Serine/threonine phosphatase PP2A is essential for optimal B cell function

Protein phosphatase 2A (PP2A), a serine/threonine phosphatase, has been shown to control T cell function. We found that in vitro–activated B cells and B cells from various lupus-prone mice and patients with systemic lupus erythematosus display increased PP2A activity. To understand the contribution of PP2A to B cell function, we generated a Cd19CrePpp2r1afl/fl (flox/flox) mouse which lacks functional PP2A only in B cells. Flox/flox mice displayed reduced spontaneous germinal center formation and decreased responses to T cell-dependent and T-independent antigens, while their B cells responded poorly in vitro to stimulation with an anti-CD40 antibody or CpG in the presence of IL-4. Transcriptome and metabolome studies revealed altered nicotinamide adenine dinucleotide (NAD) and purine/pyrimidine metabolism and increased expression of purine nucleoside phosphorylase in PP2A-deficient B cells. Our results demonstrate that PP2A is required for optimal B cell function and may contribute to increased B cell activity in systemic autoimmunity. Research Article Immunology


Introduction
B cells encounter antigens in secondary lymphoid organs and respond by proliferating and undergoing maturation and differentiation, which include class-switch recombination (CSR), Ab affinity maturation, and generation of memory B cells and plasma cells (1,2). T-independent responses are elicited by polysaccharides that engage the B cell receptor (BCR) or CpG motifs and result in polyclonal B cell activation (3)(4)(5). T-dependent responses are elicited by proteins, such as 4-Hydroxy-3-nitrophenylacetyl-chicken γ-globulin (NP-CGG), that are processed and presented to cognate Th cells for recognition by B cells (6). CD40 engagement by its ligand, which is expressed on T cells, is essential for T cell-dependent Ab responses and represents a major contributor to Ig production (7,8). In vivo, activated mature B cells proliferate and form the germinal centers (GCs) where high-affinity, long-lived plasma cells and memory B cells are generated (1,9,10). GCs develop spontaneously without exposure to foreign antigens (11)(12)(13)(14)(15) in murine lupus and human systemic lupus erythematosus (SLE), and this dysregulation results in the production of autoantibodies (12,13,(16)(17)(18)(19)(20).
Protein phosphatase 2A (PP2A) is a serine/threonine phosphatase ubiquitously expressed in eukaryotic cells and comprises a scaffold-type structural subunit (A), a catalytic subunit (C), and a regulatory subunit (B) (21). There are 2 isoforms for the structural subunit (A α and A β ), 2 isoforms for the catalytic subunit (C α and C β ), and more than 20 regulatory subunits that enable PP2A to act on a wide range of substrates and to participate in cellular processes like cell cycle, gene transcription, and apoptosis (22,23). PP2A levels and activity are increased in T cells from patients with SLE and account for decreased IL-2 production (24)(25)(26)(27). Transgenic mice that overexpress the catalytic subunit PP2A C in T cells develop susceptibility to glomerulonephritis (28), whereas PP2A is requisite for Treg function (29). In addition, studies in patients with acute (30) and chronic B lymphocytic leukemia have signified the importance of PP2A in B cell oncogenesis (31). Given the central role of PP2A in the regulation of many key cellular processes and its role in human disease pathogenesis, therapeutic modalities targeting PP2A are being developed (32). The involvement of serine/threonine phosphatases in the B cell-dependent expression of autoimmunity is not known.
Protein phosphatase 2A (PP2A), a serine/threonine phosphatase, has been shown to control T cell function. We found that in vitro-activated B cells and B cells from various lupus-prone mice and patients with systemic lupus erythematosus display increased PP2A activity. To understand the contribution of PP2A to B cell function, we generated a Cd19 Cre Ppp2r1a fl/fl (flox/flox) mouse which lacks functional PP2A only in B cells. Flox/flox mice displayed reduced spontaneous germinal center formation and decreased responses to T cell-dependent and T-independent antigens, while their B cells responded poorly in vitro to stimulation with an anti-CD40 antibody or CpG in the presence of IL-4. Transcriptome and metabolome studies revealed altered nicotinamide adenine dinucleotide (NAD) and purine/pyrimidine metabolism and increased expression of purine nucleoside phosphorylase in PP2A-deficient B cells. Our results demonstrate that PP2A is required for optimal B cell function and may contribute to increased B cell activity in systemic autoimmunity.
Here, we present evidence that PP2A activity is increased in in vitro-activated B cells and in B cells from lupus-prone mice. We also show that PP2A expression is increased in B cells from patients with SLE. The newly constructed mouse, which lacks PP2A in B cells, responds poorly to immunization with T-dependent and T-independent antigens, with decreased GC formation and plasmablast/plasma cell differentiation. When stimulated in vitro, B cells from Cd19 Cre Ppp2r1a fl/fl (flox/flox) mice show decreased Ig production with decreased CSR and plasmablast/plasma cell differentiation. This globally defective B cell response can be explained by differences in NAD and purine/pyrimidine metabolism and purine nucleoside phosphorylase (PNP) expression. We propose that B cell function depends on PP2A, and its downregulation in B cells from patients with SLE should have therapeutic value.

Results
Increased PP2A expression and function in activated B cells or B cells from lupus-prone mice and SLE patients. SLE is characterized by multiple B cell abnormalities leading to autoantibody production. Mrl.lpr, B6.lpr, and the triple-congenic NZM2410-derived SLE1.2.3 (SLE1.2.3) mice develop lupus spontaneously and display many features of human SLE, including autoantibody production and renal immune complex deposition. To examine whether PP2A has a role in B cell function in systemic autoimmunity, we measured PP2A protein levels in B cells from these mice. PP2A A and PP2A C protein levels were both increased in B cells from lupus-prone mice compared with matched controls ( Figure 1A and Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.130655DS1). We next asked whether PP2A expression was restricted in certain subsets of B cells during development and differentiation. Marginal zone B cells (MZ), follicular B cells (FO), GC B cells, and plasma B cells (PC) were sorted out from MRL/lpr mice and subjected to quantitative PCR (qPCR) (33). All subsets were found to express PP2A A , but its expression was higher in GC B cells and PC ( Figure 1B). To extend our studies to humans, peripheral blood mononuclear cells (PBMCs) from patients with SLE and matched healthy controls were isolated and stained for PP2Ac (the catalytic subunit) (34). Consistently, PP2A C levels were found increased, as determined by flow cytometry, in B cells from SLE patients; even the expression doesn't correlate well with SLE Disease Activity Index (SLEDAI; Figure 1C and Supplemental Figure 2). Of note, B cells with activating or activated phenotypes including memory B cells (CD19 + IgD -CD27 + ), double negative (DN, CD19 + IgD -CD27 -) B cells, and PC (CD19 + CD138 + IgG + ) displayed higher expression of PP2A C compared with naive B cells (CD19 + IgD + CD27 -) ( Figure 1D). We postulated that certain stimuli that activate B cells might drive the expression of PP2A. To test our hypothesis, purified B cells were stimulated with either CpG or anti-CD40, and the expression of PP2A A and PP2A C (Figure 1, E and F) was quantified by qPCR. As expected, stimulation increased the expression of both PP2A subunits. Western blot analysis validated the increased expression of PP2A A and PP2A C (Figure 1, G and H) at the protein level. Increased protein levels of PP2As were accompanied with increased enzymatic activity in B cells from lupus-prone mice compared with matched controls ( Figure 1I and Supplemental Figure 3), and in CpG or anti-CD40 stimulated B cells compared with unstimulated control ( Figure 1J). Our results imply that enhanced PP2A phosphatase activity in B cells may contribute to the pathogenesis of lupus.
PP2A is required for B cell activation and Ig production in vitro. To study the role of PP2A in B cell function, we constructed a flox/flox mouse that lacks the dominant α isoform of the scaffold PP2A A subunit in B cells and studied it along with Cd19 Cre Ppp2r1a +/+ (control) mice. Flox/flox mice did not express the scaffold PP2A A subunit and display decreased protein levels of the catalytic PP2Ac subunit ( Figure 2A). However, the B cells in the spleen (Supplemental Figure 4A) and BM (data not shown) of flox/flox mice were intact and displayed comparable ex vivo apoptosis rates (Supplemental Figure 4B). Because PP2A has been reported to affect B cell survival in a variety of B cell malignancies (31), the development of different subsets of B cells was carefully examined. We did not detect any significant differences in the percentage or the absolute numbers of transitional B cells, MZ, and FO between flox/flox and control mice (data not shown). PP2A A was efficiently deleted but the deletion was restricted in B cells, and expectedly, we did not find any differences in the percentages of total T cells, CD4 + T cells, or CD8 + T cells (data not shown). Interestingly, total serum IgM, IgG, IgG1, and IgA levels were significantly decreased in flox/flox compared with control mice ( Figure 2B), which suggested that PP2A may play a role in B cell activation and Ab production. To assess this possibility, B cells from either control or flox/flox mice were cultured in vitro in the presence of CpG and IL-4 ( Figure 2C). As expected, stimulation with CpG efficiently switched most mature IgM + mature B cells into IgG-producing cells, as indicated by higher titers of IgG but low titers of IgM released into culture medium. Of note, B cells with PP2A A deficiency displayed significantly reduced ability to differentiate into plasma cells and produce antibodies. Induction of activation-induced cytidine deaminase (AID) and Blimp-1 during B cell activation is essential for B cell differentiation and isotype switching. Not surprisingly, failure of induction of AID and Blimp-1 was observed in B cells with PP2A A deficiency ( Figure 2C). Next, we stimulated B cells with anti-CD40 in the presence of IL-4. Consistently, reduced plasma cell formation and failure of AID/Blimp-1 induction were also observed in B cells from flox/flox mice compared with control mice ( Figure 2D). Equal proliferation and surviving rate of B cells from flox/flox mice or control mice in the presence of stimulation excluded the possibility that PP2A A indirectly affects Ig production through regulation on survival and proliferation (Supplemental Figure 4, C and D). Collectively, our data indicate that, in B cells, PP2A is required for cell activation and differentiation.
PP2A is critical for B cell activation and differentiation and Ig production in vivo. Interestingly, the percentages of both spontaneous GC B cells (CD19 + FAS + PNA + ) and T follicular helper cells (Tfh) (CD4 + CXCR5 hi PD1 hi ) in the spleens of flox/flox mice with middle ages were significantly reduced compared with age-matched controls ( Figure 3A), which indicates that PP2A A contributes to B cell activation and differentiation in vivo. To validate our observation in vitro, control and flox/flox mice were immunized with sheep RBCs (SRBCs), which stimulate B cells polyclonally and induce GC formation. Confocal image analysis revealed that flox/flox mice immunized with SRBC developed significantly fewer GCs with smaller sizes in the spleens ( Figure 3B). Flow cytometry analysis further confirmed the reduction of GC B cells, Tfh cells, and CD138 + plasmablast/plasma cells ( Figure  3C) (35). Consistently, flox/flox mice immunized with either T-independent antigen NP-Ficoll ( Figure 3D) or T-dependent antigen NP-CGG ( Figure 3E) developed much lower NP-specific antibodies. Taken together, our results demonstrated that PP2A played a critical role duirng B cell activation and differentiation.
PP2A expression in B cells contributes to disease pathogenesis in pristine-induced lupus. Next, we asked whether mice that lack PP2A in B cells can develop lupus-like manifestations when injected with pristine, a substance known to induce lupus-like sydrome in normal mice. Control and flox/flox mice were administrated pristine i.p., and we recorded that the flox/flox mice displayed reduced titers of anti-ANA IgG in the circulation ( Figure 4A), with decreased IgG deposition in the kidneys ( Figure 4B). These data imply that the presence of PP2A in B cells is needed for the development of autoimmunity.
To identify PP2A downstream molecules, we sorted B cells and subjected their lysates for Western blot analysis. STAT3 is constitutively expressed in B cells and increases after activation. Since dysregulated STAT3 pathway in B cells is linked to lupus pathogenesis (36,37), we examined the expression of STAT3 and pSTAT3 in total B cells from either flox/flox or control mice. To our surprise, pSTAT3 level in B cells from flox/flox mice was signifcantly reduced compared with the controls, even though the levels of total STAT3 were comparable between the 2 mice ( Figure 4C). Furthermore, we extended this observation in human primary B cells in which we successfully silenced PP2A A using a Ppp2r1a siRNA ( Figure 4D). After stimulation with anti-CD40, both flow cytometry ( Figure 4E) and confocal image analysis (Figure 4F) revealed that, in the absence of PP2A A , there was a delayed and reduced pSTAT3 response in B cells. Collectively, our results strongly suggest that PP2A functions upstream of STAT3 during B cell activation.

PP2A in B cells inhibits mitochondrial respiration by suppressing expression of PNP, which is involved in purine and pyrimidine metabolism.
Considering that metabolic reprogramming is essential for B cell activation and Ab production (38), we decided to measure different metabolic parameters in B cells with Ppp2r1a deficiency. Seahorse metabolic analysis revealed that flox/flox B cells have increased mitochondrial respiration (maximal oxygen consumption rate [OCR] and space capacity OCR) but normal glycolysis rates ( Figure 5A and data not shown). It has been shown that mitochondria provide instructive signals for several processes of B cells, including CSR and PC differentiation (39). Therefore, PP2A might regulate B cell differentiation and isotype switching through modulation of mitochondrial function. To further trace the pathways that are affected in the absence of PP2A in B cells, we performed RNA sequencing (RNA-seq) (Supplemental Table 1) and metabolomics analysis (Supplemental Figure 5) in B cells from flox/flox mice and from control mice. Data obtained from the RNA-seq and metabolomics studies were combined to generate integrated metabolic pathway analysis, which revealed significant alterations in NAD, purine metabolism, and pyrimidine metabolism ( Figure 5B and Supplemental Figure 6). Joint analysis showed a significant increase in PNP expression in B cells from flox/flox mice, which was further confirmed by mRNA expression (qPCR) ( Figure 5C) and protein expression (Western blotting) ( Figure 5D). Consistently, human B cells treated with Ppp2r1a siRNA displayed increased mitochondrial respiration (maximal OCR and space capacity OCR), and interestingly, addition of 9-Deazaguanine (a potent inhibitor of PNP) efficiently suppressed the enhanced mitochondrial respiration mediated by PP2A deficiency back to normal level, which validate the suppression of PNP mediated by PP2A during B cell activation ( Figure 5E).

Discussion
In this study, we demonstrate that PP2A is required for optimal B cell function and that PP2A is increased in B cells of humans and mice with SLE. In the absence of PP2A, GC form poorly, and plasma cell generation is decreased. Mice with B cells deficient in PP2A cannot raise proper responses to T-dependent and T-independent antigens. Consistently, mice with PP2A deficiency in B cells are resistant to lupus in a pristine induced murine model. At the molecular level, we showed that PP2A expression in B cells is essential for proper STAT3 phosphorylation and downstream signaling during activation. We further show that deficiency of PP2A in B cells promotes oxidative phosphorylation and increases the expression of PNP, the key metabolic enzyme that is essential for purine metabolism and mitochondrial function. It has been reported that changes in mitochondrial function are essential in providing instructive and stochastic signals to activated B cells for cell fate determination; thus, PP2A expression in B cells contributes heavily to plasma cell differentiation and Ig production (39).
In agreement with our findings is the report that B cell-specific regulatory subunit G5PR (Ppp2R3C) transgenic mice have augmented generation of GC, and after aging, they produce autoantibodies (30). We did not observe reduced B cell numbers in vivo with increased apoptosis, a discrepancy with observations reported in mice lacking the regulatory subunit G5PR in B cells (40), which can be explained by the fact that PP2A regulatory subunits dictate substrate and cell function specificity (41).
Our metabolomics and genomics studies pointed to defects in the NAD, purine metabolism, and pentose phosphate pathway, which is in agreement with studies in cells from patients with acute B cell lymphoblastic leukemia (30). Increased PP2A in B cells from patients with SLE and lupus-prone mice may facilitate the expansion of B cells and their differentiation into Ig-secreting cells. Mice lacking PP2A in B cells displayed a compromised ability to generate GC needed for the completion of the immune response. Interestingly, we found that PNP expression is increased when B cells lack PP2A. Functional polymorphisms of PNP have been linked to IFN type I-linked SLE immunopathology (42), and T cells from patients with SLE -which have increased expression of PP2A (24) -have been reported to have decreased levels of PNP, suggesting a protective role for this important metabolic enzyme in autoimmunity (43). In parallel, PNP deficiency causes severe combined immunodeficiency, with a third of the patients developing humoral autoimmunity (44).
Our studies document that PP2A is required to generate a sufficient immune response to antigens and suggest that drugs or adjuvants able to increase PP2A activity in B cells may improve the efficacy of vaccines. Conversely, curtailment of the activity of PP2A in B cells may offer a novel tool to suppress autoimmune disease activity (45).

Methods
Mice. C57BL/6J (strain 000664) and Cd19 Cre (strain 006785) mice were purchased from the Jackson Laboratory. Ppp2r1a fl/fl mice (strain: 017441, the Jackson Laboratory, FVB background) were backcrossed for at least . Left: Representation of oxygen consumption rate (OCR) of human primary B cells exposed to the indicated treatments (9-Deazaguanine was applied as PNP inhibitor). Right: Dot plots indicate mitochondrial respiration in human primary B cells with PP2Aa deficiency compared with controls exposed to the indicated treatments. Paired t test, mean ± SEM.
Targeted metabolomics analysis. Metabolites were extracted from cells using 80% methanol (vol/vol). Polar metabolomics profiling (303 metabolites) was performed by using liquid chromatography-tandem mass spectroscopy (LC/MS) in the BIDMC mass spectrometry core according to the protocol (46). Once the SRM data were acquired, peaks were integrated in order to generate chromatographic peak areas used for quantification across the sample set. Metabolomics data were analyzed using MetaboAnalyst 4.0 (http://www.metaboanalyst.ca/) (47). Briefly, missing metabolite raw intensity values were filled in with the lowest detectable intensity of the respective metabolites, and all raw intensities were normalized to the sum intensity of the respective replicate. Metabolites with ≥ 2-fold changes between groups were identified and subjected to either enrichment analysis or pathway analysis. For joint pathway analysis, metabolites with ≥ 2-fold changes between groups and genes with significant alterations (P < 0.05, 2-tailed t-test) obtained from RNA-seq data were combined for the integrated metabolic pathway analysis.
Transcriptome analysis (RNA-seq). RNA was extracted from sorted cells (1 × 10 4 ) using QIAGEN RNeasy Micro Kit. RNA-seq was performed using the Smart-seq2 platform. Smart-Seq2 libraries were prepared by the Broad Technology Labs and sequenced by the Broad Genomics Platform. Transcripts were quantified by the Broad Technology Labs computational pipeline using Cuffquant version 2.2.1. All RNA-seq data were deposited in the NCBI's Gene Expression Omnibus database (GEO accession no. GSE144700; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE144700).
Statistics. Student 2-tailed t test and 2-way ANOVA were applied for statistical analysis. P < 0.05 was considered statistically significant, although lower P values are indicated in individual figures. In all graphs, data represent mean ± SEM.
Study approval. For human studies, written informed consent was obtained from all participants, and all studies were approved by the IRB (Committee on Clinical Investigations) at BIDMC. All animal studies were approved by the IACUC at BIDMC.

Author contributions
EM, HL, and GCT designed experiments, analyzed data, and wrote the manuscript. WP, MK, SY, VCK, CI, NRR, JCC, SAA, PL, JM, AS, and MGT provided expertise and reviewed the manuscript.