Inducible cAMP Early Repressor Limits Glycolytic Programming in Regulatory T Cells and Supports Control of Inflammatory Colitis

2.1. Deletion of ICER Does Not Impair Treg Development or Suppressive Function Under Homeostatic Conditions

To investigate the role of ICER in Treg function, we crossed ICER ^fl/fl mice [22] with Foxp3 ^cre mice [23] to generate the Treg‐specific ICER‐deficient line ICER ^fl/fl Foxp3 ^cre. Efficient loss of the P2‐derived ICER transcript in Tregs was confirmed by RT‐qPCR (Figure S1A). A germline‐recombined ICER allele arose during breeding and was subsequently maintained as the ICER ko allele. Germline recombination of floxed alleles has previously been reported for several Cre‐driver lines, including the Rgs9 Cre line [24]. The ICER ko allele used here was validated by locus‐specific genotyping and then used as the ICER ^ko/ko line to examine ICER function independently of Foxp3 (Figure S1A). To determine whether the loss of ICER affects Treg development, tissue distribution, or survival under homeostatic conditions, we analysed Treg frequencies and absolute numbers in secondary lymphoid organs (SLOs) and in the small intestinal lamina propria (LP) of ICER ^fl/fl Foxp3 ^cre mice in comparison with ICER ^fl/wt Foxp3 ^cre and ICER ^wt/wt Foxp3 ^cre littermates. No significant decrease in Foxp3⁺ cell frequencies was observed, and total splenic Foxp3⁺ cell numbers were increased (Figure 1A; Figure S1B,C). When we assessed Treg markers CTLA‐4, Helios, and CD39, no differences could be detected (Figure S1D–G). Consistently, in vitro iTreg differentiation assays using ICER ^ko/ko mice, together with analysis of Foxp3⁺ cell frequencies in SLOs of ICER ^ko/ko mice, indicated that the loss of ICER does not impair Treg development in either the thymus or the periphery (Figure S1H–J).

Because cAMP is central to Treg‐mediated suppression of Teff proliferation, we next investigated whether the loss of ICER affects Treg suppressive function. We performed in vitro suppression assays in which sorted nTregs from SLOs of ICER ^fl/fl Foxp3 ^cre and Foxp3 ^cre mice were co‐cultured with naïve CD45.1⁺ T cells and dendritic cells. After 96 h, Teff proliferation was assessed (Figure 1B; Figure S2A). We found that suppression of Teff proliferation was independent of ICER expression in Tregs (Figure 1C). To further test the role of ICER under conditions of enhanced cAMP signalling, we performed suppression assays in the presence of NECA, a broadly acting adenosine receptor agonist that increases cAMP signalling [25]. While NECA treatment induced Icer transcription (Figure S2B), it also increased suppression by both WT and ICER‐deficient Tregs (Figure S2C,D), suggesting that cAMP‐mediated suppression of Teff proliferation by Tregs does not require ICER.

2.2. Lack of ICER Improves Foxp3 Stability in a Th1‐Like Inflammatory Environment

As cAMP metabolism has been linked to Treg stability, we next asked whether ICER contributes to the maintenance of Foxp3 expression under inflammatory conditions. It is well‐established that exposure to inflammatory cytokines such as IFN‐γ can destabilise Foxp3 expression in Tregs, which impacts their function and contributes to autoinflammatory diseases [1]. To address this, we cultured sorted Tregs from SLOs of ICER ^fl/fl Foxp3 ^cre and Foxp3 ^cre mice for up to 12 days either under control conditions (IL‐2) or in the additional presence of IL‐12 and IFN‐γ. Surprisingly, ICER deficiency significantly improved Treg stability in the presence of IL‐12 and IFN‐γ, although no difference was observed under control conditions (Figure 1D). We tested Th1 differentiation in the absence of ICER to understand whether the general IL‐12 response could be compromised, but found comparable differentiation in WT and ICER‐KO cells (Figure S2E). To determine whether the increased stability of ICER‐deficient Tregs translated to enhanced suppressive activity, we next performed suppression assays under inflammatory conditions. For this, iTregs were generated from ICER ^fl/fl Foxp3 ^cre and Foxp3 ^cre mice (Figure 2A), and IL‐12 was added to suppression assay co‐cultures. Although the overall suppressive capacity was modestly reduced in the presence of IL‐12, no significant difference in suppression of Teff proliferation was observed between ICER‐competent and ICER‐deficient Tregs (Figure 2B).

In addition to suppressing proliferation, Tregs are also known to restrain inflammatory cytokine production by Teff cells [26]. We therefore assessed IFN‐γ production by CD45.1⁺ Teff cells after 96 h of co‐culture. CD45.1⁺ Teff‐cell IFN‐γ production was higher in co‐cultures containing ICER‐deficient Tregs than WT Tregs, particularly at low Treg:Teff ratios (Figure 2C). This difference did not appear to originate from the Treg‐intrinsic IFN‐γ production, since upregulation of IFN‐γ in response to IL‐12 stimulation was not altered in the absence of ICER (Figure 2D).

Because IL‐10 was previously shown to be involved in the regulation of IFN‐γ production [27], we next used ELISA to measure IL‐10 concentrations in suppression assay supernatants after 96 h of co‐culture. Surprisingly, we observed an increase in IL‐10 levels in the supernatants from ICER‐deficient Treg co‐cultures relative to co‐cultures containing WT Tregs (Figure 2E). These data reveal a dissociation between Foxp3 stability and inflammatory effector control. While ICER limits Foxp3 stability, it supports the ability of Tregs to restrain inflammatory cytokine production by Teff cells.

2.3. ICER Restrains Treg Glycolysis

Treg proliferation is known to be supported by glycolytic metabolism [6]. We observed that ICER‐deficient Tregs were more abundant in the spleen (Figure S1B) and, consistently, sorted ICER‐deficient nTregs proliferated more strongly during IL‐2‐driven expansion, without a reduction in cell survival (Figure 3A; Figure S3A–C). To investigate the metabolic basis of this enhanced proliferation, we performed Seahorse analysis of iTregs to assess their metabolic status. ICER‐deficient iTregs displayed significantly increased basal glycolysis and glycolytic capacity (Figures 3B,C), consistent with their enhanced proliferative behaviour. Additionally, we examined mitochondrial metabolism and found that maximal respiration was significantly increased in ICER‐deficient cells, whereas basal respiration was unchanged (Figure S3D,E). We also observed a reduced OCR/ECAR ratio in ICER‐deficient Tregs (Figure 3D), suggesting a relative shift toward glycolytic metabolism. Interestingly, increased glycolytic activity did not confer a survival disadvantage for ICER‐deficient Tregs in low glucose conditions (Figure S3F). Since mTOR signalling is a central regulator of glycolytic programming, we wondered whether altered mTOR activity could account for this phenotype. Unexpectedly, Western blot analysis of mTOR phosphorylation at Ser2448 did not reveal any difference between WT and ICER‐deficient iTregs (Figure 3E; Figure S3G). To further investigate whether changes to mTOR activity accounted for the increased glycolysis in ICER‐deficient cells, we added 100 nM rapamycin to iTreg cultures after 72 h of differentiation and performed Seahorse analysis 24 h later. Addition of Rapamycin decreased glycolytic capacity in both ICER‐competent and ICER‐deficient cells (Figure 3F; Figure S3H). However, glycolysis levels of ICER‐deficient Tregs remained higher than in WT Tregs, with glycolytic capacity still showing a significant increase, indicating an additional mTOR‐independent regulatory mechanism (Figure 3G). Enhanced mTOR signalling is known to impair Treg suppressive function and thus, the absence of an effect upon mTOR phosphorylation in ICER‐deficient Tregs is consistent with their unchanged suppressive capacity (Figure 1C). Together, these data suggest that increased glycolysis in ICER‐deficient Tregs is not fully normalised by rapamycin and is therefore not fully accounted for by the measured mTOR axis.

2.4. Deletion of ICER in Tregs Exacerbates Pathology During DSS Colitis

Finally, to test our findings in vivo, we performed dextran sodium sulphate (DSS) colitis experiments on ICER ^fl/fl Foxp3 ^cre animals and ICER ^fl/fl littermates (Figure 4A). The general disease parameters, such as body weight or colon shortening, were not markedly different between the groups (Figure 4B; Figure S4A). Histological analysis of colon sections, however, revealed more severe inflammation, greater inflammatory extent, more pronounced epithelial alterations by trend, and significantly increased disruption of mucosal architecture in ICER ^fl/fl Foxp3 ^cre mice (Figures 4C,D). These findings prompted us to investigate whether Tregs in the inflamed intestinal environment could fail to adequately control effector cytokine responses. Analysis of Tregs from the mesenteric lymph nodes (mLNs) and colonic LP revealed a modest increase in Treg frequencies and absolute numbers in both compartments, as well as a non‐significant trend towards higher IL‐10 frequencies consistent with increased Treg accumulation and/or maintenance in inflamed tissue (Figure 4E; Figure S4B–D). Further characterisation of the Treg compartment showed that neither RORγt expression nor cytokine production by Tregs nor Treg markers CTLA‐4, CD39, Helios and ST‐2 were significantly altered by ICER deficiency (Figure 4F,G; Figure S4E–J). Similarly, there was no increase in RORγt‐expressing or IL‐17A‐producing Foxp3⁻ CD4⁺ T cells, indicating no shift in Treg/Th17 balance (Figure 4H; Figure S4K), and no increase in TNF‐α production (Figure S4L). In contrast, and in line with our in vitro data, we found a significant increase in IFN‐γ⁺ Teff cells in colonic LP, indicating that ICER expression in Tregs contributes to control of inflammatory IFN‐γ production by Teff cells (Figure 4I). TIGIT‐expressing Tregs are known to suppress Th1 and Th17 responses and promote tissue repair [28, 29]. When we analysed TIGIT expression, we unexpectedly found that in ICER‐deficient mice, a higher percentage of Tregs expressed TIGIT (Figure 4J). This raises the possibility that in ICER‐deficient Tregs IFN‐γ‐control and tissue‐repair functions are impaired while TIGIT expression is retained, fitting our observation of increased tissue damage in ICER ^fl/fl Foxp3 ^cre mice.

Together, these findings suggest that ICER in Tregs contributes to control of inflammatory effector responses in vivo, consistent with ICER‐mediated regulation of Treg metabolic and suppressive function.

3.1. Data Limitations and Perspectives

One limitation of this study is the use of DSS colitis as the sole in vivo model to assess the consequences of ICER deletion in Tregs. Although this model was sufficient to reveal aggravated intestinal histopathology and impaired control of effector IFN‐γ  responses, it may not capture the full spectrum of Treg metabolic adaptation in vivo. In addition, while ICER deficiency was associated with enhanced glycolysis in Tregs, the molecular basis of this phenotype remains unresolved. Our data indicate that the increased glycolytic state is not fully explained by the mTOR‐related readouts examined here, but they do not yet establish the complete underlying mechanism. Thus, this study identifies ICER as a candidate regulator of Treg metabolic programming and inflammatory function, while the precise pathways linking ICER to metabolism and Treg effector control remain to be defined.

4.1. Animals

All mice were bred and maintained under specific pathogen‐free conditions at the animal facility of the Medical Center of Johannes Gutenberg University Mainz (Mainz, Germany) and Ribeirão Preto Medical School (Ribeirão Preto, Brazil). C57BL/6JRj mice were obtained from Janvier Labs and B6.SJL‐Ptprca Pepcb/BoyJ (CD45.1) mice were obtained from the Jackson Laboratory. ICER flox animals, described previously [22], were imported from the Institute of Infection Immunology TWINCORE—Centre for Experimental and Clinical Infection Research, Hannover, and crossed to B6129S‐Tg(Foxp3‐EGFP/icre)1aJbs/J (Foxp3cre) mice imported from Core Facility Laboratory Animal Services (CF‐LAS), Neuherberg. Full knockout mice were expanded from a spontaneous germline deletion event [24] and validated using primers: ICERko fw: 5’‐TGGAGCCTTGTTGAGACTG‐3’, ICERko rev: 5’‐ TGAAACTGTAGCCTAGTGGTG‐3’, ICERfl A fw: 5’‐GCCTTGTTGAGACTGTCATC‐3’, ICERfl A rev: 5’‐ATCTTCTCTCCCCTATCCCC‐3’, ICERfl B fw: 5’‐GCTTCCCCACAATTTAGCTC‐3’, ICERfl B rev: 5’‐TGTAGCCTAGTGGTGGTAAAC‐3’. All experiments were performed according to the guidelines of the Animal Welfare Committee of the Ribeirão Preto Medical School, University of São Paulo (protocol number: 1399/2024). For in vitro and ex vivo experiments, mice of both sexes were used in the age range of 7–20 weeks.

4.2. DSS Colitis

12–16‐week‐old male mice were offered water containing 2.5% DSS (MP Biomedicals) ad libitum for 6 days. Thereafter, animals were offered untreated water for an additional 6 days. Weight was measured daily. On the twelfth day, animals were sacrificed, and mLNs and colonic LP were prepared for flow cytometry analysis as well as for histology. Eight to twelve animals were used per group.

4.3. Histology

For haematoxylin and eosin staining, tissue samples were fixed in 4% formaldehyde (Carl Roth GmbH+Co.KG, Karlsruhe, Germany) and embedded in paraffin. Images were taken using a light microscope type Olympus BX51 (Olympus Corporation, Tokyo, Japan), and digitally using NDP.view2 (Hamamatsu) software. Scale bar: 250 µm for the detailed section and 200 µm for the overview section. Blinded histological scoring of colon sections was performed according to Erben et al. [48].

4.4. Lamina Propria Isolation

LP isolation was performed as described previously [49]. Briefly, colons or small intestines were isolated, Peyer's patches were removed, cut open longitudinally, and washed in ice‐cold PBS. Tissue was then incubated for 20 min in PBS with 30 mM EDTA and 1.5 mM DTT on ice, washed again, and incubated in PBS with 30 mM EDTA for 20 min at 37°C. Afterward, the tissue was washed in HBSS +calcium/+magnesium, minced, and digested with 0.5 mg/mL Collagenase D (Roche), 0.5 mg/mL DNase I (Applichem), and 0.04 mg/mL Dispase I (Corning) for 30 min at 37°C. Afterward, tissue was strained and resuspended in 40% Percoll. After, the suspension was overlaid on 80% and centrifuged for 20 min at 1200 x g. LP fraction at the interphase of the Percoll gradient was collected, washed, and stained for flow cytometry.

4.5. Murine T Cell Enrichment

Naïve T cells were isolated from spleen and peripheral lymph nodes using the EasySep Mouse Naive CD4⁺ T Cell Isolation Kit (Stemcell Technologies) according to the manufacturer's instructions. For sorting, the EasySep Mouse CD4⁺ T Cell Isolation Kit (Stem Cell) was used for pre‐enrichment and cells were stained with anti‐CD4 (RM4‐5; eBiosciences), anti‐CD25 (PC61.5; eBiosciences), anti‐CD44 (IM7; eBiosciences), and anti‐CD62L (MEL‐14; Tonbo Biosciences) antibodies and sorted using the Aria III (BD).

4.6. In Vitro T Cell Differentiation

Isolated naïve T cells were stimulated in RPMI (ThermoFisher) containing 10% FCS (Sigma), penicillin–streptomycin cocktail (ThermoFisher), and 50 µM β‐mercaptoethanol (Gibco). For iTreg differentiation, cells were stimulated with plate‐bound 5 µg/mL anti‐CD3 (BioLegend), soluble 1 µg/mL anti‐CD28 (BioLegend) in the presence of 200 U/mL rhIL‐2 (Peprotech) and 1 ng/mL rhTGF‐β (Peprotech) for 96 h. After 48 h, cells were treated with an additional 100 U/mL of rhIL‐2. For Th1 differentiation, cells were stimulated with plate‐bound 10 µg/mL anti‐CD3 (BioLegend), soluble 1 µg/mL anti‐CD28 (BioLegend) in the presence of 10 µg/mL anti‐IL‐4 (BioLegend) and 20 ng/mL IL‐12 (Peprotech) for 96 h. For Th0 differentiation, cells were stimulated with plate‐bound 10 µg/mL anti‐CD3 (BioLegend) and soluble 1 µg/mL anti‐CD28 (BioLegend) in the presence of 200 U/mL rhIL‐2 (Peprotech) for 96 h. For low‐glucose conditions, differentiation was performed using Seahorse XF RPMI Medium (Agilent) supplemented with glutamine and different concentrations of glucose.

4.7. nTreg Culture

nTregs were sorted as CD4⁺ Foxp3cre⁺ (GFP⁺) cells (Figure 1C,D) or CD4⁺ CD25⁺ CD127^low cells (Figure S3B) using anti‐CD4 (RM4‐5; eBiosciences), anti‐CD25 (PC61.5; eBiosciences), anti‐CD127 (A7R34; eBiosciences). Cells were then seeded at a density of 1 × 10⁶ cells/mL in 1 µg/mL anti‐CD3 and 5 µg/mL anti‐CD28 coated flat‐bottom plates in IMDM (ThermoFisher) containing 10% FCS, Penicillin–Streptomycin cocktail, 500 µM β‐mercaptoethanol and 1× nonessential amino acids (ThermoFisher). Starting on day 0, IL‐2 was replenished every 2 days. After 48 h, cells were transferred to uncoated wells. For stability assays, cells were additionally treated with 20 ng/mL rmIL‐12 (Peprotech) and 10 ng/mL rmIFN‐γ (Peprotech) every 2 days.

4.8. Flow Cytometry

For proliferation assays, T cells were stained before plating with 5 µM Cell Trace Violet (C34557) for 7 min at 37°C. For apoptosis measurement, cells were stained with Annexin V (1:200) (SKU 35‐6409‐T100) for 15 min at 37°C. PI (00‐6990‐50) was added to each sample immediately before acquisition on the cytometer. For all other stains, dead cells were excluded with the live/dead fixable dead cell stain kit (Life Technologies), and FcBlock was performed for 15 min at 4°C. For analysis of surface markers, cells were stained in PBS using the following fluorescence‐conjugated monoclonal antibodies: CD3e (145‐2C11; eBiosciences), CD45 (30‐F11; BioLegend), CD4 (GK1.5; eBiosciences), CD45.1 (A20; eBiosciences), CD45.2 (104; eBiosciences), CD39 (24DMS1; eBiosciences), TIGIT (GIGD7; eBiosciences), ICOS (7E.17G9; eBiosciences), and ST‐2 (U29‐93; BD). For intracellular staining of cytokine and transcription factors, cells were stained with IFN‐γ (XMG1.2; Tonbo biosciences), IL‐10 (JES5‐16E3; eBiosciences), IL‐17A (eBio17B7; eBiosciences), Foxp3 (FJK‐16s, eBiosciences), RORγt (B2D, eBiosciences), TNF‐α (MP6‐XT22; eBiosciences), CTLA‐4 (UC10‐4B9; eBiosciences), and Helios (22F6; BioLegend) using the Foxp3/transcription factor fixation/permeabilisation kit (eBiosciences) according to the manufacturer's instructions. For cytokine staining, cells were stimulated in vitro in the presence of phorbol‐12‐myristate‐13‐acetate (0.1 µg/mL; Sigma‐Aldrich) and ionomycin (1 µg/mL; Sigma‐Aldrich) for 2 h, followed by incubation for 2 h with brefeldin A (5 µg/mL; eBiosciences) at 1 × 10⁶ cells/mL before staining. Cells were acquired on a Cytoflex S (Beckman Coulter), and data were analysed with FlowJo software (v.10.8.1, Tree Star).

4.9. In Vitro Suppression Assay

Naïve CTV‐stained T cells obtained from Ly5.1 mice were seeded at a density of 5 × 10⁵/mL in complete RPMI in 96‐well round‐bottom plates. GM‐CSF DCs were generated from bone marrow of sex‐matched mice as described previously [50] and added at a density of 3 × 10⁴ cells/mL. To test their suppressive capacity, sorted nTregs or iTregs were added at different concentrations. Soluble anti‐CD3 was added at 1 µg/mL. For Figure S2, 10 µM NECA and for Figure 2, 10 ng/mL rmIL‐12 (Peprotech) were added at the beginning of co‐culture.

4.10. Seahorse

At 96 h of culture iTreg cells were harvested and plated on poly‐L‐lysine‐coated (Merck; 0.01% 4°C O/N) 96‐well XF cell culture microplates in XF assay medium (pH 7.4, both from Agilent) supplemented with 10 mM D‐glucose, 1 mM sodium‐pyruvate (Merck), and 2 mM L‐glutamine (Gibco) at a density of 15 × 10⁴ cells/well. Microplates were incubated for 45 min at 37°C in a non‐CO₂ incubator and subjected to real‐time analysis of OCR using an XFe96 Extracellular Flux Analyzer (Agilent). For the mitochondrial stress test analysis, the XF mitochondrial stress test was performed according to the manufacturer's instructions, using subsequent injections of oligomycin (1 µM), FCCP (1.5 µM), and rotenone/antimycin A (0.5 µM). For glycolysis analysis, the XF Glycolytic Stress Test was performed according to the manufacturer's instructions, using subsequent injections of glucose (10 mM), oligomycin (1 µM), and 2‐DG (50 mM). For rapamycin treatment, Tregs were treated 24 h before analysis with 100 nM rapamycin.

4.11. RT‐qPCR

RNA was extracted as per the manufacturer's instructions using the Monarch Spin RNA isolation kit (NEB). Reverse transcription was performed according to the manufacturer's instructions using the ImProm‐II Reverse Transcription System (Promega). cDNA was used at 8 ng per reaction, and RT‐qPCR was performed using iQ SYBR Green Supermix (Bio‐Rad). The cycling program was 95°C for 120 s followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Each sample was assayed in triplicate, and the results were normalized to the level of housekeeping gene Actb. Primers used: ICER fw: 5’‐ATGGCTGTAACTGGAGATGAAACT‐3’, ICER rev: 5’‐GACTGTGCAGGCTTCCTGGTGAG‐3’, β‐Actin fw: 5’‐TGTTACCAACTGGGACGACA‐3’, β‐Actin rev: 5’‐GGGGTGTTGAAGGTCTCAAA‐3’.

4.12. Western Blot

For immunoblotting, SDS‐PAGE was performed on whole‐cell lysates. Proteins were then transferred to a PVDF membrane (Bio‐Rad) and incubated with primary antibodies overnight. For detection anti‐phospho (Ser2448) mTOR (Cell Signaling, 5536) diluted 1:200 and anti‐mTOR (Proteintech, 66888‐1‐Ig) diluted 1:200, anti‐rabbit IgG (Cell Signaling, 7074) diluted 1:10,000 and anti‐mouse IgG (Jackson ImmunoResearch Lab Inc., 115‐036‐062) diluted 1:10,000 were used. As a loading control for protein samples, anti‐GAPDH antibody (Cell Signaling, 5174) diluted 1:5000 was applied. ImageJ was used for quantification.

4.13. ELISA

The suppression assay supernatants were collected after 96 h of co‐culture and stored at −80°C until use. IL‐10 (R&D, #DY417) cytokine concentration was directly assessed by ELISA. Optical density (OD) was measured using a microplate reader (BioTek Cytation 5, Agilent), and the differential absorbance (ΔOD_450–570) was recorded.

4.14. Statistical Analysis

Statistical analysis was performed with GraphPad Prism software (version 9; GraphPad Software Inc., San Diego, CA, USA). The p‐values were calculated using a Student's t‐test, one‐way ANOVA, or two‐way ANOVA, as indicated. Data are displayed as mean ± standard deviation (SD). A p‐value of < 0.05 was considered significant and marked by asterisks (*p< 0.05; **p < 0.01; ***p < 0.001). Data distribution was assumed to be normal, but this was not formally tested.