# Evidence digest — 2026-09-02 (week 2026-08-26..09-02)

Synthesized from the weekly collector (`scripts/collect_evidence.py`),
20 papers (PubMed + Europe PMC) + targeted follow-up. Every item carries
its source. OpenEndo agenda: *living evidence synthesis*.

## Headlines this week

### 1. Computational repurposing proof-of-concept: simvastatin + primaquine
**Oskotsky TT, Tang X, Arthurs E et al.** "A transcriptomics-based
computational drug repurposing pipeline identifies simvastatin and
primaquine as therapeutics for endometriosis." *iScience*, 2026-09-18.
PMID 42668641 · doi:10.1016/j.isci.2026.117153
- Transcriptomics-driven pipeline over approved drugs → **simvastatin**
  (statin, anti-inflammatory/anti-angiogenic) and **primaquine**
  (anti-malarial, oxidative stress) flagged as candidates
- **Template for our M3 repurposing screen** — we extend the approach
  with a ChEMBL activity-based screen over the 35 novel targets
- Next step: replicate/validate hits against expression data (GTEx /
  GEO), then clinical follow-up

### 2. SEMA3C → ferroptosis axis (mechanistic lead)
SEMA3C promotes endometriosis progression by inducing ferroptosis
(2026-09-01). GPX4 (in our novel-target list!) is a ferroptosis gatekeeper
— **direct overlap between this week's mechanism literature and our
target audit.** Worth a targeted follow-up.

### 3. Shared genetic architecture, female reproduction
(2026-09-01) — shared genetic architecture + cellular convergence between
female reproductive traits and endometriosis. Supports the registry/
genetics angle (agenda: "-omics integration").

### 4. Diversity gap in endometriosis evidence
(2026-09-01) — lack of socio-demographic participant diversity in
endometriosis evidence. Implication for the hub: our data collection
should track diversity signals; translation agenda (agenda: patient tools
in every language) is evidence-backed.

## Diagnostics (from agenda follow-up)
- **Ziwig Endotest** — CE-IVD saliva test (miRNA signature + ML) shortens
  the 7–10 year diagnostic odyssey to one swab; first AI diagnostics
  already in clinics. Benchmark for open biomarker work (agenda: biomarker
  development, made open).

## Danish angle
- Denmark's national registries (diagnoses, prescriptions, surgery,
  absence) are among the world's best for population-scale endometriosis
  research — barely used for this disease. (agenda: registry & EHR
  analytics — high impact)

## Trial landscape (site data)
- 38 recruiting trials worldwide after relevance filter (only genuine
  endo trials); Denmark trial set maintained weekly.
- Gap-analysis opportunity: disease-modifying vs symptom trials.

## M3 repurposing screen — first results (2026-09-02)

`scripts/repurpose_screen.py`: approved drugs (max_phase 4) with measured
potency (pchembl ≥ 6) against the **35 novel targets** → **9 candidate pairs**
(docs/data/repurposing_candidates.json):

| Target | Approved drug(s) | Why it matters |
|---|---|---|
| **SLC7A11** | **Sulfasalazine** | xCT/cystine-glutamate antiporter = ferroptosis gatekeeper — converges with this week's SEMA3C→ferroptosis paper; sulfasalazine is a known xCT inhibitor |
| **FKBP4** | Sirolimus, Tacrolimus, Cyclosporine | FKBP4 is the progesterone-receptor chaperone; endometriosis is progesterone-resistant — mechanism-relevant |
| **MRGPRX2** | Cetrorelix | GnRH antagonists are already used in endometriosis — supports target validity |
| **GPER1** | Estradiol | GPER1 is the rapid-signaling estrogen receptor — validates target in an estrogen-driven disease |
| **ACVR1B** | Crizotinib, Dabrafenib | TGF-β/activin signaling (fibrosis, lesion growth) |
| **SLCO2A1** | Dinoprostone (PGE2) | prostaglandin transporter — direction of effect needs review (agonist vs blockade) |

Caveats: in-vitro potency ≠ clinical efficacy; direction of effect per
target requires mechanism review; hypothesis-generation only. Next:
cross-reference expression (GTEx) + pathway direction before any clinical
follow-up. The Oskotsky simvastatin/primaquine pipeline remains the
transcriptomics-side template; this screen is the target-activity-side
complement.

---
*Sources: PubMed (PMID links in the weekly JSON snapshot:
docs/research/evidence/weekly/2026-09-02.json), Europe PMC, iScience,
openendo.ai-agenda. Research infrastructure, not medical advice.*