Lymphotoxin limits Foxp3+ regulatory T cell development from Foxp3lo precursors via IL-4 signaling

Borelli, Alexia; Santamaria, Jérémy C.; Zamit, Cloé; Apert, Cécile; Chevallier, Jessica; Pierre, Philippe; Argüello, Rafael J.; Spinelli, Lionel; Irla, Magali

doi:10.1038/s41467-024-51164-5

Lymphotoxin limits Foxp3⁺ regulatory T cell development from Foxp3^lo precursors via IL-4 signaling

Article
Open access
Published: 14 August 2024

Volume 15, article number 6976, (2024)
Cite this article

Download PDF

You have full access to this open access article

From

View current issue

Lymphotoxin limits Foxp3⁺ regulatory T cell development from Foxp3^lo precursors via IL-4 signaling

Download PDF

265 Accesses
5 Altmetric
Explore all metrics

Abstract

Regulatory T cells (T_reg) are critical players of immune tolerance that develop in the thymus via two distinct developmental pathways involving CD25⁺Foxp3⁻ and CD25⁻Foxp3^lo precursors. However, the mechanisms regulating the recently identified Foxp3^lo precursor pathway remain unclear. Here, we find that the membrane-bound lymphotoxin α₁β₂ (LTα₁β₂) heterocomplex is upregulated during T_reg development upon TCR/CD28 and IL-2 stimulation. We show that Lta expression limits the maturational development of T_reg from Foxp3^lo precursors by regulating their proliferation, survival, and metabolic profile. Transgenic reporter mice and transcriptomic analyses further reveal that medullary thymic epithelial cells (mTEC) constitute an unexpected source of IL-4. We demonstrate that LTα₁β₂-lymphotoxin β receptor-mediated interactions with mTEC limit T_reg development by down-regulating IL-4 expression in mTEC. Collectively, our findings identify the lymphotoxin axis as the first inhibitory checkpoint of thymic T_reg development that fine-tunes the Foxp3^lo T_reg precursor pathway by limiting IL-4 availability.

Essential and non-overlapping IL-2Rα-dependent processes for thymic development and peripheral homeostasis of regulatory T cells

Article Open access 04 March 2019

Recirculating IL-1R2+ Tregs fine-tune intrathymic Treg development under inflammatory conditions

Article 27 January 2020

Lymphotoxin α fine-tunes T cell clonal deletion by regulating thymic entry of antigen-presenting cells

Article Open access 28 March 2018

Introduction

Regulatory T cells (T_reg) constitute a subset of CD4⁺ T cells that specifically express the transcription factor Foxp3 and CD25, also known as IL-2Rα. Naturally occurring CD25⁺Foxp3⁺ T_reg critically prevent autoimmune and inflammatory disorders by maintaining self-tolerance through the suppression of autoreactive T cells that have escaped thymic selection. The majority of CD25⁺Foxp3⁺ T_reg emerge in the thymus from 3 days after birth¹. In particular, medullary thymic epithelial cells (mTEC) contribute to T_reg development by expressing MHC class II and CD80/86 molecules as well as a broad range of self-antigens^2,3,4. T_reg development follows a “two-step” model in the thymic medulla^5,6,7. The first step is driven by the stimulation of the T-cell receptor (TCR) and of CD28 in developing CD4⁺ single-positive (SP) thymocytes, leading to the generation of CD25⁺Foxp3⁻ precursors (CD25⁺ T_regP)^5,6,8. TCR and CD28 signals govern the expression of tumor-necrosis factor receptor (TNFR) members, GITR and OX40, which promote CD25 expression⁹. The second step is driven by γ-chain cytokines, in particular IL-2 and IL-15, that convert CD25⁺ T_regP into CD25⁺Foxp3⁺ mature T_reg⁵. In addition to Foxp3 upregulation, IL-2 also modulates GITR and OX40 expression in mature T_reg¹⁰. IL-2 was found to be provided by autoreactive CD4⁺ SP thymocytes¹¹, whereas IL-15 is produced by mTEC¹². More recently, a second CD25^-Foxp3^lo precursor (Foxp3^lo T_regP), lacking CD25 and expressing low levels of Foxp3, was identified¹³. Similarly to CD25⁺ T_regP, this precursor has the ability to generate CD25⁺Foxp3⁺ mature T_reg upon intrathymic transfer^14,15. The differentiation of Foxp3^lo T_regP into mature T_reg also follows the “two-step” model^9,15. Although Foxp3^lo T_regP make a substantial contribution to the mature T_reg pool, the mechanisms controlling this recently identified developmental pathway remain elusive. Peripheral Foxp3⁺ T_reg were shown to express the membrane-bound lymphotoxin α₁β₂ (LTα₁β₂) heterocomplex, belonging to the TNF superfamily, which interacts with its unique cognate LTβ receptor (LTβR)^16,17,18. However, it remains unknown whether LTα₁β₂ expression is induced in thymic T_reg and whether it controls their generation.

Here, we show that LTα₁β₂ gradually increases during the “two-step” model of thymic T_reg development. Its upregulation depends on TCR/CD28 stimulation in CD25⁺ and Foxp3^lo T_regP, and on IL-2 stimulation in CD25⁺Foxp3⁺ mature T_reg. Interestingly, we observed an increased development of Foxp3^lo T_regP and mature T_reg in Foxp3^eGFPxLta^−/− mice, indicating that Lta negatively regulates CD25⁺Foxp3⁺ T_reg generation from Foxp3^lo T_regP. This phenotype was associated with an increased proliferation and survival of Foxp3^lo T_regP and mature T_reg, as well as an altered metabolic profile. Strikingly, the maturational development of Foxp3^lo T_regP and CD25⁺Foxp3⁺ T_reg increased in these mice. We show that IL-4 stimulation substantially promotes the conversion of Foxp3^lo T_regP into CD25⁺Foxp3⁺ mature T_reg and that mTEC constitute an unexpected source of IL-4. We further demonstrate that the LTα₁β₂-LTβR axis negatively regulates the generation of CD25⁺Foxp3⁺ T_reg by limiting IL-4 expression in mTEC. Altogether, this study reveals that LTα₁β₂-LTβR interactions with mTEC fine-tunes the Foxp3^lo T_regP developmental pathway in an IL-4-dependent manner.

Results

LTα₁β₂ expression is upregulated during thymic T_reg development

To characterize LTα₁β₂ expression during thymic T_reg development, we used Rag2^GFPxFoxp3^Thy1.1 mice expressing the green fluorescent protein (GFP) and the membrane-bound Thy1.1 under the control of Rag2 and Foxp3 promoters, respectively^19,20. These mice allowed us to discriminate newly generated thymocytes (Rag2^GFP+) from more mature T_reg that have been retained within the thymus or have recirculated from the periphery (Rag2^GFP−)^21,22,23. Among Rag2^GFP+ thymocytes, we identified newly produced CD4⁺ SP thymocyte subsets: CD4⁺ T_conv (CD25⁻Foxp3⁻), CD25⁺ T_regP (CD25⁺Foxp3⁻), Foxp3^lo T_regP (CD25⁻Foxp3^lo) and mature T_reg (CD25⁺Foxp3⁺) (Fig. 1a). The expression of the membrane-bound LTα₁β₂ heterocomplex, detected using the soluble LTβR-Fc receptor, was higher in CD25⁺ T_regP and Foxp3^lo T_regP compared to CD4⁺ T_conv (Fig. 1b, c). This expression was even further increased in newly produced CD25⁺Foxp3⁺ T_reg. In contrast, the soluble LTβR-Fc receptor did not bind to thymic T_reg subsets from Lta^−/− mice (Fig. 1b). We also used the differential expression of CCR6, which discriminates between newly produced T_reg (CCR6⁻) and recirculating T_reg (CCR6⁺) in the thymus⁴ and confirmed Lta and Ltb upregulation during T_reg development using Foxp3^eGFP reporter mice, in which an enhanced GFP reporter construct was introduced into the Foxp3 locus (Supplementary Fig. 1). Interestingly, the gradual upregulation of LTα₁β₂ expression during T_reg development correlated with that of OX40 (Tnfrsf4) and GITR (Tnfrsf18) (Fig. 1d and Supplementary Fig. 1c). Furthermore, cells with the highest CD25 expression also had the highest Foxp3 expression in newly produced mature T_reg and LTα₁β₂ level tightly correlated with CD25 and Foxp3 levels (Fig. 1e,f). These results show that LTα₁β₂ is gradually upregulated during T_reg development in the thymus.

**Fig. 1: Membrane-bound LTα₁β₂ expression increases during thymic T_reg development concomitantly with OX40 and GITR.**

LTα₁β₂ is upregulated by instructive signals of the “two-step model” of T_reg development

Since we observed a tight correlation between LTα₁β₂ and Foxp3 levels, we then asked whether LTα₁β₂ upregulation could be related to TCR stimulation. To address this question, we used the transcription factor Nur77 as a reliable readout of TCR signal strength²⁴ and found a significant positive correlation between LTα₁β₂ and Nur77 in CD25⁺ T_regP, Foxp3^lo T_regP and mature T_reg (Fig. 2a and Supplementary Fig. 2a)_. We also analyzed LTα₁β₂ expression in T_reg subsets stimulated in vitro with increasing doses of anti-CD3ε monoclonal antibody, which upregulated LTα₁β₂ expression in both CD25⁺ T_regP and Foxp3^lo T_regP (Fig. 2b and Supplementary Fig. 2b). In contrast, CD3ε stimulation was unable to upregulate LTα₁β₂ expression in mature T_reg. We next assessed the effect of CD28 costimulation on LTα₁β₂ expression. To this end, we used Cd28^−/− mice, which showed decreased frequencies and numbers of CD25⁺ T_regP, Foxp3^lo T_regP, and mature T_reg (Supplementary Fig. 3a). Similarly to OX40 and GITR, LTα₁β₂ was reduced in these cells (Fig. 2c and Supplementary Fig. 3b,c), indicating that CD28 costimulation is implicated in LTα₁β₂ upregulation in T_reg subsets.

**Fig. 2: LTα₁β₂ expression is upregulated during thymic T_reg development by TCR/CD28 and IL-2 signals.**

Given that IL-2 and IL-15 are crucial for the conversion of T_regP into CD25⁺Foxp3⁺ mature T_reg^5,10,25_, we hypothesized that these cytokines could also be implicated in LTα₁β₂ upregulation. To address their respective contribution, we used Il2^KO or Il15^KO mice backcrossed with Rag2^GFPxFoxp3^Thy1.1 mice to analyze newly generated T_reg. In contrast to Il15^KOxRag2^GFPxFoxp3^Thy1.1 (Il15^−/−) mice (Fig. 2d), LTα₁β₂ expression was strongly altered in mature T_reg from Il2^KOxRag2^GFPxFoxp3^Thy1.1 (Il2^−/−) mice (Fig. 2e). Because thymic IL-2 is important for T_reg development and survival^13,26, we next sought to demonstrate that IL-2 directly regulates LTα₁β₂ expression in mature T_reg. To this end, purified CD25⁺ T_regP, Foxp3^lo T_regP and mature T_reg were stimulated in vitro with increasing doses of IL-2. Interestingly, LTα₁β₂ was upregulated in an IL-2 dose-dependent manner in mature T_reg (Fig. 2f and Supplementary Fig. 2c). Altogether, these results show that LTα₁β₂ upregulation is initially induced by TCR/CD28 signals in T_regP and further upregulated by IL-2 stimulation in mature T_reg and thus, controlled by the signals of the “two-step model” of T_reg development.

LTα₁β₂ limits T_reg generation through the Foxp3^lo T_regP developmental pathway

To test whether Lta could be implicated in T_reg development, we backcrossed Lta^−/− mice with Foxp3^eGFP mice (Foxp3^eGFPxLta^−/− mice), in which membrane-bound LTα₁β₂ expression is fully abolished¹⁸. Remarkably, in contrast to CD25⁺ T_regP, frequencies and numbers of Foxp3^lo T_regP and mature T_reg were substantially higher in 6-week-old Foxp3^eGFPxLta^−/− mice compared to Foxp3^eGFP controls (Fig. 3a, b). This phenotype was also observed in the thymus of 3-day-old neonates, indicating that Lta regulates the development of Foxp3^lo T_regP and mature T_reg from their emergence (Fig. 3c, d). Moreover, 4-month-old adult Foxp3^eGFPxLta^−/− mice also showed an increased representation of these cells, suggesting that Lta controls T_reg development throughout the lifespan. Since we previously reported an enhanced clonal deletion of CD4⁺ T_conv in Lta^−/− mice²⁷, we further analyzed the impact of Lta specifically on developing cells within the T_reg lineage (Supplementary Fig. 4a). In accordance with the increased cellularity of Foxp3^lo T_regP and mature T_reg (Fig. 3a–d), frequencies of these cells were enhanced in the T_reg lineage of Foxp3^eGFPxLta^−/− mice (Supplementary Fig. 4a). In line with this phenotype, frequencies and numbers of Ki67⁺ proliferating cells were increased in Foxp3^lo T_regP and mature T_reg subsets of Foxp3^eGFPxLta^−/− mice (Fig. 3e). Moreover, frequencies of annexin V⁺ were reduced in Foxp3^lo T_regP from Foxp3^eGFPxLta^−/− mice (Fig. 3f). Lower levels of the pro-apoptotic protein BIM, which is required for the apoptosis of autoreactive thymocytes^28,29, were also observed in Foxp3^lo T_regP and mature T_reg of these mice (Fig. 3g). In contrast, Ki67⁺ proliferating cells were slightly increased while annexin V⁺ cells and BIM levels were unchanged in CD25⁺ T_regP (Supplementary Fig. 4c–e). Thus, the increase of Foxp3^lo T_regP and mature T_reg in Foxp3^eGFPxLta^−/− mice likely relies on enhanced proliferation and survival.

**Fig. 3: *Lta* negatively regulates T_reg maturational development from Foxp3^lo T_regP.**

Given that cells of the T_reg lineage are heterogeneous, we further analyzed their maturational state in Foxp3^eGFPxLta^−/− mice using a multiparametric flow cytometry analysis with developmental markers (Fig. 3h). We performed an unsupervised clustering that revealed a high maturational heterogeneity within the T_reg lineage (Fig. 3i). Cell subsets were classified according to the relative expression of the maturational markers used (Fig. 3j). Subsets 1–5 corresponded to CD25⁺ T_regP, subsets 6–8 to Foxp3^lo T_regP and subsets 9–10 to mature T_reg (Fig. 3i, j). In Foxp3^eGFPxLta^−/− mice, CD25⁺ T_regP subsets (clusters 1–4) were under-represented, while two immature Foxp3^lo T_regP subsets (clusters 6,7) were unchanged (Fig. 3k). In contrast, the Foxp3^lo T_regP cluster 8 expressing high levels of H2K^b and Qa2 and low levels of CD69 and CD24, likely corresponding to more mature transitional Foxp3^lo T_regP, was over-represented in Foxp3^eGFPxLta^−/− mice. Strikingly, we found an increased representation of mature T_reg subsets (clusters 9, 10) in Foxp3^eGFPxLta^−/− mice. These results reveal a bias in the T_reg maturational development via Foxp3^lo T_regP in Foxp3^eGFPxLta^−/− mice.

Because the LTα₁β₂-LTβR axis controls thymic medulla organization^17,30, we analyzed the contribution of non-hematopoietic cells in T_reg development by generating bone marrow (BM) chimeras in which lethally irradiated CD45.2 WT or Lta^−/− recipients were reconstituted with CD45.1 Foxp3^eGFP BM cells (Supplementary Fig. 5a). Six weeks later, similar frequencies and numbers of CD25⁺ T_regP, Foxp3^lo T_regP, and mature T_reg were observed in the thymus of WT and Lta^−/− recipients (Supplementary Fig. 5b), indicating that the stromal environment of Lta^−/− mice is unlikely implicated in the enhanced T_reg development. Given that Lta is expressed by hematopoietic cells¹⁷, we next wondered whether a partial reintroducion of WT BM cells with Lta^−/− BM cells could be sufficient to reverse the phenotype observed in Lta^−/− mice. To this end, we generated mixed BM chimeras in which lethally irradiated CD45.1/2 WT recipients were reconstituted with BM cells from either CD45.1 Foxp3^eGFP or Foxp3^eGFPxLta^−/− mice in competition with CD45.2 WT BM cells (Fig. 4a). Six weeks later, as opposed to CD25⁺ T_regP, we found increased frequencies of Foxp3^lo T_regP and mature T_reg in Foxp3^eGFPxLta^−/− + WT → WT chimeras (Fig. 4b). Moreover, we observed enhanced frequencies of Foxp3^lo T_regP and mature T_reg by specifically analyzing cells of the T_reg lineage (Fig. 4c). In line with these observations, we found slightly decreased frequencies of Ki67⁺ CD25⁺ T_regP while Foxp3^lo T_regP proliferation increased (Fig. 4d). We next analyzed the maturational state of cells from the T_reg lineage using a multiparametric flow cytometry analysis (Fig. 4e). We performed an unsupervised clustering that revealed that CD25⁺ T_regP subsets (clusters 2, 5 and 6) were under-represented while clusters 7 and 8 corresponding to Foxp3^lo T_regP and clusters 10 corresponding to mature T_reg were increased in Foxp3^eGFPxLta^−/− + WT → WT chimeras (Fig. 4f–h). Altogether, these data confirm that Lta expression in thymocytes negatively regulates the Foxp3^lo T_regP developmental pathway, which could not be rescued by the reintroduction of WT cells.

**Fig. 4: Hematopoietic origin of *Lta* expression regulates thymic T_reg development.**

Lta expression controls the metabolic profile of T_reg during their thymic development

To define the impact of Lta on T_reg development, we performed 3′-end single-cell RNA sequencing (scRNA-seq) on purified T_reg lineage cells from Foxp3^eGFP and Foxp3^eGFPxLta^−/− mice (Fig. 5a). Our analysis revealed that these cells separated into five distinct clusters (Fig. 5b, c). Ccr7 expression was detected in all clusters, indicating that both T_regP and mature T_reg are located in the thymic medulla (Fig. 5d and Supplementary Fig. 6a). Clusters 0 and 1 exhibited an immature phenotype, characterized by the expression of Cd24a and Cd69, with no or poor Foxp3, Sell (CD62L) and S1pr1 expression. While cluster 0 expressed Il2rb (CD122), Il2rg (CD132) and Tnfrsf18 (GITR), cluster 1 also expressed Il2ra (CD25). This indicates that cluster 0 and 1 corresponded to T_reg pre-precursors (CD122⁺CD25⁻Foxp3⁻)³¹ and CD25⁺ T_regP, respectively. Cluster 2 expressed low levels of Foxp3 and Il2ra in addition to high levels of TCR-signaling associated genes (Nr4a1, Nr4a3, and Tnfrsf4) (Fig. 5c, d), suggesting a transitional state towards mature T_reg (transitional T_reg). Cluster 3 expressed low levels of Foxp3 while Il2ra was not detected, suggesting that cells from this cluster contained Foxp3^lo T_regP. All cells from cluster 3 expressed H2-K1 (H2K^b) while a fraction expressed Cd69 (Fig. 5d and Supplementary Fig. 6a). In line with this observation, Foxp3^lo T_regP contained mature CD69⁺H2K^b+ (M1) and CD69⁻H2K^b+ (M2) cells while CD25⁺ T_regP contained semi-mature CD69⁺H2K^b− (SM), M1 and M2 cells (Supplementary Fig. 6b), indicating that Foxp3^lo T_regP have a more mature phenotype than CD25⁺ T_regP. Accordingly, Foxp3^lo T_regP resemble mature T_reg based on the differential expression of CD69 and H2K^b. Furthermore, we cannot exclude that cluster 3 included mature cells that could exit the thymus since high levels of Sell, Klf2, and S1pr1 were detected. In line with this observation, we found that Foxp3^lo T_regP expressed high level of CD62L by flow cytometry (Supplementary Fig. 6c). Cluster 4 expressed the highest levels of Foxp3, Il2ra, Il2rb, and Il2rg, suggesting that this cluster corresponded to CD25⁺Foxp3⁺ mature T_reg (Fig. 5d and Supplementary Fig. 6a), consistently with a high expression of Sell, Klf2 and S1pr1.

**Fig. 5: Single-cell RNA sequencing reveals that *Lta* deficiency affects the metabolic program of developing T_reg subsets.**

We next performed trajectory analysis using the Monocle 3 package³² to assess the developmental trajectory within the T_reg lineage. Three major branches were identified, starting at roots Y1, Y30, and Y2 all of which converged to Y16 located in cluster 4 corresponding to mature T_reg. (Supplementary Fig. 7a). The branch from Y1 to Y16 indicated a potential conversion of cells from cluster 0, to cluster 1, then to cluster 2, and finally to cluster 4. The Y30 to Y16 branch revealed another possible developmental route from cluster 0 directly to cluster 4. The branch from Y2 to Y16 indicated a short transition from cluster 3 to cluster 4. In line with the Seurat analysis (Fig. 5b–d), the progressive upregulation of Foxp3 and Il2ra was associated with an upregulation of genes related to terminal maturation (H2-K1, Sell and S1pr1) along the pseudotime axis (Supplementary Fig. 7b).

We then compared changes in distinct cellular pathways in the different clusters of Foxp3^eGFP and Foxp3^eGFPxLta^−/− mice to analyze the impact of Lta on T_reg development (Fig. 5e). We calculated module scores of gene sets taken from the KEGG database and found metabolic pathways substantially altered by Lta deficiency. Although metabolic reprogramming regulates T_reg function in the periphery^33,34, the metabolic profile of thymic T_reg remains unknown. Our analysis revealed metabolic changes associated to T_reg development (Fig. 5f and Supplementary Fig. 7c). Interestingly, the module score for the oxidative phosphorylation (OXPHOS) gene set increased from cluster 0 (CD122⁺CD25⁻Foxp3⁻ pre-T_regP) to cluster 1 (CD25⁺ T_regP), then decreased in cluster 2 (Transitional T_reg) and further decreased in cluster 4 (Mature T_reg). We also found the module scores for glycolysis/gluconeogenesis, biosynthesis, and nucleotide metabolic pathways to be higher in cluster 1 compared to all other clusters, suggesting that CD25⁺ T_regP have a high metabolic activity. In line with these observations, cluster 1 highly expressed both OXPHOS (Ndufa4, Ndufc2, and Atp5b) and glycolysis (Ldha, Pgk1, and Tpi1) hallmark genes (Fig. 5g), while clusters 3 and 4 expressed mainly OXPHOS hallmark genes. On the other hand, there is likely no implication of the fatty acid metabolism during T_reg development suggested by the module score for this pathway (Fig. 5f). Compared to Foxp3^eGFP mice, clusters 2, 3 and 4 of Foxp3^eGFPxLta^−/− mice showed an enhanced score for OXPHOS metabolism, indicating that Lta expression regulates the metabolic profile of the most mature T_reg subsets (Fig. 5f and Supplementary Fig. 7d).

Given that scRNA-seq data suggest gene expression changes of OXPHOS-related genes during T_reg development (Fig. 5f), we then analyzed mitochondrial activity. In line with the transcriptomic analysis, CD25⁺ T_regP displayed higher mitochondrial activity compared to Foxp3^lo T_regP and mature T_reg, as shown by MitoTracker Deep Red staining (Fig. 6a). CD25⁺ T_regP and Foxp3^lo T_regP of Foxp3^eGFPxLta^−/− mice showed fewer active mitochondria compared to their respective control counterparts (Fig. 6b). Since this functional defect in mitochondrial polarization was unforeseen in scRNA-seq data, we further analyzed the metabolic profile of developing T_reg using a functional assay. Given that low numbers of T_regP can be retrieved from the thymus, we took advantage of the newly developed flow-cytometry-based method, SCENITH™ (Single-Cell Metabolism by Profiling Translation Inhibition)³⁵, to analyze the metabolic profiles of T_regP and mature T_reg. This method is based on metabolism-dependent translation rates and puromycin incorporation into nascent proteins in response to specific metabolic inhibitors. Thus, SCENITH™ permits the assessment of glucose dependency, mitochondrial dependency, glycolytic capacity as well as fatty acid and amino acid oxidation capacity in multiple rare cell types simultaneously. In line with our transcriptomic (Fig. 5f) and MitoTracker analysis (Fig. 6a), CD25⁺ T_regP showed higher levels of protein synthesis than Foxp3^lo T_regP and mature T_reg (Supplementary Fig. 8a), further confirming that CD25⁺ T_regP are more metabolically active. Strikingly, CD25⁺ T_regP and Foxp3^lo T_regP showed a distinct metabolic profile (Fig. 6c and Supplementary Fig. 8b), with Foxp3^lo T_regP being more similar to mature T_reg. Specifically, CD25⁺ T_regP showed 50% dependency on both glucose and mitochondrial OXPHOS, while Foxp3^lo T_regP showed 85% mitochondrial dependency similar to the 75% mitochondrial dependency of mature T_reg (Fig. 6c). These results are in accordance with the OXPHOS and glycolysis/gluconeogenesis module scores calculated using our scRNA-seq analysis at steady state (Fig. 5f). Interestingly, we found that Foxp3^lo T_regP and mature T_reg from Foxp3^eGFPxLta^−/− mice exhibited an exacerbated glucose dependency (Fig. 6d and Supplementary Fig. 8c, d).

**Fig. 6: LTα₁β₂ expression regulates T_reg glucose dependence during their development.**

Next, we further explored the metabolic profiles of the distinct maturational stages of T_regP and mature T_reg cells using CD69 and H2K^b developmental markers (Fig. 6e)³⁶. In wild-type conditions, the mitochondrial dependence increased while the glycolytic capacity decreased during CD25⁺ T_regP maturation (Fig. 6f). We further found a distinct metabolic profile for the M1 subset within CD25⁺ and Foxp3^lo T_regP, whereby M1 Foxp3^lo T_regP exhibited a higher mitochondrial dependence. Notably, CD25⁺ and Foxp3^lo T_regP M2 subsets had a similar metabolic profile compared to M1 and M2 mature T_reg whose profile was characterized by a high mitochondrial dependency and a low glycolytic capacity. A similar profile was also observed in the Foxp3^lo T_regP M1 subset. These results suggest that the increase in mitochondrial dependency is tightly linked to T_reg maturation. In Foxp3^eGFPxLta^−/− mice, CD25⁺ T_regP SM cells showed an increased mitochondrial dependence while the M2 cells exhibited an increased glucose dependence (Fig. 6g). Moreover, the glucose dependence was also increased in Foxp3^lo T_regP and mature T_reg at M1 and M2 stages. Thus, Foxp3^lo T_regP display a similar metabolic profile to the most mature T_reg (M2) and Lta expression negatively regulates their glucose dependency. Overall, these findings show that LTα modulates the metabolic profile of T_reg along their maturational development.

LTα₁β₂-regulated T_reg development depends on IL-4 production by mTEC

We then investigated the underlying mechanism of LTα₁β₂-LTβR-regulated T_reg development. We made the hypothesis that the LTα₁β₂-LTβR axis could regulate the expression of cytokines. Consistent with a previous report¹⁵, we found that Il4ra was preferentially expressed in Foxp3^lo T_regP compared to CD25⁺ T_regP, and its expression was further upregulated in mature T_reg at both mRNA and protein levels (Fig. 7a, b). In contrast, cells of the T_reg lineage did not produce IL-4, ruling out an autocrine stimulation (Fig. 7a). This expression pattern suggests that IL-4 may support the conversion of Foxp3^lo T_regP into CD25⁺Foxp3⁺ mature T_reg. To test this hypothesis, purified CD25⁺ T_regP and Foxp3^lo T_regP were stimulated with increasing doses of IL-4. In contrast to CD25⁺ T_regP, IL-4 stimulation promoted the robust conversion of Foxp3^lo T_regP into CD25⁺Foxp3⁺ mature T_reg (Fig. 7c and Supplementary Fig. 9). Interestingly, we found increased Il4 expression in the thymus of Lta^−/− mice, which was associated with enhanced development of IL-4-dependent Eomes⁺ memory-like CD8⁺ SP thymocytes^37,38 (Fig. 7d, e). Furthermore, frequencies and numbers of phospho-Stat6⁺ cells were substantially enhanced, further highlighting an IL-4-rich environment in the thymus of Lta^−/− mice (Supplementary Fig. 10a). Among these cells, frequencies of phospho-Stat6⁺ mature T_reg were increased in Lta^−/− mice (Supplementary Fig. 10b), suggesting that enhanced IL-4 expression affects mature T_reg. In line with enhanced Foxp3^lo T_regP and mature T_reg in Foxp3^eGFPxLta^−/− + WT → WT chimeras (Fig. 4), Il4 expression was higher in the thymus of those recipients, which was associated with increased frequencies of phospho-Stat6⁺ thymocytes (Fig. 7f and Supplementary Fig. 10c). Moreover, in line with the reintroduction of WT BM cells in these chimeras, IL-4 and phospho-Stat6 levels increased by only half of those observed in Lta^−/− mice. This partial increase in Il4 expression was sufficient to enhance the frequencies of Foxp3^lo T_regP and mature T_reg of CD45.2 WT origin (Fig. 7g). This indicates that the Foxp3^lo T_regP developmental pathway is highly sensitive to subtle variations in IL-4 levels. To investigate the role of IL-4 in vivo in T_reg development, Lta^−/− pups were injected with an anti-IL-4 neutralizing antibody during the emergence of T_regP at post-natal day 0 and 3 (Fig. 7h). Five days after birth, Eomes⁺ memory-like CD8⁺ SP thymocytes were reduced, showing the effectiveness of IL-4 neutralization (Fig. 7i). Interestingly, frequencies and numbers of Foxp3^lo T_regP and CD25⁺Foxp3⁺ mature subsets were substantially reduced while CD25⁺ T_regP were over-represented (Fig. 7j). These results suggest that IL-4 favors the Foxp3^lo T_regP developmental pathway at the expense of the CD25⁺ T_regP pathway.

**Fig. 7: IL-4 controls the Foxp3^lo T_regP developmental pathway in the thymus.**

Although type 2 natural killer T cells (NKT2) are an important source of IL-4 in the thymus^37,38,39, increased expression of Il4 in Lta^−/− mice was unlikely due to NKT2 cell enrichment (Supplementary Fig. 10d), suggesting another cellular source of IL-4. We therefore explored whether other thymic cells may express Il4. Using a scRNA-seq dataset⁴⁰, we found that Il4 was not expressed by thymic DC and macrophages (Fig. 8a). We next investigated the possibility that mTEC could produce this cytokine and found that IL-4 was mainly expressed by Aire⁺ mTEC^hi and late-Aire mTEC in scRNA-seq datasets (Fig. 8b). We confirmed by qPCR that the level of Il4 was higher in mTEC expressing high levels of the CD80 costimulatory molecule (mTEC^hi) than in mTEC expressing low levels of CD80 (mTEC^lo) and that the level in mTEC^hi was similar to that observed in NKT cells (Fig. 8c). Using IL-4 reporter mice on a C57BL/6 or BALB/c background, IL-4 expression was detected both in NKT cells and mTEC (Fig. 8d and Supplementary Fig. 10e), confirming that mTEC constitute a substantial source of IL-4 in the thymus. Given that the classical NF-κB member p65 (RelA) was reported to upregulate Il4 expression by binding to its promoter⁴¹, we assessed its expression in mTEC^hi of Lta^−/− mice and found enhanced protein levels of p65 (Supplementary Fig. 10f). This suggests that the increased expression of p65 may lead to elevated Il4 expression in mTEC^hi of Lta^−/− mice.

**Fig. 8: mTEC are a novel source of IL-4 in the thymus.**

Given that mTEC control the generation of CD25⁺Foxp3⁺ T_reg^2,3,42,43,44, we explored the expression of LTβR within mTEC subsets by reanalysing TEC scRNA-seq datasets^45,46. Ltbr was widely expressed in mTEC subsets (Supplementary Fig. 11a). We confirmed by flow cytometry this wide expression in mTEC^lo and mTEC^hi (Supplementary Fig. 11b,c). Next, we analyzed the TEC scRNA-seq with our thymic T_reg scRNA-seq datasets to identify possible cell-cell interactions via the LTα₁β₂-LTβR axis using CellChat⁴⁷. CD25⁺ T_regP and Foxp3^lo T_regP likely interact with several mTEC subsets (Supplementary Fig. 11d). These interactions seemed to be stronger with transitional and mature T_reg, consistent with the gradual increase of LTα₁β₂ expression during T_reg development (Fig. 1c and Supplementary Fig. 1c).

Next, to determine the contribution of mTEC in LTα₁β₂-LTβR-regulated T_reg development, we used OTII transgenic mice that possess CD4⁺ SP thymocytes expressing an MHCII-restricted TCR specific for the chicken ovalbumin (OVA). First, we verified that OTII cells transit through the CD25⁺ and Foxp3^lo T_regP developmental pathways. To this end, we took advantage of RipmOVAxOTIIxRag2^−/− mice in which the Rat Insulin Promoter (Rip) drives the expression of the membrane-bound OVA (mOVA) form specifically in mTEC. In these transgenic mice, high-affinity interactions between OVA-expressing mTEC and OTII thymocytes are possible in vivo. Whereas Foxp3^lo T_regP and CD25⁺Foxp3⁺ mature T_reg were undetectable in OTIIxRag2^−/− mice, these two T_reg subsets developed in RipmOVAxOTIIxRag2^−/− mice, indicating that OTII thymocytes transit through Foxp3^lo T_regP toward mature T_reg when the cognate self-antigen is expressed in mTEC (Fig. 9a). We then performed in vitro antigen-specific T_reg conversion assays in which WT mTEC loaded with the OVA_323-339 peptide were co-cultured with CD69⁺ semi-mature OTII CD4⁺ SP thymocytes, that expressed the membrane-bound LTα₁β₂ and were shown to contain cells with a high ability to differentiate into mature T_reg⁴² (Fig. 9b, c). After 18 h of co-culture, mTEC presenting the cognate antigen had the capacity to induce Foxp3^lo T_regP, further confirming that OTII thymocytes transit via a Foxp3^lo T_regP stage (Fig. 9d). To further investigate the role of Lta, we performed this experiment with CD69⁺CD4⁺ SP thymocytes purified from OTII mice backcrossed on an Lta^−/− background (OTIIxLta^−/− mice). Interestingly, the conversion in Foxp3^lo T_regP increased while that of CD25⁺ T_regP decreased with CD69⁺ CD4⁺ from OTIIxLta^−/− mice compared to OTII mice. Moreover, blocking LTα₁β₂-LTβR interactions using the soluble LTβR-Fc receptor, also resulted in an increased Foxp3^lo T_regP conversion and a decreased CD25⁺ T_regP conversion (Fig. 9e, f), indicating that the lymphotoxin axis limits the Foxp3^lo T_regP developmental pathway in favor of the CD25⁺ T_regP pathway. As previously reported⁴², mTEC had the capacity to induce antigen-specific CD25⁺Foxp3⁺ mature T_reg after 5 days of co-culture (Fig. 9g). We found a higher conversion in CD25⁺Foxp3⁺ mature T_reg when OVA_323-339-loaded mTEC were co-cultured with CD69⁺CD4⁺ SP thymocytes from OTIIxLta^−/− mice compared to OTII mice. A similar enhanced T_reg generation was observed irrespective of the mTEC maturation state (Fig. 9h). Moreover, blocking LTα₁β₂-LTβR interactions, also resulted in an increased mature T_reg conversion after 5 days of co-culture (Fig. 9i). These results show that LTα₁β₂-LTβR interactions with mTEC act as a negative regulator of CD25⁺Foxp3⁺ mature T_reg conversion via the Foxp3^lo T_regP developmental pathway. To further define whether the LTα₁β₂-LTβR axis controls the conversion of Foxp3^lo T_regP into CD25⁺Foxp3⁺ mature T_reg, we backcrossed RipmOVAxOTIIxRag2^−/− mice with the Foxp3^eGFP reporter mice, which allowed us to purified both T_regP. We performed antigen-specific co-culture assays using OVA_323-339-loaded mTEC with CD25⁺ or Foxp3^lo T_regP purified from RipmOVAxOTIIxFoxp3^eGFP mice and treated them or not with the soluble LTβR-Fc receptor (Fig. 9j and Supplementary Fig. 12). The addition in the co-culture of LTβR-Fc receptor caused a greater conversion into CD25⁺Foxp3⁺ mature T_reg by Foxp3^lo T_regP, but not by CD25⁺ T_regP (Fig. 9k, l). Altogether, these data reveal that the LTα₁β₂-LTβR axis acts as a feedback mechanism by which T_regP can limit their own differentiation.

**Fig. 9: The LTα₁β₂-LTβR axis provided by mTEC limits Foxp3^lo T_reg development in an IL-4 dependent manner.**

We next evaluated whether the LTα₁β₂-LTβR axis regulates Il4 expression in mTEC in these co-culture assays. Interestingly, we found higher levels of Il4 in OVA_323-339-loaded mTEC co-cultured with CD69⁺CD4⁺ SP thymocytes from OTIIxLta^−/− compared to OTII mice, demonstrating that Lta negatively regulates the expression of Il4 in mTEC (Fig. 9m). Consistently, Rela was also higher in co-cultures with CD69⁺CD4⁺ SP thymocytes from OTIIxLta^−/− mice. To test the hypothesis that Il4 overexpression observed in the absence of Lta may be responsible for the enhanced conversion of CD69⁺CD4⁺ SP thymocytes into mature T_reg (Fig. 9g), we used a neutralizing anti-IL-4 antibody in these co-culture assays. Strikingly, the increased conversion of CD69⁺CD4⁺ SP thymocytes from OTIIxLta^−/− mice into CD25⁺Foxp3⁺ T_reg was fully abolished by IL-4 neutralization (Fig. 9n). Overall, these results demonstrate that the LTα₁β₂-LTβR axis negatively regulates CD25⁺Foxp3⁺ T_reg conversion by repressing IL-4 expression in mTEC.

Discussion

Recent advances have revealed that CD25⁺Foxp3⁺ mature T_reg arise from two distinct developmental pathways involving CD25⁺ and Foxp3^lo T_regP^7,14,15,48. Although the CD25⁺ T_regP pathway is well studied, the more recently identified Foxp3^lo T_regP pathway remains poorly investigated. Our study reveals that the LTα₁β₂-LTβR axis plays an unexpected role in limiting CD25⁺Foxp3⁺ T_reg generation via the Foxp3^lo T_regP developmental pathway.

Previous reports have found that the first step of T_reg development is linked to the TCR signal strength, implicating transcription factors of the Nr4a family⁴⁹. Concomitantly to TCR signaling, CD28 costimulation promotes T_reg generation by inducing Foxp3 expression⁸. We observed that Cd28^−/− mice have decreased numbers of both CD25⁺ T_regP and Foxp3^lo T_regP, resulting in a drastic reduction of mature T_reg. Our data show that the stimulation of TCR and CD28 also upregulates the expression of LTα₁β₂ in CD25⁺ T_regP and Foxp3^lo T_regP. In addition to TCR/CD28 signals, γ-chain cytokines are important regulators of T_reg development since Il2^−/− and Il15^−/− mice exhibit a severe reduction of T_regP and mature T_reg in their thymus^10,50. Interestingly, we found that IL-2 controls LTα₁β₂ expression specifically in mature T_reg. Accordingly, LTα₁β₂ and CD25 levels were tightly correlated and IL-2 in vitro stimulation led to a dose-dependent LTα₁β₂ upregulation in mature T_reg. A recent report described that IL-2 also increases LTα₁β₂ expression in peripheral T_reg⁵¹, suggesting a conserved regulatory mechanism between the thymus and the periphery.

Although Foxp3^lo T_regP leads to the generation of CD25⁺Foxp3⁺ T_reg upon intrathymic transfer^14,15, the underlying mechanisms that control this newly identified developmental pathway remain unknown. The generation of Foxp3^eGFPxLta^−/− mice revealed that Lta expression limits T_reg development from Foxp3^lo T_regP. To our knowledge, LTα is the first identified molecule that fine-tunes the Foxp3^lo T_regP developmental pathway. Although CD25⁺ T_regP express LTα₁β₂ to some extent, these cells remain poorly affected compared to Foxp3^lo T_regP in Foxp3^eGFPxLta^−/− mice. We show that the increased cellularity of Foxp3^lo T_regP and mature T_reg in Foxp3^eGFPxLta^−/− mice was likely a direct consequence of enhanced proliferation and survival. Moreover, given that LTβR signaling was found to regulate the expression of chemokines involved in thymocyte trafficking¹⁷ and that the LTα₁β₂- LTβR axis has been implicated in T_reg trafficking in a model of pancreatic islet allograft^18,52, future investigations are expected to clarify whether Lta-deficiency could affect the intrathymic migration of T_reg subsets. Multiparametric flow cytometry analyses based on maturational markers revealed an unexpected heterogeneity in both T_regP and mature T_reg. Our results highlight that CD25⁺ T_regP is a highly heterogenous population, with cells at distinct maturational stages. Compared to CD25⁺ T_regP, Foxp3^lo T_regP contain more mature cells at M1 and M2 stages, similarly to CD25⁺Foxp3⁺ mature T_reg. Our transcriptomic and flow cytometry analyses highlight that CD25⁺ T_regP and Foxp3^lo T_regP subsets contain not only precursors but also CD62L⁺ mature cells, which is a reliable marker for cells about to leave the thymus⁵³. Therefore, cells commonly called T_regP likely contain cells with the ability to emigrate, although this requires further investigations. The CD69⁺CD62L⁻ Foxp3^lo T_regP population, displaying a mature phenotype characterized by high levels of H2K^b and Qa2, increased in Foxp3^eGFPxLta^−/− mice, suggesting that this putative transitional stage towards mature T_reg is regulated by Lta. Altogether, our data identified LTα as a negative regulator of T_reg maturational development.

A recent study performed scRNA-seq on thymic T_reg using pre-sorted subpopulations from TCliβxTCRα^+/−xFoxp3^GFP mice and characterized CD25⁺T_regP, Foxp3^loT_regP and mature de novo T_reg⁵⁴. In contrast, our cell-sorting strategy based on the entire T_reg lineage allowed us to also capture CD122⁺ pre-T_regP and transitional T_reg states from Foxp3^eGFP and Foxp3^eGFPxLta^−/− mice. Moreover, we characterized the metabolic profile in our scRNA-seq experiment during T_reg development, which remained unknown to date. Interestingly, whereas CD25⁺ T_regP rely on both glucose and OXPHOS mitochondrial dependencies, our transcriptomic and functional metabolic analyses revealed that like mature T_reg, Foxp3^lo T_regP are highly dependent on OXPHOS activity and have a lower dependency on glucose metabolism. These findings are consistent with the fact that CD25⁺ T_regP and Foxp3^lo T_regP have a distinct transcriptomic profile^15,54. In Foxp3^eGFPxLta^−/− mice, the glucose dependency substantially increased in both T_regP and mature T_reg, indicating that Lta regulates their metabolic profile. This higher glucose dependency may be linked to the enhanced survival and/or proliferation of T_regP and mature T_reg observed in Foxp3^eGFPxLta^−/− mice. Further investigations will elucidate whether the distinct metabolic profiles of CD25⁺ T_regP and Foxp3^lo T_regP depend on their responsiveness to γ-chain cytokines and influence their epigenetic landscape. Understanding the intricate interplay between the two T_regP and their respective stromal niches, which could modulate their metabolic profiles and influence their differentiation, represents an exciting area for future research.

T_reg conversion assays using co-cultures of CD69⁺ semi-mature OTII CD4⁺ SP thymocytes with mTEC that express LTβR revealed that the LTα₁β₂-LTβR axis acts as a negative regulator of CD25⁺Foxp3⁺ mature T_reg conversion. We found an enhanced generation of Foxp3^lo T_regP from CD69⁺ CD4⁺ SP thymocytes in the absence of Lta in CD4⁺ SP thymocytes or when LTα₁β₂-LTβR interactions were blocked with the soluble LTβR-Fc receptor. In line with these observations, this resulted in an increased conversion into CD25⁺Foxp3⁺ mature T_reg. Furthermore, T_reg conversion assays using co-cultures with T_regP showed that the LTα₁β₂-LTβR axis acts as an inhibitory feedback by which Foxp3^lo T_regP can limit their own differentiation into CD25⁺Foxp3⁺ mature T_reg. In agreement with a previous study¹⁵, we found that Foxp3^lo T_regP express higher levels of IL-4Rα than CD25⁺ T_regP both by scRNA-seq and flow cytometry. Moreover, Foxp3^lo T_regP have a higher ability to respond to IL-4 stimulation than CD25⁺ T_regP. Remarkably, Il4 expression was higher in Lta^−/− mice, which correlated with a greater frequency of Eomes⁺ memory-like CD8⁺ SP thymocytes and phospho-Stat6⁺ cells. In vivo IL-4 neutralization resulted in a substantial reduction of Foxp3^lo T_regP and CD25⁺Foxp3⁺ mature T_reg whereas CD25⁺ T_regP were over-represented, indicating that IL-4 favors the Foxp3^lo T_regP pathway at the expense of the CD25⁺ T_regP developmental pathway. Future studies are expected to define the respective contribution of T_regP and mature T_reg in LTα₁β₂-controlled IL-4 production by mTEC.

ScRNA-seq, qPCR and IL-4^eGFP reporter mice revealed that mTEC are a valuable source of IL-4 in addition to NKT2. Strikingly, mTEC produced a larger amount of IL-4 upon antigen-specific co-cultures with CD69⁺CD4⁺ SP thymocytes purified from OTIIxLta^−/− mice compared to OTII mice, indicating that Lta expression negatively regulates Il4 expression in mTEC. This higher production correlated with a greater conversion into mature T_reg, which was abrogated by a neutralizing IL-4 antibody. The classical NF-κB member p65 has the ability to bind to the P4/PRE-I and P1 sites of the Il4 promoter and upregulate its transcription^41,55. Whereas the non-classical NF-κB pathway is preferentially activated by lymphotoxin signaling at steady state^27,56, our data indicate a preferential usage of the classical NF-κB pathway characterized by an upregulation of p65 when lymphotoxin signaling is disrupted. Since IL-4 is regulated by the classical NF-κB member p65, which was overexpressed in Lta^−/− mice, this cytokine upregulation in Lta^−/− mice is likely induced by the classical NF-κB pathway. Interestingly, previous studies have shown that IL-4 acts as a proliferation and survival factor of peripheral T_reg^57,58,59. Thus, IL-4 may not only induce mature T_reg conversion but also promote their survival and proliferation. In line with this hypothesis, we found reduced expression of the pro-apoptotic protein BIM in Foxp3^lo T_regP and mature T_reg of Foxp3^eGFPxLta^−/− mice. Whereas IL-2 is crucial for the CD25⁺ T_regP developmental pathway^10,50, our findings show that IL-4 preferentially regulates the Foxp3^lo T_regP developmental pathway, indicating that these two cytokines play a non-overlapping role in T_reg development. IL-15, also implicated in T_reg development, is expressed by mTEC^10,12,14. Therefore, by their ability to produce both IL-15 and IL-4, mTEC likely constitute a privilege niche that controls T_reg development.

In summary, this study shows that instructive signals of the “two step model” induce the gradual upregulation of the membrane-bound LTα₁β₂ during T_reg development. Importantly, we identify that LTα₁β₂-LTβR interactions act as an inhibitory checkpoint that limits the Foxp3^lo T_regP developmental pathway by negatively regulating IL-4 production in mTEC (Supplementary Fig. 13). Interestingly, human thymi, routinely removed during pediatric cardiac surgery, constitute a novel source of stable T_reg for cell therapy that can be purified in large quantities^60,61. Given that CD25⁺ and Foxp3^lo T_reg P-derived T_reg likely exhibit a differential ability to protect against distinct autoimmune disorders^15,62, targeting LTα₁β₂-LTβR interactions may therefore be used to foster the Foxp3^lo T_regP pathway in order to treat specific pathologies. Our findings are, thus, expected to pave the way for therapeutic strategies tailored to treat distinct autoimmune and inflammatory disorders.

Methods

Mice

All mice - CD45.1, CD45.2 and CD45.1/2 WT (Charles River), CD45.2 Lta^−/−⁶³, Rag2^−/−⁶⁴, Cd28^−/−⁶⁵, OTII⁶⁶, Rip-mOVA⁶⁷, CD45.1 Foxp3^eGFP⁶⁸, Rag2^GFP19, Foxp3^Thy1.120, Il15^KO⁶⁹ and Il2^KO (B6.129P2-Il2tm1Hor/J, JAX laboratories) - were on a pure C57BL/6 background. The IL-4^eGFP reporter (4get) mice⁷⁰ were on a C57BL/6 or BALB/c background. CD45.2 Lta^−/− mice were backcrossed with CD45.1 Foxp3^eGFP mice (CD45.1 Foxp3^eGFPxLta^−/− mice). OTII mice were backcrossed on a Rag2^−/− background (OTIIxRag2^−/− mice) and OTIIxRag2^−/− on a Lta^−/− background (OTIIxRag2^−/−xLta^−/− mice). Rip-mOVA mice were backcrossed with OTIIxRag2^−/− mice (Rip-mOVAxOTIIxRag2^−/−). Rip-mOVAxOTIIxRag2^−/− were backcrossed with Foxp3^eGFP mice (Rip-mOVAxOTIIxFoxp3^eGFP). Foxp3^Thy1.1 mice were backcrossed with Rag2^GFP mice (Rag2^GFPxFoxp3^Thy1.1 mice) and Rag2^GFPxFoxp3^Thy1.1 mice on a Il15^KO or Il2^KO background (Il15^KOxRag2^GFPxFoxp3^Thy1.1 or Il2^KOxRag2^GFPxFoxp3^Thy1.1 mice). All mice were maintained under specific pathogen free conditions at the CIML (Marseille, France). Mice were housed under a standard 12 h/12 h light-dark cycle, with a controlled ambient temperature of 22 °C ± 2 °C, and relative humidity between 45% and 65%. Standard food and water were given ad libitum. For all experiments, male and female mice were sex- and age-matched. Unless specified, adult mice were used between 6 and 8 weeks of age. Chimeras were generated between 6 and 12 weeks of age. All experiments were done in accordance with National and European laws for laboratory animal welfare (EEC Council Directive 2010/63/UE) and approved by the Marseille Ethical Committee for Animal Experimentation no.14.

Bone marrow chimeras

Bone marrow chimeras were generated by co-injecting intravenously (i.v.) 2 × 10⁶ BM cells from CD45.2 WT mice with 2 × 10⁶ BM cells from either CD45.1 Foxp3^eGFP or Foxp3^eGFPxLta^−/− mice (ratio 1:1) into lethally irradiated CD45.1/2 WT recipients (two doses of 500 rads, 12 h apart, X-ray using a RS-2000 Irradiator, Rad Source Technologies). Similarly, 5 × 10⁶ BM cells from CD45.1 Foxp3^eGFP mice were injected into lethally irradiated CD45.2 WT or Lta^−/− recipients. Mice were analyzed 6 weeks post-reconstitution.

Thymic cell isolation

For T_reg isolation, thymi were mechanically dissociated on a 70 µm cell strainer. CD4⁺ SP thymocytes were pre-enriched by depleting CD8⁺ and CD11c⁺ cells using biotinylated anti-CD8α (clone 53.6.7; Biolegend) and anti-CD11c (clone N418; Biolegend) antibodies with anti-biotin microbeads and ran on an AutoMACS Pro separator using the Deplete program (Miltenyi Biotec). For TEC isolation, thymi were digested 10 min at 37 °C in HBSS medium containing 50 µg/ml of Liberase^TM (Roche) and 100 µg/ml DNase I (Roche) and were then subjected to vigorous pipetting. This step was repeated until complete tissue digestion. Cells were filtered through a 70 μm strainer to remove clumps. CD45⁺ hematopoietic cells were depleted using biotinylated anti-CD45 antibodies (clone 30-F11; Biolegend) with anti-biotin microbeads, and run on an AutoMACS Pro separator using the DepleteS program (Miltenyi Biotec). CD69⁺CD4⁺ SP thymocytes, T_regP, mature T_reg, NKT, and mTEC subsets were cell-sorted using a FACSAria III cell sorter (BD Biosciences).

In vivo IL-4 neutralization experiments

Lta^−/− neonates received an i.p. injection of either isotype control (clone RTK2071, Biolegend) or neutralizing anti-mouse IL-4 antibody (clone 11B11, BioXcell) on the day they were born (100 µg) and 3 days later (100 µg). Mice were analyzed 5 days post-birth.

In vitro antigen-specific T_reg conversion assays

Co-culture assays between mTEC and CD69⁺CD4⁺ SP thymocytes to assess T_reg conversion were performed⁷¹. Briefly, 1 × 10³ cell-sorted total mTEC (CD45⁻EpCAM⁺Ly51^loUEA-1^hi), mTEC^hi (CD45⁻EpCAM⁺Ly51^loUEA-1^hiCD80⁺) and mTEC^lo (CD45^-EpCAM⁺Ly51^loUEA-1^hiCD80⁻) from WT thymi were loaded or not for 90 min with OVA_323-339 peptide (5 µg/ml, PolyPeptide) at 37 °C. Cells were then co-cultured with 5 × 10³ purified CCR6⁻CD25⁻CD69⁺CD4⁺CD8⁻ SP thymocytes from OTIIxRag2^−/− mice pre-incubated or not during 1 h at 37 °C with recombinant LTβR-Fc (2 μg/ml, RnD systems) or from OTIIxRag2^−/−xLta^−/− mice in a complete medium containing D-MEM (ThermoFisher Scientific), 10 % FBS (Sigma Aldrich), 2 mM L-glutamine, 1 mM sodium pyruvate, 2.10⁻⁵ M 2-mercaptoethanol, 100 IU/ml penicillin, and 100 μg/ml streptomycin (all from ThermoFisher Scientific). In addition, total mTEC were co-cultured with 5 × 10³ purified CD25⁺ T_regP or Foxp3^lo T_regP from Rip-mOVAxOTIIxFoxp3^eGFP mice pre-incubated or not during 1 h at 37 °C with recombinant LTβR-Fc (2 μg/ml, RnD systems). Alternatively, 10 µg/mL of neutralizing anti-mouse IL-4 antibody (clone 11B11, BioXcell) was added in the co-cultures of OVA_323–339-loaded mTEC with CCR6⁻CD25⁻CD69⁺CD4⁺CD8⁻ SP from OTIIxRag2^−/− or OTIIxRag2^−/−xLta^−/− mice. The conversion into CD25⁺ T_regP, Foxp3^lo T_regP, or CD25⁺Foxp3⁺ mature T_reg was analyzed by flow cytometry 18 h, 24 h, 72 h, or 5 days later, as indicated in the legends.

In vitro T_regP conversion assays with IL-4 stimulation

1 × 10⁴ purified CD25⁺ T_regP or Foxp3^lo T_regP from Foxp3^eGFP mice were incubated in complete RPMI medium supplemented with different doses of mouse IL-4 (0.5, 1 or 100 ng/mL; Peprotech). After 72 h, cells were harvested, stained with anti-CD4 and anti-CD25 antibodies, and analyzed by flow cytometry for the conversion into CD25⁺Foxp3^eGFP+ mature T_reg.

In vitro activation of T_reg cell subsets

For CD3ε stimulation, 2 × 10⁴ purified CD25⁺ T_regP, Foxp3^lo T_regP or CD25⁺Foxp3⁺ T_reg from Foxp3^eGFP mice were cultured with different doses (0.1, 1, 2.5 or 5 µg/mL) of coated anti-CD3ε antibody (clone 2C11, BD Biosciences) in a complete RPMI medium containing soluble anti-CD28 antibody (0.5 µg/mL, clone 37.51, Biolegend) in the presence of a low dose of IL-2 (5 ng/mL, Immunotools). For IL-2 stimulation, 2 × 10⁴ purified CD25⁺ T_regP, Foxp3^lo T_regP, or mature T_reg from Foxp3^eGFP mice were cultured with a low dose of coated anti-CD3ε antibody (2 µg/mL, clone 2C11, BD Biosciences) in a complete RPMI medium containing soluble anti-CD28 antibody (0.5 µg/mL, clone 37.51, Biolegend) in the presence or not of different doses (5, 20 or 40 ng/mL) of IL-2 (Immunotools). After 16 h, cells were harvested and analyzed for LTα₁β₂ expression by flow cytometry.

Flow cytometry

Cells were stained with standard procedures using antibodies listed in Supplementary Data 1. For intracellular staining, cells were fixed, permeabilized, and stained with the Foxp3 staining kit according to the manufacturer’s instructions (ThermoFisher Scientific). Fluorescence minus one (FMO) and secondary antibody controls were used to set positive staining gates. NKT cells were detected using CD1d tetramers loaded with PBS-57 or unloaded (obtained from the National Institutes of Health Tetramer Core Facility). For LTα₁β₂ staining, cells were incubated with LTβR-Fc (1 μg/10⁶ cells, R&D systems) for 45 min on ice, and staining was visualized using an APC-conjugated donkey anti-human IgG F(ab’)₂ antibody (Jackson ImmunoResearch). For all panels, we inferred the background levels observed in the respective cell subsets from Lta^−/− mice. To evaluate mitochondrial activity, thymocytes were stained with MitoTracker Deep Red (100 nM) (Thermo Fisher) for 30 min at 37 °C in complete RPMI medium. For unbiased clustering of the cells from the T_reg lineage, data from all samples were concatenated and analyzed using the t-SNE plugin of the FlowJo software (BD Life Sciences, version 10.8.1). The PhenoGraph plugin was used for the unsupervised discrimination of T_reg subsets (with a k-nearest neighbor value of 130 for Fig. 3i and 120 for Fig. 4f), which were then visualized on the t-SNE dimensional reduction and quantified. The color code representing the expression level of each marker in heatmaps was normalized on the median intensity value for a given marker. Levels of LTα₁β₂ expression were measured in thymic T_reg subsets from Rag2^GFPxFoxp3^Thy1.1 (Fig. 1b–f), Rag2^GFPxFoxp3^Thy1.1xIl15^−/− and Rag2^GFPxFoxp3^Thy1.1xIl2^−/− (Fig. 2d, e) mice on a BD LSR Fortessa at Infinity (Toulouse, France). The other flow cytometry experiments were conducted using FACSCanto II, BD LSR II, and Symphony cytometers (BD Biosciences) at CIML (Marseille, France). Data were analyzed using the FlowJo software (BD Life Sciences).

SCENITH™ metabolic profiling

SCENITH kit containing all reagents (including anti-puromycin, clone R4743L-E8) and protocols were obtained (www.scenith.com) and the T_reg metabolic profile was analyzed³⁵. Briefly, 4 × 10⁶ thymocytes were plated at 40 × 10⁶ cells/ml in complete medium in a 96-well plate and treated with Puromycin (10 µg/mL) and DMSO (Co), 2-deoxy-glucose (2DG; 100 mM), oligomycin (O; 1 µM), a combination of 2DG and oligomycin (DGO) or harringtonine (H; 2 µg/mL) for 40 min at 37 °C, 5 % CO₂. Cells were washed in cold FACS buffer and incubated with a BD Horizon^TM Fixable Viability Stain 700 (BD Biosciences). Then, cells were stained with primary-conjugated antibodies for 25 min at 4 °C in FACS Buffer and were fixed and permeabilized with the Foxp3 staining kit according to the manufacturer’s instructions (ThermoFisher Scientific). Intracellular staining with anti-puromycin and anti-Foxp3 antibodies was performed by incubating cells during 1 h at 4 °C in Permeabilization Buffer (ThermoFisher Scientific).

Immunofluorescence staining

20-µm-thick-frozen thymic sections embedded in O.C.T (Sakura Finetek) were stained⁷². Briefly, sections were fixed with 2% paraformaldehyde (Merck), saturated with 3% BSA, and 0.01 % Triton X100 in 0.1 M Tris HCl buffer. Sections were then stained for 45 min with anti-keratin 14 and anti-NK1.1 Alexa Fluor 647 (clone PK136, Biolegend) in hybridization buffer (1% BSA and 0.02% Triton X100 in 0.1 M Tris-HCl, pH 7.4). Keratin 14 staining was revealed with a Cy3-conjugated anti-rabbit antibody (ThermoFisher Scientific). Sections were counterstained with DAPI (1 µg/mL; Biolegend) and mounted with Mowiol (Calbiochem). Confocal microscopy was performed with an LSM 880 confocal microscope (Carl Zeiss Microscopy) and analysed using Zen Blue (3.5.093).

RT-qPCR

Total RNA was isolated with TRIzol (ThermoFisher Scientific) and cDNA was synthesized with random oligo-dT primers and Superscript II reverse transcriptase (ThermoFisher Scientific). qPCR was performed with SYBR Premix Ex Taq master mix (Takara) on an ABI 7500 fast real-time PCR system (Applied Biosystem). Reactions were conducted in duplicates. Results were normalized to Actin mRNA. Primer sequences are provided in Supplementary Data 2.

scRNA-seq library preparation

CD8- and CD11c-depleted thymocytes from Foxp3^eGFP and Foxp3^eGFPxLta^−/− mice were stained with fluorescently labeled antibodies (Fig. 5a) and barcoded using TotalSeq-A antibodies (BioLegend). FACS sorted Live Dead⁻CD4⁺CD8⁻CCR6⁻CD3ε⁺ expressing either CD25, Foxp3^eGFP, or both were pooled and subsequently prepared for single-cell sequencing using the Chromium Single Cell 3′ kit v3 (10X Genomics) with a targeted number of 10,000 total cells. Libraries were prepared according to the manufacturer’s instructions. The concentration and quality of cDNA libraries were respectively determined using the Qubit dsDNA HS Assay Kit (ThermoFisher Scientific) for the Qbit fluorometer and the High Sensitivity DNA Analysis Kit (Agilent) for the bioanalyzer. Libraries were sequenced on the Illumina NextSeq2000 sequencer.

Data analysis for scRNA-seq analysis

mRNA and Cell Hashing FASTQ raw files were processed using Cell Ranger v6.0.1 (10X genomics Inc.) software with default parameters to perform alignment, filtering, barcode counting, and unique molecular identifier (UMI) counting. Reads were aligned to the mm10 genome. A total number of 4,798 cells were identified with a mean of 44,015 reads per cell and a median of 1,992 genes per cell. Data sets were analysed using the R package Seurat 4.0.5⁷³. First, cells were demultiplexed (with the Cell Hashing tags) to their original sample groups using the HTODemux function from Seurat R package with automated threshold finding to define the best quantile. 302 doublet cells and 4 negative cells were identified and removed. Quality control (QC) was performed to remove poor quality cells. We observed co-variables using the ShIVA interface and set the following thresholds⁷⁴: cells with less than 603 or more than 4128 detected genes, cells with less than 851 or more than 21,686 detected UMI, cells with more than 9.94% mitochondrial gene expression, and cells with less than 11.5% ribosomal gene expression were removed. Expression data was normalized using the NormalizeData function of the Seurat R package (logNormalize method and scale factor of 10,000). We centered the expression data from these factors using the Seurat R package ScaleData function (centering true and scaling false).

Feature selection was performed using the Seurat R package FindVariableGene function; the variance-stabilizing transform (vst) algorithm and 2,000 genes were selected. Principal component analysis (PCA) was ran using Seurat RunPCA function on those 2,000 most variable genes. UMAP dimensionality reduction was ran using Seurat RunUMAP function on the first 50 principal components from PCA. Clustering was done with Seurat FindClusters based on the first 30 principal components from PCA dimensionality reduction and with an initial resolution parameter set to 1.1. Markers for each cluster were identified using Wilcoxon Rank Sum test from Seurat FindAllMarkers with the log fold change threshold of 0.1 and an adjusted FDR threshold of 0.001. Cell-cycle scores and cell phases were calculated using CellCycleScoring with s.genes and g2m.genes from the Seurat package. After cell type analysis using marker genes, we identified several clusters to be contaminants, in proliferation or recirculating cells and excluded them. PCA, UMAP, and clustering (with resolution 0.8) were performed again on the filtered dataset.

Heatmaps of gene expression data per cluster were created using the Heatmap function of the ComplexHeatmap (v2.10.0)⁷⁵ R package. Expression was presented as the mean of the expression of cells inside each cluster. Values were scaled to make them more comparable. Module scores of gene set list from the KEGG database were calculated using the AddModuleScore function from the Seurat R package. The Nebulosa package was used to visualize features on UMAP embeddings using kernel density estimation⁷⁶.

Trajectory analysis were performed using Monocle 3³². Cells were clustered using the Monocle 3 cluster_cells function with a resolution set to 0.003. Trajectories were computed using the Monocle 3 learn_graph and order_cells function. Gene set scores were computed as previously described and plotted along the pseudotime trajectories. All regression lines were computed using the ggplot2 geom_smooth function, setting the method to “loess”.

Statistical analysis

All statistical analyses were performed in GraphPad Prism 9.5.1 and were considered statistically different at a p-value < 0.05. Unpaired Student’s t test, Mann–Whitney U test, one-way ANOVA, or Kruskal–Wallis test were calculated. Normal distribution of the data was assessed using d’Agostino-Pearson omnibus normality test. Correlation was calculated using the parametric two-tailed Pearson correlation test or the non-parametric two-tailed Spearman correlation test. Statistical comparison of module scores between conditions for each cluster was performed using the Wilcoxon test with Benjamini-Hochberg multitest correction. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001. Error bars represent mean ± SEM.

Reporting summary

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

Data availability

The scRNA-seq data generated in this study have been deposited in the GEO database under accession number GSE237589. All data supporting the findings of this study are found within the paper and its Supplementary Information. Source data are provided with this paper.

Code availability

Scripts used to analyze sequencing data have been deposited on GitHub (https://github.com/CIML-bioinformatic/MIlab_LTaTreg-Thymus) and Zenodo (https://doi.org/10.5281/zenodo.12705747)⁷⁷.

References

Fontenot, J. D., Dooley, J. L., Farr, A. G. & Rudensky, A. Y. Developmental regulation of Foxp3 expression during ontogeny. J. Exp. Med. 202, 901–906 (2005).
Article CAS PubMed PubMed Central Google Scholar
Aschenbrenner, K. et al. Selection of Foxp3⁺ regulatory T cells specific for self antigen expressed and presented by Aire+ medullary thymic epithelial cells. Nat. Immunol. 8, 351–358 (2007).
Article CAS PubMed Google Scholar
Perry, J. S. A. et al. Distinct contributions of Aire and antigen-presenting-cell subsets to the generation of self-tolerance in the thymus. Immunity 41, 414–426 (2014).
Article CAS PubMed PubMed Central Google Scholar
Charaix, J. et al. Recirculating Foxp3+ regulatory T cells are restimulated in the thymus under Aire control. Cell. Mol. Life Sci. 79, 355 (2022).
Article CAS PubMed PubMed Central Google Scholar
Lio, C. W. J. & Hsieh, C. S. A two-step process for thymic regulatory T cell development. Immunity 28, 100–111 (2008).
Article CAS PubMed PubMed Central Google Scholar
Klein, L., Robey, E. A. & Hsieh, C. S. Central CD4⁺ T cell tolerance: deletion versus regulatory T cell differentiation. Nat. Rev. Immunol. 19, 7–18 (2019).
Article CAS PubMed Google Scholar
Santamaria, J. C., Borelli, A. & Irla, M. Regulatory T cell heterogeneity in the thymus: impact on their functional activities. Front Immunol. 12, 1–8 (2021).
Article Google Scholar
Tai, X., Cowan, M., Feigenbaum, L. & Singer, A. CD28 costimulation of developing thymocytes induces Foxp3 expression and regulatory T cell differentiation independently of interleukin 2. Nat. Immunol. 6, 152–162 (2005).
Article CAS PubMed Google Scholar
Mahmud, S. A. et al. Costimulation via the tumor-necrosis factor receptor superfamily couples TCR signal strength to the thymic differentiation of regulatory T cells. Nat. Immunol. 15, 473–481 (2014).
Article CAS PubMed PubMed Central Google Scholar
Apert, C. et al. IL-2 and IL-15 drive intrathymic development of distinct periphery-seeding CD4+Foxp3+ regulatory T lymphocytes. Front Immunol. 13, 965303 (2022).
Article CAS PubMed PubMed Central Google Scholar
Hemmers, S. et al. IL-2 production by self-reactive CD4 thymocytes scales regulatory T cell generation in the thymus. J. Exp. Med. 216, 2466–2478 (2019).
Article CAS PubMed PubMed Central Google Scholar
Cui, G. et al. Characterization of the IL-15 niche in primary and secondary lymphoid organs in vivo. Proc. Natl Acad. Sci. USA 111, 1915–1920 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Tai, X. et al. Foxp3 transcription factor is proapoptotic and lethal to developing regulatory T cells unless counterbalanced by cytokine survival signals. Immunity 38, 1116–1128 (2013).
Article CAS PubMed PubMed Central Google Scholar
Marshall, D., Sinclair, C., Tung, S. & Seddon, B. Differential requirement for IL-2 and IL-15 during bifurcated development of thymic regulatory T cells. J. Immunol. 193, 5525–5533 (2014).
Article CAS PubMed Google Scholar
Owen, D. L. et al. Thymic regulatory T cells arise via two distinct developmental programs. Nat. Immunol. 20, 195–205 (2019).
Article CAS PubMed PubMed Central Google Scholar
Gommerman, J. L. & Browning, J. L. Lymphotoxin/LIGHT, lymphoid microenvironments and autoimmune disease. Nat. Rev. Immunol. 3, 642–655 (2003).
Article CAS PubMed Google Scholar
Borelli, A. & Irla, M. Lymphotoxin: from the physiology to the regeneration of the thymic function. Cell Death Differ. 28, 2305–2314 (2021).
Article CAS PubMed PubMed Central Google Scholar
Brinkman, C. C. et al. Treg engage lymphotoxin beta receptor for afferent lymphatic transendothelial migration. Nat. Commun. 7, 12021 (2016).
Article ADS PubMed PubMed Central Google Scholar
Yu, W. et al. Continued RAG expression in late stages of B cell development and no apparent re-induction after immunization. Nature 400, 682–687 (1999).
Article ADS CAS PubMed Google Scholar
Liston, A. et al. Differentiation of regulatory Foxp3 T cells in the thymic cortex. Proc. Natl Acad. Sci. USA 105, 11903–11908 (2008).
Article ADS CAS PubMed PubMed Central Google Scholar
McCaughtry, T. M., Wilken, M. S. & Hogquist, K. A. Thymic emigration revisited. J. Exp. Med. 204, 2513–2520 (2007).
Article CAS PubMed PubMed Central Google Scholar
Thiault, N. et al. Peripheral regulatory T lymphocytes recirculating to the thymus suppress the development of their precursors. Nat. Immunol. 16, 628–634 (2015).
Article CAS PubMed Google Scholar
Yang, E. J., Zou, T., Leichner, T. M., Zhang, S. L. & Kambayashi, T. Both retention and recirculation contribute to long-lived regulatory T-cell accumulation in the thymus. Eur. J. Immunol. 44, 2712–2720 (2014).
Article CAS PubMed PubMed Central Google Scholar
Moran, A. E. et al. T cell receptor signal strength in Treg and iNKT cell development demonstrated by a novel fluorescent reporter mouse. J. Exp. Med. 208, 1279–1289 (2011).
Article CAS PubMed PubMed Central Google Scholar
Apert, C., Romagnoli, P. & van Meerwijk, J. P. M. IL-2 and IL-15 dependent thymic development of Foxp3-expressing regulatory T lymphocytes. Protein Cell 9, 322–332 (2018).
CAS PubMed Google Scholar
Bayer, A. L., Lee, J. Y., de la Barrera, A., Surh, C. D. & Malek, T. R. A function for IL-7R for CD4⁺CD25⁺Foxp3⁺ T regulatory cells. J. Immunol. 181, 225–234 (2008).
Article CAS PubMed Google Scholar
Lopes, N., Charaix, J., Cédile, O., Sergé, A. & Irla, M. Lymphotoxin α fine-tunes T cell clonal deletion by regulating thymic entry of antigen-presenting cells. Nat. Commun. 9, 1–16 (2018).
Article CAS Google Scholar
Bouillet, P. et al. BH3-only Bcl-2 family member Bim is required for apoptosis of autoreactive thymocytes. Nature 415, 922–926 (2002).
Article ADS CAS PubMed Google Scholar
Gray, D. H. D. et al. The BH3-Only proteins bim and puma cooperate to impose deletional tolerance of organ-specific antigens. Immunity 37, 451–462 (2012).
Article CAS PubMed PubMed Central Google Scholar
Irla, M. et al. Three-dimensional visualization of the mouse thymus organization in health and immunodeficiency. J. Immunol. 190, 586–596 (2013).
Article CAS PubMed Google Scholar
Schuster, M., Plaza-Sirvent, C., Visekruna, A., Huehn, J. & Schmitz, I. Generation of Foxp3⁺CD25⁻ regulatory T-cell precursors requires c-rel and IκBNS. Front Immunol. 10, 1583 (2019).
Article CAS PubMed PubMed Central Google Scholar
Cao, J. et al. The single-cell transcriptional landscape of mammalian organogenesis. Nature 566, 496–502 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Shi, H. & Chi, H. Metabolic control of treg cell stability, plasticity, and tissue-specific heterogeneity. Front Immunol. 10, 2716 (2019).
Article CAS PubMed PubMed Central Google Scholar
Newton, R., Priyadharshini, B. & Turka, L. A. Immunometabolism of regulatory T cells. Nat. Immunol. 17, 618–625 (2016).
Article CAS PubMed PubMed Central Google Scholar
Argüello, R. J. et al. SCENITH: a flow cytometry-based method to functionally profile energy metabolism with single-cell resolution. Cell Metab. 32, 1063–1075.e7 (2020).
Article PubMed PubMed Central Google Scholar
Xing, Y., Wang, X., Jameson, S. C. & Hogquist, K. A. Late stages of T cell maturation in the thymus involve NF-κB and tonic type i interferon signaling. Nat. Immunol. 17, 565–573 (2016).
Article CAS PubMed PubMed Central Google Scholar
Lee, Y. J., Holzapfel, K. L., Zhu, J., Jameson, S. C. & Hogquist, K. A. Steady-state production of IL-4 modulates immunity in mouse strains and is determined by lineage diversity of iNKT cells. Nat. Immunol. 14, 1146–1154 (2013).
Article CAS PubMed Google Scholar
Lai, D. et al. KLF13 sustains thymic memory-like CD8+ T cells in BALB/c mice by regulating IL-4-generating invariant natural killer T cells. J. Exp. Med. 208, 1093–1103 (2011).
Article CAS PubMed PubMed Central Google Scholar
Wang, H. et al. Myeloid cells activate iNKT cells to produce IL-4 in the thymic medulla. Proc. Natl Acad. Sci. USA 116, 22262–22268 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Zhou, T. A. et al. Thymic macrophages consist of two populations with distinct localization and origin. Elife 11, e75148 (2022).
Article CAS PubMed PubMed Central Google Scholar
Li-Weber, M., Giasi, M. & Krammer, P. H. Involvement of Jun and Rel proteins in up-regulation of interleukin-4 gene activity by the T cell accessory molecule CD28. J. Biol. Chem. 273, 32460–32466 (1998).
Article CAS PubMed Google Scholar
Wirnsberger, G., Mair, F. & Klein, L. Regulatory T cell differentiation of thymocytes does not require a dedicated antigen-presenting cell but is under T cell-intrinsic developmental control. Proc. Natl Acad. Sci. USA 106, 10278–10283 (2009).
Article ADS CAS PubMed PubMed Central Google Scholar
Malchow, S. et al. Aire enforces immune tolerance by directing autoreactive T cells into the regulatory T cell lineage. Immunity 44, 1102–1113 (2016).
Article CAS PubMed PubMed Central Google Scholar
Malchow, S. et al. Aire-dependent thymic development of tumor-associated regulatory T cells. Science 339, 1219–1224 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Michelson, D. A., Hase, K., Kaisho, T., Benoist, C. & Mathis, D. Thymic epithelial cells co-opt lineage-defining transcription factors to eliminate autoreactive T cells. Cell 185, 2542–2558.e18 (2022).
Article CAS PubMed PubMed Central Google Scholar
Wells, K. L. et al. Combined transient ablation and single cell RNA sequencing reveals the development of medullary thymic epithelial cells. Elife 9, 1–80 (2020).
Article Google Scholar
Jin, S. et al. Inference and analysis of cell-cell communication using CellChat. Nat. Commun. 12, 1088 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Irla, M. Instructive Cues of Thymic T Cell Selection. Annu. Rev. Immunol. 26, 95–119 (2022).
Sekiya, T. et al. Nr4a receptors are essential for thymic regulatory T cell development and immune homeostasis. Nat. Immunol. 14, 230–237 (2013).
Article CAS PubMed Google Scholar
Fontenot, J. D., Rasmussen, J. P., Gavin, M. A. & Rudensky, A. Y. A function for interleukin 2 in Foxp3-expressing regulatory T cells. Nat. Immunol. 6, 1142–1151 (2005).
Article CAS PubMed Google Scholar
Piao, W. et al. Regulatory T cells condition lymphatic endothelia for enhanced transendothelial migration. Cell Rep. 30, 1052–1062.e5 (2020).
Article CAS PubMed PubMed Central Google Scholar
Saxena, V. et al. Treg tissue stability depends on lymphotoxin beta-receptor- and adenosine-receptor-driven lymphatic endothelial cell responses. Cell Rep. 39, 110727 (2022).
Article CAS PubMed PubMed Central Google Scholar
James, K. D. et al. Endothelial cells act as gatekeepers for LTβRdependent thymocyte emigration. J. Exp. Med. 215, 2984–2993 (2018).
Article CAS PubMed PubMed Central Google Scholar
Owen, D. L., Rue, R. S., La, Munro, S. A. & Farrar, M. A. Tracking regulatory T cell development in the thymus using single-cell RNA sequencing/TCR sequencing. J. Immunol. 209, 1300–1313 (2022).
Article CAS PubMed Google Scholar
Li-Weber, M. & Krammer, P. H. Regulation of IL4 gene expression by T cells and therapeutic perspectives. Nat. Rev. Immunol. 3, 534–543 (2003).
Article CAS PubMed Google Scholar
Irla, M., Hollander, G. & Reith, W. Control of central self-tolerance induction by autoreactive CD4+ thymocytes. Trends Immunol. 31, 71–79 (2010).
Article CAS PubMed Google Scholar
Pandiyan, P. & Lenardo, M. J. The control of CD4⁺CD25⁺Foxp3⁺ regulatory T cell survival. Biol. Direct 3, 6 (2008).
Article PubMed PubMed Central Google Scholar
Zhou, J. Y., Alvarez, C. A. & Cobb, B. A. Integration of il-2 and il-4 signals coordinates divergent regulatory T cell responses and drives therapeutic efficacy. Elife 10, 1–27 (2021).
Article Google Scholar
Yang, W. C. et al. Interleukin-4 supports the suppressive immune responses elicited by regulatory T cells. Front. Immunol. 8, 1508 (2017).
Article PubMed PubMed Central Google Scholar
Dijke, I. E. et al. Discarded human thymus is a novel source of stable and long-lived therapeutic regulatory T cells. Am. J. Transplant. 16, 58–71 (2016).
Article CAS PubMed Google Scholar
Bernaldo-de-Quirós, E. et al. First-in-human therapy with Treg produced from thymic tissue (thyTreg) in a heart transplant infant. J. Exp. Med. 220, e20231045 (2023).
Article PubMed PubMed Central Google Scholar
Wyss, L. et al. Affinity for self antigen selects Treg cells with distinct functional properties. Nat. Immunol. 17, 1093–1101 (2016).
Article CAS PubMed PubMed Central Google Scholar
De Togni, P. et al. Abnormal development of peripheral lymphoid organs in mice deficient in lymphotoxin. Science. 264, 703–707 (1994).
Article Google Scholar
Shinkai, Y. et al. RAG-2-deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell 68, 855–867 (1992).
Article CAS PubMed Google Scholar
Morin, S. O. et al. In the absence of its cytosolic domain, the CD28 molecule still contributes to T cell activation. Cell. Mol. Life Sci. 72, 2739–2748 (2015).
Article CAS PubMed PubMed Central Google Scholar
Barnden, M. J., Allison, J., Heath, W. R. & Carbone, F. R. Defective TCR expression in transgenic mice constructed using cDNA- based α- and β-chain genes under the control of heterologous regulatory elements. Immunol. Cell Biol. 76, 34–40 (1998).
Article CAS PubMed Google Scholar
Kurts, C. et al. Constitutive class i-restricted exogenous presentation of self antigens in vivo. J. Exp. Med. 184, 923–930 (1996).
Article CAS PubMed Google Scholar
Wang, Y. et al. Th2 lymphoproliferative disorder of lat Y136F mutant mice unfolds independently of TCR-MHC engagement and is insensitive to the action of Foxp3⁺ regulatory T cells. J. Immunol. 180, 1565–1575 (2008).
Article CAS PubMed Google Scholar
Kennedy, M. K. et al. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. J. Exp. Med. 191, 771–780 (2000).
Article CAS PubMed PubMed Central Google Scholar
Mohrs, M., Shinkai, K., Mohrs, K. & Locksley, R. M. Analysis of type 2 immunity in vivo with a bicistronic IL-4 reporter. Immunity 15, 303–311 (2001).
Article CAS PubMed Google Scholar
Borelli, A., Zamit, C. & Irla, M. Medullary thymic epithelial cell antigen-presentation assays. Bio Protoc. 13, e4865 (2023).
Article CAS PubMed PubMed Central Google Scholar
Sergé, A., Bailly, A. L., Aurrand-Lions, M., Imhof, B. A. & Irla, M. For3D: Full organ reconstruction in 3D, an automatized tool for deciphering the complexity of lymphoid organs. J. Immunol. Methods 424, 32–42 (2015).
Article PubMed Google Scholar
Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e29 (2021).
Article CAS PubMed PubMed Central Google Scholar
Aussel, R. et al. ShIVA: a user-friendly and interactive interface giving biologists control over their single-cell RNA-seq data. Sci. Rep. 13, 14377 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 32, 2847–2849 (2016).
Article CAS PubMed Google Scholar
Alquicira-Hernandez, J. & Powell, J. E. Nebulosa recovers single-cell gene expression signals by kernel density estimation. Bioinformatics 37, 2485–2487 (2021).
Article CAS PubMed Google Scholar
Spinelli, L. CIML-bioinformatic/MIlab_LTaTreg-Thymus: article publication release. Zenodo https://doi.org/10.5281/zenodo.12705747 (2024).
Article Google Scholar

Download references

Acknowledgements

We are grateful to Jacques Nunes and Geoffrey Guittard (CRCM, Marseille, France) for providing us with Cd28^−/− mice, to Bernard Malissen (CIML, Marseille, France) for Foxp3^eGFP mice, to Joost P.M. van Meerwijk (Inifinity, Toulouse, France) for Rag2^GFPxFoxp3^Thy1.1, Il2^−/− and Il15^−/− mice and to David Voehringer and Andreas Ruhl (FAU, Erlangen, Germany) for 4get mice. We also thank the National Institutes of Health Tetramer Facility for PBS57:CD1d tetramers. We thank the CIML animal facility, flow cytometry, PICSL imaging facility (ImagImm), and Genomics platform for technical support. We are grateful to Benoit Salomon (Infinity, Toulouse, France), Lena Alexopoulou (CIML, Marseille, France), and Arnauld Sergé (LAI, Marseille, France) for helpful discussions. We thank Serge van de Pavert (CIML, Marseille, France) for help in preparing the Supplementary Fig. 13. This work was supported by the ARC Foundation (PJA20171206491 to M.I.), Inserm CoPoC—proof of concept (N° MAT-PI-17326-A-01 to M.I.), Agence de Biomédecine (21GREFFE009 to M.I.) and the Agence Nationale de la Recherche (grant ANR-22-CE18-0045-01 Reality to M.I.). This work was supported by the France Génomique national infrastructure, funded as part of the “Investissements d’Avenir” program managed by the Agence Nationale de la Recherche (contract ANR-10-INBS-0009). We also acknowledge financial support from n° ANR-10-INBS-04-01 France Bio Imaging. A.B. was supported by a PhD fellowship from the Ministère de l’Enseignement Supérieur et de la Recherche et de l’Innovation (MESRI) and by the Fondation ARC pour la Recherche sur le Cancer (ARCDOC42022010004424). The Supplementary Fig. 13 was created with BioRender.com.

Author information

Cécile Apert
Present address: Microenvironment & Immunity Unit, Institut Pasteur, Paris, France

Authors and Affiliations

Aix-Marseille University, CNRS, INSERM, CIML, Centre d’Immunologie de Marseille-Luminy, Turing Centre for Living Systems, Marseille, France
Alexia Borelli, Jérémy C. Santamaria, Cloé Zamit, Jessica Chevallier, Philippe Pierre, Rafael J. Argüello, Lionel Spinelli & Magali Irla
Toulouse Institute for Infectious and Inflammatory Diseases (Infinity), INSERM UMR1291—CNRS UMR5051—University Toulouse III, Toulouse, France
Cécile Apert

Authors

Alexia Borelli
View author publications
You can also search for this author in PubMed Google Scholar
Jérémy C. Santamaria
View author publications
You can also search for this author in PubMed Google Scholar
Cloé Zamit
View author publications
You can also search for this author in PubMed Google Scholar
Cécile Apert
View author publications
You can also search for this author in PubMed Google Scholar
Jessica Chevallier
View author publications
You can also search for this author in PubMed Google Scholar
Philippe Pierre
View author publications
You can also search for this author in PubMed Google Scholar
Rafael J. Argüello
View author publications
You can also search for this author in PubMed Google Scholar
Lionel Spinelli
View author publications
You can also search for this author in PubMed Google Scholar
Magali Irla
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

A.B., J.C.S, C.A., C.Z., and M.I. conducted the experiments, analyzed and interpreted the data. L.S. and J.C. analyzed scRNA-seq data. R.J.A. provided assistance with data interpretation, discussion on SCENITH and scRNA-seq data and reviewed the manuscript. P.P. provided us anti-puromycin antibody. A.B., J.C.S., and M.I. wrote the manuscript. M.I. initiated, supervised, and conceived the study.

Corresponding author

Correspondence to Magali Irla.

Ethics declarations

Competing interests

The authors declare the existence of a financial/non financial competing interest. M.I. holds a patent related to Treg genetically modified for the lymphotoxin alpha gene (PCT/EP2020/063346; WO2020229546A1). There are restrictions to the commercial use of SCENITH^TM due to pending patent application (PCT/EP2020/060486;WO2020212362A1 to R.J.A.). All other authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Derk Amsen and the other anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

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

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1-2

Reporting Summary

Source data

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

Borelli, A., Santamaria, J.C., Zamit, C. et al. Lymphotoxin limits Foxp3⁺ regulatory T cell development from Foxp3^lo precursors via IL-4 signaling. Nat Commun 15, 6976 (2024). https://doi.org/10.1038/s41467-024-51164-5

Download citation

Received: 26 September 2023
Accepted: 31 July 2024
Published: 14 August 2024
DOI: https://doi.org/10.1038/s41467-024-51164-5
Springer Nature Limited

Lymphotoxin limits Foxp3+ regulatory T cell development from Foxp3lo precursors via IL-4 signaling

Abstract

Similar content being viewed by others

Introduction

Results

LTα1β2 expression is upregulated during thymic Treg development

LTα1β2 is upregulated by instructive signals of the “two-step model” of Treg development

LTα1β2 limits Treg generation through the Foxp3lo TregP developmental pathway

Lta expression controls the metabolic profile of Treg during their thymic development

LTα1β2-regulated Treg development depends on IL-4 production by mTEC

Discussion

Methods

Mice

Bone marrow chimeras

Thymic cell isolation

In vivo IL-4 neutralization experiments

In vitro antigen-specific Treg conversion assays

In vitro TregP conversion assays with IL-4 stimulation

In vitro activation of Treg cell subsets

Flow cytometry

SCENITH™ metabolic profiling

Immunofluorescence staining

RT-qPCR

scRNA-seq library preparation

Data analysis for scRNA-seq analysis

Statistical analysis

Reporting summary

Data availability

Code availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

Search

Navigation

Lymphotoxin limits Foxp3⁺ regulatory T cell development from Foxp3^lo precursors via IL-4 signaling

LTα₁β₂ expression is upregulated during thymic T_reg development

LTα₁β₂ is upregulated by instructive signals of the “two-step model” of T_reg development

LTα₁β₂ limits T_reg generation through the Foxp3^lo T_regP developmental pathway

Lta expression controls the metabolic profile of T_reg during their thymic development

LTα₁β₂-regulated T_reg development depends on IL-4 production by mTEC

In vitro antigen-specific T_reg conversion assays

In vitro T_regP conversion assays with IL-4 stimulation

In vitro activation of T_reg cell subsets