Immune tolerance represents the foundational principle of self-non-self-discrimination and the long-term stability of immune homeostasis. It is not a static barrier but a dynamic, multilayered regulatory architecture encompassing central, peripheral, and tissue-level mechanisms that collectively prevent self-reactivity while preserving immune adaptability [19]. The failure of this architecture underlies the pathogenesis of autoimmune rheumatic diseases, transforming transient inflammation into self-sustaining pathology [20, 21]. Understanding tolerance as a distributed and metabolically governed network rather than a series of isolated checkpoints has become a defining shift in modern immunology.

At the central level, tolerance is established during lymphocyte ontogeny in the thymus and bone marrow. In the thymus, developing T cells are exposed to a broad repertoire of self-antigens presented by medullary epithelial cells under the control of the transcription factors autoimmune regulator (AIRE) and FEZ family zinc finger 2 (FEZF2). This process eliminates highly self-reactive clones through negative selection while allowing survival of low-affinity T cells capable of peripheral regulation [22, 23]. Similarly, in the bone marrow, autoreactive B cells undergo clonal deletion or receptor editing through secondary light-chain rearrangements, recalibrating their antigen specificity [24, 25]. However, both systems operate probabilistically, permitting the escape of partially autoreactive clones into the periphery. Thus, central tolerance provides a structural foundation but not a self-sufficient guarantee of immune restraint.

Peripheral tolerance expands this foundation into a context-dependent, adaptive network involving multiple immune and stromal players. Tregs form their keystone, exerting antigen-specific suppression through contact inhibition, IL-10, and TGF-β secretion, and control of dendritic cell maturation. Complementing Tregs are regulatory B cells (Bregs), tolDCs, and anti-inflammatory macrophages that collectively modulate effector activation thresholds [26, 27]. Peripheral tolerance is therefore achieved through active regulatory processes that maintain effector-regulatory equilibrium under fluctuating antigenic conditions.

Recent systems-level studies indicate that tolerance maintenance depends on coordinated regulatory programs across immune cell populations that preserve functional stability and prevent aberrant activation [28–30]. Disruption of these regulatory networks promotes a shift toward persistent inflammatory states, reflecting a breakdown in immune system coordination. The metabolic and epigenetic mechanisms underlying these processes are discussed in detail in Epigenetic-metabolic programming of immune cell fate.

Immune tolerance is also shaped by tissue microenvironments, where stromal and parenchymal cells actively regulate immune behavior. In the synovium, fibroblasts and endothelial cells provide signals that influence immune cell activation and differentiation [31, 32]. Under conditions of chronic stress, these stromal cells can adopt pathogenic phenotypes that promote sustained inflammation and disrupt local immune regulation [33, 34].

Temporal dynamics further define tolerance architecture. Resolution of immune responses requires coordinated resetting of regulatory programs, processes that are often impaired in chronic autoimmune disease [35, 36]. This results in immune cells remaining in partially activated states, a phenomenon referred to as tolerance inertia.

Collectively, these insights position immune tolerance as a self-organizing systems equilibrium regulated by dynamic interactions between immune cells and their environment. Rheumatic autoimmunity arises when this equilibrium loses resilience—when metabolic stress and inflammatory signaling reinforce one another to fix the network in an inflammatory attractor state that resists return to homeostasis. Recognizing tolerance as a dynamic architectural system reframes the therapeutic challenge: rather than suppressing inflammation downstream, interventions must restore the metabolic and epigenetic balance that underpins the architecture of immune self-regulation. This systems-based understanding provides the conceptual groundwork for exploring the cellular and molecular participants of tolerance networks in autoimmune rheumatology (Table 1). Table 1 summarizes these core cellular modules by linking each cell type to its dominant metabolic program, epigenetic features, and functional contribution to immune tolerance.

Immune tolerance emerges from the coordinated activity of lymphoid, myeloid, and stromal lineages, which together constitute a multicellular regulatory network maintaining self-non-self discrimination and tissue homeostasis. This distributed system integrates antigen sensing, cytokine signaling, and metabolic feedback to preserve immune equilibrium under constant environmental fluctuation [51]. T cells establish regulatory polarity and enforce suppression through forkhead box P3 (FOXP3)-governed transcriptional programs; B cells calibrate humoral memory and self-reactivity thresholds; myeloid antigen-presenting cells (APCs) define the cytokine and costimulatory context that determines whether antigen encounter yields activation or tolerance; and FLS, along with other stromal cells, provide the structural and metabolic scaffolding through which these immune interactions are spatially and energetically integrated [51–53]. The orchestration of these cellular modules depends on synchronized metabolic and epigenetic states, such as oxidative phosphorylation (OXPHOS), redox homeostasis, and chromatin accessibility, that collectively define the tolerogenic milieu. Disruption of any component propagates through the network, dismantling intercellular feedback loops and precipitating the self-reinforcing inflammatory circuits characteristic of autoimmune rheumatic diseases.

Although autoimmune rheumatic diseases share common features of immune tolerance breakdown, the underlying epigenetic-metabolic dysregulation is highly disease-specific, reflecting distinct cellular drivers, tissue environments, and dominant signaling pathways. Understanding these differences is critical for developing targeted and effective therapeutic strategies.

The collapse of immune tolerance in autoimmune rheumatic diseases represents a convergent endpoint of molecular, cellular, and metabolic perturbations that erode the stability of immune homeostasis. Rather than a single initiating lesion, these disorders reflect the progressive failure of a multilayered tolerance network in which immune and stromal compartments lose reciprocal regulation. The breakdown process unfolds through intertwined mechanisms: Chronic cytokine exposure and metabolic stress destabilize regulatory lineages; epigenetic drift fixes inflammatory transcriptional states; and tissue stroma acquires autonomous inflammatory memory [12, 80]. These transitions collectively rewire immune ecosystems from dynamically self-regulating networks into an inflammatory attractor state, in which cellular and metabolic activity becomes self-sustaining and resistant to reversal even after inflammatory pathways are pharmacologically suppressed.

In RA, tolerance failure localizes to the synovial microenvironment, where autoreactive T and B cells specific for citrullinated and carbamylated self-antigens accumulate within hypoxic, cytokine-rich niches. Persistent IL-6, TNF, and granulocyte-macrophage colony-stimulating factor (GM-CSF) signaling drives glycolytic reprogramming in T cells, macrophages, and FLS, reinforcing effector differentiation and tissue invasiveness [113, 114]. Metabolomic and single-cell multi-omic profiling reveal increased lactate and succinate levels, mitochondrial depolarization, and histone hyperacetylation at inflammatory loci, indicating that metabolic and chromatin states become coupled to inflammatory persistence [79, 113]. These findings are supported by integrative RA FLS datasets combining chromatin accessibility, histone-state mapping, transcriptomics, and chromosome conformation analyses, which demonstrate that inflammatory activation in fibroblasts is accompanied by stable regulatory rewiring rather than transient cytokine responsiveness [115, 116]. Even under biologic therapy, residual synovial T cells exhibit open chromatin at IFNG and IL17A loci, while FLS maintain hypomethylated, H3K27ac-enriched promoters controlling cytokine and matrix-remodeling genes. These durable transcriptional signatures define a form of tissue-imprinted inflammatory memory, explaining the relapse-prone nature of RA [117].

In SLE, tolerance breakdown is systemic and nucleic acid-driven. Defective clearance of apoptotic debris and aberrant activation of Toll-like receptors 7/9 (TLR7/9) in pDCs sustain type I IFN signaling, leading to continuous activation of autoreactive B and T cells. Epigenetic alterations—including DNA hypomethylation at IFN-stimulated genes and acetylation of histones at IFIT and MX dynamin-like GTPase 1 (MX1) loci—enhance transcriptional responsiveness to IFN signaling. Concurrent mitochondrial dysfunction in T cells and B cells promotes ROS accumulation, loss of NAD⁺ balance, and activation of cyclic GMP-AMP synthase-stimulator of interferon genes (cGAS-STING) pathways, further amplifying inflammation [118, 119]. These processes yield a self-reinforcing IFN network in which metabolic exhaustion and chromatin permissiveness coalesce into a stable pathogenic state, resistant to standard immunosuppressive therapy.

Sjögren’s syndrome exemplifies the interface between immune tolerance failure and epithelial stress. Salivary gland epithelial cells acquire features of professional APCs under chronic viral mimicry and IFN-driven signaling, expressing major histocompatibility complex (MHC) class II and co-stimulatory molecules. This epithelial transformation is accompanied by metabolic rewiring—enhanced glycolysis and ROS generation—and by epigenetic activation of IFN-response elements. The local microenvironment promotes ectopic germinal center formation, where Tfh/Tph and B cells interact to sustain autoantibody production against Ro/SSA and La/SSB antigens [120, 121]. Similarly, in SSc, endothelial injury and stromal activation initiate fibrosis through TGF-β-driven transcriptional programs. SSc, fibroblasts undergo metabolic reprogramming, including increased glycolytic and glutaminolytic activity, which supports myofibroblast differentiation, collagen production, and progressive extracellular-matrix accumulation [107]. These metabolic changes cooperate with profibrotic signaling pathways to stabilize fibroblast activation and promote persistent tissue remodeling, a clinically important process reflected in SSc progression assessment and trial-design considerations [122]. These disease-specific manifestations illustrate a convergent pathophysiological principle in which tolerance breakdown arises from the disruption of intercellular metabolic-epigenetic coherence, rather than from a single immune-activation event.

Environmental and systemic modifiers further compound this collapse. Hypoxia, oxidative stress, microbiome-derived metabolites, xenobiotic exposure, and nutrient imbalance intersect with genetic susceptibilities in human leukocyte antigen (HLA), PTPN22, DNA methyltransferase 1 (DNMT1), and MECP2 to perturb metabolic cofactors such as acetyl-CoA, α-ketoglutarate, and NAD⁺—critical regulators of chromatin-modifying enzymes. These perturbations shift global histone acetylation and DNA methylation patterns, reducing the energy and epigenetic flexibility required for immune adaptation [123, 124]. The resulting state is one of immunologic rigidity, in which feedback loops between metabolism, chromatin architecture, and cytokine signaling are locked into chronic activation.

Collectively, the breakdown of tolerance across autoimmune rheumatic diseases reveals a unifying systems pathology in which immune dysfunction emerges when cellular metabolism and epigenetic control lose synchrony across immune and stromal compartments. What begins as a transient inflammatory adaptation gradually becomes fixed as stable transcriptional reprogramming and persistent metabolic change. Reversing this process through metabolic normalization, epigenetic remodeling, and restoration of communication between mitochondria and the nucleus represents the central therapeutic challenge for the next generation of precision immunomodulatory interventions.

This progressive loss of immune tolerance provides the mechanistic context for the epigenetic remodeling discussed in the following section, where transient inflammatory signals become fixed as durable pathogenic transcriptional states.

DNA methylation serves as a long-term stabilizer of transcriptional identity by repressing gene promoters, insulating repetitive elements, and maintaining lineage fidelity. In autoimmune rheumatic diseases, perturbations in methylation patterns have emerged as a unifying mechanism underlying tolerance loss. CD4⁺ T cells in RA exhibit broad hypomethylation at IFN-responsive and proinflammatory loci, coincident with hypermethylation of genes central to metabolic restraint and immune regulation [e.g., FOXP3, IL-2 receptor alpha (IL2RA)]. This asymmetric remodeling skews transcription toward effector differentiation while impairing regulatory lineage stability [126, 127]. Similarly, B cells in SLE display promoter hypomethylation at CD11a, CD70, and CD40L, conferring autonomous activation potential even in the absence of antigenic stimulation [128].

Mechanistically, inflammatory cytokines such as IL-6, TNF, and type I IFNs converge on DNMT1 and TET dioxygenases, altering their activity through redox-dependent post-translational modifications. Elevated mitochondrial ROS oxidize 5-methylcytosine to 5-hydroxymethylcytosine, disrupting methylation fidelity and propagating transcriptional noise. These oxidative events further disrupt methylation fidelity and reinforce transcriptional instability under chronic inflammatory conditions [129, 130]. The cumulative outcome is a progressive loss of epigenetic precision, whereby gene expression becomes uncoupled from antigenic context and tolerance checkpoints collapse. This methylation drift transforms the adaptive immune system into a metastable state primed for autoreactivity (Figure 2), an effect increasingly recognized as a molecular bridge between chronic inflammation, metabolic stress, and heritable immune mispatterning.

Histone modifications act as rapid yet stable mediators of environmental and metabolic adaptation, determining whether chromatin regions remain transcriptionally accessible or repressed. In rheumatoid arthritis synovial tissue, single-cell transcriptomic and mass-cytometry analyses have identified inflammatory immune and stromal cell states linked to persistent synovial inflammation, while lupus lymphocytes show altered histone-modification patterns, including increased activating marks at immune-response genes [13, 131]. These inflammatory transcriptional and epigenetic patterns are associated with sustained expression of cytokine, chemokine, and tissue-remodeling programs, and may be influenced by metabolite-dependent chromatin-modifying enzymes, including HATs, histone deacetylases (HDACs), and HMTs [132, 133].

In chronically inflamed tissues, sustained inflammatory signaling promotes persistent chromatin accessibility and enhancer activation at pro-inflammatory gene loci, reinforcing stable transcriptional programs [134, 135]. This metabolic encoding of chromatin accessibility integrates mitochondrial dysfunction with nuclear gene regulation, effectively linking bioenergetic imbalance to transcriptional persistence. In parallel, EZH2, the catalytic subunit of the Polycomb repressive complex 2 (PRC2), is aberrantly activated in autoreactive B cells and FLS, resulting in selective silencing of anti-inflammatory and regulatory genes (e.g., SOCS1, CDKN1A) while maintaining proliferation and matrix-degrading activity. The coexistence of localized hyperacetylation and targeted Polycomb repression exemplifies the dual-axis chromatin remodeling that sustains pathogenic transcriptional programs in rheumatic autoimmunity [69, 136]. Collectively, these findings affirm that histone modification patterns in autoimmune rheumatic disease are not passive reflections of inflammation but active, metabolically inscribed determinants of tolerance fate [137].

Beyond linear chromatin modifications, autoimmune inflammation induces profound remodeling of the 3D genome, reconfiguring enhancer-promoter interactions, compartmental organization, and long-range chromosomal topology (Figure 3). This spatial reorganization converts transient transcriptional activation into structural persistence, thereby encoding inflammatory memory within nuclear architecture [138, 139]. High-resolution multi-omic studies now provide more direct support for this concept. In RA FLS, integrated high-throughput chromosome conformation capture (Hi-C), Capture Hi-C, assay for transposase-accessible chromatin with sequencing (ATAC-seq), and RNA-seq analyses have demonstrated TNF-responsive changes in chromatin organization, including altered A/B compartment activity, differential topologically associating domain (TAD) boundary behavior, and enhancer-promoter interactions linked to inflammatory gene regulation [116, 140]. Complementary epigenomic profiling in primary RA FLS, including ATAC-seq, histone-mark mapping, RNA-seq, and whole-genome bisulfite sequencing, has further shown that pathogenic fibroblasts acquire stable enhancer and promoter programs consistent with persistent inflammatory activation [115, 141]. More recent integrated RNA-seq and ATAC-seq studies also support joint-specific chromatin accessibility states in RA FLS, indicating that disease location and inflammatory context shape fibroblast regulatory architecture [6]. These altered contact maps result in the juxtaposition of distal enhancers and promoters previously segregated into repressive compartments, creating aberrant transcriptional hubs that sustain cytokine production independently of external stimuli. Figure 3 illustrates how higher-order chromatin reorganization, including TAD disruption, compartment shifts, enhancer rewiring, and mechanotransduction-linked remodeling, converts transient inflammatory signaling into structurally persistent inflammatory memory.

In SLE, B-cell chromatin exhibits nuclear repositioning of IFN-stimulated genes (ISGs) toward transcriptionally active A-compartments, thereby amplifying responsiveness to type I IFNs and establishing a self-reinforcing IFN signature [142, 143]. Similarly, in synovial and endothelial cells exposed to chronic hypoxia and mechanical stress, lamina-associated domains (LADs) are restructured, weakening their tethering to the nuclear periphery and permitting ectopic enhancer activation. This nuclear reorganization is stabilized by persistent transcription factor occupancy—most notably NF-κB, STAT1/3, BATF-IRF4, and AP-1 complexes—which anchor chromatin loops and maintain open chromatin conformation at inflammatory loci [144, 145]. Concomitantly, actin-lamin A/C coupling transduces mechanical stress to the nucleus, reinforcing pathological chromatin topology through mechanotransductive feedback.

These findings converge on a unifying principle in which inflammatory transcriptional states become structurally encoded once chromatin topology is reprogrammed to favor sustained accessibility of effector loci. The resulting nuclear architecture, marked by altered TAD insulation, loss of compartmental segregation, and pathological enhancer-promoter connectivity, renders effector programs refractory to classical regulatory cues. This form of topological fixation creates a higher-order layer of immune memory that transcends lineage boundaries, embedding the history of inflammatory stress directly into nuclear organization and sustaining autoimmunity even when antigenic stimulation is no longer present.

The recognition that aberrant epigenetic remodeling constitutes a primary driver of autoimmune persistence has reframed therapeutic strategy from transient immunosuppression toward durable immune reprogramming. In rheumatic autoimmunity, pathogenic chromatin landscapes are increasingly viewed as druggable architectures, dynamic and reversible determinants of lineage identity. Broad-spectrum epigenetic modulators, including HDAC inhibitors (vorinostat, givinostat), BET inhibitors (JQ1, PLX51107), and EZH2 inhibitors (tazemetostat), have demonstrated the capacity to compress hyper-accessible chromatin, silence inflammatory super-enhancers, and restore regulatory gene expression in preclinical models of arthritis and lupus [146–148]. These agents suppress cytokine transcription (IL-6, TNF, IFN-γ), attenuate fibroblast invasiveness, and partially reinstate FOXP3-dependent regulatory networks. However, their systemic administration remains constrained by dose-limiting hematopoietic toxicity, interference with antiviral defense, and incomplete cellular specificity, underscoring the need for precision-guided delivery. These approaches primarily target chromatin structure and transcriptional regulation, independent of upstream metabolic modulation.

Contemporary efforts, therefore, focus on cell-type-restricted and combinatorial strategies that exploit the intrinsic coupling between metabolism and chromatin state. Nanoparticle- or liposome-mediated targeting of epigenetic drugs to synovial fibroblasts, macrophages, or autoreactive lymphocytes can achieve locus-specific remodeling while sparing quiescent tissues [149, 150]. Parallel approaches employ ex vivo metabolic conditioning of regulatory T and B cells with cofactors such as NAD⁺ precursors, α-ketoglutarate, or AMP-activated protein kinase (AMPK) agonists to stabilize FOXP3- and BLIMP-1-dependent enhancer architectures before cellular reinfusion [151–153]. These manipulations enhance mitochondrial oxidative capacity and reinforce repressive chromatin marks, generating epigenetically fortified regulatory populations capable of resisting pro-inflammatory conversion in vivo. In parallel, CRISPR-based epigenome editing platforms using dCas9-fused acetyltransferase or demethylase domains allow locus-selective activation of tolerance genes (e.g., IL10, CTLA4) or silencing of effector loci (IL17A, IFNG), introducing a programmable dimension to immunotherapy [154, 155].

The most transformative advances lie in integrative metabolic-epigenetic interventions, which re-align energetic flux with chromatin control. Activation of AMPK, enhancement of sirtuin-dependent deacetylation, and restoration of NAD⁺ balance collectively re-establish transcriptional restraint, while inhibition of glycolytic checkpoints [lactate dehydrogenase A (LDHA), 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase-3 (PFKFB3)] mitigates acetyl-CoA-driven histone hyperacetylation [156, 157]. Combining these metabolic correctives with selective epigenetic modulators yields synergistic tolerance induction, dampening effector transcription while restoring mitochondrial and redox homeostasis. Such frameworks exemplify a paradigm shift from immune inhibition to epigenetic rehabilitation of tolerance, in which therapy seeks not to extinguish immune activity but to recalibrate its regulatory grammar.

Within precision-medicine paradigms, chromatin accessibility, histone-mark dynamics, and circulating metabolomic profiles are emerging as quantifiable biomarkers of therapeutic response, enabling real-time monitoring of tolerance restoration. Collectively, these innovations define an emerging frontier: the epigenetic re-education of the immune system, where targeted manipulation of chromatin-metabolism circuits holds the potential to convert episodic remission into sustained immunologic equilibrium.

Despite promising preclinical efficacy, the translational potential of BET inhibitors remains constrained by safety and tolerability concerns. Early-phase clinical trials in oncology have reported dose-limiting toxicities, including thrombocytopenia and broader hematopoietic suppression, reflecting the global role of BET proteins in transcriptional regulation [158, 159]. In addition, off-target epigenetic effects and limited therapeutic windows raise concerns regarding long-term use in chronic autoimmune diseases such as RA. These limitations highlight the need for improved target selectivity, optimized dosing strategies, and patient stratification to balance efficacy with safety.

The immune system relies on dynamic metabolic programming to balance energy expenditure with biosynthetic demand.

Immune activation in autoimmune rheumatic disease is associated with increased glycolytic flux, enhanced glucose uptake via glucose transporter-1 (GLUT1), and activation of key metabolic regulators, including mTORC1 and HIF-1α [160]. These changes support biosynthetic demand and inflammatory signaling. Concurrently, suppression of oxidative pathways, including OXPHOS and FAO, is frequently observed, particularly under conditions of hypoxia and chronic cytokine exposure. Disruption of AMPK signaling and mitochondrial regulatory networks further contributes to metabolic imbalance across immune and stromal compartments. These pathway-level alterations establish a bioenergetic environment that favors sustained inflammatory activity (Figure 4) [83, 161, 162]. As illustrated in Figure 4, these metabolic shifts are not uniform across diseases: glycolytic dominance is especially prominent in rheumatoid synovial fibroblasts and inflammatory macrophages, whereas glutaminolytic and mitochondrial rewiring are more prominent in lupus B cells and plasmablasts.

Multi-omic profiling of RA, SLE, and SSc reveals a convergent pattern of mitochondrial dysfunction and glycolytic bias. RA T cells exhibit diminished respiratory-chain capacity and increased mitochondrial ROS, which activate HIF-1α and sustain glycolytic flux despite ATP inefficiency. In RA synovial fibroblasts, TNF-related mitochondrial stress and altered mitophagy have been implicated in synovial inflammation, whereas mitochondrial abnormalities and ROS accumulation have been described in SLE lymphocytes [163, 164]. More broadly, mitochondrial DNA release can activate cGAS-STING-type I IFN pathways in inflammatory contexts, although the specific cell-type and disease-context relationships require careful distinction [165].

In SLE, B cells and plasmablasts rely on enhanced glutaminolysis and tricarboxylic acid (TCA)-cycle anaplerosis, producing excess α-ketoglutarate and fumarate that modulate histone demethylase activity and stabilize antibody-secreting programs. Systemic-sclerosis fibroblasts undergo a glycolytic conversion mediated by pyruvate dehydrogenase kinase 1 (PDK1)-dependent pyruvate shunting and TGF-β-driven repression of mitochondrial genes, culminating in fibrotic matrix overproduction [107, 166]. Collectively, these lesions define metabolic reprogramming as a core convergent mechanism of chronic inflammation, in which impaired mitochondrial quality control translates energetic instability into transcriptional fixation of pathogenic states.

Specific metabolites accumulate within inflamed tissues and directly modulate inflammatory signaling pathways. Lactate, abundant in hypoxic synovium, can signal through GPR81 and influence immune-cell function, including dendritic-cell maturation and inflammatory responses [167]. Succinate stabilizes HIF-1α by inhibiting prolyl hydroxylases, thereby amplifying IL-1β production in inflammatory macrophages [168]. In contrast, itaconate, generated via immune-responsive gene 1 (IRG1), exerts counter-regulatory control by alkylating KEAP1 and activating nuclear factor erythroid 2-related factor 2 (NRF2)-dependent antioxidant transcription, limiting ROS-driven cytokine cascades [169, 170].

Beyond carbohydrates and lipids, amino-acid metabolism forms a third regulatory tier that links nutrient flux to epigenetic fidelity. The methionine-one-carbon cycle generates SAM, the universal methyl donor for DNA and HMTs. Limitation of methionine or serine availability, a common feature of nutrient-restricted or inflamed microenvironments, reduces SAM pools and undermines methyltransferase activity, which promotes transcriptional noise. At the same time, glutamine-derived α-ketoglutarate fuels TET and lysine demethylase (KDM) dioxygenases and supports demethylation and chromatin plasticity [171, 172]. These amino-acid-dependent pathways illustrate the way nutritional context becomes epigenetically encoded within immune and stromal compartments.

Collectively, these metabolite-mediated processes integrate environmental sensing with nuclear regulation, allowing metabolic perturbations to be transcribed into chromatin landscapes. The reciprocal nature of this system explains why metabolic correction—through restoration of NAD⁺/NADH ratios or mitochondrial redox balance—can re-establish tolerance even without direct genomic intervention (Table 4).

Therapeutic redirection of cellular metabolism is emerging as a promising strategy for restoring immune homeostasis. Metformin, through AMPK activation, re-establishes mitochondrial integrity, limits ROS, and promotes FAO-driven Treg persistence [191]. Peroxisome proliferator-activated receptor gamma (PPAR-γ) agonists (pioglitazone, rosiglitazone) enhance oxidative metabolism and reduce fibroblast glycolysis [192], while NAD⁺ precursors (nicotinamide riboside, nicotinamide mononucleotide) reactivate sirtuin-dependent deacetylation and transcriptional restraint. mTOR inhibitors (rapamycin, everolimus) suppress anabolic signaling and have demonstrated clinical benefit in refractory lupus and vasculitis [193, 194].

Combination frameworks are under investigation that merge metabolic correction with epigenetic modulation. For example, SIRT1 activation paired with BET or HDAC inhibition can work synergistically to recalibrate immune networks [178]. Parallel translational approaches include bioenergetic conditioning of adoptive cell therapies in which ex vivo programming of CAR-Tregs or tolDCs under oxidative culture conditions increases mitochondrial mass and improves in vivo stability [195, 196]. Together, these interventions signal a shift from symptomatic immune suppression to metabolic restoration of tolerance architecture and align molecular therapy with systems-level homeostasis.

However, the clinical translation of NAD⁺-targeting strategies has yielded mixed results. While preclinical models suggest restoration of immune regulation, early clinical studies in RA have reported limited or no significant therapeutic benefit, with inconsistent effects on Treg function in human cohorts [177, 197]. These discrepancies likely reflect differences in disease stage, metabolic heterogeneity, and insufficient target engagement in vivo, underscoring the need for biomarker-guided patient selection and improved pharmacodynamic monitoring.

Across immune and stromal lineages, tolerance stability arises from the synchronization of energetic and redox states across tissues. Immune, endothelial, and mesenchymal cells engage in continuous exchange of metabolites, oxygen, and signaling intermediates that collectively maintain systemic homeostasis. In health, balanced oxidative metabolism preserves chromatin restraint and transcriptional adaptability. In disease, compartmentalized hypoxia, lactate buildup, and mitochondrial fragmentation disrupt this coherence, producing spatially heterogeneous yet functionally unified inflammation [198, 199].

Emerging spatial metabolomics and single-cell fluxomics have revealed that rheumatic tissues function as metabolic ecosystems, where dysregulation in one cellular subset propagates through shared NAD⁺, succinate, or ROS pools [46, 200]. Restoring these intercellular networks requires integrative strategies that align metabolic, epigenetic, and biomechanical correction. From a systems perspective, autoimmune rheumatic tissues function as interconnected metabolic environments in which alterations in one cellular compartment propagate through shared metabolite pools, including lactate, succinate, and ROS. Restoring metabolic balance across these networks represents a key requirement for durable disease control. These findings further support the concept that metabolic reprogramming is not cell-intrinsic alone but operates at the tissue and system levels.

Metabolism and chromatin form an interdependent regulatory circuit in which shared intermediates act as biochemical currencies that couple cellular energetics to gene regulation. At this interface, metabolic intermediates function not as isolated biochemical inputs but as integrative signals that couple cellular energy state to gene-regulatory architecture. Their significance lies in coordinating transcriptional stability across cells rather than in their individual enzymatic roles. Acetyl-CoA fuels HATs, SAM provides methyl donors for DNA and HMTs, α-ketoglutarate supports TET and Jumonji-domain demethylases, and NAD⁺ activates sirtuin deacetylases, collectively coupling metabolic flux to the chromatin landscape. In parallel, the promoters and enhancers of metabolic genes are themselves regulated by DNA methylation and histone marks, generating bidirectional feedback loops that integrate energy metabolism with transcriptional control [132, 203, 204].

Several metabolic enzymes are chromatin-associated. ATP-citrate lyase (ACLY) generates local acetyl-CoA at sites of active transcription; SIRT6 and poly(ADP-ribose) polymerase 1 (PARP1) function as chromatin-bound redox sensors; and EZH2, the catalytic core of PRC2, integrates SAM abundance to modulate histone methylation. Perturbations in nutrient availability, oxidative stress, or mitochondrial signaling destabilize this biochemical synchrony, producing transcriptional noise and lineage infidelity [205–207].

Emerging evidence further highlights nuclear-mitochondrial coordination as a stabilizing layer of this interface. Retrograde communication through NAD⁺/sirtuin-peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC-1α) pathways and mitochondrial transcription factor A (TFAM) links nuclear chromatin acetylation to mitochondrial gene expression, ensuring bidirectional alignment between nuclear and organellar epigenomes (Figure 5). When this coordination fails—as observed in T cells and FLS from RA and lupus—metabolic stress responses decouple from chromatin regulation, initiating persistent inflammatory transcription [208–210].

Fluctuations in cellular energy flux reshape chromatin accessibility in real time. Elevated acetyl-CoA promotes histone hyperacetylation at cytokine and chemokine loci, whereas NAD⁺ depletion limits sirtuin-mediated deacetylation, collectively sustaining open, transcriptionally active chromatin [211, 212]. Restoration of OXPHOS increases NAD⁺ availability, re-engaging sirtuin activity and favoring chromatin compaction. AMPK and mTOR function as metabolic rheostats for this process [213, 214]. AMPK activation promotes mitochondrial biogenesis and histone deacetylation, while mTORC1 activity maintains anabolic flux and supports chromatin relaxation [213, 215]. Importantly, these relationships operate as dynamic control systems, where fluctuations in energy flux continuously recalibrate chromatin accessibility rather than statically determining cell identity.

Redox signaling and oxygen tension exert additional control. ROS modify DNA and histone residues, perturb nucleosome positioning, and disrupt LADs. Chronic hypoxia promotes the migration of inflammatory gene clusters into transcriptionally active A-compartments [216, 217]. Transcription factors such as HIF-1α, PGC-1α, FOXP3, and c-Myc integrate metabolic sensing with chromatin remodeling, linking energy utilization to lineage-specific transcription [12, 173]. Through these coupled circuits, energy flux does not merely dictate cell fate but continuously reshapes the probability landscape of immune-state transitions.

Within inflamed rheumatic tissues, immune, stromal, and endothelial cells form interconnected metabolic-epigenetic ecosystems. Metabolites such as lactate, succinate, and NAD⁺ diffuse across the microenvironment and act as paracrine cues that synchronize chromatin states among neighboring cells [218]. Lactate derived from glycolytic fibroblasts suppresses dendritic-cell maturation yet stabilizes Th17 cells, while macrophage-derived succinate amplifies HIF-1α activity and IL-1β production in adjacent fibroblasts [168, 219]. Spatial metabolomics and single-cell ATAC-seq studies demonstrate coordinated enhancer activation across cellular compartments, revealing cross-cell synchronization of metabolic and chromatin landscapes [185, 220]. This intercellular exchange creates a distributed regulatory system in which no single cell type controls inflammation; instead, collective metabolic-epigenetic synchronization determines tissue-level behavior.

Tolerance failure arises when this synchronization collapses, fragmenting the system into locally stable but globally pathological states. Once immune and stromal cells become locked into mutually reinforcing glycolytic and inflammatory programs, the tissue transitions into a pathological attractor state that is self-sustaining, metabolically polarized, and epigenetically stabilized, persisting independently of antigenic stimulus.

Integration of ATAC-seq, chromatin immunoprecipitation sequencing (ChIP-seq), metabolomics, and transcriptomics now enables quantitative modeling of metabolic and epigenetic coupling. Dynamic Bayesian inference and network-entropy analyses identify energy-regulatory attractors that correspond to tolerant or inflammatory states. Machine-learning models applied to single-cell multi-omic datasets can predict fate transitions based on metabolic and epigenetic signatures. Machine-learning models applied to integrated transcriptomic, epigenomic, and metabolomic datasets may help identify immune-metabolic patterns, disease-associated cellular programs, and treatment-responsive states. Although broader multi-omic studies illustrate the value of computational integration across metabolism, immune features, and gene-regulatory datasets [221, 222], direct validation of specific metabolite ratios, chromatin marks, and effector-cell persistence in autoimmune rheumatic diseases remains an important area for future investigation.

These computational frameworks conceptualize tolerance as a low-entropy basin within immune-state space, characterized by energy efficiency and chromatin order, while inflammation corresponds to a high-entropy regime of metabolic inefficiency and transcriptional chaos. Such modeling provides a quantitative basis for forecasting how targeted interventions might shift cellular ensembles back toward the low-entropy, tolerant state. These models collectively support a view of immune regulation as a dynamic systems process governed by state transitions rather than fixed cellular identities.

At the systems level, therapeutic intervention can be reframed as the deliberate re-synchronization of coupled metabolic-epigenetic networks rather than the suppression of isolated inflammatory pathways. Agents that concomitantly modulate both metabolic and chromatin axes such as AMPK activators in combination with BET or HDAC inhibitors, NAD⁺ augmentation coupled with EZH2 modulation, or PPAR-γ agonists integrated with chromatin re-educators, embody rational designs aimed at restoring durable immune equilibrium [223, 224]. Rather than transiently silencing cytokine effector cascades, these interventions seek to reconstruct the energetic and epigenomic architectures that physiologically uphold tolerance, thereby enabling stable remission through the re-establishment of regulatory setpoints.

At the diagnostic frontier, the convergence of epigenomic and metabolomic profiling is giving rise to a new generation of integrative biomarkers capable of capturing the multidimensional state of immune regulation. Quantitative indices such as the acetyl-CoA/NAD⁺ ratio, SAM/S-adenosylhomocysteine (SAH) flux, mitochondrial redox potential, chromatin accessibility signatures, and single-cell transcriptional states offer mechanistic readouts of metabolic-epigenetic coherence [59, 225]. However, most of these candidate biomarkers remain research tools rather than clinically validated assays, and none are currently standardized for routine clinical decision-making in autoimmune rheumatology [225, 226]. Their current limitations include inter-platform variability, lack of assay harmonization, uncertain threshold definitions, and insufficient prospective validation across disease stages and treatment settings [59, 226, 227]. These constraints underscore the need for multicenter validation studies and clinically deployable biomarker panels before precision tolerance-reprogramming strategies can be broadly implemented.

At the conceptual level, this framework reframes autoimmunity as a disorder of epigenetic-metabolic desynchronization, wherein the loss of temporal and functional coherence between cellular energy homeostasis and gene-regulatory architecture entrenches the system within a chronic, self-sustaining inflammatory state. Reinstating tolerance thus necessitates the coordinated realignment of oxidative metabolism, redox equilibrium, and chromatin restraint across immune and stromal compartments. Within this systems-theoretic framework, immune homeostasis emerges not as a static equilibrium but as an actively regulated, programmable attractor state, a coupled configuration of metabolic efficiency and epigenetic stability that can, in principle, be rationally recalibrated through precision interventions informed by the logic of metabolic-epigenetic integration.

From this perspective, durable remission represents not the elimination of inflammation, but the restoration of system-wide coherence across metabolic and epigenetic dimensions.

Contemporary immunotherapy is shifting from nonspecific immunosuppression toward immune re-education, a strategy that seeks to retrain rather than silence pathogenic circuits. Conventional agents that block cytokine signaling or lymphocyte proliferation attenuate inflammation transiently but rarely restore the regulatory architecture required for durable tolerance. In contrast, tolerance reprogramming refers to the deliberate recalibration of immune homeostasis through coordinated modulation of metabolic, epigenetic, and transcriptional checkpoints that determine cellular identity and plasticity.

At the mechanistic level, tolerance reprogramming realigns the molecular interfaces linking energy metabolism with chromatin state. The AMPK-mTOR-sirtuin axis serves as a central rheostat integrating nutrient and redox cues to dictate whether immune cells adopt glycolytic effector profiles or oxidative, quiescent phenotypes [228, 229]. Parallel remodeling of histone acetylation, DNA methylation, and enhancer accessibility consolidates these metabolic states into heritable transcriptional programs [230, 231]. Restoring alignment across these layers re-establishes oxidative metabolism, NAD⁺ sufficiency, and repressive chromatin landscapes that stabilize FOXP3⁺ Tregs, IL-10⁺ Bregs, and reparative stromal subsets.

From a systems perspective, tolerance reprogramming constitutes an integrative therapeutic philosophy that unifies pharmacologic, biologic, and cellular interventions under a shared mechanistic framework (Figure 6). As summarized in Figure 6, translational implementation of immune tolerance reprogramming requires coordinated integration of therapeutic strategies, tissue-level target engagement, multi-omic biomarker frameworks, and computational modeling to guide adaptive treatment optimization. Rather than targeting single cytokines or receptors, emerging strategies act on the network topology of the immune system, coordinating chromatin accessibility, metabolic flux, and signaling feedback loops to restore systemic equilibrium. This conceptual transition—from suppression to re-education—defines the translational frontier of modern rheumatology and immunotherapy.

The successful translation of these strategies will depend not only on mechanistic efficacy but also on validated biomarkers, rational patient stratification, long-term safety assessment, and clinically feasible delivery systems.

A new generation of therapeutics exploits epigenetic and metabolic coupling to re-establish tolerance (Figure 6). Epigenetic modulators such as HDAC inhibitors (e.g., givinostat, vorinostat) suppress pro-inflammatory gene networks and favor expansion of Tregs. BET inhibitors (e.g., PLX51107) selectively dismantle super-enhancers controlling TNF, IL6, and CXCL loci, thereby dampening sustained transcriptional activation. BET inhibitors have been investigated for suppressing inflammatory transcriptional programs, whereas EZH2 inhibitors and low-dose DNMT-targeting strategies are being explored as approaches to modulate pathogenic chromatin states and regulatory gene programs in selected experimental contexts [232–235]. Collectively, these agents exemplify how selective chromatin remodeling can re-educate immune networks toward regulatory stability.

In parallel, metabolic immunomodulators are being repositioned as tools to tune energy flux and redox homeostasis. Metformin activates AMPK and enhances OXPHOS, indirectly augmenting sirtuin-mediated histone deacetylation. PPAR-γ agonists recalibrate lipid metabolism and repress NF-κB-driven transcription, while NAD⁺ precursors such as nicotinamide riboside and nicotinamide mononucleotide restore mitochondrial function and sustain regulatory-cell persistence [213, 236, 237]. Agents targeting the mTOR-AMPK-SIRT network thus provide metabolic leverage points through which cellular bioenergetics can be re-aligned with transcriptional tolerance programs.

Increasingly, combinatorial regimens are being designed to exploit this reciprocity. Metabolic pre-conditioning through AMPK activation or NAD⁺ supplementation may augment the durability of chromatin-directed therapies by stabilizing repressive histone marks and limiting effector relapse [11, 35]. Nevertheless, rational combination therapy design remains a major translational challenge. Key unresolved issues include optimal sequencing of metabolic and epigenetic agents, avoidance of synergistic toxicities, and alignment of specific combinations with disease-specific pathogenic mechanisms [238, 239]. For example, combinations that may be suitable in stromal-dominant RA may not translate directly to IFN-driven SLE or fibrosis-dominant SSc [240, 241]. These considerations highlight the need for biomarker-guided trial design and adaptive dosing strategies rather than empiric combination approaches (Table 5). Figure 6 should be read in conjunction with the therapeutic limitations summarized in Table 5, as translational feasibility depends not only on mechanistic rationale but also on safety, scalability, and disease-specific applicability. Table 5 summarizes the major therapeutic classes currently under investigation, together with their mechanistic rationale, development stage, and key translational constraints relevant to clinical implementation.

Despite this promise, targeted delivery platforms face important clinical limitations. These include variable synovial penetration, low bioavailability in inflamed tissues, batch-to-batch manufacturing complexity, and high production costs [264, 265]. In addition, delivery efficiency may differ substantially across disease sites and tissue compartments, limiting the consistency of therapeutic exposure [239, 264, 266]. Next-generation biomaterial and nanoparticle systems may improve localization and reduce systemic toxicity, but their clinical scalability and regulatory feasibility remain incompletely established.

Importantly, therapeutic strategies must be aligned with disease-specific epigenetic-metabolic architectures (Disease-specific epigenetic-metabolic dysregulation in autoimmune rheumatology), as targeting stromal metabolism in RA, IFN signaling in SLE, epithelial activation in Sjögren’s syndrome, or fibroblast reprogramming in SSc may require distinct intervention approaches.

Table 6 summarizes how epigenetic and metabolic therapeutic strategies may be differentially prioritized across RA, SLE, Sjögren’s syndrome, and SSc according to dominant disease mechanisms.

Cellular immunotherapies embody tolerance reprogramming in a living form. Treg therapies, particularly CAR-Tregs, are engineered to recognize disease-specific antigens and deliver localized immune control through contact-dependent suppression and IL-10 or TGF-β secretion [267]. Preclinical models of RA and SLE show that CAR-Tregs maintain FOXP3 expression and resist pro-inflammatory conversion when metabolically preconditioned to favor OXPHOS and adequate NAD⁺ reserves [268, 269]. Engineered tolDCs, modified to express IL-10 and IDO1, promote peripheral tolerance by presenting self-antigens under metabolically quiescent conditions that favor Treg induction [270, 271].

Beyond cellular re-engineering, mesenchymal stem cell (MSC)-derived exosomes and extracellular vesicles function as nanoscale mediators of metabolic and epigenetic recalibration. These vesicles deliver microRNAs, metabolites, and chromatin-modifying enzymes that remodel the metabolic landscape of recipient immune cells, fostering regulatory phenotypes within inflamed tissues. Metabolic pre-conditioning of MSCs through AMPK activation or NAD⁺ enrichment enhances the tolerogenic composition of their secreted vesicles by reinforcing FOXP3 and BLIMP-1 chromatin programs [272, 273]. Early-phase clinical studies in RA, SLE, and type 1 diabetes have confirmed safety and feasibility, signaling a shift toward living biologics capable of implementing immune reprogramming in situ [274].

Despite encouraging preclinical results, significant challenges remain for CAR-Treg therapies. Recent studies indicate that CAR-Tregs may exhibit limited in vivo persistence and phenotypic instability within inflammatory microenvironments, with potential conversion toward effector-like states under sustained cytokine exposure [84, 275]. These risks raise concerns regarding durability, safety, and functional fidelity, particularly in chronically inflamed tissues such as rheumatoid synovium. Addressing these barriers will require improved CAR design, stabilization of FOXP3 expression, and microenvironment-aware engineering strategies.

Rapid advances in multi-omic profiling are redefining how immune tolerance can be quantified, monitored, and predicted. The next generation of biomarkers moves beyond cytokine titers or autoantibody levels toward integrated indicators of the epigenetic-metabolic architecture that defines immune equilibrium.

Chromatin accessibility profiling by ATAC-seq and Cleavage Under Targets and Tagmentation (CUT&Tag) enables the derivation of tolerance signatures that reflect the re-closure of inflammatory enhancers and the stabilization of FOXP3- and BLIMP-1-associated regulatory elements. Integration of these data with histone-mark landscapes (H3K27ac/H3K27me3 balance) and DNA-methylation clocks yields dynamic indices of epigenomic restraint and functional maturity of regulatory networks [276, 277].

Metabolomic and redox biomarkers complement these readouts by linking bioenergetic efficiency to chromatin control. Ratios such as acetyl-CoA: NAD⁺, SAM: SAH, and mitochondrial ROS flux provide quantitative proxies for histone-acetyltransferase and demethylase activity. Circulating metabolites that reflect NAD⁺ salvage, β-oxidation, and tricarboxylic-acid-cycle activity correlate with T-regulatory-cell frequency and remission probability in early interventional trials, underscoring their translational potential [278, 279].

Minimally invasive liquid-biopsy approaches extend these tools to clinical practice. Exosomes and extracellular vesicles isolated from plasma or synovial fluid contain microRNAs, histone fragments, and metabolic cofactors that mirror the transcriptional and energetic state of tissue-resident immune and stromal cells. Longitudinal profiling of vesicle cargo enables dynamic assessment of tolerance restoration without invasive sampling [280].

Emergent computational biomarkers employ machine-learning models trained on integrated epigenomic, transcriptomic, and metabolomic datasets to infer chromatin openness, metabolic flux, and treatment responsiveness. These predictive systems support individualized modeling of tolerance trajectories. Their adoption, however, requires rigorous regulatory standardization, encompassing sample harmonization, normalization pipelines, and cross-platform reproducibility. As these digital readouts mature, they are poised to become integral to adaptive clinical-trial design and real-time immunotherapy monitoring.

Tolerance reprogramming converges with the principles of systems medicine, which interprets immune function as an emergent property of dynamically coupled molecular and cellular networks. Within this framework, AI-driven patient stratification and multi-omic immune atlases allow mechanistic subtyping of autoimmune disorders that were previously defined solely by clinical phenotype, supporting biomarker-guided therapeutic decision-making and precision immunomodulation [281].

By jointly mapping metabolic state and chromatin configuration, clinicians can distinguish patient subsets dominated by glycolytic inflammation from those exhibiting mitochondrial insufficiency or epigenetic rigidity. This precision enables rational matching of metabolic modulators or chromatin-targeted agents to underlying molecular pathophysiology. The resulting concept of immune-network re-entrainment envisions therapy as restoration of synchronized metabolic and transcriptional oscillations across immune and stromal compartments, rather than unidirectional cytokine blockade.

This systems perspective also aligns tolerance reprogramming with the biology of regeneration and aging. Restoration of redox equilibrium, mitochondrial quality control, and chromatin fidelity parallels the rejuvenation of immune plasticity observed in caloric-restriction and sirtuin-activation paradigms. Integrating metabolic and senolytic interventions with tolerance reprogramming could therefore unify the therapeutic treatment of chronic inflammation, autoimmunity, and age-associated immune decline within a single conceptual scaffold of immunologic rejuvenation.

Translation of such complexity demands robust ethical and regulatory governance. Off-target chromatin effects, potential epigenetic inheritance of therapeutic modifications, and the persistence of engineered or metabolically conditioned cells necessitate proactive oversight [282]. In parallel, digital biomarkers introduce additional challenges concerning algorithmic transparency, data privacy, and interpretability [283]. Establishing validated standards and oversight mechanisms will be essential to ensure that precision immunology evolves responsibly and maintains public trust.

Emerging technologies now make it feasible to design immune tolerance as a programmable property of living systems. Synthetic-biology platforms are constructing gene-circuit modules capable of sensing inflammatory metabolites and autonomously activating regulatory transcriptional programs. These self-regulating constructs enable cells to maintain homeostasis within defined biochemical thresholds [284], offering an unprecedented degree of precision in controlling immune dynamics.

Simultaneously, longitudinal digital-twin frameworks are being developed to simulate immune reprogramming in silico. By integrating continuous omic, metabolic, and clinical data streams, these models can predict therapeutic efficacy, optimize dosing, and anticipate relapse before clinical manifestation. The resulting feedback systems transform immunotherapy into an adaptive, data-driven process of continual recalibration.

Future progress will depend on cross-disciplinary integration spanning metabolism, neuroscience, synthetic biology, and regenerative medicine. Such collaboration will clarify how systemic energy balance, neuronal signaling, and tissue repair interface with immune homeostasis. The synthesis of these domains will define the foundations of tolerance engineering. This includes the deliberate design of immunologic stability through coordinated manipulation of metabolic, epigenetic, and signaling networks.

Ultimately, immune tolerance should be viewed not as a passive outcome but as an actively maintainable physiological state that can be induced, stabilized, and monitored through rationally designed interventions. Achieving this vision would signal a paradigm shift in rheumatology and immunotherapy, transforming disease management from symptomatic control to the proactive cultivation of immune resilience.

While emerging epigenetic, metabolic, and cellular therapies offer promising avenues for restoring immune tolerance, several translational challenges remain. Across therapeutic classes, discrepancies between preclinical efficacy and clinical outcomes highlight the complexity of human autoimmune disease.

Epigenetic modulators, including BET and HDAC inhibitors, demonstrate potent anti-inflammatory effects in preclinical models; however, their broad regulatory roles introduce risks of off-target effects, hematologic toxicity, and limited therapeutic windows [158, 285]. Achieving selective modulation of pathogenic programs without disrupting essential cellular functions remains a central challenge.

Metabolic interventions, such as NAD⁺ augmentation or AMPK activation, have shown the capacity to rebalance immune cell function in experimental systems. However, clinical responses have been variable, likely reflecting inter-patient metabolic heterogeneity, disease chronicity, and incomplete target engagement [177]. These findings emphasize the need for biomarker-driven stratification and real-time metabolic monitoring.

Cellular therapies, including CAR-Treg approaches, provide high specificity but face challenges related to stability, persistence, and scalability. Inflammatory microenvironments may compromise regulatory phenotypes, while manufacturing complexity and cost remain significant barriers to widespread implementation [286–288].

Importantly, these emerging strategies must be evaluated in the context of existing standard-of-care therapies, including biologics and JAK inhibitors, which offer well-established efficacy and safety profiles. While next-generation approaches may provide deeper or more durable immune reprogramming, their comparative effectiveness, long-term safety, and cost-efficiency remain to be fully established.

Together, these considerations highlight that successful translation will require not only mechanistic innovation but also careful integration of safety, patient selection, and therapeutic positioning within current clinical frameworks.

The translation of epigenetic-metabolic tolerance reprogramming into clinical practice faces several unresolved barriers. First, biomarker validation remains limited. Although candidate measures such as chromatin accessibility profiles, metabolic ratios, and multi-omic signatures offer mechanistic resolution, most remain research-grade tools without standardized clinical thresholds, assay harmonization, or prospective validation [289, 290].

Second, patient stratification remains underdeveloped. It is still unclear how best to identify patients most likely to benefit from metabolic correction, epigenetic modulation, or cellular therapies. Distinguishing glycolysis-dominant, mitochondrial-dysfunction-dominant, IFN-driven, or fibrosis-dominant disease states will likely be essential for successful therapeutic selection [289, 291].

Third, rational combination therapy design remains a major challenge. While combined metabolic and epigenetic interventions are mechanistically attractive, their implementation requires careful optimization of dose, sequence, timing, and toxicity management. Matching specific combinations to disease-specific pathogenic architectures will be critical to avoid empiric and potentially harmful regimens [238, 239, 290].

Fourth, long-term safety requires far greater attention, particularly for epigenetic modulators. Chronic manipulation of chromatin regulators raises concerns regarding off-target transcriptional remodeling, malignant transformation, impaired host defense, and potentially durable epigenetic effects in non-target tissues [238, 239, 292]. These concerns are especially relevant in non-malignant rheumatic diseases, where therapies may be used for prolonged periods.

Finally, targeted delivery remains a major practical obstacle. Nanoparticle and ligand-directed platforms may reduce systemic toxicity, but limitations in tissue penetration, bioavailability, manufacturing complexity, and cost continue to constrain clinical translation [265, 293]. Together, these barriers indicate that successful translation will require not only mechanistic innovation but also clinically scalable biomarker platforms, disease-informed patient stratification, safety monitoring, and practical delivery solutions.

Beyond mechanistic and trial-design challenges, real-world implementation of tolerance-reprogramming strategies will depend on practical clinical considerations. Compared with established standard-of-care therapies such as biologics and JAK inhibitors, many emerging epigenetic, metabolic, and cellular interventions currently carry greater uncertainty regarding cost-effectiveness, scalability, and health-system integration [294, 295].

Cellular therapies such as CAR-Tregs and tolDCs may offer highly specific and durable immunoregulation, but their manufacturing requirements, individualized production workflows, and reimbursement challenges may limit broad accessibility [288, 295, 296]. Similarly, widespread use of multi-omic profiling for patient stratification remains constrained by cost, turnaround time, technical complexity, and uneven access across care settings [289, 290].

These realities suggest that the future adoption of tolerance-reprogramming therapies will depend not only on efficacy and safety, but also on whether they can be deployed in a cost-conscious, scalable, and equitable manner within routine rheumatology practice.