Cytokines and Inflammatory Mediators
Cytokines are signaling proteins that coordinate immune responses. Cytokine abnormalities in ME/CFS have been extensively studied, though findings vary considerably across studies.
1 Pro-inflammatory Cytokines
1.1 Interleukin-1 (IL-1)
IL-1 is a master regulator of inflammation, with IL-1\(\beta\) often elevated in ME/CFS. Its effects include fever, fatigue, muscle breakdown, and the acute phase response. Notably, IL-1 produces “sickness behavior” in the central nervous system that closely resembles ME/CFS symptoms, and levels may correlate with symptom severity.
1.2 Interleukin-6 (IL-6)
IL-6 has both pro- and anti-inflammatory effects and is frequently elevated in ME/CFS, particularly in early illness. This cytokine induces acute phase proteins, promotes B cell differentiation, and crosses the blood-brain barrier to affect central nervous system function. IL-6 correlates with fatigue in other conditions, suggesting a mechanistic link to this cardinal ME/CFS symptom.
1.3 Tumor Necrosis Factor-Alpha (TNF-α)
TNF-\(\alpha\) is a central inflammatory cytokine elevated in some ME/CFS studies. It causes fatigue, malaise, and cognitive dysfunction while also affecting mitochondrial function and promoting muscle wasting (cachexia) via ubiquitin-proteasome pathway activation. A meta-analysis of 42 studies confirmed significantly elevated TNF-\(\alpha\) in CFS versus controls (effect size 0.274, \(p < 0.001\)), alongside elevated IL-2, IL-4, TGF-\(\beta\), and CRP, but found 12 other measured proteins did not differ between groups (Strawbridge et al. 2019). Variable findings across studies may reflect patient heterogeneity within the ME/CFS population (Strawbridge et al. 2019). Nutritional anti-catabolic strategies that inhibit the ubiquitin-proteasome pathway — such as HMB, a leucine metabolite that preserved lean mass in a bed rest RCT — may partially offset TNF-\(\alpha\)-driven muscle catabolism in immobilized patients; this is discussed in Section Exercise Dose-Response in ME/CFS Is Bifurcated: Maintenance Below Threshold, Deterioration Above (Deutz et al. 2013).
1.4 Interferons
Type I interferons (IFN-\(\alpha\), IFN-\(\beta\)) are antiviral cytokines elevated in some ME/CFS patients. These interferons cause profound fatigue (as known from their therapeutic use in other conditions) and may indicate ongoing viral activation. Interferon-induced gene expression patterns have been observed in ME/CFS. Type II interferon (IFN-\(\gamma\)) activates macrophages and promotes Th1 responses, though findings in ME/CFS are variable; levels may be elevated or reduced depending on disease stage.
1.5 Interleukin-2 (IL-2)
IL-2 is a critical cytokine for T cell function and immune regulation:
- T cell proliferation: Essential for clonal expansion of activated T cells
- Regulatory T cell maintenance: Required for Treg development and suppressive function
- NK cell activation: Enhances NK cell cytotoxicity
- Memory T cell formation: Supports long-term immunity
- Therapeutic use: Low-dose IL-2 used in autoimmune diseases to boost Tregs; high-dose IL-2 used in cancer immunotherapy
IL-2 signaling requires three receptor subunits (CD25/CD122/CD132) and activates JAK/STAT pathways. Dysregulation can lead to either immune deficiency (insufficient IL-2 or receptor expression) or autoimmunity (Treg dysfunction). Recent evidence suggests IL-2 pathway abnormalities in ME/CFS (see hypothesis below).
1.6 Cytokine Patterns Across Disease Duration
Hornig et al. identified distinct immune signatures in ME/CFS that vary dramatically by disease duration. In a cohort of 298 ME/CFS patients and 348 healthy controls, early-stage patients (illness duration \(<\) 3 years, n=52) showed prominent activation of both pro- and anti-inflammatory cytokines, with elevated levels of IL-1\(\alpha\), IL-8, IL-10, IL-12p40, IL-17F, IFN-\(\gamma\), CXCL1 (GRO-\(\alpha\)), CXCL9 (MIG), and IL-5 (all p\(<\) 0.05, FDR-corrected). A 17-cytokine panel distinguished early ME/CFS from controls with high diagnostic accuracy.
In stark contrast, patients with longer disease duration (\(>\) 3 years, n=246) had cytokine profiles that normalized to control levels, with no significant differences for most cytokines. This finding represents the first large-scale evidence that ME/CFS immunopathology evolves over time, potentially from initial immune activation to exhaustion or adaptation.
Single study (Hornig 2015, n=298 total, but only n=52 in the critical early-stage subgroup). The duration-dependent pattern has not been independently replicated. The small early-stage sample limits confidence in the 3-year inflection point. Cross-sectional design cannot distinguish true longitudinal evolution from cohort effects. Longitudinal studies tracking individual patients over time are needed.
Implications of Duration-Dependent Cytokine Changes
The Hornig et al. findings have profound implications:
- Therapeutic windows: Early-stage disease may respond better to immunomodulatory therapies targeting active inflammation
- Study heterogeneity: Failure to stratify by disease duration explains contradictory findings in previous cytokine studies
- Biomarker utility: Cytokine profiling is most useful as a diagnostic tool within the first 3 years of illness
- Disease progression: Normalization may reflect immune exhaustion, regulatory adaptation, or shift to different pathological mechanisms
Hornig et al. found that illness duration was more strongly predictive of cytokine patterns than symptom severity in their cross-sectional analysis, suggesting that immune changes primarily reflect disease stage . However, this group-level observation does not preclude severity-related gradients within early-stage or late-stage patients (see following section).
1.7 Cytokine-Severity Correlations
Montoya et al. demonstrated dose-response relationships between cytokines and symptom severity in 192 ME/CFS patients compared to 392 healthy controls. Although only two cytokines differed overall between patients and controls (TGF-\(\beta\) higher and resistin lower), 17 cytokines showed statistically significant upward linear trends correlating with disease severity. Thirteen of these 17 are proinflammatory, including CCL11 (Eotaxin-1), CXCL1 (GRO-\(\alpha\)), CXCL10 (IP-10), IFN-\(\gamma\), IL-4, IL-5, IL-7, IL-12p70, IL-13, IL-17F, G-CSF, GM-CSF, and TGF-\(\alpha\).
This dose-response relationship—rather than simple binary patient-control comparison—provides stronger evidence that immune activation tracks with symptom burden. The findings suggest cytokine profiling could stratify patients for clinical trials and identify individuals likely to benefit from anti-inflammatory therapies.
The cytokine-severity correlation pattern is directionally consistent with Hornig 2015’s findings, providing convergent support from a different analytic approach. However, specific cytokine identities differ between studies, and neither has been independently replicated with the same panel and methodology.
Notably, CXCL9 (MIG) inversely correlated with fatigue duration, showing higher levels in early disease and lower levels in chronic disease . This continuous inverse correlation mirrors Hornig’s group-level finding of elevated early-disease cytokines, providing convergent support from a different analytic approach (within-group correlation versus cross-sectional comparison of early vs. late subgroups).
1.8 Sex-Specific Cytokine Dysregulation
Recent work by Che et al. in a large multi-center cohort revealed that hyperinflammatory cytokine responses are particularly pronounced in women over 45 years of age with diminished estradiol levels. Using multi-omics analysis including microbial stimulation assays (heat-killed Candida albicans), the study demonstrated exaggerated production of IL-6 and other proinflammatory cytokines in ME/CFS patients, with responses amplified before and especially after exercise.
The sex- and hormone-specific pattern provides mechanistic insight into the female predominance of ME/CFS (approximately 3:1 female-to-male ratio) and suggests potential therapeutic interventions, such as estrogen supplementation for post-menopausal women with evidence of immune hyperactivation. This sex-specific finding complements the NIH deep phenotyping study’s observation of distinct immune abnormalities in male versus female patients (Walitt et al. 2024), underscoring that ME/CFS pathophysiology may differ fundamentally between sexes.
1.9 Integrated Model: Duration, Severity, and Sex
Combining findings from Hornig , Montoya , and Che , an integrated model of cytokine dysregulation emerges:
- Disease duration: Early disease (\(<\) 3 years) shows high cytokines at the group level; late disease (\(>\) 3 years) shows normalized group-level cytokines
- Disease severity: Within patient cohorts, severe patients show higher proinflammatory cytokines than mild patients through dose-response relationships
- Sex and hormones: Women, particularly post-menopausal women with low estradiol, show more pronounced immune activation
Reconciling Duration and Severity Effects
The Hornig and Montoya findings are not contradictory but complementary. Hornig examined group differences between early-stage and late-stage patients, finding that the early-stage group as a whole had elevated cytokines. Montoya examined severity gradients within their cohort (which included both early and late patients), finding that more severe patients had higher cytokines regardless of duration. These observations can coexist: early disease may be characterized by overall immune activation (shifting the entire distribution upward), while severity effects create gradients within both early and late subgroups. The interaction between duration and severity has not been directly tested in a study stratified by both factors simultaneously.
Clinical Application
This integrated model suggests personalized treatment approaches, though these represent theoretical predictions requiring validation:
- Early + severe + female + low estradiol: Predicted to have highest cytokines; most likely to benefit from immunomodulatory therapies (extrapolated from individual studies)
- Late + severe + female: May have severity-driven inflammation despite duration-dependent normalization; immune status assessment needed
- Late + mild + male: Predicted to have lowest cytokines; may require therapeutic strategies targeting mechanisms beyond acute immune activation
- All other phenotypes: Require individualized immune profiling before treatment selection
The implications of patient heterogeneity for treatment stratification and the concept of distinct ME/CFS subtypes are discussed in Chapter Integrative Models and Multi-System Pathophysiology, Section Research Questions and Future Directions.
No study has yet examined all three factors (duration, severity, sex/hormones) simultaneously in a fully stratified design. The clinical predictions above are extrapolations from separate studies and require prospective validation.
The integrated duration/severity/sex model combines findings from Hornig 2015, Montoya 2017, and Che 2025—three studies using different cohorts, cytokine panels, analytic methods, and stratification criteria. No study has simultaneously examined all three factors. The clinical predictions (e.g., “early + severe + female + low estradiol” = highest cytokines) are logical extrapolations, not empirical findings. The interaction effects between duration, severity, and sex may be non-additive in ways not predictable from individual studies.
A 41-study meta-analysis of inflammatory markers in dementia found that excluding patients with comorbid psychiatric and physical inflammatory conditions removed the classically reported elevations of IL-6 and TNF-\(\alpha\) in Alzheimer’s disease, while elevations of IL-17A, IL-1\(\alpha\), IL-10, G-CSF, GM-CSF, and IL-3 survived this exclusion (Kuring et al. 2026). This is a methodological caution that applies to any chronic condition including ME/CFS: if included ME/CFS populations are not screened for common comorbid inflammatory conditions (arthritis, diabetes, chronic infections, autoimmune disease), apparent IL-6 or TNF-\(\alpha\) signals may be inflated by comorbidity rather than reflect disease-inherent inflammation. The finding also reinforces the specificity of the ME/CFS cytokine signal: the markers that survive comorbidity-exclusion in dementia (IL-17A, G-CSF, GM-CSF, IL-1\(\alpha\)) overlap with the ME/CFS severity-correlated markers reported above (G-CSF, GM-CSF, IL-17F, IL-17A-adjacent Th17 axis (Kuring et al. 2026)), suggesting a possible shared Th17/myeloid-axis signal — but the dementia markers are not directly transferable to ME/CFS, and replication in comorbidity-screened ME/CFS cohorts is needed before any such inference is drawn.
(Certainty: 0.45 — methodological parallel from a dementia population; specific marker transfer to ME/CFS unsupported. Severity coverage: unknown — dementia populations not ME/CFS-severity-stratified.)
Consequence: For patients and readers, this is a caution about how inflammatory-marker research can overstate a signal that is really driven by other coexisting illnesses; it does not change any current ME/CFS treatment or diagnostic practice.
IL-2 as Emerging Biomarker Target
Certainty: 0.45. Two independent methodological approaches (extracellular vesicle proteomics and epigenetic chromosome conformation) converge on IL-2 pathway dysregulation, lending moderate confidence. However, whether this reflects a causal role or an epiphenomenon of chronic immune activation, and whether the two findings reflect the same underlying process, remain unresolved.
Two independent methodologies implicate the IL-2 pathway in ME/CFS, though through different mechanisms. Giloteaux et al. found significantly elevated IL-2 specifically in extracellular vesicles from ME/CFS patient plasma (n=49 patients, n=49 controls; q=0.007 after multiple comparison correction), with proinflammatory cytokines CSF2 and TNF\(\alpha\) correlating with physical and fatigue symptom severity. Independently, Hunter et al. used epigenetic profiling (EpiSwitch® technology) of chromosome conformation in 47 ME/CFS patients versus 61 controls, identifying IL-2 signaling among dysregulated pathways in a 200-marker panel (92% sensitivity, 98% specificity in validation).
The convergence—extracellular vesicle cytokine content in one study, epigenetic regulation in another—suggests the IL-2 pathway warrants focused investigation. However, several questions remain: Do elevated IL-2 levels in extracellular vesicles reflect the same process as epigenetic dysregulation of IL-2 signaling? Are ME/CFS cells producing excess IL-2, responding abnormally to normal IL-2, or both? Does IL-2 dysfunction contribute causally to symptoms or merely correlate with disease? Further studies measuring IL-2 receptor expression, downstream signaling (JAK/STAT pathway), and functional T-cell responses to exogenous IL-2 could clarify the pathway’s role and therapeutic potential.
2 Anti-inflammatory Cytokines
2.1 Interleukin-10 (IL-10)
IL-10 is a potent immunosuppressive cytokine with variable findings in ME/CFS. Levels may be elevated (potentially reflecting an attempt to control inflammation) or reduced (which would permit inflammation to continue). IL-10 is important for resolving immune responses and is produced by regulatory T cells and other cell types.
2.2 Transforming Growth Factor-Beta (TGF-β)
TGF-\(\beta\) is the most consistently elevated cytokine in ME/CFS. A systematic review of 38 papers and 77 cytokines found TGF-\(\beta\) elevated in 5 of 8 studies that measured it (63%)—the highest consistency of any cytokine (Blundell et al. 2015). The Montoya et al. Stanford study (192 patients, 392 controls, 51 cytokines) confirmed this elevation (\(p = 0.0052\)), with TGF-\(\beta\) and resistin being the only two analytes consistently different at the population level .
TGF-\(\beta\) has immunosuppressive and tissue remodeling functions. This elevation may represent an attempt to control inflammation, though chronic elevation can promote fibrosis. TGF-\(\beta\) is also important for regulatory T cell development. Critically, TGF-\(\beta\) is a potent suppressor of NK cell cytotoxicity through multiple established mechanisms: direct transcriptional repression of perforin (approximately 57% reduction) and granzyme B (approximately 38% reduction) via Smad pathway signaling, inhibition of mTOR (reducing NK cell metabolic capacity), and downregulation of activating receptors NKG2D and NKp30.
Roerink et al. (2018) raised a serious methodological challenge: TGF-\(\beta\) 1 measurements in ME/CFS may be confounded by platelet contamination during sample processing (Roerink et al. 2018). Platelets contain 40–100 times more TGF-\(\beta\) 1 than other cell types. Lower centrifuge g-force produces higher platelet activation (p-selectin correlated \(r = 0.79\) with TGF-\(\beta\) 1), which artificially inflates measured concentrations. After controlling for platelet contamination in their cohort, no authentic TGF-\(\beta\) 1 elevation distinguished ME/CFS patients from controls. This methodological concern applies to all prior studies that did not standardize centrifugation protocols or measure platelet activation markers. The Montoya 2017 study used standardized multiplex plasma assays and remains the strongest evidence that TGF-\(\beta\) elevation is real, but even this single-site study cannot fully exclude centrifugation effects.
Certainty: 0.40. Mechanistically well-supported by cancer immunology literature; not yet tested in ME/CFS.
The two most consistently replicated immune findings in ME/CFS are (1) elevated TGF-\(\beta\) and (2) reduced NK cell cytotoxicity. These are routinely treated as independent observations. We propose they may represent a single finding and its downstream consequence: chronic TGF-\(\beta\) elevation suppresses NK cell function.
Mechanistic basis (established in cancer immunology):
- TGF-\(\beta\) directly represses perforin and granzyme B transcription via Smad/ATF1
- TGF-\(\beta\) inhibits mTOR in NK cells, reducing metabolic capacity for sustained killing
- TGF-\(\beta\) downregulates NKG2D and NKp30 activation receptors
- Prolonged TGF-\(\beta\) exposure may cause persistent NK dysfunction through epigenetic remodeling of IRF, T-bet, and EOMES binding sites—meaning NK cells remain dysfunctional even after TGF-\(\beta\) normalizes
Key evidence gap: No published ME/CFS study has measured both TGF-\(\beta\) levels and NK cell cytotoxicity in the same cohort. This co-measurement study is the critical missing experiment.
Implications if confirmed:
- Reduces the number of independent immune abnormalities in ME/CFS from two to one
- Shifts therapeutic focus from NK cells to whatever drives TGF-\(\beta\) elevation
- Explains why interventions targeting NK cells directly (BioBran, isoprinosine, interferon-\(\alpha\)) have failed—they address the consequence, not the cause
- Connects to the Roerink artifact concern: if TGF-\(\beta\) elevation is partly artifactual, the NK finding may be even more independent than assumed
Falsification criteria:
- If TGF-\(\beta\) levels do not correlate with NK cytotoxicity within ME/CFS patients in a properly controlled co-measurement study
- If in vitro TGF-\(\beta\) neutralization does not rescue NK cytotoxicity in ME/CFS patient samples
- If the TGF-\(\beta\) elevation is entirely explained by centrifugation artifact (Roerink concern), eliminating the upstream driver
Certainty: 0.30. Extends the TGF-\(\beta\) unifying hypothesis with emerging epigenetic evidence. The epigenetic locking mechanism is from a single 2026 preprint (not yet peer-reviewed); application to ME/CFS is entirely speculative.
The TGF-\(\beta\) Rosetta Stone speculation (Speculation TGF-\(\beta\) as Unifying Explanation for the Two Most Replicated Immune Findings) assumes ongoing TGF-\(\beta\) elevation suppresses NK cells. But the Roerink artifact concern and Hornig duration-dependent findings raise the possibility that TGF-\(\beta\) elevation is transient—present during early disease and normalising in chronic ME/CFS. If so, how does NK dysfunction persist?
A 2026 preprint demonstrated that prolonged TGF-\(\beta\) exposure causes persistent epigenetic remodeling of IRF, T-bet, and EOMES binding sites in NK cells—and critically, the dysfunction persists even after TGF-\(\beta\) is withdrawn. We propose a “hit-and-run” mechanism:
- Initial infection triggers transient TGF-\(\beta\) elevation (normal immune regulation)
- In susceptible individuals, TGF-\(\beta\) exposure exceeds a duration threshold sufficient for epigenetic locking
- NK cell chromatin is permanently remodeled: perforin/granzyme loci become inaccessible
- TGF-\(\beta\) normalizes (explaining null measurements in chronic patients)
- NK cells remain dysfunctional indefinitely (explaining persistent cytotoxicity deficit)
- Dysfunctional NK cells cannot clear herpesvirus reactivations
- Chronic low-level viral stimulation may trigger further TGF-\(\beta\) from regulatory T cells, locking new NK cells as they mature from progenitors
This model elegantly reconciles: (a) consistent NK dysfunction in chronic ME/CFS, (b) inconsistent TGF-\(\beta\) elevation (present early, absent late), (c) the Roerink artifact concern (even if current measurements are artifactual, past elevation was real), and (d) the failure of interventions targeting NK cells directly (the epigenetic lock cannot be reversed by providing BioBran or interferon).
Testable prediction: Epigenetic profiling (ATAC-seq, ChIP-seq for H3K27me3) of NK cells from ME/CFS patients should show persistent chromatin modifications at EOMES/T-bet/IRF loci characteristic of TGF-\(\beta\) exposure, even in patients with normal current TGF-\(\beta\) levels. If this chromatin signature is absent, the hit-and-run model is falsified.
2.3 Balance Between Pro- and Anti-inflammatory Signals
The key issue in ME/CFS may not be absolute cytokine levels but rather the balance between pro- and anti-inflammatory signals. Patients may exhibit imbalanced pro-/anti-inflammatory ratios, inappropriate cytokine responses to stimuli, and failure to resolve inflammation properly. This results in chronic low-grade immune activation.
3 Chemokines
Chemokines direct immune cell migration to sites of infection or inflammation:
3.1 Recruitment Patterns
Several chemokines show altered levels in ME/CFS. CCL2 (MCP-1), which recruits monocytes, is often elevated. CCL5 (RANTES) recruits T cells and NK cells, while CXCL8 (IL-8) recruits neutrophils. CXCL10 (IP-10), an interferon-induced chemokine, recruits T cells to sites of inflammation.
3.2 Tissue Infiltration
Elevated chemokines may promote immune cell infiltration into tissues such as muscle, brain, and gut, leading to local inflammation and tissue damage. This infiltration generates symptoms through inflammatory mediators acting at sites of tissue involvement.