Fructose intake driven glycolysis-ROS-EGFR axis specifically promotes the generation and pathogenicity of Th17 cells

Abstract

Th17 cells are quite heterogeneous. Treating Th17-related inflammatory disorders requires understanding the functionally diverse subtypes in the context of tissue homeostasis, which is shaped by nutrient availability among other factors. Here, we show that increased consumption of fructose exacerbates colitis and experimental autoimmune encephalomyelitis (EAE), via pathogenic Th17 cells. Fructose selectively enhances the differentiation and function of this pathogenic subtype of Th17 cells, which are induced by a combination of IL1β, IL-6 and IL-23 (pTh17). In contrast, TGFβ1and IL-6-induced homeostatic, non-pathogenic Th17 cells remain unaffected. Notably, fructose enhances metabolic activity in pTh17 cells, leading to increased ROS production and subsequently promoting pathogenic-Th17 cell differentiation. N-acetyl cysteine (NAC), a ROS scavenger, specifically impaired pathogenic-Th17 cell immunity and mitigated high-fructose regulated colitis and EAE disease. Mechanistically, ROS accumulation results in elevated EGFR expression and phosphorylation, which leads to increased nuclear translocation. Nuclear EGFR binds to STAT3, enhancing its transcriptional activity at the CNS6 and CNS9 regions of Rorc. In summary, our work describes here a mechanism through which high fructose intake specifically exacerbates pathogenic Th17-cell-related pathologies and provides potential therapeutic targets for pTh17-mediated diseases.

Th17 cells play roles in maintaining immune homeostasis but also potentially harmful inflammatory processes. Here, authors show that homeostatic and pathological Th17 cells are distinct subtypes, responding to different cytokine combinations and metabolic cues, with high fructose intake stimulating exclusively the pathological Th17 cells in vitro and in tissues.

Introduction

Th17 cells are a distinct CD4⁺ T cell subset characterized by the expression of the transcription factor retinoic acid receptor-related orphan receptor-γt (RORγt) and the production of interleukin-17A (IL-17A)^1–3. Th17 cells have prominent roles in autoimmunity and tissue inflammation². However, Th17 cells have been shown to exhibit heterogeneity, adopting both pathogenic and non-pathogenic phenotypes, in a content-dependent manner^4,5. IL-6 + IL-23 + IL-1β drive pathogenic-Th17 (pTh17) cells to promote inflammatory disorders such as inflammatory bowel disease and multiple sclerosis^6–8. In contrast, TGF-β1 + IL-6 is capable to induce non-pathogenic-Th17 cell differentiation and endeavors non-pathogenic signature to promote tissue homeostasis^9,10. To advance our understanding of Th17 cell heterogenicity and improve the treatment for pTh17-related inflammatory disorders, we must determine how pTh17 cells are specifically regulated. However, current understanding of Th17 cell differentiation and heterogeneity regulation primarily arises from studies on intracellular molecules or genetic factors^11–14. Of note, changes in environmental factors have been extensively implicated in the shaping T cell responses during autoimmunity and inflammatory disorders^15,16. Recent studies also reveal that T cell subsets are strongly influenced by nutrients uptake from their environment, which impacts their metabolism and function^17–21. However, little is known about which and how nutrients that directly influence the dichotomous of Th17 cells in immunopathology.

Immune system could be regulated by nutritional factors²². A western-style diet, which is high in sugars, salt, and saturated fat acids, is now well recognized for increasing the risk of inflammatory disorders^23,24. Fructose is a major natural dietary sugar present in fruits, honey, sweetened beverages, and processed foods containing high glucose-fructose syrup. Of note, high fructose intake disrupts immune homeostasis by directly targeting innate immune cells such as macrophages and monocytes, or indirectly modulating CD8⁺ T cells anti-tumor immunity^25–27. Whether fructose has a direct effect on T cells and the underlying mechanisms in autoimmunity remains unclear. Fructose-rich diet, accounting for over 40% of all added caloric sweeteners^28,29, has been shown to accelerate the development of pathogenic-Th17 cell-mediated inflammatory diseases, including colitis, rheumatoid arthritis, and type 2 diabetes^30–32. This correlation, along with evidence that fructose intake increases serum IL-17A levels in rats with a spontaneous hypertension model³³, suggests that fructose may contribute to Th17 cell immunity. Nonetheless, the effects of fructose on pathogenic Th17 cell responses in the context of inflammatory disorders remain largely unknown and require further investigation.

Here, we demonstrate that high-fructose consumption specifically enhances the differentiation of pathogenic-Th17 cells while has no effect on TGFβ1 induced non-pathogenic Th17 cells, thereby exacerbating inflammatory disorders in experimental colitis and EAE models. Additionally, fructose intake enhances metabolic levels, promoting reactive oxygen species (ROS) accumulation, which subsequently facilitates pathogenic-Th17 cell generation and function. Mechanistically, fructose specifically promotes pTh17 cell differentiation via a ROS-EGFR-STAT3 axis. Targeting this hub can suppress pathogenic-Th17 cell immunity and ameliorate inflammatory disorders. Collectively, this study elucidates how fructose contributes to the pathogenesis of pathogenic-Th17 cell-mediated autoimmunity.

Results

High-fructose consumption exacerbated inflammatory disorders in colitis disease model

We firstly investigate the impact of high-fructose intake on experimental inflammatory disorders, which mediated by pathogenic Th17 cells in vivo⁶. We used a well-established inducing experimental colitis model by adoptive transfer of CD4⁺CD25⁻CD45RB^hi naive T cells into Rag1^−/− recipient mice³⁴. The Rag1^−/− mice were fed with 10% fructose in water (Fru group) or normal water only (Ctrl group) 2 days before T cell transfer, as illustrated in Fig. 1A. Of note, an adult mouse drinks approximately 4 to 6 milliliters of water per day. This led to approximately 10% of the total daily caloric intake for the mice, mirroring the estimated energy intake from fructose consumption in a typical human daily diet. This fructose amount reflects roughly equivalent to a 12-ounce sugar-sweetened beverage such as cola plus naturally occurring fructose from sources like fruits³⁵. As expected, T cell-transferred mice with the high-fructose diet exhibited more severe weight loss, shorten colon length and more severe pathology than the mice given regular drinking water (Fig. 1B–E). Consistently, high-fructose consumption resulted in a specific increased pathogenic Th17 cell responses in diseased mice. The percentage of IL-17A⁺ and GM-CSF⁺ CD4⁺ T cells were elevated in the lamina propria of the colon and/or mesenteric lymph nodes (mLN) in T cell transferred Rag1^-/- mice supplemented with fructose than that with water (Figs. 1F and S1A, B), while Th1, Th2, and Treg cells were largely unaffected (Fig. S1C–H). These results therefore suggested excessive consumption of fructose accelerated colitis disease and specifically increasing Th17 cell responses.

Fig. 1. High fructose intake exacerbates colitis disease and pTh17 cell responses.

A Schematic description of adoptive transfer of CD4⁺CD25⁻CD45RB^hi naïve T cells to elicit colitis in Rag1^−/− mice (male, 6 weeks). B Weight of mice as depicted in (A), was monitored through the study. mean ± SD. n = 5 mice for Ctrl; n = 6 mice for Fru; **p < 0.01; two-way multiple-range ANOVA test. C–E Colonic length (C, D) and histopathology of colon (E) in Rag1^−/− mice transferred with CD4⁺CD25⁻CD45RB^hi naïve T cells as in (A), with 10% fructose in drinking water (Fru) or without (Ctrl). n = 5 mice; *p < 0.05; two-sided t-test. F Flow cytometry to assess the expression of IL-17A by CD4 ⁺ T cells infiltrating in the colon of mice with colitis as depicted in (A–E). n = 5 mice; ***p < 0.001; two-sided t-test. G Schema of DSS-elicited colitis in male C57BL/6 mice (6 weeks). H, I Comparison of the body weight (H) and disease activity index (I) of mice that depicted in (G). mean ± SD. n = 5 mice; *p < 0.05; two-sided t-test. J–M On day 7 post 3% DSS, all mice depicted in (G) were sacrificed, the colon (J), the length (K), weight (L), and the histopathology (M) of colon from diseased mice with fructose consumption (Fru) or without (Ctrl). n = 5 mice; *p < 0.05, **p < 0.01; two-sided t-test. N–P Flow cytometry to assess the IL-17A (N), RORγT (O), and GM-CSF (P) expression by CD4 ⁺ T cells infiltrating in the colon of mice with colitis as depicted in (G). n = 7 mice for (N), n = 5 mice for (O) and (P); **p < 0.01, ***p < 0.001; two-sided t-test. Q EAE was induced in female C57BL/6 mice (6 weeks) by MOG peptide/CFA and disease scores were evaluated. mean ± SEM. n = 5 mice; *p < 0.05; two-way multiple-range ANOVA test. R, S Flow cytometry to assess the IL-17A (R), IFN-γ (S) produced by CD4⁺ T cells infiltrating in the colon from EAE mice. n = 5 mice; *p < 0.05; two-sided t-test. Data are representative of at least three independent biological experiments. Each point represents a biological replicate. Error bars indicate standard error of the mean and centers indicate the mean values. The exact P values can be found in the Source data files. Source data are provided as a Source Data file.

To mimic the effect of fructose absorption in humans, immune component WT mice were given fructose containing drinking water for 7 days and then acute colitis was induced by 3% dextran sulfate sodium (DSS), as illustrated in Fig. 1G. Similar to results observed in T cell transfer colitis model, WT mice with high-fructose consumption aggravated colitis disease, as evidenced by greater weight loss, higher disease index score, shorter colon, more severe histopathology, and increased lymphocytes infiltration in colon (Fig. 1H–M). Analysis of colon-infiltrating T cells in diseased mice with high-fructose consumption revealed a substantial increase in the Th17 cell population and proinflammatory GM-CSF⁺ expression which is instrumental for Th17 cell pathogenicity (Fig. 1N–P), but Th1 cells, Th2 cells, Treg cells, B cells dendritic cells (DC), and macrophage were largely unaltered (Fig. S1I–N). were largely unaltered (Fig. S1H–M). Fructose intake could affect microbiota homeostasis, which contributes to the development of colitis^36,37. To exclude the possibility that fructose may promoted colitis disease via microbiota, mice received antibiotic treatment and fructose intake during DSS-induced colitis (Fig. S2A). Interestingly, fructose intake still enhanced colitis disease in microbiota-depleted mice as evidenced by severe body weight loss, shortened colon length, and histopathology (Fig. S2B–E). Furthermore, fructose was able to promote Th17 cell response in mice received antibiotic treatment, indicating fructose promoted Th17-mediated inflammatory disorders in microbiota-independent manner (Fig. S2F, G).

To understand if fructose has a broad effect on inflammatory disorders, WT mice were induced with EAE disease, a mouse model of human multiple sclerosis. Mice with fructose intake exhibited more severe EAE disease (Fig. 1Q). In addition, EAE diseased mice that received fructose revealed increased IL-17A, RORγT, and GM-CSF in spinal cord-infiltrating CD4 T cells, while IFNγ expression was unchanged (Figs. 1R, S and S1O, P). Altogether, these results confirmed that high-fructose intake exacerbated autoimmunity and specifically promoted a pathogenic Th17 cell response in vivo.

Fructose intake aggravated inflammatory disorders through pathogenic Th17 cells

The question remains whether the exacerbation of inflammatory disorders after fructose intake was dependent on pathogenic Th17 cells. We then established the T cell-transfer model of colitis with Il17^+/+ and Il17^−/− CD4⁺CD25⁻CD45RB^hi naive T cells in Rag1^−/− mice and fed with 10% fructose containing drinking water (Fig. 2A). As expected, high-fructose absorption led to greater weight loss, reduced colon length, increased inflammation (Fig. 2B–E) and enhanced Th17 cell generation in Il17^+/+ T cell-transferred Rag1^−/− mice compared to those given normal drinking water (Fig. 2F). However, the heavier disease severity in Il17^+/+ T cell-transferred Rag1^−/− mice with high-fructose uptake was fully rescued by IL-17 knockout (Fig. 2B–E). Despite high-fructose condition, the weight loss, shorter colon length, and tissue inflammation were significantly alleviated in Il17^−/− T cell-transferred Rag1^−/− mice compared to those with Il17^+/+ T cells (Fig. 2B–E). The improvement in colitis was clearly attributed to the absence of IL-17A, as the distribution of Th1, Th2, and Treg cells in the colon and mLN remained unaffected across these groups (Figs. 2F–I and S3A–D). Collectively, these findings demonstrate that fructose promotes autoimmunity through the intensified pTh17 cell generation and function.

Fig. 2. Fructose consumption promotes autoimmunity by increasing pTh17 cell generation.

A Schema of adoptive transfer of CD4⁺CD25⁻CD45RB^hi naïve T cells from WT or Il17a^−/− mice to elicit colitis in Rag1^−/− male mice (6 weeks). B Comparison of the weight change of mice that depicted in (A). n = 6 mice for WT+Fru; n = 5 mice for others. Data are presented as mean ± SD.; **p < 0.01, ***p < 0.001; two-way multiple-range ANOVA test. C–E The morphology (C), length (D), and histopathology (E) of colon from diseases Rag1^−/− mice transferred with WT or Il17a^−/− naïve T cells, with fructose intake or not. n = 5 mice; ns not significant, *p < 0.05, **p < 0.01; two-sided t-test. F Flow cytometry to assess the IL-17A production by CD4 ⁺ T cells infiltrating in the colon of mice with colitis as in (A–E). n = 5 mice; ns not significant, ***p < 0.001; two-sided t-test. G–I Flow cytometry to assess the Foxp3 (G), IL-4 (H), and IFN-γ (I) expressed by CD4 ⁺ T cells infiltrating in colon from mice as in this figure (A). n = 5 mice; ns not significant; two-sided t-test. Data are representative of at least three independent biological experiments. Each point represents a biological replicate. Error bars indicate standard error of the mean and centers indicate the mean values. The exact P values can be found in the Source data files. Source data are provided as a Source Data file.

High-fructose specifically promoted the differentiation of pathogenic Th17 cells

To examine the role of fructose in regulating Th17 cell differentiation in vitro, IL-6 + IL-1β + IL-23-polarized pathogenic Th17 (hereafter as pTh17) cells and IL-6 + TGF-β1-polarized, non-pathogenic Th17 cells (hereafter as βTh17) were cultured with the addition of fructose. βTh17 cells with fructose addition generated similar percentages of IL-17A⁺, GM-CSF⁺ and RORγT⁺ cells, with largely normal levels of Th17-related inflammatory genes such as Il17a, Rorc, Il23r and Csf2 when compared to control cells (Figs. 3A–C and S4A). Interestingly, we observed high fructose induced higher IL-17A and RORγT expression in pTh17 cells (Figs. 3D, E and S4B). Moreover, GM-CSF, a critical cytokine that drives the pathogenicity of pTh17 cells³⁸, was also increased in pathogenic Th17 cells in response to high-fructose addition (Fig. 3F). Similarly, fructose-treated pTh17 cells increased proinflammatory genes, including Il17a, Rorc, Il23r, and Csf2⁸ (Fig. 3G). The effect was predominantly specific to pTh17 cells, as no comparable enhance phenotype was observed during the differentiation of βTh17, Th1, Th2, or Treg cells (Fig. S4B). Furthermore, the differential effects of fructose for pTh17 cell generation were not due to a difference in T cell proliferation or survival as determined by 7-AAD/Annexin-V and CFSE labeling (Fig. S4C–F).

Fig. 3. Fructose specifically enhances the differentiation of pathogenic-Th17 cells while has no effect on βTh17 cells.

A, B Flow cytometry to assess the IL-17A (A) and RORγT (B) production by βTh17 treaded with or without fructose. n = 8 mice; ns not significant; two-sided t-test. C mRNA expression of Th17-related genes expressed by βTh17 as in (A). n = 5 mice; ns not significant; two-sided t-test. D–F Flow cytometry to assess the IL-17A (D), RORγT (E), and GM-CSF (F) production by pTh17 treaded with or without fructose. n = 9 mice for (D); n = 8 mice for (E); n = 11 mice for (F), ***p < 0.001; two-sided t-test. G mRNA expression of Th17-related genes in pTh17 as described in (D). n = 5 mice; ns not significant, **p < 0.01, ***p < 0.001; two-sided t-test. H mRNA levels of Slc2a5 in βTh17 and pTh17, assayed by qRT-PCR. n = 5 mice; **p < 0.01; two-sided t-test. I Immunoblotting of Glut5 in T cells under indicated polarizing conditions. Results are representative of 3 independent experiments. J Relative fructose uptake capacity of βTh17 and pTh17. n = 5 mice; *p < 0.05; two-sided t-test. K Relative fructose uptake capacity of βTh17 and pTh17 under indicated conditions. n = 5 mice; ns not significant, *p < 0.05, **p < 0.01; two-sided t-test. L Flow cytometry to assess the percentage of IL-17A by CD4 ⁺ T cells activated as in (K). n = 5 mice for Ctrl-sh-NC, n = 6 for others; **p < 0.01, ***p < 0.001; two-sided t-test. M mRNA expression of Th17-related genes in CD4 ⁺ T cells as in (K), assayed by qRT-PCR. n = 5 mice; **p < 0.01, ***p < 0.001; two-sided t-test. N Comparison of the clinical score of female mice (6 weeks) that depicted in (Fig. S4J). n = 5 mice; mean ± SEM, **p < 0.01, two-way multiple-range ANOVA test. O, P Flow cytometry to assess the IL-17A (O) and GM-CSF (P) production by CD4 ⁺ T cells infiltrating in the spinal cords of mice with EAE as in (O, P). n = 8 for (O); n = 6 mice for (P); **p < 0.01; two-sided t-test. Data are representative of at least three independent biological experiments. Each point represents a biological replicate. Error bars indicate standard error of the mean and centers indicate the mean values. The exact P values can be found in the Source data files. Source data are provided as a Source Data file.

Fructose is specifically and passively transported by GLUT5 (encoded by the Slc2a5 gene), a dedicated fructose transporter, or by GLUT8 (encoded by the Slc2a8 gene), a facilitative transporter for both glucose and fructose³⁹. We found that pTh17 cells exhibited higher GLUT5 expression, but not GLUT8 (Figs. 3H, I and S4G). Moreover, fructose addition further augmented GLUT5 protein level in Th17 cells (Fig. S4H). In agreement with these findings, pTh17 cells demonstrated a greater fructose uptake capacity compared to βTh17 cells, and dependent on GLUT5 (Figs. 3J, K and S4I), highlighting a prospective role of fructose transportation in promoting pTh17 cell polarization. Indeed, GLUT5 knockdown in T cells diminished the effect of fructose on pTh17 cell differentiation, as shown by decreased IL-17A percentage and pTh17-related proinflammatory genes expression, such as Il17a Rorc, Csf2, Il23r (Fig. 3L, M). Next, we employed Th17 cells adoptive transfer model to elicit experimental autoimmune encephalomyelitis (EAE) in mice. CD4⁺ T cells were differentiated into pathogenic-Th17 cells in the presence of fructose or not, and i.v adoptively transferred into irradiated wild-type mice to elicit EAE disease (Fig. S4J). Mice transferred with fructose-intake pTh17 cells, not control pTh17 cells, exhibited more severe EAE (Fig. 3N). As anticipated, T cells expressed more IL-17A and GM-CSF in the spinal cords from EAE diseased mice that received fructose-treated pTh17 cells (Fig. 3O, P). These findings therefore suggest that high-fructose intake specifically and directly promotes the generation and pathogenicity signature of pTh17 cells.

Pathogenic-Th17 cells but not βTh17 cells were highly-responsive to fructose addition in transcriptional and metabolic levels

To further interrogate the impact of fructose on pTh17 cells, we conducted genome-wide bulk RNA-seq analysis to evaluate the gene expression profiles of pTh17 cells under high-fructose condition. Gene set enrichment analysis (GSEA) confirmed that pTh17 cells directed a stronger pathogenic-Th17 cell program under high-fructose conditions, as most core pro-inflammatory signatures, including Il17a, Gzmb, IL23r, Csf2, Ccr6 were highly upregulated (Fig. S5A–C). Notably, principal component analysis (PCA) revealed that high-fructose addition significantly altered the gene profiles of pTh17 cells compared to standard medium (Fig. S5D). Nonetheless, the effect of fructose on βTh17 cell transcriptome was minimal (Fig. S5D). In line with these findings, over 2582 genes were differentially expressed in fructose-treated pTh17 cells, compared to fewer than 727 genes in βTh17 cells (Fig. S5E). Thus, these observations indicate a remarkable transcriptomic alteration in pTh17 cells but not in βTh17 cells in the presence of fructose.

Fructose treatment often induced cellular metabolic rewiring, which is crucial in regulating divergent T cell responses^40–42. Then, unbiased liquid chromatography-tandem mass spectrometry (LC-MS)-based metabolomics was performed to examine the metabolic profiles in Th17 cells with fructose intake. Of interest, we found that fructose induced a more pronounced metabolic change in pTh17 cells. On one hand, high-fructose addition resulted in distinct metabolite profiles in pTh17 cells compared to the control group, while high-fructose condition had little effects on the metabolite profiles of βTh17 cells (Fig. S5F). On the other hand, only 29 metabolites were differential regulated in βTh17 with fructose addition, while 88 molecules were identified differentially expressed in pTh17 cells cultured with fructose (Fig. S5G). Interestingly, we found serine and its downstream fates-choline production, were elevated in fructose-elicited pTh17 cells, while comparable in βTh17 cells after fructose addition (Fig. S5H). Previous compass findings that serine pathway may be involved in pTh17 cell immunity^43,44. KEGG analysis also confirmed that fructose resulted in the enrichment of serine metabolism and cholinergic synapse pathways in pTh17 cells (Fig. S5I), adding further granularity to the role of serine metabolism in pTh17 cells. Overall, our data indicate that pTh17 cells but not βTh17 cells, are highly responsive to fructose addition, at both transcriptome and metabolic levels.

Fructose promoted pTh17 cell differentiation in both aerobic glycolysis- and OXPHOS-dependent manner

Cell metabolism is highly dynamic and meticulously reprogrammed to support the demand of T cell responses in response to varying environments^20,27,40,45. Intrigued by above findings, we then investigated the metabolic pathways through which high-fructose intake promotes Th17 cell generation (Supplementary Data 1). Carbon metabolism was the most upregulated differential pathways in high fructose-treated pTh17 cells (Figs. 4A and S6A). Of note, carbon metabolism comprises three primary pathways, including glycolysis, tricarboxylic acid (TCA) cycle, and pentose phosphate pathways (PPP), is the most fundamental metabolic pathway and involved in modulating T cell responses^46,47. Because PPP signal had no effect on fructose-increased pTh17 differentiation using PPP pathway inhibitor (Fig. S6B, C), we therefore were prompted to focus on role of glycolysis and TCA cycle, the central hub of oxidative phosphorylation (OXPHOS). pTh17 cells with high-fructose uptake resembled increased glycolysis pathway and OXPHOS, and exhibited upregulated related-intermediate metabolites (Figs. 4B, C, S5I and S6A). Consistently, lactate, the by-products of glycolysis, were elevated in fructose-promoted pTh17 cells (Fig. S6D). Additionally, pyruvate acid, the final product of glycolysis and then catalyzed into acetyl-CoA, the main input for TCA, and fumaric acid, a key metabolite in TCA cycle^44,47, were more abundant in pTh17 cells in the presence of fructose (Fig. S6E, F). These results therefore suggest fructose promotes both aerobic glycolysis and OXPHOS in pTh17 cells.

Fig. 4. Fructose enhances pTh17 cell generation by promoting aerobic glycolysis and OXPHOS.

A GSEA for differentially expressed metabolites in pTh17 cells with fructose addition in comparison to control cells, using KEGG pathway database. B Heatmap of the expression of glycolysis-related metabolites in pTh17 with fructose addition. C GSEA of the oxidative phosphorylation and glycolysis genesets in βTh17 and pTh17 cells with or without fructose addition, performed on RNA-seq datasets. The normalized enrichment score (NES) and false discovery rate (FDR) were included. D Glycolytic rates were measured as in proton efflux rates (PER) in real time under basal conditions and in response to rotent/antimycin A and 2-DG in T cells activated per indicated conditions. Mean ± SD. n = 3 mice. E Basal glycolysis, compensatory glycolysis, and acidification levels post 2-DG of T cells activated as in (D). n = 3 mice; *p < 0.05, **p < 0.01, ***p < 0.001; two-sided t-test. F Oxygen consumption rates (OCR) were measured in real time under basal conditions and in response to indicated mitochondrial inhibitors in T cells activated under indicated conditions. Mean ± SD. n = 3 mice. G Basal respiration, maximal respiration, and spare respiratory capacity of T cells activated under indicated conditions as in (F), performed on OCR datasets. n = 3 mice; *p < 0.05, **p < 0.01, ***p < 0.001; two-sided t-test. H Ratio of basal glycolysis to basal respiration rate as in (F). n = 3 mice; ***p < 0.001; two-sided t-test. I, J Flow cytometry to assess the IL-17A production in pTh17 (I) or βTh17 (J), 50 mM fructose and 2 mM 2-DG was added accordingly. n = 5 mice; ns not significant, *p < 0.05, **p < 0.01, ***p < 0.001; two-sided t-test. K, L Flow cytometry to assess the percentage of IL-17A production in pTh17 (K) or βTh17 (L), 50 mM fructose, and dosage of Rot/AA addition. n = 5 mice; ns not significant, **p < 0.01, ***p < 0.001; two-sided t-test. Data are representative of at least three independent biological experiments. Each point represents a biological replicate. Error bars indicate standard error of the mean and centers indicate the mean values. The exact P values can be found in the Source data files. Source data are provided as a Source Data file.

We then sought to validate the above findings through Seahorse Bioanalyzer, including the glycolysis rate assay and mitochondrial stress test. Compared to βTh17, pTh17 cells exhibited higher glycolysis proton exchange rate (PER) as demonstrated by increased basal glycolysis rate, and reserved glycolytic level, all of which were further amplified in response to high-fructose condition (Fig. 4D, E). In addition, the basal oxygen consumption rate (OCR), as well as the spare and maximal respiratory capacities, were also significantly higher in pTh17 cells than in βTh17 cells, and were enhanced by fructose addition (Fig. 4F, G). Moreover, high-fructose increased the mitochondrial mass of pTh17 cells as determined by MTG assay (Fig. S6G) and enhanced the mitochondrial membrane potential in pTh17 cells as assessed by Tetramethylrhodamine methyl ester (TMRM) labeling and ATP production (Fig. S6H, I). These results together determined a higher OXPHOS activity in pTh17 cells with high-fructose intake. Of interest, we noted that pTh17 cells exhibited higher ECAR/OCR ratio while βTh17 cells displayed higher OCR/ECAR ratio, suggesting pTh17 cell differentiation may rely more on aerobic glycolysis and βTh17 cells favor OXPHOS, respectively (Fig. 4H). Our findings suggest a differential metabolic adaption between pTh17 cells and βTh17 cells. More importantly, both aerobic glycolysis and OXPHOS were reinforced in pTh17 cells by fructose treatment.

Although Th17 cell metabolism is primarily glycolytic, OXPHOS activity is also essential for its effector function^48–50. However, the relative contributions of glycolysis and OXPHOS in regulating Th17 cell dichotomy remain unclear. To dissect the role of metabolic fitness in fructose-elicited pTh17 cells, 2-DG, the nonmetabolizable glucose analog that acts as a competitive inhibitor of glycolysis by inhibiting the critical enzyme hexokinase, or with rotenone/antimycin A, the inhibitors of OXPHOS metabolism by blocking mitochondrial respiratory chains, were introduced during pTh17 cell differentiation. In the presence of 2-DG at a low dose (2 mM) without completely disrupting glycolysis⁵¹, IL-17A and RORγT production were downregulated in pTh17 cells which cultured with fructose (Figs. 4I and S6J). On the contrary, the inhibition of glycolysis led to enhanced βTh17 cell differentiation (Figs. 4J and S6K), indicate a distinct role of glycolysis in pTh17 and βTh17 cells. Furthermore, OXPHOS was essential for both βTh17 and pTh17 cells generation, as indicated by decreased IL-17A levels in fructose-intake cells cultured with rotenone/antimycin A, the respiratory chain inhibitors (Fig. 4K, L). In agreement with the finding in Fig. 4H, βTh17 appeared more susceptible to OXPHOS inhibition as a relative lower dose of rotenone/antimycin A potently reduced βTh17 generation but had no effect on pTh17 cells (Fig. 4K, L). These metabolic differences underscore the potential to target glycolysis selectively to modulate pTh17 cell-related inflammatory diseases. Overall, our findings suggest that high-fructose intake reshapes the metabolic fitness of pTh17 cells, facilitating the generation and function of pTh17 cells.

ROS accumulation by fructose intake specifically dictated pathogenic Th17 cell differentiation and function

We then sought to investigate how metabolic adaption contributed to increased pTh17 cells by fructose. ROS, the by-products of cellular metabolisms, has been widely investigated for its role in the pathogenesis of T cell mediated chronic inflammatory and inflammatory disorders^52–54. We found that pTh17 cells resembled enhanced ROS-related signature in comparison to TGFβ1 + IL-6-induced βTh17 cells, as indicated by GSEA analysis, which was further promoted by fructose intake (Fig. 5A). We next investigated the mitochondrial ROS (mtROS) production by measuring MitoSOX fluorescence and the cytosolic ROS accumulation using DCFH-DA probe. We found pTh17 cells exhibited higher cytosolic ROS and mtROS accumulation in comparison to βTh17 cells (Fig. S7A, B). Moreover, fructose addition elicited significantly higher amounts of both mtROS and cytosolic ROS in pTh17 cells, indicating a potential role of ROS in fructose-regulated pTh17 cell immunity (Fig. 5B, C). Meanwhile, the increased ROS accumulation in pTh17 cells following fructose addition was due to elevated glycolysis and OXPHOS activity, as confirmed by diminished ROS levels upon 2-DG or rotenone/antimycin A treatment (Fig. S7C, D). The effects of ROS in pTh17 cell differentiation were further investigated by N-acetyl cysteine (NAC) treatment, a ROS scavenger⁵⁵ (Fig. S7E, F). NAC treatment completely impaired the high fructose-increased differentiation and function of pTh17 cells as evident by a lower frequency of IL-17A⁺, RORγT⁺, and GM-CSF⁺ cells, but had no effect on βTh17 cells differentiation (Figs. 5D and S7G–I). In line with these observations, proinflammatory Th17 signature genes such as Il17a, Csf2, Rorc, and Il23r in fructose-exposed pTh17 cells were also downregulated by NAC treatment (Fig. 5E). With Mitoquinone (MitoQ), a specific scavenger of mtROS⁵⁵, we demonstrated that the mtROS also contributed to fructose-elicited pTh17 cell differentiation (Fig. S7J, K). Altogether, these results suggest the elevated ROS levels after fructose addition specifically dictate pTh17 cell polarization.

Fig. 5. High fructose promotes the differentiation of pathogenic-Th17 cells through ROS accumulation.

A GSEA of the reactive oxygen species (ROS) gene set in βTh17 and pTh17, performed on unbiased RNA-seq analysis. The normalized enrichment score (NES), false discovery rate (FDR). B Flow cytometry to assess cytosolic ROS in pTh17, with or without fructose, by DCFH labeling. n = 6 mice; ***p < 0.001; two-sided t-test. C Flow cytometry to assess the mitochondrial ROS level in pTh17, with or without fructose. n = 6 mice; ***p < 0.001; two-sided t-test. D Flow cytometry to assess the IL-17A production by pTh17, in the presence of 2.5 mM NAC for 4 days. n = 6 mice; **p < 0.01, ***p < 0.001; two-sided t-test. E mRNA level of Th17-related genes expressed by CD4⁺ T cells, as described in (D). n = 5 mice; ns not significant, *p < 0.05, **p < 0.01, ***p < 0.001; two-sided t-test. F Schema of adoptive transfer of CD4⁺CD25⁻CD45RB^hi T cells to elicit colitis in female Rag1^−/− mice (6 weeks). Mice were fed with 10% fructose and 2 mg/ml NAC accordingly. G Body weight of Rag1^−/− mice that depicted in (F) was recorded. Data are presented as mean ± SD. n = 5 mice; **p < 0.01; two-way multiple-range ANOVA test. H–J Colonic length (H, I) and colon histopathology (J) of mice as in (F). n = 5 mice; *p < 0.05, **p < 0.01; two-sided t-test. K–M Flow cytometry to assess the IL-17A (K), RORγT (L), and GM-CSF (M) production by CD4⁺ T cells infiltrating in colon from Rag1^−/− mice as depicted in (F). n = 5 mice; *p < 0.05, **p < 0.01, ***p < 0.001; two-sided t-test. N EAE was induced in C57BL/6 female mice (6 weeks) disease scores were evaluated. mean ± SD. n = 5 mice; *p < 0.05; two-way multiple-range ANOVA test. O, P Flow cytometry to assess the IL-17A (O), IFN-γ (P) produced by CD4⁺ T cells infiltrating in the spinal cord from EAE mice n = 5 mice; ns not significant, **p < 0.01; two-sided t-test. Data are representative of at least three independent biological experiments. Each point represents a biological replicate. Error bars indicate standard error of the mean and centers indicate the mean values. The exact P values can be found in the Source data files. Source data are provided as a Source Data file.

The forehead findings prompted us to investigate if ROS scavenging attenuates fructose-elevated autoimmunity in vivo. To address this, WT mice were pre-given 10% fructose-containing drinking water for one week and acute colitis was then induced by 3% DSS, with or without NAC, as schemed in Fig. S8A. The ROS scavenger significantly alleviated the severity of colitis in mice driven by fructose consumption, as documented by increased body weight, reduced disease scores, increased colon weight, and longer colon length (Fig. S8B–F). NAC also reduced lymphocytes infiltration and alleviated the histopathology in colon of mice with fructose intake (Fig. S8G, H). Consistent with the in vitro results, NAC administration suppressed ROS accumulation (Fig. S8I) and reduced the expression of IL-17A, RORγT, and GM-CSF in colonic CD4⁺ T cells from mice fed with high fructose-containing drinking water (Fig. S8J–L).

Furthermore, with a T cell-induced colitis model (Fig. 5F), we confirmed that T cell-derived ROS contributes to the severe colitis disease with high-fructose consumption, as determined by increased body weight, longer colon length, and attenuated histopathology after NAC treatment (Fig. 5G–J). Indeed, NAC administration diminished IL-17A, RORγT, and GM-CSF production in Rag1^−/− mice transferred with T cells (Fig. 5K–M). In the meantime, the proportion of Th1, Th2, and Treg cells in colon of diseased mice remained unaffected regardless of ROS elimination (Fig. S8M–O). Notably, NAC administration reduced the EAE disease in mice with fructose intake (Fig. 5N). Additionally, ROS elimination diminished IL-17A production in spinal-cord infiltrating CD4T cells, while has no effect on IFNγ expression (Fig. 5O, P). Taking together, these data suggest that high-fructose consumption promotes pTh17 cell differentiation and immune pathology through increased ROS accumulation. Eradication of ROS could mitigate the impact of fructose on pTh17 cell generation and autoimmunity.

ROS promoted pTh17 cell differentiation via EGFR-STAT3 signal

We then sought to determine by which ROS promoted pTh17 cell generation in high-fructose condition. ROS has been acknowledged as second messenger to trigger cell signals during tissue development, cell migration, cell differentiation, and many more biological processes⁵⁶. By comparative bio-informatic analysis of the RNA-seq datasets between fructose-intake pTh17 cells and control pTh17 cells, we found that epidermal growth factor receptor (EGFR)-related pathway, but not MAPK PATHWAY, NFKB PATHWAY, HIF1A PATHWAY, PI3K AKT mTOR PATHWAY, which could be elicited by ROS⁵⁷, was enriched in fructose-treated pTh17 cells (Figs. 6A and S9A). Because EGFR has been recently implicated in controlling T cells proliferation and function^58–60, we interrogated if EGFR signal was also involved in Th17 cells. Higher total EGFR expression and activation (phosphorylation-EGFR) were observed in pTh17 cells than in βTh17 cells and were further escalated by fructose intake (Figs. 6B and S9B). In addition, increased EGFR in pTh17 cells appeared to be ROS-dependent. NAC remarkably suppressed EGFR expression and activation in pTh17 cells with fructose addition (Fig. 6C). Of interest, afatinib, an irreversible inhibitor of EGFR activation (EGFRi)⁶¹ impaired fructose-induced increase in pTh17 cell generation and function (Fig. 6D–F). Yet, βTh17 cells differentiated normally with unperturbed expression of IL-17A following EGFRi treatment (Fig. S9C). These findings suggest a critical role of ROS-elicited EGFR in regulating pTh17 cell differentiation under high-fructose conditions.

Fig. 6. ROS promoted pTh17 cell differentiation via EGFR-STAT3-RORC axis upon fructose addition.

A Enrichment of EGFR pathway in pTh17-ctrl vs pTh17-Fru by GSEA of RNA-Seq data sets. The normalized enrichment score (NES), false discovery rate (FDR). B Immunoblotting of EGFR and phosphorylation of EGFR in pTh17, with or without fructose for 2 days. Results are representative of 3 independent experiments. n = 5 mice; *p < 0.05; two-sided t-test. C Immunoblotting of EGFR and phosphorylation of EGFR in pTh17, after treating with 50 mM fructose or 2.5 mM NAC for 2 days. Results are representative of 3 independent experiments. D–F Flow cytometry to assess the IL-17A (D), RORγT (E), and GM-CSF (F) expression by pTh17 with 50 mM fructose or 0.1 μM EGFR inhibitor (Afatinib) added accordingly for 4 days. n = 5 mice; ns not significant, **p < 0.01, ***p < 0.001; two-sided t-test. G The interactions between EGFR with STAT3, detected by pEGFR co-immuno-precipitation in wild-type CD4⁺ T cells activated for 2 days under pTh17 cell polarizing condition. Results are representative of three experiments. H The interactions between STAT3 with EGFR, detected by pSTAT3 co-immunoprecipitation in CD4 ⁺ T cells activated for 2 days under pTh17 cell-polarizing condition, and with fructose addition. I The interactions between STAT3 with EGFR, detected by pSTAT3 co-immunoprecipitation in fructose-treated pTh17 cells, in the presence of ROS scavenger NAC for 2 days. J ChIP analysis of the binding activities of STAT3 at the CNS6 and CNS9 regions on Rorc promoter in CD4 ⁺ T cells, as described in (H). n = 5 mice; **p < 0.01; two-sided t-test. K ChIP analysis of STAT3 at the CNS6 and CNS9 regions on Rorc promoter in pTh17 cells under high fructose condition, in the presence of EGFR inhibitor. n = 5 mice; *p < 0.05, **p < 0.01, ***p < 0.001; two-sided t-test. Data are representative of at least three independent biological experiments. Each point represents a biological replicate. Error bars indicate standard error of the mean, and centers indicate the mean values. The exact P values can be found in the Source data files. Source data are provided as a Source Data file.

EGFR may exert cellular functions by activating ERK signaling or by acting as a co-transcriptional factor to enhance STAT3 activity in nuclei^62–64, both ERK1/2 and STAT3 activation are also important in pTh17 cell immunity^13,65–67. However, ERK1/2 activation was unaffected with fructose addition (Fig. S9D). Additionally, given the unchanged phosphorylation levels of STAT3 in pTh17 cells with fructose (Fig. S9E) and increased EGFR in nuclei (Fig. S9F), we then investigated whether EGFR cooperates with STAT3 to enhance its transcriptional activity on Rorc expression, considering the increased Rorc mRNA observed in fructose-treated pTh17 cells. Of interest, EGFR physically interacted to STAT3 in pTh17 cells, a binding further strengthens by fructose intake (Fig. 6G, H). In addition, ROS elimination with NAC impaired the interaction of EGFR with STAT3 (Fig. 6I). STAT3 could bind conserved non-coding sequences (CNSs) 6 and 9 at the Rorc gene, which are essential for its expression during Th17 cell differentiation⁶⁸. Moreover, we found that fructose promoted the binding of STAT3 on both CNS6 and CNS9 regions of Rorc locus, indicating an increased transcriptional activity on Rorc loci in pTh17 cells, as determined by ChIP assay (Fig. 6J). EGFRi strongly suppressed the binding of STAT3 on Rorc loci in fructose-treated pTh17 cells (Fig. 6K). In sum, these data suggest that ROS promotes the cooperation of EGFR with STAT3, to enhance STAT3 transcriptional activity at the Rorc locus, thereby promoting pTh17 differentiation under high-fructose condition.

In this study, we found that increases in fructose intake could specifically augment pTh17 cell generation and function through enhanced metabolic fitness and ROS-EGFR-STAT3 axis. Targeting this hub could curb pTh17 cell-related immune-pathologies. Collectively, these findings emphasize the importance of environmental factors in shaping pTh17 cell responses and their implications for disease prevention and treatment (Supplementary Fig. 10).

Discussion

Th17 cells exert both beneficial or detrimental effects under context-specific conditions⁵. To develop more effective interventions for autoimmunity, it is imperative to dissect the dichotomy of Th17 cells. Research has shown that genetic predisposition accounts for approximately 30% of autoimmune diseases, underscoring the significant role of non-genetic environmental factors-such as dietary components, chemicals, stress, and microbiota-in driving the global rise of autoimmune disorders^69–73. Although dietary monosaccharides such as glucose and mannose have been implicated in T cell immunology^20,55, whether other monosaccharides, such as fructose directly regulates T cell responses remain unknown. Here, we further reveal a critical dietary supplement, fructose, could specifically enhance mouse pTh17 cell generation and function through ROS accumulation, without affecting on βTh17 cells, highlighting a fructose-restrict diet is important during treating inflammatory disorders. These data suggest that various monosaccharides play a distinct function in T cell subsets. Our study not only elucidates the mechanisms that fructose leads to increased autoimmunity, but also expands our understanding in how environmental factors intake dictate Th17 cell heterogeneity.

While some studies report that IL-17 deficiency in CD4⁺ T cells fails to rescue mice from IBD, others have found that IL-17A deletion significantly alleviates disease severity in both DSS-induced acute colitis and T cell transfer-induced chronic colitis models^55,74–76. Here, we also found the heavier disease severity in Il17^+/+ T cell-transferred Rag1^−/− mice with high-fructose uptake was fully rescued by IL-17 knockout. Evidences revealed increased IL-23 and pTh17 signature in inflamed colon and attribute to colitis disease^77–79. These quite controversial effects of IL-17 suggest the intricate effects of Th17 immunity in gut and indicating the microenvironment of Th17 cells may affect the outcome of IL-17A depletion. Emerging evidence suggest that pathogenic and non-pathogenic Th17 cells are functionally distinct and controlled by unique molecular network^4,5. Neutralizing antibodies targeting or genetic ablation IL-23 have demonstrated efficacy in mitigating inflammatory disorders and TH17 responses⁷⁹, suggesting pathogenic-Th17 may be a promising target during the intervention of colitis, without affecting TGFβ-induced non-pathogenic Th17 cells.

Multiple mechanisms may contribute to the differential effects of fructose on pathogenic-Th17 cell and βTh17 cells. One could be that there is a difference in the responsiveness of Th17 cells to fructose during activation. Indeed, pTh17 cells exhibited higher responsiveness to fructose at transcriptome and metabolism levels. GLUT5, a fructose transporter, was highly expressed on pTh17 cells, which made pTh17 cells much more sensitive to fructose addition. Another mechanism could be that fructose engage in different downstream pathways in pTh17 cells and βTh17 cells. We found that fructose intake led to dramatic enhanced ROS-EGFR signal, which is essential for pTh17 cell differentiation. We therefore proposed that fructose intake dictates pTh17 cell-mediated inflammatory responses without interfering homeostatic βTh17 cells.

It is intriguing that excessive consumption of fructose increases the risk of both autoimmune diseases and tumor development^26,30,31, two seemingly opposing processes. In addition, fructose indirectly promotes anti-tumor function of CD8⁺T cell by the leptin secretion of adipocytes in lung carcinoma²⁷. Whereas, fructose directly inhibited M1-like macrophage polarization via Glut5 during another CRC model²⁶. Here, our findings suggest fructose addition may has broader and specific implication in pTh17 cell-mediated inflammatory disorders, including colitis and EAE in mice. These findings suggest that fructose’s role in immune cells is manifold and varies depending on the cell types and disease contexts. However, other immune cell subsets such as B cells, DC cells, macrophage, and Treg cells were also involved in the occurrence and development of colitis^80–82. The question remaining is whether these immune cell subsets contributed to colitis disease under high-fructose conditions. Results may suggest the opposite as fructose has no effects on the proportion of these immune cell subsets in colitis models. The specificity and importance of Th17 cells was further demonstrated by the amelioration of colitis in Rag1^-/- mice receiving Il17^−/− CD4⁺ T cells, both in the presence and absence of fructose. However, given the diversity of immune cell subsets and function in autoimmune environments, fructose may function by modulating their function or plasticity, rather than their distribution, warranting further investigation.

Our study further addressed the contribution of ROS in the dichotomy of Th17 cells. We found that pathogenic Th17 cells exhibited higher ROS levels in comparison to βTh17 cells. ROS scavenger NAC specifically reduced pTh17 cell differentiation but not βTh17 cell differentiation. Of note, recent studies have detected the role of ROS in regulating Th17 cell responses, albeit their effects remain a topic of debate^83–86. While moderate ROS is critical for Th17 cell generation, superfluous ROS levels by pre-oxidants may suppress Th17 cell differentiation. Likewise, ROS exerts differential role in the function and stability of other cells, such as Treg cells. ROS is essential for the suppressive function of Treg cells, whereas excessive ROS in Treg cells disrupts its function and integrity^84,87–90. Mechanistically, low levels of ROS act as key messengers for metabolism and immune signaling transduction in immune cells, whereas high levels of ROS are cytotoxic, causing mitochondrial dysfunction and DNA damage^91,92. These findings underscore the pivotal role of ROS in shaping T-cell immunity in a context- and dose-dependent manner, where they can exert either pro-inflammatory or regulatory effects^93,94. Considering the duo role of antioxidant activity in T cell immunity, caution is needed when using antioxidants to treat inflammatory disorders like colitis.

In summary, we elucidated the adverse effect of fructose consumption on pTh17 cell-mediated inflammatory disorders, and found that high-fructose intake specifically promoted pTh17 cell differentiation and function via a reactive oxygen species (ROS)-EGFR-STAT3 axis.

Methods

Mice

RAG1^−/−, IL-17A^−/−, CD45.1 and CD45.2 congenic mice were used on a C57BL/6 background. Rag1^−/− mice were purchased from GemPharmatech (#T004753). Il17a^−/− mice which were generated as previously described and provided by Dr. Chaojie Hu (USTC)⁹⁵. Littermates were used unless stated otherwise. Mice were maintained under specific pathogen-free (19–23 °C and 45–65% humidity) conditions with a 12-h light-dark cycle at the Medical Research Institute of Wuhan University. They were provided standard chow ad libitum and received water that was treated according to the experimental protocols. All animal experiments were performed in accordance with the protocols approved by the Animal Care and Use Committee of the Medical Research Institute, Wuhan University. The mice were euthanized by inhalation of carbon dioxide.

T cell activation, differentiation, and proliferation in vitro

CD4⁺ T cells were isolated by mouse CD4 magnetic beads (Miltenyi Biotec) manufacture’s protocols. Isolated T cells were activated with plates coated with 10 μg/ml anti-CD3 (145-2C11, BioXCell) and 10 μg/ml anti-CD28 mAb (37.51, BioXCell) and cultured in serum-free X-VIVO 20 medium (Lonza). For pathogenic-Th17 cell differentiation, 20 ng/ml IL-1β (Biolegend), 40 ng/ml IL-6 (Biolegend), 50 ng/ml IL-23 (Biolegend), 20 µg/ml anti-IL-4 (11B11, BioXcell), and 20 μg/ml anti-IFN-γ (XMG1.2, BioXcell) were added to the culture. For TGF-β1 + IL-6 induced Th17 cell differentiation, 1 ng/ml TGF-β1 (Biolegend), 40 ng/ml IL-6 (Biolegend), and 20 µg/ml anti-IFN-γ (XMG1.2, BioXcell) were added to the culture. For Th1 cell conditions, 20 ng/ml IL-12 (Biolegend) and 10 μg/ml anti-IL-4 were used. For Th2 cell conditions, 40 ng/ml IL-4 (Biolegend) and 10 μg/ml anti-IFNγ (XMG1.2, BioXcell) were used. For Treg cell differentiation, 1 ng/ml TGF-β1 (Biolegend), 5 ng/ml IL-2 (Biolegend), 20 μg/ml anti-IL-4 (11B11, BioXcell), and 20 μg/ml anti-IFN-γ (XMG1.2, BioXcell) were added to the culture. N-Acetyl Cysteine NAC (Solarbio, 616-91-1), Mitoquinone (Focus Biomolecules, 10-1363), 2-DG (sigma, D8375), Rotenone (sigma, 45656), antimycin A (GLPBIO,1397-94-0), G6PDi-1 (MCE, HY-W107464), and afatinib (MCE, HY-10261) were added when needed.

To assess proliferation, isolated CD4⁺ T Cells were labeled with 5 μM carboxyfluorescein diacetate succinimidyl ester (CFSE, AnaSpec) for 5 min at the room temperature. Labelled T cells were activated under various Th17 differentiation conditions as indicated. The T cell proliferation was assessed 72 h post-activation based on CFSE dilution by flow-cytometry.

Flow cytometry

Fluorescence-conjugated anti-CD4 (RM4-5), anti-CD8 (53-6.7), anti-CD25 (PC61), anti-CD45RB (C363-16A), anti-CD45.1 (A20), anti-CD45.2 (104), anti-IFN-γ (XMG1.2), anti-IL-17A (TC11-18H10.1), anti-GM-CSF (MP1-22E9), anti-IL-10 (JES5-16E3), anti-IL-4 (11B11), anti-Foxp3 (150D), anti-CD11c (N418), anti-CD11b (M1/70), anti-B220 (RA3-6B2), TMRM, MitoTracker^TMGreen (MTG), Annexin-V and 7-amino-actinomycin D (7-AAD) from BioLegend were used. anti-RORγT (Q31-378) was from BD Biosciences.

For surface staining, cells were stained with antibody solutions for 15 min at 4 °C in the dark. For intracellular staining, cells were stimulated with ionomycin (1 μM), phorbol 12-myristate 13-acetate (50 ng/ml), and Brefeldin A (5 μg/ml) at 37 °C for 4 h. For apoptosis assays, cells were stained with Annexin-V (BioLegend) and 7-AAD (BioLegend) in dilution buffer for 30 min at room temperature in the dark.

For cytosolic ROS staining, cells were stained with 2 μM DCFH-DA (Sigma, D6883) in prewarmed DMEM for 30 min at 37 °C. For mtROS assay cells were stained with 2.5 μM MitoSox Red (Thermo, M36008) in prewarmed DMEM medium for 1 h at 37 °C. For mitochondrial content and membrane potential, cells were stained with TMRM (25 nM) and MTG (50 nM) in prewarmed DMEM for 30 min at 37 °C in the dark.

Stained cells were recorded on a BD FACSCelesta^TM Cell Analyzer (BD Biosciences) or CytoFLEX LX Flow Cytometer (Beckman) and analyzed with FlowJo software (Tree Star). The gate strategy is provided in Supplementary Fig. 11.

Quantitative RT-PCR (qRT-PCR) analysis

For qRT-PCR analysis, total RNA from T cells was extracted using Trizol reagent (TakaRa) and reverse-transcribed into cDNA with the TRUEscript RT Master Mix kit (Aidlab) according to the manufacturer’s instructions. Quantitative PCR (qPCR) was performed on the CFX384 Touch real-time PCR Detection System (BioRad). Relative mRNA amounts were then analyzed using the comparative ΔΔCt method with housekeeping gene β-actin as the internal control.

RNA-seq analysis

Total RNA was extracted from cultured T cells using the Direct-zol RNA Microprep Kit (ZYMO Research). RNA-seq libraries were prepared by the TruSeq RNA Sample Prep Kit (Illumina, San Diego, CA) with 1 µg of RNA as input, followed by poly(A) enrichment. Sequencing was performed in 150-bp paired-end mode on a NovaSeq 6000 platform (Novogene) according to the manufacturer’s protocol.

Total RNA was extracted from cultured T cells using the Direct-zol RNA Microprep Kit (ZYMO Research). RNA integrity was assessed using RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Total RNA was used as input material for the RNA sample preparations.

Raw data (raw reads) of fastq format were processed through fastp software. Reference genome and gene model annotation files were downloaded from genome website directly. Index of the reference genome was built using Hisat2 v2.0.5 and paired-end clean reads were aligned to the reference genome using Hisat2 v2.0.5. PCA was performed to assess variance between sample groups using FactoMineR (v2.1.1) and factoextra (v1.0.7) in R (v4.3.3). Differentially expressed genes (DEGs) were identified based on a log2 fold change ≥ 0.5 when the p value was less than 0.05. Visualization of DEGs was conducted using heatmap (v1.0.12) and ggplot2 (v3.5.1).

Functional enrichment analysis of selected gene sets was performed using the Broad Institute’s GSEA tool (v4.3.3). KEGG pathway enrichment analysis was conducted using the clusterProfiler package (v4.10.1) in R (v4.3.3). RNA-seq data are deposited in GEO database under ID code: GSE285450.

Metabolomics assay

Th17 cell lysate was prepared by 80% methanol suspension and sonication. LC-MS/MS analyses were performed using an ExionLC™ AD system (SCIEX) coupled with a QTRAP® 6500+ mass spectrometer (SCIEX) in Novogene Co., Ltd. (Beijing, China). Samples were injected onto a Xselect HSS T3 (2.1 × 150 mm, 2.5 μm) using a 20-min linear gradient at a flow rate of 0.4 mL/min for the positive/negative polarity mode. The detection of the experimental samples using Multiple Reaction Monitoring were based on Novogene in-house database. Metabolites were annotated using the KEGG database (http://www.genome.jp/kegg/), HMDB database (http://www.hmdb.ca/), and LipidMaps database (http://www.lipidmaps.org/). The data files generated by HPLC-MS/MS were processed using the SCIEX OS Version 1.4. The main parameters were set as follows: minimum peak height, 500; signal/noise ratio, 5; Gaussian smooth width, 1. The area of each peak represents the relative content of the corresponding substance. metaX software was used for metabolomics data processing. Differential metabolites were identified based on the following criteria: VIP > 1, p value < 0.05, and Fold Change (FC) ≥ 1 or ≤ −1.

ShRNA-mediated gene knock down

For shRNA-mediated gene knockout in T cells, lentiviral shRNA vectors carrying puromycin resistance genes were constructed. Lentiviruses were prepared by transfecting HEK293T cells. The lentiviral shRNA constructs, psPAX2, and pMD2.G were transfected at a 3:2:1 ratio using calcium chloride (CaCl₂) method. After 48 h, viruses were harvested, filtered through a 0.45 μm syringe filter, and stored in a −80 °C freezer. Twenty-four hours after activation, T cells were spin-inoculated at 1500 × g with the indicated recombinant virus in the presence of 8 μg/ml polybrene (Sigma-Aldrich) and 10 mM HEPES buffer (Gibco) at 30 °C for 90 min. Antibiotic selection was performed, when applicable, by adding 2 μg/ml puromycin to the culture medium for 4 days. The transduced cells were identified based on puromycin resistance. Target sequence for Glut is: CCCAATCGTTTGAGCTAATAA.

Immuno-blotting analysis and immunoprecipitation

For immunoblotting, proteins were extracted using RIPA lysis buffer (Sigma) and resolved by home-made SDS-page gels. The proteins were then transferred to a polyvinylidene fluoride membrane (Millipore) and blocked with 5% skimmed milk at room temperature for 1 h. The membrane was analyzed by immunoblotting with the following antibodies: anti-EGFR (CST, 4267T), anti-p-EGFR (CST, 3777T), anti-STAT3 (Proteintech, 10253-2-AP), anti-p-STAT3 (Proteintech, 28945-1-AP), anti-GLUT5 (Abclonal, A13650), anti-GAPDH (Abclonal, A19056), Histone H3 (Cat: A2348, Abclonal), and anti-β-actin antibodies (AC026, Abclonal).

For immunoprecipitation (IP), cells were lysed in IP lysis buffer (10 mM HEPES, pH 7.5, 1.5 mM MgCl₂, 0.2 mM EDTA, and 150 mM NaCl, containing 1% NP40) and sonicated using a Bioruptor PICO. Then, cell lysates were incubated with 50 μl of Agarose beads (Cytiva) that had been conjugated with indicated antibody. The beads were washed three times with lysis wash buffer after overnight incubation at 4 °C. The associated proteins were eluted with 2xLaemmli sample buffer (Bio-Rad) and incubated at 95 °C for 5 min. The eluted samples were separated by SDS-PAGE and analyzed by immunoblotting. Antibody dilutions are provided as Supplementary Data 2.

ChIP-qPCR assay

Cells were harvested and cross-linked with 1% formaldehyde at room temperature. Then, the reaction was terminated by adding glycine. The cells were washed with ice-cold PBS. The precipitate was successively lysed in cell lysis buffer-A [10 mM Tris [pH 8.0], 10 mM EDTA (pH 8.0), 0.5 mM EGTA (pH 8.0), 0.25% Triton X-100, 0.5% NP-40, and 1× protease inhibitor)], cell lysis buffer-B [10 mM Tris (pH 8.0), 1 mM EDTA (pH 8.0), 0.5 mM EGTA (pH 8.0), 0.2 M NaCl, and 1× protease inhibitors], and sonication buffer [10 mM Tris pH 8.0, 1 mM EDTA pH 8.0, 0.5 mM EGTA pH 8.0, 1× protease inhibitors, and SDS], and sonicated with Bioruptor PICO. Then, cell lysates were incubated with 50 μl of Agarose beads (Cytiva) pre-bound with antibodies. After overnight incubation, the beads were washed with low salt immune complex wash buffer, high salt immune complex wash buffer, LiCl immune complex wash buffer, and 1 × TE. Subsequently, the beads were incubated with elution buffer, and de-crosslinked at 65 °C for 4 h, followed by treatment with RNase, proteinase K, EDTA, and Tris (pH 6.5). The DNA obtained was purified using a commercial kit (Nanjing Better Biotechnology Co., Ltd.) and the enrichment was detected by q-PCR. The primer sequences of CNS6 promoter are as follows: F: GCCCACCAATCTGTCCCCACTA; R: TTCTGACTCGCTTTCCTCCCAG. The primer sequences of CNS9 promoter are as follows: F: GCACAGGACCAGAACAAGCAGG; R: AGCACAAGAAACGGGG AGACTGA.

T cell-transfer colitis

CD4⁺CD8⁻CD25⁻CD45RB^hi T cells, which were sorted from WT mice or Il17^−/− mice, and 4 × 10⁵ naïve T cells were intravenously transferred into Rag1^−/− mice. The mice were treated with normal water or water containing 10% fructose. Then, the development of colitis was monitored for every two or three days. When specified, the mice were also given 2 mg/ml of NAC in their drinking water.

DSS-induced colitis

To induce acute colitis, mice were treated with 3% dextran sodium sulfate (DSS) containing drinking water for 7 days. Body weight and disease severity was measured daily throughout the experiment. At the end of the experiment, the mice were euthanized, and the colon was removed for measurement of length and weight. Additionally, histological and immunological analyses were performed on colon and mLN tissues. For NAC treatment, the mice were given 2 mg/ml of NAC in their drinking water. For antibiotic treatment, mice were treated with antibiotic (Vancomycin, Sangon, 1 g/L; Metronidazole, Sangon, 1 g/L) containing drinking water for across the study.

Lamina propria leukocytes isolation

Colon was mechanically dissected and flushed with ice-cold PBS. The tissue was cut into pieces and incubated in the presence of 1 mM DTT and 5 mM EDTA at 37 °C for 30 min. The digested pieces were passed through a 100 μm cell strainer. The cell suspension was discarded. The remaining pieces were cut into 1 mm pieces and further digested in RPMI 1640 medium with collagenase D (1 mg/ml collagenase D, Sangon), DNase I (100 μg/ml, Sigma), Liberase TL (0.2 mg/ml, Roche), and 10% FBS at 37 °C for 1 h. The LPL were isolated by centrifuging at 2000 rpm for 20 min with 40% and 80% discontinuous Percoll density gradient (Biosharp). Isolated lymphocytes were then subjected to subsequent analysis.

EAE induction in mice

For EAE induction, 50 μg murine myelin oligodendrocyte glycoprotein (MOG) peptide 35–55 (AnaSpec) was emulsified in complete Freund’s adjuvant (CF A) that consists incomplete Freund’s adjuvant (Difco Laboratories) and 5 mg/ml of Mycobacterium tuberculosis H37RA (Difco Laboratories). MOG/CFA was injected sub-cutaneous (s.c.) on day 0. 200 ng Pertussis toxin (List Biological) was injected intra-peritoneal (i.p.) on day 0 and day 2. The severity of EAE was monitored and graded on a clinical score of 0 to 5: 0 = no clinical signs; 1 = Limp tail; 2 = Para-paresis (weakness, incomplete paralysis of one or two hind limbs); 3 = Paraplegia (complete paralysis of two hind limbs); 4 = Paraplegia with forelimb weakness or paralysis; 5 = Moribund or death.

For passive transfer induced EAE, CD45.1.1 + CD4 + T cells were isolated from spleen, activated and differentiated under pTh17 cell-polarizing condition for 4 days with or without fructose, then re-stimulated in 24-well plates pre-coated with 2 μg/ml anti-CD3 and 2 μg/ml anti-CD28 in fresh medium for 2 days. 1.5 × 10⁶ live pTh17 cells were then purified, suspended in 200 μL PBS and injected intravenously (i.v.) into recipient CD45.2.2+ mice that had received 450 cGy irradiation. 200 ng pertussis toxin (List Biological) in 200 μl PBS was injected intraperitoneally (i.p.) at day 0 and day 2 after cell transfer.

After EAE induction, mice were sacrificed and perfused with ice-cold phosphate-buffered saline containing 20 U/ml heparin. After excision of all tissues, the spinal cord was separated, dissected, and digested with 1 mg/ml collagenase D (Sigma) for 60 min at 37 °C in a water bath, with shaking every 15 min. The digested tissue was then centrifuged at 2000 rpm for 20 min in 38% Percoll (Sigma) to isolate mononuclear cells. Residual red blood cells were lysed with ACK lysis buffer. The lymphocytes were then washed with PBS and subjected to subsequent analysis.

The severity of EAE was monitored by clinical scoring, ranging from 0 to 5: 0 = no clinical symptoms; 1 = tail paralysis; 2 = partial paralysis of hind limbs; 3 = complete paralysis of hind limbs; 4 = hind limb paralysis + 75% body paralysis; 5 = moribund.

Fructose uptake test

After chemotaxis of T cells in vitro for 48 h 50 mM fructose was added for 6 h. The cells were centrifuged, collected, washed, resuspended in extraction solution, and ultrasonically treated. Then, the fructose intake in T cells was analyzed using a fructose content kit (PYRAM) according to the manual instruction.

Measurement of the cell metabolism

GlycoPER and OCR measurements were performed using an XF24 Extracellular Flux analyzer (Agilent Technologies). Polarized Th17 cells were collected and resuspended in XF assay medium (1 mM pyruvate, 2 mM glutamine, and 10 mM glucose), and inoculated into the Seahorse XFe24 cell culture plate (Agilent Technologies; 101085-004) at 5 × 10⁵ cells per well. OCR was determined using the Cell Mito Stress Test Kit (Agilent Technologies, 103015-100). GlycoPER rates were measured with the Glycolytic Rate Assay Kit (Agilent Technologies). Data analysis and interpretation were performed using WAVE 2.6 software according to the manufacturer’s manual (Agilent Technologies).

Detection of lactic acid

After chemotaxis of pTh17 cells in vitro with or without fructose for 4 days, the supernatant from the cell culture was collected after centrifugation and lactic acid content was determined using a lactic acid test kit (Nanjing Jiancheng Bioengineering Institute, A020-2-2) according to the manufacturer’s instructions.

Statistical analysis

A Student’s t test or two-way ANOVA was used for two-group comparison. p values of <0.05 were considered significant. In the figures, *, **, and *** were used to indicate p < 0.05, p < 0.01, and p < 0.001, respectively. All results shown are mean ± s.d. unless stated otherwise.

Reporting summary

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

Author contributions

X.Y.L. and W.H.H. contributed equally to the design and implementation of the cellular, molecular, biochemical, and animal experiments and the writing of the manuscript; J.S. contributed to the bioinformatic analysis; B.W. conceived the project, designed the experiments, analyzed the data, and wrote the manuscript.

Peer review

Peer review information

Nature Communications thanks Bernd Lepenies and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

RNA-seq data are deposited in GEO database under ID code: GSE285450 with link: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE285450. Metabolomics data are provided as Supplementary Data 1. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-66064-5.