Abstract

   Malignant B cells rely on B cell receptor (BCR) signals for their
   survival and growth. Besides the Immunoglobulin M (IgM) BCR, lymphoma
   cells can also express non–IgM (IgG) BCRs; however, the role of IgG
   BCRs in malignant B cell is not well understood. Here, we report poorer
   disease outcomes in diffuse large B cell lymphoma (DLBCL) expressing
   high IgM versus those expressing IgG1. Using isogenic lymphoma cells
   expressing distinct BCRs, we found that IgM expressing cells strongly
   outcompete their IgG1 counterparts. Mechanistically, IgG1 BCR is
   associated with a dysfunctional mitochondrial state and reduced cell
   survival. We show that mitochondrial dysfunction is triggered by
   accentuated calcium responses downstream of IgG1 BCR. Genetic reversal
   of IgG1 to IgM, pharmacological dampening of calcium signaling, or
   treatment with interleukin-21 can correct mitochondrial defects and
   rescue IgG1 survival. Our findings demonstrate that distinct BCR
   isotypes are inherently unique and can differentially affect B cell
   lymphoma pathogenesis.
     __________________________________________________________________

   B cell receptor isotypes play a crucial role in B cell physiology and
   related pathologies.

INTRODUCTION

   Mature B cell lymphomas are thought to arise from the malignant
   transformation of B cells undergoing activation in specialized
   microanatomical sites called the germinal centers (GCs) ([52]1–[53]3).
   B cells in the GC rapidly proliferate and mutationally diversify their
   immunoglobulin (Ig) genes through the activity of activation-induced
   cytidine deaminase (AID) ([54]1, [55]4). AID deaminates cytosines to
   uracils and, through mutagenic DNA repair pathways, promotes somatic
   hypermutation (SHM) events to generate high-affinity antibodies ([56]1,
   [57]4). In addition, AID activity directed to a high density of
   hotspots in the Ig constant regions, known as switch regions, are
   processed as DNA double-strand breaks (DNA DSBs) to mediate genetic
   switching from IgM (or its splice variant IgD) to IgG, IgA, or IgE
   isotypes ([58]1, [59]2, [60]4). AID activity is often misdirected to
   non-Ig genes, leading to unwanted genetic lesions that drive the
   malignant transformation of activated B cells ([61]2, [62]4). One of
   the features of mature B cell–derived malignancies is their reliance on
   B cell receptor (BCR) signaling for survival, proliferation, and growth
   ([63]5–[64]7). Similar to the essential functions of BCR signaling in B
   cell development, activation, and differentiation, malignant B cells
   hijack these pathways to propagate growth and survival signals
   ([65]5–[66]7). Notably, small molecule inhibitors targeting proximal
   components of BCR signaling, such as Bruton’s tyrosine kinase (Btk) and
   phosphoinositide 3-kinase (PI3K), have been approved for the treatment
   of B cell neoplasms ([67]8). Although most studies investigating the
   role of BCR signaling in malignant B cells have focused on
   membrane-bound IgM, in line with the GC origin of these neoplasms,
   expression of non-IgM BCR isotypes have also been reported ([68]9,
   [69]10). The non-IgM isotypes, such as IgG and IgA, differ greatly in
   their signaling potential in comparison to IgM ([70]5), but how they
   contribute to the pathogenesis of mature B cell neoplasms is poorly
   understood.

   Among non-IgM BCR isotypes, studies have primarily compared the
   signaling potential of the IgG subclass of BCRs to IgM ([71]11–[72]14).
   IgG BCRs can be divided into IgG1, IgG2, IgG3, and IgG4 subclasses in
   humans and IgG1, IgG2a, IgG2b, and IgG3 subclasses in mice, with IgG1
   being the most dominant of the non-IgM isotypes in the serum. IgG BCRs
   harbor more flexible hinge regions, distinct transmembrane domains, and
   longer cytoplasmic tails in comparison to IgM ([73]15). Moreover, the
   observed differences in signaling between IgM and IgG1 BCRs have been
   mostly attributed to the longer cytoplasmic tail in membrane-bound IgG1
   ([74]12, [75]14, [76]16, [77]17). Like IgG BCRs, the membrane-bound
   forms of IgE and IgA isotypes carry longer cytoplasmic tails
   ([78]18–[79]21). The shorter cytoplasmic tail of IgM makes these
   receptors more reliant on cytoplasmic signaling domains of Ig
   alpha/beta (Igα/β), whereas the longer cytoplasmic tail of IgG1 confers
   less dependence on Igα/β ([80]14). Moreover, compared to IgM BCR, IgG1
   BCR mediates stronger membrane proximal signaling events soon after BCR
   engagement and forms larger signaling microclusters ([81]13). IgG1 BCRs
   also display a lower force threshold to trigger signaling than IgM BCRs
   ([82]16). These key differences in signaling between IgM and IgG1 BCRs
   are shown to influence fate decisions in differentiating B cells, with
   IgG1 favoring plasma cell differentiation and IgM preferred during
   generation of long-term memory cells ([83]22–[84]25). In the context of
   malignant B cells, both antigen-independent tonic BCR signaling as well
   as (self) antigen-dependent chronic BCR signaling have been proposed to
   drive pathogenesis ([85]6, [86]26, [87]27). Related to this, IgM and
   IgG1 BCRs have been reported to differ in the context of both active
   and tonic BCR signaling ([88]28) and therefore might be pertinent
   regardless of the mode of BCR signaling in B cell neoplasms.

   Mature B cell neoplasms are broadly classified into Hodgkin’s lymphoma
   (~9000 new cases/year) and the more common non–Hodgkin’s lymphoma
   (B-NHL ~76,000 new cases/year), most of which arise from GCs. Among
   B-NHL, diffuse large B cell lymphoma (DLBCL) is the most common adult
   lymphoid malignancy, accounting for ~40% of all B-NHL cases. Initial
   classification of DLBCL based on transcriptional signatures defined two
   main DLBCL subtypes, the activated B cell type (ABC) DLBCL and GC B
   cell type (GCB) DLBCL ([89]29, [90]30). The classification of patients
   into these subtypes has immense prognostic value as ABC-DLBCL is
   associated with a more aggressive disease course and poorer clinical
   outcomes, whereas GCB-DLBCL follows a less aggressive disease course
   with better patient prognosis ([91]29, [92]30). More recent studies
   have used deep sequencing to genetically classify DLBCL into subgroups,
   including BCL6 and NOTCH2 alterations (BN2/Cluster1; ABC-like), MYD88
   and CD79b mutations (MCD/Cluster 5; ABC-like), EZH2 and BCL2 mutations
   (EZB/Cluster 3; GCB-like), SGK1 and TET2 mutations (ST2), and NOTCH1
   mutations ([93]31–[94]33). Even with the most recent iteration of the
   LymphGen genetic classification, a large proportion of DLBCL cases
   (~37%) remains unclassified ([95]32, [96]33). Besides the genetic and
   mutational signatures, the GCB-DLBCL and ABC-DLBCL subtypes also differ
   in the BCR isotypes that they express, with ABC-DLBCL mainly expressing
   IgM and GCB-DLBCL having a higher predilection to express non-IgM
   isotypes, such as IgG and IgA ([97]7, [98]9). The expression of non-IgM
   BCR isotypes have also been frequently reported in other B cell
   malignancies, such as, mantle cell lymphoma (MCL), chronic lymphocytic
   leukemia (CLL), and double-hit lymphomas ([99]34–[100]36). A previous
   study in a small cohort of patients with DLBCL highlighted that the
   expression of non-IgM isotypes, specifically IgG1 and IgA, are
   associated with better disease prognosis than those expressing IgM
   ([101]9). However, whether distinct BCR isotypes actively affect the
   differential pathogenesis of B cell lymphomas is an open question.

   We sought to investigate the role of distinct BCR isotypes in the
   pathogenesis of mature B cell lymphomas. We focused on the IgG1
   subclass and found that, among human patients with DLBCL, IgG1
   expression is associated with better disease outcomes compared to IgM.
   Using isogenic human lymphoma cells, a mouse model of GC-derived
   lymphoma, and normal activated B cells, we show that IgM receptors are
   favored over IgG1 across multiple contexts. Our studies further reveal
   that IgG1 expressing cells display reduced survival associated with
   mitochondrial dysfunction and metabolic perturbations, caused by
   enhanced calcium responses in IgG1 cells compared to IgM. Overall, our
   work highlights mechanisms by which distinct BCR isotypes can
   differentially influence B cell fates and the pathogenesis of B cell
   lymphoma.

RESULTS

IgM BCR is associated with more aggressive pathogenesis of B cell lymphoma
compared to IgG1 BCR

   To investigate the role of IgM and non-IgM BCR isotypes in lymphoma
   pathogenesis, we focused on the IgG1 BCR isotype, which is the most
   common non-IgM/IgD isotype expressed on B cells. To study this in the
   context of human B cell lymphoma, we categorized 481 patients with
   DLBCL from The Cancer Genome Atlas (TCGA) into those with high IGHM
   (n = 111) or IGHG1 (n = 63) (mRNA) expression ([102]Fig. 1A). The
   patients with DLBCL with high IGHM expression showed significantly
   poorer disease prognosis as evidenced by a shorter time to
   progression-free survival (median survival: 4.59 years), in comparison
   to patients with high IGHG1 expression (median survival: 6.68 years)
   ([103]Fig. 1B). These differences in patient outcomes could be due to
   underlying distinctions in DLBCL subtypes within IgM and IgG1 groups.
   Therefore, we further dissected our dataset into distinct DLBCL genetic
   subtypes in both IgM high and IgG1 groups. As reported previously
   ([104]33), IgM high samples were enriched for N1, MCD, and BN2 genetic
   clusters ([105]Fig. 1C), whereas high IgG1 expression was mainly
   associated with the EZB and BN2 clusters ([106]Fig. 1C). A large
   fraction of the patient samples in both IgM and IgG1 high groups did
   not belong to any known genetic clusters and was designated as
   unclassified or “other” genetic subtype ([107]Fig. 1C). We compared the
   impact of IgM and IgG1 expression on disease prognosis within BN2, EZB,
   and other clusters, all of which were composed of substantial
   proportions of IgM and IgG1 high samples. Compared to the IgM high
   cohort, the samples with high IgG1 expression were associated with
   better progression-free survival and lower clinical stage of disease in
   BN2 and Other subtypes but not the EZB subtype ([108]Fig. 1, D to F,
   and fig. S1A). The EZB genetic subtype is known to be associated with
   IgG1 expression ([109]33); hence, the similar patient prognosis between
   IgM and IgG1 cohorts within the EZB subtype might be attributed to the
   genetic mutations coevolving with IgG1 BCRs in this subgroup. We also
   compared the impact of IgM and IgG1 expression on patient outcomes in a
   patient cohort predominantly expressing the IGVH3-23 gene family, which
   presumably reflects comparable antigen-driven selection (fig. S1B). The
   IgG1 expression within the VH3-23 cohort also appeared to have better
   progression-free survival when compared to patients with high IgM
   expression, although these differences were not statistically
   significant (fig. S1B). In aggregate, we identify a strong association
   of IgG1 expression with improved survival in DLBCL compared to IgM, and
   these findings are consistent with a previous report ([110]9).
   Moreover, our results further highlight the differential impact of IgM
   and IgG1 BCRs on DLBCL pathogenesis, particularly in a cohort of DLBCL
   patients which remain genetically unclassified.

Fig. 1. IgG1 expression is associated with less aggressive lymphoma
pathogenesis.

   [111]Fig. 1.
   [112]Open in a new tab

   (A) IGHM (left) and IGHG1 (right) expression as transcripts per million
   (TPM) in IgM high (n = 111) or IgG1 high (n = 63) samples. (B)
   Kaplan-Meir curves displaying progression-free survival (probability)
   in IgM and IgG1 high samples. (C) Distribution of distinct DLBCL
   subtypes in IgM and IgG1 high groups. (D to F) Kaplan-Meir curves for
   progression-free survival (probability) in IgM high and IgG1 high in
   BN2 (D), EZB (E), and Other (F) genetic subtypes. sgRNA, single guide
   RNA. (G) Engineering of switched isogenic human lymphoma cell lines
   using CRISPR-Cas9. (H) IgM and IgG1 expression in BJAB, OCI-Ly7. and
   HBL-1 cells before and after CRISPR-Cas9 editing. (I) Tumor growth of
   IgM or IgG1 cells (1 million per mouse) subcutaneously injected into
   NSG mice (n = 5 mice). (J) Tumor growth curves of IgM or IgG1
   expressing BJAB subcutaneously injected into NSG mice. (K) Surface
   expression of Ig-kappa light chain (Igκ) in IgM BJAB cells (gray), IgM
   tumors (gray dashed), and IgG1 BJAB cells (red), tumors which grew from
   IgG1 BJAB cells (red dashed) and BJAB IgM cells stained with secondary
   Ab (no Ab control, black). (L) Cotransplantation of IgG1 and IgM
   lymphoma cells mixed at 1:1. Bottom; Representative plots of IgM and
   IgG1 at days 0 and 21 posttransplant. (M) Quantification of IgM and
   IgG1 cell frequency in tumors after 21 days (n = 4 for Mino IgG1, Mino
   IgM, and BJAB IgG1; n = 5 for BJAB IgM). (N and O) Bar graphs
   quantifying percent IgM and IgG1 BJAB (N) and Mino (O) cells after
   mixing (day 0) and upon coculture for 5 or 8 days. Statistical
   significance calculated by log-rank test [(B) and (D) to (F)], t test
   (I), and two-way ANOVA [(M) to (O)]. Error bars represent ± SEM;
   *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. [(G) and (L)]
   Created in BioRender. Shukla, V. (2025)
   [113]https://BioRender.com/1mfdxwl. See also fig. S1.

   To directly test the role of IgM and IgG1 BCR isotypes in B cell
   lymphoma pathogenesis, we used clustered regularly interspaced short
   palindromic repeats (CRISPR)–Cas9–based genetic engineering to induce
   isotype switching from IgM to IgG1 in three different cell lines, each
   derived from lymphoma patients of distinct subtypes: BJAB
   (ABC-DLBCL/Burkitt-like lymphoma), OCI-Ly7 (GCB-DLBCL), and HBL-1
   (ABC-DLBCL) ([114]Fig. 1, G and H). We used lentiviral transduction to
   deliver three different guide RNAs targeting the 5′ and 3′ of the Sμ
   region and the 3′ of the Sγ1 in the IGH locus ([115]37) in BJAB, HBL-1,
   and OCI-Ly7 cells, which normally express IgM BCR. CRISPR-Cas9
   targeting followed by single-cell plating yielded isogenic lymphoma
   cells that were switched to express IgG1 BCRs, without treatment with
   mitogens or exogenous cytokines ([116]Fig. 1H). We tested the
   tumorigenic potential of IgM or IgG1 expressing isogenic cells in
   xenograft models using the NOD-SCID (nonobese diabetic–severe combined
   immunodeficient) gamma chain–deficient (NSG) mice. The IgM expressing
   BJAB, OCI-Ly7, and HBL-1 formed aggressive tumors upon subcutaneous
   injections in mice and rapidly progressed to humane endpoints for
   euthanasia ([117]Fig. 1I and fig. S1C). However, to our surprise, the
   IgG1 expressing isogenic cells mostly failed to form tumors in vivo
   ([118]Fig. 1I and fig. S1C). In follow-up experiments with IgM and IgG1
   expressing isogenic BJAB cells, at least two mice originally injected
   with IgG1 BJAB cells grew tumors ([119]Fig. 1J). Intriguingly, the
   tumors arising from the IgG1 BJAB cells in these two mice had no
   surface or intracellular expression of Igκ light chain and were BCR
   negative (Ig−) ([120]Fig. 1K). Retransplantation of both IgM and Ig−
   BJAB cells again led to the formation of tumors, which were IgM
   positive and Ig negative, respectively (fig. S1D). In addition to the
   DLBCL cell lines (BJAB, OCI-Ly7, and HBL-1), we also induced switching
   from IgM to IgG1 in Mino cells, an MCL-derived cell line believed to
   originate from memory-like B cells that have exited the GCs. The
   CRISPR-Cas9–based approach could not efficiently induce switching to
   IgG1 in Mino cells, presumably due to altered genomic architecture of
   the IGH locus in these cells. However, stimulation of Mino cells with
   anti-CD40, lipopolysaccharide (LPS), and interleukin-4 (IL-4) generated
   IgG1-switched cells (fig. S1E). Comparable to our results with other
   lymphoma lines, the IgG1 expressing Mino cells did not form tumors in
   vivo, whereas their IgM expressing counterparts again formed aggressive
   tumors (fig. S1F).

   Next, we performed studies in which IgM and IgG1 expressing lymphoma
   cells, BJAB, and Mino, were mixed and cotransplanted in NSG mice, with
   tumors analyzed as they became palpable ([121]Fig. 1, L and M, and fig.
   S1G). The IgM expressing BJAB and Mino cells outcompeted their IgG1
   expressing counterparts and resulted in tumors that primarily expressed
   IgM and lacked IgG1 ([122]Fig. 1, L and M, and fig. S1G). We further
   performed similar cellular competition assays with 1:1 mixing of IgM
   and IgG1 expressing BJAB and Mino cells in culture. As seen in vivo,
   the IgM expressing BJAB and Mino cells rapidly outcompeted their IgG1
   expressing counterparts within 8 days of coculture ([123]Fig. 1, N and
   O). To examine the effects of BCR signaling on cellular kinetics and
   proliferation, we performed 5-bromo-2′-deoxyuridine (BrdU) pulse
   labeling in IgM or IgG1 expressing BJAB and OCI-Ly7 cells. In addition,
   we also generated BCR negative [Ig(−)] BJAB and OCI-Ly7 cell lines to
   study the contribution of BCR signaling to cellular proliferation.
   Notably, the BJAB and OCI-Ly7 cells expressing IgM showed the highest
   cell numbers over 3 days in culture, followed by IgG1 and Ig(−) cells
   [IgM > IgG1 > Ig(−)] (fig. S1H). The frequency of cells undergoing S
   phase, measured by BrdU labeling, followed a subtle but similar trend
   in OCI-Ly7 cells; and, in BJAB cells, IgG1 expression was associated
   with the lowest amount of BrdU labeling (fig. S1I). These findings
   suggest that differences in cellular kinetics mediated by tonic BCR
   signaling are not solely due to changes in proliferation. However, in
   the context of induced BCR signaling with an anti-kappa light chain
   antibody, IgG1 expressing BJAB cells demonstrated a pronounced
   reduction in BrdU labeling compared to their IgM counterparts (fig.
   S1J). Together, from these results, we conclude that the IgM BCRs are
   strongly favored over IgG1 in human lymphoma cells.

IgM expressing cells outcompete IgG1 expressing counterparts in a
TET-deficient model of B cell lymphoma

   One possible caveat of using established lymphoma cell lines could be
   the presence of genetic alterations that are more complementary to and
   reliant on the IgM BCR. To address this, we used a mouse model in which
   B cell–specific deletion of Tet2 and Tet3 (CD19 Tet2/Tet3 DKO) leads to
   the development of a GC-derived lymphoma, which recapitulates several
   features of human disease ([124]38). Moreover, TET2 (and occasionally
   TET3) is recurrently mutated in DLBCL, and patients with TET2 mutations
   are classified into the ST2 DLBCL cluster, which frequently expresses
   IgG1 BCRs ([125]33).

   We isolated CD19 DKO B cells at early stages of lymphomagenesis
   (6-week-old mice) and cultured them on a CD40 ligand (CD40L)– and B
   cell activation factor (BAFF)–expressing fibroblast cell line (40LB
   cells) in the presence of IL-4 to induce isotype switching to IgG1
   ([126]Fig. 2A) ([127]39). Four days of CD19 DKO B cell coculture with
   40LB cells consistently yielded 30 to 40% of IgG1 expressing cells
   ([128]Fig. 2B). Upon depletion of 40LB cells, CD138+ plasmablasts and
   plasma cells, and a population of IgE-switched cells, we transferred
   CD45.2-expressing CD19 DKO B cells into sublethally irradiated CD45.1
   expressing immunocompetent host mice. Four weeks posttransfer, the
   CD45.2+ CD19 DKO B cells expressing a Cre recombinase driven yellow
   fluorescent protein (YFP) reporter were readily detectable in blood.
   Beginning at the 4-week time point, the frequency of IgG1 expressing
   CD19 DKO (YFP+) B cells in blood steadily declined in comparison to IgM
   expressing B cells ([129]Fig. 2C). At 10 weeks posttransfer, the host
   mice were euthanized to examine the transferred CD19 DKO B cells in the
   spleen. Compared to the IgM expressing populations, the frequency and
   the numbers of IgG1 expressing CD19 DKO B cells were significantly
   reduced in the spleen ([130]Fig. 2, D and E). The remaining IgG1
   expressing TET-deficient B cells displayed a higher turnover than their
   IgM expressing counterparts, as evidenced by higher frequency of
   proliferating (Ki67 high) as well as apoptotic (cleaved caspase-3+)
   cell populations ([131]Fig. 2, F and G). It is also worth noting that,
   after 10 weeks, a substantial proportion of transferred TET-deficient B
   cells in the spleen was IgM and IgG1 double-negative cells (fig. S2A).
   Among the IgM and IgG1 double-negative cells, a large fraction
   expressed IgG2b or IgD and a smaller fraction expressed IgG3 or IgA BCR
   isotypes (fig. S2, B and C). We also tested the effect of IgM and IgG1
   BCR in the context of wild-type primary B cells stimulated and switched
   on 40LB cultures. For this, we used congenically marked (CD45.1)
   wild-type B cells, stimulated them on 40LB cultures in the presence of
   IL-4, labeled them with CellTrace Violet (CTV), and transferred them
   into CD45.2-expressing host mice (fig. S2D). Although the frequency and
   numbers of IgM and IgG1 expressing (CD45.1) B cells in spleen 2 weeks
   posttransfer were much lower in comparison to CD19 DKO B cells, the
   cell numbers, frequencies, survival (cleaved caspase-3) and
   proliferation (CTV dye dilution) of IgG1 expressing cells were again
   lower than their IgM expressing counterparts (fig. S2, E to I).
   Together, these studies show that IgG1 expressing normal and
   pre–malignant B cells are more apoptotic than the IgM expressing
   counterparts. This provides further evidence that IgM cells are
   selected over IgG1 cells in a mouse model of GC-derived B cell lymphoma
   driven by TET deficiency.

Fig. 2. IgM expressing TET-deficient B cells outcompete their IgG1 expressing
counterparts.

   [132]Fig. 2.
   [133]Open in a new tab

   (A) Schematic of the experimental design. Cd19 cre Tet2/Tet3-deficient
   DKO B cells (CD45.2+) were isolated from the spleen and cultured for 4
   days using an iGB in vitro culture system. TET-deficient B cells (2
   million) were transplanted into sublethally irradiated CD45.1+
   recipient mice. (B) Representative flow cytometry plot showing the
   frequency of IgM and IgG1 before transplantation in TET-deficient YFP+,
   CD45.1− B cells 4 days after iGB culture. (C) Relative frequency of IgM
   and IgG1 expressing YFP+, CD45.2+ B cells in blood over time. IgM and
   IgG1 frequencies are normalized to their respective pretransplantation
   frequency (n = 9 from two independent experiments). (D) Representative
   flow cytometry plot of YFP+, CD45.1− splenic B cells at week 10
   posttransplantation (left). Quantification of IgM and IgG1 frequency
   (right) (n = 9 from two independent experiments). (E) Quantification of
   the absolute number of IgM and IgG1 splenic B cells gated from YFP+,
   CD45.1− B cells. Cell numbers are normalized to IgM (n = 9 from two
   independent experiments). (F) Representative flow cytometry plots of
   proliferating IgM and IgG1 cells with Ki67 staining from YFP+, CD45.1−
   B cells (left). Quantification of Ki67+ from YFP+, CD45.1− splenic B
   cells (right) (n = 6 from two independent experiments). (G)
   Representative flow cytometry plots of apoptotic IgM and IgG1 cells
   measured by cleaved caspase-3 staining from YFP+, CD45.1− cells (left).
   Quantification of cleaved caspase-3 from YFP+, CD45.1− splenic B cells
   (right) (n = 6 from two independent experiments). Statistical
   significance is calculated by ordinary two-way ANOVA test with Šídák
   multiple comparisons test (C) and paired t test [(D) to (G)]. Error
   bars represent means ± SEM; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.
   Schematic in (A) created in BioRender. Shukla, V. (2025)
   [134]https://BioRender.com/1mfdxwl. See also fig. S2.

IgG1 expressing B cells are more sensitive to BCR-induced cell death than IgM
expressing cells

   Because IgM lymphoma cells show increased survival compared to their
   IgG1 counterparts, we asked whether similar differences are evident in
   IgM and IgG1 expressing cells during normal B cell activation. To
   induce B cell activation and class switching in vivo, we immunized
   wild-type mice with sheep red blood cells (SRBCs) to generate
   T-dependent polyclonal GC responses ([135]Fig. 3A). Immunization with
   SRBCs led to robust induction of GC responses with efficient isotype
   switching to IgG1 ([136]Fig. 3B). Consistent with our observations in
   TET-deficient lymphoma cells, the IgG1 expressing GC B cells were more
   apoptotic than IgM expressing GC B cells on day 14 postimmunization
   ([137]Fig. 3B, right).

Fig. 3. IgG1 expressing B cells are more sensitive to BCR-induced apoptosis.

   [138]Fig. 3.
   [139]Open in a new tab

   (A) Schematic for SRBC immunization. (B) Representative plots for
   gating GC B cells (left). Quantification of cleaved caspase-3 from IgM
   and IgG1 GC B cells (right). Connecting lines show data points within
   each mouse. (C) Schematic of iGB cocultures. (D) Representative plots
   for IgM and IgG1 frequency from CD19+CD138−, 12 hours poststimulation
   (BCR) with anti-kappa antibody (1 μg/ml). (E) Quantification of IgG1
   and IgM B cell frequency with or without BCR stimulation. IgM and IgG1
   are normalized to respective frequencies under untreated conditions
   (n = 7, three independent experiments). (F) Quantification of IgM and
   IgG1 B cell numbers (n = 7, three independent experiments). (G)
   Representative flow plots (left) and quantification of cleaved
   caspase-3 (right). Cleaved caspase-3 is normalized to IgM cells in the
   untreated group (n = 17 from seven independent experiments). (H) Genome
   browser tracks of RNA-seq data in the IgH locus (Ighg1 and Ighm genes
   are highlighted). Black and red tracks are from IgM and IgG1 cells from
   untreated and BCR-stimulated groups. Tracks are averaged from three
   replicates. (I) Volcano plot displaying DEGs in IgG1- versus IgM
   expressing cells from untreated (left) and BCR-stimulated conditions
   (right). (J) Up-regulated (red) and down-regulated (gray) pathways
   identified from DEGs in IgG1 compared with IgM using Metascape. The x
   axis represents the z score. (K) Gene set enrichment analysis (GSEA) of
   apoptosis and abnormal mitochondrial morphology gene sets. The y axis
   denotes enrichment score. NES, normalized enrichment score; P nominal P
   value. Statistical significance is calculated by the paired t test (B)
   and ordinary two-way ANOVA with Šídák multiple comparisons test [(E)
   and (F)]. Error bars represent ± SEM; *P ≤ 0.05; **P ≤ 0.01;
   ***P ≤ 0.001; ****P ≤ 0.0001. Schematics in (A) and (C) created in
   BioRender. Shukla, V. (2025) [140]https://BioRender.com/1mfdxwl. See
   also fig. S3.

   To further study the effects of IgM and IgG1 BCRs on B cell survival in
   a more tractable system, we isolated naive B cells from wild-type mice
   and stimulated them on 40LB cultures with IL-4, mimicking in vivo GCs
   (induced GC B cell or iGB culture). Following 3 days of activation on
   iGB cultures, we depleted 40LB stromal, CD138+ plasmablasts and plasma
   cells, and IgE+ cells and further cultured these B cells with (BCR) or
   without (untreated) treatment with an anti-kappa light chain antibody
   to induce active BCR signaling ([141]Fig. 3C). Twelve hours post–BCR
   stimulation, the IgG1+ cells showed a significant decrease in both
   frequencies and total cell numbers compared with IgG1+ cells under the
   untreated condition ([142]Fig. 3, C to E). Although both IgM and IgG1
   expressing non-plasmablast B cells (B220+CD19+CD138−) demonstrated a
   significant increase in apoptosis upon BCR engagement, this effect
   appeared to be stronger in IgG1 than IgM expressing cells ([143]Fig.
   3G). The IgG1 expressing human lymphoma cells (OCI-Ly7 and BJAB) also
   responded similarly by exhibiting a heightened sensitivity to BCR
   engagement, leading to higher levels of apoptosis in comparison to IgM
   expressing lymphoma cells (fig. S3, A to E). In sum, these studies
   demonstrate enhanced sensitivity of IgG1 expressing normal B cells and
   lymphoma cells to BCR-induced apoptosis.

BCR engagement in IgM and IgG1 expressing cells leads to prominent
transcriptional changes

   To mechanistically interrogate the differential responses of IgM and
   IgG1 to BCR signaling, we performed transcriptional profiling (in
   triplicate) with or without BCR engagement. We activated B cells on iGB
   cultures for 3 days, then depleted 40LB, CD138+, and IgE+ cells, and
   further cultured IgM and IgG1 expressing B cells in the absence or
   presence of anti-kappa antibody (1 μg/ml) to stimulate BCR signaling.
   Following 12 hours of culture with or without BCR stimulation, RNA
   sequencing (RNA-seq) was performed on flow-sorted populations of IgM or
   IgG1-positive cells (B220+ CD19+ CD138− and IgE−). Reassuringly, cells
   sorted for IgM expression under both untreated and BCR-stimulated
   conditions showed high expression of the Ighm transcript, whereas cells
   sorted for IgG1 expression displayed high Ighg1 transcripts and minimal
   Ighm expression ([144]Fig. 3H). We then performed comparative
   expression analysis between different groups to identify differentially
   expressed genes (DEGs). The highest number of DEGs was observed between
   BCR-stimulated IgG1 cells compared with BCR-stimulated IgM cells [299
   up-regulated and 374 down-regulated; log[2] fold change > ±1 and FDR
   (false discover rate) < 0.05] ([145]Fig. 3I). The down-regulated genes
   in stimulated IgG1 cells (compared to stimulated IgM) included
   non–receptor tyrosine kinase Syk, E3 ubiquitin ligase Cblb,
   costimulatory molecule IcosL, and components of Notch signaling
   ([146]Fig. 3I and fig. S3G). The up-regulated genes in stimulated IgG1
   cells included plasma cell differentiation regulators: Prdm1, Xbp1, and
   Irf4, negative regulators of BCR signaling: Spry2, Tigit, and Pdcd1,
   and metabolic enzymes: Shmt1, Shmt2, Mthfd1, and Ppat ([147]Fig. 3I and
   fig. S3, H to K). IgG1 expressing B cells have previously been shown to
   have an increased propensity toward plasma cell differentiation
   ([148]22, [149]23), and our findings are consistent with those
   observations. Moreover, the expression of Aicda, the gene encoding AID,
   which mediates SHM and CSR in activated B cells, was comparable in
   BCR-stimulated IgM and IgG1 B cells (fig. S3L). Considerably fewer
   genes were altered between IgG1 and IgM expressing cells in the
   non–BCR-stimulated (untreated) groups (45 up-regulated and 145
   down-regulated). We also assessed the effects of BCR stimulation within
   IgM or IgG1 cells in comparison to untreated conditions ([150]Fig. 3I).
   The IgM expressing cells showed no significant alterations in gene
   expression upon BCR stimulation for 12 hours, whereas BCR-stimulated
   IgG1 cells showed similar changes in gene expression to those seen in
   comparison with BCR-stimulated IgM cells (up-regulated: Prdm1, Irf4,
   and Xbp1; down-regulated: Notch2, Hes1, and Cblb) ([151]Fig. 3I).

   We next used pathway enrichment analysis to further study the DEGs
   identified in comparison between BCR-stimulated IgG1 and IgM cells. The
   down-regulated pathways in BCR-stimulated IgG1 cells were associated
   with B cell activation, differentiation, and B cell–mediated immunity
   ([152]Fig. 3J). On the other hand, the most highly enriched
   up-regulated pathways in IgG1 cells included one-carbon metabolism
   associated with nucleotide biosynthesis as well as glycine, glutamine,
   and folate metabolism ([153]Fig. 3J). These observations are
   intriguing, particularly in light of previous studies that linked high
   metabolic stress and mitochondrial dysfunction to increased dependency
   on one-carbon metabolism ([154]40, [155]41). Prompted by these data, we
   specifically assessed changes in gene signatures associated with
   mitochondrial morphology and function. We noted a significant
   enrichment of gene signatures associated with abnormal mitochondrial
   morphology in BCR-stimulated IgG1 cells ([156]Fig. 3K, top). In
   addition, we observed an enrichment of gene signatures linked to
   apoptosis in BCR-stimulated IgG1 cells, in line with the increased
   tendency of IgG1 cells to undergo apoptosis ([157]Fig. 3K, bottom). In
   sum, these studies identified altered differentiation and metabolic
   pathways as the predominant transcriptional changes associated with BCR
   engagement in IgG1 B cells.

IgG1 expressing cells display lower mitochondrial mass and function than IgM
expressing counterparts

   Our results thus far demonstrated that IgG1 expressing cells are
   outcompeted by IgM cells, with the transcriptomics data linking BCR
   stimulation of IgG1 to altered cellular metabolism. Therefore, we asked
   whether engagement of IgG1 and IgM BCRs could affect mitochondrial
   dynamics and function. To this end, we measured the mitochondrial mass
   using a cell-permeant fluorescent dye, MitoTracker Green, in IgM and
   IgG1 expressing GC B cells from SRBC-immunized mice. IgG1 expressing GC
   B cells exhibited significantly lower mitochondrial mass than IgM
   expressing GC B cells ([158]Fig. 4A). To confirm these findings with an
   analogous approach, we used a flow cytometry–based imaging method
   (ImageStream) to stain for the mitochondrial membrane protein,
   translocase of outer mitochondrial membrane 20 (Tomm20), in GC B cells
   ([159]Fig. 4B). The IgG1 GC B cells showed reduced Tomm20 staining,
   indicative of lower mitochondrial mass than IgM GC B cells ([160]Fig.
   4B). Furthermore, IgG1 cells switched on iGB cultures displayed lower
   mitochondrial mass compared with IgM expressing cells, and this
   decrease in mitochondrial mass in IgG1 cells was magnified upon BCR
   engagement ([161]Fig. 4C). These changes were not an outcome of altered
   cell size as normalization by cell size in IgM or IgG1 cells did not
   affect these findings ([162]Fig. 4C, right). Similar to mouse B cells,
   the human OCI-Ly7 and BJAB lymphoma cells expressing IgG1 BCRs showed
   lower mitochondrial mass, particularly upon BCR stimulation in BJAB
   cells (fig. S4, A and B).

Fig. 4. IgG1 expressing B cells display lower mitochondrial mass.

   [163]Fig. 4.
   [164]Open in a new tab

   (A) Representative plots of IgM and IgG1 GC B cells (Fas+, CD38low)
   (left) 14 days postimmunization with SRBCs. Histograms show MitoTracker
   Green staining in IgM and IgG1 cells (middle). Median fluorescence
   intensity of MitoTracker Green in IgM and IgG1 cells (right) (n = 8).
   (B) Representative ImageStream analysis of IgM and IgG1 GC B cells
   stained with DAPI (nucleus), Tomm20 (PE: mitochondria), IgG1 (FITC),
   and IgM (PE-Cy7). (C) Histograms showing MitoTracker Green in IgM and
   IgG1 cells from iGB cultures stimulated with anti-kappa (1 μg/ml) for
   12 hours (left). Quantification of relative MitoTracker Green mean
   fluorescence intensity (MFI) (middle). Values are normalized to
   untreated IgM from respective experiments. Relative MitoTracker Green
   MFI also normalized to cell size (right) (n = 6 from two independent
   experiments). (D) Flow cytometry histograms showing MitoTracker Green
   staining (top) and quantification of MitoTracker Green MFI (bottom) in
   IgM and IgG1 splenic B cells from heterozygous IgH^γ1μ/+ mice (n = 3).
   (E) Representative confocal microscopy images of splenic B cells from
   heterozygous IgH^γ1μ/+ mice (left) and quantified as mean fluorescence
   intensity (right). (F) Experimental design with IgG1 cells isolated
   from expressing homozygous IgH^γ1μ/γ1μ mice and treated with TAT-Cre
   for 2 hours and stimulated with anti-kappa antibody. (G) Representative
   flow plots of IgM and IgG1 staining of IgH^γ1μ/γ1μ splenic B cells
   before TAT-Cre treatment and 24 and 48 hours after TAT-Cre treatment.
   (H) Histograms showing MitoTracker Green staining (left) and
   quantification of MitoTracker Green MFI (right) in IgM and IgG1 B cells
   from IgH^γ1μ/γ1μ mice 48 hours post-BCR anti-kappa stimulation (2
   μg/ml) (n = 5). Statistical significance is calculated by the paired t
   test (A), ordinary two-way ANOVA with Šídák multiple comparisons test
   [(C), (D), and (H)], and unpaired t test (E). Error bars represent ±
   SEM; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. See also fig.
   S4.

   Isotype switching is tightly linked to cell activation and
   proliferation, and we questioned whether the decrease in mitochondrial
   mass in IgG1 cells could be unique to their activation history and
   kinetics in comparison to IgM B cells. To account for this, we used a
   transgenic mouse line in which immunoglobulin heavy chain (IgH) locus
   is engineered to harbor an Ighg1 allele with Ighm constant region
   placed in the opposite orientation, flanked by loxP sites (IgH^γ1μ)
   ([165]14). This mouse line allows reverse switching from IgG1 to IgM
   upon Cre-mediated recombination ([166]14). In the IgH^γ1μ/γ1μ
   homozygous mice, all B cells express IgG1 BCRs following VDJ
   recombination in the bone marrow. In the IgH^γ1μ/+ heterozygous mice, a
   population of both IgG1 and IgM expressing B cells is readily
   identified without the need for immunization or exogenous stimulation
   of B cells. Analysis of IgM and IgG1 expressing B cells from IgH^γ1μ/+
   mice again showed lower mitochondrial mass (defined by MitoTracker
   Green) in both IgG1 expressing naive and BCR-stimulated cells in
   comparison with IgM cells ([167]Fig. 4D). Confocal imaging of B cells
   from IgH^γ1μ/+ mice showed a clear reduction in mitochondrial mass
   (Tomm20 staining) in IgG1 compared to IgM expressing B cells ([168]Fig.
   4E). We further tested whether the decreased mitochondrial mass is a
   direct consequence of IgG1 BCR signaling rather than the distinct
   developmental dynamics of IgG1 expressing B cells in this model. Toward
   this, we isolated B cells from IgH^γ1μ/γ1μ homozygous mice and induced
   a reverse switch from IgG1 to IgM upon treatment with TAT-Cre in vitro
   ([169]Fig. 4F). Two hours of TAT-Cre treatment and washout was
   sufficient to induce substantial switching to IgM by 48 hours
   ([170]Fig. 4G). Under these conditions, the cells were either left
   untreated or were stimulated with anti-kappa antibody to induce BCR
   signaling. Notably, a switch from IgG1 to IgM was sufficient to
   increase the mitochondrial mass and rescue the mitochondrial defects
   observed in IgG1 B cells ([171]Fig. 4H).

   To assess the functional consequences of reduced mitochondrial mass in
   IgG1 expressing cells, we performed a Seahorse mitochondrial stress
   test to measure mitochondrial function. Consistent with the reduced
   mitochondrial mass, the IgG1 expressing B cells isolated from IgH^γ1μ/+
   mice displayed lower levels of basal respiration [oxygen consumption
   rate (OCR)] and reduced maximal respiratory capacity [following
   carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) treatment]
   than IgM expressing cells ([172]Fig. 5A). The IgG1 expressing BJAB
   cells also showed lower basal respiration and respiratory capacity than
   IgM expressing isogenic cells ([173]Fig. 5B). It is worth pointing out
   that BCR engagement led to a substantial decrease in basal respiration
   measured in real time in both IgM and IgG1 BJAB cells, indicating a
   negative impact of active BCR signaling on mitochondrial respiration
   ([174]Fig. 5B).

Fig. 5. IgG1 signaling impairs mitochondrial function in normal B and
lymphoma cells.

   [175]Fig. 5.
   [176]Open in a new tab

   (A) Changes in OCR measured over time in IgM (black) and IgG1 (red)
   IgH^γ1μ/+ B cells stimulated with anti-kappa antibody, treated with
   oligomycin, phenylhydrazone (FCCP), and antimycin A and rotenone. (B)
   Changes in OCR measured over time in IgM and IgG1 BJAB lymphoma cells
   treated with anti-Igκ antibody, oligomycin, FCCP, antimycin A, and
   rotenone. (C) Representative flow cytometry plots of IgM and IgG1
   splenic GC B cells (Fas+, CD38 low) stained with MitoTracker Green (x
   axis) and TMRM (y axis). Gated population indicates dysfunctional
   mitochondria (left), identified as lower TMRM and positive MitoTracker
   Green staining. Quantification of the frequency of dysfunctional
   mitochondria in IgM and IgG1 GC B cells (right) (n = 3). (D)
   Representative flow cytometry plots of IgM and IgG1 B cells from in
   vitro iGB cultures stained with MitoTracker Green (x axis) and TMRM (y
   axis) (left). Quantification of the frequency of dysfunctional
   mitochondria measured by identifying as lower TMRM and positive
   MitoTracker Green staining (right) (n = 5 from two independent
   experiments). (E) Frequency of cleaved caspase-3–positive cells in IgM
   and IgG1 B cells from in vitro iGB cultures (n = 5 from two independent
   experiments). (F and G) Representative flow cytometry plots of OCI-Ly7
   (F) and BJAB (G) IgM and IgG1 lymphoma cells stained with MitoTracker
   Green and TMRM after BCR treatment for 24 hours (left). Quantification
   of the frequency of dysfunctional mitochondria (right) (n = 6 from two
   independent experiments (F) and n = 3 (G) (right). Statistical
   significance is calculated using a paired t test (C) and ordinary
   two-way ANOVA with Šídák multiple comparisons test [(D) to (G)]. Error
   bars represent means ± SEM; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001;
   ****P ≤ 0.0001. Schematic in (F) created in BioRender. Shukla, V.
   (2025) [177]https://BioRender.com/1mfdxwl. See also fig. S4.

   Functionally intact mitochondria maintain electrical potential in their
   inner membranes for efficient energy production through oxidative
   phosphorylation. Therefore, as another readout for mitochondrial health
   and function, we stained cells with tetramethylrhodamine (TMRM), a
   fluorescent dye that labels mitochondria in a membrane
   potential–dependent manner. IgG1 cells in GCs displayed a higher extent
   of mitochondrial depolarization than the IgM GC B cells, as measured by
   lower TMRM staining ([178]Fig. 5C). This depolarized mitochondrial
   state was noticeable even in IgG1 cells compared with IgM cells
   cultured under iGB conditions and was strongly associated with BCR
   stimulation ([179]Fig. 5D). BCR stimulation following iGB cultures led
   to an ~5-fold increase in IgG1 cells with depolarized mitochondria
   compared to those with IgM ([180]Fig. 5D). For reference, the
   difference in IgG1 GC cells with depolarized mitochondria in comparison
   to IgM GC B cells was more subtle (~1.6-fold) ([181]Fig. 5, C and D).
   This led us to hypothesize that IgG1 expressing cells in GCs could be
   somewhat protected from accumulating dysfunctional mitochondria. To
   test for signals that might contribute to this, we assessed the role of
   IL-21, a prominent GC-associated cytokine, which is shown to amplify
   BCR or CD40-dependent activation of AKT (phospho-AKT) and the mammalian
   target of rapamycin (mTOR) target, phosphorylated S6 ([182]42). BCR
   stimulation combined with IL-21 rescued the depolarized mitochondrial
   state in IgG1 cells ([183]Fig. 5D). The protective effect of IL-21 on
   mitochondrial dysfunction in IgG1 cells was further associated with a
   rescue of cell survival ([184]Fig. 5E). These results indicate that
   certain GC associated cytokines could at least partially protect
   against IgG1-induced mitochondrial dysfunction.

   Similarly, IgG1 lymphoma cells also showed enhanced mitochondrial
   depolarization in IgG1 cells in comparison to IgM cells upon BCR
   engagement ([185]Fig. 5, F and G). Although OCI-Ly7 cells were more
   sensitive to BCR-induced mitochondrial depolarization than BJAB cells,
   the IgG1 expressing cells consistently showed more mitochondrial
   depolarization upon BCR engagement than the IgM counterparts ([186]Fig.
   5, F and G). These studies demonstrate that IgG1 BCR signaling
   profoundly affects mitochondrial dynamics, leading to a depolarized
   mitochondrial state, which is associated with reduced mitochondrial
   mass and function in comparison to IgM BCR.

IgG1 BCR-induced calcium flux drives mitochondrial dysfunction and apoptosis

   A recent study showed that chronic BCR stimulation drives mitochondrial
   dysfunction in a calcium-dependent manner ([187]43). Hence, to test
   whether altered calcium dynamics could contribute to mitochondrial
   depolarization and reduced survival in IgG1 cells, we measured
   intracellular calcium flux downstream of BCR engagement. In primary B
   cells ([188]Fig. 6A) and all four isogenic lymphoma cell lines
   ([189]Fig. 6B and fig. S5, A to C) that we analyzed, BCR engagement of
   IgG1 receptors induced a higher and more sustained calcium flux than
   their IgM expressing counterparts. These results are consistent with a
   previous report that has shown enhanced calcium release upon IgG1 BCR
   engagement ([190]14). We next asked whether dampening of this enhanced
   calcium flux downstream to IgG1 BCR engagement could rescue defects in
   mitochondrial depolarization and survival. To interrogate this, we used
   a cell-permeant calcium chelator, BAPTA AM (BAPTA), to soak up
   intracellular calcium. For these studies, we used wild-type B cells
   activated under iGB conditions for 4 days, followed by BCR stimulation
   in the presence or absence of BAPTA ([191]Fig. 6C). As shown before,
   BCR stimulation led to enhanced mitochondrial depolarization in IgG1
   cells compared to IgM cells ([192]Fig. 6, D and E). In addition,
   treatment with calcium chelator BAPTA (2 μM) under BCR-stimulated
   conditions partially reversed the accumulation of dysfunctional
   mitochondria in IgG1 cells when compared with vehicle [dimethyl
   sulfoxide (DMSO)] control ([193]Fig. 6, D and E). The decrease in
   dysfunctional mitochondria upon BCR engagement and BAPTA treatment was
   accompanied by a rescue in cell survival of IgG1 cells in comparison to
   BCR engagement alone ([194]Fig. 6, F and G). In summary, these data
   demonstrate that IgG1 BCRs, in comparison to IgM BCRs, trigger a
   stronger calcium flux, which promotes mitochondrial depolarization and
   reduced cell survival ([195]Fig. 6H). Our findings also show that
   blunting of calcium responses through chelation of intracellular Ca^2+
   at least partially rescues mitochondrial function and survival in IgG1
   expressing B cells ([196]Fig. 6H).

Fig. 6. IgG1 BCR engagement triggers stronger calcium flux and impairs
mitochondrial function.

   [197]Fig. 6.
   [198]Open in a new tab

   (A) Calcium flux measured by fluorescence intensity of Fluo4 (y axis)
   over time in IgM and IgG1 IgH^γ1μ/+ splenic B cells in response to BCR
   (anti-Igκ) stimulation. (B) Calcium flux measured by fluorescence
   intensity of Fluo4 in IgM and IgG1 OCI-Ly7 lymphoma cells in response
   to BCR (anti-Igκ) stimulation. (C) Schematic of the iGB in vitro
   culture system with treatment groups. (D) Representative flow cytometry
   plots of IgM and IgG1 B cells derived from iGB in vitro cultures
   stained with MitoTracker Green (x axis) and TMRM (y axis) 12 hours
   posttreatment. (E) Quantification of dysfunctional mitochondria
   frequency in IgM and IgG1 B cells from iGB in vitro cultures (n = 6
   from two independent experiments). (F) Representative flow cytometry
   plots of cleaved caspase-3 from IgM and IgG1 B cells from iGB in vitro
   cultures 12 hours posttreatment. (G) Quantification of relative
   frequency of cleaved caspase 3 from IgM and IgG1 B cells from iGB in
   vitro cultures 12 hours poststimulation. Values are normalized to DMSO
   IgM and IgG1. (H) Proposed model showing that IgG1 B cell signaling
   triggers greater calcium flux, leading to increased mitochondrial
   dysfunction, which contributes to reduced cell survival. Dampening of
   calcium flux with BAPTA-AM rescues the survival of IgG1 expressing B
   cells. Statistical significance is calculated by repeated measures
   two-way ANOVA with Tukey’s multiple comparisons test [(E) and (G)].
   Error bars represent means ± SEM; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001;
   ****P ≤ 0.0001. Schematics in (C) and (H) created in BioRender. Shukla,
   V. (2025) [199]https://BioRender.com/1mfdxwl. See also fig. S5.

DISCUSSION

   Our studies here highlight how distinct BCR isotypes govern discrete B
   cell fates and could influence the pathogenesis of mature B
   cell–derived neoplasms. A previous study had linked distinct BCR
   isotypes to different patient outcomes in DLBCL ([200]9), and our
   studies provide mechanistic insights into how IgM and IgG1 isotypes
   might alter lymphoma pathogenesis. Although BCR signaling is known to
   be critical for the growth and survival of mature B cell malignancies,
   we identify important distinctions between IgM and IgG1 BCRs in
   lymphoma and normal B cells. The finding that mouse and human lymphoma
   cells expressing IgM outcompete those expressing IgG1 was unexpected to
   us, given that IgG1 BCRs signal more strongly and have been shown to
   better support proliferative responses in GCs ([201]13, [202]16,
   [203]23, [204]44). In addition to the increased proliferation of IgG1
   cells reported previously, we consistently observed reduced survival in
   IgG1 cells in comparison to IgM. On the basis of these results, we
   propose a model in which AID activity in GC-derived lymphomas may
   frequently induce class switching to IgG1 (and other non-IgM) isotypes;
   however, IgM expressing cells would outcompete those expressing IgG1.
   That said, under the “right” mutational background (for instance, in
   the EZB DLBCL), non-IgM receptors may be favored, and it will be
   interesting to further parse out the mutational signatures enriched in
   lymphomas expressing IgG1 BCRs. Our studies suggest that categorization
   of lymphoma patients based on BCR isotypes could add substantial
   prognostic value for predicting patient outcomes in combination with
   the current genetic classification. Of specific importance is the clear
   association of IgG1 expression to better patient outcomes in the cohort
   of patients with DLBCL, which remain unclassified (Other subtype) even
   when the most comprehensive genetic classifications are applied.

   GCB-DLBCL and ABC-DLBCL (and other B cell lymphomas) are proposed to
   rely on antigen-independent “tonic” and antigen-dependent “active” BCR
   signaling, respectively ([205]6, [206]26, [207]27). For most of our
   studies, we used relatively low concentrations (1 μg/ml) of anti-kappa
   light chain antibody to induce signaling. This allowed us to more
   easily ascribe functional changes in cell survival, proliferation,
   transcription, and mitochondrial function to active BCR signaling.
   Under these conditions, the IgG1 cells are clearly more sensitive to
   BCR engagement, leading to enhanced cellular turnover, profound
   transcriptional changes, and increased mitochondrial dysfunction
   compared to IgM expressing cells. In human DLBCL, the chronically
   activated BCR signaling is presumed to be driven by recognition of
   self-antigens ([208]27). Even in the TET-deficient model of B cell
   lymphoma, our previous study identified features of BCR-driven
   selection ([209]38). Furthermore, some of our studies do highlight that
   the effects of IgG1 BCR are also relevant in the context of tonic BCR
   signaling. For example, in human xenograft studies, we did not
   exogenously stimulate the BCR, and although we cannot rule out
   recognition of specific antigen(s) in the NSG mice, our data suggest
   that IgM BCR is favored over IgG1 even in this seemingly tonic BCR
   context. This is further supported by a study in which IgG1 BCRs are
   shown to differ from IgM under conditions of both tonic and active BCR
   signaling ([210]28). Moreover, patients with different DLBCL subtypes
   show heterogeneous clinical responses to BCR targeting therapies. For
   instance, a study observed only a 5% response rate among GCB-DLBCL
   treated with the Btk inhibitor, ibrutinib, compared with patients with
   ABC-type DLBCL that showed 37% response rate ([211]45). ABC-DLBCL is
   shown to be enriched for IgM BCR, whereas GCB-DLBCL displays IgG1 BCR
   expression ([212]9, [213]27). Given that IgM and IgG1 BCRs greatly
   differ in their signaling potential, it will be intriguing to examine
   whether IgM and IgG1 BCRs could mediate differential sensitivities and
   clinical responses to BCR targeting therapies.

   The IgM BCR is preferred over IgG1 during normal B cell development in
   the bone marrow ([214]14). The IgM BCRs are also favored for the
   generation of long-lived memory cells as memory cells expressing
   switched IgG BCRs are either purged over time due to high mutational
   burden or taken over by high avidity IgM B cells from the naive
   repertoire ([215]22, [216]24, [217]25). In contrast to these scenarios,
   the IgG1 BCRs are better at promoting plasma cell differentiation than
   IgM ([218]14, [219]22). Studies by Sundling et al. used an isotype
   switchable BCR transgenic mouse specific to hen egg lysozyme (HEL) and
   showed that HEL-specific GC B cells expressing IgG1 are selected over
   IgM ([220]44). However, removal of the cytoplasmic tail in IgG1 or
   fusion of the IgG1 cytoplasmic tail to IgM, which is responsible for
   enhanced BCR signaling, could still not reverse IgG1 selection over IgM
   in these studies ([221]44). With these results, the authors suggested
   that perhaps another aspect of IgG receptor structure (such as antigen
   capture), rather than BCR signaling itself, might drive IgG1 selection
   in the GCs ([222]44). Besides this explanation, the higher signaling
   potential and the lower activation threshold of the IgG1 BCR may
   provide a selective advantage to IgG1 cells over IgM under conditions
   where BCR specificities are restricted, and antigen is limiting. In
   such scenarios, additional mechanisms should exist to tolerate
   pro-apoptotic BCR signaling outputs of IgG1 receptors. In support of
   this premise, a recent report showed that, in GCs, IgG1 cells depend on
   transcription factor Miz1 and its downstream target, an anti-apoptotic
   protein Tmbim4, to safeguard against Ca^2+-induced mitochondrial
   dysfunction ([223]46). Similarly, the GC cytokine IL-21, produced by T
   follicular helper cells, enhances the levels of phospho-AKT and the
   mTOR signaling target, phospho-S6 downstream of BCR and CD40 ([224]42).
   mTOR activity is a known regulator of mitochondrial biogenesis
   ([225]47), and we speculate that the protective effect of IL-21 on IgG1
   survival observed in our studies is possibly an outcome of increased
   mitochondrial biogenesis. In alignment, another recent study showed
   that IL-21 better promotes the growth of IgG1 cells compared to IgM
   ([226]48). Therefore, in the context of the GC response, distinct
   protective mechanisms may exist to withstand the unwanted consequences
   of enhanced BCR signaling through IgG1. On the basis of our work and
   published literature, we postulate that IgG1 BCR would be favored in
   several physiological contexts, particularly under conditions where T
   cell help is available in the form of survival signals, such as IL-21.

   Unlike conventional GCs, survival signals such as IL-21 are likely
   limiting in the lymphoma microenvironment, and therefore IgG1 may not
   provide a strong selective advantage over IgM for malignant B cells.
   This advantage of IgM over IgG1 was generally true for most DLBCL
   genetic subtypes, except for EZB DLBCL, where IgG1 BCR showed no
   disadvantage when compared to IgM. We believe that this is likely due
   to the overexpression of anti-apoptotic protein BCL2 as a consequence
   of IGH-BCL2 translocations, which occur frequently in EZB DLBCL.
   Another important implication of our work is the link between distinct
   BCR isotypes and the unique metabolic features they impart on normal
   and malignant B cells. Our studies highlight that, due to their
   metabolic vulnerabilities, IgG1 expressing normal and malignant B cells
   may require additional survival signals, such as sustained T cell help
   (IL-21), for their expansion and persistence. In future work, it will
   be interesting to investigate how the lymphoma microenvironment differs
   between ABC-DLBCL and GCB-DLBCL subtypes or genetic subtypes of DLBCL,
   which show differential enrichment of IgM versus IgG BCRs.

   For our studies with human lymphoma cells, we primarily relied on
   CRISPR-Cas9–based editing to induce isotype class switching. One
   potential limitation of using this method would be that the stable
   expression of CRISPR-Cas9 could lead to additional mutations that may,
   in turn, affect cellular fitness. Therefore, to complement our studies
   with human lymphoma cells, we used primary murine B cells to confirm
   our findings in the context of normal B cell physiology wherever
   possible. In our studies with TET-deficient pre–malignant B cells, we
   also found that IgM cells outcompete IgG1. Nonetheless, our work
   warrants further investigation to carefully examine the role of
   different BCR isotypes of both restricted and diverse specificities in
   the context of normal GC and non–GC B cell responses, as well as
   autoimmune and malignant B cell pathologies, using additional genetic
   models.

   Although we focused on the role of IgG1 BCRs during lymphoma
   pathogenesis, GC-derived B cells lymphomas can express other IgG
   (subclasses) and IgA BCRs. Most studies characterizing the signaling
   differences between IgM and non-IgM BCR isotypes have mainly focused on
   IgG1 BCRs, and relatively less is known about the signaling properties
   of IgA, IgE, IgD, and other IgG subclasses, such as IgG2(s), IgG3, and
   IgG4. The differences in signaling potential of IgG1 to IgM are mainly
   attributed to its longer cytoplasmic tail ([227]12–[228]15, [229]17).
   In addition to a longer cytoplasmic tail, the IgG BCRs also harbor more
   flexible hinge regions, which are likely to support better antigen
   recognition. Furthermore, IgG1 BCRs are known to be less reliant on Ig
   α/β heterodimers for their signaling than IgM BCRs ([230]14). This
   could be an important distinction for lymphoma pathogenesis as
   gain-of-function mutations in the CD79B gene are common in the MCD
   genetic subtype, which are enriched in the ABC-DLBCL subgroup
   expressing IgM BCRs ([231]7, [232]31, [233]32). Together, IgG1 BCRs
   exhibit lower activation thresholds ([234]16), stronger BCR clustering
   ([235]13), and higher calcium responses ([236]14) and are associated
   with increased plasma cell differentiation ([237]23); these features of
   IgG1 were evident in our work. Similar features of enhanced calcium
   signaling and plasma cell differentiation have been reported for IgE
   and IgA BCRs ([238]18–[239]21, [240]49). It is worth pointing out that,
   in our in vivo studies with TET-deficient pre–malignant B cells, we
   observed populations of IgG2b-expressing and minor populations of IgA-,
   IgG3-, and IgD-expressing cells. IgG2 antibody responses are known to
   be directed toward polysaccharide bacterial antigens, and perhaps under
   conditions where BCR specificity is restricted to such antigens, IgG2
   might be a relevant isotype to investigate further ([241]15).

   Besides more pronounced calcium responses downstream to IgG1 BCRs, it
   is quite likely that other differences intrinsic to IgG1 and IgM BCRs
   may functionally contribute to distinct cellular phenotypes ([242]14,
   [243]50). Enhanced calcium flux is shown to have a detrimental effect
   on mitochondrial function in several cellular systems. Mitochondria
   function as important cellular buffers of calcium, and excessive
   calcium buildup in mitochondria can cause opening of the mitochondrial
   permeability transition pore (mPTP), leading to cell death ([244]51).
   Studies from Akkaya and colleagues demonstrated a role of BCR-induced
   calcium responses in driving mitochondrial dysfunction ([245]43), and
   our studies highlight key differences in calcium responses downstream
   of IgM and IgG1 BCRs. We show that calcium-induced mitochondrial
   dysfunction in IgG1 cells renders them metabolically less fit in
   comparison to IgM. We suspect that the enrichment of transcriptional
   changes associated with one-carbon metabolism upon IgG1 stimulation are
   an adaptation to this dysfunctional mitochondrial state. Previous
   studies have identified a higher dependency on one-carbon metabolism
   through serine catabolism in cells with mitochondrial dysfunction
   ([246]40, [247]41). One-carbon metabolism is critical for growth of
   cancer cells, which heavily rely on glycolysis to provide energy and
   fuel anabolic pathways ([248]52, [249]53). The precise role of
   one-carbon metabolism in B cells expressing distinct BCR isotypes is an
   intriguing question for future studies. The ability of IgM cells to
   outcompete their IgG1 counterparts in these studies is likely related
   to the metabolic stress experienced by IgG1 cells. Pathways associated
   with calcium signaling are commonly mutated in B cell lymphomas
   ([250]54), and additional work will be needed to elucidate how these
   mutations alter the metabolic landscapes in B cell lymphomas. In
   summary, we provide exciting evidence that IgM and IgG BCRs are
   inherently distinct in their activities and differentially control the
   pathogenesis of B cell lymphomas.

MATERIALS AND METHODS

Mice

   Tet2fl/fl and Tet3fl/fl mice were generated as previously described
   ([251]55, [252]56). C57BL/6J (000664), Cd19-cre (006785), CD45.1 mice
   (002014, Ptprca), and NOD/SCID IL-2Rγ^−/− (NSG) (JAX#005557) were
   obtained from the Jackson Laboratory. IgH^γ1μ mice ([253]14) were
   provided by A.W. (Institute for Molecular Medicine, University of
   Mainz) to R.C.R. Experiments were performed with mice 8 to 12 weeks of
   age of both sexes, and experimental mice were age and sex matched.
   Animals were bred and housed at the specific pathogen–free facility of
   Sanford Burnham Prebys Medical Discovery Institute and Northwestern
   University. All procedures were performed under approved protocols by
   the Institutional Animal Care and Use Committee (IACUC).

Primary B cell isolation and cell cultures

   The protocol for the in vitro iGB culture system was adapted from
   Haniuda and Kitamura ([254]57). 40LB feeder cells were irradiated with
   30 gray. Irradiated 40LB cells were plated at a density of 3 × 10^6
   cells per 10-cm dish and incubated overnight in Dulbecco’s modified
   Eagle’s medium (10-013-CV, Corning) supplemented with 10% fetal bovine
   serum (FBS) (35-016-CV, Corning), 10 mM Hepes (pH 7.4), 2 mM GlutaMAX,
   and 55 μM 2-mercaptoethanol (all from Life Technologies) at 37°C with
   5% CO[2] to allow for adhesion. 40LB cells were obtained from the D.
   Kitamura laboratory (Tokyo University of Science). Primary naive B
   cells were isolated from mouse spleen using the EasySep Mouse B Cell
   Isolation Kit (19854, STEMCELL Technologies) and seeded on the
   irradiated 40LB feeder layer at a density of 600K cells per 10-cm dish
   with rmIL-4 (1 ng/ml; 214-14, PeproTech) in RPMI 1640 medium
   (11875-093, Gibco) supplemented with 10% heat-inactivated FBS, 10 mM
   Hepes (pH 7.4), 2 mM GlutaMAX, 1 mM sodium pyruvate, and 55 μM
   2-mercaptoethanol (Life Technologies) at 37°C with 5% CO[2]. B cells
   were expanded for 3 or 4 days, depending on the experiment. To isolate
   IgM and IgG1 B cells from the culture, 40LB, IgE+, and CD138+ cells
   were depleted using biotinylated anti-H2Kd, IgE, and CD138 antibodies
   at a concentration of 1:200 with the EasySep Mouse Streptavidin
   RapidSpheres Isolation Kit (19860, STEMCELL Technologies).

   For stimulation experiments, the isolated IgG1 and IgM B cells were
   then stimulated for 12 hours with anti-kappa light chain antibody (1
   μg/ml; 75861S, Cell Signaling Technology) in supplemented RPMI as
   previously described at a density of –2 million cells/ml. For IL-21
   experiments, mouse IL-21 (210-21, PeproTech) was additionally added at
   a concentration of 10 ng/ml. For BAPTA experiments, mouse primary cells
   were treated with 2 μM BAPTA-AM (27-872-5, Tocris Bioscience) for 12
   hours in RPMI media.

Engineering of isogenic lymphoma cell lines, viral transduction, and cell
cultures

   Lentiviral vectors expressing Cas9 and guide RNAs were obtained from
   T.-C.C. and R.C. ([255]37). Lentivirus was produced by transfecting
   293T cells with pLenti lentiviral vectors targeting Sμ 5′ (2.5 μg) and
   Sy1 3′ (2.5 μg) and pMD2G (1.25 μg) and psPax2 (3.75 μg) packaging
   vectors. The lentiviral supernatants were added to BJAB, OCI-Ly7, and
   HBL-1 cells plated at a density of 200 × 10^3/ml of media and
   centrifuged at 2000 rpm at >20°C for 2 hours with polybrene (8 μg/ml).
   Transduced cells were selected starting at 48 hours posttransduction
   with puromycin (2 mg/ml) for at least 72 hours. Single clones were
   expanded to generate class-switched cell lines.

   Switching in Mino cells was induced using LPS (25 μg/ml), IL-4 (10
   ng/ml), and anti-CD40 (200 ng/ml) for 5 days. Single clones were
   expanded to generate class-switched cell lines.

   All cell lines were routinely monitored and immunophenotyped with
   anti-IgM and anti-IgG1 antibodies to ensure their BCR expression.
   Isogenic lymphoma cell lines were cultured in RPMI supplemented as
   described for primary B cell cultures at 37°C with 5% CO[2] for all
   experiments.

   For stimulation experiments, BJAB and OCI-Ly7 human lymphoma cell lines
   were stimulated with anti-kappa antibody (2060-01, Southern Biotech) at
   a final concentration of 5 and 1 μg/ml, respectively, in supplemented
   RPMI media at a density of 0.25 million cells/ml for 24 hours
   (mitochondria staining) or 48 hours (cleaved caspase 3 staining).

Transplantation experiments

   For Cd19 double knockout (DKO) transplantation studies, naive B cells
   from the spleen of Cd19 DKO mice were cultured using the described iGB
   culture system for 4 days. CD45.1 mice were sublethally irradiated with
   600 centigray 24 hours before transfer of 500,000 Cd19 DKO B cells
   through the retro-orbital sinus. For wild-type transplantation studies,
   naïve B cells were isolated from CD45.1 mice and cultured using the
   described iGB culture system for 4 days. C57/BL6 recipient mice
   (CD45.2+) were injected with cyclophosphamide (10 mg/kg) 24 hours
   before transfer of 9 million CD45.1+ B cells through the retro-orbital
   sinus.

Tumor engraftment

   For tumor engraftment studies with isogenic cell lines, 10^6 cells were
   injected subcutaneously into NSG mice. Mice health was monitored daily.
   Tumor volume was measured by (length × width^2/2 [cm^3]). Mice were
   euthanized when the total tumor volume reached 2 cm^3 or if they were
   in distress.

Immunizations

   For immunizations with SRBCs, citrated SRBCs (31102, Colorado Serum
   Company) were washed two times with sterile phosphate-buffered saline
   (PBS) and injected intraperitoneally at a dose of 200 × 10^6 SRBCs on
   day 0, followed by a boost on day 5 of 10^9 SRBCs. Mouse spleens were
   collected and analyzed on day 14.

Flow cytometry

   Primary cells and in vitro cultured cells were stained with appropriate
   antibodies in fluorescence-activated cell sorting (FACS) buffer [0.5%
   bovine serum albumin (BSA), 1 mM EDTA, and 0.05% sodium azide in PBS]
   for 30 min on ice. Cells were washed and fixed with 4% paraformaldehyde
   (PFA; Affymetrix) for 12 min on ice. For intracellular stains, cells
   were subsequently permeabilized with the eBioscience
   Foxp3/Transcription Factor Staining Buffer Set (00-5523, Invitrogen) at
   4°C for 2 hours to overnight before staining with appropriate
   antibodies for 45 min on ice. Flow cytometry experiments were analyzed
   using the Cytek Aurora spectral cytometer, BD Canto II, and BD LSR
   Fortessa X20 18-color analytical flow cytometer. Antibodies were from
   BioLegend, eBioscience, BD Pharmingen, Miltenyi, and Cell Signaling
   Technology, as listed in [256]Table 1. Data were analyzed with FlowJo
   10.9.0 software. Gating for flow cytometry analysis is described in the
   figure legends.

Table 1. Antibodies used for flow cytometry staining.

   Antibodies (clone) Company Catalog no.
   Biotin rabbit anti-active caspase-3 BD Biosciences 550557
   Anti-human IgG1 PE (IS11-12E4.23.20) Miltenyi Bio 130-119-946
   Anti-human IgG1 Bio (IS11-12E4.23.20) Miltenyi Bio 130-119-858
   lot no. 5220500500
   Anti-human IgM PB BioLegend 314513
   Biotin anti-mouse IgE BioLegend 406903
   Biotin anti-mouse CD138 BioLegend 142511
   Biotin anti-mouse H2kd BioLegend 116604
   Anti-mouse IgM APC BioLegend 406509
   Anti-mouse IgM BV711 BioLegend 406539
   Anti-human kappa antibody Southern Biotech 2060-01
   Anti-mouse kappa antibody Cell Signaling Technology 75861S
   LEAF purified anti-mouse CD40 antibody BioLegend 102810
   Anti-mouse IgG1 PE/Cy7 BioLegend 406613
   Anti-mouse Ki67 BV 421 BD Biosciences 562899
   PE streptavidin BioLegend 405203
   Anti-mouse IgD AF-700 BioLegend 405730
   Anti-mouse IgA AF-647 Southern Biotech 1040-31
   Anti-mouse IgG2b PE BioLegend 406707
   Anti-mouse IgG3 PE/Cy7 Southern Biotech 1100-17
   Anti-mouse FAS PE/Cy/7 BD Biosciences 557653
   Anti-mouse IgG1 BV421 BioLegend 406616
   AF-647 streptavidin BioLegend 405237
   Anti-mouse IgG1 APC/Cy7 BioLegend 406619
   Anti-mouse IgG1 PE BioLegend 406608
   Anti-mouse CD138 PE/Cy7 BioLegend 142513
   Anti-mouse CD19 PB BioLegend 115526
   eBioscience Fixable Viability Dye eFluor 450 Thermo Fisher Scientific
   65-0863-18
   Anti-mouse B220 BV650 BioLegend 103241
   Anti-mouse CD19 BV510 BioLegend 115546
   Anti-mouse CD45.1 APC/Cy7 BioLegend 110716
   Anti-mouse CD19 FITC BioLegend 115506
   Streptavidin PerCP/Cy5.5 BioLegend 405214
   Anti-mouse IgG1 FITC (A85-1) BD Biosciences 553443
   Anti-mouse IgM PE/Cy7 (II/41) eBioscience 25-5790-82
   Tomm20 (D8T4N) Cell Signaling Technology 42406S
   Anti-mouse IgK (187.1) BD Pharmingen 559750
   Anti-mouse CD38 PE (90) BD Pharmingen 553764
   Anti-mouse IgM FITC (II-41) eBioscience 11-5790-82
   Active/Pro-Caspase-3 AF647 (C92-605) BD Biosciences 560626
   Anti-human IgM PE (PJ2-22H3) Miltenyi Bio 130-122-930
   Anti-human IgG1 APC (#IS11-12E4.23.20) Miltenyi Bio 130-119-857
   Anti-mouse CD45.1 PECy7 BD Pharmingen 560578
   Anti-mouse B220 APC (RA3-6B2) BD Pharmingen 553092
   Anti-mouse CD45.2 Alexa Fluor 700 (104) BioLegend 109822
   Anti-mouse CD45.2 APCCy7 BD Pharmingen 560694
   [257]Open in a new tab

Measurement of mitochondrial mass and membrane potential

   For mitochondrial mass staining, cells were incubated with 50 nM
   MitoTracker Green FM (7514, Invitrogen) in culture media for 25 min at
   37°C. Cells were washed twice with FACS buffer (described above) and
   rapidly run on a flow cytometer. For experiments with primary cells,
   cells were subsequently stained with surface markers for 30 min and
   then rapidly run on a flow cytometer. For mitochondrial mass and
   membrane potential staining, cells were simultaneously incubated with
   50 nM MitoTracker Green FM and 500 nM tetramethylrhodamine, methyl
   ester, perchlorate (TMRM) (T668, Invitrogen) following the same
   protocol. Dysfunctional mitochondria population was gated on TMRM low
   and MitoTracker Green–positive cells.

Cell proliferation assays

   The BrdU incorporation assay (552598, BD Biosciences) was used to
   measure proliferation in the isogenic lymphoma cell lines. A total of 1
   × 10^6 cells from each BJAB or OCI-Ly7 cell line were incubated with
   BrdU for 25 or 15 min, respectively, at a concentration of 10 μM in
   supplemented RPMI. Following incubation, cells were fixed and
   permeabilized according to kit instructions and then treated with 30 μg
   of DNase in PBS for 1 hour. Cells were stained with anti-BrdU APC
   antibody for 30 min and then run on a flow cytometer. Data were
   analyzed with FlowJo 10.9.0 software. BrdU+ gating was used to
   calculate the frequency of cells in the S phase.

ImageStream X

   Cells from the spleen of SRBC-immunized mice were stained and processed
   as described above for intracellular staining. Samples were run on an
   Amnis ImageStreamX Mark II imaging flow cytometer at the Sanford
   Burnham Prebys Flow core facility. Data were analyzed using Inspired
   acquisition software.

Confocal microscopy and analysis

   IgM and IgG1 expressing cells from IgH^γ1μ mice were incubated on glass
   slides precoated with poly-l-lysine for 15 min. Cells were stained with
   1:100 anti-kappa antibodies for 25 min. Cells were washed twice and
   fixed with 4% PFA for 10 min at room temperature (RT), followed by two
   washes with PBS for 5 min. Cells were next permeabilized with 0.5%
   Triton X-100 for 10 min. Permeabilized cells were washed again two
   times with PBS for 5 min and then blocked in 500 μl of 1% BSA with PBS
   for 30 min. The blocking solution was replaced with 250 μl of Tomm20
   (Cell Signaling Technology), 1:500 diluted in 1% BSA, and incubated
   overnight at 4°C. After three washes, cells were incubated in secondary
   antibody anti-rabbit (PE) 1:2000 in 1% BSA for 1 hour at RT. Cells were
   counterstained with 4′,6-diamidino-2-phenylindole (DAPI) for 10 min and
   then washed three times with PBS. A drop of antifade fluorescence
   mounting medium was added, and a coverslip was placed. All steps were
   performed in a humidified chamber at 37°C unless otherwise specified.
   Confocal images were acquired with a Leica SP8 confocal microscope and
   analyzed in Leica Application Suite X software v3.1.5.16308. Mean
   fluorescence intensity was quantified using Fiji6.

Calcium flux assay

   Cells were stained with Ca^2+ indicator Fluo-4 ([258]F14201, Thermo
   Fisher Scientific). Briefly, cells were prepared according to the
   manufacturer’s guide in HBSS (Hanks’ balanced salt solution) with 1x
   Hepes, and an equal amount of 2X Fluo-4 was added to the cells. Cells
   were then incubated for 30 min at 37°C in a water bath. Fluorescence
   intensity of Fluo-4 was analyzed by flow cytometry using a BD Canto II
   flow cytometer. Calcium influx was initiated by injection of anti-kappa
   antibody (5 μg/ml). Cells were kept in a water bath throughout the
   assay unless acquiring the data.

Seahorse

   OCR was measured using the XF24 extracellular flux analyzer (Agilent)
   according to the manufacturer’s guidelines. Briefly, cells were
   harvested and plated at concentrations of 500,000 cells per well (BJAB
   cell line) or 1.5 million (primary IgH^γ1μB cells) in Seahorse medium
   with 2 mM l-glutamine and 1 mM sodium pyruvate. Cells were then spun at
   300g for 5 min with brakes off to create a monolayer with no gaps in
   XF24 plates. Injection of compounds occurred at the time points
   indicated in figures. Changes in OCR measured over time in IgM and IgG1
   IgH^γ1μ/+ splenic B cells treated with oligomycin, phenylhydrazone
   (FCCP), and antimycin A and rotenone. Changes in OCR measured overtime
   in IgM and IgG1 BJAB lymphoma cells treated with anti-Igκ (5 μg/ml),
   oligomycin, FCCP, and antimycin A and rotenone.

TCGA analysis

   We downloaded 481 preprocessed tsv files containing the transcriptomic
   profile, in transcripts per million (TPM) values of patients with DLBCL
   from TCGA (Project ID: NCICCR-DLBCL). We parsed the files on the basis
   of their IGHM and IGHG1 mRNA expression and identified 111 samples with
   high IGHM (IGHM TPM representing >50% of all IGH transcripts) 63
   samples with high IGHG1 (IGHG1 TPM representing >20% and IGHM TPM no
   more than 20% of all IGH transcripts). We applied these filtering
   criteria taking into account that IGHM are default transcripts
   expressed on all B cells and will therefore be highly represented in a
   mixed pool (lymphoma and contaminating naïve cells) of B cells used in
   these studies.

RNA sequencing

   For RNA-seq experiments, naive B cells were isolated from wild-type
   C57BL/6J mice spleen and cultured using the in vitro iGB culture system
   and stimulated with anti-kappa antibody (1 μg/ml) for 12 hours as
   previously described. After stimulation, B cells were FACS sorted from
   live cells, CD19+ and CD138− into IgG1+ and IgM+ using BD FACSAria 5L
   Sorter at the Northwestern University RHLCCC Flow Cytometry Facility.
   Total RNA was isolated using the RNeasy Plus Kit (74134, Qiagen)
   according to the manufacturer’s instructions. Quality of RNA was
   assessed using the Agilent Bioanalyzer RNA 6000 Pico Chip Bioanalyzer.
   The RNA-seq libraries were prepared using purified RNA (40 ng) with the
   NEBNext Ultra II RNA Library Prep Kit (E7770, New England Biolabs)
   according to the manufacturer’s instructions. Samples were sequenced on
   the Illumina NovaSeq 6000 platform [50–base pair (bp) reads].

RNA-seq analysis

   Paired-end (50 bp) reads were mapped to the mouse genome mm10/GRCm38
   using STAR ([259]58) (2.7.10a) (--genomeLoad LoadAndRemove
   --outFilterMismatchNmax 4 --outFilterMultimapNmax 100
   --winAnchorMultimapNmax 100). Counts were obtained with featureCounts
   ([260]59) (subread v2.0.1) (-g gene_name -s 0). DEGs were calculated
   with DESeq2 ([261]60), filtering out genes that did not have any count
   under any condition; cutoff to define DEGs was an adjusted P value of
   0.05 and a log[2] fold change of ≥±1. The enrichment processes and
   pathway analysis of DEGs in gene ontology function was performed using
   Metascape ([262]http://metascape.org).

Statistical analysis

   Statistical analyses were performed with GraphPad Prism 10.1.0. The
   statistical tests used to determine significance are described in the
   corresponding figure legends. Data are represented as means ± SE. No
   statistical methods were used to predetermine sample sizes. Parametric
   tests [t tests, analysis of variance (ANOVA)] were used in experiments
   where normal distribution could be assumed. No data points were
   excluded from the analysis, and appropriate animals/samples for each
   experiment were chosen randomly.

Acknowledgments