Metabolic reprogramming of Th17/Treg imbalance in autoimmune thyroid diseases

Abstract

Autoimmune thyroid diseases (AITDs), including Hashimoto’s thyroiditis and Graves’ disease, arise from thyroid-specific autoimmunity driven by a breakdown of immune tolerance and dysregulated T-cell responses. Within this immune network, imbalance between T helper 17 (Th17) cells and regulatory T (Treg) cells has emerged as a major determinant of persistent inflammation and defective immune restraint. These two subsets are supported by distinct but interconnected metabolic programs. Th17 cells preferentially engage glycolytic and anabolic pathways to sustain inflammatory activity, whereas Treg cells rely more strongly on oxidative metabolism and mitochondrial fitness to preserve lineage stability and suppressive function. In AITDs, these intracellular programs are further reshaped by disease-associated microenvironmental cues, including excess iodine, oxidative stress, lactate accumulation, inflammatory cytokines, and tissue-derived stromal signals. This review summarizes how glucose, lipid, mitochondrial, and amino acid metabolism collectively regulate Th17 and Treg differentiation and function. We further examine how these pathways are altered in AITDs and distorted in thyroid and orbital tissues to amplify immune disequilibrium. Finally, we discuss emerging therapeutic strategies aimed at targeting immune metabolic circuits to restore immune homeostasis.

Introduction

Autoimmune thyroid diseases (AITDs) represent the most prevalent group of organ-specific autoimmune disorders, mainly comprising Hashimoto’s thyroiditis (HT) and Graves’ disease (GD). HT is primarily characterized clinically by hypothyroidism, whereas GD typically presents with hyperthyroidism and thyroid-associated ophthalmopathy [1]. Pathologically, the former is dominated by autoimmune-mediated thyroid destruction, while the latter is primarily driven by autoimmune stimulation leading to thyroid hyperfunction [2] Epidemiological data indicate that AITDs affect approximately 5%–10% of the general population, with a striking female predominance and a with a steadily increasing global incidence [1]. Genetic susceptibility and environmental triggers, such as excessive iodine intake, infections, psychological stress, and xenobiotic exposure, converge to initiate disease [3].However, the fundamental pathogenic driver is the breakdown of immune tolerance toward thyroid-specific autoantigens, including thyroid peroxidase (TPO), thyroglobulin (TG), and the thyroid-stimulating hormone receptor (TSHR), leading to persistent autoreactive T- and B-cell response [4].

The immunopathogenesis of AITDs involves a coordinated yet dysregulated interplay among multiple CD4⁺ T-cell subsets, cytotoxic T cells, B cells, and innate immune cells [5]. Traditionally, HT is characterized as a Th1-dominant disorder. Elevated IFN-γ, TNF-α, and IL-2 promote macrophage activation, upregulate MHC class II molecules on thyroid follicular cells, and trigger CD8⁺ T-cell–mediated apoptosis, culminating in progressive glandular destruction and hypothyroidism [1]. In contrast, GD exhibits a more dynamic immune profile. Active disease phases are associated with Th1-skewed inflammation, whereas remission stages often demonstrate relative Th2 predominance. Th2 cytokines, including IL-4, IL-5, and IL-13, enhance B-cell activation and immunoglobulin class switching, facilitating the production of thyroid-stimulating antibodies (TRAb), which drive hyperthyroidism [6].

Beyond the classical Th1/Th2 paradigm, additional CD4⁺ T-cell subsets play critical roles in AITDs. Follicular helper T (Tfh) cells, characterized by CXCR5, PD-1, and Bcl-6 expression, secrete IL-21 and promote germinal center formation [7]. Increased circulating and thyroid-infiltrating Tfh cells correlate positively with TRAb titers, sustaining pathogenic B-cell responses [8]. Notably, IL-21 derived from Tfh cells not only facilitates B cell differentiation but also sustains Tfh expansion through an autocrine loop, establishing a pathogenic positive feedback circuit [9]. Similarly, the balance between T helper 17 (Th17) cells and regulatory T (Treg) cells constitutes an essential immunological checkpoint governing the equilibrium between inflammation and tolerance. Th17 cells differentiate in response to IL-6, IL-1β, and IL-23, express RORγt, and produce IL-17 A, IL-21, and IL-22, thereby amplifying inflammatory cascades and promoting tissue remodeling [10]. Differentiating in response to TGF-β and IL-2, Treg cells are characterized by stable Foxp3 expression and high levels of CD25 and CTLA-4, and suppress excessive immune activation through IL-10, TGF-β, and inhibition of antigen-presenting cell activity [11]. A growing body of evidence demonstrates that AITDs exhibit a significant expansion of Th17 cells in peripheral blood and thyroid tissue, accompanied by elevated levels of IL-17 A, IL-21, IL-22, and other pro-inflammatory cytokines, which promote the recruitment of neutrophils and macrophages. In contrast, the proportion of Treg cells is reduced, leading to impaired immunosuppressive function. These alterations result in excessive activation of autoreactive effector T cells and sustained production of TPO, TG, and TSHR antibodies, thereby driving chronic inflammatory processes and contributing collectively to follicular destruction and fibrotic remodeling. This imbalance between Th17 and Treg cells is particularly pronounced in Graves’ orbitopathy (GO). Consequently, the Th17/Treg ratio correlates positively with the titres of disease-related antibodies and the severity of orbital inflammation in AITDs [12, 13]. Such dysregulated Th17/Treg dynamics are not unique to AITDs but represent a common immunological signature shared across multiple autoimmune diseases, including systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) [14]. The balance between these two subsets therefore acts as a critical checkpoint that determines the balance between inflammation and tolerance in autoimmune conditions.

Recent immunometabolic research has revealed that T cell fate determination relies not only on cytokine and transcription factor regulation, but also on modulation by TCR activation signals, co-stimulatory or inhibitory molecules and the metabolic state of the local microenvironment. Dynamic metabolic reprogramming provides T cells with energy and biosynthetic precursors, further determining their activation levels, differentiation pathways, and effector functions, a process termed metabolic reprogramming [15]. Upon antigen stimulation, naïve T cells transition from a metabolically quiescent state primarily dependent on oxidative phosphorylation (OXPHOS) to a highly active aerobic glycolytic phenotype. This shift provides rapid ATP generation and biosynthetic precursors to support the intense proliferative and functional demands of activated T cells [16]. Distinct T cell subsets display unique metabolic signatures. Th17 cells rely heavily on glycolysis and fatty-acid synthesis (FAS) to sustain their pro-inflammatory phenotype, while Treg cells predominantly depend on OXPHOS and fatty-acid oxidation (FAO) to preserve their suppressive function and lineage stability [17]. In AITDs, excessive iodine, reactive oxygen species accumulation, lactate enrichment, and chronic inflammatory signaling reshape the local metabolic landscape and influence key nodes including mTOR, AMPK, and HIF-1α, thereby skewing Th17/Treg differentiation [2, 18]. Thus, immunometabolic remodeling provides a mechanistic bridge linking thyroid-specific environmental stress to systemic immune imbalance.

While the metabolic programming of Th17 and Treg cells has been extensively studied in cancer and systemic autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus, evidence specifically addressing these processes in AITDs remains fragmented. Therefore, this narrative review provides a focused synthesis to distinguish general immunometabolic principles from thyroid-specific regulatory mechanisms. Furthermore, we explore how thyroid-specific environmental factors influence T-cell metabolism, thereby driving immune imbalance. Finally, we discuss emerging evidence supporting metabolic-targeted therapeutic strategies in AITDs. This integrated perspective not only deepens insight into disease pathogenesis but also provides a rational basis for developing metabolism-oriented immunotherapies, ultimately contributing to precision medicine approaches for AITDs.

Metabolic characteristics and regulatory mechanisms of Treg and Th17 Cells

Metabolic reprogramming has emerged as a central mechanism governing the differentiation and functional balance of Th17 and Treg cells. The mechanistic evidence in this section comes mainly from systemic autoimmune models such as RA, and SLE. We also draw selectively on findings from tumor microenvironments, where the metabolic logic can either diverge or align with what is seen in autoimmunity. This contrast will prove especially relevant when we turn to AITDs. Rather than serving as passive adaptations to activation, metabolic pathways actively shape lineage commitment, effector function, and cellular stability. In this context, Th17 and Treg cells adopt distinct yet dynamically regulated metabolic strategies that support their opposing immunological roles. These metabolic programs are not isolated processes but are organized into interconnected networks involving glucose utilization, lipid metabolism, mitochondrial function, and amino acid handling. Through coordinated regulation of energy production, biosynthetic pathways, and signaling intermediates, these metabolic layers collectively determine the balance between inflammatory and regulatory responses. Understanding how these pathways operate in parallel provides a mechanistic framework for interpreting immune dysregulation in autoimmune settings.

Glucose metabolism

Glucose metabolism is a major dynamically regulated axis that distinguishes Th17 and Treg programs and helps determine lineage stability and inflammatory output. In Th17 cells, activation- and cytokine-derived cues converge on mechanistic target of rapamycin complex 1 (mTORC1), which establishes a glycolysis-permissive transcriptional state by inducing c-Myc and hypoxia-inducible factor 1 alpha (HIF-1α) [2, 18]. Consequently, Th17 differentiation is commonly accompanied by a marked reliance on aerobic glycolysis to support rapid expansion and proinflammatory effector function. Early glycolytic commitment is initiated by c-Myc, which upregulates the key glucose transporter glucose transporter 1 (GLUT1) and core glycolytic enzymes such as hexokinase 2 (HK2), thereby increasing glucose uptake and generating the energy and biosynthetic intermediates needed for effector differentiation [19]. As the program progresses, HIF-1α further amplifies glycolytic gene expression and functionally biases fate decisions by reinforcing RORγt-associated transcription while antagonizing Foxp3 stability [20, 21].

Notably, glycolytic reliance is not uniform across Th17 states. IL-23-driven pathogenic Th17 cells exhibit heightened glycolytic dependency and preferentially upregulate the high-affinity transporter glucose transporter 3 (GLUT3), facilitating increased glucose acquisition and reinforcing pathogenic effector programming [22]. GLUT3 has been proposed to engage a glycolytic and epigenetic circuit that shapes inflammatory cytokine gene expression, suggesting that glucose handling can influence effector identity beyond bioenergetic supply. Within this pathway, lactate dehydrogenase A (LDHA), a terminal glycolytic enzyme that converts pyruvate to lactate while regenerating NAD+, supports continued glycolytic flux and helps sustain phosphoinositide 3-kinase (PI3K)-Akt signaling in pathogenic Th17 cells [23]. Other glycolytic nodes also exert signaling effects. Pyruvate kinase M2 (PKM2), which catalyzes the final rate-limiting step from phosphoenolpyruvate (PEP) to pyruvate, can translocate to the nucleus and enhance signal transducer and activator of transcription 3 (STAT3)-dependent transcriptional activity, thereby strengthening RORγt-driven gene expression [24]. Because PEP also functions as a signaling intermediate that can fine-tune Th17 transcriptional output, glycolysis should be viewed as a step-specific regulatory pathway rather than a process defined only by total flux [25]. In addition, phosphoglycerate mutase (PGAM), another glycolytic enzyme, has been linked to constraining pathogenic Th17 programming [26].

In contrast, Treg cells preserve suppressive stability by restraining glycolytic intensity rather than eliminating glucose utilization [27]. Strong mTORC1 activity favors Th17 glycolytic commitment, whereas excessive mTORC1 signaling can destabilize Treg lineage maintenance, underscoring the need for balanced mTOR activity in Treg metabolic homeostasis [28]. Foxp3 directly and indirectly represses c-Myc and GLUT1 expression, thereby preventing sustained high glycolytic amplification that would erode lineage stability [29]. This restraint is stage- and context-dependent. A limited and transient increase in glycolysis may support Treg proliferation and migration, whereas prolonged or excessive glycolytic drive more often compromises long-term suppressive function [30, 31]. Collectively, pathogenic Th17 cells rely heavily on glycolysis, whereas functional Tregs are characterized by restrained glycolytic activity. This contrasting metabolic pattern has been consistently observed in experimental autoimmune encephalomyelitis (EAE) and RA models and provides a useful mechanistic framework for interpreting glucose metabolic alterations in organ-specific autoimmune diseases including AITDs [32].

Lactate, the end product of glycolysis, acts as a prominent microenvironmental metabolite signal that can consolidate Th17-Treg divergence. In inflammatory settings, lactate can stabilize HIF-1α and reinforce Th17 polarization and IL-17 production, whereas lactate-associated acidification can in some contexts destabilize Foxp3 and weaken Treg suppressive capacity [33]. However, Treg responses to lactate are context dependent and shaped by metabolite availability, extracellular pH, and lactate transport capacity [34, 35]. In parallel, Treg stability has been linked to inhibitory receptor signaling and autophagy-mediated restraint of excessive glycolytic activity, while lactate uptake through monocarboxylate transporter 1 (MCT1) may sustain activated Treg function in low-glucose milieus [34, 36].

Lipid metabolism

Th17 and Treg cells diverge in lipid metabolism by prioritizing anabolic lipid synthesis versus oxidative utilization and lipid homeostasis, which in turn shapes lineage stabilization, effector output, and persistence under inflammatory stress.

Th17 cells preferentially initiate de novo fatty acid synthesis (FAS) because rapid expansion and sustained IL-17 A-centered output require continuous membrane lipid supply [37]. This anabolic commitment is governed by acetyl-CoA carboxylase 1 (ACC1), which converts acetyl-CoA to malonyl-CoA and thereby enables downstream lipogenesis. Coordinated activity of fatty acid synthase (FASN) and stearoyl-CoA desaturase 1 (SCD1) expands the repertoire of membrane lipids, sustaining phospholipid availability for membrane growth and the organization of lipid rafts that facilitate robust T-cell receptor (TCR) signaling and help maintain RORγt-dependent Th17 effector function [38]. When this anabolic mode predominates, malonyl-CoA can limit carnitine palmitoyltransferase 1 (CPT1)-dependent mitochondrial fatty acid entry, keeping lipid handling aligned with biosynthetic needs during effector formation [39]. Notably, Th17 lipid utilization is not uniformly anabolic, and fatty acid oxidation (FAO) can be engaged in specific inflammatory niches or at later stages of differentiation, where lipid catabolism modulates cytokine output in a context-dependent manner [40]. A predictable consequence of enhanced desaturation and polyunsaturated fatty acid (PUFA) enrichment is increased susceptibility to oxidative lipid stress. Th17 cells counterbalance this liability through diacylglycerol O-acyltransferase 1 (DGAT1)-linked triglyceride synthesis and lipid droplet formation, which sequester peroxidizable lipids and restrain excessive lipid peroxidation, thereby supporting survival in autoimmune microenvironments [41]. In the same membrane-centered framework, de novo sphingolipid biosynthesis is required for Th17 metabolic rewiring and differentiation, whereas cholesterol precursors and selected oxysterols further reinforce Th17 identity by enhancing RORγt activity [42, 43].

In contrast, Treg cells are often described as relying more heavily on FAO under quiescent or nutrient-limited conditions, with CPT1A-mediated fatty-acid entry frequently linked to Foxp3 stability and suppressive function alongside AMPK- and PPAR-associated inputs [37]. Whether FAO is universally required across contexts remains debated, as pharmacologic and genetic perturbations have not yielded fully concordant dependence across models [44]. In tissue-conditioned environments such as tumors or chronic inflammation, Tregs commonly integrate exogenous lipid acquisition, including CD36-associated uptake, with sterol and lipid homeostasis control, and cholesterol handling with SCAP–SREBP-associated regulation has been implicated in sustaining suppressive fitness under nutrient competition [45, 46].

Lipid-derived metabolites provide an additional regulatory layer. Bile acid derivatives such as 3-oxoLCA and isoalloLCA can counteract RORγt activity and support Foxp3-associated regulation, thereby favoring a regulatory direction. β-hydroxybutyrate can also modestly bias toward regulatory phenotypes through histone deacetylase inhibition [47, 48].

Mitochondrial metabolism

Mitochondrial metabolism supports T-cell lineage stability and functional persistence through mitochondrial respiration, OXPHOS, tricarboxylic acid cycle activity, and redox control. In the Th17-Treg axis, these processes actively shape lineage stability and functional persistence.

During the long-term activation process of Treg cells, their stability mainly depends on the oxidative metabolic process of mitochondria to maintain the stability of Foxp3 and the persistent inhibitory function [29]. By contrast, mitochondrial involvement in Th17 cells is more state dependent, with pathogenic Th17 populations showing stronger reliance on sustained mitochondrial support than homeostatic or barrier-associated Th17 cells [49, 50]. Accordingly, disruption of mitochondrial function often weakens Th17 inflammatory fitness while relatively favoring Treg maintenance, indicating that mitochondrial metabolism can shift the Th17-Treg balance [51, 52].

The tricarboxylic acid cycle provides an important signaling layer because its intermediates connect mitochondrial activity to epigenetic regulation and transcriptional control. In Treg cells, relatively complete cycle flux coupled to OXPHOS helps maintain the supply of alpha-ketoglutarate, a tricarboxylic acid cycle intermediate repeatedly linked to Foxp3 stability [53–55]. By contrast, accumulation of succinate and fumarate can oppose alpha-ketoglutarate-dependent dioxygenase activity, alter DNA and histone demethylation, and weaken the chromatin support required for stable regulatory programs under inflammatory stress [56]. In Th17 cells, changes in intermediate balance are more often discussed as a mechanism that adjusts effector magnitude and persistence during sustained activation.

Mitochondrial structural control also contributes to lineage regulation because fusion and fission affect respiratory efficiency and metabolic adaptation. Treg cells tend to maintain fusion-associated mitochondrial architecture supported by optic atrophy 1 (OPA1) and mitofusin 1 or mitofusin 2 (MFN1 or MFN2), which preserves cristae integrity and supports efficient OXPHOS [57–59]. By contrast, pathogenic Th17 and other activated effector states are more often associated with increased dynamin-related protein 1 (DRP1)-dependent fission, which facilitates rapid metabolic reprogramming [60, 61]. Consistent with this view, perturbation of fusion or fission regulators can shift the Th17-Treg balance.

Redox signaling is tightly coupled to respiratory chain status and tricarboxylic acid cycle rewiring. Succinate-associated mitochondrial reactive oxygen species (mROS) can stabilize HIF-1α signaling and thereby support inflammatory transcription in Th17 cells [62]. In this setting, a moderate increase in mROS promotes IL-17-associated effector function, whereas inadequate antioxidant buffering in Treg cells is more likely to erode Foxp3 stability and suppressive capacity [62, 63]. When mitochondrial stress exceeds cellular control, release of mitochondrial DNA can further amplify inflammatory signaling through the cyclic GMP-AMP synthase and stimulator of interferon genes (cGAS-STING) pathway, thereby pushing the balance away from regulation [64].

These mechanistic insights come from different experimental settings. Examples include inflammatory bowel disease models, tumor microenvironments, and in vitro T cell polarization systems [62, 63, 65]. Several core features are consistently observed across autoimmune conditions, namely succinate-driven mROS accumulation, OPA1-dependent maintenance of Treg fitness, and cGAS‑STING‑mediated inflammatory amplification. These features help interpret mitochondrial dysregulation in thyroid-specific autoimmunity.

Amino acid metabolism

Beyond serving as biosynthetic precursors, amino acids tune the Th17–Treg balance by modulating nutrient sensing, mitochondrial bioenergetics, redox buffering, and the epigenetic landscape [36]. This control is organized through three linked layers, transport-defined substrate availability, lineage-biased catabolic routing, and the generation of signaling metabolites that stabilize lineage-specific transcriptional programs.

Early lineage bias is dictated by differential amino acid uptake and the activation of starvation-sensing checkpoints. SLC7A5-mediated import of large neutral amino acids is essential for robust Th17 differentiation, as it facilitates the mTORC1 activation required for Th17 effector programming and inflammatory output. In contrast, Foxp3 induction during iTreg differentiation remains relatively resilient to restricted amino acid import under TGF-beta and IL-2 signaling [66, 67]. This gatekeeping effect in Th17 cells is reinforced by coordinated transporter activity, where ASCT2 (SLC1A5) sustains intracellular glutamine pools and, together with LAT1-dependent leucine import, supports effector expansion [68, 69]. When these pools are compromised, the GCN2–eIF2α–ATF4 axis functions as an asymmetric sensor that attenuates Th17 inflammatory output while largely preserving regulatory potential [70, 71].

Following uptake, lineage-biased routing of specific amino acids provides metabolic and signaling support for Th17 pathogenicity. Glutaminase 1 (GLS1)-driven glutaminolysis supplies carbon and nitrogen for clonal expansion and supports downstream biosynthetic demand [72]. It can also contribute to alpha-ketoglutarate generation, thereby linking amino acid catabolism to downstream biosynthetic and signaling support for inflammatory T-cell function [53]. Similarly, branched-chain amino acid catabolism through branched-chain amino acid transaminase 1 (BCAT1) reinforces Th17 inflammatory competence by potentiating HIF-1α-associated effector circuits [66].

The stability of the Th17–Treg axis is further supported by amino-acid-driven redox control and nitrogen-related signaling. Cysteine is indispensable for glutathione (GSH) synthesis, enabling Treg cells to safeguard Foxp3 stability against the destabilizing effects of oxidative stress [73]. Furthermore, the coupling of methionine and cysteine through the transsulfuration pathway generates hydrogen sulfide (H₂S), which promotes Treg stability via sulfhydration-mediated regulation of epigenetic machinery upstream of Foxp3 [74, 75]. Arginine availability further influences this balance, not only by supporting proliferative expansion but also through diversion into the polyamine axis, restricting polyamine flux has been consistently shown to suppress Th17 programs while favoring acquisition of regulatory features [76].

Finally, the serine-glycine one-carbon pathway and the methionine cycle connect amino acid flux to the epigenetic regulation of T-cell fate. In inflammatory Th17 states, methylenetetrahydrofolate dehydrogenase 2 (MTHFD2) is upregulated and supports in vivo persistence, whereas its inhibition alleviates autoimmune inflammation with comparatively limited effects on Treg programs in reported models [77, 78]. Downstream, the methionine cycle supplies S-adenosylmethionine (SAM) for DNA, histone, and RNA m6A methylation. SAM-dependent m6A writers such as methyltransferase-like 3 (METTL3) fine-tune the Th17-Treg balance by regulating the stability of transcripts encoding signaling constraints such as suppressor of cytokine signaling (SOCS) family members [79]. Completing this control layer, tryptophan catabolism engages the aryl hydrocarbon receptor (AhR), through which kynurenine metabolites and microbiota-derived indoles modulate the plasticity and stability of the Th17-Treg equilibrium according to cytokine context [80].

Integrated signaling axes governing the dynamic metabolic balance between Th17 and Treg cells

Rather than operating independently, the metabolic programs described above are coordinated through a limited set of integrative signaling nodes. Their functional significance lies not in any single pathway alone, but in how they redistribute carbon flux, lipid utilization, mitochondrial activity, and amino acid-derived inputs toward either inflammatory expansion or regulatory stability. The Th17–Treg axis is therefore shaped by an integrated metabolic network in which anabolic demand, oxidative support, and stress adaptation are continuously rebalanced.

A central component of this network is the mTOR–AMPK rheostat, which determines whether metabolism is directed toward anabolic expansion or oxidative restraint [32]. Under nutrient-replete conditions, mTORC1 combines activation signals and amino acid sufficiency to sustained glycolytic engagement, promoting the entry of glucose-derived carbon into the tricarboxylic acid cycle and supporting citrate export for acetyl-CoA generation [81]. This routing links carbohydrate utilization to ACC1-dependent fatty acid synthesis, thereby favoring membrane biogenesis and effector expansion [82]. At the same time, increased malonyl-CoA restrains CPT1-dependent fatty acid entry into mitochondria, limiting fatty acid oxidation as a major bioenergetic route [83]. In this way, glucose metabolism, lipid synthesis, and mitochondrial substrate selection are aligned with the anabolic requirements of Th17 cells, supporting proliferative growth, inflammatory output, and lineage stabilization. By contrast, this pattern is less compatible with Treg maintenance, which depends more strongly on metabolic flexibility, sustained lipid oxidation, and controlled energetic intensity than on persistent biosynthetic escalation [32].

This bias is counterbalanced by AMPK-associated oxidative adaptation. Under energetic stress, AMPK restrains ACC activity, lowers malonyl-CoA abundance, and restores CPT1-mediated fatty acid transport into mitochondria [32, 84]. As a result, lipid utilization shifts away from biosynthetic expansion and toward mitochondrial β-oxidation, allowing OXPHOS to sustain ATP production under conditions of restrained glycolytic drive [85]. In parallel, AMPK supports mitochondrial fitness, autophagy-associated homeostatic maintenance, and redox preservation, thereby generating a metabolic configuration that is more compatible with Treg persistence than with high-intensity inflammatory activity [86]. Within this framework, the mTOR–AMPK axis functions as an allocation system that determines whether integrated metabolic resources are committed to Th17-associated anabolic inflammation or to the oxidative stability required for sustained Treg function.

A second major axis is the HIF-1α–AhR–redox pathway, which links metabolic and inflammatory stress to lineage-selective adaptation. HIF-1α reinforces a state in which glycolytic support, altered pyruvate routing, and mitochondrial stress are coupled to inflammatory transcriptional maintenance, thereby strengthening the metabolic background that favors Th17 persistence while weakening Foxp3 stability [21]. In parallel, AhR acts less as an isolated amino acid-associated pathway than as a metabolite-sensitive interface through which extracellular metabolite availability, cytokine context, and intracellular stress states are converted into lineage plasticity [87]. Acid-derived signals are integrated with broader mitochondrial and inflammatory metabolic conditions rather than acting alone. These effects converge at the level of redox control, where glycolytic intensity contributes biosynthetic pressure, mitochondrial activity determines reactive oxygen species burden, lipid composition influences susceptibility to peroxidation, and amino acid metabolism provides antioxidant buffering capacity [88].

Taken together, these signaling hubs indicate that the metabolic balance between Th17 and Treg cells is dynamic rather than fixed. When mTOR-dominant anabolic routing and HIF-1α-associated stress adaptation prevail, metabolic resources are preferentially directed toward inflammatory persistence, favoring Th17 stability while weakening regulatory resilience. Conversely, when AMPK-linked oxidative support, mitochondrial fitness, and effective redox buffering are preserved, the same network becomes more compatible with Foxp3 maintenance and sustained Treg function. The Th17–Treg axis is therefore better understood as a dynamic metabolic state that is continually reshaped by nutrient availability, energetic stress, and local metabolite signals. Importantly, these integrated signaling axes represent conserved immunometabolic principles characterized across cancer and systemic autoimmune models. Yet their precise configuration in AITDs is uniquely reshaped by the thyroid-specific microenvironment. Ultimately, it is this tissue-specific metabolic adaptation that subverts regulatory networks and drives the characteristic Th17/Treg imbalance central to AITD pathogenesis (Fig. 1).

Fig. 1.

Integrated metabolic landscape governing Th17 and Treg cell fate decisions. Upon activation, naïve CD4⁺ T cells undergo lineage-specific polarization guided by distinct cytokine microenvironments. IL-6, IL-1β, and IL-23 drive differentiation toward the inflammatory Th17 lineage, whereas TGF-β and IL-2 orchestrate the induction of immunosuppressive Treg cells. Following these initial triggers, the two subsets engage divergent metabolic programs to sustain their functional identities. Treg cells maintain OXPHOS and FAO via CPT1A, supported by AMPK and PPAR signaling. Foxp3 restrains glycolysis by repressing c-Myc and GLUT1. TCA-derived α-KG and NAD⁺/Sirtuin coupling reinforce Foxp3-associated chromatin stability. GSH and transsulfuration-derived H₂S buffer oxidative stress. Restricted amino acid uptake via ASCT2/LAT1 activates the GCN2–eIF2α–ATF4 axis, preserving Treg potential while attenuating Th17 output.Th17 cells rely on mTORC1-driven aerobic glycolysis, with LDHA sustaining NAD⁺ regeneration and nuclear PKM2 amplifying STAT3/RORγt transcription. DRP1-dependent fission and mtDNA-cGAS-STING signaling facilitate inflammatory reprogramming. Succinate/fumarate-derived mROS stabilize HIF-1α and antagonize Foxp3. ACC1/FASN/SCD1-coordinated lipogenesis supports membrane expansion, while GLS1 and BCAT1 fuel biosynthesis. METTL3-mediated m6A methylation fine-tunes RORγt, and tryptophan-derived AhR ligands modulate lineage plasticity.The central interface depicts the mTOR–AMPK rheostat and AhR pathway as shared nodes governing the suppressive-to-inflammatory gradient (For visual clarity, the dashed nucleus represents a schematic projection of the functional nuclear compartment to avoid overlapping pathways)

Metabolic reprogramming mechanisms of Treg and Th17 cells in AITDs

Phenotypic disequilibrium and functional skewing of Th17 and Treg Cells in AITDs

AITDs are consistently associated with disrupted immune homeostasis within the CD4⁺ T-cell compartment. Although GD and HT differ in clinical presentation and dominant pathological outcome, both converge on a disturbed Th17-Treg axis. This disequilibrium is reflected not only in altered subset abundance, but also in disturbances in lineage stability and functional output, thereby providing an immunological basis for persistent thyroid inflammation and sustained autoantibody production.

At the quantitative level, multiple flow cytometric studies show that circulating Th17 cells are consistently increased in patients with GD and HT, approximately 2–5 times higher than in healthy controls, whereas Treg cells are reduced by approximately 20–40% in several studies, together driving an elevated Th17/Treg ratio [13]. This shift is accompanied by increased levels of Th17-associated cytokines, including IL-17 A, IL-21, and IL-22, as well as a transcriptional balance that tilts from FOXP3 toward RORγt [89]. Overall, the direction of Th17 skewing is relatively consistent across AITD, but the reduction in circulating Treg cells appears more reproducible in GD, whereas findings in HT remain more variable, likely because of differences in disease stage, treatment status, and the criteria used to define Treg subsets [90]. This immune deviation is not restricted to the circulation. Thyroid tissue from HT and orbital tissue from GD and GO likewise show enrichment of IL-17-producing CD4⁺ T cells together with relative paucity of FOXP3⁺ regulatory cells, indicating that the Th17–Treg imbalance is also maintained within target tissues [91–93]. The degree of imbalance further correlates with disease activity and autoantibody burden, supporting its close association with both intrathyroidal and extrathyroidal manifestations of AITDs [93].

Beyond changes in cell number, Th17 and Treg cells in AITDs also exhibit abnormalities in lineage stability and effector function. Treg cells often show weakened suppressive capacity, reflected by reduced IL-10 production after stimulation and impaired ability to restrain effector T-cell proliferation. Instability of FOXP3 expression further points to defective maintenance of regulatory identity rather than a simple numerical loss [94]. Recent single-cell multi-omic studies in Graves’ orbitopathy further suggest that, under persistent inflammatory and antigen-experienced conditions, activated Treg populations may lose canonical regulatory features and acquire cytotoxic-like characteristics, consistent with ex-Treg deviation [95, 96]. Concomitantly, abnormalities in the Th17 compartment are not limited to numerical expansion, but also involve sustained reinforcement of pathogenic programs, including maintenance of IL-23–IL-17-associated signaling and persistence of migration- and activation-related phenotypes such as CCR6 and CXCR3 [97, 98]. Enhanced responsiveness of thyroid follicular cells and orbital stromal cells to IL-17-associated signaling further links this immune skewing to local inflammatory amplification [91]. Together, impaired regulatory restraint and persistent pathogenic Th17 activity provide a cellular basis for chronic inflammation and tissue injury in AITDs.

Mechanism of metabolic-driven functional polarization of Treg and Th17 cells in AITDs

In AITDs, intracellular metabolic reprogramming provides an important mechanistic basis for persistent disequilibrium between Th17 and Treg cells. Current evidence does not accumulate evenly across all disease subtypes. The clearest and most integrated mechanistic support is presently available in HT, whereas findings from GD and GO remain more fragmentary and are often derived from related tissue contexts rather than direct lineage-resolved analyses. The metabolic landscape in this field is broad and still evolving. However, alterations in glucose, lipid, mitochondrial, and amino acid metabolism consistently converge to enhance inflammatory Th17 activity and compromise Treg stability.

In HT, the clearest evidence for disease-associated metabolic reprogramming currently comes from enhanced glycolysis in CD4⁺ T cells. Peripheral CD4⁺ T cells from patients with HT show increased expression of GLUT1, HK2, PKM2, and LDHA, together with activation of the mTOR–HIF-1α glycolytic pathway [99]. In this setting, heightened glycolytic activity is associated with inflammatory T-cell polarization, whereas preservation of FOXP3 stability in Treg cells appears to remain more dependent on oxidative metabolism [100, 101]. Consistent with this pattern, HIF-1α activation in HT accompanies enhanced RORγt-driven IL-17 production and impaired maintenance of the Treg program, indicating that increased glucose uptake and glycolytic activation are closely linked to the Th17/Treg imbalance [102]. Pharmacological suppression of glycolysis correspondingly reduces Th17-associated responses and improves the frequency and suppressive function of Treg cells, supporting a functional contribution of glycolytic reprogramming to immune disequilibrium in HT [99].

Evidence from experimental autoimmune thyroiditis (EAT) further places this process in an upstream signaling context. In EAT mice, increased p-AKT and p-mTOR expression accompanies Th17 expansion and higher IL-17 A levels, and PI3K inhibition attenuates these changes while alleviating thyroid inflammation [103]. In parallel, Notch-driven Th17 polarization in autoimmune thyroiditis has been mechanistically connected to PI3K–AKT–mTORC1 activation and is similarly restrained by PI3K blockade [103]. Collectively, clinical and experimental evidence identifies a PI3K–Akt–mTOR–glycolysis axis that drives Th17-skewed inflammation while destabilizing Tregs in HT.

This pattern is not restricted to HT. In GO, recent evidence indicates that WTAP-mediated m6A modification of THBS1 promotes disease progression through glycolysis-dependent disturbance of the Th17/Treg balance [104]. Although these data derive from an orbital manifestation of GD, they still suggest that enhanced glycolytic reprogramming can contribute to pathogenic Th17/Treg imbalance across the AITD spectrum.

Lipid metabolic remodeling provides another intracellular basis for the persistent Th17/Treg imbalance in HT. Evidence from peripheral CD4⁺ T cells shows increased CPT1A expression together with reduced intracellular levels of several medium- and long-chain fatty acid substrates, supporting enhance FAO [105]. Experimental EAT models further indicate activation of a lipogenic program, with increased expression of ACC1 and FASN together with abnormal upregulation of CPT1A [105]. Considered together, these findings suggest that activated CD4⁺ T cells in HT do not simply alter substrate consumption, but undergo coordinated changes involving both FAS and FAO. This pattern is relevant to Th17/Treg disequilibrium because ACC1-dependent lipid synthesis supports Th17 expansion and effector function, whereas disturbed lipid oxidative homeostasis can undermine the metabolic fitness required for Treg maintenance [37]. Accordingly, lipid abnormalities in HT are consistent with persistence of Th17 predominance and impaired preservation of the Treg compartment.

Experimental intervention studies support a functional role for this lipid metabolic shift. In EAT, inhibition of the activated mTOR–ACC1–CPT1A program reduces the proportion of Th17 cells, increases the Treg compartment, and alleviates thyroid lymphocytic infiltration, indicating that lipid metabolic reprogramming contributes to distortion of the Th17/Treg axis [105]. Additional support is provided by tissue-level findings in HT. In HT thyroid tissue, increased total lipid levels, reduced CPT1 expression, increased TNF-α, upregulation of CD36 and phospholipase A2, downregulation of 15-LOX, and impaired pro-resolving lipid mediator signaling together indicate that local lipid remodeling is coupled to defective resolution of inflammation [106]. By contrast, direct evidence for T-cell intrinsic lipid reprogramming in GD or GO remains limited. Evidence regarding lipid profiles in GO is currently conflicting. Certain studies associate elevated LDL levels with increased disease severity, whereas other research fails to support such a correlation [107, 108]. Overall, lipid dysregulation may participate in GD, but its impact on Th17/Treg imbalance has not yet been defined with the same mechanistic clarity available in HT.

Mitochondrial metabolic disturbance in HT is currently reflected more clearly by intensified oxidative stress and redox imbalance than by direct measurements of respiratory flux. In patients with HT, NOX4 and PKM2 levels are elevated and positively correlate with TPOAb, IFN-γ, and IL-17 [109]. In EAT, increased ROS accompanies more severe thyroid inflammation and stronger Th17-associated responses. Treatment with an oxidative stress inhibitor reduces ROS and inflammatory cytokines, decreases IL-17 A⁺ Th17 cells, increases FoxP3⁺ Treg cells [110]. These disease-associated changes indicate that redox dysregulation in HT is accompanied by enhanced Th17 activity and more difficult maintenance of the Treg compartment.

Similarly, oxidative stress responses are recognized as crucial components of GD pathology. While GD lacks the detailed thyroid-specific evidence seen in HT, it clearly exhibits comparable redox abnormalities systemically and particularly within GO. In GO, oxidative stress is a well-established disease feature, orbital fibroblasts show increased sensitivity to oxidative injury, and HIF-1α-associated pathways are linked to glycolysis, adipogenesis, and fibrotic remodeling in orbital tissue [111, 112]. These findings place redox-associated metabolic disturbance within a broader AITD context, even though the most direct linkage to Th17 and Treg dysfunction has been delineated in HT.

Among amino acid pathways, the strongest disease-specific evidence in HT currently comes from dysregulated tryptophan metabolism. Clinical studies show reduced serum tryptophan levels in HT, and experimental indicates that this abnormality is functionally linked to immune imbalance rather than representing an isolated systemic metabolic fluctuation [113]. In EAT, tryptophan supplementation alleviates thyroid injury and inflammatory cytokine production, improves T-cell subset balance, and suppresses PI3K–Akt pathway activation, whereas inhibition of tryptophan metabolism aggravates disease progression and enhances inflammatory signaling [113]. Taken together, these findings support the view that impaired tryptophan metabolism in HT weakens regulatory support for Treg stability while permitting stronger Th17-associated responses. Additional amino acid abnormalities are also beginning to emerge in HT. Recent metabolomic studies indicate that TPOAb-positive and TgAb-positive HT are associated with distinct metabolic profiles involving glycine, serine, and threonine metabolism, suggesting that amino acid-related reprogramming may vary across antibody-defined subgroups [114].

Related evidence is also accumulating in GD. A growing body of research suggests that tryptophan-derived metabolites can reduce inflammation and proliferation in orbital fibroblasts, and broader work on the gut-thyroid axis further indicates that microbiota-derived metabolites, including tryptophan-related products, may modulate immune responses in GD [115]. Therefore, these findings suggest that amino acid-related immunometabolic changes are relevant across the AITD spectrum.

Thyroidal and orbital microenvironmental amplification of Th17/Treg disequilibrium in AITDs

Once Th17/Treg disequilibrium is established in AITD, the target tissue microenvironment does not simply receive infiltrating immune cells. It actively amplifies and entrenches the imbalance through redox stress, metabolite retention, persistent antigenic stimulation, and progressive stromal remodeling. These tissue-level conditions matter because they influence whether inflammatory effector programs remain dominant or whether regulatory recovery is still possible. Excess iodine, stressed thyrocytes, activated orbital fibroblasts, and locally retained metabolites together form a microenvironmental network that sustains Th17-associated activity while gradually wearing down the metabolic conditions on which Treg stability depends.

Iodine excess

Excess iodine is one of the most firmly established environmental factors associated with AITD, supported by both epidemiological observations and experimental models [116]. In genetically susceptible hosts, chronic iodine overload is associated with increased thyroid autoantibody production and greater Th17 expansion among thyroid-infiltrating lymphocytes, suggesting that iodine excess acts as a thyroid-centered trigger of local immune disequilibrium rather than a nonspecific environmental insult.

The thyroid gland is particularly vulnerable to iodine-induced oxidative stress because thyroid hormone synthesis depends on iodine-related redox reactions. In the NOD-H-2h4 mouse model, excess iodine induces follicular cell injury and excessive ROS generation, establishing a sustained oxidative environment within the gland. This oxidative burden extends well beyond epithelial damage and directly disrupts local immune regulation [117]. Recent evidence further indicates that iodine can activate the NLRP3 inflammasome in immune cells from patients with AITD and promote Th1 and Th17 differentiation, suggesting that genetic susceptibility may partly reflect heightened cellular sensitivity to iodine-induced inflammatory stress [118]. Under these conditions, local immune activation couples more readily to metabolic programs that drive inflammatory effector responses than to those that maintain regulatory function.

Iodine excess also disrupts antioxidant defense through the Keap1–Nrf2 pathway. When this axis is dysregulated, mitochondrial function becomes more susceptible to oxidative injury, glycolytic reprogramming is reinforced, and conditions become more favorable for inflammatory T-cell differentiation [119]. At the same time, reduced antioxidant capacity impairs ROS-scavenging in Treg cells, increases mitochondrial stress, and destabilizes Foxp3 expression [120]. Recent transcriptomic data suggest that the thyroidal response to iodine overload is shaped by baseline Keap1/Nrf2 activity, which may partly explain why excess iodine produces more pronounced inflammatory injury in some individuals than others. The same oxidative environment can thus intensify effector activity while simultaneously undermining the metabolic conditions required for Treg maintenance.

Excess iodine also upregulates MHC class I and II expression on thyroid epithelial cells, strengthening local antigen presentation. This sustained antigenic stimulation adds a continuing metabolic burden on infiltrating lymphocytes and further amplifies abnormal immune programming within the gland. In this way, iodine excess links the oxidative demands of thyroid hormone synthesis to persistent Th17/Treg disequilibrium within the target organ [121].

Thyrocytes and orbital fibroblasts

Far from being passive casualties of autoimmune attack, tissue-resident epithelial and stromal cells function as active hubs of local immunometabolic amplification. Within the thyroid, follicular thyrocytes continuously translate inflammatory signals into sustained oxidative and immunological pressure. Because these cells operate under high mitochondrial activity and an inherently demanding redox environment, they are particularly sensitive to cytokines released by infiltrating immune cells, including IL-17 and IFN-γ [122]. Under inflammatory stress, thyrocytes increase ROS production and release pro-inflammatory mediators and chemoattractants such as IL-6, IL-8, and MCP-1, promoting further immune recruitment and prolonging intrathyroidal T-cell activation [102, 123]. Thyrocytes can also acquire antigen-presenting features under these conditions, though professional APCs such as dendritic cells likely play a more prominent role at earlier disease stages, with thyrocyte HLA-II expression serving mainly to amplify rather than initiate immune responses [124].

The consequences of this thyrocyte activation differ between disease subtypes. In HT, the result is a microenvironment in which inflammatory effector activity is continuously renewed while the conditions needed for regulatory recovery are progressively lost. Sustained thyrocyte stress increases susceptibility to epithelial injury, perpetuates antigen exposure, and maintains a cycle in which tissue damage and immune activation reinforce each other [1]. Over time, this leads not only to persistent inflammatory infiltration but to gradual disruption of glandular architecture, progressive follicular loss, and eventual fibrotic replacement of damaged tissue [125]. Emerging evidence suggests that stromal remodeling may begin earlier than previously thought, adding further to matrix alteration and inflammatory persistence.

In GD, thyrocyte activation is coupled less to follicular destruction than to sustained immune collaboration within the gland. Persistent epithelial stress, antigen presentation, and cytokine release help maintain a thyroidal environment that supports continued T-cell activation, B-cell help, and autoantibody production [126]. Here the thyrocyte acts as a local amplifier of the intrathyroidal immune network that sustains TRAb-driven disease [127]. Thus, while HT and GD differ in their pathological outcomes, both involve thyrocyte-dependent reinforcement of tissue-retained immune disequilibrium, and in both the local metabolic environment remains more compatible with inflammatory persistence than with full restoration of regulatory control.

A parallel amplification system operates in GO, where orbital fibroblasts serve as the principal stromal organizers of local metabolic remodeling. A central pathogenic mechanism involves activation of the TSHR/IGF-1R signaling complex on orbital fibroblasts by autoantibody-driven stimulation [128]. This receptor crosstalk couples inflammatory signaling to downstream metabolic remodeling, which is consistent with the clinical efficacy of IGF-1R-targeted therapy in this disease [129]. Once activated, orbital fibroblasts increase glycolytic flux, lactate output, and profibrotic signaling. Recent evidence identifies PDK2 as an upstream driver of the glycolysis–IL-11 axis in GO, whereby enhanced glycolysis promotes IL-11 expression and fibrotic remodeling, and inhibition of this pathway reduces fibrosis in experimental models [130]. HIF-1α also appears to play an upstream role, coordinating glycolytic, adipogenic, and profibrotic transcriptional programs within these cells [131].

Orbital fibroblasts are also major effectors of hyaluronan synthesis, adipogenic expansion, and fibrotic remodeling, which together underlie the tissue enlargement and structural distortion seen in GO [132]. Their activation shapes an orbital niche marked by persistent cytokine exposure, metabolic byproduct accumulation, extracellular matrix expansion, and sustained stromal activity. Emerging single-cell analyses reveal considerable heterogeneity among orbital fibroblast subsets, which may help explain the variable clinical presentation and fibrotic phenotypes of this disease [133]. Hyaluronan accumulation, adipose expansion, and fibrosis therefore reflect linked outcomes of the same locally reinforced inflammatory-metabolic circuit. For infiltrating immune cells, this stromal environment is not neutral, it retains inflammatory signals, supports effector persistence, and makes restoration of Treg function progressively harder to achieve [134].

In this respect, thyrocytes in HT and GD and orbital fibroblasts in GO all function as local amplifiers, each through disease-specific patterns of tissue remodeling. The shared consequence is a tissue environment that favors Th17 persistence and works against Treg-mediated restraint.

Metabolite gradients and redox niches

A defining feature of target tissues in AITD is the accumulation of local metabolite gradients that spatially stabilize inflammatory phenotypes once T-cell lineage imbalance has been established. Lactate is particularly relevant in this regard. It is produced not only by glycolytically active immune cells but also, in GO, by metabolically reprogrammed orbital fibroblasts [135]. Rather than a simple metabolic end-product, lactate functions as a tissue-retained signal that reinforces existing immune skewing. When locally accumulated, it promotes extracellular acidification, sustains inflammatory programming, and weakens regulatory fitness, creating an environment in which effector cells are retained more readily than suppressive ones [135]. There is also emerging evidence that lactate influences immune-cell behavior through lactylation-dependent mechanisms, offering an additional route by which these metabolically stressed niches stabilize inflammatory phenotypes and impede regulatory recovery [136, 137].

The spatial organization of these gradients reflects disease-specific tissue architecture. In HT, metabolite retention is closely tied to chronic follicular disruption, epithelial oxidative stress, and ongoing metabolic exchange between damaged thyrocytes and infiltrating immune cells, all of which cumulatively hinder regulatory recovery [138]. In GD, sustained thyrocyte stimulation, immune-cell recruitment, and antibody-driven inflammation similarly favor local retention of inflammatory metabolites within the thyroid, even though the structural consequences differ [139]. In GO, lactate accumulation is especially marked because activated orbital fibroblasts provide a sustained stromal source and hypoxia-associated signaling further consolidates this metabolically biased niche [130, 135]. In each setting, the target tissue provides a bioenergetic environment that favors inflammatory persistence over regulatory restoration.Redox heterogeneity adds further spatial structure to these tissues. In all three disease contexts, oxidative burden is not uniformly distributed, instead, it generates focal niches in which inflammatory cells, activated stromal or epithelial cells, and injured parenchymal cells continuously exchange damaging metabolites and stress signals [140].

Self-reinforcing progression of immune-metabolic disequilibrium

Once Th17/Treg disequilibrium is established in AITDs, it rarely remains a transient immune disturbance. Instead, it progressively develops into a self-reinforcing state in which inflammatory polarization, metabolic dysfunction, and tissue stress continuously sustain one another [110]. Enhanced effector activity intensifies local cytokine production, oxidative burden, and metabolic disturbance, whereas impaired regulatory control weakens the capacity to restrain and resolve these responses. Under these conditions, the initial imbalance is not simply maintained, but repeatedly re-amplified, and the target tissue increasingly evolves into a microenvironment that favors persistent immune dysregulation and chronic inflammation. Although this self-reinforcing progression is shared across AITDs, its pathological expression differs by disease context [141]. In HT, it promotes progressive follicular injury and fibrotic replacement. Thus, once immune and metabolic disequilibrium becomes established, the local disease process is increasingly locked into a state that is difficult to reverse ( Fig. 2).

Fig. 2.

Metabolic reprogramming mechanisms of Treg and Th17 Cells in AITDs. Upstream triggers (top panel), including iodine excess, altered amino acid availability and genetic/environmental factors, converge to initiate thyroid autoimmunity. Following these triggers, disease-specific intracellular metabolic reprogramming bifurcates into distinct tissue pathways. In Hashimoto’s thyroiditis (HT, left panel), activation of the PI3K–Akt–mTOR–HIF-1α axis drives robust glycolytic and lipid reprogramming (evidenced by elevated GLUT1, HK2, PKM2, ACC1, and CPT1A), accompanied by NOX4-mediated reactive oxygen species (ROS) production. Conversely, in Graves’ disease and orbitopathy (GD/GO, right panel), WTAP-mediated m6A modification drives a PDK2–IL-11 glycolytic axis, leading to pathological lactate accumulation, while TSHR/IGF-1R signaling amplifies orbital fibroblast activation.These distinct metabolic routes converge centrally to induce pathogenic Th17 expansion and Treg destabilization within hypoxic, ROS-, and lactate-enriched niches, amplifying a cascade of pro-inflammatory cytokines (e.g., IFN-γ, IL-6, IL-17, IL-11). Downstream tissue-specific consequences are depicted in the lower panels. In HT (bottom left), follicular loss and fibrosis disrupt the Keap1/Nrf2 antioxidant pathway, leading to mitochondrial damage, ROS accumulation, and the release of pro-inflammatory mediators (e.g., IL-6, IL-8, MCP-1), along with elevated TPOAb production. Dysregulation of CPT1, 15-LOX, CD36, PLA2, and TNF-α further reflects impaired lipid resolution and chronic inflammation within the thyroid parenchyma. In GD/GO (bottom right), HLA-II-driven TRAb production fuels hyperthyroidism, while activated orbital fibroblasts orchestrate hyaluronan accumulation and adipogenic expansion. Crucially, sustained oxidative stress, lactate accumulation, and cytokine release form a self-reinforcing metabolic-inflammatory cycle, progressively locking the target tissues into a state of chronic immune disequilibrium that resists immunological restoration

Therapeutic potential of metabolic targeting interventions

The central role of T cell metabolism in the pathogenesis of AITDs has increasingly made metabolic targeting a focus of research. Current intervention strategies leverage the metabolic plasticity of Th17 and Treg cells by aligning with the natural biological hierarchy of these pathways. At the apex are upstream metabolic regulators that dictate broad immunometabolic programming and cell fate decisions. Further down the cascade, downstream interventions target the specific enzymatic steps and metabolic fluxes that directly fuel effector functions. Intervening across these distinct regulatory nodes establishes a comprehensive therapeutic framework. This approach is geared towards restoring the Th17/Treg balance, mitigating local inflammation, and opening novel treatment avenues for AITDs.

Targeting upstream metabolic regulators

Upstream metabolic regulators provide the most direct therapeutic entry points because they influence whether inflammatory and regulatory immune programs become metabolically stabilized. In AITDs, intervention at this level is relevant not only to the Th17/Treg balance itself, but also to the thyroid and orbital microenvironments. Current therapeutic evidence does not accumulate evenly across all targets. Some upstream nodes already show disease-facing relevance in AITD-oriented systems, whereas others remain primarily supported by mechanistic rationale and cross-disease immunometabolic evidence.

The inhibition of mTOR signal is among the most mature strategies in this category. Representative agents include rapamycin (sirolimus), everolimus, temsirolimus, and dioscin, all of which act through mTOR-centered signaling, particularly mTORC1. Across immune-mediated disease models, these compounds suppress Th17 polarization while preserving or enhancing Treg-associated features [142]. In the NOD.H-2h4 spontaneous autoimmune thyroiditis model, dioscin reduced thyroid lymphocytic infiltration and autoantibody levels, providing direct support in an AITD-relevant setting [143]. Parallel evidence firmly establishes the therapeutic relevance of this axis in GO. In human orbital fibroblasts, mTOR inhibition with rapamycin directly suppresses adipogenesis, hyaluronic acid accumulation, and myofibroblast-driven fibrotic contraction [144]. Furthermore, in adenovirus-induced GO mouse models, rapamycin significantly ameliorates orbital tissue remodeling by suppressing pathogenic CD4 + cytotoxic T lymphocytes [145]. Clinically, low-dose sirolimus has recently demonstrated notable efficacy in steroid-refractory GO patients, improving clinical activity scores and ocular motility restriction [146].

The activation of AMPK pathway provides a second major upstream avenue. Metformin and AICAR are the most relevant examples, with metformin carrying the strongest disease-facing support. In thyroglobulin-induced EAT models, metformin reduced TgAb, diminished thyroid lymphocytic infiltration, and suppressed Th17-associated responses while improving regulatory balance [147]. Across multiple clinical studies in Hashimoto thyroiditis, metformin has also been associated with lower TSH and thyroid autoantibody levels [148]. Beyond HT, the therapeutic potential of the AMPK axis extends directly to the orbital microenvironment in GO. In patient-derived orbital fibroblasts, AMPK activation by metformin or AICAR potently suppresses adipocyte differentiation, curbs extracellular hyaluronic acid accumulation, and attenuates pro-inflammatory cytokine, notably IL-6, production by antagonizing downstream pathogenic signaling [149]. These findings give the AMPK axis firmer translational relevance in thyroid autoimmunity than many upstream targets that are still grounded mainly in non-thyroid autoimmune models.

Targeting HIF-1α represents another important upstream strategy, particularly in disease settings characterized by tissue stress. This axis is especially relevant in GO, where orbital tissues form mechanically constrained inflammatory niches in which hypoxia-like signaling, lactate accumulation, and stromal-immune reciprocity may sustain local disease activity. Crucially, targeted HIF-1α inhibition in GO fibroblasts effectively blunts IL-6 expression [150]. Targeting in HT remains underexplored, its microenvironment exhibits a parallel signature of oxidative stress and glycolytic bias that favors HIF-1α stabilization, making it a logical intervention node [102]. Representative HIF-1α-directed compounds include echinomycin, PX-478, and YC-1. In immune-mediated disease models outside AITD, these agents have been shown to reduce inflammatory cytokine production and attenuate tissue injury, consistent with suppression of glycolysis-associated pathogenic programs [151, 152]. While such findings support the mechanistic relevance of this axis, their applicability to AITD remains to be established. The translational significance of this pathway is further reinforced by teprotumumab, an IGF-1R-blocking monoclonal antibody. Although teprotumumab is not itself a direct HIF-1α inhibitor, its efficacy in GO is consistent with disruption of a fibroblast-centered pathogenic program that includes HIF-1α stabilization and glycolytic adaptation [153].

SIRT1-activating strategies exert their effects via a different upstream logic by stabilizing regulatory function under chronic metabolic and oxidative stress. Resveratrol remains the historically cited prototype, however, more recent investigations have shifted focus toward broader NAD+-supportive strategies, including β-nicotinamide mononucleotide (NMN) and its related precursors. In multiple disease models, activation of the NAD+/SIRT1 axis has been associated with reduced inflammation and lower oxidative stress, supporting the view that SIRT1-centered intervention may enhance Treg resilience rather than simply suppress inflammatory expansion [154, 155]. In AITDs, where thyroid and orbital tissues are exposed to sustained redox imbalance, this strategy is mechanistically attractive, because it links intracellular energetic state to immunoregulatory stability. But direct disease-specific validation remains limited, so these approaches are better regarded as emerging rather than established therapeutic options. More speculatively, modulation of NAD+-consuming enzymes such as CD38 may also reshape intracellular NAD+ availability and thereby influence sirtuin-dependent immune homeostasis [156]. Recent work nevertheless suggests that the CD38-NAD+-SIRT1 axis is strongly context-dependent, particularly in local Treg adaptation, and its relevance to AITD remains undefined [157].

AhR-directed intervention should also be placed in this upstream category because AhR functions as a metabolite-responsive transcriptional regulator rather than merely a downstream consequence of tryptophan metabolism. The tryptophan-kynurenine-AhR axis is closely linked to Treg stability, and AhR activation can suppress pathogenic Th17-associated programs while favoring regulatory differentiation [80]. In this context, laquinimod provides a representative example. Its active metabolite appears to function less by directly suppressing pathogenic T cells than by engaging AhR-expressing antigen-presenting cells, especially dendritic cells, to induce an IDO-associated tolerogenic program that promotes regulatory differentiation and dampens pathogenic T-cell responses [158]. This makes AhR a conceptually coherent upstream target for repairing a metabolic-immunological defect that is already implicated in AITD-associated immune dysregulation.

Additional upstream regulators remain valuable as exploratory nodes because they sit near the metabolic commitment point of inflammatory T-cell activation. c-Myc inhibitors such as Omomyc and 10,058-F4 suppress glycolytic and glutaminolytic enzymes to limit rapid effector proliferation and inflammatory activity in a KRAS-driven NSCLC mouse model and in Phase I patients receiving the Omomyc derivative OMO-103 [159–162]. Although this direction is still supported mainly by non-thyroid immune disease studies, it remains noteworthy because c-Myc sits at a central regulatory junction connecting nutrient uptake, biosynthetic programming, and pathogenic immune expansion [163, 164]. For this reason, c-Myc is best viewed as a future-oriented regulatory node with potential relevance to AITD metabolic intervention.

Targeting downstream metabolic effectors

If the upstream regulatory level determines the direction of immune-metabolic polarization, the downstream effector level sustains its execution. This level includes the core metabolic pathways that directly support inflammatory or regulatory function, together with the downstream metabolites and stromal products through which these pathways shape thyroid and orbital microenvironments. In AITDs, persistent disease is therefore maintained not only by cell-intrinsic metabolic flux, but also by lactate accumulation, oxidative burden, lipid mediators, and fibroblast-derived outputs that amplify local inflammation.

Glycolysis-directed intervention remains one of the most direct ways to restrain pathogenic effector activity. Various classical glycolysis inhibitors, including 2-deoxy-D-glucose (2-DG), lonidamine, FX11, have been developed to target glucose utilization or specific enzymes such as GLUT1, HK2, LDHA, and PFKFB3 [165]. Among these, the therapeutic potential of broad glycolytic blockade is directly validated in EAT models, where 2-DG robustly decreases Th17 cell frequencies and ameliorates thyroid glandular destruction [166]. However, because blanket suppression carries systemic toxicity risks, recent focus has shifted toward selective nodal effectors like PKM2. When PKM2 dissociates into dimers and translocates to the nucleus, it acts as a nonmetabolic coactivator for STAT3-dependent Th17 programming [24]. Although selective PKM2 modulators await extensive in vivo testing in AITD, indirect evidence comes from EAE models. In these systems, small molecules such as TEPP-46, which lock PKM2 in its transcriptionally inactive tetrameric state, suppress pathogenic Th17 differentiation and alleviate neuroinflammation without disrupting basal cellular metabolism [167]. Given the shared STAT3/Th17 reliance in HT, stabilizing PKM2 represents a translationally viable strategy for targeted immune correction. Furthermore, because glycolytic flux critically supports the inflammatory activation of orbital fibroblasts, targeting nodal regulators like PKM2 may eventually yield therapeutic benefits extending to GO.

Lipid metabolism, particularly FAS, provides another major downstream execution target. Inflammatory T cells rely heavily on FAS for membrane biogenesis and clonal expansion, with ACC1 acting as the rate-limiting gatekeeper. Multiple pharmacological strategies exist to modulate lipid handling, but ACC1-specific inhibitors such as Soraphen A and TOFA represent the most robustly validated interventions [168, 169]. While direct ACC1 inhibitors have yet to enter AITD clinical trials, their efficacy in related autoimmune models provides a strong translational rationale. In murine models of EAE and colitis, treatment with these agents selectively starves Th17 cells of crucial lipid building blocks, effectively attenuating tissue damage while simultaneously promoting Treg cell generation [170]. Because the persistence of HT is fundamentally driven by a Th17/Treg imbalance sustained in a lipid-rich microenvironment, targeting ACC1 offers a mechanistically rational downstream strategy. Importantly, the relevance of this axis likely extends to GO. Since pathogenic orbital tissue expansion heavily depends on localized adipogenesis and lipid accumulation, limiting FAS may provide a dual mechanism to restrain both systemic T-cell activation and local orbital remodeling.

Importantly, this lipid-centric framework extends beyond internal cellular synthesis to include systemic lipid mediators such as the short-chain fatty acid butyrate. Although AITD patients often exhibit a depletion of butyrate-producing bacteria, supplemental restoration of this lipid metabolite has shown remarkable efficacy in multiple autoimmune paradigms [171, 172]. As a naturally occurring HDAC inhibitor, butyrate directly enforces Foxp3 expression and stabilizes Treg lineage commitment even under inflammatory stress [173]. For AITD, pharmacologically replenishing systemic butyrate levels may limit both systemic T-cell activation in HT and localized orbital remodeling in GO.

Mitochondrial metabolism and redox control constitute a critical downstream execution layer. In AITD, the thyroid gland is uniquely susceptible to oxidative injury during hormone synthesis. When mROS continuously accumulate, they not only fuel inflammatory T-cell polarization but also drive severe lipid peroxidation in thyroid tissues. In experimental autoimmune settings, mitochondria-targeted antioxidants such as MitoTEMPO, as well as broader ROS scavengers like N-acetylcysteine, have been shown to attenuate mROS signaling, reduce inflammatory T-cell differentiation, and preserve regulatory function [174]. Clinical observations further support the relevance of this redox axis. In GO selenium supplementation has been shown to improve clinical outcomes in patients with mild disease [175]. Mechanistically, selenium acts not merely as a generic antioxidant, but as the essential co-factor for glutathione peroxidase 4 (GPX4), the master enzyme that clears lipid peroxidesn [176]. This redox-centric therapeutic mechanism is directly corroborated in EAT models. Targeted pharmacological intervention with agents like ursolic acid alleviates thyroid tissue destruction and restores the Th17/Treg balance by suppressing pathogenic ACSL4 expression and rescuing GPX4 activity [110]. Therefore, managing mitochondrial redox homeostasis and its downstream peroxidative consequences offers a validated intervention capable of directly arresting thyroidal autoinflammation.

Amino acid metabolism provides another crucial downstream layer. Glutamine metabolism supports inflammatory T-cell function through biosynthetic supply, signaling support, and maintenance of redox homeostasis. Representative inhibitors include BPTES, CB-839, DON and V-9302, which target GLS1, ASCT2, or related glutamine-utilizing processes [177]. In models such as EAE and colitis, these agents reduce Th17 polarization while favoring a relative Treg bias [33]. Crucially, the relevance of glutamine blockade extends deeply into GD because TRAb production depends strictly on the metabolic expansion of Tfh cells and germinal center B cells, both of which are highly reliant on glutaminolysis [178, 179]. Parallel to glutamine, tryptophan metabolism through the kynurenine pathway offers an additional execution-level target for immune modulation. Although direct AhR-targeted therapies for AITD remain exploratory, evidence from EAE models demonstrates that pharmacological activation of this pathway with L‑kynurenine can successfully shift the T‑cell balance and reduce tissue‑specific inflammation [180]. Within the AITD spectrum, enhancing kynurenine-driven regulatory signals represents a mechanistically rational strategy to counteract the T-cell-mediated destruction characteristic of HT while potentially limiting the orbital inflammatory remodeling seen in GO. By integrating glutamine restriction with the modulation of tryptophan-derived metabolites, these amino acid-centered strategies provide a comprehensive approach to interrupting pathogenic persistence in AITD.

At the local tissue level, downstream intervention must also intercept the stromal effector outputs that physically sustain AITD tissue remodeling. This is paramount in GO, where orbital fibroblasts execute disease progression by massively overproducing hyaluronan. Targeting hyaluronan synthases, specifically HAS2, provides a direct execution-level blockade [181]. Strong preclinical evidence demonstrates that treatment with 4-methylumbelliferone (4-MU), a classic pharmacological inhibitor of HA synthesis, effectively halts pathological HA accumulation and arrests orbital fibroblast proliferation in patient-derived GO models [179]. For example, in patient-derived GO orbital fibroblasts, knocking down ATF6 with small interfering RNA strongly blocks PDGF-induced HAS2 expression [181]. Similarly, overexpressing METTL3 also strongly blocks It. There’s no direct evidence for GO yet, but a METTL3/14 activator called C4 has been shown to improve motor function and protect dopaminergic neurons in the rat 6-OHDA model of Parkinson’s disease. This hints that METTL3-activating compounds could be worth exploring as a treatment for GO [182]. By shifting the therapeutic focus from upstream systemic immunosuppression to the direct pharmacological or epigenetic blockade of HAS2, these stromal-targeted strategies offer a highly precise mechanism for reversing the physically constrained orbital tissue expansion characteristic of GO.

Ultimately, metabolic intervention in AITDs can be conceptualized as an integrated, hierarchical framework rather than a collection of isolated targets. The mechanistic continuity between upstream regulatory sensors and downstream metabolic executors dictates both the initial polarization and the persistence of pathogenic immune states. Crucially, this therapeutic paradigm extends beyond intrinsic T-cell dysfunction, such as Th17/Treg imbalance. It directly addresses the thyroidal and orbital microenvironments that amplify local inflammation. By targeting these interconnected nodes, current strategies provide a comprehensive mechanism for dismantling the metabolic-inflammatory circuits underlying AITDs (Table 1).

Table 1.

Therapeutic potential of metabolic targeting interventions

	Metabolic level/ strategy	Representative agents	Direct target / pathway	Main AITD-relevant effects	Disease context / model	Ref.
Upstream regulators	mTOR inhibition	Rapamycin (sirolimus), everolimus, temsirolimus, dioscin	mTORC1 signalling	Suppresses Th17 polarization; reduces thyroid infiltration and TgAb; suppresses orbital adipogenesis, HA accumulation, and fibrotic contraction	NOD.H-2h4 mice; GO orbital fibroblasts; adenovirus-induced GO models; steroid-refractory GO patients	[142–146]
	AMPK activation	Metformin, AICAR	AMPK signalling axis	Reduces TgAb and Th17 responses; suppresses orbital adipocyte differentiation, HA accumulation, and IL-6 production	EAT models; HT patients; patient-derived GO fibroblasts	[147–149]
	HIF-1α inhibition	Echinomycin, PX-478, YC-1	HIF-1α stabilization	Blunts IL-6 expression; attenuates hypoxia/glycolysis-associated tissue injury	GO orbital fibroblasts; HT microenvironmental rationale	[102, 150–152]
	IGF-1R blockade	Teprotumumab	IGF-1R/TSHR crosstalk	Disrupts fibroblast pathogenic programmes; curtails downstream HIF-1α/glycolytic adaptation	GO patients (clinical trial)	[153]
	SIRT1/NAD⁺ support	Resveratrol, NMN, CD38 modulators	NAD⁺/SIRT1 axis; CD38	Enhances Treg resilience; lowers inflammation and oxidative stress	AITD redox imbalance rationale; non-AITD models	[154–157]
	AhR-directed intervention	Laquinimod	Trp–Kyn–AhR axis	Induces IDO-associated tolerogenic programme via APCs; promotes regulatory differentiation	AITD immune dysregulation rationale	[80, 158]
	c-Myc inhibition	Omomyc,10,058-F4	c-Myc	Suppresses glycolytic/glutaminolytic reinforcement; limits rapid effector proliferation	KRAS-driven NSCLC mouse model; phase I patients	[159–164]
Downstream effectors	Glycolysis-directed intervention	2-DG, lonidamine, FX11, TEPP-46	Glucose utilization (GLUT1, HK2, LDHA, PFKFB3); PKM2 tetramer stabilization	Decreases Th17 frequencies; ameliorates thyroid destruction; suppresses STAT3-dependent Th17 programming	EAT models; EAE models	[24, 165–167]
	Lipid metabolism targeting	SoraphenA, TOFA, butyrate	ACC1 (FAS); HDAC inhibition	Starves Th17 of lipids; attenuates tissue damage; promotes Treg/enforces Foxp3 expression; limits localized orbital remodelling	EAE, colitis models; AITD/HT/GO mechanistic rationale	[168–173]
	Mitochondrial metabolism and redox control	MitoTEMPO, NAC, selenium, ursolic acid	mROS; GPX4/ACSL4 axis	Attenuates mROS signalling; clears lipid peroxides; alleviates thyroid tissue destruction and restores Th17/Treg balance	Experimental autoimmune settings; mild GO patients; EAT models	[110, 174–176]
	Amino acid metabolism targeting	BPTES, CB-839, DON, V-9302, L‑kynurenine	GLS1, ASCT2 (glutaminolysis); kynurenine pathway	Reduces Th17 bias; interrupts TRAb production; shifts T-cell balance; counteracts HT destruction and GO remodelling	EAE, colitis models	[33, 177–180]
	Stromal output blockade	4-MU, ATF6 siRNA, METTL3/14 activator (C4)	HAS2/hyaluronan programme	Halts pathological HA accumulation; arrests orbital fibroblast proliferation; blocks PDGF-induced HAS2 expression	Patient-derived GO models; Parkinson’s disease model (C4 rationale)	[179, 181, 182]

AITD Autoimmune thyroid disease, GO Graves’ ophthalmopathy, HT Hashimoto’s thyroiditis, EAT Experimental autoimmune thyroiditis, EAE Experimental autoimmune encephalomyelitis, HA Hyaluronic acid, TgAb Thyroglobulin antibody, TRAb Thyrotropin receptor antibody, mROS mitochondrial reactive oxygen species, APCs Antigen-presenting cells, NSCLC Non-small-cell lung cancer, IDO Indoleamine 2,3-dioxygenase, Treg regulatory T cell

Conclusions and perspectives

The persistence of AITDs is driven not merely by aberrant immune activation, but also by progressive metabolic remodeling that stabilizes pathogenic T-cell states and compromises immune regulation. Within this framework, Th17 and Treg disequilibrium should not be viewed as a simple numerical imbalance accompanied by secondary metabolic changes. Intracellular metabolic rewiring actively shapes lineage commitment and functional plasticity. It also drives immune persistence. Glucose, lipid, mitochondrial, and amino acid metabolism act as interconnected regulatory layers that collectively influence inflammatory output, suppressive fitness, and lineage stability. Within the thyroid and orbit, this intracellular remodeling is further amplified by tissue-specific stressors, including excess iodine, oxidative burden, lactate retention, inflammatory cytokines, and stromal activation, thereby creating a self-reinforcing immune–metabolic environment that promotes ongoing disease progression.

At the same time, the current evidence base remains uneven across AITD subtypes and tissue settings. Metabolically driven Th17/Treg dysfunction is best documented in HT, with less complete and still-emerging mechanisms in GD and GO. This distinction is critical, as it precludes the uncritical extrapolation of general immunometabolic principles across all AITD subtypes. A more precise distinction is still needed between broad T-cell metabolic programs, disease-specific immune remodeling, and tissue-restricted mechanisms operating within the thyroid gland or orbital lesions. In particular, the spatial and temporal heterogeneity of metabolic reprogramming across disease initiation, progression, and chronic persistence remains insufficiently defined.

From a translational perspective, the significance of metabolic targeting lies not in suppressing a single pathway in isolation, but in interrupting coordinated pathogenic circuits linking immune polarization, redox stress, stromal remodeling, and tissue injury. This framework helps explain why upstream regulators such as mTOR, AMPK, HIF-1α, SIRT1, and AhR, as well as downstream metabolic effectors involving glycolysis, lipid handling, mitochondrial stress responses, amino acid utilization, and lactate-associated signaling, may all hold therapeutic relevance in different disease settings. However, the clinical promise of such strategies will depend on stronger causal validation, clearer disease-stage stratification, and improved identification of biomarkers that reflect immune–metabolic states in individual patients.

Future studies should therefore prioritize mechanistic validation in disease-relevant human systems and animal models, spatially resolved analysis of thyroidal and orbital metabolic niches, and integration of immunological, metabolic, and stromal readouts across different AITD subtypes and disease stages. A more refined understanding of these processes will not only clarify how immune imbalance is sustained in AITDs, but may also provide a stronger foundation for metabolism-informed precision intervention strategies.

Authors’ contributions

MZ, drafted the manuscript and drawn the table and figure. WW, YL, and SW conducted the project guidance. HP designed the study and revised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Jiangsu Provincial Medical Key Discipline Cultivation Unit (Grant No. JSDW202241), Zhenjiang science and technology planning project (Grant No. SH2025033).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.