The role of LBK1/

AMPK as tumor suppressors in lymphoma cells is in stark contrast
to the previously discussed dependence of leukemic cells on these factors. In order to better understand the role of AMPK/LBK1 in B cell-derived malignancies, it would be important to elucidate how the maturation status or the transforming oncogene influences how metabolic homeostasis is maintained in B cells.
A similarly complex picture emerges when studying the role of GSK3 in B cell lymphoma. As discussed above, GSK3 is a metabolic regulator, needed to maintain cellular quiescence. Upon GSK3 inhibition, numerous pro-survival and pro-proliferation factors are stabilized. In some cell types, GSK3 can act as a tumor suppressor; however, GSK3 has also been implicated to have tumor promoter effects in certain types of cancer. 78 Although GSK3 deletion in mature B cells leads to increased proliferation, these cells are more sensitive to metabolic stress and fail to participate in the GC response. 37 In contrast to GC B cells, a recent study suggests that Myc-driven lymphoma benefits from GSK3 inhibition. 79 Varano et al demonstrate that lymphoma cells lacking surface BCR expression display an altered metabolic profile and are less competitive than BCR-sufficient cells. The authors suggest that BCR-derived signaling attenuates GSK3 activity thereby supporting a beneficial metabolic profile.
Using a pharmacological GSK3 inhibitor, the authors show that GSK3 inhibition can rescue competitiveness of BCR-deficient lymphoma cells. 79 These findings suggest that GSK3 plays a tumor suppressor role in B cell lymphoma. On the other hand, it has been shown that the GSK3 inhibitor 9-ING-41 induces cell death in numerous lymphoma lines in vitro and in vivo. 80 Moreover, increased GSK3 expression in B cell lymphoma patients correlates with reduced overall survival arguing for a tumor promoter role of GSK3. 80 Since GSK3 has many different target molecules and affects a plethora of biological processes, it is possible that the signaling context of the given cell determines the outcome of GSK3 inhibition. A model that would help predict under which conditions GSK3 inhibition induces cell death rather than boosting metabolic activity is crucial in order to be able to use GSK3 inhibitors to treat lymphoma patients.
In summary, the discussed studies emphasize the complexity of metabolic homeostasis regulation in B cells. In order to successfully expand, transformed B cells need to bypass different checkpoints.
They need to evade senescence induction while maintaining high levels of metabolic activity without completely depleting their energy stores. The signaling context of the given malignancy and the physiological niches these tumors occupy in vivo could determine how the cells respond to metabolic stress. A better understanding of the specific roles different metabolic regulators play in lymphoma biology would be rewarding in improving cancer treatment and disease stratification.

| ME TABOLI C HOMEOS TA S IS IN PL A S MA CELL S AND MEMORY B CELL S
Plasma cells are terminally differentiated effector cells that secrete large amounts of antibodies to neutralize pathogens. Although these cells do not proliferate, their metabolic profile differs dramatically from naive B cells. Since plasma cells can secrete thousands of antibody molecules per second, they require enormous quantities of different biosynthetic precursor molecules and additional carbon sources to generate energy. 9 The metabolic program of plasma cells varies depending on the niches they occupy and depending on their longevity. Both short-and long-lived plasma cells have been shown to take up glucose and to use it for protein glycosylation. 81 Longlived plasma cells are able to use glucose-derived pyruvate to fuel the TCA cycle, and this metabolic feature plays an important role in promoting their survival. 81 Although it is currently unclear how these different metabolic programs are established in plasma cells, their dependence on glucose metabolism highlights potential targets amenable to medical intervention. In addition to glucose, plasma cells also consume amino acids to enable biomass synthesis. The activity of the transcriptional repressor protein, Blimp1, which is important for the establishment of the plasma cell transcription program, favors the expression of amino acid transporters, and at the same time represses the transcription of AMPK activators sestrins 1 and 3. 82 Although some insight exists on which signaling pathways are needed to supply plasma cells with nutrients, much less is known about how plasma cells maintain metabolic homeostasis and how they cope with metabolic dysregulation. One process involved in regulating metabolic homeostasis that has been shown to play an important role in plasma cells is autophagy. During autophagy, misfolded proteins and defective organelles are engulfed in specialized vesicles called autophagosomes, which fuse with lysosomes leading to the degradation of the autophagosome contents. This process requires the coordinated action of a set of specialized proteins to recruit cargo, to assemble the autophagosome, and to induce the fusion of autophagosomes and lysosomes. 83 Originally, autophagy has been viewed as a means to gain energy during periods of starvation.
It is becoming increasingly apparent however that autophagy is also needed in order to remove damaged organelles and toxic protein aggregates. Plasma cells face increased proteotoxic and oxidative stress due to their high levels of protein synthesis. The lack of an essential autophagy molecule, Atg5, results in an expansion of the endoplasmic reticulum (ER) and increased antibody secretion in plasma cells. 84 However, these cells show reduced energy levels and a shortened life span. Thus, autophagy represents a trade-off between antibody generation and viability in plasma cells. 84 In a similar manner, autophagy has been found to be essential for the survival of B1a cells. B1a cells are a unique long-lived B cell subset providing first-line defenses against many common pathogens by secreting the so-called "natural antibodies." The deletion of Atg7 in B cells has been found to result in the loss of B1a cells, without affecting conventional B cells. 10 Autophagy has also been shown to play a crucial role in the disrupted. An inability to inhibit energy-consuming processes or a failure to increase energy production may simply result in an energetic collapse. Insufficient ATP supply will halt many vital cellular functions and cells will ultimately succumb to apoptosis. In many instances however, the absence of metabolic regulators does not result in an acute depletion of energy stores, but rather impacts on viability over a longer period of time. Increasing evidence suggests that ROS production provides a link between cell metabolism and cellular life span. A significant portion of intracellular ROS is produced as a consequence of normal mitochondrial function 87 (Figure 3). The electron transport chain (ETC) transfers F I G U R E 3 Intracellular ROS production and antioxidant defense mechanisms in B cells. The primary sources of ROS in B cells are the ER, the ETC of the mitochondria, and membrane-bound NOX and DUOX enzymes. Redox balance is maintained by the GSH/GSSG and the Trx red /Trx ox system. NADPH is the ultimate electron donor in redox cascades involving both systems. Cystine is imported via the glutamate/ cysteine antiporter system x c --and reduced to cysteine. Cysteine is used for the biosynthesis of proteins and glutathione. Pre-B cells and naive B cells lack the x c system and depend on exogenous cysteine provided to them by other cell types (not shown). The components of the x c-system are expressed upon BCR stimulation so that activated B cells can take up the more prevalent cystine. 90  III of the ETC resulting in a partial reduction of oxygen to superoxide. 87 Mitochondrial dismutase subsequently converts superoxide to hydrogen peroxide, which is free to diffuse into the cytosol.
Consistent with ROS production being associated with normal mitochondrial function, enhanced oxygen consumption often correlates with increased ROS production in B cells. Various cellular stresses that affect mitochondrial health, the rate of electron flux through the ETC, the concentration of electron carriers, mitochondrial membrane potential, oxygen availability, and the type of substrate being oxidized can also augment ROS production. [87][88][89] A second source of ROS that may play an important role in B cell biology is protein folding in the ER (Figure 3). Formation of immunoglobulins, which are rich in disulfide bonds, can lead to significant ROS production. In professional secretory cells such as plasma cells, the ER can represent a major source of intracellular ROS. 90 Additional sources of ROS include plasma membrane-associated NADPH oxidases (NOX) or dual oxidases (DUOX) 90 (Figure 3).  103 HuR is a RNA-binding protein that affects the splicing of a vast array of mRNA transcripts and is essential for GC B cell responses. The exact mechanism of ROS accumulation in the absence of HuR is not known yet, but may include reduced TCA cycle activity due to impaired splicing of mRNA that encodes dihydrolipoamide S-succinyltransferase. 103 Autophagy deficiency has also been reported to result in ROS accumulation