Autoimmunity in ME/CFS
Evidence increasingly supports autoimmune mechanisms in at least a subset of ME/CFS patients.
1 Autoantibodies Identified
1.1 Anti-Nuclear Antibodies
Anti-nuclear antibody (ANA) prevalence is elevated in ME/CFS, with 15–25% of patients testing positive compared to 5–10% in healthy individuals (Nishikai 2007). Various ANA patterns are observed, with positive ANA possibly indicating general immune dysregulation rather than a specific autoimmune disease.
Whether elevated ANA titres in ME/CFS patients represent a pathogenic autoimmune process, an epiphenomenon of systemic immune dysregulation, or a prognostically relevant subgroup marker remains unresolved.
Given the strong correlation between calprotectin and NET burden (Hetland et al. 2022) and the role of NETosis in post-viral thrombo-inflammation (Krinsky et al. 2023), serial calprotectin measurement before and after standardized exertion protocols (e.g., 2-day CPET) could test whether NET release mediates post-exertional malaise. If calprotectin rises predictably 6–24 hours post-exertion and correlates with subsequent symptom severity, it would provide an inexpensive, objective, real-time biomarker for the signature feature of ME/CFS.
Son et al. (Son et al. 2023) tracked 106 COVID-19 convalescent patients longitudinally and found that anti-nuclear antibody (ANA) positivity at 3 months post-infection predicted persistent symptoms (fatigue, dyspnoea) at 12 months. Specific antibodies targeting U1-snRNP and SS-B/La correlated with fatigue severity; elevated TNF-\(\alpha\) independently predicted both ANA persistence and symptom continuation. The autoantibody profile resembled early connective tissue disease, suggesting post-viral autoimmunity as a distinct endotype of long COVID with implications for ME/CFS subgroups. Study: (\(n=106\), prospective cohort 3–12 months post-infection, European Respiratory Journal, certainty: 0.65, partially replicated). The temporal relationship between NETosis and autoantibody emergence is now supported by converging evidence: viral-triggered NETosis generates citrullinated autoantigens (Monsalve et al. 2025), and anti-nuclear autoantibodies detectable at 3 months post-infection—the expected timeframe for a germinal centre reaction initiated by NET-derived antigens—predict symptom persistence at 12 months (Son et al. 2023). The causal relevance of these autoantibodies is established by passive transfer experiments: purified IgG from Long COVID patients reproduces neurological symptoms in recipient mice (Guedes de Sá et al. 2024). Taken together, these findings position NETosis as the upstream driver and autoantibodies as the downstream effectors of a unified post-viral autoimmune cascade.
1.2 G-Protein-Coupled Receptor (GPCR) Autoantibodies
GPCR autoantibodies represent one of the most well-studied autoantibody classes in ME/CFS, with substantial evidence for their pathogenic role. The B cell abnormalities described in Section Research Priorities for Validating the CD8+ T-Cell Mitochondrial Exhaustion Hypothesis likely contribute to autoantibody generation. For comprehensive coverage of GPCR autoantibodies—including initial discovery, validation across cohorts, correlation with symptom severity, downstream mechanisms, and therapeutic targeting through immunoadsorption, plasma cell depletion, and direct neutralization—see the detailed discussion in Section Research Priorities for Validating the CD8+ T-Cell Mitochondrial Exhaustion Hypothesis.
1.3 Anti-Neuronal Antibodies
Antibodies targeting nervous system components have been identified in ME/CFS, including anti-ganglioside antibodies, antibodies against voltage-gated ion channels, and anti-neuronal surface antigen antibodies. These autoantibodies may contribute to the neurological symptoms observed in the condition.
2 Autoimmune Mechanisms
2.1 Molecular Mimicry
Molecular mimicry occurs when structural similarity between pathogen and self-antigens leads antibodies or T cells generated against an infection to cross-react with self-tissues. This phenomenon has been documented for several viruses associated with ME/CFS and may explain the link between infection and subsequent autoimmunity.
2.2 Epitope Spreading
Epitope spreading occurs when tissue damage exposes new antigens to the immune system. The initial immune response causes tissue injury, releasing self-antigens that trigger new autoimmune responses. This leads to progressive expansion of autoimmune targets over time.
2.3 Loss of Self-Tolerance
Loss of self-tolerance occurs when regulatory mechanisms fail. Treg dysfunction permits autoreactive cells to escape suppression, while B cell tolerance checkpoints fail to eliminate autoreactive B cells. Chronic inflammation further promotes autoimmunity by creating a permissive environment for autoimmune responses.
Molecular mimicry, epitope spreading, and loss of self-tolerance are well-characterized mechanisms in established autoimmune diseases (SLE, rheumatoid arthritis, type 1 diabetes). Their role in ME/CFS is inferred by analogy, not demonstrated:
- No specific molecular mimicry target (pathogen epitope \(\to\) self-antigen cross-reaction) has been identified in ME/CFS.
- Epitope spreading has not been documented longitudinally in ME/CFS cohorts.
- Treg dysfunction is reported in ME/CFS but has not been causally linked to loss of tolerance versus being a downstream consequence of chronic immune activation.
- The proportion of ME/CFS patients with a genuinely autoimmune etiology (as opposed to immune dysregulation without autoimmunity) remains unknown.
If ME/CFS neutrophils share the stress-responsive mitochondrial failure phenotype described in depression (Cullen et al. 2026), NETosis initiation may proceed normally but completion may fail due to energy insufficiency. This “frustrated NETosis” would release citrullinated autoantigens (Monsalve et al. 2025) without effective pathogen trapping, reconciling the suppressed neutrophil activation proteome observed by Hoel et al. (Hoel et al. 2026) with persistent NET-driven thrombo-inflammation (Krinsky et al. 2023) (Romano et al. 2022).
Testable prediction: Ex vivo NETosis assays in ME/CFS neutrophils should show reduced NET completion rates (lower percentage of fully extruded NETs) but equivalent or elevated citrullinated histone release compared to healthy controls.
Treatment implication: If confirmed, DNase I therapy (Veras et al. 2023) would be particularly important for clearing incompletely formed NETs that the body cannot resolve normally; DNase deficiency documented in severe post-viral states by Garcia et al. (Garcia et al. 2024) provides mechanistic rationale for this approach.
Study: (cross-paper synthesis, certainty: 0.35, untested).
3 Developmental Immune Tolerance and Recovery
The autoimmune mechanisms described above—molecular mimicry, epitope spreading, and loss of self-tolerance—operate in both pediatric and adult ME/CFS patients. Yet pediatric patients recover at dramatically higher rates (54–94%) than adults (\(\leq\) 22%), often despite similar autoantibody profiles at disease onset. This paradox suggests that developing immune systems may possess unique mechanisms for eliminating aberrant immune memory that adult systems lack.
Certainty: 0.50. Pediatric recovery from ME/CFS may be facilitated by developmental immune tolerance mechanisms that actively delete or reprogram aberrant immune memory cells—a process termed “immune memory pruning.” The developing immune system, particularly during puberty and adolescence, appears to undergo quality control checkpoints that can eliminate autoreactive B cells, exhausted T cells, and pathological memory populations. The certainty level reflects: (1) well-characterized developmental changes in peripheral B and T cell tolerance mechanisms; (2) evidence for thymic output through adolescence; (3) documented sex hormone effects on immune regulation; (4) however, the specific mechanisms enabling clearance of ME/CFS-associated autoantibodies remain unclear; (5) the distinction between resolution via immune pruning versus other pediatric recovery mechanisms (glial plasticity, HSC regeneration, recovery capital) cannot yet be resolved.
Pediatric recovery from ME/CFS may be facilitated by developmental immune tolerance mechanisms that actively delete or reprogram aberrant immune memory cells—a process we term “immune memory pruning.”
Mechanistic basis:
The developing immune system is not merely a smaller version of the adult system; it is qualitatively different in its capacity for self-correction. Several mechanisms may contribute to immune memory pruning:
Peripheral B cell tolerance checkpoints. Autoreactive B cells that escape central tolerance in the bone marrow normally undergo peripheral tolerance mechanisms including anergy (functional unresponsiveness), receptor editing (modification of the B cell receptor to eliminate autoreactivity), and deletion (apoptosis of autoreactive cells) (Meffre and O’Connor 2024) (Mohtashami and Zúñiga-Pflücker 2024). These peripheral checkpoints are most active during immune development and may decline with age. Recent evidence demonstrates that impaired clearance of autoreactive B cells due to disruption of negative selection barriers causes autoantibody-producing B cells to mature and subsequently differentiate into plasma cells (Meffre and O’Connor 2024). In children with ME/CFS, autoantibodies generated during acute illness may be subject to these active tolerance mechanisms, leading to gradual elimination of autoreactive B cell clones. In adults, with attenuated peripheral tolerance, these clones persist.
Thymic contribution to T cell repertoire. The thymus remains highly active through adolescence, continuously generating naive T cells that have undergone rigorous negative selection against self-antigens. This ongoing thymic output may dilute and eventually replace dysfunctional T cell populations (exhausted cells, aberrantly activated cells) that accumulate during ME/CFS. Adult thymic involution eliminates this regenerative capacity, leaving dysfunctional T cell populations to persist.
Pubertal immune reorganization. Puberty involves substantial reorganization of immune function, driven by sex hormones and growth factors. This reorganization period may include “quality control” checkpoints that assess immune memory and eliminate populations that fail to meet tolerance criteria. Adolescents who develop ME/CFS before completing this reorganization may benefit from these checkpoints; those who develop ME/CFS after puberty have already passed this window.
Regulatory T cell plasticity. Tregs in children show greater plasticity and proliferative capacity than adult Tregs. Pediatric ME/CFS patients may mount more effective regulatory responses that suppress autoreactive populations, eventually leading to their deletion through lack of antigenic stimulation.
Testable predictions:
This hypothesis generates several falsifiable predictions:
Autoantibody trajectory differences: In recovering pediatric patients, autoantibody titers (particularly GPCR autoantibodies) should decline over time, whereas in non-recovering adults, titers should remain stable or increase. Longitudinal measurement of autoantibody levels in pediatric versus adult cohorts would test this prediction.
Tolerance gene signatures: Peripheral blood from recovering pediatric patients should show gene expression signatures of active tolerance mechanisms—upregulation of genes involved in receptor editing (RAG1, RAG2), anergy (CBLB, ITCH), and deletion (FAS, BIM)—that are absent in adults. RNA-seq comparing pediatric recoverers versus non-recoverers versus adults could identify these signatures.
Thymic output markers: Recent thymic emigrants (CD31+ naive CD4 T cells, T cell receptor excision circles or TRECs) should correlate with recovery probability. Patients with higher thymic output, regardless of age, should have better prognosis.
B cell subset dynamics: Recovering patients should show declining frequencies of CD21lo atypical memory B cells (associated with chronic immune activation) and autoantibody-secreting plasmablasts, with increasing naive B cell proportions.
Pubertal timing effects: Among pediatric patients, those who develop ME/CFS pre-pubertally (Tanner stages 1–2) should have higher recovery rates than those developing ME/CFS post-pubertally (Tanner stages 4–5), as the former still have pubertal immune reorganization ahead of them.
Treatment implications:
If immune memory pruning explains pediatric recovery advantage, several therapeutic strategies could be explored to induce analogous processes in adults:
Tolerance-inducing therapies. Low-dose antigen administration can induce peripheral tolerance in autoimmune conditions. If specific autoantigens driving ME/CFS are identified (e.g., GPCR peptides), tolerance induction protocols could be developed.
Regulatory T cell therapy. Expansion and infusion of autologous Tregs could enhance regulatory mechanisms that suppress autoreactive populations.
Timed B cell depletion. Rituximab showed mixed results in ME/CFS trials, but the immune memory pruning hypothesis suggests that B cell depletion might be more effective in younger patients or when combined with therapies that prevent re-emergence of autoreactive clones during B cell reconstitution. The timing of B cell depletion relative to immune developmental stage may be critical.
Thymic rejuvenation. Experimental approaches to restore thymic function (IL-7 therapy, sex steroid ablation, thymic transplantation) could potentially restore the naive T cell output that enables immune repertoire renewal.
Limitations and caveats:
This hypothesis remains speculative and requires validation. Alternative explanations for the pediatric recovery advantage exist, including greater metabolic reserves, neural plasticity, psychosocial factors, and ascertainment bias (milder cases preferentially diagnosed in children). The hypothesis does not explain why some children do not recover or why rare adults do recover. Additionally, even if the hypothesis is correct, translating developmental tolerance mechanisms into adult therapeutics presents substantial challenges.
Enterochromaffin-Vagal Pathway: Gut-Immune-Brain Axis
Recent work by Wirth and Scheibenbogen (2025) on the enterochromaffin-vagal pathway provides novel insights into gut-immune-brain interactions that may influence autoimmune processes in ME/CFS. The full mechanistic evidence chain—butyrate-enterochromaffin serotonin-vagal signaling—is developed in Section Replication Status: Not Yet Replicated of Chapter Gastrointestinal and Microbiome Dysfunction; here we focus on its immune dysregulation implications:
Certainty: 0.35. The enterochromaffin-vagal pathway may contribute to immune dysregulation in ME/CFS through multiple mechanisms . Gut dysbiosis with reduced butyrate production impairs enterochromaffin cell serotonin synthesis , leading to reduced vagal afferent signaling (Barton et al. 2023) . This vagal dysfunction may disrupt immune regulation through:
Potential immune mechanisms:
- Reduced cholinergic anti-inflammatory pathway activity: Vagal efferents inhibit cytokine production via the cholinergic anti-inflammatory pathway (Barton et al. 2023). Reduced vagal tone may impair this regulatory mechanism.
- Altered gut immune homeostasis: Vagal dysfunction may disrupt gut-associated lymphoid tissue (GALT) regulation, contributing to intestinal inflammation and bacterial translocation .
- Neuro-immune feedback disruption: The vagus nerve provides afferent input to brain regions that regulate immune function, including the hypothalamus and brainstem .
- MCAS bidirectional amplification: Mast cells and enterochromaffin cells interact in the gut lamina propria . Reduced vagal tone removes cholinergic inhibition of mast cell degranulation (Barton et al. 2023), potentially amplifying mast cell activation syndrome (MCAS) in susceptible patients . Conversely, chronic histamine elevation may plausibly impair enterochromaffin function and worsen vagal afferent input (speculative; no direct ME/CFS evidence). This bidirectional loop may contribute to the observed co-occurrence of MCAS and autonomic dysfunction in ME/CFS .
Implications for autoimmunity:
- Provides a mechanistic link between gut dysbiosis and systemic immune dysregulation
- Suggests that vagal stimulation therapies may have immunomodulatory effects
- Explains how gut-directed therapies (probiotics, butyrate) might influence systemic inflammation
- Offers testable hypotheses for future research on gut-immune-brain interactions
Limitations:
- Preprint status (not peer-reviewed)
- Direct evidence for vagal-immune interactions in ME/CFS is limited
- Causality and directionality remain uncertain
- Clinical implications require validation in controlled studies
Certainty: 0.25. EBV reactivation is documented in ME/CFS (Hwang et al. 2023). EBV infection induces interferon-\(\gamma\) (Huang et al. 2020), which upregulates IDO2, diverting tryptophan into the kynurenine pathway (Phair, Davis, and Kashi 2019) and depleting the substrate available for enterochromaffin serotonin synthesis . Simultaneously, kynurenine pathway products—particularly quinolinic acid—may impair mitochondrial function via PARP-mediated NAD+ depletion (Kavyani et al. 2022) (Dehhaghi et al. 2022) , creating both substrate depletion and cellular dysfunction in enterochromaffin cells.
This chain (EBV reactivation \(\to\) IFN-\(\gamma\) \(\to\) IDO2 \(\to\) tryptophan depletion \(\to\) impaired enterochromaffin serotonin \(\to\) vagal afferent dysfunction) may explain why post-EBV ME/CFS patients often have prominent autonomic dysfunction (Hwang et al. 2023). If correct, patients with high EBV antibody titers or active reactivation would show higher kynurenine/tryptophan ratios, lower platelet serotonin, and more severe autonomic impairment (lower HRV) than non-EBV cases. 5-HTP supplementation, which bypasses the IDO2 diversion point , may be particularly relevant for this subgroup.
Limitations: The EBV \(\to\) IDO2 \(\to\) gut-vagal chain involves multiple extrapolation steps, each with limited direct ME/CFS evidence. Quinolinic acid effects on enterochromaffin cells have not been studied in ME/CFS. Causal directionality between EBV reactivation and IDO2 upregulation in ME/CFS is unestablished.
Certainty: 0.35. This is a novel mechanistic reinterpretation of existing GPCR autoantibody data. While consistent with published correlational and therapeutic findings , the core claim—that autoantibodies increase effort-per-output rather than simply blocking or activating receptors—has not been directly tested. Functional assays distinguishing partial agonism from complete blockade in ME/CFS are absent, and the “effort-to-output ratio” metric does not yet have a validated measurement instrument.
The standard model of autoantibody pathogenesis posits binary effects: agonist antibodies overstimulate receptors, antagonist antibodies block them. However, emerging evidence suggests GPCR autoantibodies in ME/CFS may operate through a more subtle mechanism—increasing the physiological “effort” required per unit of functional output rather than preventing output entirely.
Conceptual Framework.
Consider two scenarios:
Complete blockade: A receptor antagonist prevents signal transduction. To achieve a given output, the system compensates by upregulating receptor expression, increasing ligand production, or activating alternative pathways. This produces a stable new equilibrium with normal output at higher baseline cost.
Partial interference: An autoantibody binds receptors with partial agonist/antagonist activity (Wallukat et al. 2021), causing stochastic signal degradation. Some signals succeed, others fail. The system cannot establish stable compensation because the interference is probabilistic rather than deterministic. To achieve reliable output, the system must increase signal redundancy, pathway cross-talk, and error-correction mechanisms—dramatically raising metabolic cost per successful signal.
The autoantibody inefficiency hypothesis proposes that ME/CFS GPCR autoantibodies function primarily through scenario 2: they increase the effort-to-output ratio rather than blocking output capacity.
Mechanistic Basis.
Multiple mechanisms could contribute to effort-per-output elevation:
Partial agonism and desensitization. GPCR autoantibodies can act as partial agonists, activating receptors less efficiently than endogenous ligands while simultaneously triggering desensitization cascades (Wallukat et al. 2021). This creates a state of chronic low-level activation paired with reduced receptor availability. Functional output requires overcoming both the partial antagonism and the desensitization state, necessitating increased ligand release, receptor cycling, and downstream amplification.
Receptor internalization and trafficking dysregulation. Autoantibody binding can trigger receptor internalization through arrestin-mediated pathways . Unlike normal agonist-induced internalization (which promotes receptor recycling), autoantibody-mediated internalization may produce abnormal trafficking patterns, sequestering receptors in non-productive compartments. The cell must continuously synthesize new receptors to maintain surface expression, increasing basal metabolic demand.
Compensatory pathway activation. When primary signaling pathways operate inefficiently due to autoantibody interference, cells activate compensatory mechanisms . In monocytes, GPCR autoantibody binding triggers production of MIP-1\(\delta\), PDGF-BB, and TGF-\(\beta\) 3—cytokines that activate alternative inflammatory and stress-response pathways. These compensatory cascades consume ATP, generate reactive oxygen species, and demand continuous cellular attention, raising the metabolic cost of maintaining homeostasis.
Stochastic signaling noise. If autoantibodies cause intermittent rather than consistent receptor modulation, downstream systems cannot adapt through simple upregulation. Instead, they must implement redundancy: multiple parallel pathways, error-checking mechanisms, and repeated signal verification. This transforms efficient point-to-point signaling into expensive consensus-based communication, analogous to the difference between a direct phone call and a committee meeting requiring majority vote.
Predictions and Testable Hypotheses.
The inefficiency model generates distinct predictions from the simple blockade/activation model:
Effort-symptom dissociation: Autoantibody-positive patients should show higher perceived effort for equivalent objective tasks compared to autoantibody-negative patients with similar disability levels.
Titer correlation with effort, not capacity: GPCR autoantibody titers should correlate more strongly with effort-per-task measures than with absolute functional capacity. High-titer patients might maintain near-normal performance briefly but report extreme subjective effort and experience rapid decompensation.
Metabolic cost elevation: During controlled tasks, autoantibody-positive patients should show elevated oxygen consumption, lactate production, or other metabolic markers for equivalent work output compared to autoantibody-negative patients.
Immunoadsorption reduces effort-per-output: The primary benefit of immunoadsorption should be reduced subjective effort for equivalent tasks, measurable before changes in absolute capacity. Stein et al. showed SF-36 improvements , but effort-normalized metrics were not reported.
PEM pattern differences: If autoantibodies increase baseline effort, PEM in autoantibody-positive patients should follow an effort-accumulation model (gradual exhaustion of compensatory reserves) rather than a damage-repair model (discrete injury requiring recovery).
CNS vs peripheral manifestations: If the inefficiency mechanism operates primarily through adrenergic and muscarinic receptors, symptoms should concentrate in systems dependent on rapid, high-frequency receptor signaling: autonomic regulation, attention/vigilance, fine motor control. Gross muscle strength might be relatively preserved despite profound fatigue.
Existing Evidence Alignment.
Several published findings align with the inefficiency model:
Correlation patterns: Sotzny et al. found GPCR autoantibody levels correlated with fatigue severity, muscle pain, cognitive impairment, and autonomic dysfunction —all effort-dependent symptoms that worsen with sustained activity.
Heterogeneous treatment response: In the Stein et al. immunoadsorption study, 70% responded but 30% did not . The inefficiency model predicts that patients with primary autoantibody-driven disease should respond dramatically, while those with additional pathology may show limited benefit.
Delayed recovery after antibody removal: While autoantibody titers drop immediately (77% reduction in Stein et al.), clinical improvement evolves over weeks to months. This delay is inconsistent with simple receptor unblocking but consistent with gradual normalization of compensatory mechanisms.
Monocyte reprogramming persistence: Hackel et al. showed GPCR autoantibodies reprogram monocyte cytokine production . This inflammatory signature might persist even after antibody removal if monocytes have been epigenetically reprogrammed.
Clinical Implications.
If the inefficiency model is correct, treatment strategies should focus on:
Early intervention: Before compensatory mechanisms become maladaptive, autoantibody removal might produce rapid recovery.
Combination therapy: Immunoadsorption or plasma cell depletion (daratumumab ) to remove autoantibodies, paired with therapies supporting cellular energy metabolism (CoQ10, NADH) to offset elevated effort costs during recovery.
Effort-based rather than capacity-based pacing: For autoantibody-positive patients, effort-tracking (heart rate variability, subjective effort scales) might be more predictive of crashes than activity level alone.
Patient selection for trials: The inefficiency model predicts a subgroup—autoantibody-positive, relatively acute onset, without severe mitochondrial damage—should show dramatic response to antibody removal.
Relationship to Other Hypotheses.
The autoantibody inefficiency hypothesis complements rather than contradicts other ME/CFS mechanisms:
Vicious cycles: Elevated effort-per-output increases metabolic demand, potentially triggering mitochondrial stress, oxidative damage, and further dysfunction . The inefficiency becomes self-reinforcing.
Autonomic dysfunction: If autoantibodies target adrenergic receptors with partial agonist effects, the autonomic nervous system operates in a state of chronic low-level activation paired with reduced responsiveness.
Immune memory: The immune memory pruning hypothesis (Hypothesis Immune Memory Pruning in Development) suggests pediatric patients might clear autoreactive plasma cell populations, while adults cannot—explaining age-related prognosis differences in autoantibody-driven disease.
Limitations and Uncertainties.
Mechanistic heterogeneity: Different autoantibodies likely have different effects. Lumping all GPCR autoantibodies together may obscure subtype-specific mechanisms.
Assay limitations: As noted by Vernino et al. , current ELISA assays measure antibody binding but not functional effects. Functional assays measuring receptor activation, desensitization kinetics, and downstream signaling are needed.
Causal uncertainty: Autoantibodies correlate with symptom severity but correlation is not causation. They may be a consequence rather than a cause of ME/CFS.
Quantification challenges: “Effort per output” is difficult to measure objectively. Developing validated effort-efficiency metrics is essential for testing this hypothesis.
Alternative explanations: Elevated perceived effort could arise from CNS dysfunction (altered interoception, motivation circuits) rather than peripheral receptor interference.
Despite these limitations, the inefficiency hypothesis provides a coherent framework for understanding why approximately 30% of ME/CFS patients have elevated GPCR autoantibodies, why these correlate with specific symptom domains, why immunoadsorption produces gradual rather than immediate improvement, and why outcome heterogeneity exists.