Stability and progressive differentiation of TFR cells are intrinsically and extrinsically controlled by TFH programs

Lee, Jeong-Mi; Raeder, Paulo Lisboa; Rashid, Ridwan B.; Zhang, Hengcheng; Nelson, Cody S.; Richardson, Sierra G.; Burke, Kelly P.; Chandrakar, Pragya; Podestà, Manuel A.; Gempler, Maya G.; Batty, Cole; Kong, Dejun; Gavzy, Samuel J.; Shirkey, Marina W.; Yang, Cyrus; Bromberg, Jonathan S.; Haigis, Marcia C.; Marasco, Wayne A.; Sharpe, Arlene H.; Sage, Peter T.

doi:10.1038/s41590-025-02373-7

Download PDF

Article
Open access
Published: 13 January 2026

Stability and progressive differentiation of T_FR cells are intrinsically and extrinsically controlled by T_FH programs

Nature Immunology (2026)Cite this article

4524 Accesses
12 Altmetric
Metrics details

Subjects

Abstract

Follicular regulatory T (T_FR) cells restrain follicular helper T (T_FH) cell-mediated B cell responses to optimize humoral immunity while limiting autoimmunity. Here we assessed the developmental dynamics of T_FR cells. We found that T_FR cells undergo progressive differentiation through progenitor, early effector and late effector stages. Late effector T_FR cells possessed inherent instability, and could lose expression of FoxP3 to become ExT_FR cells. Expression of effector T_FH programs in T_FR cells preceded instability, a process that was mediated by Tcf7. A subset of ExT_FR could be redeemed by re-expression of FoxP3. Extrinsically, T_FH cells enhanced late effector T_FR cell differentiation by diverting cells away from a default Prdm1/Blimp-1 fate to express Bcl6. Together, these data indicate that T_FR cells are a dynamic and plastic cell subset, the differentiation of which is controlled by intrinsic and extrinsic programs that work together to form a feedback loop to control humoral immunity.

Divergent cytokine and transcriptional signatures control functional T follicular helper cell heterogeneity

Article Open access 09 September 2025

The endogenous repertoire harbors self-reactive CD4⁺ T cell clones that adopt a follicular helper T cell-like phenotype at steady state

Article 09 February 2023

A T follicular helper cell origin for T regulatory type 1 cells

Article Open access 27 March 2023

Main

Antibodies protect against infection, but B cell dysregulation can lead to pathogenic antibodies. To prevent this, the immune system uses overlapping mechanisms to ensure antibodies are tightly controlled. Follicular helper T (T_FH) cells promote antibody production by activating B cells and the germinal center (GC) reaction. In contrast, follicular regulatory T (T_FR) cells limit B cells and GCs^1,2,3,4. T_FR cells attenuate foreign antigen-specific antibodies as well as pathogenic antibodies in autoimmunity, allergies and other diseases^5,6,7. In addition, T_FR cells regulate tertiary lymphoid structures (TLS) and control the efficacy of immune checkpoint blockade^8,9. However, in some settings, T_FR cells enhance antibodies¹⁰. This discrepancy can be partially explained by the finding that T_FR cells may promote affinity maturation by limiting clonal competition within GCs¹¹.

T_FR cells differentiate from natural T regulatory (T_reg) cell precursors, although some may arise from induced T_reg cells^12,13,14. Both dendritic cells and B cells are required for initial T_FR differentiation, as is costimulation through CD28 and ICOS^15,16,17,18. Similar to T_FH, a subset of T_FR cells circulate in the blood as memory cells^14,16,19. T_FR cells in GCs have lower CD25 and higher programmed cell death protein 1 (PD-1) expression indicating activation^15,20,21,22. Despite clonal expansion, T_FR cells proliferate less substantially than T_FH cells^12,23. Some T_FR cells can also enter the GC reaction relatively late to contract GCs^6,23. Like other T_reg subsets, T_FR coopt transcription factors of their target cells, such as T_FH, to establish programs^24,25. Some T_FR can downregulate FoxP3 to become ExT_FR with reduced suppressive capacity²⁵. Recent data indicate T_FH cells undergo progressive differentiation from stem-like progenitor to effector stages—a process highly regulated by the immune system^26,27. Progenitor T_FH cells are similar to stem-like populations of CD8 T and Th17, which sustain long-term responses^28,29. Whether T_FR cells undergo similar development is unknown.

Here we used longitudinal sampling, fate mapping and subset deletion to fully elucidate the dynamics and signals for T_FR development and stability. We find T_FR cells require progressive development through progenitor, early effector and late effector stages to generate functional T_FR cells in both mice and humans. Moreover, we find that only late effector T_FR cells have inherent instability and form ExT_FR cells, a process controlled by Tcf7/TCF-1. Moreover, ExT_FR cells persist longer than effector T_FR and can be redeemed through re-expression of FoxP3. FoxP3 instability is heterogeneous and acquisition of effector T_FH programs in T_FR cells precedes instability. Using inducible deletion strategies, we further find that T_FH cells extrinsically redirect late effector stage T_FR cells from a default Prdm1/Blimp-1-expressing, to Bcl6-expressing, late effector state. Together these data reveal how progressive development generates effector T_FR cells and how intrinsic and extrinsic mechanisms control humoral immunity.

Results

T_FR cells expand clonally throughout lymphoid organs after immunization

T_FR cells undergo clonal expansion, which is poorly understood^12,21. We isolated the superior and inferior regions of a spleen 10 days after immunization (Extended Data Fig. 1a). These regions had similar frequencies of total follicular T cells (CD4⁺CXCR5⁺PD-1⁺), T_FH cells (CD4⁺CXCR5⁺PD-1⁺FoxP3⁻), T_FR cells (CD4⁺CXCR5⁺PD-1⁺FoxP3⁺) and PD-1^hi T_FR cells (CD4⁺CXCR5⁺PD-1^hiFoxP3⁺) (Fig. 1a). Moreover, GC B cells (CD19⁺CD38⁻GL7⁺) were similar (Fig. 1b). To assess whether T_FR cells expand clonally we performed T cell receptor (TCR) sequencing (TCR-seq) on sorted T_FR cells from the superior and inferior regions. When we compared all T_FR TCR sequences we found ~1,000 T_FR clones per mouse with ~40% of these being expanded (Fig. 1c). Within expanded clones, most were minimally expanded, although some clones appeared more than ten times (Fig. 1d). To assess whether T_FR cells can move between distinct regions of the spleen during expansion we compared TCR sequences and found the same T_FR clones in both the superior and inferior regions (Fig. 1e–g and Extended Data Fig. 1b). Of all expanded T_FR clones, ~10% had evidence of ‘traveling’ from one region of the spleen to another (Fig. 1f). In some cases, clones had a higher proportion in one of the regions, possibly where they first developed (Fig. 1g). Together these data indicate that T_FR cells undergo clonal expansion and can migrate to neighboring regions of the same lymphoid organ.

**Fig. 1: T_FR cells expand clonally throughout the spleen after immunization.**

Effector T_FR cells require progressive differentiation from stem-like progenitor to effector T_FR cells

We next studied the dynamics of T_FR differentiation while incorporating assessment of FoxP3 stability. Foxp3^CreERT2GFPRosa26^{LoxSTOPLoxTdtomato} mice (cells previously FoxP3⁺ are Tdtomato⁺, cells currently FoxP3⁺ are GFP⁺) were vaccinated and assessed on day 24 (during GC contraction) (Extended Data Fig. 2a). Around 30% of T_FR cells lost FoxP3 to become ExT_FR cells (Fig. 2a). We next assessed T_FR cell developmental dynamics incorporating transcriptional programming (differentiation state), FoxP3 history (stability and ExT_FR (ref. ²⁵)) and TCR sequence (clonal dynamics) in a longitudinal sampling setting. We performed a survival hemisplenectomy surgery (used in T cell studies^23,30) in which the inferior region of the spleen was removed surgically at day 10 after immunization (initial T_FR response) and the superior region profiled 2 weeks later (GC contraction) (Fig. 2b). We sorted T_FR cells (CD4⁺CXCR5⁺GFP⁺Tdtomato⁺) at days 10 and 24, and ExT_FR (CD4⁺CXCR5⁺GFP⁻Tdtomato⁺) on day 24, and performed single-cell RNA sequencing (scRNA-seq)/TCR-seq. All T_FR cells together segregated into ten distinct clusters in uniform manifold approximation and projection (UMAP) space (Fig. 2c). Using published T_FR cell genes, we separated clusters into three distinct differentiation stages: progenitor, early effector and late effector (Fig. 2c–e). The progenitor T_FR cell stage included a single cluster ‘Prog’ that expressed Sell/CD62L and S1pr1, which was notable since other studies used Sell to label naive T cells²³. The early effector stage included four clusters; a circulating-like cluster (expressing Klf2, Crip1, Vim, Cxcr3 and S100a6), an interferon (IFN)-responsive cluster (expressing Ifit1, Isg15 and Ifit3), an early effector 1 cluster (expressing Lars1 and Klf2) and an early effector 2 cluster (expressing Izumo1r and Il2ra). The late effector stage contained five clusters including CD25⁺ effector and TNFR⁺ effector clusters (expressing Il2ra, Tnfrsf9/4-1BB, Tnfrsf4/OX40, Cd69), late effector 1 and 2 (both expressing Pdcd1, Maf, Tigit, Tox2 and Sostdc1) and memory (expressing Ikzf2, Camk1d, Jarid2 and Lars2). Pseudotime analysis demonstrated progressive development from progenitor to early effector to late effector stages with three terminal outcomes; CD25⁺ effector, late effector 2 or memory T_FR cells (Fig. 2d). In addition, a fourth terminal outcome of progenitor to Circulating T_FR differentiation was found. Stages could be separated based on Pdcd1/PD-1 and Il2ra/CD25 expression with progenitor T_FR being Il2ra⁺Pdcd1^low, early effector T_FR being Il2ra⁺Pdcd1⁺ and late effector being Il2ra⁻Pdcd1^hi (Fig. 2e,f and Extended Data Fig. 2b,c). Network analysis revealed Bcl6 as a transcription factor driving stage transitions (Extended Data Fig. 2d). To confirm progenitor potential, we adoptively transferred progenitor T_FR cells and found these differentiated into Effector T_FR cells with enhanced expansion (Extended Data Fig. 2e). Moreover, progenitor T_FR cells that were activated in vitro had suppressive capacity (Extended Data Fig. 2f).

**Fig. 2: Effector T_FR cells require progressive differentiation from stem-like progenitor to late effector T_FR cells.**

T_FR cells at day 10 had more progenitor and early effector cells, whereas day 24 T_FR had increased late effector T_FR cells (Fig. 2g and Extended Data Fig. 2g). Using TCR sequences, we found almost all expanded T_FR clones at day 10 were also found expanded at day 24 (Fig. 2h). The substantial overlap of clones from day 10 and day 24 suggests expanded T_FR clones originate from progressive differentiation. Together, these data indicate that T_FR cells undergo progressive differentiation, that it occurs over the course of weeks, and that these clones can travel to distinct regions of the spleen during the differentiation process.

Human lymph node T_FR cells undergo progressive differentiation, but only early effector T_FR can be found in the circulation

To confirm progressive differentiation of T_FR in humans, we assessed T_FR cells from lymph nodes (LNs) and blood. Total follicular T cells (live⁺CD4⁺CXCR5⁺CD19⁻) from iliac LNs and similarly gated cells from the blood of another two participants were profiled (Fig. 3a). T_FR cells were separated from T_FH cells using FOXP3 expression (Extended Data Fig. 3a,b). T_FR cells formed five distinct clusters (Fig. 3b). Cluster C4 was found only in blood. Clusters C1/C2 and C3, enriched in both LN and blood, were consistent with progenitor/early effector T_FR and early-to-late transitioning cells, respectively (using gene modules from Fig. 2) (Fig. 3c). Cluster C5 was the only late effector cluster and highly expressed PDCD1/PD-1 and TOX (Fig. 3c,d). Although most blood T_FR were in cluster C4, some were also found in clusters C1, C2 and C3. However, no blood T_FR cells were found in cluster C5. This trend was consistent for each donor (Extended Data Fig. 3c). The early effector clusters C1–C3 expressed KLF2, which promotes egress of T cells into the circulation³¹, whereas the late effector cluster C5 lacked expression of circulating genes (SELL/CD62L, S1PR1 and CCR7) (Extended Data Fig. 3d).

**Fig. 3: Human T_FR cells from LNs undergo progressive differentiation but late effector T_FR do not enter the circulation.**

To confirm appearance of early effector T_FR cells in blood after vaccination we analyzed a scRNA-seq dataset of human CXCR5⁺CD4⁺ cells³² (Fig. 3e). A cluster expressing FOXP3 was further interrogated as T_FR cells (Extended Data Fig. 3e,f). Total T_FR cells formed five distinct clusters (Fig. 3f and Extended Data Fig. 3g). Clusters B2/B4 were less abundant at baseline but increased after vaccination before being attenuated at day 14 (Fig. 3f). People with B2 and B4 responses had a higher overall T_FR:T_FH ratio (Extended Data Fig. 3h). The vaccine-elicited clusters B2/B4 were enriched for early effector T_FR gene modules as well as programs for cellular activation, metabolic flux and motility (Fig. 3g,h). Genes expressed in B2/B4 included: CXCR4 (expressed on activated T_FH³³), CD69, ID3 (which controls T_FR programs³⁴) and RGS1 (required for T_FH differentiation³⁵). Together, these data indicate that human T_FR cells undergo progressive differentiation within LNs and that a small proportion of early effector T_FR cells (but not late effector T_FR cells) gain access to the blood and maintain programs for at least a short amount of time, where they are indicators of T_FR responses to vaccination.

ExT_FR cells originate from unstable late effector T_FR stages but maintain transcriptional programming

Next, we assessed the developmental origins of ExT_FR cells and found these were contained within annotated T_FR clusters, but with over-representation in circulating and late effector (late effector 1/2, TNFR⁺ effector and memory) stages (Fig. 4a,b). ExT_FR cells maintained transcriptional programs including expression of Pdcd1/PD-1, Bcl6 and Icos (Fig. 4c and Extended Data Fig. 4a,b). However, ExT_FR cells did have altered expression of some genes within each cluster, notably attenuation of Foxp3, Il2ra (directly regulated by FoxP3³⁶) and Cd74 (a T_reg stabilizing factor³⁷) (Fig. 4d). Moreover, the T cell stemness and T_FH differentiation gene Tcf7^38,39, the TCR signaling gene Themis, the GC T_FH and exhaustion regulator Tox2⁴⁰, and the costimulatory molecule Cd40lg were upregulated in ExT_FR cells compared to T_FR in most clusters.

**Fig. 4: ExT_FR cells originate from unstable late effector T_FR cells but maintain T_FR transcriptional programming.**

To assess potential clonal origins we compared the TCR sequences. Most expanded clones in T_FR were shared with ExT_FR cells, suggesting that ExT_FR originate from expanded T_FR cells (Fig. 4e). Clonal expansion was similar between T_FR and ExT_FR within each stage, with the exception of Circulating stage cells, which were more expanded in ExT_FR (Fig. 4f,g). Moreover, all of the most expanded clones in T_FR and ExT_FR cells were from clones also found at day 10 (Fig. 4h). Together, these data suggest that ExT_FR cells arise from clonally expanded late effector T_FR cells, and largely maintain developmental stage transcriptional signatures.

We next assessed whether the ExT_FR state is terminal or if ExT_FR cells could be ‘redeemed’ through reacquisition of FoxP3. In vitro cultures of ExT_FR cells showed some FoxP3 re-expression indicating redemption—a process that was augmented in presence of T_FH cells (Extended Data Fig. 4c–f). Furthermore, ExT_FR cells persisted longer than late effector T_FR cells and underwent less apoptosis, indicating superior fitness (Extended Data Fig. 4g–i). The superior longevity combined with redemption capabilities indicate ExT_FR may resupply late effector T_FR cells over time.

Acquisition of effector T_FH cell transcriptional programs in T_FR cells precedes instability

We next assessed transcriptional predictors of T_FR cell instability by stratifying clones based on stability. We designated T_FR clones as ‘Stable’ if the clonal frequency was higher in T_FR versus the ExT_FR compartments at day 24 (Fig. 5a). Likewise, ‘Unstable’ clones had fewer cells in T_FR versus ExT_FR. Both Stable and Unstable clones were found in each mouse, the latter of which were found in late effector stages (Fig. 5b,c and Extended Data Fig. 5a).

**Fig. 5: Acquisition of follicular T cell programs in T_FR cells precedes and predicts instability.**

To uncover factors preceding and contributing to stability, we assessed differentially expressed genes between stable and unstable clones, but only when cells still expressed FoxP3 in the T_FR stage. Unstable T_FR clones had increased expression of effector T_FH genes (Sostdc1, Ascl2, Tox2, Il21, Pou2af1, Plekho1 and Gzmk) and stable clones had increased expression of negative regulators of T_FH (Il2ra, Bach2, Klf2, Il7r) (Fig. 5d). To determine whether unstable clones had enrichment for effector T_FH genes, we performed gene set enrichment analysis (GSEA) utilizing gene sets for effector T_FH (effector versus progenitor T_FH²⁷, Bcl6-expressing T_FH⁴¹). We found unstable T_FR clones had enrichment of both effector T_FH gene modules, suggesting unstable T_FR clones obtain an effector-like T_FH program before downregulating FoxP3 (Fig. 5e). Moreover, unstable T_FR clones were also enriched for genes in cell division and cell cycle, suggesting enhanced activation (Fig. 5f).

We further assessed how T_FR cells transition between T_FR and ExT_FR cells on an individual clone basis. Both stable and unstable clones were dominated by late effector T_FR cells at day 24 and largely maintained a similar distribution of developmental stages when they transitioned to ExT_FR cells (Fig. 5g and Extended Data Fig. 5b). Two distinct clones illustrated divergent fates as ExT_FR cells. Stable clonotype 9 was found in late effector clusters as T_FR cells, but had reduced conversion to ExT_FR. In contrast, unstable clonotype 3 expanded as ExT_FR even though these cells had similar distribution in the late effector stage. Furthermore, we identified a stable clone (clonotype 31) that did not advance to the most terminal late effector 2 stage either as T_FR or ExT_FR. Both stable and unstable clones had evidence of ExT_FR cells in circulating stages even though circulating cells were not found in ancestor T_FR cells.

We also assessed our human dataset to assess which LN T_FR cells had evidence of instability utilizing an ‘unstable T_FR gene module’ predicting descendent instability. We found that late effector-like cluster C5 from the human LN dataset had the most enrichment (Fig. 5h). When we assessed shared genes between mouse unstable T_FR cells and human cluster C5 T_FR cells we found a number of effector T_FH genes such as PDCD1/PD-1, TOX2, BCL6, ICOS, BATF, CTLA4 and CD40LG, suggesting late effector T_FR cells in human LN have evidence of instability. Together, these data indicate that acquisition of effector T_FH cell programs in late effector T_FR cells precedes and predicts descendent instability and that when T_FR cells lose FoxP3 to become ExT_FR, they reside in similar developmental stages.

T_FH cells extrinsically redirect late effector T_FR cell development to generate Bcl6-expressing cells

Next we sought to understand how T_FH cells control T_FR cell differentiation. We required a system in which T_FH cells could be acutely deleted without affecting T_FR cells. We used F/Ex-DTR mice which delete effector T_FH cells without eliminating T_FR^27,42 (Extended Data Fig. 6a). We vaccinated F/Ex-DTR (Il21^CreCxcr5^{LoxSTOPLoxDTR}Rosa26^{LoxSTOPLoxYFP}) or control (Il21^CreCxcr5^wtRosa26^{LoxSTOPLoxYFP}) mice then administered diphtheria toxin (DT) on days 7 and 9 to delete T_FH before collection on day 10 (Fig. 6a). Deleting T_FH cells acutely minimally affected GC B cells so effects on T_FR could be studied (Extended Data Fig. 6b). T_FR cells (CD4⁺CXCR5⁺PD-1⁺GITR⁺) were sorted from control or T_FH-deleted mice and scRNA-seq/TCR-seq performed. T_FR cells separated into eight distinct clusters, which were annotated for T_FR cell developmental stages (Fig. 6b,c). Milo neighborhood analysis indicated over-representation of C4 (early effector 2/memory stage) and C5 (Prdm1^hi/late effector 1/Circulating stage), as well as under-representation of C7 (Bcl6^hi/late effector 2 stage) in T_FR cells from T_FH-deleted mice (Fig. 6d and Extended Data Fig. 6c). We further confirmed that C5 and C7 were consistent with late effector 1 and 2 stages, respectively, due to mutually exclusive expression of Bcl6/Prdm1 (Fig. 6e and Extended Data Fig. 6d,e). Together, these data indicate that the default developmental pathway in T_FR cells is from progenitor > early effector > Blimp-1⁺ late effector stage 1 > CD25⁺ or TNFR⁺ effector stages. However, when T_FH cells are present, T_FR are redirected to a late effector 2/Bcl6^hi T_FR stage. To determine whether Blimp-1 was required to limit C7 genes, we performed bulk RNA-seq on T_FR cells (CD4⁺ICOS⁺CXCR5⁺FoxP3-YFP⁺) from control (Foxp3^CrePrdm1^wt) or Blimp-1cKO (Foxp3^CrePrdm1^floxed) mice. C5 genes (Snx9, Il2ra, Ccr2, Arl5a, Ccr5, Gbp6 and Vim) had reduced expression in Blimp-1-deleted T_FR cells (Fig. 6f and Extended Data Fig. 6f). Likewise, C7 genes were upregulated in Blimp-1-deleted T_FR cells, indicating Blimp-1 limits late effector 2 stage differentiation.

Next we determined whether T_FH cells control T_FR cell stability. We required a strategy to delete T_FH cells inducibly while allowing FoxP3 fate mapping. We crossed an Il21^LoxDTRLox strain⁴³ to FoxP3 fate mapping alleles. In this way, effector T_FH, which produce IL-21, will be eliminated with DT and T_FR cells assessed for FoxP3 stability. In addition, since the DTR cassette is flanked by LoxP sites, any cell that has expressed FoxP3 (T_FR/ExT_FR) will be insensitive to deletion regardless of IL-21 expression. We immunized IL21DTR (Il21^LoxDTRLoxFoxp3^CreERT2GFPRosa26^{LoxSTOPLoxTdtomato}) or control Il21^wtFoxp3^CreERT2GFPRosa26^{LoxSTOPLoxTdtomato}) mice and deleted T_FH with DT, and assessed ExT_FR. ExT_FR were higher in T_FH-deleted versus nondeleted mice, suggesting that T_FH cells promote T_FR stability (Fig. 6g). Further profiling revealed Circulating-like T_FR (CXCR3⁺PD-1⁻) were enhanced, whereas the late effector 2 population (CXCR3⁻PD-1⁺) were attenuated upon T_FH deletion (Extended Data Fig. 6g).

To further investigate mechanisms of T_FR stability, we focused on C7 genes also expressed in ExT_FR, including Tcf7 (Fig. 6e,f). In T_FR cells, Tcf7 was differentially expressed in C7 compared to C5 cells (Fig. 6e and Extended Data Fig. 6h). At the protein level, TCF-1 (encoded by Tcf7) was highest in progenitor T_FR cells, but was still expressed in effector stages (Fig. 6h). To assess the role of TCF-1 in T_FR stability we vaccinated Foxp3^ERT2CreTcf7^fl/fl mice and administered tamoxifen to delete Tcf7. T_regΔTcf7 mice had less total T_FR, as reported³⁹ (Extended Data Fig. 6i). We cultured T_FR from control or T_regΔTcf7 mice with B cells and assessed conversion to ExT_FR cells. Tcf7-deficient T_FR cells were more stable compared to Tcf7-sufficient T_FR cells, indicating Tcf7 promotes instability (Fig. 6i). Taken together, T_FH cells redirect progressive differentiation of T_FR cells away from a default Blimp-1⁺ late effector 1 pathway toward a Bcl6-dominated late effector 2 pathway without inducing instability (Extended Data Fig. 6j). In this manner, the GC reaction contains negative feedback loops in which increased T_FH cells redirect sequential T_FR cell differentiation toward late effector T_FR cells to limit humoral immunity (Extended Data Fig. 6k).

Discussion

T_FR cells potently regulate the GC reaction in health and disease, yet factors controlling T_FR cell differentiation and stability are poorly understood. Here we uncovered three main developmental stages in T_FR cells; progenitor, early effector and late effector stages, with substages contained within each main stage. The stem-like progenitor stage in T_FR cells is similar to the progenitor stage in T_FH cells²⁷, with both having expression of Sell/CD62L and can rapidly differentiate and expand in vivo. The current prevailing paradigm suggests that T_FR cells invade GCs only late during immune responses, and regulation of GCs is dictated by de novo recruitment of T_FR cells. However, studies supporting this paradigm utilized Sell expression to identify naive T cells²³. Our work suggests this strategy probably marked progenitor T_FR cells as naive CD4 T cells. Our data support a paradigm in which T_FR cells enter the GC reaction early as progenitor cells but require weeks for sequential development to occur to generate effector T_FR cells. In support of this concept, we find little evidence of new T_FR clones entering the GC reaction after a few weeks. This prolonged presence in lymphoid organs may allow the immune system several opportunities to regulate the T_FR response.

We observed similar developmental stages in human LNs and peripheral blood after seasonal influenza vaccination. A recent finding similarly suggests heterogeneity in human tonsillar T_FR cells through varying expression of ICOS and PD-1⁴⁴. Our study indicates human late effector T_FR cells do not gain access to the blood, but some progenitor/early effector-like T_FR cells can. These data support previous reports that human blood T_FR cells are not fully developed¹⁹. Our data indicate that progenitor/early effector T_FR cells peak in blood within a week after influenza vaccination. Therefore, identifying and analyzing progenitor/early effector T_FR cells in the easily accessible blood may offer a window into lymphoid organ T_FR cell responses during human studies. However, these cells seem to quickly lose their LN progenitor/early effector programming, probably to generate memory T_FR cells, which mount quicker responses upon antigen exposure¹⁶.

Although ExT_FR cells have been reported, the signals conferring instability are poorly studied²⁵. We found that late effector T_FR cells can downregulate FoxP3, whereas progenitor and early effector T_FR are more stable. The transcriptional similarities between T_FR and ExT_FR cells in each stage were an unexpected finding. This observation underscores that the transcriptional program controlling T_FR cell identity is complex and probably involves factors beyond FoxP3 and may be distinct for each stage. In a similar way, tissue-resident T_reg cells can adapt in response to fluctuating signals within tissues^45,46,47. Comparison of Unstable and Stable T_FR clones revealed that the acquisition of T_FH cell transcriptional programs preceded and predicted downstream T_FR instability. Although previous work suggested ExT_FR cells have distinct transcriptional programs as a bulk population²⁵, our current single-cell data show that ExT_FR cells largely maintain their transcriptional programs but are found in later developmental stages. Nevertheless, we did find some key genetic changes in ExT_FR cells, including upregulation of Tox2, Il-21, Tcf7 and Cd40lg, as well as loss of Cd74. Due to these features, ExT_FR may be identified as T_FH cells in studies lacking FoxP3 fate mapping. Moreover, in limited settings some T_reg may lose FoxP3 and undergo T_FH differentiation; however, the lack of ExT_FR in progenitor/early effector stages in our study suggest this is not a dominant pathway^48,49. Although we cannot assess ExT_FR cells in human lymphoid organs in vivo due to the lack of fate mapping strategies, our finding of enriched Unstable T_FR gene modules in human LN late effector T_FR cells suggests this process also occurs in humans. The superior persistence of ExT_FR along with reduced cell death and ability to regain FoxP3 expression indicate this population may be an important source of T_FR cells during prolonged GC reactions. Relatedly, reports of T_FH cells upregulating FoxP3 during GC contraction^50,51 may contain some ExT_FR cells that re-acquire FoxP3. The possible contribution of Foxp3⁻ T_reg or ExT_reg cells to ExT_FR through further differentiation also remains uncertain⁵².

Our work also uncovers extrinsic factors that control T_FR development. The accumulation of Blimp-1⁺ Effector T_FR cells when T_FH are acutely deleted suggests that development of T_FR cells until the Blimp-1⁺ Effector stage is largely intrinsic, and later stages conditional on T_FH cells. Specifically, T_FH cells redirect T_FR cells to a fully suppressive Bcl6⁺ late effector 2 T_FR stage. This arrangement allows for the formation of partially developed T_FR cells as a resource pool that can be converted quickly to suppressive effector T_FR cells as T_FH cells accumulate. The precise mechanisms by which T_FH cells redirect late effector T_FR cell differentiation and control stability are unclear, but Bcl6 is involved. Tcf7, a stemness marker in some T cells^29,53, was highly expressed in ExT_FR cells. Tcf7-deficient T_FR cells had greater stability, suggesting Tcf7 promotes T_FR instability. Paradoxically, T_FH cells induce Tcf7 in T_FR but also promote T_FR stability. Therefore, we hypothesize T_FH provide other stabilization factors to counteract Tcf7/TCF-1 effects. Taken together, these data indicate that T_FR cells are a dynamic population that undergoes sequential development and that the immune system incorporates several intrinsic and extrinsic strategies to control T_FR cells as a negative feedback loop to control humoral immunity.

Methods

Mice

Foxp3^CreERT2GFP, Rosa26^{Lox-STOP-Lox-TdTomato} and Tcf7^fl/fl were from Jackson Laboratories. Il21^Cre mice were a kind gift from U. Hoepken. Cxcr5^{LoxSTOPLoxDTR} mice have been published previously^5,11. Il21^LoxDTRLox was made by CRISPR-based knock-in of a LoxP-IRES-DTR-LoxP cassette into the 3′ UTR of the Il21 gene⁴³. Both male and female mice, aged 6–10 weeks, were included in the study. Mice were maintained on a 12-h/12-h dark/light cycle at 22 °C with 42% humidity and were fed a 5053 PicoLab rodent diet 20 (LabDiet). All animals were handled in accordance with the policies of the Brigham and Women’s Hospital Institutional Animal Care and Use Committee and the guidelines of the National Institutes of Health.

Immunizations and treatments

Mice were immunized with 100 μg NP-OVA (Biosearch Technologies) emulsified (1:1) in complete Freund adjuvant (CFA, Sigma-Aldrich) for subcutaneous immunization or emulsified in incomplete Freund adjuvant (IFA, Sigma-Aldrich) for intraperitoneal immunization. For deletion experiments, mice were administered 0.5 μg of diphtheria toxin intraperitoneally in PBS at indicated timepoints. In experiments with Foxp3 fate mapping, tamoxifen (2.5 mg per dose, Sigma-Aldrich) was injected intraperitoneally for 5 consecutive days before immunization to induce the activity of Cre recombinase.

Adaptive TCR sequencing

T_FR cells (gated as CD4⁺CD19⁻CXCR5⁺FoxP3⁺) were sorted from superior and inferior regions of spleens from NP-OVA immunized mice at day 12 postimmunization. From the sorted T_FR cells, genomic DNA was extracted according to the sample preparation guidelines provided for ImmunoSEQ assays (Adaptive Biotechnologies). TCR B regions were sequenced and TCRs with identical VDJ CDR3 amino acid sequences were considered clonotypes.

Human LNs and blood samples

LNs were collected from consenting patients before kidney transplantation. All patients had renal failure but had not received any immunosuppression. LN tissue was mashed through 70-µm strainers. Total follicular T cells, gated as CD4⁺CXCR5⁺, were sorted and subjected to scRNA-seq. Peripheral blood samples were collected from healthy people (Research Blood Components). For human blood analysis after influenza vaccination, data were reanalyzed from a published dataset³². In this dataset, blood was collected at indicated timepoints from influenza vaccine recipients between the ages of 18 and 85 years.

Mouse hemisplenectomy

Inferior parts of spleens from two NP-OVA immunized FoxP3FM mice (Foxp3^ERT2Cre-eGFPRosa26^{Lox-STOP-Lox-TdTomato}) were collected 10 days post-immunization with a survival hemisplenectomy, and the remaining superior parts of spleens were collected 24 days post-immunization from the matched mice for longitudinal assay. Briefly, a laparotomy was performed following anesthesia. The spleen was exposed by making a midline abdominal incision. The lieno-pancreatic vein and artery were ligated and disconnected close to the spleen with silk suture. The spleen was ligated twice around its mid-body with silk suture and cut off from the middle. The inferior part of the spleen was then removed, and the superior part was placed back into the peritoneum. The surgical site was kept sterile throughout the procedure. After the laparotomy, the mice were placed on a heating pad for recovery.

scRNA-seq of mouse T_FR/ExT_FR cells

Spleen or inguinal LN T_FR cells (gated as DAPI⁻CD4⁺CD19⁻CXCR5⁺FoxP3FM-tdTomato⁺FoxP3-GFP⁺ and DAPI⁻CD4⁺CD19⁻CXCR5⁺GITR⁺ for the hemisplenectomy experiments in Fig. 2 and Fig. 4 and the F/Ex-DTR experiment in Fig. 6, respectively) were sorted. Before sorting, CD4⁺ T cells were enriched through magnetic selection (Miltenyi Biotec, mouse CD4(L3T4) microbeads). Enriched CD4⁺ T cells in single-cell suspensions were stained with fluorochrome-labeled cell surface antibodies as well as barcoded antibodies (Cell-Hashing antibody, TotalSeq-C, Biolegend) following the manufacturer’s protocol. Subsequently, T_FR cells and/or ExT_FR cells were sorted per gating strategies above using BD FACS Symphony S6. The sorted cells were loaded onto a single lane of a Chromium chip K (10x Genomics) and encapsulated in lipid droplets using the Single Cell 5′ kit V2 (10x Genomics) at the Brigham and Women’s Hospital Single Cell Genomics Core. cDNA and library generation were performed according to the manufacturer’s protocol. The gene expression (GEX) library and the V(D)J library were sequenced using Illumina Novaseq. Reads were processed with CellRanger, and quantification was performed using the STAR aligner against the mm10 transcriptome.

scRNA-seq of human T_FR cells

From cryopreserved peripheral blood mononuclear cells and external iliac LN single-cell suspensions, CD4⁺ T cells were enriched magnetically according to the manufacturer’s instructions (Miltenyi Biotech). Samples were barcoded with TotalSeq-C antibodies from Biolegend. CD4⁺ CXCR5⁺ CD19⁻ cells were sorted and loaded onto a single lane of a Chromium chip K (10x Genomics) and encapsulated in lipid droplets using the Single Cell 5′ kit V2 (10x Genomics) at the Brigham and Women’s Hospital Single Cell Genomics Core. Subsequently, cDNA synthesis and library preparation were executed as per the manufacturer’s instructions. The 5′ mRNA library was sequenced on an Illumina Novaseq. Postsequencing reads underwent processing using CellRanger, and quantified using the STAR aligner against the human transcriptome. CellRanger output data were imported into the R programming environment for analysis using the Seurat package.

scRNA-seq data processing

The raw data output from CellRanger were imported and analyzed using the Seurat v.5.0 in R 4.3.1. Initially, sample demultiplexing and doublet exclusion were conducted to ensure that only singlets were selected for downstream analysis. Additional quality control measures were applied based on unique molecular identifier counts, gene counts, log-transformed genes per unique molecular identifier and mitochondrial RNA content. Non-T cell contaminants were identified and excluded using the cell annotation tool SingleR (v.2.4.1)⁵⁴, referencing the ImmGen database for mouse data and HumanCellAtlas for human data. For mouse hemisplenectomy data and human blood kinetics, data from different timepoints were preprocessed independently in separate Seurat objects then merged and integrated using the IntegrateLayers function, employing the anchor-based CCA (canonical correlation analysis) or Harmony integration method. Postintegration, the data were subjected to SCT normalization to generate a merged object, FindNeighbors, FindClusters and RunUMAP for visualization using the new dimensional reduction (integrated.cca or harmony). Afterwards, the layers of each timepoint were rejoined for further differential expression analyses. Monocle pseudotime analysis and Milo differential abundance analysis were performed as described^55,56. Gene regulatory network inference was performed using R SCENIC (v.1.1.2) as described⁵⁷.

Area under the curve analysis and gene set enrichment analysis

Area under the curve (AUC) analysis was utilized to identify corresponding T_FR developmental stages of each cluster from F/Ex-DTR scRNA-seq data presented in Fig. 6. AUC scores were calculated to assess the enrichment of gene sets representing different T_FR stages in a cluster using R AUCell (v.1.24.0)⁵⁷. These gene sets were curated using the FindMarker function on mouse hemisplenectomy experiments, with parameters set to a log fold change threshold of 0.1 and a min.pct of 0.1. GSEA was performed to assess the gene signature of mouse Unstable T_FR clones and blood human follicular T cells. For mouse Unstable T_FR clones, of the in-house curated T_FH/T_FR-related gene modules and gene ontology categories (C5) and immunological conditions (C7) gene modules from the mSigDB (Molecular Signature Database), significantly enriched gene sets were selected based on the NES (normalized enrichment score). For human blood follicular T cells, the T_FH versus T_con and T_FH versus T_FR gene modules from previous work²⁵ were used to elucidate the identity of C3.

TCR clonotype analysis

The 10x Genomics CellRanger VDJ output, comprising indexed FASTA files and contig annotations, was processed using the R package Platypus (v.3.6.0)⁵⁸ to construct a VGM (V(D)J-GEX-Merged) project. This project generated a matrix of TCR clonotype information on a per-cell barcode basis. VGM projects for day 10 T_FR, day 24 T_FR and day 24 exT_FR were merged into a single matrix, and public clonotypes were assigned across the three datasets using the VDJ_clonotype function. Clonotyping was based on CDR3 amino acid sequences. To address cells with atypical chain numbers, a single-chain-based hierarchical clonotyping approach was implemented. Cells with two VDJ chains were excluded, whereas those with only one chain (either VDJ or VJ) were incorporated into preexisting clones if they met the clonotyping criteria. Cells from the GEX library were annotated with their corresponding clonotypes based on their barcodes.

Flow cytometry

Single-cell suspensions were treated with antibody cocktails targeting surface markers for 30 min at 4 °C, typically using a 1:200 dilution unless specified otherwise: anti-CD4 (BD Horizon, cat. no. GK1.5), anti-CD19 (BD Horizon, cat. no. 1D3), anti-PD-1 (BioLegend, cat. no. RMP1-30), Biotin anti-CXCR5(BioLegend, cat. no. L138D7, 1:100), anti T cell and B cell activation antigen (BD Biosciences, cat. no. GL7), anti-CD62L (BioLegend, clone MEL-14), anti-CD25 (BioLegend, cat. no. PC61), anti-CD38 (BD OptiBuild, 90), anti-IgG1 (BioLegend, cat. no. A85-1, 1:100), anti-CD357 (BioLegend, cat. no. DTA-1), anti-Foxp3 (Invitrogen, cat. no. FJK16-s), anti-TCF-1 (BD Biosciences, cat. no. S33-966). For CXCR5 detection, cells were further stained with Streptavidin-BV421 (BioLegend, 1:300) for 20 min at 4 °C. Intracellular staining involved fixation with Foxp3 Fix/Perm buffer (eBiosciences) according to the manufacturer’s instructions, followed by incubation with intracellular antibody cocktail in permeabilization buffer for 30 min. For apoptosis analysis, Annexin V Apoptosis Detection Kit with 7-AAD (BioLegend) was used per manufacturer’s instructions. Analysis was performed using FlowJo v.4.0 software.

Statistical analyses

Statistical analyses were performed using GraphPad Prism v.9.0 or R v.4.3.1. Student’s two-tailed unpaired t-test was used for comparisons between two groups, while one-way analysis of variance (ANOVA) was used for comparisons involving more than two groups. For DEG analysis between clusters in scRNA-seq data, the Wilcoxon rank sum test was applied. For bulk RNA-seq analysis, DESeq2 was used with the default settings, employing the Wald test, and the Benjamini–Hochberg method was applied to adjust P values for multiple testing. The number of mice per group, the number of replicates per experiment, summary statistics and measures of dispersion are indicated in the legend of each figure.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The sequencing data generated in this study have been deposited in the GEO database under GSE306818 and GSE307004. All data are available from the corresponding author on reasonable request. Source data are provided with this paper.

References

Sage, P. T. & Sharpe, A. H. The multifaceted functions of follicular regulatory T cells. Curr. Opin. Immunol. 67, 68–74 (2020).
Article PubMed PubMed Central Google Scholar
Wing, J. B., Lim, E. L. & Sakaguchi, S. Control of foreign Ag-specific Ab responses by Treg and Tfr. Immunol. Rev. 296, 104–119 (2020).
Article PubMed Google Scholar
Fonseca, V. R., Ribeiro, F. & Graca, L. T follicular regulatory (Tfr) cells: dissecting the complexity of Tfr-cell compartments. Immunol. Rev. 288, 112–127 (2019).
Article PubMed Google Scholar
Sage, P. T. et al. Suppression by TFR cells leads to durable and selective inhibition of B cell effector function. Nat. Immunol. 17, 1436–1446 (2016).
Article PubMed PubMed Central Google Scholar
Clement, R. L. et al. Follicular regulatory T cells control humoral and allergic immunity by restraining early B cell responses. Nat. Immunol. 20, 1360–1371 (2019).
Article PubMed PubMed Central Google Scholar
Botta, D. et al. Dynamic regulation of T follicular regulatory cell responses by interleukin 2 during influenza infection. Nat. Immunol. 18, 1249–1260 (2017).
Article PubMed PubMed Central Google Scholar
Fu, W. et al. Deficiency in T follicular regulatory cells promotes autoimmunity. J. Exp. Med. 215, 815–825 (2018).
Article PubMed PubMed Central Google Scholar
Eschweiler, S. et al. Intratumoral follicular regulatory T cells curtail anti-PD-1 treatment efficacy. Nat. Immunol. 22, 1052–1063 (2021).
Article PubMed PubMed Central Google Scholar
Noel, G. et al. Functional Th1-oriented T follicular helper cells that infiltrate human breast cancer promote effective adaptive immunity. J. Clin. Invest. 131, e139905 (2021).
Article PubMed PubMed Central Google Scholar
Lu, Y. et al. CD4+ follicular regulatory T cells optimize the influenza virus-specific B cell response. J. Exp. Med. 218, e20200547 (2021).
Article PubMed Google Scholar
Cavazzoni, C. B. et al. Follicular T cells optimize the germinal center response to SARS-CoV-2 protein vaccination in mice. Cell Rep. 38, 110399 (2022).
Article PubMed PubMed Central Google Scholar
Maceiras, A. R. et al. T follicular helper and T follicular regulatory cells have different TCR specificity. Nat. Commun. 8, 15067 (2017).
Article PubMed PubMed Central Google Scholar
Aloulou, M. et al. Follicular regulatory T cells can be specific for the immunizing antigen and derive from naive T cells. Nat. Commun. 7, 10579 (2016).
Article PubMed PubMed Central Google Scholar
Kumar, S. et al. Developmental bifurcation of human T follicular regulatory cells. Sci. Immunol. 6, eabd8411 (2021).
Article PubMed Google Scholar
Sage, P. T., Francisco, L. M., Carman, C. V. & Sharpe, A. H. The receptor PD-1 controls follicular regulatory T cells in the lymph nodes and blood. Nat. Immunol. 14, 152–161 (2013).
Article PubMed Google Scholar
Sage, P. T., Alvarez, D., Godec, J., von Andrian, U. H. & Sharpe, A. H. Circulating T follicular regulatory and helper cells have memory-like properties. J. Clin. Invest. 124, 5191–5204 (2014).
Article PubMed PubMed Central Google Scholar
Sage, P. T., Paterson, A. M., Lovitch, S. B. & Sharpe, A. H. The coinhibitory receptor ctla-4 controls B cell responses by modulating T follicular helper, T follicular regulatory, and T regulatory cells. Immunity 41, 1026–1039 (2014).
Article PubMed PubMed Central Google Scholar
Wing, J. B., Ise, W., Kurosaki, T. & Sakaguchi, S. Regulatory T cells control antigen-specific expansion of Tfh cell number and humoral immune responses via the coreceptor CTLA-4. Immunity 41, 1013–1025 (2014).
Article PubMed Google Scholar
Fonseca, V. R. et al. Human blood Tfr cells are indicators of ongoing humoral activity not fully licensed with suppressive function. Sci. Immunol. 2, eaan1487 (2017).
Article PubMed PubMed Central Google Scholar
Jandl, C. et al. IL-21 restricts T follicular regulatory T cell proliferation through Bcl-6 mediated inhibition of responsiveness to IL-2. Nat. Commun. 8, 14647 (2017).
Article PubMed PubMed Central Google Scholar
Ritvo, P. G. et al. High-resolution repertoire analysis reveals a major bystander activation of Tfh and Tfr cells. Proc. Natl Acad. Sci. USA 115, 9604–9609 (2018).
Article PubMed PubMed Central Google Scholar
Wing, J. B. et al. A distinct subpopulation of CD25(-) T-follicular regulatory cells localizes in the germinal centers. Proc. Natl Acad. Sci. USA 114, E6400–E6409 (2017).
Article PubMed PubMed Central Google Scholar
Merkenschlager, J. et al. Continually recruited naive T cells contribute to the follicular helper and regulatory T cell pools in germinal centers. Nat. Commun. 14, 6944 (2023).
Article PubMed PubMed Central Google Scholar
Panduro, M., Benoist, C. & Mathis, D. Tissue Tregs. Annu. Rev. Immunol. 34, 609–633 (2016).
Article PubMed PubMed Central Google Scholar
Hou, S. et al. FoxP3 and Ezh2 regulate Tfr cell suppressive function and transcriptional program. J. Exp. Med. 216, 605–620 (2019).
Article PubMed PubMed Central Google Scholar
Feng, H. et al. A novel memory-like Tfh cell subset is precursor to effector Tfh cells in recall immune responses. J. Exp. Med. 221, e20221927 (2024).
Article PubMed Google Scholar
Podesta, M. A. et al. Stepwise differentiation of follicular helper T cells reveals distinct developmental and functional states. Nat. Commun. 14, 7712 (2023).
Article PubMed PubMed Central Google Scholar
Im, S. J. et al. Defining CD8+ T cells that provide the proliferative burst after PD-1 therapy. Nature 537, 417–421 (2016).
Article PubMed PubMed Central Google Scholar
Schnell, A. et al. Stem-like intestinal Th17 cells give rise to pathogenic effector T cells during autoimmunity. Cell 184, 6281–6298 (2021).
Article PubMed PubMed Central Google Scholar
Merkenschlager, J. et al. Dynamic regulation of T(FH) selection during the germinal centre reaction. Nature 591, 458–463 (2021).
Article PubMed PubMed Central Google Scholar
Matloubian, M. et al. Lymphocyte egress from thymus and peripheral lymphoid organs is dependent on S1P receptor 1. Nature 427, 355–360 (2004).
Article PubMed Google Scholar
Nelson, C. S. et al. The inflammaging microenvironment induces dysfunctional rewiring of Tfh cell differentiation. JCI Insight 10, e187271 (2025).
Article PubMed PubMed Central Google Scholar
Estes, J. D. et al. Follicular dendritic cell regulation of CXCR4-mediated germinal center CD4 T cell migration. J. Immunol. 173, 6169–6178 (2004).
Article PubMed Google Scholar
Miyazaki, M. et al. Id2 and Id3 maintain the regulatory T cell pool to suppress inflammatory disease. Nat. Immunol. 15, 767–776 (2014).
Article PubMed PubMed Central Google Scholar
Caballero-Franco, C. & Kissler, S. The autoimmunity-associated gene RGS1 affects the frequency of T follicular helper cells. Genes Immun. 17, 228–238 (2016).
Article PubMed PubMed Central Google Scholar
Marson, A. et al. Foxp3 occupancy and regulation of key target genes during T-cell stimulation. Nature 445, 931–935 (2007).
Article PubMed PubMed Central Google Scholar
Bonnin, E. et al. CD74 supports accumulation and function of regulatory T cells in tumors. Nat. Commun. 15, 3749 (2024).
Article PubMed PubMed Central Google Scholar
Wu, T. et al. TCF1 is required for the T follicular helper cell response to viral infection. Cell Rep. 12, 2099–2110 (2015).
Article PubMed PubMed Central Google Scholar
Choi, Y. S. et al. LEF-1 and TCF-1 orchestrate T(FH) differentiation by regulating differentiation circuits upstream of the transcriptional repressor Bcl6. Nat. Immunol. 16, 980–990 (2015).
Article PubMed PubMed Central Google Scholar
Xu, W. et al. The transcription factor Tox2 drives T follicular helper cell development via regulating chromatin accessibility. Immunity 51, 826–839 e825 (2019).
Article PubMed Google Scholar
Liu, X. et al. Bcl6 expression specifies the T follicular helper cell program in vivo. J. Exp. Med. 209, 1841–1852 (2012).
Article PubMed PubMed Central Google Scholar
Zhang, H. et al. IL-21-producing effector Tfh cells promote B cell alloimmunity in lymph nodes and kidney allografts. JCI Insight 8, e169793 (2023).
Article PubMed PubMed Central Google Scholar
Chandrakar, P. et al. Progressively differentiated T(FH)13 cells are stabilized by JunB to mediate allergen germinal center responses. Nat. Immunol. 26, 473–483 (2025).
Article PubMed PubMed Central Google Scholar
Ribeiro, F. et al. PD-1 and ICOS are coexpressed in T follicular helper cells but define three stages of maturation of T follicular regulatory cells. Sci. Adv. 11, eadt8901 (2025).
Article PubMed PubMed Central Google Scholar
Cipolletta, D. et al. PPAR-gamma is a major driver of the accumulation and phenotype of adipose tissue Treg cells. Nature 486, 549–553 (2012).
Article PubMed PubMed Central Google Scholar
Wang, Y., Su, M. A. & Wan, Y. Y. An essential role of the transcription factor GATA-3 for the function of regulatory T cells. Immunity 35, 337–348 (2011).
Article PubMed PubMed Central Google Scholar
Sjaastad, L. E., Owen, D. L., Tracy, S. I. & Farrar, M. A. Phenotypic and Functional Diversity in Regulatory T Cells. Front. Cell Dev. Biol. 9, 715901 (2021).
Article PubMed PubMed Central Google Scholar
Tsuji, M. et al. Preferential generation of follicular B helper T cells from Foxp3+ T cells in gut Peyer’s patches. Science 323, 1488–1492 (2009).
Article PubMed Google Scholar
Gaddis, D. E. et al. Apolipoprotein AI prevents regulatory to follicular helper T cell switching during atherosclerosis. Nat. Commun. 9, 1095 (2018).
Article PubMed PubMed Central Google Scholar
Jacobsen, J. T. et al. Expression of Foxp3 by T follicular helper cells in end-stage germinal centers. Science 373, eabe5146 (2021).
Article PubMed PubMed Central Google Scholar
Le Coz, C. et al. Human T follicular helper clones seed the germinal center-resident regulatory pool. Sci. Immunol. 8, eade8162 (2023).
Article PubMed PubMed Central Google Scholar
Feng, Y. et al. Control of the inheritance of regulatory T cell identity by a cis element in the Foxp3 locus. Cell 158, 749–763 (2014).
Article PubMed PubMed Central Google Scholar
Chen, Z. et al. TCF-1-centered transcriptional network drives an effector versus exhausted CD8 T cell-fate decision. Immunity 51, 840–855 e845 (2019).
Article PubMed PubMed Central Google Scholar
Aran, D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat. Immunol. 20, 163–172 (2019).
Article PubMed PubMed Central Google Scholar
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Article PubMed PubMed Central Google Scholar
Dann, E., Henderson, N. C., Teichmann, S. A., Morgan, M. D. & Marioni, J. C. Differential abundance testing on single-cell data using k-nearest neighbor graphs. Nat. Biotechnol. 40, 245–253 (2022).
Article PubMed Google Scholar
Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017).
Article PubMed PubMed Central Google Scholar
Cotet, T. S. et al. ePlatypus: an ecosystem for computational analysis of immunogenomics data. Bioinformatics 39, btad553 (2023).
Article PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We would like to thank U. Hoepken for kindly providing mice. We would like to thank the Brigham and Women’s Hospital Center for Cellular Profiling, the Beth Israel Deaconess Medical Center Flow Core, the Harvard Medical School Biopolymer facility and Harvard Medical School Flow Core for their support. This work was supported by grants from the NIH (R01AI153124 to P.T.S., R01AI158413 to P.T.S., U01AI165442 to W.A.M.), the São Paulo Research Foundation (grant number: 2022/07035-3 to P.L.R.) and TaiwanBio Therapeutics.

Author information

Authors and Affiliations

Department of Medicine, Transplantation Research Center, Division of Renal Medicine, Brigham and Women’s Hospital, Boston, MA, USA
Jeong-Mi Lee, Paulo Lisboa Raeder, Ridwan B. Rashid, Hengcheng Zhang, Cody S. Nelson, Sierra G. Richardson, Pragya Chandrakar, Manuel A. Podestà, Maya G. Gempler & Peter T. Sage
Department of Immunology, Institute of Biomedical Sciences, University of São Paulo, São Paulo, Brazil
Paulo Lisboa Raeder
Department of Medicine, Division of Allergy and Clinical Immunology, Brigham and Women’s Hospital, Boston, MA, USA
Cody S. Nelson
Department of Medical Oncology, Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA, USA
Kelly P. Burke
Department of Immunology, Blavatnik Institute, Harvard Medical School, Boston, MA, USA
Kelly P. Burke & Arlene H. Sharpe
Unit of Nephrology, Dialysis and Renal Transplantation, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Milan, Italy
Manuel A. Podestà
Department of Cancer Immunology and Virology, Dana Farber Cancer Institute, Boston, MA, USA
Cole Batty & Wayne A. Marasco
Department of Surgery and Center for Vascular and Inflammatory Diseases, University of Maryland School of Medicine, Baltimore, MD, USA
Dejun Kong, Samuel J. Gavzy, Marina W. Shirkey & Jonathan S. Bromberg
TaiwanBio Therapeutics, Taipei, Taiwan
Cyrus Yang
Department of Cell Biology, Blavatnik Institute, Harvard Medical School, Boston, MA, USA
Marcia C. Haigis
Gene Lay Institute of Immunology and Inflammation, Brigham and Women’s Hospital, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA
Arlene H. Sharpe

Authors

Jeong-Mi Lee
View author publications
Search author on:PubMed Google Scholar
Paulo Lisboa Raeder
View author publications
Search author on:PubMed Google Scholar
Ridwan B. Rashid
View author publications
Search author on:PubMed Google Scholar
Hengcheng Zhang
View author publications
Search author on:PubMed Google Scholar
Cody S. Nelson
View author publications
Search author on:PubMed Google Scholar
Sierra G. Richardson
View author publications
Search author on:PubMed Google Scholar
Kelly P. Burke
View author publications
Search author on:PubMed Google Scholar
Pragya Chandrakar
View author publications
Search author on:PubMed Google Scholar
Manuel A. Podestà
View author publications
Search author on:PubMed Google Scholar
Maya G. Gempler
View author publications
Search author on:PubMed Google Scholar
Cole Batty
View author publications
Search author on:PubMed Google Scholar
Dejun Kong
View author publications
Search author on:PubMed Google Scholar
Samuel J. Gavzy
View author publications
Search author on:PubMed Google Scholar
Marina W. Shirkey
View author publications
Search author on:PubMed Google Scholar
Cyrus Yang
View author publications
Search author on:PubMed Google Scholar
Jonathan S. Bromberg
View author publications
Search author on:PubMed Google Scholar
Marcia C. Haigis
View author publications
Search author on:PubMed Google Scholar
Wayne A. Marasco
View author publications
Search author on:PubMed Google Scholar
Arlene H. Sharpe
View author publications
Search author on:PubMed Google Scholar
Peter T. Sage
View author publications
Search author on:PubMed Google Scholar

Contributions

P.T.S. and J.-M.L. conceptualized the project. J.-M.L., C.S.N. and M.A.P. developed bioinformatics. J.-M.L., P.L.R., H.Z., S.G.R., R.B.R., K.P.B., P.C. and M.G.G. performed the experiments. C.S.N., C.B., D.K., S.J.G., M.W.S. and J.S.B. provided resources. P.T.S. and J.-M.L. wrote the paper, which was reviewed and edited by all authors. P.T.S., A.H.S., W.A.M., M.C.H. and C.Y. provided technical support and funding.

Corresponding author

Correspondence to Peter T. Sage.

Ethics declarations

Competing interests

C.Y. is an employee of TaiwanBio Therapeutics. The other authors declare no competing interests.

Peer review

Peer review information

Nature Immunology thanks Luis Graca and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editor: S. Houston, in collaboration with the Nature Immunology team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Clonality of Tfr cells in superior and inferior regions of the spleen.

a) Schematic of sampling superior and inferior regions of the spleen after immunization. Mice were immunized with NP-OVA and the superior/inferior regions of spleens were analyzed on day 10. b) Pie charts depicting Tfr clonality in inferior and superior regions in each mouse (n = 3). TCR clones were identified based on identical CDR3 amino acid sequences. The number at the center of each pie chart indicates the total number of TCR clones detected. Black segments represent grouped singleton clones.

Extended Data Fig. 2 Additional analysis of Tfr stage progression over time.

a) Flow cytometry data from a kinetics experiment profiling Tfh and Tfr cells. Mice were administered NP-OVA adjuvanted with IFA intraperitoneally to induce follicular cells (D0, n = 4; D10, n = 3; D17, n = 5; D24, n = 5). Tfr cells were gated as CD4⁺CXCR5⁺Foxp3⁺, and Tfh cells were gated as CD4⁺Foxp3⁻CXCR5⁺PD-1⁺. P-value indicates results from one-way ANOVA. b) Identification of Tfr developmental stages by flow cytometry (n = 7). c) Differential gene expression analyses between indicated stage transitions. From the mouse hemisplenectomy scRNA-Seq data in Fig. 2, only the day 24 Tfr data was used for analyses using FindMarkers function of R Seurat package with the parameter setting of log2fc threshold = 0.5 and min.pct = 0.1. The dashed line indicates the threshold for a log2 fold change of 1 and an adjusted p-value of 0.05. d) Transcription factor regulatory network analysis. Day 24 Tfr and ExTfr cells from Fig. 4 were subject to SCENIC pipeline. The AUC scores of top 30 regulons (that is, transcription factor and its target genes) that have high variance of activities across all the cells are shown. e) Adoptive transfer of Progenitor/Effector Tfr cells. WT CD45.1 mice were vaccinated with NP-OVA and on day 10 spleens were used to sort Progenitor Tfr cells (CD4⁺CXCR5⁺GITR⁺CD62L⁺) and Effector Tfr cells (CD4⁺CXCR5⁺GITR⁺CD62L⁻PD-1⁺) (ProgTfr + Tfh, n = 2; Eff Tfr + Tfh, n = 3). Cells were transferred along with effector Tfh cells (CD4⁺CXCR5⁺GITR⁻PD-1⁺) into TCRa^−/− mice that were vaccinated with NP-OVA on the same day. Lymph nodes were assessed on d7 post vaccination. f) In vitro suppression assay in which progenitor or effector Tfr cells (gated as above) were cultured with Tfh and B cells (n = 3 wells per experimental group). GL7 expression was assessed on B cells. Statistical analyses were performed using one-way ANOVA for group comparisons. g) UMAP cluster distribution of Tfr cells at day 10 or day 24 in Fig. 2g separating replicate mice.

Extended Data Fig. 3 Additional assessment of human Tfr cell developmental stages in lymph nodes and blood.

a) Gating strategy and representative FACS plots for sorting follicular T cells from human lymph nodes and blood. b) (left) UMAP plots showing thirteen clusters of total CD4⁺CXCR5⁺ cells. (right) Feature plot of FOXP3 expression. Clusters 2 and 11 were defined as FOXP3⁺ (n = 692 cells) and used for downstream analyses of lymph node/blood Tfr. c) Relative abundance of Tfr clusters C1-C5 in each donor for human lymph nodes and blood as in Fig. 3a-b. d) Expression of indicated genes in human LN Tfr clusters taken from Fig. 3b. e) (left) UMAP plots showing eight clusters of total CD4⁺CXCR5⁺ cells from human blood vaccination dataset³². (right) Feature plot of FOXP3 expression. Cluster 7 (n = 434 cells) and the Tfr Cluster 3 (n = 5,471 cells) identified as FOXP3+ are indicated. The Tfr cluster 3 was used for downstream analyses. f) GSEA showing enrichment of TfhvsTcon or TfhvsTfr gene modules²⁵ in Cluster 7 versus Cluster 3. g) Dot plots showing gene markers for clusters B0-B5. h) (top) Representative flow cytometry plots of human blood Tfr cells gated as CD3⁺CD4⁺CXCR5⁺CD45RA⁻Foxp3⁺. (bottom) Relative percentages of total Tfr as a frequency of total CD4 + T cells or of total follicular T cells at indicated timepoints. Number indicates donor. The donors that showed the increase of vaccine-elicited clusters B2 and B4 are marked in dark red.

Extended Data Fig. 4 Additional assessment of ExTfr transcriptional programs and in vitro characteristics.

a) Heatmap showing key marker gene expression in developmental stages in Tfr and ExTfr cells. Marker genes identified using FindAllMarkers on day 24 Tfr data. Top 8 markers per stage (avg_log2FC > 2 and pct.1 > 0.2) displayed for side-by-side Tfr and ExTfr comparison. b) Pathway analysis of differentially expressed genes between Tfr and ExTfr in LateEffector or CD25 + /TNFR+Effector stages. DEGs with significance threshold 0.05 (Benjamini-Hochberg FDR) queried in gProfiler. c-f) FoxP3 reacquisition by ExTfr cells. Tfr/ExTfr cells from day 17 post NP-OVA vaccination were cultured. c) ExTfr with B cells plus anti-IgM (5 µg/mL) and anti-CD3/28 Dynabeads on day 4 (n = 3 wells). d) ExTfr with Tfh cells plus anti-CD3/28 Dynabeads on day 4 (n = 3 wells). e) ExTfr with B cells ± Tfh cells plus anti-CD3 (2 µg/mL) and anti-IgM (5 µg/mL) on day 6 (n = 3 wells). f) Tfr and ExTfr cultures with anti-CD3/28 Dynabeads (n = 4 ExTfr; n = 3 Tfr). Foxp3 mRNA measured by RT-PCR on day 4. g) In vitro Tfr instability. Tfr cultured with B cells anti-IgM (5 µg/mL), and varying anti-CD3/CD28 concentrations (0.2 n = 2; 2 n = 3; 2/0.5 n = 3; 2/5 n = 3) on day 4. (left) Representative Foxp3 expression in live CD4⁺MHC-II⁻ cells. (right) Quantification. h) B cell suppression assay. Tfr (CD4⁺CXCR5⁺tdTomato⁺GFP⁺) and ExTfr (CD4⁺CXCR5⁺tdTomato⁺GFP⁻) sorted from NP-OVA immunized Foxp3-fatemapper mice; Tfh cells (CD4⁺Foxp3⁻CXCR5⁺ICOS⁺) from Foxp3-GFP mice. Cultures contained 30,000 Tfh, 50,000 B cells, and 15,000 Effector Tfr or ExTfr cells with anti-CD3 (2 μg/mL) and anti-IgM (5 μg/mL) (No Tfh/Tfr n = 3; No Tfr n = 3; Effector Tfr n = 3; ExTfr n = 3), day 6. i) Apoptosis analysis. 30,000 Progenitor Tfr (CD62L⁺ Tfr), Effector Tfr (CD62L⁻PD-1⁺ Tfr), or ExTfr (CD4⁺CXCR5⁺tdTomato⁺GFP⁻) from immunized Foxp3-fatemapper mice cultured with anti-CD3/CD28 Dynabeads (Prog Tfr n = 2; Effector Tfr n = 2; ExTfr n = 3). Annexin-V and 7-AAD assayed day 3. (left) Representative plots. (right) Quantification.

Extended Data Fig. 5 Relative proportions and clonal features of Stable and Unstable Tfr clones.

a) (top) Relative proportions of Stable and Unstable clones of all TCRs detected at day 24 encompassing Tfr and ExTfr. (bottom) Relative proportions of Stable and Unstable clones for traveling clones (found in both d10 and d24 datasets). b) Individual Tfr clones at distinct timepoints and in Tfr and ExTfr compartments. The highlighted cells in each plot denote the cells from each time point that share the same TCR sequence.

Extended Data Fig. 6 Additional analysis of Tfr cells in Tfh-deleted mice.

a) Characterization of F/Ex-DTR mice. Control (Il21^CreRosa^YFPCxcr5^wt) or F/Ex-DTR (Il21^CreRosa^YFPCxcr5^{LoxSTOPLoxDTR}) were immunized, given diphtheria toxin on days 7-9, and dLN and spleen assessed on d10 as in Fig. 6a (Ctrl n = 5, F/Ex-DTR n = 3). b) Flow cytometry showing Tfh deletion and GC B cell frequencies. Mice vaccinated with NP-OVA, followed by diphtheria toxin treatment on days 7 and 9, and lymph nodes collected on day 10 (Ctrl n = 5, F/Ex-DTR n = 11). Unpaired Student’s t-test for statistical analysis. c) Flow cytometry of Tfr cells from F/Ex-DTR mice as in Fig. 6a-d and Extended Data Fig. 6a (Ctrl n = 4, F/Ex-DTR n = 5). Gating strategy shown in Extended Data Fig. 2b was used. Statistical significances determined using two-tailed Mann-Whitney test. d) Expanded clones of Tfr cells shown as blue on UMAP plot from Fig. 6a-d. Expanded clones defined as TCR clones found more than twice. e) Violin plots showing area under curve (AUC) scores for Late Effector 1 (Late1) and Late Effector 2 (Late2) in clusters C1-C8. f) Volcano plot showing differentially expressed genes between Control and ΔBlimp-1 Tfr cells. Genes identified as significant are highlighted in light orange, while C5 and C7 marker genes are marked in red. Horizontal dotted line represents significance threshold at padj = 0.1, and vertical lines indicate fold change thresholds of 1.5. g) Flow cytometry of Tfr subsets in Tfh-deleted experiments as in Fig. 6i (Control n = 7; Tfh-deleted n = 6). h) Feature plot of Tcf7 from the scRNAseq data of F/Ex-DTR described in Fig. 6a. i) Foxp3^ERT2CreTcf7^floxed mice immunized with NP-OVA, and lymph nodes were taken at day 10 post immunization for Tfr profiling (+/+ n = 9; fl/+ n = 7; fl/fl n = 5). Unpaired Student’s t-test for statistical analysis. j) Illustration summarizing the role of Tfh cell in controlling Tfr cell developmental stage progression and stability. k) Schematic summary of progressive differentiation of Tfr cells as well as roles of Tfh cells or Tfh programs in controlling fate decisions. For a, b, g, and i, unpaired two-tailed Student’s t-test was used for statistical analysis. Illustrations in j and k created using BioRender.com.

Supplementary information

Reporting Summary

Peer Review File

Source data

Source Data Figs. 1–6

Statistical source data.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Lee, JM., Raeder, P.L., Rashid, R.B. et al. Stability and progressive differentiation of T_FR cells are intrinsically and extrinsically controlled by T_FH programs. Nat Immunol (2026). https://doi.org/10.1038/s41590-025-02373-7

Download citation

Received: 09 September 2024
Accepted: 14 November 2025
Published: 13 January 2026
Version of record: 13 January 2026
DOI: https://doi.org/10.1038/s41590-025-02373-7