Ion Channel Hypotheses

1 TRPM3 Channelopathy

Certainty: 0.55. The TRPM3 channel defect is the most replicated ion channel finding in ME/CFS (six independent cohorts, (Cabanas et al. 2021)). TRPM3 is a non-selective cation channel gated by PIP2, pregnenolone sulfate, and heat. Its primary physiological role: calcium influx in NK cells (cytotoxicity), pancreatic Ξ²-cells (insulin secretion), vascular smooth muscle (tone), and sensory neurons (heat/pain detection). Multi-system dysfunction is the expected consequence of impaired calcium influx across these cell types. The cascade traces PIP2-dependent gating failure through to terminal consequences, with each node offering a pharmacologically distinguishable probe.

1.1 Cascade: PIP2 β†’ TRPM3 β†’ five-arm end-organ dysfunction

Cascade:

  • PIP2 depletion (GPCR AAb β†’ PLC-Ξ² hydrolysis)
  • TRPM3 calcium influx failure
    1. NK cell cytotoxicity loss β†’ impaired viral clearance, immune surveillance
    1. pancreatic Ξ²-cell insulin secretion impairment β†’ postprandial energy dysregulation
    1. neurotransmitter release impairment at autonomic ganglia β†’ orthostatic intolerance, GI dysmotility
    1. vascular smooth muscle calcium dysregulation β†’ impaired vasoconstriction and vasodilation
    1. sensory neuron TRPM3 dysfunction β†’ altered pain and temperature perception
  • Each arm has a distinct probe.

Step A1: PIP2 depletion (membrane-level trigger β€” upstream of all TRPM3 dysfunction)

  • Mechanism: GPCR autoantibodies (Ξ²2-AR, M2/M4) activate GΞ±q β†’ PLC-Ξ² β†’ PIP2 hydrolysis to IP3 + DAG. PIP2 is the essential cofactor for TRPM3 gating β€” without PIP2 in the inner membrane leaflet, TRPM3 cannot open regardless of agonist (pregnenolone sulfate, heat). PIP2 depletion is the most parsimonious single mechanism for simultaneous TRPM3, TRPV1, TRPM7, and TRPC channel dysfunction β€” all are PIP2-dependent.
  • Intercept: Lithium (low-dose, 2–20 mg elemental; IMPase inhibitor, mood stabilizer at microdose) β€” inhibits IMPase, the enzyme that degrades IP2 β†’ IP1, conserving the inositol pool for PIP2 resynthesis Lithium Safety: Drug Interactions and Contraindications. This is the most upstream probe in the entire chapter β€” lithium intercepts before any channel, at the membrane phospholipid level.
ImportantFinding: PIP2 depletion is rate-limiting.

Lithium works β†’ PIP2 depletion is rate-limiting. The lesion is at membrane phospholipid regulation β€” upstream of all ion channels. GPCR autoantibody-driven PLC hyperactivity is the likely upstream driver. If lithium works AND immunoadsorption works β†’ the causal chain is confirmed at both ends (remove the AAb that drives PLC, and separately conserve the PIP2 that PLC depletes).

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: PIP2 depletion is not the bottleneck.

Lithium does NOT work β†’ PIP2 depletion is not the bottleneck. Either (a) PIP2 levels are normal (AAb are not driving sufficient PLC activity), (b) PIP2 resynthesis capacity is saturated, (c) TRPM3 has additional defects beyond PIP2-dependence (direct channel phosphorylation, altered expression, or channel-blocking autoantibodies), or (d) the lesion is downstream of PIP2 availability β€” calcium handling inside the cell, not calcium entry at the membrane.

Certainty
Low to Medium
Level of action
Partial root cause
  • Lithium dosing gradient: If benefit at 2–5 mg but lost at 20 mg β†’ inverted U (IMPase inhibition at low dose, inositol depletion at higher doses). The dose at which benefit inverts is consistent with inositol pool size β€” inverts earlier with a smaller pool, later or never with a larger one.

Step A2: TRPM3 calcium influx failure (channel-level defect)

  • Mechanism: Even with adequate PIP2, TRPM3 may fail to gate due to: (a) direct channel-blocking autoantibodies (anti-TRPM3 IgG, not yet documented but mechanistically possible), (b) altered channel expression (TRPM3 gene silencing via promoter methylation), (c) competitive inhibition at the pregnenolone sulfate binding site by endogenous antagonists, or (d) channel desensitization from chronic low-grade activation (heat, inflammation β†’ sustained TRPM3 opening β†’ calcium-dependent inactivation).
  • Intercept: Pregnenolone sulfate (positive allosteric TRPM3 modulator; neurosteroid) β€” shifts the channel’s activation curve leftward so lower PIP2/heat can open it) Pregnenolone + LDN Synergy; Naltrexone (LDN, 1–4.5 mg; low-dose opioid receptor antagonist, TLR4 antagonist) β€” in vitro TRPM3 restoration via TLR4 antagonism reducing channel-destabilizing inflammatory signaling (Cabanas et al. 2021).
ImportantFinding: TRPM3 dysfunction is present and rate-limiting.

LDN or pregnenolone works β†’ TRPM3 dysfunction is present AND rate-limiting. The lesion is at the channel level (primary channelopathy) or just above (PIP2-dependent gating failure). Distinguishing: if lithium also works β†’ the lesion is at PIP2; TRPM3 can gate when given PIP2. If lithium does NOT work but LDN/pregnenolone works β†’ the channel itself is the bottleneck β€” PIP2 depletion is not the rate-limiting node.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: TRPM3 is the primary LDN target.

LDN works within 48 hours β†’ consistent with TRPM3 mechanism (immediate channel gating effect). Onset latency suggests TRPM3 as the primary LDN target in this patient β€” however, placebo and expectation effects also onset within this window limiting specificity.

Certainty
Low
Level of action
Partial root cause
ImportantFinding: LDN benefit is from endorphin/TLR4 mechanisms, not TRPM3.

LDN works but takes 3–6 weeks β†’ the mechanism is endorphin upregulation or TLR4-mediated anti-neuroinflammatory effect β€” NOT TRPM3. The ion channel hypothesis is not confirmed. This dissociation is clinically relevant: a patient with LDN benefit at 6 weeks may have neuroinflammation, not ion channel dysfunction, and different follow-up probes are indicated.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: TRPM3 ion channel dysfunction is absent, not LDN-responsive, not rate-limiting, or downstream machinery damaged.

LDN/pregnenolone does NOT work β†’ TRPM3 ion channel dysfunction is either (a) absent in this patient, (b) present but not LDN-responsive (direct channel-blocking autoantibodies that LDN cannot displace, requiring immunoadsorption), (c) not rate-limiting β€” other channel defects dominate (TRPV1, TRPM7, or calcium channel autoantibodies), or (d) the downstream cellular machinery is so damaged that restoring Ca²⁺ influx does not restore function. A null TRPM3 response combined with a null lithium response effectively eliminates the PIP2β†’TRPM3 ion channel axis.

Certainty
Low
Level of action
Partial root cause

Step A3: NK cell cytotoxicity loss (immune consequence)

  • Mechanism: NK cells require Ca²⁺ influx through TRPM3 for lytic granule polarization and exocytosis β€” the killing mechanism. TRPM3 failure β†’ impaired NK cell degranulation β†’ reduced perforin/granzyme release β†’ impaired viral clearance and tumour surveillance. This is one of the most consistently replicated immune findings in ME/CFS.
  • Intercept: Immunoadsorption (extracorporeal IgG removal; autoantibody depletion) β€” removes GPCR autoantibodies upstream, potentially restoring PIP2 and thus TRPM3; IVIG (pooled donor immunoglobulins; broad immunomodulation, Fc receptor modulation) β€” broad immunomodulation including NK cell enhancement via Fc receptor modulation; Cimetidine (H2 receptor antagonist; T-cell enhancer, CYP450 inhibitor) β€” enhances T-cell and NK cell function through H2 receptor blockade on suppressor T cells (disinhibits cytotoxic function); Inosine pranobex (Isoprinosine, immunostimulant; NK cell enhancer) β€” enhances NK cell activity through undefined mechanism.
ImportantFinding: Immune dysfunction is at T-cell/H2-receptor level, not NK/TRPM3.

Cimetidine works but LDN does NOT β†’ the immune dysfunction is at T-cell/H2-receptor level, not NK/TRPM3-dependent. Cimetidine disinhibits cytotoxic function by blocking H2-mediated suppression β€” the pathway is intact but suppressed. Consistent with histamine-driven immune suppression rather than ion-channel-driven NK failure.

Certainty
Low
Level of action
Partial root cause
ImportantFinding: Autoantibody-mediated pathology is likely dominant.

IVIG works β†’ autoantibody-mediated pathology (upstream of TRPM3) is likely dominant. The channelopathy is secondary β€” restoring PIP2 through AAb removal restores TRPM3 function, and NK cell cytotoxicity follows. IVIG response + LDN response = multi-level immune restoration.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: Full PIP2β†’TRPM3β†’NK axis is causal.

Both IVIG and LDN work β†’ both upstream (autoantibody β†’ PIP2 restoration via IA) and downstream (TRPM3 gating via LDN) interventions are effective. Therapeutic priority: treat the autoantibody source for root cause (rituximab, IA, BC007), LDN for symptomatic channel support while awaiting source treatment to take effect. The dual response confirms the full PIP2β†’TRPM3β†’NK axis as causal.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: The NK defect is at the channel level β€” channelopathy is primary.

IVIG does NOT work but LDN works β†’ the NK defect is at the channel level β€” LDN restores TRPM3 gating directly at NK cells (where TRPM3 and TLR4 are co-expressed). The channelopathy is primary, not secondary to AAb. Immunoadsorption would not help because PIP2 is not the limiting factor.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: NK defect is structural or downstream of Ca²⁺ influx β€” exhausted NK cells.

Neither LDN nor IVIG works β†’ the NK cell defect is either (a) structural β€” NK cell numbers are reduced (common in ME/CFS), not just functionally impaired; restoring TRPM3 gating cannot rescue cytotoxicity if there are too few NK cells, or (b) the lesion is downstream of Ca²⁺ influx β€” the lytic granule machinery (Rab27a, Munc13-4, syntaxin-11) is intact but not responsive to calcium signal. This pattern is consistent with exhausted/depleted NK cells, not channel-inhibited NK cells.

Certainty
Low
Level of action
Partial root cause

Step A4: Neurotransmitter release impairment (neurochemical consequence)

  • Mechanism: Calcium-dependent vesicle exocytosis at autonomic ganglia and CNS synapses requires TRPM3-mediated Ca²⁺ entry (and Cav2.x channels). While many synapses rely on voltage-gated calcium channels for fast transmission, TRPM3 provides sustained Ca²⁺ elevation for vesicle pool replenishment and neuropeptide release (substance P, CGRP, NPY β€” pain, vasodilatory, and sympathetic co-transmitters respectively). TRPM3 failure β†’ impaired sustained neurotransmitter release β†’ ganglionic transmission failure β†’ orthostatic intolerance (sympathetic ganglia), GI dysmotility (enteric and vagal), and cognitive dysfunction (CNS).
  • Intercept: Pyridostigmine (Mestinon, 30–60 mg; acetylcholinesterase inhibitor, peripheral cholinergic) β€” acetylcholinesterase (AChE) inhibitor, enhances cholinergic signalling by increasing synaptic acetylcholine (ACh) dwell time, compensating for reduced release probability at autonomic ganglia; Atomoxetine (norepinephrine reuptake inhibitor; NRI) β€” NRI, increases norepinephrine (NE) at synapse by blocking reuptake, compensating for reduced NE release; 5-HTP (serotonin precursor) β€” serotonin precursor, bypasses tryptophan-dependent synthesis if serotonin depletion coexists.
ImportantFinding: ACh-dependent synaptic transmission at autonomic ganglia is impaired.

Pyridostigmine works β†’ acetylcholine-dependent synaptic transmission at autonomic ganglia is impaired β€” consistent with TRPM3-mediated Ca²⁺-dependent vesicle release failure at the synapse. Pyridostigmine partially compensates by increasing the probability that any single release event produces postsynaptic activation. The improvement is functional, not corrective β€” the underlying Ca²⁺ deficit remains.

Certainty
Low to Medium
Level of action
Symptom management
ImportantFinding: TRPM3 lesion is away from autonomic ganglia β€” LDN benefit is non-synaptic.

Pyridostigmine does NOT work and LDN works β†’ TRPM3 dysfunction exists (LDN responsive) but the bottleneck is upstream of cholinergic transmission β€” LDN benefit is through a non-synaptic mechanism: TRPM3 on immune cells (NK cells, mast cells), pancreatic Ξ²-cells, or vascular smooth muscle. The synaptic cholinergic system is either intact (sufficient Ca²⁺ from Cav channels) or too damaged to respond to AChE inhibition. This dissociation localizes the TRPM3 lesion away from autonomic ganglia.

Certainty
Low
Level of action
Partial root cause
ImportantFinding: Neurotransmitter release deficit is noradrenergic-specific.

Atomoxetine works but pyridostigmine does NOT β†’ the neurotransmitter release deficit is specific to the noradrenergic system β€” NE reuptake inhibition compensates. Cholinergic synapses are intact (pyridostigmine null means ACh release is adequate). This pattern distinguishes NE-specific from global Ca²⁺-dependent release failure. Consistent with selective noradrenergic defect rather than global synaptic TRPM3 dysfunction.

Certainty
Low to Medium
Level of action
Symptom management
ImportantFinding: Neurotransmitter deficit is serotonergic, not noradrenergic.

5-HTP works but atomoxetine does NOT β†’ the neurotransmitter deficit is serotonergic, not noradrenergic β€” consistent with TRPM3 on raphe neurons (serotonin synthesis/release) being more affected than TRPM3 on locus coeruleus neurons (NE). The TRPM3 phenotype has cell-type specificity.

Certainty
Low
Level of action
Symptom management

Step A5: Vascular smooth muscle calcium dysregulation (orthostatic consequence)

  • Mechanism: Vascular smooth muscle contraction depends on Ca²⁺ influx through L-type Ca²⁺ channels (Cav1.2) and TRPM3 (which provides sustained Ca²⁺ for myosin light chain kinase activation). TRPM3 is co-expressed with Cav1.2 in resistance arteries. TRPM3 failure β†’ impaired sustained vasoconstriction β†’ venous pooling, orthostatic hypotension, and compensatory tachycardia.
  • Intercept: Midodrine (Ξ±1-adrenergic receptor agonist, 2.5–10 mg; vasoconstrictor) β€” directly constricts vasculature through GΞ±q β†’ PLC β†’ IP3 β†’ Ca²⁺ release from sarcoplasmic reticulum, bypassing membrane Ca²⁺ channels entirely; Ivabradine (I_f channel blocker; SA node heart rate reduction) β€” slows HR through If current block at SA node, no Ca²⁺ channel dependence; Fludrocortisone (mineralocorticoid; volume expansion, sodium retention) β€” volume expansion, reduces the vasoconstrictive demand.
ImportantFinding: Vasculature is responsive β€” deficit is at membrane Ca²⁺ entry.

Midodrine works β†’ the vasculature is responsive to direct Ξ±1 agonism β€” the contractile machinery (myosin light chain kinase, actin-myosin) is functional. The deficit is at the membrane level: insufficient Ca²⁺ entry for sustained contraction. Either sympathetic output is low (CNS catecholamine deficiency β†’ reduced Ξ±1 activation β†’ reduced Cav/TRPM3 opening) or vascular Ca²⁺ handling is directly impaired (TRPM3 dysfunction). The positive midodrine response confirms that the postsynaptic vasculature can contract β€” the problem is the signal to contract, not the contractile apparatus.

Certainty
Low to Medium
Level of action
Symptom management
ImportantFinding: Multi-channel dysfunction from PIP2 depletion β€” consider immunoadsorption.

Midodrine works AND LDN works β†’ both Ca²⁺-handling (TRPM3 at vascular smooth muscle) and vascular tone are abnormal. Consistent with multi-channel dysfunction from PIP2 depletion β€” the same PIP2 deficit affects TRPM3 (on NK cells, neurons, and smooth muscle) and potentially other PIP2-dependent channels. Immunoadsorption should be considered β€” if PIP2 depletion is driving multi-channel dysfunction, removing the AAb that drives PLC should restore PIP2 and all channel functions simultaneously.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: Vasculature is unresponsive to Ξ±1 stimulation.

Midodrine does NOT work β†’ the vasculature is unresponsive even to direct Ξ±1 stimulation. This is a severe finding β€” it means either (a) Ξ±1 receptors are absent or blocked (GPCR AAb against Ξ±1-AR), (b) the contractile machinery is damaged (smooth muscle atrophy from chronic disuse, nitric oxide overproduction from endothelial dysfunction), or (c) severe hypovolemic orthostatic intolerance where vasoconstriction is inadequate because there is insufficient blood volume to distribute. If fludrocortisone works where midodrine fails β†’ hypovolemia is dominant; the vasculature is functional but empty. If neither works β†’ the vasculature is structurally unable to contract; endothelial dysfunction or receptor-level blockade is likely.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: Inappropriate sinus tachycardia is dominant, not vascular failure.

Ivabradine works for HR but midodrine does NOT work for BP β†’ inappropriate sinus tachycardia is the dominant hemodynamic symptom, not vascular failure. The SA node is responding to a baroreflex signal that says β€œpressure is low,” but the vasculature cannot increase pressure. Slowing the HR reduces myocardial oxygen demand without worsening perfusion (because perfusion was already inadequate). This pattern is consistent with baroreflex reset (GPCR AAb at NTS) combined with vascular unresponsiveness.

Certainty
Low
Level of action
Symptom management

Consequence: TRPM3 is the most pharmacologically probe-able ion channel in ME/CFS. Ten drugs at five distinct steps (lithium at PIP2, LDN/pregnenolone at channel, cimetidine/IVIG at NK cells, pyridostigmine/atomoxetine/5-HTP at neurotransmitter release, midodrine/ivabradine at vasculature) map the lesion along the PIP2β†’TRPM3β†’cell-type axis. The pattern of which drugs work and which do not localizes whether the defect is upstream (PIP2, treat AAb source), at the channel (TRPM3 gating, treat with LDN/pregnenolone), or cell-type-specific (NK cells β†’ immunomodulation; synapses β†’ neurotransmitter support; vasculature β†’ vasoconstrictors). The onset latency of LDN response (under 48h suggests TRPM3, 3–6 weeks suggests endorphin/TLR4 β€” with limited specificity; placebo and expectation effects also onset within 48h) further distinguishes channel from inflammatory mechanisms. Origin: mechanistic-pathway-tracing, Gerlier 2026-07-21 prompt.

2 TRPV1-Mediated Arteriolar Vasoconstriction (PEM-Driven Pain/Ischemia)

Certainty: 0.50. TRPV1 is functionally expressed on skeletal muscle arteriolar smooth muscle (Cavanaugh 2011) and on sensory C-fibers innervating muscle. Its canonical activators β€” protons (H⁺ from lactic acid), heat (from muscle activity), ROS (from mitochondrial electron leak), and lipid peroxides (from membrane damage) β€” are all produced during exertion. In healthy muscle, TRPV1 activation on sensory fibers signals β€œmuscle is working” to the CNS. In ME/CFS, the hypothesis is: TRPV1 on arteriolar smooth muscle causes vasoconstriction (paradoxical in working muscle, which should vasodilate), producing ischemia. The ischemia produces more metabolites β†’ more TRPV1 activation β†’ further vasoconstriction β†’ a positive feedback loop of ischemia and pain that constitutes the PEM sensation. COX-2/PGE2 amplification loop is well-established in pain biology β€” PGE2 sensitizes TRPV1, lowering its activation threshold. Direct ME/CFS measurement of arteriolar TRPV1 pending.

2.1 Cascade: Exercise β†’ TRPV1 vasoconstriction β†’ ischemic PEM loop

Cascade:

  • Exercise β†’ (1) lactate/H⁺ production, (2) heat generation, (3) mitochondrial ROS, (4) lipid peroxidation from membrane damage
  • TRPV1 activation on skeletal muscle arteriolar smooth muscle
  • Paradoxical vasoconstriction (TRPV1-mediated Ca²⁺ influx β†’ contraction)
  • Tissue ischemia β†’ more metabolite accumulation
  • Further TRPV1 sensitization via COX-2/PGE2 (PGE2 lowers activation threshold)
  • Positive feedback loop: ischemia β†’ metabolites β†’ TRPV1 β†’ vasoconstriction β†’ more ischemia
  • Severe localized muscle pain, prolonged recovery (the PEM sensation), impaired oxygen delivery during subsequent exertion

2.2 Step B1: TRPV1 activation on arteriolar smooth muscle

Mechanism: TRPV1 on vascular smooth muscle is different from TRPV1 on sensory neurons. On sensory neurons, TRPV1 activation produces pain (the burning sensation of capsaicin). On vascular smooth muscle, TRPV1 activation produces contraction (Ca²⁺ influx β†’ myosin light chain kinase β†’ vasoconstriction). The drug probes must distinguish these two TRPV1 populations β€” a drug that blocks TRPV1 on sensory neurons (reducing pain) may also block TRPV1 on vascular smooth muscle (preventing vasoconstriction β€” beneficial), or it may only block one population depending on tissue distribution and drug access.

Probes:

  • Capsaicin (topical, 0.025–0.075%, applied to painful muscles): TRPV1 agonist at low concentration β€” activates then desensitizes TRPV1 on sensory C-fibers. Repeated application depletes substance P and CGRP from sensory terminals, producing prolonged analgesia. If topical capsaicin reduces localized muscle pain after exertion β†’ the pain is TRPV1-mediated on sensory fibers. Does NOT address the vascular TRPV1 component (topical capsaicin does not reach arteriolar smooth muscle in meaningful concentrations). Capsaicin response = sensory TRPV1 confirmed. Capsaicin null = pain is not TRPV1-mediated on sensory fibers (it is from ischemia, acidosis directly, or central sensitization).
  • NEO6860 (selective TRPV1 antagonist, Phase II for osteoarthritis): Blocks TRPV1 without the initial activation that capsaicin causes. Systemic administration would block both sensory and vascular TRPV1. If NEO6860 reduces PEM severity and duration β†’ TRPV1 on BOTH sensory and vascular smooth muscle is rate-limiting. If NEO6860 reduces pain but not PEM duration β†’ TRPV1 is involved in pain perception but not in the ischemic feedback loop β€” vascular TRPV1 is not the dominant mechanism.
  • If capsaicin works and NEO6860 is unavailable: This is the clinical reality for most patients. Capsaicin response confirms TRPV1-mediated pain but cannot assess vascular TRPV1. The distinction matters for treatment: if pain is purely sensory TRPV1 β†’ topical capsaicin is sufficient. If vascular TRPV1 is involved β†’ systemic TRPV1 antagonism (NEO6860 or future drugs) would be needed for the ischemic component.
  • If capsaicin/NEO6860 does NOT reduce PEM: TRPV1-mediated vasoconstriction is not rate-limiting. Other vasoconstrictive mediators may dominate: angiotensin II (RAAS dysregulation), endothelin-1 (endothelial dysfunction), or ischemia is from pre-capillary obstruction (microclots, RBC stiffness) rather than arteriolar constriction. Distinguish: if pyridostigmine improves PEM but capsaicin does not β†’ the ischemia is flow/delivery-limited (microvascular), not TRPV1-mediated. If midodrine improves PEM but capsaicin does not β†’ the ischemia is perfusion pressure-limited.

2.3 Step B2: COX-2/PGE2 amplification (the feed-forward loop)

Mechanism: COX-2 is induced by inflammatory cytokines (IL-1Ξ², TNF-Ξ±) and by TRPV1 activation itself (Ca²⁺ influx β†’ PLA2 activation β†’ arachidonic acid release β†’ COX-2 substrate). COX-2 produces PGE2 β†’ PGE2 binds EP1 and EP4 receptors on TRPV1-expressing neurons and smooth muscle β†’ PKA and PKC phosphorylation of TRPV1 β†’ lowers TRPV1 thermal and proton activation threshold. The result: after one episode of exertion-induced TRPV1 activation, subsequent TRPV1 activation occurs at lower stimulus intensities β†’ the PEM threshold drops with repeated exertion (the Day 2 CPET phenomenon).

Probes:

  • Celecoxib (COX-2 inhibitor, 100–200 mg/day): Blocks COX-2 selectively, reducing PGE2 production and the sensitization loop. If celecoxib reduces PEM severity on Day 2 of a 2-day CPET β†’ COX-2/PGE2 sensitization is driving the PEM exacerbation. If celecoxib reduces pain during exertion but not PEM recovery β†’ COX-2 is involved in pain signaling (PGE2 sensitizes nociceptors) but not in the ischemic feedback loop β€” pain and PEM are distinct TRPV1 consequences.
  • Etoricoxib (30–60 mg/day, COX-2 selective, longer half-life): Similar to celecoxib but once-daily dosing. If etoricoxib works where celecoxib does not β†’ dosing schedule matters β€” continuous COX-2 suppression is needed for the sensitization loop.
  • CAUTION β€” COX-2 and resolution: COX-2 also produces pro-resolution lipid mediators (lipoxins, resolvins) from omega-3 fatty acids. Blocking COX-2 may impair inflammation resolution. If celecoxib improves pain acutely but worsens fatigue over weeks β†’ the COX-2-derived pro-resolution mediators were limiting neuroinflammation, and blocking them allows inflammation to accumulate. The temporal pattern (acute benefit, chronic worsening) distinguishes COX-2 as pathogenic (benefit persists) vs. COX-2 as protective (benefit fades, symptoms worsen).
  • LDN (TLR4 antagonist): Indirect COX-2 suppression β€” LDN reduces TLR4-mediated NF-ΞΊB activation β†’ reduces COX-2 transcription. If LDN reduces PEM β†’ the TLR4β†’NF-ΞΊBβ†’COX-2β†’PGE2 pathway is driving TRPV1 sensitization. LDN response + celecoxib response = confirmation that the inflammatoryβ†’COX-2β†’TRPV1 axis is causal.
  • Omega-3 fatty acids (EPA/DHA, 2–4 g/day): Provide substrate for pro-resolution lipid mediators (resolvins, protectins) through COX-2 and LOX pathways. If omega-3 + celecoxib is more effective than celecoxib alone β†’ the COX-2 pathway was being used for both pro-inflammatory (PGE2) and pro-resolution (resolvin) synthesis, and providing omega-3 substrate shifts the balance toward resolution even with COX-2 partially blocked.

2.4 Step B3: Ischemia-induced metabolite accumulation (lactate, ROS, protons, lipid peroxides)

Mechanism: Ischemic muscle produces the same metabolites that activate TRPV1. Even if TRPV1 is the initial trigger for vasoconstriction, the ischemia it produces generates metabolites that further activate TRPV1 at the same or adjacent arterioles β†’ spatial spread of vasoconstriction β†’ larger muscle regions become ischemic β†’ the PEM sensation spreads beyond the initially exerted muscle group.

Probes:

  • Pyridostigmine (30–60 mg): Increases tissue perfusion β†’ reduces ischemia β†’ fewer metabolites produced. If pyridostigmine reduces lactate and PEM β†’ the ischemic cycle is perfusion-limited β€” improving oxygen delivery breaks the feedback loop. If pyridostigmine does NOT reduce PEM but capsaicin does β†’ TRPV1 activation is the primary driver, and perfusion improvement is insufficient because the arterioles are constricted regardless of blood flow.
  • CoQ10 (100–300 mg) + NAC (600–1200 mg): CoQ10 improves mitochondrial efficiency β†’ reduces ROS production. NAC provides glutathione β†’ scavenges ROS. If CoQ10/NAC reduces PEM β†’ ROS-mediated TRPV1 activation is a significant component of the feedback loop. The loop is: exertion β†’ mitochondrial ROS β†’ TRPV1 activation β†’ vasoconstriction β†’ ischemia β†’ more ROS. Breaking the ROS component (antioxidants) should reduce TRPV1 activation even without TRPV1 antagonism.
  • CoQ10/NAC vs. capsaicin differential: If CoQ10/NAC reduces PEM but capsaicin does NOT β†’ the ischemic feedback loop is driven by ROS (metabolic) rather than by TRPV1 directly (pharmacological). The mitochondria are the dominant ROS source, not TRPV1-mediated vasoconstriction. If capsaicin reduces PEM but CoQ10/NAC does NOT β†’ TRPV1 is the dominant driver, and ROS reduction without TRPV1 blockade is insufficient. If both work β†’ dual mechanism: TRPV1 activation AND ROS production drive the feedback loop, and both must be addressed simultaneously.

Consequence: TRPV1 is the most feed-back-driven cascade in the chapter β€” it is a positive feedback loop, not a linear cascade. The diagnostic logic must account for loop dynamics: blocking any single node (TRPV1, COX-2, ROS, perfusion) may produce partial benefit, but complete interruption of the loop requires blocking multiple nodes simultaneously. A patient who gets 30% PEM reduction from capsaicin + 30% from celecoxib + 30% from CoQ10/NAC β†’ triple-node blockade is needed for full PEM suppression. The loop structure means that monotherapy trials may produce false-negatives β€” a single-node block is insufficient to overcome the positive feedback. The chapter’s drug-drug interaction framework (Diurnal Response Window as Circadian Pharmacodiagnostic Probe) is essential for TRPV1: capsaicin + celecoxib + CoQ10/NAC as a three-drug combination should be trialed before concluding that TRPV1 is not rate-limiting. Origin: mechanistic-pathway-tracing.

3 Piezo2 Mechanosensitive Channel Dysfunction in hEDS Overlap

Certainty: 0.30. Piezo1/Piezo2 are established mechanosensitive channels; their role in ME/CFS hEDS overlap is mechanistically inferred. Not directly measured in ME/CFS.

3.1 Cascade: hEDS laxity β†’ Piezo2/1 overactivation β†’ multi-system dysfunction

Cascade:

  • Connective tissue laxity (hEDS) β†’ abnormal mechanical forces on sensory neurons, mast cells, and vascular endothelium
  • Piezo2/Piezo1 overactivation β†’ pathologically elevated mechanosensitive calcium influx
  • Proprioceptive degradation, chronic mast cell degranulation, impaired endothelial mechanotransduction
  • fatigue, positional pain, orthostatic intolerance, brain fog

Step C1: Piezo2 overactivation on sensory neurons (proprioceptive degradation)

  • Intercept: Physical therapy (cervical stabilization, joint bracing) β€” reduces mechanical forces; Low-dose amitriptyline (5–10 mg; tricyclic antidepressant at microdose, NMDA/NE/5-HT effects) β€” NMDA antagonism downstream
ImportantFinding: Bracing/stabilization works β†’ Piezo2-driven proprioceptive degradation

Piezo2-driven proprioceptive degradation is causally contributing. Mechanical forces β€” not spontaneous channel dysfunction β€” drive symptoms.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: Bracing does NOT work β†’ central or persistent Piezo2 dysfunction

Piezo2 dysfunction persists even without mechanical input, or proprioceptive degradation is central (cerebellar, cingulate), not peripheral.

Certainty
Low to Medium
Level of action
Partial root cause

Step C2: Piezo1/Piezo2 overactivation on mast cells (mechanically-driven degranulation)

  • Intercept: Ketotifen (1–2 mg BID; mast cell stabilizer with H1 antihistamine activity) β€” mast cell stabilizer; Cromolyn (100–200 mg QID; mast cell stabilizer, GI-/nasal-localized); LDN (low-dose naltrexone; TLR4 antagonist, opioid receptor antagonist) β€” TLR4 antagonism
ImportantFinding: Ketotifen/cromolyn works β†’ mast cell degranulation occurring

Mast cell degranulation is occurring. Does not distinguish Piezo-driven from IgE-driven.

Certainty
Low to Medium
Level of action
Symptom management
  • Differential logic with bracing: If bracing works AND ketotifen works β†’ Piezo2-driven mast cell degranulation. If bracing works but ketotifen does NOT β†’ mechanosensitive sensory neuron activation, not mast cells.

Step C3: Piezo1 on vascular endothelium (impaired NO bioavailability)

  • Intercept: Midodrine (Ξ±1-adrenergic receptor agonist; vasoconstrictor); Fludrocortisone (mineralocorticoid; volume expansion) β€” volume expansion
ImportantFinding: Fludrocortisone works better than midodrine β†’ hypovolemic vascular lesion

Vascular lesion is likely hypovolemic, not endothelial.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: Midodrine works but fludrocortisone does NOT β†’ vasomotor failure

Vasomotor failure (endothelial NO deficiency from Piezo1 dysfunction) is dominant.

Certainty
Low to Medium
Level of action
Partial root cause

Consequence: Bracing response combined with ketotifen distinguishes mechanosensitive from IgE-mediated mast cell activation. If bracing works alone, Piezo2-driven proprioceptive degradation is rate-limiting β€” treat with mechanical stabilization. Origin: mechanistic-pathway-tracing.

4 Ion Channel Convergence: PIP2 Depletion Model

Certainty: 0.35. PIP2 convergence derived from TRPM3/PIP2 co-localization (Eaton-Fitch 2021) and GPCR autoantibody-driven PLC hyperactivity. Multi-channel simultaneous measurement not yet performed in ME/CFS.

4.1 Cascade: GPCR AAb β†’ PIP2 depletion β†’ multi-channel calcium collapse

Cascade:

  • GPCR autoantibody (Ξ²2-AR, M2/M4) β†’ persistent PLC-Ξ² activation
  • PIP2 hydrolysis β†’ simultaneous loss of PIP2-dependent channel function (TRPM3, TRPM7, TRPV1, TRPC)
  • Multi-channel convergence on calcium dysregulation β†’ multi-system dysfunction

This convergence hypothesis makes a specific differential prediction: if GPCR autoantibodies drive PIP2 depletion, then immunoadsorption should restore TRPM3 function, reduce TRPV1 sensitization, and improve Piezo2-driven symptoms simultaneously β€” because all three channels depend on PIP2.

  • Intercept: Immunoadsorption (extracorporeal IgG removal; autoantibody depletion) β€” removes GPCR autoantibodies; BC007 (DNA aptamer; GPCR autoantibody neutralization) β€” neutralizes GPCR autoantibodies
ImportantFinding: Immunoadsorption works and improves all channel-dependent symptoms β†’ PIP2 convergence model supported

PIP2 convergence model is supported. Root cause is GPCR autoantibodies; ion channel dysfunction is secondary.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: Immunoadsorption works but only improves one symptom domain β†’ independent channel dysfunction

Channels have independent dysfunction mechanisms β€” not PIP2-dependent convergence.

Certainty
Low to Medium
Level of action
Partial root cause
ImportantFinding: Immunoadsorption does NOT work β†’ GPCR autoantibodies not causal

GPCR autoantibodies are either absent, non-pathogenic, or not upstream of channel dysfunction.

Certainty
Low to Medium
Level of action
Partial root cause

Consequence: The strongest test of PIP2 convergence: if immunoadsorption simultaneously improves TRPM3, TRPV1, and Piezo2-dependent symptoms, GPCR AAb→PIP2 depletion is the shared upstream mechanism. Otherwise, independent channel dysfunction requires channel-specific treatment. Origin: mechanistic-pathway-tracing.

References

Cabanas, Helene, Katsuhiko Muraki, Natalie Eaton-Fitch, Donald R Staines, and Sonya Marshall-Gradisnik. 2021. β€œLow Dose Naltrexone Restores TRPM3 Ion Channel Function in Natural Killer Cells from Myalgic Encephalomyelitis/Chronic Fatigue Syndrome Patients.” Frontiers in Immunology 12: 687806. https://doi.org/10.3389/fimmu.2021.687806.