Tet2 deficiency promotes IgG1+ B‐cell expansion and differentiation blockade through deregulation of the Nfkbia–c‐Rel axis

Abstract

The germinal center (GC) reaction is essential for orchestrating humoral immunity by producing plasma cells (PCs) and memory B cells (MBCs). TET2, an α‐ketoglutarate‐dependent dioxygenase, plays a critical role in B‐cell exit from the GC and in plasma cell differentiation. Moreover, TET2 functions as a tumor suppressor in diffuse large B‐cell lymphoma (DLBCL), with mutations frequently observed in the ST2 DLBCL subgroup, which is marked by elevated NF‐κB and PI3K signaling and predominant expression of IgG B‐cell receptors (BCRs). We used a combination of in vivo mouse models and in vitro differentiation systems to investigate the effects of Tet2 deficiency on IgG1+ GC B‐cells. We performed flow cytometry, gene expression, and DNA methylation analysis to assess differentiation, proliferation, and molecular alterations. Tet2‐deficient IgG1+ GC B‐cells displayed impaired differentiation into both PCs and MBCs, accompanied by enhanced proliferation. These cells exhibited hypermethylation and repression of the Nfkbia locus, increased activation of the NF‐κB subunit c‐Rel, and sustained high levels of surface IgG1. Upon recall immunization, Tet2‐deficient IgG1+ MBCs failed to efficiently differentiate into PCs, resulting in their accumulation and further GC expansion. These findings demonstrate that Tet2 is essential for balancing proliferation and terminal differentiation of IgG1+ GC B‐cells during the humoral response. The impaired regulation of this balance due to Tet2 loss provides mechanistic insight into a contributory pathway that may facilitate DLBCL transformation in TET2‐mutated cases.

BACKGROUND

The Ten‐Eleven Translocation (TET) genes play a crucial role in gene expression and chromatin structure regulation by oxidizing 5‐methylcytosine to 5‐hydroxymethylcytosine (5hmC), 5‐formylcytosine (5fC), or 5‐carboxylcytosine (5caC). ¹ The TET2 gene is recurrently inactivated by somatic mutations in a variety of myeloid and lymphoid malignancies and in clonal hematopoiesis. ¹ , ² , ³ In mature B‐cell malignancies, TET2 inactivating mutations occur in 6%–10% of diffuse large B‐cell lymphoma (DLBCL) cases. ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹

DLBCL originates from B cells transitioning through the germinal center (GC) reaction, a key step in B‐cell differentiation, facilitating the selection of high‐affinity antibodies and the generation of an efficient humoral immune response. ¹⁰ After antigen encounter, activated B cells form the GC in which they undergo AID‐mediated immunoglobulin (Ig) gene somatic hypermutation and class switch recombination, and become memory B‐cell (MBC) or plasma cell (PC) secreting antigen‐specific antibodies. ¹⁰ , ¹¹ Class switch recombination, also known as isotype switching, involves the swapping of the IgM constant immunoglobulin domains of the B‐cell receptor (BCR) to those of IgG, IgE, or IgA, resulting in changes in the BCR structure, protein interactions, and intra‐cellular signalization and allowing specialized cellular functions. ¹² , ¹³ While memory B cells differ from their naïve counterparts in various aspects, the Ig isotype plays a crucial role in their inherent functional distinctions. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Specifically, naive B cells express an IgM and IgD BCR, whereas memory B cells may undergo isotype‐switching, allowing expression of IgG, IgA, or IgE. The isotype transition has been demonstrated to contribute significantly to robust secondary antibody responses upon reexposure to the inducing antigen, mainly mediated by IgG‐expressing MBCs. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ The isotype of the BCR has been shown to influence B‐cell differentiation fate during the humoral immune response: IgE and IgG1 MBCs show a propensity to differentiate into plasma cells, whereas IgM MBCs have a higher propensity to form GC. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵

GC B‐cells undergo massive chromatin remodeling, along with massive modification of DNA methylation, H3K27 acetylation, and gene expression profiles. ²⁶ , ²⁷ , ²⁸ We have shown that TET2 is required for an efficient humoral immune response and is a tumor suppressor in DLBCL. ²⁹ Tet2 inactivation in mice, whether induced in hematopoietic stem cells, B‐cell progenitors, or GC B‐cells, impairs the exit of B cells from the GC and plasma cell differentiation, and predisposes to lymphoma development. ²⁹ Tet2 deficiency in GC B‐cells leads to hypermethylation of regulatory elements and repression of associated genes. ²⁹ , ³⁰ Tet2 inactivation in human and mouse B‐cells also impairs isotype switching from IgM to IgG. ²⁹ , ³¹ Furthermore, Tet2‐deficient mice exhibit a significant reduction in the titer of antigen‐specific IgG1 antibodies in the serum and in antigen‐specific IgG1‐producing plasma cells in the bone marrow, for both low‐ and high‐affinity IgG1 upon immunization. ²⁹

Surprisingly, Tet2‐deficient mice produce similar numbers of antigen‐specific IgM‐producing plasma cells in the bone marrow as wild‐type mice, regardless of whether the antibodies are of low or high affinity. ²⁹ This finding indicates a discrepancy between the IgM and IgG1 responses. Similar results were observed in immunized mice with concurrent inactivation of both Tet2 and Tet3 in GC B‐cells. ³² Moreover, plasma cells generated in vitro from human memory B‐cells harboring a germline Tet2‐inactivating mutation secreted similar levels of IgM but had reduced IgG secretion compared to wild‐type cells. ³¹ Interestingly, the TET2‐mutant DLBCL subtype was shown to be distinctive in its preferential expression of IgG BCR. ⁷ Based on these observations, we conducted a detailed analysis of the effects of Tet2 deficiency on the biology of IgG1+ GC B‐cells. We show that the loss of Tet2 leads to hypermethylation of Nfkbia, increased activity of the NF‐kB subunit c‐Rel, and sustained IgG1 cell surface levels in GC B‐cells. This dysregulation disrupts differentiation processes and promotes proliferation during primary immunization. Specifically, Tet2‐deficient IgG1+ GC B‐cells predominantly generate CD80low IgG1+ MBCs, which are known to preferentially re‐enter the GC during recall responses. In secondary responses, Tet2‐deficiency impedes the differentiation of these IgG1+ MBCs into plasma cells, resulting in their accumulation and further re‐entry into the GCs. This contributes to GC hyperplasia and an increased subset of IgG1+ GC B‐cells in secondary GCs. These findings highlight Tet2's crucial role in maintaining the balance between differentiation and proliferation during both primary and secondary immune responses. We propose that these mechanisms may contribute to the transformation of TET2‐mutated GC B‐cells in DLBCL.

MATERIALS AND METHODS

Mice and immunization

Mice carrying constitutive and conditionally inactivated Tet2 alleles have been previously described. ³ , ³³

For the analysis of GC formation in T cell‐dependent immune response, Tet2‐KO and WT age‐matched 2‐ to 3‐month‐old mice were immunized with sheep red blood cells (SRBCs; Atlantis) (1 × 10⁸ cells per mouse) via intraperitoneal injection. For primary immune response analysis, mice were sacrificed 10 days after immunization. For secondary immune response analysis, mice received a second SRBC injection between 28 and 30 days after the primary injection and were analyzed 5 days later.

Flow cytometry and cell sorting

Single‐cell suspensions from spleen were stained after ammonium chloride (Stemcell Technologies) lysis of red blood cells in PBS (Invitrogen) supplemented with 2% fetal bovine serum (FBS; Invitrogen). The following anti‐mouse antibodies were purchased from Biolegend: CD38 (clone 90), ebioscience: IgG1 (M1‐14D12), IgM (II/41), CD19 (1D3), B220/CD45R (RA3‐6B2) or BD Bioscience: IgD (11‐26 c.2a), IgM (R6‐60.2), IgE (R35‐72), CD43 (R2‐60), CD80 (16‐10A1), CD138 (281‐2), Blimp1 (5E7), CD95/FAS (Jo2), GL7 (GL7), CD4 (RM4‐5), CD8 (53‐6.7) and Gr1 (RB6‐8C5). Anti‐phospho‐Akt pS473 (M89‐61), ‐Akt pY308 (J1‐223.371), ‐Erk pT202/pY204 (20 A), ‐Btk pY223 (N35‐86), ‐Syk pY352 (17 A/P‐ZAP70), and ‐mTor pS2448 (O21‐404) were purchased from BD Bioscience. Anti‐c‐Rel (1RELAH5) antibody was purchased from eBioscience, and anti‐Phospho‐p65 pS536 (93H1) was purchased from Cell Signaling Technology. Antibodies were conjugated with fluorescein isothiocyanate (FITC), phycoerythrin (PE), Peridinin‐Chlorophyll‐Protein (PerCP)‐Cyanin (Cy)5.5, PE‐Cy7, allophycocyanin (APC), APC‐Cy7, Pacific blue (Pb), APCeFluor780, brilliant violet 421 (BV421), brilliant violet 605 (BV605), Alexa‐Fluor 647 (AF647), or brilliant violet 510 (BV510). Sytox blue (ThermoFisher) or Fixable Viability Stain 450 (FVS450) (BD Horizon) dyes were used to exclude dead cells. For memory B‐cell gating, T and myeloid cell lineages were excluded using anti‐CD4, anti‐CD8, and anti‐Gr1 (Lin‐). Memory B cells were defined as Lineage– IgD– GL7– CD19+ CD38+ (Supporting Information S1: Figure 11), a standard strategy excluding naïve, GC, and extrafollicular B cells. ³⁴ , ³⁵ While CD80 and PD‐L2 together best define MBC subsets, CD80 alone is a validated and reliable proxy within IgG1+ cells. ³⁵ , ³⁶ The analyzed population was IgG1+, consistent with a GC origin, though rare GC‐independent switching may occur. ³⁷ We did not use a hapten‐based immunization approach to avoid bias toward high‐affinity clones and to capture the full polyclonal pool. ³⁸ To capture the early GC‐derived MBC wave during first immunization, the mice were analyzed at Day 10 as an early wave of MBCs peaks between Days 6 and 8, and ~50% of the total IgG1+ MBC pool already formed by Day 11. ³⁹ For c‐Rel intracellular staining, cells were fixed and permeabilized using the Transcription Factor Buffer Set (BD Biosciences) following manufacturer's instructions. Intranuclear c‐Rel staining was performed as previously described. ⁴⁰ , ⁴¹ For phospho‐flow, cells were fixed with BD Cytofix Fixation Buffer (BD Biosciences) and then permeabilized with BD Phosflow Perm Buffer II (BD Biosciences), following the manufacturer's instructions. Flow cytometry analysis was performed using a FACS Canto II, Canto X, or LSR Fortessa Flow Cytometer (BD Biosciences), and cell sorting was performed using a FACSAriaIII or Influx (BD Biosciences). Immunophenotypic data were analyzed using the FlowJo Version 10 software (TreeStar, Inc.). The gating of cells was based on the isotype control and fluorochrome minus one (FMO) setting. Gating strategies are shown in Supporting Information S1: Figure 11.

In vitro generation of induced germinal center B‐cells (iGCB) and induced plasma cells (iPC)

In vitro‐iGCB, iPC, and induced memory B cells (iMBC) were generated and defined according to established protocols and nomenclature. ⁴² , ⁴³

Splenic murine naïve B‐cells were isolated from 2‐ to 3‐month‐old mice using the mouse B‐cell isolation kit (CD43 (Ly‐48) MicroBeads; Miltenyi Biotec) according to the manufacturer's instructions (yield >90% CD19+ B220+). Naïve B‐cells were cultured in the presence of 40LB‐cells expressing CD40L and producing BAFF as described previously. ²⁹ , ⁴² Briefly, purified B‐cells (110 × 10³ cells per well) were cultured in a 6‐well plate in the presence of a confluent layer of 40LB‐cells that had been irradiated with 80 Gy of γ‐ray. rmIL‐4 (1 ng/mL; Peprotech) was added to the primary culture for 4 days, and on Day 4, the cells were replated onto a new irradiated feeder layer and cultured with rmIL‐21 (10 ng/mL; Peprotech). Cells were harvested at Days 4 (D4) or 8 (D8) and processed for flow cytometric analysis and RNA isolation. Cells were cultured in B‐cell medium: RPMI‐1640 (Invitrogen) supplemented with 10% FBS (Invitrogen), 10 mM HEPES, 1 mM sodium pyruvate, 1 mM nonessential amino acids, 0.055 mM 2‐ME, 100 U Penicillin/Streptomycin, and 0.3 mg/mL glutamine (all from Invitrogen). Induced memory B‐cells (iMBC) differentiation was performed as previously described. ⁴³ For continuous culture of the cells until Day 24, cells were plated at 110 × 10³ cells per well on a new layer of 40LB feeder cells every 4 days until Day 12 and then every 6 days until Day 24.

For the analysis of PC propensity of the iGCB expressing specific isotype, cells were sorted by FACS on Day 4 as CD138− CD19+ GL7+ and IgM+ or IgG1+ or IgE+. Cells were then plated at 110 × 10³ cells per well on a new layer of 40LB feeder cells in the presence of rmIL‐21 for 4 days. Cells were analyzed on Day 8 for the frequency of CD138+ iPC.

For cell culture in the presence of vitamin C (l‐ascorbic acid; Sigma Aldrich) or 5‐azacytidine (Sigma Aldrich), a final concentration of 250 µM vitamin C or of 400 nM 5‐azacytidine was added on Day 0 and Day 4.

Production of MSCV containing shRNA

To produce IRES‐GFP‐MSCV (Murine Stem Cell Virus) particles, we used the jetPRIME kit (Polyplus). On Day 0, a mixture of 6 µg of MSCV plasmid containing the shRNA sequence (obtained from ⁴⁴ ) and 4 µg of psi eco plasmid was prepared in 500 µL of jetPRIME buffer. Subsequently, 20 µL of jetPRIME was added, and the mixture was vortexed for 10 s, briefly centrifuged, and incubated at room temperature for 10 min. The prepared mix (500 µL) was then added drop by drop to a Petri dish containing HEK 293 T cells at 60% to 80% confluency. The Petri dish was gently rocked and placed in the cell incubator. On Day 1, the cell culture medium was replaced with 6 mL of complete DMEM medium (Invitrogen). On Day 2, the MSCV containing cell culture medium was filtered through a 0.22 µm low protein adsorption filter and aliquoted into 1 mL tubes. These tubes were promptly frozen in liquid nitrogen and stored in a −80°C freezer.

Transduction of iGCB

Naïve B‐cells were isolated from the spleens of 2‐ to 3‐month‐old mice using the mouse B‐cell isolation kit (CD43 (Ly‐48) MicroBeads; Miltenyi Biotec) with a yield exceeding 85% CD19+ B220+. These B‐cells were cultured in a 6‐well plate with irradiated 40LB‐cells (confluent, irradiated with 80 Gy of γ‐ray) in B‐cell medium (8 mL per well). On Day 0, naïve B‐cells (110 × 10³ cells per well) were cultured with rmIL‐4 (1 ng/mL) (Peprotech) for 4 days. On Day 4, the B‐cells were harvested and then cultured (110 × 10³ cells per well) with rmIL‐21 (10 ng/mL) (Peprotech). On Day 7, the B‐cells were harvested again, and CD138−, CD19+, GL7+, IgM− cells (IgM−iGCB) were sorted and transduced with MSCV. The sorted IgM‐iGCB were cultured in a 6‐well plate at 0.5 × 10⁶ cells/mL in 3 mL of transduction medium (B‐cell medium with rmIL‐21 [10 ng/mL], HEPES [30 mmol/L] [Invitrogen], and polybrene [4.8 µg/mL]; Sigma‐Aldrich). One milliliter of MSCV virus‐containing supernatant was added, and 6‐well plates were centrifuged at 2500 RPM for 90 min at 33°C. After centrifugation, IgM‐iGCB were cultured in a 6‐well plate (500 × 10³ cells in 8 mL per well) with irradiated 40LB‐cells in B‐cell medium containing rmIL‐21 (10 ng/mL). On Day 11, B‐cells were harvested, and GFP + B‐cells (transduced B‐cells) were sorted. The sorted transduced B‐cells were cultured in a 6‐well plate (110 × 10³ cells in 8 mL per well) with irradiated 40LB in B‐cell medium with rmIL‐21 (10 ng/mL) until Day 14. At Days 11 and 14, transduced B‐cells were counted with trypan blue (Invitrogen) and analyzed by flow cytometry.

For Blimp1 transduction, a PRDM1 (human) cDNA fragment was obtained from Genescript and was subcloned into an IRES‐GFP‐MSCV retrovirus backbone. Virus production and cell transduction were performed as previously described. ²⁹

Validation of MSCV containing shRNA

The efficiency of MSCV containing shRNA against c‐Rel was validated using an in vitro model similar to the one used for transduction of iGCB. At Day 0, splenic murine naïve B‐cells from a Tet2‐KO mouse were isolated from 3 to 4‐month‐old mice using the mouse B‐cell isolation kit (CD43 (Ly‐48) MicroBeads; Miltenyi Biotec) following the manufacturer's instructions (yield >90% CD19+ B220+). Naïve B‐cells were cultured on Petri dishes with confluent 40LB‐cells irradiated with 80 Gy of γ‐ray. For one Petri dish, 1 million isolated B‐cells were cultured in 48 mL of B‐cell medium with rmIL‐4 (1 ng/mL) (Peprotech). On Day 3, B‐cells were harvested and transduced with MSCV. B‐cells were cultured in a well of a 6‐well plate at 0.5 × 10⁶ cells/mL in 3 mL of transduction medium (B‐cell medium with rmIL‐4 [1 ng/mL], HEPES [30 mmol/L] (Invitrogen), and polybrene (4.8 µg/mL); Sigma‐Aldrich). One milliliter of MSCV containing supernatant was added, and 6‐well plates were centrifuged at 2500 RPM for 90 min at 33°C. After centrifugation, B‐cells were cultured on Petri dishes with confluent irradiated 40LB‐cells. We seeded 3.98 × 10⁶ B‐cells in 48 mL of B‐cell medium with rmIL‐4 (1 ng/mL) per Petri dish. On Day 7, B‐cells were harvested, and GFP + B‐cells were sorted. GFP+ transduced B‐cells were processed for RNA or protein extraction.

The efficiency of MSCV containing shRNA against Nfkbia was validated in the A20 cell line. A20 cells were transduced with MSCV in a well of a 6‐well plate; 1.6 million A20 cells were cultured in 3 mL of complete RPMI medium (Invitrogen) containing HEPES (10 mmol/L) and polybrene (1.6 µg/mL). One milliliter of MSCV containing supernatant was added, and 6‐well plates were centrifuged at 2500 RPM for 90 min at 33°C. After centrifugation, 6‐well plates were placed in the cell incubator for 4 h. Then, A20 cells were centrifuged and cultured in 10 mL of complete RPMI medium. Two days after, transduced GFP + A20 cells were sorted and cultured in complete RPMI medium. Transduced A20 cells in the expansion phase were processed for RNA.

RNA extraction and quantitative real‐time PCR

Total RNA was extracted using the RNeasy Mini Kit (Qiagen) or the Total RNA Purification Plus kit (Norgen) with DNase (Norgen) treatment. RNA concentration was determined using Qubit (LifeTechnologies), and integrity was verified using Agilent 2100 Bioanalyzer. cDNA synthesis from RNA was performed using the SuperScript IV Reverse Transcriptase kit (Thermo Scientific). All TaqMan probes were purchased from Applied Biosystems: Prdm1 (Mm00476128), Nfkbia (Mm00477798_m1), Tet2 (Mm00524395), and Rel (Mm01239661). Gene expression was normalized to Polr2a (Mm00839493), Abl1 (Mm00802029_m1), and Gapdh (Mm99999915) using the ΔCT method. Expression was detected using TaqMan Universal PCR Master Mix (Applied Biosystems) on an ABI PRISM 7500 (Applied Biosystems) or a QuantStudio 5 (Applied Biosystems).

Quantification of DNA methylation

The CpG methylation‐sensitive digestion of genomic DNA was carried out with the EpiJET DNA Methylation Analysis Kit, based on MspI/HpaII digestion (Thermo Scientific), following the manufacturer's instructions using 100 ng purified genomic DNA. qPCR was performed using SYBR Green Select Master Mix (Thermo Scientific) following the manufacturer's instructions. The primers for common PCR were as follows: forward: 5′‐CTCTGACTCTGGTCTGAAGT‐3′; reverse: 5′‐GTCTCCTGCTTCGTGTTATC‐3′. The methylation level was quantified using the 2^−ΔCt method and presented as the percentage of cytosine methylation at the target‐gene site.

Western blotting

Cell lysates were prepared by direct lysis into Laemmli 2× (Biorad) + DTT (Sigma‐Aldrich), heated 10 min at 95°C, sonicated, and stored at −80°C. Just before migration, protein lysates were thawed and heated for 5 min at 95°C. Protein lysates were separated on a NuPAGE 10% Bis‐Tris Gel (Invitrogen) and then transferred to a 0.22 µM nitrocellulose membrane (Biorad). Proteins were stained with the following primary antibodies: anti‐c‐Rel (sc‐6955; Santa Cruz Biotechnology) and anti‐HSC70 (sc‐7298; Santa Cruz Biotechnology). HSC70 protein was used as an internal control. HRP‐conjugated antibody (anti‐mouse; Jackson ImmunoResearch) was used as a secondary antibody. Pierce™ ECL Plus Western Blotting Substrate and SuperSignal™ West Atto Ultimate Sensitivity Substrate kits (Thermo Fisher Scientific) were used for detection. Signals were acquired using an ImageQuant^TM LAS 4000 system (GE Healthcare), and expression was quantified using Fiji software. ⁴⁵

Statistical analysis

Statistical significance of differences between the results was assessed using a two‐tailed unpaired Student's t‐test with Welch's correction or a paired RM one‐way ANOVA performed using Prism (version 5.03 or 10.0.1; GraphPad software). Statistically significant P values: *P < 0.05, **P < 0.01, ***P < .001. Error bars displayed throughout the paper represent SEM or SD as indicated in figure legends. Mutual exclusivity of mutations in human DLBCL patients' cohorts was computed using Gitools. ⁴⁶ No statistical method was used to predetermine sample size. No blinding and no randomization of samples were applied. No data was excluded.

RESULTS

Tet2 deficiency affects the generation of IgG1+ plasma cells and memory B‐cells in primary immunization and results in the expansion of IgG1 + GC and memory B‐cells upon secondary challenge

We have previously shown that Tet2‐deficiency leads to GC hyperplasia and a reduced isotype switching toward IgG1 upon immunization. ²⁹ We first reproduced these observations using sheep red blood cells (SRBC), a classical T cell‐dependent immunization (Figure 1A for primary and secondary immunization schemes). Flow cytometry analysis of primary and secondary immunization revealed higher frequencies and cell numbers of GC B‐cells in Tet2‐KO compared to Tet2‐WT mice (Supporting Information S1: Figure 1A). Upon primary immunization, the monitoring of IgM and IgG1 positive GC B‐cells revealed an increase in IgM+ and a decrease in IgG1+ GC B‐cells in Tet2‐KO mice with respect to wildtype (Figure 1B and Supporting Information S1: Figure 1B). Given our NP‐CGG immunization data, ²⁹ indicating a plasma cell differentiation defect of IgG1+ B‐cells but not of IgM+ B‐cells, we anticipated that this differentiation blockade may lead to accumulation of IgG1+ cells after a second immunization. Indeed, five days after a second immunization, we observed a significantly higher frequency and cell numbers of IgG1+ GC B‐cells in Tet2‐KO mice compared to Tet2‐WT mice (Figure 1B and Supporting Information S1: Figure 1B).

Figure 1.

Tet2 deficiency impairs memory B‐cell and plasma cell differentiation of IgG1+ germinal center B‐cells. (A) Schematic diagram of the protocol of SRBC primary and secondary immunizations. (B) Representative cytometry plot and quantification of IgM+ and IgG1+ GC B‐cells (CD19+CD95+GL7+) from Tet2‐WT and Tet2‐KO mice upon primary (n = 8) and secondary immunization (n = 18) with SRBC. (C) Representative cytometry plot and quantification of IgM+ and IgG1+ MBCs, (Lin (CD4, CD8, Gr1)‐IgD‐GL7‐CD19+CD38+) from Tet2‐WT and Tet2‐KO mice upon primary and secondary immunization with SRBC (primary immunization: Tet2‐WT [n = 7] and Tet2‐KO [n = 9]; secondary immunization: Tet2‐WT [n = 9] and Tet2‐KO [n = 10]). (D) Representative cytometry histograms and quantification of the expression of CD80 on the cell surface of IgM+ and IgG1+ MBCs from Tet2‐WT (n = 7) and Tet2‐KO (n = 9) mice upon primary immunization with SRBC. (E) Representative cytometry plot of PBs (CD19+CD138+CD43+) from Tet2‐WT (n = 4) and Tet2‐KO (n = 5) mice upon secondary immunization with SRBC. (F) Representative cytometry plot and quantification of IgM+ and IgG1+ PBs (CD19+CD138+CD43+) from Tet2‐WT (n = 4) and Tet2‐KO (n = 5) mice upon secondary immunization with SRBC. All P values were calculated using an unpaired two‐tailed t‐test, *P < 0.05, **P < 0.01, ***P < 0.001 and ns, not significant, in all experiments. GC, germinal center; MBC, memory B cell; PBs, plasmablasts; SRBC, sheep red blood cells.

We next explored the role of Tet2 in memory B‐cell differentiation within the GC during the primary immune response. Tet2‐KO mice exhibited a significant decrease in the percentage of IgG1+ MBCs (CD19+ IgD− GL7− CD38+ IgG1+), while no significant change was observed for IgM+ MBCs (CD19+ IgD− GL7− CD38+ IgM+) (Figure 1C). To assess the cell‐autonomous impact of Tet2 deficiency on MBC differentiation, we examined CD19‐Cre/Tet2‐KO mice, revealing a consistent decrease in the percentage of IgG1+ MBCs, mirroring findings in constitutive Tet2‐KO mice (Supporting Information S1: Figure 2A). To further establish the cell‐autonomous nature of this phenotype, we used an antigen‐free in vitro MBC differentiation system. ⁴³ In this system, the culture of naïve B‐cells from Tet2‐KO and Tet2‐WT mice resulted in significantly lower percentages of induced memory B‐cells (iMBCs) from Tet2‐KO mice at Day 9 (D9) (Supporting Information S1: Figure 2B). These results underscore a B‐cell autonomous mechanism driving the alteration of MBC differentiation upon Tet2 loss.

We next conducted a detailed analysis of the MBC subpopulations, delineated by CD80 expression levels, which characterize functionally distinct MBC subsets. CD80hi MBCs preferentially differentiate into plasma cells, while CD80lo MBCs tend to re‐enter GC B‐cells during a recall response. ³⁵ , ³⁶ , ⁴⁷ Our analysis uncovered that IgG1+ MBCs from immunized Tet2‐KO mice exhibited a significantly lower proportion of CD80hi MBCs and a higher proportion of CD80lo MBCs compared to their counterparts from immunized Tet2‐WT mice (Figure 1D). In contrast, IgM+ MBCs from immunized Tet2‐KO mice showed a significantly higher percentage of CD80hi MBCs and a lower percentage of CD80lo MBCs (Figure 1D). These findings highlight the impact of Tet2 deficiency on MBC differentiation, influencing the composition and function of MBC compartments.

Given the impact of Tet2 deficiency on plasma cell and MBC differentiation in the primary response, we investigated MBC behavior in the secondary immune response. MBC analysis showed a tendency toward a decrease in IgM+ MBCs and a significant increase in the percentage and the cell numbers of IgG1+ MBCs in Tet2‐KO mice, potentially indicative of impaired differentiation and subsequent accumulation (Figure 1C and Supporting Information S1: Figure 1C). We next assessed the frequency of plasmablasts (PBs) in the spleens of mice after secondary immunization. The analysis revealed a significant reduction in plasmablasts in Tet2‐KO mice (Figure 1E and Supporting Information S1: Figure 2C). Further examination of the Ig isotypes expressed by these PBs demonstrated a significant decrease in IgG1‐expressing PBs and a concurrent increase in IgM‐expressing PBs in Tet2‐KO mice (Figure 1F).

Overall, our data suggests that Tet2 loss disrupts IgG1+ memory B‐cell differentiation, reducing their generation and altering their differentiation potential. This results in the replenishment of GCs, contributing to GC hyperplasia ²⁹ (Supporting Information S1: Figure 1A). Combined with the observed effects on plasmablasts, these results underscore the critical role of Tet2 in terminal B‐cell fate determination during humoral immune responses.

The expansion and differentiation blockade of IgG1+ Tet2‐KO GC B‐cells are cell‐autonomous and can be reproduced in vitro

To ascertain the cell‐autonomous nature of the previous observation, we employed an in vitro differentiation model in which naïve B‐cells undergo substantial proliferation, efficient switching to IgG1 and IgE, and differentiation into plasma cells ⁴² (Figure 2A for scheme). Equal numbers of Tet2‐WT and Tet2‐KO naïve B‐cells were grown on 40LB feeder cells supplemented with IL‐4 for 4 days. On Day 4 (D4), equivalent numbers of B‐cells were seeded with IL‐21 for an additional 4 days (Day 8, D8). We confirmed the IgG1 isotype switching defect in Tet2‐KO iGCB at D4 (Figure 2B). The cell frequencies of IgG1+ B‐cells were significantly lower in Tet2‐KO cells compared to Tet2‐WT cells (Figure 2B); however, a slight decrease was noticed in the cell number of IgG1+ Tet2‐KO but did not reach statistical significance (Supporting Information S1: Figure 1D). On D8, we observed significantly higher frequencies and cell numbers of IgG1+ B‐cells derived from Tet2‐KO cells (Figure 2B and Supporting Information S1: Figure 1E). D8 IgM+ iGCB were virtually absent in these cultures of both genotypes (Figure 2B and Supporting Information S1: Figure 1E).

Figure 2.

The expansion and differentiation blockade of IgG1+ Tet2 ‐KO GC B‐cell are cell‐autonomous. (A) Schematic diagram of the in vitro differentiation system. (B) Representative cytometry plot and quantification of IgM+ and IgG1+ in vitro iGCB (CD19+ CD95+ GL7+) from Tet2‐WT and Tet2‐KO mice at D4 (n = 18) and D8 (n = 15). (C) Schematic illustration of the in vitro culture system of iGCB expressing a specific Ig isotype. (D) Number of live Tet2‐WT and Tet2‐KO B‐cells cultured on 40LB with IL21 on Day 8 (D8) (n = 4). (E) Representative cytometry plots of iPC on Day 8. (F) Quantification of flow cytometry data corresponding to D8 iPC (n = 4). (G) Number of live iPC (CD19+ CD138+) Tet2‐WT and Tet2‐KO B‐cells on Day 8 (D8) (n = 4). All P values were calculated using an unpaired two‐tailed t‐test, *P < 0.05, **P < 0.01, ***P < 0.001 and ns, not significant, in all experiments. D, Day; iGCB, induced germinal center B‐cells, iPC, induced plasma cells; SRBC, sheep red blood cells.

We next assessed the propensity of iGCB cells to differentiate into plasma cells based on their specific Ig isotype using the in vitro model. For this experiment, at D4, unswitched IgM+ B‐cells and IgG1+ and IgE+ switched B‐cells were flow‐sorted, and equal numbers of cells were plated on 40LB‐cells in the presence of IL‐21 for another 4 days (Figure 2C for schema). At D8, we observed significantly higher cell numbers of IgG1+ Tet2‐KO cells compared to IgG1+ Tet2‐WT cells (Figure 2D). In contrast, the cell numbers were similar for both IgM+ and IgE+ cells (Figure 2D). The frequency and cell numbers of iPC (CD19+ CD138+) were significantly lower in Tet2‐KO IgG1+ cell cultures with respect to Tet2‐WT IgG1+ cell cultures (Figure 2E–G). In contrast, the frequencies and cell numbers for IgE+ and IgM+ Tet2‐KO cells were similar to controls (Figure 2E–G). Consistent with the findings of previous studies, ¹³ , ¹⁴ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²³ plasma cell differentiation was significantly higher in cultures from Tet2‐WT IgG1+ cells compared to those from Tet2‐WT IgM+ cells (Figure 2E–G).

The observed block in plasma cell differentiation in Tet2‐KO IgG1+ GC B‐cells, but not in Tet2‐KO IgM+ GC B‐cells, potentially indicates that Tet2 is needed to initiate the plasma cell differentiation program specifically within IgG1+ GC B‐cells. Supporting this proposition, an analysis of previously published and publicly available data revealed higher transcription levels of Tet2 in light zone IgG1+ GC B‐cells, particularly those with active BCR signaling (Nurr77+), compared to light zone IgM+ GC B‐cells. ¹⁷ Our analyses similarly show that Tet2 expression is higher in D4 IgG1+ Tet2‐WT iGCB compared to D4 IgM+ Tet2‐WT cells, with even greater levels observed in D8 IgG1+ Tet2‐WT iGCB (Supporting Information S1: Figure 3). Together, these findings underscore the crucial role of Tet2 in the terminal differentiation of IgG1+ GC B‐cells.

In summary, our in vitro findings highlight that Tet2 loss in IgG1+ GC B‐cells leads to their expansion with a concurrent blockade of differentiation into plasma cells, as observed in vivo, and demonstrate the cell‐autonomous nature of these observations.

High cell surface levels of IgG1 on Tet2‐deficient GC B‐cells correlate with impaired plasma cell differentiation

Cell‐surface BCR density and signaling influence B‐cell fate, including plasma cell differentiation. ¹⁸ , ²² , ²⁵ , ⁴⁸ To delve into the mechanisms of plasma cell differentiation block in Tet2‐deficient IgG1+ GC B‐cells, we first analyzed the cell surface expression of IgM and IgG1 on GC B‐cells from Tet2‐WT and Tet2‐KO mice immunized with SRBC, as well as on in vitro generated GC B‐cells at D4 and D8. IgM expression levels were not different between GC cells from Tet2‐WT and Tet2‐KO in primary and secondary SRBC immunization (Figure 3A,B). While the expression levels of IgG1 were similar on Tet2‐KO and Tet2‐WT GC B‐cells upon primary immunization, they were significantly higher on Tet2‐KO GC B‐cells upon secondary immunization (Figure 3A,B).

Figure 3.

Tet2 deficiency induces upregulation of IgG1 expression on the cell surface of germinal center B‐cells. (A) Representative cytometry histograms and quantification of the MFI of IgM and IgG1 staining on the cell surface of GC B‐cells (CD19+ CD95+ GL7+) from Tet2‐WT (n = 9) and Tet2‐KO (n = 10) mice upon primary immunization with SRBC. (B) Representative cytometry histograms and quantification of the MFI of IgM and IgG1 staining on the cell surface of GC B‐cells (CD19+ CD95+ GL7+) from Tet2‐WT (n = 9) and Tet2‐KO (n = 9) mice upon secondary immunization with SRBC. (C) Representative cytometry histograms and quantification of the MFI of IgM and IgG1 staining on the cell surface of Tet2‐WT (n = 11 [IgM] and n = 13 [IgG1]) and Tet2‐KO (n = 11 (IgM) and n = 13 (IgG1)) iGCB (CD19+ CD95+ GL7+) at D4. (D) Representative cytometry histograms and quantification of the MFI of IgG1 staining on the cell surface of Tet2‐WT (n = 15) and Tet2‐KO (n = 15) iGCB (CD19+ CD95+ GL7+) at D8. (E) Quantification of the MFI of IgG1 staining on the cell surface of iGCB (CD19+ CD95+ GL7+) from Tet2‐WT (n = 11) and Tet2‐KO (n = 11) mice between D4 and D8. All P values were calculated using an unpaired two‐tailed t‐test, *P < 0.05, **P < 0.01, ***P < 0.001 and ns, not significant, in all experiments. GC, germinal center; iGCB, induced germinal center B‐cells; MFI, mean fluorescence intensity.

In the in vitro GC culture system, IgM and IgG1 expression levels on D4 iGCB showed no difference between both genotypes (Figure 3C). The expression level of IgM on iGCB at D8 was not assessed because these cells were virtually absent from the culture (below 0.3%) (Figure 2B). At D8, IgG1 expression levels were significantly higher on Tet2‐KO iGCB cells (Figure 3D), consistent with the in vivo results. The representation of IgG1 expression over time between D4 and D8 indicated a significant decrease in IgG1 levels at the cell surface in the WT condition, while in the KO condition, IgG1 levels significantly increased between D4 and D8 (Figure 3E). IgE expression levels at D4 or D8 iGCB showed no difference between both genotypes (Supporting Information S1: Figure 4A). No significant differences in the IgM and IgG1 expression levels on MBCs from Tet2‐WT and Tet2‐KO mice during primary and secondary SRBC immunization were observed (Supporting Information S1: Figure 4B,C).

Our data suggests a connection between Tet2 loss, plasma cell differentiation, and IgG1 expression levels. To explore this connection further, we FACS‐sorted IgG1+ Tet2‐KO iGCB cells as higher and lower IgG1 expression levels (IgG1high and IgG1low) and assessed their differentiation abilities into PC (Supporting Information S1: Figure 5A). The IgG1high cells showed diminished differentiation into PC, suggesting a direct impact of IgG1 expression levels on PC differentiation (Supporting Information S1: Figure 5B). To gain a dynamic perspective, we monitored the cell culture for 24 days, analyzing cells at D4, D8, D12, D18, and D24 time points (Supporting Information S1: Figure 5C for schema). At each time‐point, cells were seeded at the same cell numbers and analyzed for proliferation by trypan blue exclusion cell count, as well as for cell‐surface expression of CD138 (for PC differentiation), GL7 and Fas (for iGCB), and IgG1. At each time‐point, we consistently observed significantly higher cell numbers in Tet2‐KO compared to Tet2‐WT cultures (Supporting Information S1: Figure 5D). Flow cytometry analysis showed significantly lower percentages of iPC and significantly higher iGCB in Tet2‐KO compared to Tet2‐WT conditions throughout all the time points (Supporting Information S1: Figure 5E,F). The frequencies of iPC progressively increased over time in cultures from both genotypes, with a slower increase observed in the Tet2‐KO condition (Supporting Information S1: Figure 5E). Throughout the 24‐day culture period, except at D4, analysis of IgG1 expression consistently showed higher IgG1 levels on Tet2‐KO iGCB compared to Tet2‐WT (Supporting Information S1: Figure 5G). Pearson's correlation test revealed a negative correlation between the frequency of CD138+ cells and IgG1 expression levels in Tet2‐KO cells (coefficient r = −0.9322) (Supporting Information S1: Figure 5H).

Overall, our results established a strong link between elevated IgG1 expression and impaired plasma cell differentiation in Tet2‐deficient GC B‐cells.

c‐Rel activity contributes to impaired plasma cell differentiation in Tet2‐deficient IgG1 + iGCB

Due to the difference in IgG1 expression between Tet2‐WT and Tet2‐KO GC B‐cells, we assessed phosphorylation of downstream BCR‐signaling kinases in iGCB cells at D4 and D8. No difference was observed in the phosphorylation levels of Akt (S473 and Y308), Spleen tyrosine kinase (Syk) (Y352), Btk (Y223), extracellular signal‐regulated kinase (Erk) (T202/Y204), and mammalian target of rapamycin (mTor) (S2448) at baseline (D0 naïve B‐cells) as well as at D4 and D8 in IgG1+ iGCB (Supporting Information S1: Figure 6A–C).

Our previously published analysis of the transcriptional profile of Tet2‐KO GC B‐cells revealed a significant enrichment of NF‐κB target genes. ²⁹ BCR activates several downstream signaling pathways, including the NF‐κB pathway. ⁴⁹ NF‐κB activity plays an important role in the regulation of GC reaction formation, maintenance, and differentiation into plasma and memory B‐cells. ⁵⁰ , ⁵¹ , ⁵² Conditional deletion of c‐Rel and RelA in GC B‐cells revealed that c‐Rel is required to maintain GC populations, whereas RelA promoted Blimp1‐mediated plasma cell development. ⁵¹ Specifically, c‐Rel promotes activated B‐cell proliferation and inhibits the transition to plasma cells by repressing Blimp1 expression. ⁵² , ⁵³ To explore NF‐κB activation, we analyzed phosphorylated P65/RelA (pP65) levels and c‐Rel nuclear localization by flow cytometry. ⁴⁰ In D4 iGCB IgM+ cells, no difference was observed in c‐Rel expression, intranuclear c‐Rel, or pP65 levels for both genotypes (Figure 4A,B). Remarkably, P65 phosphorylation levels were significantly higher in D4 IgM+ compared to D4 IgG1+ iGCB (P = 0.0002) (Figure 4A,B). At D4, IgG1+ iGCB showed no difference in c‐Rel total expression levels between Tet2‐WT and Tet2‐KO cells, but a tendency toward increased c‐Rel nuclear localization in Tet2‐KO cells (P = 0.0555) (Figure 4A,B). However, at D8, significantly higher levels of c‐Rel expression and c‐Rel nuclear translocation were found in IgG1+ Tet2‐KO iGCB compared to their WT counterparts (Figure 4C,D). No significant difference in pP65 was observed for D4 and D8 iGCB for both BCR isotypes (Figure 4A–D). Nevertheless, the mean fluorescence intensity (MFI) for pP65 tended to be higher in D8 IgG1+ iGCB Tet2‐KO compared to Tet2‐WT and displayed a broader distribution of values (Figure 4D). These data suggest that the c‐Rel higher activity might contribute to the differentiation block in IgG1+Tet2‐KO B‐cells.

Figure 4.

Tet2 deficiency induces higher c‐Rel activity in IgG1+ germinal center B‐cells. (A, B) Representative cytometry plots (A) and quantification (B) of intracellular and nuclear localization of c‐Rel and phosphorylated p65 (Ser536) in sorted IgM+ and IgG1+ Tet2‐WT and Tet2‐KO B cells at D4 of the in vitro GC model (n = 3). (C, D) Representative cytometry plots (C) and quantification (D) of intracellular and nuclear localization of c‐Rel and phosphorylated p65 (Ser536) in sorted IgG1+ Tet2‐WT and Tet2‐KO B cells at D8 of the in vitro GC model (n = 3). (E) Expansion factor compared to empty MSCV condition between D11 and D14 and quantification of MFI of IgG1 in IgG1+ iGCB (GFP+ CD19+ CD138− Fas+ GL7+ IgG1+) at D14 of the iGCB transduction model, after transduction of Tet2‐KO B‐cells by an empty MSCV, a MSCV containing shRNA Rel 908, or a MSCV containing shRNA Rel 1869 (n = 3). For (B, D), P values were calculated using an unpaired two‐tailed t‐test. For (E), P values were calculated using a paired RM one‐way ANOVA test. *P < 0.05, **P < 0.01, ***P < 0.001 and ns, not significant in all experiments. ANOVA, analysis of variance; GC, germinal center; iGCB, induced germinal center B‐cell; MFI, mean fluorescence intensity; MSCV, murine stem cell virus.

To confirm the involvement of c‐Rel in the Tet2‐KO GC B‐cells phenotype, we transduced Tet2‐KO iGCB cells with shRNA targeting Rel (coding for c‐Rel protein) (Supporting Information S1: Figure 7A for schema). RT‐qPCR and western blot showed a substantial reduction (over 80%) of the transcript and the protein levels of c‐Rel, respectively (Supporting Information S1: Figure 7B,C and Supporting Information S1: Figure 12). At day 14, we observed a significant decrease in proliferation and a substantial reduction in IgG1 cell surface expression on IgG1+ iGCB for Tet2‐KO B‐cells transduced with shRNA against Rel compared to those transduced with an empty MSCV (Figure 4E and Supporting Information S1: Figure 8A). These results confirmed that increased c‐Rel activity contributes to increased proliferation and IgG1 cell surface expression in Tet2‐KO GC B‐cells.

Repression of Nfkbia contributes to the phenotype of Tet2‐deficient B‐cells

To uncover potential mechanisms contributing to the increased activation of c‐Rel in Tet2‐deficient GC B‐cells, we analyzed the impact of Tet2 loss on the expression of genes belonging to the NF‐κB pathway using RNA‐seq data from mouse GC B‐cells. ²⁹ We identified a significant downregulation of Nfkbia (NF‐κB Inhibitor Alpha) gene expression in Tet2‐KO GC B‐cells (Figure 5A), a finding confirmed by RT‐qPCR in IgG1+ iGCB at D8 (Figure 5B). Nfkbia encodes IκBα, an inhibitor of inducible NF‐κB activity that retains the P65 and c‐Rel proteins in the cytoplasm during the resting state. We further explored the connection between this abnormal silencing of the Nfkbia gene and alterations in DNA methylation profiles at the Nfkbia locus. Our previous DNA methylation profiling in purified GC B‐cells from Tet2‐KO and Tet2‐WT mice revealed hypermethylation in the promoter region and an enhancer of Nfkbia in Tet2‐KO GC B‐cells. ²⁹ , ³⁰ These findings suggest that reduced Nfkbia expression in Tet2‐KO GC B‐cells, resulting from DNA methylation changes, contributes to the higher c‐Rel activity observed in these cells.

Figure 5.

Tet2 deficiency induces downregulation of Nfkbia expression in germinal center B‐cells. (A) Normalized gene expression level of Nfkbia by RNAseq of Tet2‐WT and Tet2‐KO GC B‐cells, data from. ²⁹ (B) Gene expression levels by RT‐PCR of Nfkbia in Tet2‐WT and Tet2‐KO GC B‐cells at D8 (IgG1+) and D4 (IgM+ and IgG1+) (n = 4). (C) Quantification of iPC (GFP+ CD19+ CD138+), iGCB (GFP+ CD19+ CD138− Fas+ GL7+), IgG1+ iGCB (GFP+ CD19+ CD138− Fas+ GL7+ IgG1+), and MFI of IgG1 in IgG1+ iGCB at D11 of the iGCB transduction model, after transduction of Tet2‐WT B‐cells by a MSCV containing sh Renilla, a MSCV containing shRNA Nfkbia 1224, or a MSCV containing shRNA Nfkbia 1485 (n = 4). For A and B, P values were calculated using an unpaired two‐tailed t‐test. For (C) P values were calculated using a paired RM one‐way ANOVA test, *P < 0.05, **P < 0.01, ***P < 0.001 and ns, not significant in all experiments. ANOVA, analysis of variance; iGCB, induced germinal center B‐cells; iPC, induced plasma cells; MFI, mean fluorescence intensity; MSCV, murine stem cell virus.

To confirm the role of Nfkbia downregulation in the higher activity of c‐Rel and in the observed phenotype in Tet2‐KO GC B‐cells, we transduced Tet2‐WT IgM‐iGCB cells on day 7 with IRES‐GFP MSCV containing shRNA targeting Nfkbia (Supporting Information S1: Figure 7C). At Day 11, we observed a decrease in the iPC population (significant for one of the two shRNAs), an increase in the iGCB population, an elevated IgG1+ iGCB population, and a tendency toward increased cell surface expression of IgG1 on IgG1+ iGCB compared to Tet2‐WT B‐cells transduced with control shRNA (sh Renilla) (Figure 5C and Supporting Information S1: Figure 8B). These results confirm the involvement of Nfkbia in the increased activity of c‐Rel and in the phenotype of Tet2‐KO GC B‐cells.

Enhancing DNA demethylation activity restores plasma cell differentiation, and Nfkbia and IgG1 expression to Tet2‐WT levels in Tet2‐KO cells

Several studies have shown that vitamin C (VitC) promotes TET‐dependent DNA demethylation and increases 5hmC in several cell types. ⁵⁴ Vitamin C was shown to promote plasma cell differentiation by enhancing TET2 and TET3‐mediated DNA demethylation in vitro and in vivo. ⁵⁵ , ⁵⁶ We hypothesized that vitamin C treatment, by enhancing the catalytic activity of the remaining TET proteins (mainly Tet3), might compensate for Tet2 loss and restore plasma cell differentiation to Tet2‐WT levels. To assess this, cells were treated with Vitamin C or the DNA‐methylation inhibitor 5‐azacytidine (5‐aza) at D0 and D4 of the in vitro differentiation system (Figure 6A and Supporting Information S1: Figure 10A for scheme). Using methylation‐sensitive restriction enzyme qPCR, we first confirmed that vitamin C treatment decreased methylation at a Prdm1 regulatory element locus in intron 2 (Supporting Information S1: Figure 9A) previously demonstrated to be hypermethylated Xin Tet2‐KO cells. ²⁹ This result is in agreement with previously published data of genome‐wide 5hmC modifications showing Xsignificant DNA demethylation of multiple elements at the Prdm1 locus in induced GC B cells after treatment with Vitamin C. ⁵⁶ Our analysis of these publicly available data (Accession number: GSE183679) on the Prdm1 locus is shown in Supporting Information S1: Figure 9B.

Figure 6.

The effect of Vitamin C on plasma cell differentiation of Tet2 ‐WT and Tet2 ‐KO B‐cells. (A) Schematic illustration of Vitamin C treatment in the in vitro iGCB culture system. (B) Number of live Tet2‐WT and Tet2‐KO cells cultured on Day 4 (D4) and Day 8 (D8), in the presence and absence of vitamin C (n = 5). (C) Graph shows iPC (CD19+ CD138+) percentages at D8 upon treatment of Tet2‐WT (n = 6) and Tet2‐KO (n = 6) B‐cells with Vitamin C (250 µM) or vehicle. (D) Quantification of the MFI of IgG1 staining on the cell surface of D8 iGCB (CD19+ CD95+ GL7+) from Tet2‐WT and Tet2‐KO mice, in the presence and absence of vitamin C (n = 9). (E) Gene expression levels by RT‐PCR of Nfkbia in Tet2‐WT and Tet2‐KO iGCB cells at D8 in the presence and absence of vitamin C (n = 3). (F) Snapshot from UCSC genome browser showing 5hmC mark distribution at the Nfkbia locus (32,380 bp, mm10, chr12: 55.474.813−55.507.192) in D4 iGCB cells, in the presence and absence of vitamin C. All P values were calculated using an unpaired two‐tailed t‐test, *P < 0.05, **P < 0.01, ***P < 0.001 and ns, not significant, in all experiments. 5hmC, 5‐hydroxymethylcytosine; D, Day; iGCB, induced germinal center B‐cells; iPC, induced plasma cells; MFI, mean fluorescence intensity.

We next characterized the PC differentiation in the presence of vitamin C or 5‐azacytidine. In Tet2‐KO cells, vitamin C treatment restored cell proliferation (Figure 6B), plasma cell differentiation (Figure 6C), and IgG1 expression (Figure 6D) to Tet2‐WT levels. Similar results were obtained when cells were treated with 5‐azacytidine (Supporting Information S1: Figure 10B,C). We note that vitamin C and 5‐aza treatments also caused shifts in WT cells, likely by enhancing endogenous TET activity and modulating gene expression in these cells.

The changes in PC differentiation in Tet2‐KO cells were associated with a restoration of Prdm1/Blimp1 expression to Tet2‐WT levels (Supporting Information S1: Figures 9C,D and 10D). Blimp1 is known to repress c‐Rel expression by directly binding to c‐Rel locus and this repression is required for plasma cell differentiation. ⁵² Indeed, our previous findings demonstrated that restoring BLIMP1 expression in Tet2‐KO iGCB rescued plasma cell differentiation in these cells. ²⁹ Importantly, analysis of the IgG1 expression levels in BLIMP1‐transduced Tet2‐KO iGCB at D8 showed comparable expression levels to Tet2‐WT (Supporting Information S1: Figure 9E).

We next analyzed the expression levels of Nfkbia and observed that vitamin C treatment also restored Nfkbia levels to WT levels (Figure 6E). Our analysis of genome‐wide 5hmC modifications in iGC B cells treated with vitamin C from publicly available data (Accession number: GSE183679 ⁵⁶ ) identified several sites with increased 5hmC within the Nfkbia locus (Figure 6F).

Overall, these observations reveal that the restoration of DNA demethylation activity through vitamin C treatment can compensate for Tet2 deficiency, normalizing both plasma cell differentiation and the expression levels of Nfkbia and IgG1 to those observed in Tet2‐WT cells.

DISCUSSION

We have recently demonstrated that Tet2‐deficient mice showed GC hyperplasia and impaired plasma cell differentiation upon T‐dependent immunization. ²⁹ Unlike IgG1, the humoral immune response of IgM was, however, unaffected. Further investigation of Tet2‐deficiency GC B‐cells and memory B‐cell differentiation shows that Tet2 inactivation reduces the generation of IgG1+ GC B‐cells during the primary immune response. Additionally, Tet2 deficiency hinders IgG1+ GC B‐cells differentiation into both plasma cells and memory B‐cells, while simultaneously enhancing the proliferative response of IgG1+ GC B‐cells to external stimulations, leading to their accumulation during subsequent immunizations.

While Tet2 loss partly hinders the generation of IgG1+ B‐cells following primary immunization with SRBC, a notable accumulation of IgG1+ GC B‐cells and IgG1+ MBCs occurs upon secondary immunization in Tet2‐KO mice. This phenomenon can be attributed to both impaired differentiation into plasma cells and an increased proliferative capacity of Tet2‐KO IgG1+ GC B‐cells. Indeed, Tet2‐deficient memory B‐cells generated following primary immunization exhibit low CD80 expression levels, indicating a greater tendency to re‐enter GCs during recall responses and a diminished capacity to differentiate into PCs. ³⁵ , ³⁶ , ⁴⁷ Similarly, in vitro experiments show a decrease in the frequency of Tet2‐KO IgG1+ iGCBs at Day 4, followed by an increase at Day 8 compared to WT cells. However, the percentage of Tet2‐KO IgM+ iGCBs is close to zero on Day 8, suggesting that Tet2‐KO cells continue to undergo isotype switching. This indicates that the observed defect in IgG1+ cells generation from Tet2‐KO GC B cells is likely due to delayed isotype‐switching. These data suggest that this switching defect is associated with a blockade in differentiation and an increased proliferative capacity in response to external stimuli (such as IL‐21 in vitro and secondary immunization in vivo), allowing these cells to expand to a greater extent compared to Tet2‐WT cells. Furthermore, the impairment in IgG1 class‐switching and plasma cell differentiation in Tet2‐KO B‐cells occurred despite a consistently heightened proliferative capacity observed in our in vitro assays (Supporting Information S1: Figure 5D). This indicates that these defects are not a secondary consequence of reduced cell division, but rather a direct result of the disrupted molecular program.

BCR expression is crucial for normal B‐cell survival, activation, and proliferation. The BCR can signal in the absence of an antigen (tonic signaling) or after an antigen encounter. ⁵⁷ , ⁵⁸ Both signaling modes were shown to be implicated in proliferation, survival, and differentiation of B‐cells, as well as in lymphomagenesis. ⁴⁹ The strength of BCR signaling, influenced by BCR density, ¹⁸ , ²² , ²⁵ , ⁴⁸ isotypes, and interactions ¹² , ¹³ affects B‐cell fate with IgG expression associated with enhanced plasma cell differentiation compared to IgM. ¹¹ , ¹³ , ¹⁴ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²³ Weak BCR signaling facilitates PC differentiation, while elevated tonic signaling blocks this process, a blockade that can be rescued by lowering surface BCR expression or inhibiting downstream signaling. ⁴⁸ Our results show a negative correlation between plasma cell differentiation and IgG1 expression level in Tet2‐KO GC B‐cells, consistent with the notion that antigen‐independent BCR signaling strength influences plasma cell fate.

Our data suggests that the defect in PC differentiation in Tet2‐deficient IgG1+ GC B‐cells is at least partially linked to increased c‐Rel activity and the repression of the Nfkbia gene, encoding a retro‐control of NF‐κB. NF‐κB activity plays an important role in the regulation of GC reaction formation, maintenance, and outcome. ⁵⁰ , ⁵¹ , ⁵² Conditional deletion of c‐Rel or RelA in GC B‐cells has revealed their distinct roles in maintaining GC populations and promoting plasma cell development, respectively. ⁵¹ c‐Rel, which is specifically upregulated at the GC B‐cell stage, regulates the switch from B‐cell proliferation to plasma cell differentiation by promoting B‐cell proliferation and blocking the transition to plasma cells through the repression of Blimp1 expression. ⁵² , ⁵³ Furthermore, c‐Rel seems to play a direct role in increased immunoglobulin transcription of IgG1. ⁵³ , ⁵⁹ Transgenic mouse models have shown that c‐Rel overexpression in B‐cells or GC B‐cells induces spontaneous GC formation and hyperplasia, providing a strong competitive advantage in the GC. ⁵³ The significance of the Nfkbia‐cRel axis in the proliferation of IgG1+ GC B‐cells is underscored by a CRISPR‐Cas9 screen aimed at identifying transcription factors essential for the survival and/or proliferation of these cells during selection. ⁶⁰ , ⁶¹ Analysis of the published screen data reveals Nfkbia being ranked second among genes with depletion resulting in a positive score, while Rel emerges as one of the top genes with depletion leading to a negative score, suggesting that the Nfkbia‐cRel axis plays a critical role in IgG1+ B‐cells proliferation or survival. We identified the Nfkbia locus as one of the hypermethylated loci in Tet2‐KO GC B‐cells, ²⁹ , ³⁰ suggesting that TET2 is required to induce or maintain Nfkbia expression. Interestingly, recent evidence shows that c‐Rel overexpression in B‐cells upregulates Nfkbia, forming a negative feedback loop that restrains NF‐κB activity. In contrast, Tet2 loss disrupts this autoregulatory feedback by downregulating Nfkbia, ⁶² thereby removing the inhibitory brake on c‐Rel and resulting in sustained nuclear accumulation and transcriptional activity of c‐Rel. Furthermore, c‐Rel overexpression in AID‐imprinted GC B‐cells was recently shown to confer a strong competitive advantage within GCs, allowing persistent GC activity, continuous recirculation of c‐Rel‐overexpressing B cells, and the establishment of a long‐term pool of lymphoma precursor cells. ⁶³ Thus, in Tet2‐deficient GC B‐cells, the increased presence of c‐Rel protein, coupled with reduced levels of its inhibitory factor, Nfkbia, leads to increased nuclear c‐Rel. This balance between these two genes, which controls the survival and/or proliferation of IgG1+ GC B‐cells, is deregulated in the absence of TET2. A positive feedback loop between elevated c‐Rel activity and high IgG1 expression disrupts the delicate balance between proliferation and terminal differentiation. This disruption is amplified through the repression of Blimp1 expression, resulting in a blockade of the differentiation process and an enhanced proliferative capacity of IgG1+ B‐cells in response to external stimuli (Figure 7).

Figure 7.

Diagram depicting the effects of Tet2 inactivation on IgG1 + GC B cells during primary and secondary immunizations, alongside a proposed model of the underlying molecular mechanisms. At the cellular level, Tet2‐deficient mice exhibit germinal center (GC) hyperplasia and impaired plasma cell differentiation following T‐dependent immunization. During primary immunization, Tet2 inactivation reduces the generation of IgG1+ GC B‐cells by impairing isotype switching to IgG1 and hindering their differentiation into plasma cells and memory B‐cells. This deficiency enhances the proliferative response of IgG1+ GC B‐cells to external cues, leading to their accumulation during subsequent immunizations and a persistent differentiation blockade. At the molecular level, Tet2 deficiency directly impacts Prdm1 (encoding Blimp1) expression through hypermethylation of its regulatory elements and induces sustained DNA methylation at the Nfkbia locus. This results in the downregulation of Nfkbia, leading to increased c‐Rel activity, which further represses Blimp1 expression. Elevated IgG1 BCR expression likely sustains c‐Rel activation. Consequently, there is an increased proliferation capacity of IgG1+ GC B‐cells, coupled with a blockade in memory and plasma cell differentiation.

Recent studies on germline TET2 mutations in humans further support our findings. ³¹ , ⁶⁴ Individuals with TET2‐inactivating mutations display reduced class‐switched memory B‐cells. Primary human TET2‐inactivated B‐cells exhibit diminished in vitro memory to plasma cell differentiation propensity and impaired IgG but not IgM secretion compared to wildtype cells. ³¹ These human data align with our observations of Tet2's role in IgG1+ GC B‐cell and MBC differentiation in mice.

Serial rounds of GC reaction have been shown to fuel lymphomagenesis due to the accumulation of genotoxic stress and mutagenesis in DLBCL and follicular lymphoma. ⁶⁵ , ⁶⁶ As Tet2‐inactivation increases IgG1+ GC B‐cells proliferation and CD80hi MBC generation, the lymphomagenic potential of Tet2‐mutated MBC would be further enhanced by the interplay with other genetic aberrations and a pro‐oncogenic microenvironment, including IL6 and CD40L. ²⁹

DLBCL is characterized by NF‐κB signaling pathway activation, with most ABC‐DLBCL cases carrying somatic mutations leading to aberrant NF‐κB activation, whereas only a small fraction of GCB‐DLBCLs show NF‐κB activation. Genetic aberrations include somatic mutation in BCR (e.g., CARD11, CD79A, and CD79B) and Toll‐like receptor (e.g., MYD88) signaling pathways, in negative regulator of the NF‐κB pathway (e.g., TNFAIP3, NFKBIA, and NFKBIE), as well as amplification of the REL locus. ⁵ , ⁴⁹ , ⁶⁷ However, autoantigens also drive BCR‐dependent activation of NF‐κB in ABC‐DLBCL. ⁶⁸ , ⁶⁹ TET2 mutations and REL amplifications are predominantly identified in the GCB‐DLBCL subtype. ⁴ , ⁵ , ⁷ , ⁶⁷ , ⁷⁰ Recent genetic classification of DLBCL subtypes include TET2 mutations in the ST2 subtype (characterized by mutations in SGK1 and TET2), predominantly composed of GCB‐DLBCL. ⁶ , ⁷ , ⁸ , ⁹ , ⁷⁰ Our Tet2‐KO IgG1+ GC B‐cells mirror characteristics of TET2‐mutated DLBCLs. At least 40% of ST2 subtype cases present genetic alterations targeting the BCR‐dependent NF‐κB pathway or negative regulators of proximal BCR signaling. ⁷ Expression profiling shows that ST2 DLBCL subtype presents a high NF‐κB and PI3K activity and is distinctive by the predominant use of IgG BCRs. ⁷ BCR‐isotype expression was shown to significantly associate with specific molecular DLBCL subtype with IgG expression linked with GCB subtype and IgM expression linked with ABC subtype. ²⁴ In addition, we identified mutual exclusivity of TET2 mutations with CD79B, BTK mutations, and REL amplification in DLBCL (Supporting Information S1: Table 1), indicating that these mutations share a common mechanism of action. In the context of lymphomagenesis, c‐Rel hyperactivity induced by TET2 loss‐of‐function may sustain the proliferation and differentiation blockade of IgG+ GC B‐cells. This impediment, coupled with memory B‐cell orientation to re‐enter the GC, could prompt the occurrence of additional mutations, ultimately leading to cellular transformation.

Targeting of REL emerges as a potential therapeutic strategy for TET2‐mutated DLBCL. Furthermore, studies indicate that vitamin C facilitates DNA demethylation in individuals with TET2 germline mutations, reducing the proportion of hypermethylated sites. Vitamin C treatment mirrors the effects of TET2 restoration in a Tet2‐deficient mouse leukemia model, promoting DNA demethylation, cell differentiation, and suppressing leukemia progression. ⁷¹ Two ongoing clinical trials are investigating vitamin C alone or in combination with azacytidine in patients with myelodysplastic syndromes and acute myeloid leukemia carrying TET2 mutations (NCT03433781 and NCT03397173). In a year‐long clinical trial involving daily oral supplementation of 1 g of vitamin C in both control and TET2‐mutated individuals from a family with lymphoma predisposition, a decrease in the proportion of hypermethylated loci was observed, along with a reduction in gene expression divergence between TET2‐mutated and non‐mutated individuals. ⁷² In light of the multiple TET enzymes and pathways influenced by vitamin C, we acknowledge that its effects are likely pleiotropic and may involve compensatory activation of other TET family members, such as TET3. While our findings implicate the Nfkbia–c‐Rel axis as a major target of vitamin C‐mediated modulation in this setting, we recognize that this represents one critical contributor rather than the sole pathway affected. Given that most TET2‐mutant DLBCL appears heterozygous and retains expression of TET1 and TET3, vitamin C emerges as a potential hypomethylating agent for therapeutic interventions.

Finally, this study adds to our previous research regarding the role of Tet2 in GC reactions by demonstrating that Tet2 specifically regulates the proliferation and differentiation of IgG1+ GC B‐cells through the Nfkbia‐c‐Rel axis. The repression of c‐Rel by Blimp1 is required for plasma cell differentiation. ⁵² Our prior work showed that Tet2‐deficiency represses Prdm1 (coding Blimp1) through hypermethylation of its regulatory elements. c‐Rel and Blimp1 have opposite effects on cell proliferation and plasma cell differentiation. ⁵² , ⁷³ Thus, c‐Rel activation and Blimp1 repression would synergize in impairing plasma cell differentiation and stimulating the cell cycle of Tet2‐deficient IgG1+ GC B‐cells (Figure 7).

CONCLUSION

Our study demonstrates that Tet2 is essential for balancing proliferation and terminal differentiation of IgG1+ GC B‐cells. Tet2 deficiency leads to impaired plasma cell and memory B‐cell formation and enhances B‐cell proliferation, partly through repression of Nfkbia and increased c‐Rel activity. These findings highlight the Nfkbia–c‐Rel axis as one major mechanistic pathway by which Tet2 regulates proliferation and differentiation balance and provide mechanistic insight into how TET2 mutations may contribute to lymphomagenesis, particularly in IgG‐expressing DLBCL subtypes. This work suggests potential therapeutic avenues targeting NF‐κB signaling or epigenetic regulation in TET2‐mutated B‐cell malignancies.

AUTHOR CONTRIBUTIONS

Hussein Ghamlouch: Conceptualization; investigation; funding acquisition; writing—original draft; writing—review and editing; visualization; validation; methodology; formal analysis; project administration; supervision; data curation. Michaël Degaud: Investigation; writing—original draft; writing—review and editing; visualization; methodology; formal analysis; data curation. Veronique Della‐Valle: Investigation; writing—review and editing; methodology; visualization; formal analysis; data curation. Alexandre Eeckhoutte: Investigation; software; methodology; writing—review and editing. Marine Armand: Investigation; writing—review and editing; methodology. Amina Joudat: Investigation; writing—review and editing; methodology. Camille Decaudin: Investigation; writing—review and editing; methodology. Pilar M. Dominguez: Methodology; writing—review and editing; investigation. Wojciech Rosikiewicz: Investigation; writing—review and editing; methodology; software. Patrycja Pawlikowska: Investigation; writing—review and editing; methodology. Walaa Darwiche: Investigation; writing—review and editing; methodology. Enguerran Mouly: Investigation; writing—review and editing; methodology. Sheng Li: Investigation; writing—review and editing; methodology; project administration. Ari M. Melnick: Investigation; writing—review and editing; methodology; project administration. Said Aoufouchi: Writing—original draft; investigation; conceptualization; methodology; validation; writing—review and editing; data curation; supervision; funding acquisition; project administration; formal analysis. Olivier A. Bernard: Conceptualization; investigation; funding acquisition; writing—original draft; methodology; validation; writing—review and editing; data curation; supervision; project administration; formal analysis.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENT

Animal experiments were conducted according to the Gustave Roussy Institutional guidelines and authorized by the Direction Départementale des Services Vétérinaires du Val de Marne.

FUNDING

This work is supported by the INSERM, CNRS, Institut National du Cancer (INCa‐DGOS‐Inserm‐ITMO Cancer_18002), the Fondation ARC pour la recherche sur le cancer, SIRIC‐EPICURE, and la Ligue Contre le Cancer. H.G. was supported by the Fondation ARC pour la recherche sur le cancer (Projet Fondation ARC 2018, PJA 20181208056).

Supporting information

Supplementary Table1.

Supplementary Figures.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.