Integrative Models
1 Energy Ratchet / Hysteresis Model
Certainty: 0.40. Progressive loss of work capacity across PEM episodes is clinically observed. Epigenetic scarring and recovery capital depletion are mechanistic inferences. No direct longitudinal measurement in ME/CFS.
1.1 Cascade: PEM → residual damage → capacity ratchet → treatment resistance
Cascade:
- Repeated PEM episodes → progressive lowering of crash threshold
- Residual damage (epigenetic marks, mitochondrial damage, microglial priming, HSC exhaustion)
- Recovery incomplete → work capacity ratchets downward → treatment resistance increases with illness duration
Step S1: PEM episode → residual damage accumulation
- Intercept: PEM prevention (pacing, HR monitoring); Antioxidants (CoQ10, NAC; mitochondrial support) during exertion; Rapamycin (mTORC1 inhibitor; autophagy enhancer, Treg stabilization) — enhances mitophagy between episodes
The ratchet is PEM-event-driven.
- Certainty
- Low to Medium
- Level of action
- Partial root cause
Step S2: Epigenetic scarring → treatment resistance
- Intercept: Early intervention (before scarring accumulates) per Integrative Speculations
- Testable prediction: Shorter illness duration patients should respond to interventions that fail in longer-duration patients.
Consequence: Temporal probe: same intervention succeeding in recently-ill but failing in long-duration patient suggests the ratchet has hardened the disease state. Argues for early aggressive intervention. Origin: mechanistic-pathway-tracing.
2 Energy Triage / Selective Energy Dysfunction
Certainty: 0.45. Pattern of preserved baseline function with catastrophic demand failure documented in ME/CFS (2-day CPET: normal VO₂peak on Day 1, significantly reduced on Day 2). CNS energy crisis model CNS Energy Crisis as Primary Dysfunction — the brain is an obligate glucose consumer with minimal glycogen reserves, consuming ~20% of total body ATP. Under energy shortage, ATP allocation is triaged: essential functions preserved (basal metabolism, vital organ function, hair/nail growth — note that hair and nails are often reported as growing normally in ME/CFS despite severe functional impairment), non-essential functions deprioritized (exercise capacity, immune adaptation, orthostatic tolerance, cognitive effort, emotional regulation). The diagnostic task is to identify WHICH demand-response systems are being ATP-deprioritized — the pattern of which pharmacological bypass agents work reveals the triage pattern in that patient.
2.1 Cascade: Energy shortage → AMPK → ATP triage → demand failure
Cascade:
- Cellular energy shortage (ETC dysfunction, PDH block, NAD⁺ depletion, or delivery failure) → AMPK activation
- Selective ATP allocation → deprioritization order: exercise capacity → immune adaptation → orthostatic tolerance → cognitive effort → emotional regulation
- Preservation of essential functions: basal metabolism, cardiac output, respiratory drive, renal function
- Clinical pattern: appears well at rest, catastrophically fails under demand
2.2 Step T1: Total ATP pool — is the pool expandable or fixed?
Probes for ATP pool expansion:
- CoQ10 (100–300 mg) + B-vitamins (B1, B2, B3, B5) + L-carnitine (500–1000 mg): Cofactor package for mitochondrial ATP production. CoQ10 shuttles electrons from Complex I/II to Complex III. B-vitamins are ETC cofactors (B2 = FAD, B3 = NAD⁺). L-carnitine shuttles fatty acids into mitochondria for β-oxidation.
The total ATP pool is expandable — mitochondrial capacity was limited by cofactor availability, not by structural damage. This is the most favorable finding: the energy deficit is nutritional/mild-functional, and supplementation increases the pool. Record magnitude: 10% improvement → cofactors are marginal; 50% improvement → cofactor deficiency was the dominant bottleneck.
- Certainty
- Low to Medium
- Level of action
- Partial root cause
The ATP pool is fixed — cofactor supplementation cannot increase production. Either: (a) mitochondrial structural damage (ETC complexes are broken, not cofactor-starved) → mitophagy + biogenesis needed (rapamycin, urolithin A); (b) the bottleneck is upstream of ETC (substrate delivery — pyruvate, fatty acids, oxygen) → delivery probes needed (midodrine, pyridostigmine); or (c) the bottleneck is at ATP utilization (ANT export, creatine phosphate shuttle) → ATP is produced but cannot reach the cellular machinery → creatine should work.
- Certainty
- Low to Medium
- Level of action
- Partial root cause
- Creatine monohydrate (5 g/day): Bypasses the ATP synthesis machinery entirely — creatine phosphate acts as a spatial ATP buffer, shuttling high-energy phosphate from mitochondria to cytosol. If creatine improves functional capacity where cofactors did not → ATP is being produced (mitochondria work) but cannot reach the cellular machinery that needs it (ANT export or creatine phosphate shuttle failure). The bottleneck is ATP distribution, not ATP production.
2.3 Step T2: Pattern diagnosis of ATP triage — which systems are deprioritized?
The core probe: pharmacological bypass of each demand-response signaling pathway. Each drug provides the neurotransmitter or receptor activation that ATP shortage should have funded. The pattern of which bypass works identifies which signaling system is being ATP-triage-deprioritized:
- Midodrine works → adrenergic signaling (α1) ATP-triage-deprioritized. The sympathetic nervous system requires ATP for NE synthesis (tyrosine hydroxylase requires BH4, which requires GTP), NE vesicular loading (VMAT2 requires ATP), and NE release (Ca²⁺-dependent exocytosis requires ATP for Ca²⁺-ATPase). If the sympathetic system’s ATP budget is cut, NE signaling fails. Midodrine bypasses the entire NE pathway — it directly activates α1 receptors, requiring no ATP for synthesis, storage, or release. The positive midodrine response means α1 receptors are present and responsive, but endogenous NE is insufficiently produced or released — consistent with ATP-triage-deprioritized adrenergic signaling.
- Pyridostigmine works → cholinergic signaling ATP-triage-deprioritized. ACh synthesis (choline acetyltransferase requires acetyl-CoA, which requires ATP for pyruvate dehydrogenase and citrate synthase), vesicular loading (VAChT requires proton gradient, indirectly ATP-dependent), and release. Pyridostigmine increases synaptic ACh by blocking its degradation — it makes whatever ACh IS released last longer. Pyridostigmine response means ACh is being released at some level (the synapse is functional), but the amount is inadequate (consistent with reduced synthesis from ATP shortage). AChE inhibition amplifies the signal without requiring more ATP.
- Aripiprazole works → dopaminergic signaling ATP-triage-deprioritized. DA synthesis (tyrosine hydroxylase is BH4- and ATP-dependent), VMAT2 loading (ATP-dependent proton gradient), and DA release are all ATP-expensive. Aripiprazole bypasses all of it — a partial agonist directly activates D2/D3 receptors regardless of endogenous DA levels. Aripiprazole response means the postsynaptic DA system is intact but presynaptic DA production cannot keep up with ATP demand.
- Guanfacine works → prefrontal noradrenergic signaling ATP-triage-deprioritized. PFC pyramidal neurons require NE for working memory maintenance through α2A receptor activation → cAMP suppression → HCN channel closure → strengthened synaptic connectivity. Guanfacine directly activates α2A receptors — bypassing the entire LC-NE synthesis and release pathway. Guanfacine response means the PFC circuits are intact and waiting for a signal that ATP shortage prevents the LC from sending.
- Pitolisant or modafinil works → histaminergic/orexin signaling ATP-triage-deprioritized. Orexin neurons are among the most metabolically active in the brain — their firing rate is directly proportional to ATP availability. Under energy shortage, orexin neurons reduce firing → reduced wakefulness drive. Pitolisant releases histamine (bypassing orexin), modafinil increases DA (bypassing orexin and histamine). Both provide wakefulness signals that the energy-depleted orexin system cannot generate.
Diagnostic pattern — the triage fingerprint:
- Midodrine + aripiprazole + pyridostigmine all work: Broad multi-system triage. Adrenergic, dopaminergic, AND cholinergic signaling are all ATP-deprioritized. This is the CNS energy crisis pattern — the brain has cut neurotransmitter production across multiple systems to conserve ATP for vital functions. The drug combination restores all three signaling pathways — but note the metabolic cost (Confirmation Bias in Patient Self-Reported Cycle-Phase Drug Response): each drug has its own ATP cost, and the combination may exceed the energy budget even though each individual drug produces benefit.
- Midodrine works alone (aripiprazole and pyridostigmine do not): Selective adrenergic triage. The ATP shortage is specifically affecting sympathetic NE synthesis — consistent with DBH being ATP-sensitive (DBH converts DA → NE, requiring ascorbate and copper, indirectly ATP-dependent for vesicular pH). DA and ACh systems have sufficient ATP budget; NE system does not. This is the neuropathic POTS pattern — selective sympathetic denervation, not global energy failure.
- Aripiprazole works alone (midodrine and pyridostigmine do not): Selective dopaminergic triage. The ATP shortage is specifically affecting DA synthesis — consistent with tyrosine hydroxylase being BH4-dependent and BH4 synthesis being GTP-dependent (GTP is ATP-derived). DA neurons are particularly vulnerable to ATP shortage because they have high baseline firing rates and require continuous BH4 regeneration.
- None of the above works: The dysfunction is NOT energetic — it is pharmacological (receptor blockade by autoantibodies, receptor desensitization, or structural receptor loss). The signaling systems have adequate ATP but the receptors cannot respond. This eliminates the energy triage model and directs diagnostic attention to GPCR AAb, channelopathy, or receptor-level pathology.
- Pacing reduces the number of bypass drugs needed: If pacing increases the ATP pool (by reducing demand → mitochondrial recovery → more ATP available for the next day) → fewer signaling systems are ATP-deprioritized. The patient who needs midodrine + aripiprazole + pyridostigmine while working full-time may only need midodrine when pacing reduces ATP demand. The reduction in required bypass drugs with pacing quantifies the ATP deficit — if pacing eliminates the need for two of three drugs → the energy deficit was demand-driven and partially reversible. If pacing changes nothing → the energy deficit is production-driven (structural mitochondrial damage) — ATP production is fixed regardless of demand.
Consequence: The energy triage model is the only cascade in this chapter that is NOT hypothesis-specific — it applies across all energy-deficit hypotheses (mitochondrial, delivery, substrate). The triage fingerprint (which bypass drugs work) identifies which signaling systems are ATP-depleted in that patient, and pacing’s effect on the fingerprint quantifies the degree to which the deficit is demand-driven vs. production-fixed. The most clinically actionable inference: if the patient requires 3 bypass drugs to function, their energy deficit is severe and their PEM budget for additional drug trials is minimal — prioritize low-energy-cost interventions. If pacing eliminates the need for bypass drugs, the treatment is pacing — not pharmacology. Origin: mechanistic-pathway-tracing.
3 Hormetic Multi-Target Dose-Response Optimization
Certainty: 0.25. The hormetic dose-response framework extends the cascade logic from single-mechanism tracing to multi-target drugs where individual mechanisms have non-overlapping dose optima. Classic cascade diagnostics assume a drug hits one target at one dose — LDN demonstrates the methodological problem: four mechanisms (TLR4/Nrf2, opioid compensatory, TRPM3, orexin) engage at different doses, and the dose-dependent diagnostic signal is as informative as the binary response/non-response. Triggered by Kevin Lee (personal communication, July 2026).
3.1 Cascade: LDN multi-target dose-response -> non-overlapping mechanism optima
Cascade:
- LDN (0.5–4.5 mg) hits four targets with different concentration-response curves:
- TLR4 partial blockade -> Nrf2-mediated M1->M2 microglial priming (hormetic peak at 0.5–1.5 mg)
- mu/delta opioid receptor blockade -> compensatory endorphin upregulation (ceiling at 1.5–3.0 mg set by precursor expression)
- TRPM3 calcium flux restoration (likely requiring 3.0–4.5 mg; single-concentration data only)
- Hypothalamic TLR4 microglial suppression -> reduced PGE2/TNF-alpha -> orexin disinhibition (tracks TLR4 dose-optimum)
- Dose-response is the diagnostic probe — the dose where benefit peaks identifies which mechanism is rate-limiting
Overall falsification condition: A prospective four-arm within-range dose-response trial (0.5, 1.5, 3.0, 4.5 mg, n at least 30, crossover, 8 weeks per dose) in ME/CFS must show (a) more than 50% of participants with non-monotonic individual dose-response curves AND (b) group-averaged decline from mid-dose to high-dose (3.0 to 4.5 mg) in at least 1 primary outcome. Falsified if more than 90% of individual curves are monotonic AND no group-level decline is observed at any dose transition — under these conditions, the multi-target dose-optimum model is wrong.
Step M1: TLR4/Nrf2 hormetic window — priming at low dose, loss at higher dose
- Mechanism: Partial TLR4 blockade at 0.5–1.5 mg removes a fraction of TLR4 activation — enough to reduce NF-kB-driven pro-inflammatory cytokine production (IL-1beta, TNF-alpha), but not enough to eliminate the basal TLR4 tone that maintains Nrf2 priming. The net effect is a shift from M1 (pro-inflammatory) to M2 (anti-inflammatory/repair) microglial phenotype, with a metabolic switch from glycolysis to OXPHOS (Kučić et al. 2021). At 3.0–4.5 mg, TLR4 blockade removes too much basal TLR4 tone — the Nrf2-mediated compensatory anti-inflammatory programme loses its priming signal and collapses, and the M1 to M2 shift reverses despite higher drug concentration. (Calabrese and Kozumbo 2021) (Calabrese and Giordano 2021)
- Intercept: Dose-response curve shape reveals Nrf2 transcriptional reserve. If benefit peaks at 0.5 mg and is lost by 1.5 mg -> narrow hormetic window, low Nrf2 reserve. If benefit persists from 0.5 to 3.0 mg -> broad hormetic window, high Nrf2 reserve. The width of the window predicts response to other Nrf2-activating drugs (sulforaphane, lithium, melatonin).
- Diagnostic: If the patient benefits at 0.5–1.5 mg and worsens at 3.0–4.5 mg -> TLR4/Nrf2 hormetic pathway is the dominant therapeutic mechanism. Neuroinflammation is TLR4-driven and the microglial population is Nrf2-responsive. The optimal dose is below 2 mg. This patient’s dose-response IS the diagnostic readout for their Nrf2 transcriptional reserve.
The hormetic window has been crossed → Nrf2 reserve is depleted by TLR4 over-blockade. This is NOT a failed titration — it identifies that the therapeutic mechanism is TLR4 hormetic priming, not TRPM3 restoration. Maintain at the effective low dose.
- Certainty
- Low to Medium
- Level of action
- Partial root cause
Step M2: Opioid compensatory upregulation — ceiling set by endogenous precursor expression
- Mechanism: LDN’s overnight opioid receptor blockade triggers compensatory beta-endorphin and met-enkephalin upregulation via transcriptional feedback. This mechanism has a biological ceiling: POMC and proenkephalin mRNA have maximum transcription rates set by promoter strength. (Boyadjieva et al. 2004) (Kreek 1992)
- Intercept: The dose at which benefit plateaus identifies the endorphin ceiling. If benefit rises from 0.5 to 1.5 mg and plateaus -> endorphin ceiling reached.
- Diagnostic: If benefit plateaus at 1.5 mg and is preserved at 4.5 mg (no worsening) -> opioid compensatory is the dominant mechanism, NOT TLR4 hormetic (which would worsen at 4.5 mg). Downgrade the dose to the plateau point.
Step M3: TRPM3 restoration — the dose-response unknown
- Mechanism: LDN restores TRPM3-mediated calcium flux in ME/CFS NK cells in vitro (Cabanas et al. 2018) — but the dose-response relationship for this mechanism is unknown. All published experiments used a single naltrexone concentration. Whether TRPM3 restoration requires low-dose TLR4 blockade (indirect, through reduced NF-kB-mediated miR-204), PIP2 preservation (Eaton-Fitch et al. 2021), or a direct naltrexone effect is unresolved.
- Intercept: If TRPM3 restoration requires TLR4 blockade -> TRPM3 dose-optimum maps to the TLR4 hormetic window (0.5–1.5 mg) — patients whose dominant mechanism is TRPM3 may need LOW doses. If TRPM3 restoration is PIP2-dependent and plateaued at any LDN dose that fully blocks TLR4 -> flat dose-response. If a direct naltrexone effect independent of TLR4/opioid -> unknown, may require concentrations above the clinical LDN range.
- Diagnostic: If LDN benefit requires 4.5 mg and no lower dose works -> TRPM3 or orexin mechanism is dominant. BUT: without a TRPM3 dose-response curve, this inference is tentative.
Step M4: Orexin disinhibition — hypothalamic microglial dose-response
- Mechanism: Hypothalamic microglia release PGE2 and TNF-alpha which tonically suppress orexin neuron firing. LDN’s TLR4 antagonism reduces this suppression. This mechanism’s dose-response tracks the TLR4 dose-response in the hypothalamus — which may differ from cortical or brainstem TLR4 populations.
Cascade resolution: The LDN dose-response is a four-dimensional optimization problem. A patient who benefits at 0.5 mg only -> TLR4/Nrf2 hormetic window dominant, narrow reserve. Monotonic benefit to 4.5 mg -> TRPM3 or opioid dominant. Biphasic benefit (M-shaped: benefit at 0.5, lost at 1.5, returns at 3.0) -> two non-overlapping optima: TLR4 hormetic at low dose + TRPM3 at higher dose. The dose-response curve SHAPE is diagnostic.
Consequence: LDN dose-finding is diagnostic, not just therapeutic. The dose-response curve maps which of LDN’s four mechanisms is rate-limiting in that patient. A patient whose optimal dose is 0.5 mg has a different disease subtype (TLR4/Nrf2-dominant) than one requiring 4.5 mg (TRPM3-dominant). Multi-target drugs with non-overlapping dose optima cannot be optimized by “start low, go slow” alone. The single largest evidence gap: TRPM3 dose-response data for LDN do not exist. (Dara et al. 2023) (Toljan and Vrooman 2018)
Origin: Kevin Lee clinical observation (July 2026) + literature integration.
3.1.1 Cascade: Time-dependent hormetic windows — why pulsing preserves therapeutic optima
Trigger: The existing LDN cascade traces dose-response curves (Steps M1–M4) but does not address the temporal dimension of those curves. The time-dependent hormesis literature (Calabrese corpus: (Sun et al. 2018) (Mushak 2016) (Calabrese 2016)) and GPCR resensitization kinetics ((Costa-Neto and Parreiras-E-Silva 2025) (Gupta, Mohan, and Naga Prasad 2018) (Kliewer, Reinscheid, and Schulz 2017)) jointly establish that the therapeutic window is a 3D volume (dose × time × response) — and that continuous daily dosing may extinguish it by filling the resensitization interval that the pathway requires.
Overall falsification condition: A prospective pulsed-vs-continuous dosing trial for at least one Nrf2-activating drug (sulforaphane, NAC, or melatonin) in ME/CFS must show superior maintenance of benefit at 12 weeks in the pulsed arm compared to the continuous arm. Falsified if no difference AND GPCR dephosphorylation half-life data for the relevant target confirm that the off-period is shorter than the half-life — i.e., the pulsing interval was too short to allow resensitization.
Cascade:
- Step T1 (Nrf2 extinguishment): Continuous Nrf2 activator dosing → tonic Keap1 cysteine oxidation → Nrf2/ARE transcriptional programme maintained → Nrf2 adaptive response desensitizes → protective gene expression declines despite continued drug presence
- Step T2 (GPCR resensitization blockade): Continuous agonist/receptor occupancy → receptors internalize in desensitized state → drug-free interval absent → dephosphorylation/recycling never occurs → receptor population shifts toward desensitized pool → tachyphylaxis
- Step T3 (Differential off-rate principle): Therapeutic target and adverse target have different dephosphorylation half-lives → intermittent dosing allows adverse receptor to resensitize during pause while therapeutic receptor remains modulated → therapeutic window maintained
- Step T4 (Oxidative stress amplification): High basal oxidative tone → ROS-mediated receptor oxidation accelerates desensitization → tachyphylaxis onset faster in ME/CFS than in general population → pulsed dosing more important in sicker patients
Step T1: Nrf2 adaptive response desensitization under continuous activator dosing
- Mechanism: The Keap1-Nrf2-ARE pathway is adaptive — it evolved to detect transient stress signals and mount a time-limited protective response. Under continuous activation (daily sulforaphane, NAC, or melatonin), the sensor desensitizes: Nrf2 protein half-life shortens, ARE-driven transcription saturates, and the anti-inflammatory gene programme plateaus or declines despite continued drug presence. This is distinct from the dose-dependent collapse described in Step M1 — here, the dose is correct (below 3.0 mg LDN) but the timing is wrong. The cell adapts to the chronic signal and downregulates its response. (Son, Camandola, and Mattson 2008) (Calabrese 2016)
- Intercept: If a patient loses benefit on an Nrf2 activator after weeks of daily dosing despite maintaining the correct dose, and benefit returns after a 2–3 day drug holiday → Nrf2 adaptive desensitization, not dose mismatch. The drug holiday serves as a resensitization interval.
- Diagnostic: Loss of benefit at unchanged dose over time NOT accompanied by worsening → adaptive desensitization (reversible). Loss of benefit at the SAME time across multiple dose escalations → dose-dependent extinction (Step M1). Loss of benefit that is permanent after discontinuation → irreversible desensitization or disease progression (exceeds this framework).
The cell’s Nrf2 machinery can resensitize. Intermittent dosing (e.g., 5 days on / 2 days off, or every other day) would preserve the therapeutic window. Nrf2 adaptive desensitization is documented in vitro; no clinical trial has tested pulsed Nrf2 activator dosing in any condition.
- Certainty
- Low to Medium
- Level of action
- Partial root cause
Either (a) the off-period was too short for Nrf2 resensitization, (b) the mechanism of benefit was not Nrf2-dependent (and tachyphylaxis is from a different pathway), or (c) the disease has progressed beyond the window where Nrf2 activation can compensate.
- Certainty
- Low
- Level of action
- Partial root cause
Step T2: GPCR resensitization blockade — the receptor-level mechanism
- Mechanism: GPCRs follow a conserved cycle: agonist binding → G protein activation → GRK phosphorylation → β-arrestin recruitment → receptor internalization → endosomal dephosphorylation by protein phosphatases → recycling to membrane (Gupta, Mohan, and Naga Prasad 2018) (Kliewer, Reinscheid, and Schulz 2017). Continuous receptor occupancy by a daily agonist blocks the dephosphorylation step: internalized receptors remain arrested in endosomes, unable to recycle to the membrane in a drug-naive state. The receptor population shifts from a drug-responsive surface pool to a desensitized internal pool. The critical parameter is the dephosphorylation half-life — if the drug-free interval is shorter than this half-life, resensitization never occurs. (Gupta, Mohan, and Naga Prasad 2018) (Kliewer, Reinscheid, and Schulz 2017)
- Intercept: Drugs whose therapeutic target is a GPCR (all catecholamine cluster drugs: modafinil/DAT, duloxetine/NET, beta-blockers/β1/β2, guanfacine/α2A, LDA/D2) are susceptible to GPCR resensitization blockade with continuous daily dosing. The resensitization half-life of each receptor determines whether “drug holidays” are feasible: half-life under 6 hours → overnight pause sufficient; half-life 6–24 hours → weekend holidays viable; half-life above 24 hours → prolonged holidays required, high withdrawal risk.
- Diagnostic: If benefit from a catecholamine-cluster drug declines over weeks at a stable dose, and a 1–2 day drug holiday restores benefit → GPCR resensitization blockade (reversible). If a drug holiday does NOT restore benefit AND the receptor has a known short dephosphorylation half-life (below 6 hours) → the mechanism is not GPCR desensitization.
β1 receptor dephosphorylation half-life is ~4 hours → off-period was adequate for resensitization. Pulsed dosing (5 days on / 2 days off) would maintain benefit without dose escalation. GPCR resensitization kinetics are well-established in vitro but have not been translated to clinical dosing trials in ME/CFS.
- Certainty
- Low
- Level of action
- Partial root cause
- Off-target confound: Drug holidays may produce withdrawal or rebound in susceptible patients (Hodding, Jann, and Ackerman 1980). Beta-blocker withdrawal can cause tachycardia and sympathetic overactivity — the drug holiday that restores receptor sensitivity is the same interval that triggers withdrawal symptoms. This is a genuine tension in the framework: the resensitization interval and the withdrawal risk interval overlap. Agents with low withdrawal risk (modafinil, guanfacine) are better candidates for pulsed dosing than agents with high withdrawal risk (beta-blockers, corticosteroids). (Hodding, Jann, and Ackerman 1980)
Step T3: Differential off-rate principle — rapamycin as paradigm
- Mechanism: Rapamycin binds FKBP12, and the FKBP12-rapamycin complex inhibits mTORC1 with high affinity and mTORC2 with lower affinity. After a single dose, mTORC1 inhibition persists for days (slow dissociation from FKBP12-mTORC1), while mTORC2 inhibition resolves faster. Intermittent dosing (every 5 days) exploits this differential: mTORC1 remains suppressed throughout the interval, but mTORC2 recovers between doses, preserving insulin sensitivity and immune function that continuous daily dosing impairs. (Arriola Apelo et al. 2016) (Konopka et al. 2023)
- Intercept: Intermittent rapamycin dosing (1–3 mg/week) preserves the therapeutic:mTORC2 adverse effect ratio observed in preclinical studies. Continuous daily dosing (reported in organ transplant protocols) produces immunosuppression and metabolic syndrome that intermittent dosing avoids.
- Diagnostic: If a patient tolerates weekly rapamycin but develops metabolic side effects on daily dosing → mTORC2 recovery during the off-period was preventing adverse effects. The therapeutic window is temporal, not dose-dependent — reducing the daily dose would not reproduce the benefit of intermittent dosing.
- Generalization principle: The differential off-rate model applies to any drug with dual targets that have different temporal recovery kinetics. Candidates include: corticosteroids (therapeutic GR modulation recovers faster than HPA axis suppression?), H1 antihistamines (peripheral H1 recovers faster than CNS H1?), DORAs (partial orexin blockade kinetics vs complete blockade). Each is a testable hypothesis; none has been tested.
Step T4: Oxidative stress amplification of tachyphylaxis — ME/CFS-specific vulnerability
- Mechanism: ROS generated by receptor activation can oxidize receptor cysteine residues, producing distinct functional outcomes — including accelerated desensitization that is independent of the canonical β-arrestin pathway (Teyani, Moghaddam, and Moniri 2024). In healthy tissue, antioxidant defenses clear these ROS before they oxidize the receptor. In ME/CFS, where baseline oxidative stress is elevated (documented across glutathione, SOD, lipid peroxidation, and isoprostane studies), receptor-level ROS may accumulate faster, accelerating tachyphylaxis. This creates a self-reinforcing loop: oxidative stress → faster receptor desensitization → reduced drug efficacy → symptom worsening → increased oxidative stress.
- Intercept: ME/CFS patients with the highest oxidative stress biomarkers (low GSH, high 8-isoprostane, high TBARS) may be the fastest to tachyphylax and the most in need of pulsed dosing strategies. This is a mechanistic prediction — the oxidative stress → tachyphylaxis link has been demonstrated for β2-adrenergic receptors in airway disease but not tested in ME/CFS for any receptor class.
- Diagnostic: If a patient with objectively elevated oxidative stress loses drug benefit faster than predicted by receptor pharmacology alone → ROS-mediated receptor oxidation may be contributing. Pulsed dosing reduces cumulative ROS exposure at the receptor — testing this prediction requires a head-to-head pulsed vs continuous trial with oxidative stress biomarkers as covariates.
- Consequence: If oxidative stress accelerates tachyphylaxis in ME/CFS, then the same drug at the same dose benefits a patient with low oxidative burden for months but a patient with high oxidative burden for only weeks. This would make oxidative stress a predictor of pulsing need — patients above a threshold benefit from intermittent dosing; patients below can maintain on continuous dosing. Certainty: 0.15 — ROS-mediated receptor tachyphylaxis is documented for one receptor class (β2AR) in airway disease; generalization to ME/CFS drug targets is speculative.
Cascade resolution: The dose-response curve is a 3D volume, not a 2D function. Time is the missing axis in the existing hormesis framework. For any drug whose therapeutic mechanism is hormetic (Nrf2 cluster) or GPCR-mediated (catecholamine cluster, LDA), the question is not just “what dose?” but “at what interval?” The resensitization clock is receptor-specific, and the optimal pulsing interval is the minimum off-period that allows the therapeutic receptor to resensitize without triggering withdrawal at the adverse receptor. This framework classifies drugs by resensitization kinetics and withdrawal risk — it is a scaffold for empirical testing, not a dosing algorithm.
Consequence: Pulsed dosing is not a fringe strategy — it is the logical consequence of taking GPCR resensitization kinetics and time-dependent hormesis seriously as pharmacological principles. The framework predicts that some ME/CFS patients who report “the drug stopped working after a few months” have not developed true tolerance but have filled the resensitization interval through continuous daily dosing. If a 2–3 day drug holiday restores benefit, the mechanism was adaptive desensitization, not disease progression. This is a testable, actionable, zero-cost intervention: stop the drug temporarily before concluding it has failed.
4 Innate Immune Hyperactivation → Systemic Inflammation → Fatigue/PEM
Certainty: 0.58. The Lipkin multi-site metabolomic + proteomic + ex-vivo microbial-stimulation study (Che 2025) (Che et al. 2025), consistent with the earlier Columbia plasma-immune-signature work (Hornig 2015) (Hornig et al. 2015), documents an exaggerated innate immune response to microbial antigens in ME/CFS that worsens after exercise and correlates with symptom severity. This entry routes that evidence through the differential-diagnostic cascade framework: heightened innate immune activation is one candidate driver of chronic inflammation, fatigue, and post-exertional malaise, and must be distinguished — pharmacologically — from the metabolically-primitive energy-triage model (Pacing prevents progressive decline) and from CNS-restricted neuroinflammation (Neuroinflammatory Hypotheses). The diagnostic value lies in which interception drugs work, not in the presence of inflammatory markers alone. Cross-reference the mechanistic background in Innate Immunity, Immune Activation and Inflammation, TLR4/NF-\(\kappa\)B Activation as PEM Cytokine Amplifier, and Integrated Hypothesis: The Multi-Lock Trap.
Consequence: Recognizing systemic innate-immune-driven inflammation as a possible upstream driver — and separating it from metabolic-energy-triage and brain-only inflammation — changes which interception drugs a clinician would trial first, and gives a research hypothesis for why some ME/CFS patients have elevated inflammatory markers while others do not.
4.1 Cascade: Innate immune hyperactivation → systemic inflammation → fatigue/PEM
#clinical-caution()
Cascade:
- Infectious / microbial-antigen trigger → exaggerated innate immune response (ex-vivo stimulation, Che 2025 (Che et al. 2025))
- Persistent TLR/NF-\(\kappa\)B + NLRP3-inflammasome / complement activation (TLR4/NF-\(\kappa\)B Activation as PEM Cytokine Amplifier, Innate Immunity)
- Systemic pro-inflammatory state (cytokines, lipid-linked inflammation, ECM/DAMP release) (Che et al. 2025)
- Downstream amplification: impaired TCA/beta-oxidation/urea-cycle energy production (Integrated Hypothesis: The Multi-Lock Trap), tryptophan→kynurenine diversion, redox imbalance
- Exercise precipitates exacerbation → fatigue + post-exertional malaise
- Differential task: distinguish systemic innate-immune-driven inflammation from metabolic energy-triage (Pacing prevents progressive decline) and from CNS-restricted neuroinflammation (Neuroinflammatory Hypotheses)
Step U1: Exaggerated innate immune response to microbial antigen
- Mechanism: ME/CFS leukocytes mount an exaggerated innate (TLR-driven) cytokine response to microbial stimulation that worsens after exercise and tracks symptom intensity; this sits upstream of — and may drive — the systemic inflammatory state documented across the Lipkin multi-site program (Che et al. 2025),(Hornig et al. 2015). The compartment caveat applies: evidence is measured ex-vivo in blood cells and plasma, inferred to represent systemic tissue inflammation. (Evidence source: PBMC/plasma — Inference target: systemic inflammation. Link is indirect.)
- Intercept: Corticosteroids (prednisone; broad suppression of NF-\(\kappa\)B-dependent cytokines) — suppresses the amplified response but is non-selective and has serious long-term risk in ME/CFS (medication reference); NSAIDs/celecoxib (COX-2, partial NF-\(\kappa\)B) — milder, symptomatic.
A partial response to a broad anti-inflammatory (corticosteroid or COX-2 inhibitor) — symptomatic relief without resolution — is consistent with an ongoing innate-immune-inflammatory driver but does NOT localize it: the same response would occur in metabolic-triage if inflammation is downstream (see Step U3). Used as a crude first probe only, given corticosteroids’ risks.
- Certainty
- Low
- Level of action
- Hypothesis
If the amplified response is insensitive to broad anti-inflammatories, systemic-innate-immune-driven inflammation is unlikely to be the rate-limiting driver — weigh metabolic energy-triage (Pacing prevents progressive decline) or CNS neuroinflammation (Neuroinflammatory Hypotheses) instead.
- Certainty
- Low
- Level of action
- Hypothesis
Step U2: Inflammasome / IL-1 axis (NLRP3, IL-1β)
- Mechanism: NLRP3-inflammasome assembly and downstream IL-1β drive sterile inflammation; complement activation is documented in ch08 (Innate Immunity, Immune Activation and Inflammation). Interception at this node is more mechanism-specific than the broad anti-inflammatory step above.
- Intercept: Colchicine (microtubule/NLRP3-inflammasome) — rate-limiting on inflammasome assembly, but narrow GI window; Anakinra (IL-1 receptor antagonist) — post-receptor IL-1 blockade; Canakinumab (anti-IL-1β monoclonal) — selective IL-1β blockade; all cross-reference medication reference.
Colchicine improves symptoms → the inflammasome is a dominant driver. If colchicine GI ceiling prevents dosing, anakinra/canakinumab response isolates whether IL-1β specifically is the effector. Failure of all three weights against an inflammasome-centric inflammatory driver.
- Certainty
- Low
- Level of action
- Hypothesis
Corticosteroid-sensitive but colchicine/anakinra/canakinumab null → the inflammation is cytokine-driven via a non-NLRP3, non-IL-1 path (e.g., TNF/NF-\(\kappa\)B dominant). Do NOT escalate biologics without this data.
- Certainty
- Low
- Level of action
- Hypothesis
Step U3: Downstream amplification (metabolic + CNS coupling)
- Mechanism: Innate-immune inflammation feeds the energy-crisis model (TCA/β-oxidation/urea-cycle impairment; tryptophan→kynurenine diversion from serotonin; redox/ECM imbalance) (Che et al. 2025),Integrated Hypothesis: The Multi-Lock Trap, and engages CNS microglial activation (Neuroinflammatory Hypotheses). This step connects the peripheral trigger to the symptom generators — but it is modeled as consequence, not independent cause.
- Intercept: LDN (low-dose naltrexone; TLR4 antagonist, systemic and glial) medication reference; Minocycline (microglial inhibitor, CNS-penetrant) medication reference. These probe the degree to which the peripheral innate-immune state propagates to CNS neuroinflammation.
LDN produces benefit and minocycline augments it → peripheral immune activity propagates to CNS microglial activation, linking the systemic trigger to the neuroinflammatory cascade.
- Certainty
- Low
- Level of action
- Partial root cause — systemic-to-CNS coupling
If LDN/minocycline are unhelpful or the fatigue is purely metabolic (energy-triage dominates), then inflammation is a marker of the energy crisis, not its driver — treat the metabolism and triage pattern, not the cytokines.
- Certainty
- Low
- Level of action
- Hypothesis
Consequence: This cascade gives clinicians a triaged, drug-testable sequence to separate three overlapping drivers of fatigue/PEM — systemic innate-immune inflammation (broad anti-inflammatory → inflammasome/IL-1 → LDN/minocycline), metabolic energy-triage (Pacing prevents progressive decline), and CNS neuroinflammation (Neuroinflammatory Hypotheses). Differential drug response is a research/clinical-judgment tool here, NOT a validated diagnostic instrument (see the caution in pharmacodiagnostic matrix); in the absence of a reliably stratifying inflammatory-marker panel it is one candidate way to separate these drivers, and none of these drugs should be deployed without the monitoring, interaction, and stopping guidance in medication reference. Severity applicability: unknown — the Che 2025 study population severity distribution is not stated; drug tolerability is poorest in severe/very-severe patients (worst for colchicine GI ceiling, corticosteroids, and biologics), so cross-reference ch14a before any trial.
Origin: /integrate-topic innate-immunity-che2025 (scoped micro-add — cascade-routing of already-integrated Che 2025 evidence; cross-references existing content, does not duplicate it).
5 Central Motor-Drive Fatigability Cascade
Certainty: 0.45. The Bedard 2026 multimodal-neuroimaging study (n=15 ME/CFS vs n=19 controls; grip force + EMG + EEG + fMRI + corticomuscular coherence) (Bedard et al. 2026) showed that ME/CFS patients generate equal maximum voluntary force yet fatigue earlier than controls. The key group difference was where the output did not change: healthy controls steadily increased muscle-fiber recruitment (Dimitrov index), cortical (EEG power), and subcortical (BOLD) activity across the sustained task and declined only slowly after exhaustion, whereas ME/CFS showed minimal fluctuation — an absent up-regulation of central motor output in the face of rising effort demand. This entry routes that evidence through the differential-diagnostic cascade framework: the earlier fatigability is modeled as a failure of the brain to increase motor drive to match demand (cortical output → corticospinal drive → spinal motoneuron → neuromuscular → muscle fiber → afferent feedback), rather than peripheral muscle exhaustion. It is distinguished — pharmacologically — from the metabolically-primitive energy-triage model (Pacing prevents progressive decline), from dopaminergic-effort loss of motivation (Cervical collar works — structural CCI component), and from brain-only neuroinflammation (Neuroinflammatory Hypotheses), because the association is central-output failure, not metabolic depletion, anhedonia, or inflammation per se. The diagnostic value lies in which central-acting interception drug restores motor-drive up-regulation — the cascade is a hypothesis and a research probe, NOT an established causal mechanism (the authors explicitly disclaim causation; only TMS or a drug-intervention study could establish direction). Cross-reference the objective-measurement finding in Central Neural Origin of Fatigability and the motor-imaging background in CSF Proteomics Without Controls.
Consequence: Reframing ME/CFS fatigability as a brain-level failure to drive muscle — rather than a purely peripheral muscle problem — changes which interception drugs a clinician would consider as research probes and gives a testable, drug-response framework for where in the motor path the bottleneck lies, but no drug is recommended on this associational evidence alone.
5.1 Cascade: Central motor drive fails to up-regulate under sustained effort
#clinical-caution()
Cascade:
- Sustained effort demand (fatiguing grip task) → cortical motor-output computation (effort required vs effort delivered) (Bedard et al. 2026)
- Corticospinal drive: motor-cortical and subcortical output should increase to maintain force — in ME/CFS it does not up-regulate (absent EEG-power and BOLD increase, minimal fluctuation)
- Spinal motoneuron recruitment / muscle-fiber recruitment: no compensatory increase in the Dimitrov index (unlike controls)
- Afferent feedback (corticomuscular coherence) fails to drive the output increase that would match demand
- Result: equal maximum voluntary force but earlier performance decline (fatigue onset)
- Differential task: distinguish central motor-drive failure (this cascade) from metabolic energy-triage (Pacing prevents progressive decline), dopaminergic-effort/motivation loss (Cervical collar works — structural CCI component), and CNS neuroinflammation (Neuroinflammatory Hypotheses)
Step C1: Cortical motor-output computation / central drive generation
- Mechanism: In healthy controls, sustained effort elicits a progressive increase in motor-cortical (EEG power) and subcortical (BOLD) activity that drives muscle-fiber recruitment to hold force; ME/CFS patients instead show minimal fluctuation in these signals — the brain fails to increase motor output to match rising effort demand (Bedard et al. 2026). This is an association observed during a fatiguing task, not an established causal mechanism; it localizes the failure to central output generation but does not identify why (energy, dopamine/effort, neuroinflammation, or maladaptive pacing are all upstream candidates). (Evidence source: scalp EEG + fMRI BOLD during a fatiguing grip task — Inference target: cortical/subcortical motor-output up-regulation. The compartment is intracranial, measured non-invasively; the link from reduced signal to reduced spinal drive is inferred.)
- Intercept: Central-acting wakefulness/effort promoters as research probes only — modafinil/armodafinil (weak DAT/NET inhibition, histamine/orexin activation; medication reference, (Minzenberg and Carter 2008)) and methylphenidate (stronger DAT/NET blockade; medication reference). A normalized motor-drive trajectory under a central probe would support a central-output bottleneck.
If a central-acting probe (modafinil, methylphenidate) restores the grip-force/EEG/BOLD trajectory — later fatigue onset, normalized up-regulation — central motor-output generation is implicated as the bottleneck; the specific drug that works then localizes the sub-node (see Step C2). A null response does NOT rule out central drive, because the failure may be upstream of, or unreachable by, the probed receptor class.
- Certainty
- Low
- Level of action
- Hypothesis
If a dopamine-targeting probe fails but the ME/CFS fatigability remains (equal MVC, early decline), central drive is not rate-limited by the probed dopaminergic/effort node — weight energy-triage (Pacing prevents progressive decline) or neuroinflammation (Neuroinflammatory Hypotheses) as the upstream cause, or probe a non-dopaminergic node (Step C3 amantadine).
- Certainty
- Low
- Level of action
- Hypothesis
Step C2: Dopaminergic / effort node (DAT/NET and D2/D3)
- Mechanism: Sustained effort engages dopaminergic circuits that convert “effort worth it” into motor output. A bottleneck here would make the brain under-prioritize sustained physical output even though maximum capability is intact (equal MVC) — matching the Bedard observation of an absent output up-regulation rather than a ceiling (Bedard et al. 2026). The dopaminergic-effort node is independently developed in the Gerlier four-pathway model Cervical collar works — structural CCI component and the dopamine-convergence hypothesis; it is the most pharmacologically accessible interception point in the motor-drive cascade.
- Intercept: Methylphenidate (DAT blockade; medication reference), modafinil/armodafinil (weak DAT/NET; medication reference), amphetamines/dextroamphetamine/lisdexamfetamine (DAT/NET + VMAT2 reversal; medication reference) probe reuptake/effort; bromocriptine/rotigotine (D2/D3 agonism; medication reference) probe the postsynaptic receptor. Each targets a distinct sub-node, so the pattern of response discriminates reuptake-blockade vs postsynaptic-receptor vs release.
If methylphenidate restores motor-drive up-regulation (later fatigue, normalized trajectory) while a D2/D3 agonist (bromocriptine/rotigotine) does not → the bottleneck is presynaptic reuptake/effort signaling rather than postsynaptic D2/D3 responsiveness. If a D2/D3 agonist works but reuptake blockade does not → the bottleneck is postsynaptic receptor sensitivity. Pattern discrimination, per the sec-12 pharmacodiagnostic entries.
- Certainty
- Low
- Level of action
- Hypothesis
If neither reuptake blockade nor D2/D3 agonism changes the fatigability trajectory → the motor-drive failure is not rate-limited at the dopaminergic-effort node; probe the glutamatergic node (Step C3) or re-weight toward energy-triage/neuroinflammation.
- Certainty
- Low
- Level of action
- Hypothesis
Step C3: Glutamatergic / release node (amantadine)
- Mechanism: Amantadine enhances presynaptic DA release (and NMDA modulation) rather than blocking reuptake, so it discriminates a release deficit from a reuptake bottleneck medication reference. A release-deficit motor-drive failure would show an absent output up-regulation that amantadine (but not a reuptake blocker) restores.
- Intercept: Amantadine (presynaptic DA release; medication reference).
If amantadine restores the motor-drive trajectory but methylphenidate does not → presynaptic release (not reuptake) is the bottleneck; if methylphenidate works but amantadine does not → reuptake is the bottleneck, release machinery intact.
- Certainty
- Low
- Level of action
- Hypothesis
If neither amantadine nor methylphenidate alters fatigability → the presynaptic dopaminergic machinery is not the rate-limiting step; the failure is upstream (energy/neuroinflammation) or postsynaptic-but-D2/D3-unresponsive — step outside the dopaminergic-effort node.
- Certainty
- Low
- Level of action
- Hypothesis
Step C4: Spinal motoneuron → neuromuscular → muscle-fiber recruitment
- Mechanism: Downstream of cortical drive, the absence of a Dimitrov-index increase indicates that spinal motoneuron recruitment / muscle-fiber recruitment does not compensate for declining force (Bedard et al. 2026). This is the efferent expression of the central failure: the muscle itself can generate maximum force (equal MVC) but is not being progressively driven to hold sub-maximal force under fatigue.
- Intercept: None directly at this node — peripheral ergogenic aids should NOT restore the trajectory if the bottleneck is central (a discriminating prediction). Cross-reference ch30 muscle/energy cascades for the peripheral-exhaustion alternative (Pacing prevents progressive decline).
If peripheral ergogenic aids (e.g., ergogenic supplements or peripheral conditioning) do not restore the fatigability trajectory, but a central-acting probe does → the bottleneck is central (this cascade). If a peripheral intervention alone restores performance → the peripheral-exhaustion model is favored, falsifying a purely central bottleneck.
- Certainty
- Low
- Level of action
- Hypothesis
Step C5: Afferent feedback / corticomuscular coherence
- Mechanism: The corticomuscular-coherence (afferent) arm of the Bedard data reflects that motor output is not being corrected by sensory feedback from the fatiguing muscle; the loop that would normally up-regulate drive fails (Bedard et al. 2026). This is the least-falsified, least-probed node and the most speculative step of the cascade.
- Intercept: None established; no drug directly targets afferent-feedback gain. This node is a research-measurement target (corticomuscular coherence) rather than an interception point.
Corticomuscular-coherence changes under a central probe would indicate whether the afferent-feedback correction of drive is a separate, drug-sensitive node or simply a readout of the reduced output. Untested; measurement-only.
- Certainty
- Low
- Level of action
- Hypothesis
Consequence: This cascade gives researchers and clinicians a triaged, drug-testable sequence to separate a brain-level failure to drive muscle (cortical output → dopaminergic-effort node → glutamatergic-release node) from peripheral muscle exhaustion, metabolic energy-triage (Pacing prevents progressive decline), and CNS neuroinflammation (Neuroinflammatory Hypotheses). Differential drug response is a research tool here, NOT a validated diagnostic instrument (see the caution in pharmacodiagnostic matrix); each probe requires the monitoring, interaction, and stopping guidance in medication reference, and none of these drugs is recommended on this associational evidence alone. Severity applicability: unknown — the Bedard 2026 NIH intramural cohort is not stratified by severity; the probes discussed (stimulants, dopamine agonists) are poorest-tolerated in severe/very-severe patients, so cross-reference ch14a before any trial.
Origin: /integrate-topic bedard2026-central-fatigability-cascade (gap-fill of the prior fatigue-cycle ch30-tier NONE decision; routes the already-integrated Central Neural Origin of Fatigability finding through the ch34 cascade framework without duplicating it).
The objective-measurement finding (Central Neural Origin of Fatigability, cert 0.60) and this motor-drive cascade (cert 0.45) converge on a single framing: ME/CFS fatigability is centrally mediated — the brain fails to increase motor output to match rising effort demand, rather than the muscle being exhausted. The convergence is functional and not merely statistical: equal maximum voluntary force (muscle capable), earlier fatigue (performance declines sooner), and an absent up-regulation of cortical/subcortical output and muscle-fiber recruitment (the brain does not push harder) jointly argue against peripheral muscle exhaustion as the primary limiter. The same protocol’s differential-drug logic then gives a research pathway to localize why: dopaminergic/effort-node interception (methylphenidate, modafinil, amphetamines, bromocriptine/rotigotine) and glutamatergic release (amantadine) each test a distinct cascade node, while a peripheral ergogenic aid tests the negative (central) prediction. This is a research and diagnostic-testing model, NOT a validated mechanism or a treatment basis — causation remains unestablished (the authors disclaim it), the upstream cause (energy, dopaminergic effort, neuroinflammation, or maladaptive pacing) is unresolved, and severity applicability is unknown. Its value is to give ME/CFS fatigability a central, drug-testable, falsifiable structure that separates central-drive failure from metabolic energy-triage (Pacing prevents progressive decline) and CNS neuroinflammation (Neuroinflammatory Hypotheses).
Consequence: If this convergent central-drive model holds, ME/CFS fatigability is a brain-level phenomenon whose bottleneck can be probed by which central-acting agent (if any) restores motor-drive up-regulation — a testable research direction that steers treatment research toward the central nervous system, though it changes no clinical recommendation today.