DecodeME-Stratified Pharmacogenomic Trial Platform

1 Background and Rationale

The DecodeME GWAS identified three convergent biological themes—neuronal/synaptic, immune-ambiguous, and autophagy/mitophagy—at genome-wide significant loci (Section:decodeme-genes, Chapter:genetics-epigenetics). If these genetic themes define biologically distinct patient subgroups, treatment response should vary by genetic architecture. This proposed trial platform tests whether polygenic risk scores (PRS) constructed from DecodeME loci can predict treatment response, enabling a pharmacogenomic approach to ME/CFS.

2 Study Design

Design: Multi-arm adaptive platform trial with PRS-based stratification. Population: ME/CFS patients meeting CCC criteria (n=300 minimum, ideally 600 for adequate subgroup power). All participants genotyped; PRS calculated for three gene clusters: (a) neuronal-synapse PRS (CA10, SHISA6, SOX6, LRRC7, DCC, UNC13C, BARHL2 loci), (b) autophagy-mitophagy PRS (FBXL4, CCPG1, ARFGEF2/CSE1L/STAU1 loci), (c) immune PRS (RABGAP1L, BTN2A2, TAOK3, HLA-DQA1 loci). Arms:

  • Arm A (Glutamatergic modulation): Low-dose memantine (5–20 mg) for 12 weeks. Primary hypothesis: benefit concentrated in patients with high neuronal-synapse PRS (Speculation:glutamatergic-autonomic-bridge, Chapter:integrative-models).
  • Arm B (Mitophagy enhancement): Urolithin A (500–1000 mg) for 12 weeks. Primary hypothesis: benefit concentrated in patients with high autophagy-mitophagy PRS (Hypothesis:mitophagy-vulnerability, Chapter:genetics-epigenetics).
  • Arm C (Low-dose amitriptyline): 10–25 mg for 12 weeks. Primary hypothesis: benefit correlates with neuronal-synapse PRS (Speculation:amitriptyline-signal, Chapter:integrative-models).
  • Arm D (Placebo): Matched placebo for each arm. Primary outcome: Composite ME/CFS severity score (fatigue, PEM frequency, cognitive function, autonomic symptoms). Key analysis: PRS \(\\times\) treatment interaction in each arm. The trial succeeds if treatment effect differs significantly by PRS stratum in at least one arm—demonstrating that genetic architecture predicts treatment response. Adaptive features: Interim analysis at 50% enrollment. Drop arms showing no PRS interaction; expand promising arms. Add new arms as additional genetic data (SequenceME rare variants) become available.

3 Expected Impact

If any PRS \(\\times\) treatment interaction reaches significance, this would: (1) validate genetic subtyping as clinically actionable in ME/CFS, (2) identify the first genetically-stratified ME/CFS treatment, (3) provide a framework for future precision medicine trials, and (4) resolve the longstanding question of whether ME/CFS treatment failures reflect genuine inefficacy or population heterogeneity masking subgroup-specific benefit.

4 Feasibility

Requires access to DecodeME genotype data (planned for sharing), collaboration with ME/CFS clinical trial networks, and pharmaceutical industry or foundation funding. PRS construction is computationally straightforward once genotype data are available. Memantine, urolithin A, and amitriptyline are all available and have established safety profiles. The adaptive platform design reduces total sample size requirements compared to separate trials for each intervention.