Neurosteroid Cross-Talk with Mitochondrial Bioenergetics
A speculative but mechanistically grounded connection links the allopregnanolone axis to the mitochondrial dysfunction documented in this chapter. Allopregnanolone has been reported to modulate mitochondrial bioenergetics beyond its well-characterized GABA-A effects.
Certainty: 0.18. Allopregnanolone and progesterone have been shown in preclinical studies to affect mitochondrial membrane potential, respiratory chain complex activity, and ATP production — effects independent of GABA-A receptor activation. If these mitochondrial effects are operative in ME/CFS, neurosteroid bridge therapy (Oral Micronized Progesterone as Immunological and Neurosteroid Bridge Therapy in ME/CFS, Chapter Epidemiological and Outcomes Research) may improve ME/CFS partly via direct mitochondrial bioenergetic support, rather than solely through HPA calming (allopregnanolone route) or immune modulation (PIBF route).
This would be particularly relevant given that ME/CFS mitochondrial dysfunction may resist cofactor supplementation alone; a hormonal signal acting at the level of mitochondrial membrane dynamics could address a distinct node in the dysfunction network.
Falsifiable prediction: Seahorse extracellular flux assays (measuring PBMC oxygen consumption rate and extracellular acidification rate) will improve in ME/CFS patients on cycle-synchronized oral micronized progesterone, independently of changes in immune phenotype markers. If mitochondrial respiration does not improve while immune markers do change, the mitochondrial route is not operative.
Limitations: Preclinical mitochondrial effects of allopregnanolone are from non-ME/CFS models. No study has measured PBMC mitochondrial function during progesterone supplementation in ME/CFS. The effect size is expected to be modest; this is proposed as a contributing factor, not a primary mechanism.
Certainty: 0.15. Allopregnanolone and progesterone modulate sigma-1 receptor (Sig-1R) activity. Sig-1R is a ligand-activated chaperone at the mitochondria-associated endoplasmic reticulum membrane (MAM) — the interface between ER and mitochondria. Sig-1R regulates ER stress response, mitochondrial-ER calcium transfer, and mitochondrial biogenesis; its dysfunction contributes to ER stress and impaired ATP production.
If some progesterone-axis benefits in ME/CFS operate via Sig-1R at the MAM rather than via GABA-A, this would explain why the brexanolone (pure GABA-A modulator without Sig-1R activity) proof-of-concept may underestimate the efficacy of oral micronized progesterone in ME/CFS: oral progesterone reaches Sig-1R, while brexanolone does not. It also connects the neurosteroid hypothesis to the broader ER stress and mitochondrial dysfunction framework developed in this chapter and in Sigma-1 Receptor Mechanisms and Fluvoxamine Therapy (Chapter Emerging and Investigational Therapies).
Falsifiable prediction: Selective Sig-1R agonists (fluvoxamine at high dose, pridopidine) will produce an overlapping symptom improvement signature with progesterone bridge therapy in ME/CFS, particularly for the mitochondrial and cognitive fatigue components — while brexanolone-like pure GABA-A agonists will improve sleep/HPA symptoms but not cognitive fatigue.
Limitations: Sig-1R modulation by progesterone is documented in vitro; ME/CFS-specific Sig-1R function is entirely unstudied. The predicted signature overlap between Sig-1R agonists and progesterone therapy is untested.
Certainty: 0.40. TSPO biology supports dual role; ME/CFS-specific evidence is lacking. Mechanistic inference from established TSPO function.
TSPO (18 kDa translocator protein, formerly peripheral benzodiazepine receptor) is located on the outer mitochondrial membrane and participates in cholesterol transport, steroidogenesis, and apoptosis (VanElzakker, Brumfield, and Lara Mejia 2019). While TSPO-PET is widely used as a marker of neuroinflammation (microglial activation), its mitochondrial location raises an important ambiguity for interpreting ME/CFS findings:
- Peripheral TSPO elevation: If whole-body TSPO PET studies in ME/CFS show increased signal in postural muscles, shoulders, neck, and bone marrow (as suggested by conference data from Michelle James, Stanford, though this study is not yet published), this could reflect (1) activated peripheral immune cells (monocytes, macrophages) in these tissues, or (2) mitochondrial dysfunction in muscles and bone marrow, or (3) both simultaneously.
- Brain TSPO signal: The null finding of ME/CFS TSPO-PET studies (Raijmakers et al. 2021) (no increased brain binding) may indicate (1) absence of central microglial activation, or (2) BBB transport confounding masking microglial activation (Barzon et al. 2026), or (3) different primary pathology (peripheral mitochondrial dysfunction without central immune involvement).
TSPO’s dual role complicates biomarker interpretation: TSPO-PET cannot distinguish between inflammation and mitochondrial dysfunction without additional data (plasma cytokines, flow cytometry, or mitochondrial function assays).
Testable prediction: In ME/CFS patients, peripheral TSPO signal (muscle, bone marrow) will correlate more strongly with mitochondrial function assays (phosphocreatine recovery, ATP production) than with immune cell activation markers (CRP, cytokines). If TSPO primarily tracks mitochondrial dysfunction in ME/CFS, mitochondrial biomarkers should predict TSPO signal better than inflammatory markers.
Limitations: No published whole-body TSPO PET study in ME/CFS exists. The Michelle James conference data (mentioned in briefing) is not yet peer-reviewed. TSPO’s dual role is established in general biology; ME/CFS-specific validation is entirely lacking. Replication status: dual role established; ME/CFS application not tested.
This metabolic dysfunction interacts bidirectionally with immune dysfunction (Chapter Immune System Dysfunction) and neurological abnormalities (Chapter Neurological and Neurocognitive Dysfunction): inflammation impairs metabolism, metabolic dysfunction impairs immune cell function, and energy deficits affect brain function. Chapter Integrative Models and Multi-System Pathophysiology synthesizes these bidirectional interactions into comprehensive models of ME/CFS pathophysiology, examining how metabolic dysfunction participates in vicious cycles (Section Unifying Mechanisms Across Systems) and contributes to the multi-lock state that perpetuates chronic illness (Section Multi-System Integration and Synthesis).
1 Emerging Metabolic Mechanisms: 2026 Update
Certainty: 0.42. Gottschalk et al. reported ATG13 — an autophagy protein essential for autophagosome initiation — markedly elevated in ME/CFS serum across independent cohorts, with exogenous ATG13 triggering inflammatory signaling in vitro. This is significant because ATG13 is not normally circulating at high levels; its presence extracellularly may function as a DAMP (damage-associated molecular pattern), activating pattern recognition receptors (TLRs, RAGE) and amplifying the Cell Danger Response independently of ongoing viral protein production. (Watton and Prusty 2026)
The upstream mechanism controlling ATG13 availability is now well-characterized: mTORC1 phosphorylates ATG13 at Ser258 (pSer258-ATG13), preventing assembly of the ULK1/ATG13/FIP200 complex required for autophagosome initiation (Section The Biogenesis Trap: ME/CFS Cells Waste Energy on Dysfunctional Mitochondrial Production). Thus, the elevated circulating ATG13 documented in ME/CFS likely reflects mTORC1-driven disassembly of the autophagy initiation complex, with non-complexed ATG13 accumulating intracellularly before extracellular release. This interpretation is supported by the Drosen et al. (2025) murine model, in which mTOR-driven ATG13 inactivation triggered M1 macrophage infiltration, IL-6/RANTES production via STAT3, and PEM-like exercise intolerance (Drosen et al. 2025). Ruan et al. reported reduced pSer258-ATG13 following rapamycin treatment in their pilot study, consistent with mTORC1 de-repression of autophagy initiation (Ruan et al. 2025).
Impaired mitophagy (mitochondrial autophagy) provides a biologically plausible pathway linking infection-triggered immune activation to sustained redox stress, reduced bioenergetic capacity, and amplified inflammatory signaling — without requiring major tissue damage or ongoing productive viraemia. When mitophagy is inefficient, damaged mitochondria accumulate, producing excess ROS and releasing mtDNA that activates innate immune pathways (cGAS-STING). This creates a self-reinforcing loop: immune activation impairs mitophagy → mitochondrial damage accumulates → mtDNA release amplifies immune activation → further mitophagy impairment. (Watton and Prusty 2026) (Certainty: 0.42)
Therapeutic implication. Autophagy/mitophagy restoration may be a tractable therapeutic target. Hochecker et al. demonstrated that whole-body hyperthermia (WBH, 39°C) in ME/CFS patients (n=9) reduced autophagy marker LC3-II by 17.84% (p=0.0065) and improved mitochondrial respiratory capacity across all measures — basal respiration (+66.60%), ATP production (+61.41%), maximal respiration (+97.88%), and spare respiratory capacity (+112.35%) — in PBMCs (Hochecker et al. 2025). HSPA5 (HSP70) mRNA increased 48.33%, suggesting that WBH acts through HSP70-mediated chaperone buffering that shifts resources from chronic autophagic quality control toward oxidative phosphorylation. Pharmacological mTORC1 inhibition (rapamycin/sirolimus) represents a direct approach: Ruan et al. (2025) reported improved fatigue and PEM in an uncontrolled pilot study (n=86, no control group) with exploratory biomarker changes (↓pSer258-ATG13, ↑BECLIN-1) (Ruan et al. 2025). AMPK activators (metformin, berberine) offer an indirect route via mTORC1 inhibition; Brown et al. demonstrated AMPK rescue by metformin in ME/CFS myotubes (Brown et al. 2018), and Fineberg et al. (2025) propose metformin for ME/CFS and Long COVID via AMPK→mTOR→autophagy restoration (Fineberg, Moreau, and Schneider-Futschik 2025). None of these approaches has been tested in a randomized controlled trial in ME/CFS. (Watton and Prusty 2026)
Falsifiable prediction. pSer258-ATG13/total ATG13 ratio is elevated in ME/CFS vs control PBMCs at baseline (≥ 1.5-fold, n=30 per group); rapamycin exposure ex vivo (100 nM, 24h) reduces this ratio and restores autophagic flux (LC3-II turnover assay with bafilomycin A1 block) in ME/CFS but not control cells. Additionally, neutralizing ATG13 with specific antibodies in ex vivo ME/CFS PBMC cultures will reduce inflammatory cytokine production (IL-1\(\beta\), TNF-\(\alpha\)) and shift mitochondrial morphology from fragmented to interconnected networks.
Certainty: 0.45. Autophagy is under strong circadian regulation, peaking during the overnight fasted period when mTORC1 activity is naturally suppressed by AMPK activation. In ME/CFS, if mTORC1 is constitutively active and AMPK is tonically suppressed by sympathetic PKA signaling (Section The Biogenesis Trap: ME/CFS Cells Waste Energy on Dysfunctional Mitochondrial Production), the normal nocturnal autophagy surge would be blunted. The consequence is that cellular waste — damaged proteins, dysfunctional mitochondrial fragments, aggregated protein complexes — accumulates during the day (from normal metabolic activity) but is not cleared overnight. The patient wakes with the same cellular burden carried over from the previous day, creating the subjective experience of unrefreshing sleep despite adequate sleep duration and architecture.
This is distinct from sleep architecture theories of unrefreshing sleep (alpha-delta intrusions, reduced slow-wave sleep). Those theories explain why sleep quality may be objectively poor; this theory explains why even objectively adequate sleep fails to restore function. The two are compatible: disrupted sleep architecture reduces the time available for autophagy, while mTORC1 hyperactivation blunts the efficacy of whatever autophagy window exists. The prediction is that pSer258-ATG13 in PBMCs should fail to show the normal nighttime decline in ME/CFS (serial 24-hour sampling, inpatient setting), and that interventions restoring circadian autophagy (TRE, metformin, rapamycin) should improve subjective sleep recovery scores independently of changes in total sleep time or sleep stage distribution.
Falsifiable prediction. PBMC pSer258-ATG13 will show a normal nighttime trough (>40% decline from daytime peak) in healthy controls but a flattened rhythm (nocturnal decline < 20%) in ME/CFS during 24-hour inpatient sampling. Rapamycin (morning dosing, peak effect at night) will restore the nocturnal pSer258-ATG13 trough and improve subjective unrefreshing sleep scores (Visual Analogue Scale) by week 4 of treatment.
Certainty: 0.50. The phosphorylation state of ATG13 at Ser258 is a direct readout of mTORC1 activity at the autophagy initiation complex. Using the ratio of pSer258-ATG13 to total ATG13 normalises for individual-level expression differences and provides a single number reflecting mTORC1 “tone” — the fraction of the autophagy gate that is actively held closed. A tone index exceeding 0.5 (more than half of ATG13 molecules phosphorylated) predicts non-functional autophagy initiation. Normalising to a post-24-hour-fasted ratio provides the “autophagy reserve” score: how much residual gate closure occurs after the strongest physiological stimulus (fasting), reflecting the depth of the mTORC1 lock.
This ratio has several properties making it suitable as a clinical biomarker: (a) it is mechanistically proximal (one step from functional autophagy), (b) it normalises inter-individual variation in ATG13 expression, (c) it is dynamic (changes within hours of mTORC1 modulation), and (d) it has preliminary clinical signal (Ruan et al. reported reduced pSer258-ATG13 after rapamycin (Ruan et al. 2025)). In practice, measurement requires paired PBMC immunoblotting for phospho- and total ATG13 — a moderate-cost assay (USD 200–300 per sample) feasible in research labs but not yet standardized for clinical use.
Falsifiable prediction. The mTOR tone index (fed/fasted pSer258-ATG13 ratio) will discriminate ME/CFS from healthy controls with AUC > 0.75 in a blinded comparison (n=50 per group). Patients with tone index > 2.0 (i.e., more than twice the expected fasted ratio) will show ≥ 2-fold higher response rate to rapamycin or metformin compared to those with tone index ≤ 1.5. Tone index will show test-retest reliability with ICC > 0.7 over a 4-week interval in stable patients.
Certainty: 0.33. Recent ME/CFS-specific work identifies SMPDL3B (sphingomyelin phosphodiesterase acid-like 3B), a membrane-associated regulator of sphingolipid balance, as a potential biomarker and mechanistic mediator. Altered SMPDL3B may change lipid raft fluidity and receptor organization, modifying immune responses to inflammatory and antiviral stimuli. (Watton and Prusty 2026)
Convergent findings in ME/CFS and Long-COVID implicate membrane lipid organization in immune dysfunction. Lipid rafts function as signaling platforms that regulate receptor clustering and activation thresholds, including Toll-like receptor signaling. The lipid raft framework links commonly observed “omics-level” lipid abnormalities in ME/CFS to cellular immunophenotypes, since raft composition depends on sphingolipid and cholesterol handling and can be remodeled by systemic metabolic cues and chronic inflammatory signaling. (Watton and Prusty 2026)
A critical theme emerges: lipid biology is not an adjunct topic in ME/CFS but a potential mechanistic substrate linking immune signaling, mitochondrial quality control, redox buffering, and endothelial stability. Lipids provide structure to membranes and substrates for inflammatory mediators; they govern receptor microdomains while regulating mitochondrial dynamics. Consequently, lipid dysregulation could plausibly yield a persistent “mis-set” immune state after infection. The SMPDL3B pathway offers a testable entry point for investigating this lipid-immune-mitochondrial axis. (Certainty: 0.33)
Falsifiable prediction. SMPDL3B knockdown or overexpression in primary immune cells will predictably alter (1) lipid raft receptor distribution (TRPM3 surface localization), (2) ceramide levels in mitochondrial membranes, and (3) mitophagy flux (LC3-II accumulation, PINK1/Parkin recruitment).
Certainty: 0.45. Moreau’s group identified haptoglobin (Hp) quantity and proteoform structure as relevant to core ME/CFS symptoms including PEM and cognitive dysfunction. Hp scavenges free haemoglobin released during hemolysis and thus buffers against heme-mediated oxidative stress and endothelial injury. Reduced functional Hp capacity permits exaggerated oxidative and vascular stress during or after exertion, with downstream effects on microvascular regulation and symptom relapse. (Watton and Prusty 2026)
Taken together, ATG13/autophagy perturbation, SMPDL3B/lipid raft regulation, and haptoglobin proteoform variation converge on a common theme: lipid and protein quality control systems buffer physiological stress, and their impairment in ME/CFS creates vulnerability that becomes clinically apparent under exertion. This is consistent with the broader model of state-dependent, stress-revealed pathology described in Chapter Integrative Models and Multi-System Pathophysiology (Section Time-Dependent Dosing — The Missing Temporal Axis in the Hormesis Framework). (Certainty: 0.45)
Falsifiable prediction. Specific Hp proteoforms will correlate with cognitive test scores (processing speed, working memory) during PEM, endothelial function markers (VCAM-1, E-selectin), and time to symptom resolution, independent of total Hp concentration.
2 Speculative Metabolic Interventions
Certainty: 0.48. Probability of clinically meaningful efficacy in ME/CFS: 0.08. PINK1/Parkin pathway activators (urolithin A, nicotinamide riboside) could enhance damaged mitochondria clearance, reduce oxidative burden, and allow metabolic recovery from the CDR state. Impaired mitophagy sustains redox stress and inflammation per the unified model. (Watton and Prusty 2026) PINK1/Parkin activators are in preclinical/early clinical development for other conditions; no ME/CFS data exists. Urolithin A has human safety data from muscle aging trials. Falsifiable: mitophagy enhancer treatment of ME/CFS cells will increase LC3-II accumulation and PINK1/Parkin recruitment to mitochondria, reduce ROS, and improve OCR.
Certainty: 0.30. Probability of clinically meaningful efficacy in ME/CFS: 0.04. Spermidine, a well-established dietary autophagy inducer, could pharmacologically recapitulate the hyperthermia effect demonstrated by Hochecker et al. (autophagy markers decreased 17.84%, mitochondrial respiration increased 61-112% (Hochecker et al. 2025)), potentially shifting cells from stalled CDR to adaptive metabolic state. Spermidine-induced autophagy is well-established in aging biology; no ME/CFS data exists. Falsifiable: spermidine supplementation in ME/CFS PBMCs will increase autophagic flux (LC3-II ratio, p62 degradation) and improve mitochondrial respiration.
Certainty: 0.35. Probability of clinically meaningful efficacy in ME/CFS: 0.06. Butyrate is an HDAC inhibitor that can epigenetically reprogramme mitochondrial gene expression. SCFA deficits (Chapter Gastrointestinal and Microbiome Dysfunction) plus CDR-induced epigenetic silencing may be counteracted by butyrate supplementation, restoring PGC-1\(\alpha\) and NRF1-dependent mitochondrial biogenesis. Butyrate HDAC inhibition and mitochondrial biogenesis upregulation are established in vitro; no ME/CFS-specific data. Falsifiable: butyrate treatment of ME/CFS immune cells will increase histone acetylation at PGC-1\(\alpha\) promoters and upregulate TFAM/mtDNA copy number.
Certainty: 0.25. Probability of clinically meaningful efficacy in ME/CFS: 0.03. Taurine modulates calcium signaling, stabilizes mitochondrial membranes, and reduces oxidative stress. Given TRPM3 dysfunction and calcium-mitochondria crosstalk deficits, taurine supplementation could support both calcium homeostasis and mitochondrial resilience. Taurine’s calcium-modulating and mitochondrial-protective effects are well-characterized; no ME/CFS data. Falsifiable: taurine treatment of ME/CFS cells will normalise calcium influx kinetics, stabilize mitochondrial membrane potential, and reduce mPTP opening.
Certainty: 0.35. Probability of clinically meaningful efficacy in ME/CFS: 0.08. CDR involves redox dysregulation. NAC replenishes glutathione, modulates redox-sensitive signaling (NF-\(\kappa\)B, Nrf2), and may facilitate CDR resolution by restoring thiol redox balance. NAC’s glutathione-replenishing effects are well-established clinically; no ME/CFS-specific trial data. The probability is higher than most nutraceuticals because NAC has demonstrated objective efficacy in related conditions (COPD, psychiatric disorders) via glutathione restoration. Falsifiable: NAC treatment will increase GSH/GSSG ratio in ME/CFS PBMCs, shift Nrf2-dependent gene expression, and reduce inflammatory cytokine production.
Certainty: 0.25. Probability of clinically meaningful efficacy in ME/CFS: 0.01. ⚠ THERMOREGULATORY SAFETY CAVEAT. Cold exposure activates brown adipose tissue, increases mitochondrial biogenesis via PGC-1\(\alpha\), and modulates inflammatory signaling. However, ME/CFS patients have documented thermoregulatory dysfunction, autonomic instability, and cold hypersensitivity that contraindicate standard protocols. The existing cryotherapy literature shows 45% attrition in the only CFS whole-body cryotherapy study (Kujawski 2023), suggesting poor tolerability. Any protocol would require extremely gradual titration under clinical supervision. Falsifiable: gradual cold exposure will increase PGC-1\(\alpha\) and NRF1 expression in PBMCs and be tolerated without PEM at slow titration rates.
Certainty: 0.30. Hochecker et al. (2025, n=9, within-subject pre-post, single center, unreplicated) demonstrated that WBH increases spare respiratory capacity by 112.35% in ME/CFS PBMCs (Hochecker et al. 2025). Critical caveat: whether PBMC spare respiratory capacity correlates with thermoregulatory tissue (eccrine gland, skeletal muscle, brown adipose, vascular smooth muscle) capacity is entirely untested — PBMCs are immune cells, not thermoeffectors. The inference from cellular respirometry to whole-body thermoregulation is a major extrapolation. Thermoregulation is energetically expensive: sweating requires sustained ATP for eccrine gland ion transport, shivering requires burst mitochondrial output, and vasomotion requires constant smooth muscle ATP turnover. Spare respiratory capacity — the difference between maximal and basal respiration — theoretically represents the mitochondrial reserve available for these thermoregulatory demands, but this has never been measured in thermoeffector tissues. Predictions (a, b, c) are tentative and depend on the PBMC-to-thermoeffector extrapolation being valid. Falsifiable: spare respiratory capacity in ME/CFS PBMCs will correlate with heat tolerance time (r > 0.6) and cold tolerance time (r > 0.5); improving reserve via WBH will increase both tolerances.
Certainty: 0.40. LC3-II is elevated in ME/CFS PBMCs, suggesting chronic selective autophagy (mitophagy, ER-phagy) attempting to clear damaged organelles. Hochecker et al. showed that WBH reduces LC3-II by 17.84% (p=0.0065) while simultaneously increasing HSP70 and mitochondrial respiration (Hochecker et al. 2025). This pattern suggests that heat shock triggers a switch from energy-intensive selective autophagy to more efficient bulk autophagy via HSP70 chaperone redistribution: HSP70 released from autophagosome membranes becomes available for mitochondrial protein import and folding support, simultaneously reducing selective autophagy markers and improving oxidative phosphorylation. This model explains the apparently paradoxical finding that decreasing autophagy (LC3-II) improves mitochondrial function — autophagy was not “good” autophagy clearing healthy mitochondria, but “stress” autophagy attempting to compensate for proteotoxic overload. Falsifiable: ME/CFS PBMCs will show elevated selective autophagy markers (Parkin, BNIP3) and reduced bulk autophagy flux (p62 degradation); WBH will decrease selective markers and increase bulk flux; HSP70 knockdown will block the WBH-induced switch.
Certainty: 0.40. Probability of clinically meaningful efficacy in ME/CFS: 0.10. Hypoxia preconditioning upregulates HIF-1\(\alpha\), enhancing mitochondrial efficiency via PGC-1\(\alpha\) co-activation and improving stress tolerance. The HIF pathway is increasingly recognized as a central hub in ME/CFS pathophysiology: HIF-1\(\alpha\) stabilization drives the metabolic shift toward glycolysis at the expense of oxidative phosphorylation, while chronic HIF-2\(\alpha\) activation contributes to endothelial barrier dysfunction (Ribeiro et al. 2026) (Section Immune Complex–Endothelial Injury as a Central Vascular Mechanism in Cardiovascular Dysfunction for the expanded mechanism). Rather than avoiding hypoxia altogether, controlled brief hypoxic exposures may harness the HIF programme’s adaptive arm — inducing EPO, VEGF, glycolytic enzymes, and mitophagy regulators — while avoiding the maladaptive persistence of HIF-mediated metabolic suppression that may underlie PEM (Kaczmarek 2023). The key distinction is intermittency: brief hypoxia cycles (3–5~min) with intervening normoxia may train the HIF response to activate and deactivate appropriately, restoring the dynamic range lost in ME/CFS. This framework integrates the HIF inertia hypothesis (HIF Pathway Inertia: Failure of Hypoxic Adaptation in ME/CFS) with intermittent hypoxic training data from other conditions (Gangwar et al. 2019) (Janssen Daalen et al. 2025).
The mechanistic coherence is higher than the original estimate (0.25) because: (a) the HIF pathway now has direct ME/CFS relevance via miRNA and ECM evidence, (b) intermittent hypoxic training has demonstrated safety and efficacy in Parkinson’s disease, and (c) the dual-face of HIF (adaptive vs maladaptive) provides a clear framework for distinguishing beneficial from harmful protocols. However, clinical efficacy probability remains modest (0.10) pending direct ME/CFS data. Hypoxia still carries risks of triggering PEM in metabolically fragile patients, and the protocol parameters (FiO2, cycle timing, session frequency) are entirely untested in ME/CFS.
Falsifiable predictions: + Normobaric hypoxia preconditioning (FiO2 0.16, 5~min cycles × 6, 3×/week × 4 weeks) increases HIF-1\(\alpha\) protein by \(>=1.5\)-fold and PGC-1\(\alpha\) mRNA by \(>=2\)-fold in PBMCs vs sham + Dynamic HIF response range (peak-to-baseline fold-change of HIF target gene expression after acute hypoxia challenge) increases by \(>=50\)% vs pre-training baseline + Peak work rate on CPET increases by \(>=10\)% without \(>=2\)-point increase in PEM incidence at 48~h vs sham
Limitations. No ME/CFS preconditioning data exist. Protocol parameters are extrapolated from neurorehabilitation literature. Individual variation in hypoxic ventilatory response and baseline HIF activity may determine responders vs non-responders. The risk of PEM induction cannot be eliminated without empirical dose-finding.
Mitochondrial fragmentation is central to ME/CFS (Section The Energy Chain: Ten Steps from Substrate to ATP). High-content imaging of PBMC mitochondria — morphology, membrane potential, respiration — could provide quantifiable biomarkers of CDR state, track disease activity during PEM cycles, and assess treatment response objectively. Existing mitochondrial fragmentation assays (MitoTracker, Seahorse) are mature technologies; the open question is whether PBMC mitochondrial morphology produces clinically useful disease-activity metrics. Falsifiable: PBMC mitochondrial fragmentation indices will correlate with disease severity and change detectably during PEM cycles before symptom-scale changes. Probability of diagnostic utility: 0.15.
Does ATG13 utilize distinct receptors from HMGB1/mtDNA, or does it synergise with conventional DAMPs to amplify inflammatory signaling (Ibrahim, Wasim, and Rahman 2026) (Chen et al. 2026)? Characterising ATG13 receptor usage (TLRs, RAGE, STING) and downstream pathway convergence (NF-\(\kappa\)B, cGAS-STING) will clarify its unique contribution to CDR amplification and whether ATG13 neutralization would produce effects additive to or redundant with standard anti-inflammatory approaches. Falsifiable: ATG13 will utilize distinct receptors from HMGB1/mtDNA and synergise with conventional DAMPs to amplify cytokine production via converging on NF-\(\kappa\)B and STING pathways. Probability of novel therapeutic target: 0.10.
Certainty: 0.60. Moschini et al. demonstrated that HIF-1alpha drives tendinopathy independently of VEGF in a murine model, establishing HIF-1alpha as a direct mechanistic driver of connective tissue pathology (Moschini, Mohanan, et al. 2026). (Sahin et al. 2012) This finding, when integrated with established ME/CFS findings on mitochondrial dysfunction and HIF-1alpha elevation, suggests a self-reinforcing pathogenic triad.
Mechanism. HIF-1alpha suppresses mitochondrial oxidative phosphorylation and shifts metabolism toward glycolysis—the “Warburg-like” metabolic phenotype documented in ME/CFS immune cells (Chapter Immune System Dysfunction) and muscle. That infection can trigger host HIF-1\(\alpha\) stabilisation has an in vitro proof-of-principle: Bartonella henselae stabilises HIF-1\(\alpha\) in cell lines, producing cellular hypoxia and decreased ATP (Kempf et al. 2005) — the first such demonstration in a bacterial infection, though it is a single unreplicated cell-line study and does not by itself establish the effect in primary cells or in ME/CFS. miRNA-mediated silencing of HIF target genes may further impair the adaptive transcriptional response to tissue hypoxia, blunting EPO and VEGF induction and locking the system in a maladapted state (Kaczmarek 2023). Mitochondrial dysfunction increases reactive oxygen species (ROS) production, which stabilizes HIF-1alpha through prolyl hydroxylase inhibition, creating a vicious cycle. Both HIF-1alpha-driven metabolic shift and mitochondrial ROS impair extracellular matrix homeostasis: HIF-1alpha directly upregulates matrix metalloproteinases (MMPs) and downregulates collagen synthesis, while ROS damages existing ECM components and fibroblasts. The triad becomes self-reinforcing: HIF-1alpha drives ECM pathology; mitochondrial dysfunction impairs ECM repair; ROS from mitochondria stabilizes HIF-1alpha.
ME/CFS context. ME/CFS tissue samples should show correlated abnormalities in HIF-1alpha, mitochondrial function, and ECM markers. The high prevalence of hypermobility syndromes (hEDS, HSD) in ME/CFS—up to 70% in some cohorts—may represent the clinical manifestation of this triad operating in connective tissue. (Ganesh and Munipalli 2024) Ramirez-Paesano et al. (2023)
Cross-references. Chapter Integrative Models and Multi-System Pathophysiology synthesizes cross-system pathophysiological loops from HIF-1alpha elevation documented in ME/CFS stress response studies.
Falsifiable prediction. ME/CFS tissue samples (muscle biopsy, skin, or tendon when clinically indicated) will show: (1) elevated HIF-1alpha protein levels compared to controls; (2) correlated mitochondrial dysfunction (reduced oxidative phosphorylation, elevated ROS); (3) ECM abnormalities (elevated MMP-3/MMP-9, abnormal collagen/elastin ratios). Within ME/CFS patients, HIF-1alpha levels will correlate with both mitochondrial dysfunction severity and hypermobility scores (Beighton, GJH).
Limitations. Moschini 2026 is a murine tendinopathy model; human tendon data are limited. Tissue sampling in ME/CFS is rare; skin or muscle biopsy may not fully recapitulate tendon pathology. The directionality of the triad components in established disease is unclear—any node could be primary in individual patients.
Certainty: 0.45. Angiotensin II receptor blockers (ARBs; losartan, telmisartan) may restore capillary basement membrane structure and improve microcirculation in ME/CFS by reducing TGF-\(\beta\) signaling, which drives excessive ECM production at capillary basement membranes. (Wüst et al. 2024) (Wirth 2026)
Mechanism. Wust et al. demonstrated massive collagen IV deposition and basement membrane thickening in ME/CFS muscle capillaries via electron microscopy—a structural abnormality that creates a mechanical bottleneck independent of mitochondrial function (Wüst et al. 2024). Basement membrane thickening impairs oxygen and nutrient diffusion, creating pre-mitochondrial hypoxia. ARBs reduce TGF-\(\beta\) signaling and ECM production, with demonstrated efficacy in diabetic nephropathy where they normalize glomerular basement membrane thickness and preserve filtration function. In ME/CFS, ARBs could thin basement membranes, restore diffusion gradients, and improve microcirculation without directly targeting mitochondria.
Clinical rationale. ARBs are well-tolerated, widely available, and have favorable safety profiles. The TGF-\(\beta\) elevation documented in ME/CFS (Chapter Immune System Dysfunction) provides mechanistic support. The bidirectional connective tissue disorder-ME/CFS relationship established by Wirth et al. suggests that connective tissue pathology is not merely a comorbidity but may contribute to disease maintenance (Wirth 2026).
Dose titration requirement. The hypothesis requires careful dose titration: low-dose ARBs may reduce TGF-\(\beta\) without clinically significant blood pressure effects, while standard antihypertensive doses may cause orthostatic intolerance in POTS-comorbid patients. Telmisartan has partial PPAR-\(\gamma\) agonism, providing additional anti-inflammatory benefit; losartan has stronger uric acid-lowering effects relevant to metabolic comorbidities.
Falsifiable prediction. ME/CFS patients treated with low-dose ARBs will show: (1) reduced basement membrane thickening on serial muscle biopsy (feasible in interventional trials with pre/post sampling); (2) improved capillary diffusion metrics (oxygen microsensor, contrast-enhanced ultrasound); (3) reduced exercise-induced lactate at given workloads (reflecting improved oxygen delivery); (4) symptom improvement correlating with basement membrane thinning rather than blood pressure changes.
Limitations. Basement membrane measurements require muscle biopsy—feasibility limited to interventional trials. TGF-\(\beta\) has pleiotropic functions; systemic reduction may have unintended consequences. Not all ME/CFS patients may have basement membrane pathology; the hypothesis predicts subgroup specificity.
Certainty: 0.30. ME/CFS metabolic dysfunction may impair fibroblast glycogen storage and utilization, reducing ECM quality and contributing to connective tissue pathology. This novel connection links the metabolic trap hypothesis (Chapter Energy Metabolism and Mitochondrial Function) to connective tissue pathophysiology.
Mechanism. Fibroblasts store glycogen as an energy reserve for ECM synthesis—collagen, elastin, and proteoglycans are energetically expensive to produce. Glycogen provides local ATP for the extensive protein synthesis and crosslinking required for ECM assembly. If ME/CFS metabolic dysfunction impairs fibroblast glycogen storage or utilization—through impaired glycolysis, glycogen synthase dysfunction, or abnormal glycogen branching—fibroblasts may produce reduced quantities or structurally abnormal ECM. This could explain the high prevalence of hypermobility, skin hyperextensibility, and tissue fragility in ME/CFS.
Connection to metabolic trap. Phair’s metabolic trap hypothesis—impaired CD39/CD73 activity reducing adenosine production—creates a systemic bioenergetic deficit that would disproportionately affect fibroblasts, which are long-lived, slowly-proliferating cells with limited regenerative capacity compared to immune cells. Fibroblasts cannot rapidly compensate for chronic energy deficit through proliferation; instead, they would downregulate energy-intensive ECM synthesis, producing progressive connective tissue weakness.
Tissue specificity. Different fibroblast populations may be differentially affected based on glycogen content and turnover rates: skin fibroblasts (high glycogen, visible as PAS-positive granules in normal histology) may show early abnormalities; tendon fibroblasts (low baseline glycogen) may be more vulnerable to minor deficits; vascular adventitial fibroblasts may contribute to capillary basement membrane pathology when glycogen-depleted.
Falsifiable prediction. ME/CFS fibroblasts (skin biopsy, easily accessible) will show: (1) reduced glycogen content compared to controls (PAS staining quantification, biochemical glycogen assay); (2) impaired glycogen synthase or branching enzyme activity; (3) reduced collagen production in culture (hydroxyproline assay) that normalizes when supplemented with pyruvate or other ATP precursors; (4) abnormal collagen crosslinking patterns (enzymatic vs. non-enzymatic crosslinks).
Limitations. This is a novel, speculative connection with no direct ME/CFS data. Fibroblast glycogen measurement is not routine; methodology would need development. The metabolic trap hypothesis itself remains untested. Alternative explanations for connective tissue abnormalities in ME/CFS include genetic collagen defects (hEDS) or immune-mediated ECM degradation.
Certainty: 0.40. NAD depletion is documented in ME/CFS (Chapter Energy Metabolism and Mitochondrial Function, Section Step 6: Key Cofactors as Cross-Cutting Failure Modes). Reduced NAD impairs mitophagy via SIRT1-PINK1/Parkin deacetylation failure, leading to mitochondrial depolarization and mtDNA release into the cytosol. Cytosolic mtDNA activates cGAS-STING, producing type I interferons (IFN-\(\beta\)). IFN-\(\beta\) induces indoleamine 2,3-dioxygenase (IDO), which consumes tryptophan along the kynurenine pathway, further depleting NAD via nicotinamide phosphoribosyltransferase (NAMPT) upregulation — a futile compensatory loop that accelerates NAD consumption.
Mechanism. Step 1: NAD deficiency impairs SIRT3-mediated deacetylation of mitochondrial proteins, reducing complex I and complex III activity (documented in ME/CFS muscle). Step 2: Impaired electron transport chain increases ROS and triggers mitochondrial permeability transition pore (mPTP) opening. Step 3: mPTP opening releases fragmented mtDNA into the cytosol. Step 4: Cytosolic mtDNA engages cGAS, producing cGAMP which activates STING. Step 5: STING-TBK1-IRF3 signaling drives IFN-\(\beta\) transcription. Step 6: IFN-\(\beta\) upregulates IDO, shunting tryptophan to kynurenine and reducing NAD salvage synthesis. Step 7: NAD drops further, worsening mitophagy failure — loop closes.
ME/CFS context. Elevated IFN signatures are documented in ME/CFS (Chapter Immune System Dysfunction). Kynurenine pathway activation is reported. NAD levels are low in ME/CFS muscle and PBMCs. cGAS-STING activation has been demonstrated in post-COVID syndromes. The loop explains chronicity: once initiated, it is self-sustaining because each arm (NAD deficit, mtDNA release, IFN production, IDO activation) reinforces the next.
Falsifiable predictions. (1) ME/CFS PBMCs will show elevated cytosolic mtDNA levels (quantitative PCR for mtDNA-encoded genes in cytosolic fractions). (2) STING phosphorylation (p-STING Ser366) will be elevated in ME/CFS immune cells and correlate with NAD/NADH ratio. (3) IDO activity (kynurenine/tryptophan ratio) will correlate inversely with cellular NAD levels. (4) Pharmacological STING inhibition (H-151) in ME/CFS PBMC cultures will reduce IFN-\(\beta\) production and partially restore NAD levels within 48 h.
Limitations. The complete loop has not been demonstrated in any single ME/CFS study — the components are assembled from separate studies across different tissues and cohorts. cGAS-STING in ME/CFS has not been directly measured. mtDNA in cytosol is technically challenging to quantify without artefacts from cell fractionation. The loop may operate in some tissues (immune cells) but not others (muscle, neurons), with measurement compartment determining detectability.
Certainty: 0.50. Chronic inflammation in ME/CFS raises hepcidin via IL-6-STAT3 signaling (documented elevated IL-6, Chapter Immune System Dysfunction), shifting the iron homeostasis setpoint toward functional iron deficiency. This produces a hypoferremic state that limits erythropoiesis (contributing to anemia of chronic disease) and impairs mitochondrial respiration (iron-sulfur cluster biogenesis for complexes I–III), while simultaneously trapping iron in macrophages and hepatocytes. Macrophage iron retention promotes intracellular oxidative stress and inflammatory polarization (M1 skewing), creating a two-compartment problem: insufficient iron for energy metabolism in muscle and brain, excessive iron driving oxidative damage in immune cells.
Mechanism. IL-6 induces hepcidin transcription via STAT3 binding to the hepcidin (HAMP) promoter. Elevated hepcidin binds ferroportin (FPN1) on enterocytes, macrophages, and hepatocytes, triggering its internalization and degradation. Enterocyte FPN degradation blocks dietary iron absorption. Macrophage FPN degradation traps recycled iron from senescent RBCs inside the phagolysosome. The resulting low serum iron (hypoferremia) starves muscle mitochondria of iron-sulfur cluster substrates (reduced aconitase activity, complex I/II/III dysfunction — documented in ME/CFS). Simultaneously, macrophage iron overload drives Fenton chemistry, producing hydroxyl radicals that oxidize membrane lipids → ferroptosis in iron-loaded cells. Ferroptotic cell death releases DAMPs (HMGB1, mtDNA) that amplify inflammation (Ibrahim, Wasim, and Rahman 2026), further raising IL-6 and hepcidin — a self-sustaining iron trap.
Bridge to ferroptosis. The hypoferremia-ferroptosis trap connects two previously separate observations in ME/CFS: (1) functional iron deficiency (low ferritin in some, elevated soluble transferrin receptor indicating cellular iron need) and (2) oxidative stress markers (elevated malondialdehyde, F2-isoprostanes, 4-HNE). If iron is sequestered in macrophages while muscles are iron-starved, serum markers would paradoxically show low iron availability while tissue markers show iron-driven oxidative damage — reconciling apparently contradictory findings.
Falsifiable predictions. (1) ME/CFS patients will show elevated hepcidin (serum ELISA) correlating with IL-6 levels and inversely with serum iron and transferrin saturation. (2) Macrophages isolated from ME/CFS PBMCs will show elevated intracellular iron (ferritin heavy chain, labile iron pool by calcein-AM fluorescence) and increased lipid peroxidation (C11-BODIPY) compared to controls. (3) Muscle biopsies from ME/CFS patients will show reduced iron-sulfur cluster proteins (aconitase, SDHB) despite normal total muscle iron — indicating functional iron deficiency within the myocyte despite adequate total tissue iron. (4) In vitro, IL-6 treatment of healthy macrophages induces hepcidin-dependent iron retention; ME/CFS patient macrophages will show blunted ferroportin expression and heightened sensitivity to ferroptosis inducers (erastin, RSL3).
Limitations. Hepcidin has not been systematically measured in ME/CFS. Macrophage iron content has not been assessed. Ferroptosis markers in ME/CFS tissue are absent. The model predicts a compartment-specific iron paradox that may not be detectable in standard serum iron panels, explaining why routine iron studies in ME/CFS are often normal. The prediction that muscle is iron-deficient while macrophages are iron-loaded requires direct tissue measurement, which is invasive.