Flawed translation triggers oncogenic B–T cell communication

B cell lymphomas originate from defined stages of B cell differentiation after the acquisition of numerous genetic and epigenetic alterations. Most lymphomas arise from germinal center (GC) or post-GC B cells, due to several oncogenic features of the unique GC environment. The acquired alterations favor survival, proliferation, and immune escape of the developing lymphoma cells, often by affecting the cross talk to cells in their microenvironment (Weniger et al., 2025). A key population in the GC microenvironment is T follicular helper (TFH) cells, a specialized CD4 T cell subset that supports B cell responses and differentiation (Ueno et al., 2015). CD4 T and especially TFH cells play important roles in sustaining human lymphomagenesis, for example, in follicular lymphoma (FL) (Pangault et al., 2010; Amé-Thomas et al., 2012), B cell chronic lymphocytic leukemia (Os et al., 2013), and early gastric mucosa-associated lymphoid tissue lymphoma (Hussell et al., 1996).

Lin et al. (2026) made an intriguing finding in this context while investigating the eIF3e component of the eIF3 translation initiation factor complex by specifically ablating it in murine GCB cells. For this purpose, they employed the Cγ1Cre knock-in model, which initiates excision of loxP-flanked eIF3e alleles in early GCB cells that received cognate CD4 T cell help via CD40L–CD40 interactions and cytokines including IL-4 (Casola et al., 2006). Loss of eIF3e led to highly prevalent clonal lymphomagenesis in naive mice starting at 8 months of age. The majority of the lymphomas were of B cell origin and expressed the GC markers Fas and GL-7, while some derived from T cells, which were not further investigated. Surprisingly, all lymphomas retained eIF3e expression, even though its loss constituted the lymphoma-initiating genetic alteration.

To resolve this conundrum, Lin et al. monitored the effects of GCB cell–specific eIF3e deficiency over time. Under steady-state conditions in specific pathogen–free animal facilities, naive mice contain only very small numbers of sporadically arising GCs. The first striking discernible change detected by Lin and colleagues was a progressive increase in the numbers of eIF3e-deficient pre-GCB cells (tracked by a YFP reporter), paralleled by an increase in TFH-like CD4 T cells from 8 to 16 weeks of age. These TFH cells overproduced IL-4 already at 8 weeks of age, reminiscent of human FL, as high expression of IL-4 was the most pronounced difference between FL-derived and normal tonsillar TFH cells. In addition, autologous FL-derived TFH cells as well as CD40L/IL-4 treatment enhanced FL cell survival in vitro (Amé-Thomas et al., 2012) and vicinity to TFH cells correlated with evidence for active IL-4 signaling in FL cells on patient-derived sections (Pangault et al., 2010). The authors then conducted a series of elegant in vivo bone marrow chimera experiments, ex vivo cell culture assays, RNA sequencing, and mass spectrometry–based proteomics, designed to distinguish cell-intrinsic from cell-extrinsic effects and to gain mechanistic insights into the underpinnings of the lymphoproliferative disorder preceding lymphomagenesis. Together, the results strongly suggested the following scenario: loss of eIF3e in pre-GCB cells led to enhanced levels of proteins mediating B–T cell cross talk, including MCHII pathway and co-stimulatory proteins. Of the latter, an enhanced translation efficiency in absence of eIF3e was suggested for ICOSL, which is essential for driving TFH and GC responses via its receptor ICOS (Ueno et al., 2015). On the other hand, several genes regulating the cell cycle had reduced translation efficiency in eIF3e-deficient cells. In sum, the direct and indirect effects of eIF3e-deficiency in GCB cell precursors triggered the instruction of strongly IL-4–producing TFH cells, which in turn initiated GC differentiation and Cre-expression in more B cells, followed by eIF3e ablation, resulting in a feed-forward cycle. However, loss of eIF3e also reduced proliferation and increased cell death, leading to the virtual absence of eIF3e-deficient mature GCB cells. The chronic progressive B–T cell cross-activation ultimately resulted in lymphomagenesis of B cells that somehow escaped Cre-mediated recombination and, less frequently, of T cells. Whether the observed clonal T cell expansions resemble human TFH-derived angioimmunoblastic T cell lymphoma remains to be elucidated.

The present study underscores the highly oncogenic potential of aberrant B–T cell interactions, which here drive lymphomagenesis in unmodified bystander B cells in more than half of the animals within 13 months. This is remarkable, as Cγ1Cre-mediated induction of major GCB-derived lymphoma-associated oncogenic alterations, including overexpression of c-Myc (Sander et al., 2012), or loss of Blimp1 (Mandelbaum et al., 2010) or of Kmt2d (Zhang et al., 2015), did not lead to overt lymphomas up to 13 months of age and longer. The identity and acquisition sequence of oncogenic alterations during bystander B cell lymphomagenesis in the model of Lin et al. could yield valuable insights, linked to the assessment of whether and when the developing lymphomas become independent of T cell help.

Given the progressively enhanced B–T cell communication mediated by increased antigen presentation capabilities, elevated surface levels of co-stimulatory molecules and cellular cross-activation, the nature of the antigens involved in this interaction becomes a key question. In this context, recent elegant work by the Bogen laboratory demonstrated that the presentation of idiotypic (B cell receptor V region–derived) peptides, a phenomenon relevant to human lymphoma (Khodadoust et al., 2017), by a small number of B cells induces autoimmunity and subsequent B and T lymphomagenesis (at 12–22 months of age), when T cells recognizing the idiotypic peptide were also present (Gopalakrishnan et al., 2026). This links to the question of whether and which autoantigens, possibly including idiotypic peptides, are presented by the eIF3e-deficient B cells to T cells, thereby triggering their differentiation to highly IL-4–producing TFH cells. Autoimmune phenotypes were not reported by Lin and co-workers, which might relate to the experimental system, where class-switching B cells ablate eIF3e, leading to reduced proliferation and cell death, precluding the production of (high-affinity) autoantibodies. The generation of chimeras reconstituted with eIF3e^fl/fl Cγ1Cre and wild-type bone marrow, where the induced TFH cells would activate wild-type B cells without inducing eIF3e ablation, could help resolve this issue.

It is tempting to speculate that the death of eIF3e-deficient pre-GCB cells contributes to providing autoantigenic stimulation. eIF3-deficient pre-GCB cells have superior antigen-presentation and co-stimulation capabilities compared with wild-type pre- and mature GCB cells, possibly enabling them to also activate rare and presumably low-affinity autoantigen-specific T cells that escaped central tolerance. Such T cells could be stimulated by B cell–derived endogenous, maybe aberrantly expressed autoantigens, including idiotypic peptides and/or autoantigens picked up from debris derived from other dying eIF3e-deficient B cells. Once the process is started, the activated T cells induce the progressive induction of more eIF3e-deficient pre-GCB cells, their subsequent death, and further release of autoantigens culminating in lymphoproliferative followed by malignant disease. Additional or alternative, not necessarily mutually exclusive mechanisms, including the participation of microbial antigenic triggers in the context of the hygiene status of the animal facility and other antigen-presenting cells, are also possible.

A further interesting aspect relates to the polarization of the B cell–driven T cell response. The eIF3e-deficient pre-GCB cells trigger the differentiation of potently IL-4–producing TFH cells that strongly amplify B cell differentiation toward a GC fate. The underlying mechanisms were linked to increased MHCII expression and co-stimulation. However, the same principles were reported to facilitate B cell–induced cytotoxic T cell differentiation in the context of the expression of the Epstein-Barr virus–derived latent membrane protein 1, a highly potent constitutively active CD40 mimic or of amplified endogenous CD40 signals through loss of TNFAIP3/A20 and gain of the NF-κB–inducing kinase. In both cases, the induced cytotoxic T cells sufficiently eliminated the instructing, aberrantly activated B cells to prevent lymphomagenesis for the lifetime of the investigated organisms (Choi et al., 2018; Diehl et al., 2024). It will be interesting to decipher, which elements on the respective activated B cells drive helper versus cytotoxic T cell differentiation. One contributing candidate to these differential outcomes might be CD70, which is highly upregulated in cytotoxicity instructing (Choi et al., 2018; Diehl et al., 2024) but only marginally in the high IL-4–expressing TFH-instructing eIF3e-deficient B cells.

The lymphoma model of Lin and colleagues most likely depends on the continuous and progressively increasing Cre-loxP–mediated generation of eIF3e-deficient pre-GCB cells generated from a pool of naive B cells, which in turn are replenished from the bone marrow. In humans, a constant supply of mutant cells could in principle be provided through mutations in hematopoietic stem and progenitor cells. However, EIF3E mutations have not been reported in human lymphoma, and this scenario would lack the oncogenic feedforward cycle. In addition to implicating eIF3e as a major regulator of B cell activation, this study therefore establishes eIF3e^fl/fl Cγ1Cre mice as a valuable novel model for the mechanistic dissection of lymphoma-inducing B–T cell interactions in the context of chronic antigenic stimulation under inflammatory conditions.