Stone2019 - T-bet Promotes ASC Differentiation by Limiting IFN-gamma Inflammation

Full citation: Stone, S. L., Peel, J. N., Scharer, C. D., Risley, C. A., Chisolm, D. A., Schultz, M. D., Yu, B., Ballesteros-Tato, A., Wojciechowski, W., Mousseau, B., Misra, R. S., Hanidu, A., Jiang, H., Qi, Z., Boss, J. M., Randall, T. D., Brodeur, S. R., Goldrath, A. W., Weinmann, A. S., Rosenberg, A. F., & Lund, F. E. (2019). T-bet transcription factor promotes antibody-secreting cell differentiation by limiting the inflammatory effects of IFN-γ on B cells. Immunity, 50(5), 1172–1187.e7. https://doi.org/10.1016/j.immuni.2019.04.004

Note: the file in raw/ is the accepted in-press version (self-paginated “Immunity 50, 1–16”); the final page range above is taken from the queue record in state.md and has not been independently re-verified here. CrossRef was unreachable during this ingest, so citations_crossref is null; the pending refresh should confirm the page range at the same time.

Raw file: [[raw/stone2019.pdf]]

Summary

This is the wiki’s mechanism paper for what T-bet actually does in a B cell committing to the antibody-secreting-cell (ASC) fate — and it is a negative result at its core. The authors set out expecting T-bet to drive ASC differentiation by inducing the canonical plasma-cell transcription factors. It does not. T-bet does not regulate the early expression or activity of Blimp-1, IRF4, XBP1 or Pou2af1, and it does not repress the B-cell-identity factors (Pax5, Bcl-6) whose removal ASC commitment requires. What T-bet does instead is repress the IFN-γ-induced inflammatory gene programme — NF-κB, TLR and STAT/IRF networks — which, if left running, locks an activated B cell into an inflammatory effector state that is incompatible with terminal differentiation.

The system is the Lund lab’s Be1/Be2 co-culture: naive mouse B cells cultured with polarised Th1 (→ Be1) or Th2 (→ Be2) effectors. The dissection separates three requirements cleanly. Blimp-1 is required (Cd19^cre/+^Prdm1^fl/fl^ Be1 cells fail to form ASCs). IFN-γR is required (Ifngr1^−/−^ Be1 cells fail to upregulate Prdm1 and form no ASCs). T-bet is required for ASC output but is not the route by which IFN-γ induces Prdm1 — deleting Tbx21 leaves day-2 Prdm1 expression and Blimp-1-motif chromatin accessibility untouched while abolishing ASC formation downstream. Sustained NF-κB or TLR7/9 signalling in wild-type Be1 cultures reproduces the Tbx21^−/−^ defect, establishing the inflammatory programme as the causal brake rather than a correlate.

In vivo, B-cell-intrinsic T-bet is required for long-lived bone-marrow ASCs and IgG2c after influenza but is dispensable after Heligmosomoides polygyrus (an IL-4-dominated Th2 infection) — so T-bet’s role is contingent on the cytokine environment, not universal. Inducible deletion of Tbx21 from established memory B cells leaves the memory pool intact but cuts NP-specific ASC recall roughly 10-fold, dissociating memory maintenance from memory differentiation. The authors’ own summary position is that T-bet, unlike Blimp-1 or IRF4, “is not a master regulator of ASC development.”

Study Design

• Type: Mouse in vitro T–B co-culture + epigenomics (ATAC-seq) + transcriptomics (RNA-seq, microarray) + two in vivo infection models with conditional and inducible gene deletion
• Sample size: ≥2–4 independent experiments per in vitro panel; ATAC-seq 3 independent samples/group; RNA-seq 3 samples/group/timepoint; microarray 7 samples/group/timepoint; in vivo 3–6 mice/group (5–8 for some panels), 5 independent experiments for ELISPOT/titre panels
• Setting: In vitro Be1/Be2 allogeneic co-culture (days 0–4); in vivo influenza A/PR/8/34 (PR8) primary infection, A/Aichi/68 (X31) heterologous challenge, and Heligmosomoides polygyrus (Hp) nematode infection
• Population: Mouse. C57BL/6, Tbx21^−/−^, Ifngr1^−/−^, Cd19^cre/+^Prdm1^fl/fl^, Tbx21^fl/fl^hCD20-TAM-cre, B6.Blimp1-YFP and B6.T-bet-ZsGreen reporters, and μMT bone-marrow chimeras (B-WT vs B-Tbx21^−/−^)
• Cell definitions: ASC = CD138⁺CD93⁺ (in vitro); FOB = CD19⁺CD38⁺PNA^lo^; GCB = CD19⁺CD38^lo^PNA^hi^; BMEM = NP⁺CD38^hi^IgD^neg^
• Data availability: GEO GSE84948 (microarray), GSE83697 (RNA-seq), GSE118984 (ATAC-seq)

Key Findings

T-bet does not drive the plasma-cell transcription factor programme

• The Blimp-1 chromatin dissociation — the paper’s cleanest internal control. Chromatin accessibility in the 100 bp surrounding Blimp-1 binding motifs (n=871 motif-containing DARs) is strongly opened by Th1 priming (day 2 B6 Be1 vs B6 Be2, p=3.8 × 10⁻⁹⁰, Fig. 1K) but is unchanged when T-bet is deleted (B6 Be1 vs Tbx21^−/−^ Be1, ns, Fig. 3G) — while T-bet’s own motifs (n=963) are strongly differential in the same panel. The same assay, same motif set, opposite answers for the two perturbations.
• Prdm1 mRNA is equivalent between day-2 B6 and Tbx21^−/−^ Be1 cells (Fig. 3H). By day 4 it is lower without T-bet — i.e. T-bet indirectly supports Blimp-1 at the point of differentiation and never represses it.
• The authors state the negative result explicitly: “Although we fully expected to find that T-bet induced B cell differentiation by promoting TFs that initiate ASC commitment, we realized that T-bet did not regulate the early expression or activity of any of the well-described ASC-associated TFs (Nutt et al., 2015) including Blimp1, which can be modulated in a T-bet-dependent fashion in T cells (Oestreich et al., 2012; Xin et al., 2016). Likewise, T-bet did not function to repress TFs like Pax5 that maintain B cell identity and prevent ASC differentiation.” (Discussion, p.13.) Both citations attached to the T-bet→Blimp-1 modulation claim are T cell papers.
• Blimp-1 itself is required: Cd19^cre/+^Prdm1^fl/fl^ Be1 cells drop from 7.9% to 1.7% CD138⁺CD93⁺ ASC, with antibody secretion falling to near zero (Fig. 1L–1N).
• IFN-γR is required, and it is the route to Prdm1: Ifngr1^−/−^ Be1 cells show significantly lower Prdm1 at day 2, keep Prdm1, Irf4, Pou2af1 and Xbp1 low through day 4, form almost no ASC (15.3% → 0.6%) and secrete no antibody (Fig. 2I–2M). The IFN-γ → PRDM1 axis is therefore T-bet-independent.

The actual mechanism: T-bet as a repressor of the inflammatory programme

• T-bet is a transcriptional repressor in Be1 cells. >2,000 DEGs separate day-4 Tbx21^−/−^ from B6 Be1 cells; 611 DARs separate day-2 Be1 from Be2 (FDR<0.05, Fig. 1J), and 561 DARs separate B6 Be1 from Tbx21^−/−^ Be1 versus only 30 for the Be2 comparison (Fig. 3F) — T-bet’s chromatin role is Th1-context-specific.
• Tbx21^−/−^ Be1 cells maintain an activated inflammatory signature that wild-type Be1 cells normally extinguish by day 4: sustained STAT1/IRF3/IRF7 targets, sustained NF-κB family members, and increased TLR and TNFR family receptors and ligands (Fig. 4).
• Causal test — the brake is sufficient. Adding an NF-κB activator (betulinic acid) or TLR7/TLR9 ligands (R848, CpG) to wild-type B6 Be1 cultures from day 2 significantly suppressed ASC formation and IgG-secreting cell numbers (Fig. 5A–5D). This was not a proliferation artefact — cells expanded equally well.
• T-bet represses Ifng and Ifngr2, breaking an autocrine inflammatory loop. Both loci carry T-bet-motif-containing DARs (Fig. 5I–5M). Tbx21^−/−^ Be1 cells produced more IFN-γ and IL-6 after TLR7/9 + anti-CD40/anti-Ig stimulation (Fig. 5N–5O).
• Overlap with the T cell T-bet repressome is partial: of 1,375 T-bet-repressed genes in Be1 cells, only 155 are shared with the 275 T-bet-repressed genes in Th1 cells (Iwata et al., 2017); 1,220 are B-cell-unique (Fig. 5E). The B cell programme is not a copy of the T cell one.

In vivo: cytokine-context dependence and the memory dissociation

• Required for flu, dispensable for nematode. In μMT bone-marrow chimeras, B-cell-intrinsic T-bet was required for flu-specific IgG, bone-marrow ASC and IgG2c responses after PR8, but Hp-specific IgG and IgG1 responses were similar in both groups (Fig. 6I–6O). T-bet is a cytokine-context-dependent, not universal, ASC factor.
• Isotype specificity: NP-specific IgG2c antibody and IgG2c⁺ memory B cells were missing in B-Tbx21^−/−^ mice, while NP-specific IgG2b antibody and IgG2b⁺ memory B cells were intact (Fig. 6O, 6P–6R). The authors attribute the IgG2b escape to TGF-β-driven switching, which suppresses IFN-γ signalling.
• Memory maintenance and memory differentiation dissociate. Tamoxifen-induced deletion of Tbx21 from established day-90 flu memory B cells (Tbx21^fl/fl^hCD20-TAM-cre) did not change memory B cell numbers over a 10-day period, but after X31 challenge NP⁺ ASC numbers fell ~10-fold versus challenged controls (Fig. 7F, 7I, 7M).
• CXCR3 is a functional T-bet target on memory B cells: CXCR3 expression fell on total B cells and NP⁺ memory B cells after inducible T-bet deletion, while T cell CXCR3 was unaffected (Fig. 7B–7E, S7C–S7D).

T-bet expression across ASC compartments — an anatomical gradient

• T-bet reporter is expressed by GC B cells, ASCs and memory B cells after flu (Fig. 6C) — T-bet⁺ B cells are not GC-excluded.
• Splenic ASCs are enriched for the T-bet reporter; bone-marrow long-lived ASCs are not. At 60 dpi, splenic ASCs were 32.5 ± 19.4% Tbx21-ZsGreen⁺ versus 9.6 ± 2.5% for splenic B cells, whereas bone-marrow ASCs were only 0.87 ± 0.8% versus 1.11 ± 0.7% for B cells (Fig. 6D–6E). Most flu-specific memory B cells in the draining LN continued to express ZsGreen and intracellular T-bet protein at 60–90 dpi (Fig. 6F–6H).

Methods Used

ATAC-seq, RNA Sequencing, In Vitro B Cell Stimulation, T-B Coculture Assay, ELISpot

Also used but not separately updated: Conventional Flow Cytometry (including Tbx21-ZsGreen and Blimp1-YFP reporter readout and intracellular T-bet staining), FACS Sorting (sort-purified day-4 Be1 subsets by CD138/CD93), ELISA (flu and Hp antibody titres), Affymetrix microarray, quantitative RT-PCR, GSEA, HOMER motif analysis and Ingenuity Pathway Analysis upstream-regulator prediction.

Entities Mentioned

T-bet, BLIMP-1, IFN-gamma, IRF4, CXCR3, Plasmablast, Atypical B Cell, DN2 B Cell, TLR7

Discussed but not separately updated: XBP1 and Bcl-6 (measured by qPCR; unaffected by T-bet early, low in Ifngr1^−/−^), TLR9 (CpG ODN1826 used alongside R848 in the ASC-suppression experiments), CD138 and CD38 and IgD (ASC and memory gating markers), IgG (IgG2c/IgG2b isotype split), Type I Interferon (T-bet prevents an Ifna autocrine circuit), Age-Associated B Cell and Switched Memory B Cell (framing only).

Concepts Addressed

Atypical B Cell Effector Output, Toll-like Receptor Signaling in B Cells, Germinal Center, Class Switch Recombination, Memory B Cell

Relevance & Notes

This paper resolves the wiki’s longest-standing flagged assertion, in the opposite direction from how it was recorded. Cancro2020 - Age-Associated B Cells asserted, without a supporting reference, that “there is evidence that T-bet represses Blimp-1.” The wiki flagged this on T-bet and BLIMP-1 as author synthesis requiring verification. Stone tested exactly that relationship in B cells and found no repression — and the two citations Stone attaches to the modulation claim (Oestreich 2012, Xin 2016) are T cell papers, which is the most likely provenance of Cancro’s unreferenced sentence. The flag is now upgraded from “unsourced” to “contradicted in B cells by primary data, with probable T-cell provenance identified.”

The second half of Cancro’s sentence fares better than the first. Cancro also wrote “few if any plasma cells express T-bet,” concluding that PC formation from ABCs “likely involves the loss of T-bet expression.” Stone’s reporter data make this an anatomical gradient rather than a binary: splenic ASCs are enriched for the T-bet reporter (32.5% vs 9.6% of B cells) while bone-marrow long-lived ASCs are not (0.87% vs 1.11%). Cancro’s conclusion is roughly right for the terminal bone-marrow compartment and wrong at the splenic ASC stage. Since the curator’s cohort work reads peripheral blood only, this gradient is a caution against extrapolating a single T-bet/ASC relationship across compartments.

It offers a proposed reconciliation of the Sutton2021 ↔ Jenks2018 contradiction, which remains formally open. If T-bet’s role is permissive (removing an inflammatory brake) rather than instructive (inducing the PC programme), then both wiki-tracked observations hold simultaneously: Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection finds no PC-programme genes in resting atypical B cells because T-bet does not induce them, and Jenks2018 - DN2 B Cells and EF Pathway in SLE finds DN2 cells are efficient plasmablast precursors on stimulation with the PRDM1 locus open because the brake has been lifted. This is a mechanism proposed, not a contradiction closed — Stone is mouse Be1/Be2 culture plus influenza, and does not touch human atypical B cells. The human observations on both sides still stand unreconciled at the level they were made.

Relation to Zumaquero2019 - IFN-gamma Programs T-bet-hi B Cells for ASC Differentiation (same lab; Stone is second author there). The two are complementary rather than overlapping: Zumaquero is the human paper showing IFN-γ licenses naive B cells by remodelling chromatin and raising IL-21R; Stone is the mouse paper showing what T-bet does downstream of that IFN-γ signal. Read together they separate two things the wiki had been treating as one: IFN-γ opens the PC programme; T-bet does not. Note that Zumaquero’s wiki page carries a partial-ingest banner, so the argument above is anchored entirely in Stone’s own internal controls.

Bearing on the wiki’s “T-bet is confirmatory, not defining” position. Stone adds mouse primary data to the point made on consensus grounds by Eisenbarth2025 - A Roadmap for Defining Extrafollicular B Cell Responses: T-bet reporter⁺ B cells include GC B cells, so T-bet positivity is not evidence of GC-independence. It also adds a functional dimension the marker debate lacked — T-bet may be dispensable for forming a CD11c⁺ ABC (per Du 2019 / Levack 2020 via Glaros2025 - Multilayered Identity of B Cell Memory) while still being required for that cell to become an ASC in a type-1 environment. Necessity for formation and necessity for output are different questions and the wiki should not use evidence about one to settle the other.

Limitations. Mouse throughout. The in vitro system is an allogeneic MLR with polarised Th1/Th2 effectors — a strong, artificial cytokine polarisation rather than a physiological infection microenvironment. The human relevance of the T-bet⁺ DN2 link is asserted in the Introduction and Discussion by citation (Jenks2018 - DN2 B Cells and EF Pathway in SLE, Knox 2017, Lau 2017, Wang 2018), not tested. No dengue or flavivirus data.

Questions Raised

• Does the permissive (brake-release) model hold in human atypical B cells? Testing it would require showing that sustained NF-κB/TLR signalling suppresses ASC output from sorted human DN2 cells — the direct human analogue of Fig. 5A–5D, which no wiki source currently supplies.
• Acute dengue is a strongly type-1, TLR7-activating environment with high IFN-γ and high TLR ligand availability. Stone’s model predicts these pull in opposite directions on ASC output — IFN-γ licensing it, sustained TLR signalling braking it. Does the well-documented massive dengue plasmablast expansion (Wrammert2012 - Plasmablast Responses in Acute Dengue) indicate that the brake is overcome, bypassed, or simply not operative in human acute infection?
• Is the splenic-vs-bone-marrow T-bet gradient present in humans, and if so, what does a blood plasmablast’s T-bet status actually report — recent EF origin, or compartment of transit?
• If T-bet is required for memory B cell differentiation but not maintenance, is the reverse true for the DN2/atypical pool? The wiki has no source addressing whether atypical B cell persistence and atypical B cell output have separable transcriptional requirements.
• Does the IgG2b escape from T-bet dependence (attributed to TGF-β) have a human isotype analogue relevant to dengue subclass distributions?

Related Pages

Zumaquero2019 - IFN-gamma Programs T-bet-hi B Cells for ASC Differentiation, Cancro2020 - Age-Associated B Cells, Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, Jenks2018 - DN2 B Cells and EF Pathway in SLE, Eisenbarth2025 - A Roadmap for Defining Extrafollicular B Cell Responses, Glaros2025 - Multilayered Identity of B Cell Memory