Date: March 2026
Status: Phase I — In Silico Pipeline (Partial
Completion)
This report documents the completed and pending steps of the in silico drug repurposing pipeline for identifying receptor-level TGF-β pathway inhibitors in LR-MDS.
Total: 30 genes across 4 functional levels
| Level | Count | Key Members |
|---|---|---|
| Receptor | 8 | TGFBR1 (ALK5), TGFBR2, ACVR1B (ALK4), ACVR2A, ACVR2B, ACVR1, BMPR1A, BMPR2 |
| Signal | 7 | SMAD2, SMAD3, SMAD4, SMAD7, SMAD1, SMAD5, SMAD6 |
| Ligand | 6 | GDF11, GDF8, INHBA, INHBB, TGFB1, TGFB2, TGFB3 |
| Co-regulator | 9 | SMURF1, SMURF2, ZFYVE9 (SARA), SKI, SKIL, BAMBI, EP300, CREBBP |
HIGH priority seeds (directly from proposal hypothesis): TGFBR1, ACVR1B, ACVR2B, SMAD2, SMAD3, SMAD7, GDF11, GDF8, INHBB
Files generated: - seed_gene_list.tsv —
Full annotated seed list with UniProt IDs, Entrez IDs, priorities -
STRING_query_genes.txt — Ready to paste into STRING
Multiple Proteins interface
| Property | Value |
|---|---|
| ChEMBL ID | CHEMBL4439 |
| UniProt | P36897 |
| Gene Symbol | TGFBR1 |
| Target Type | SINGLE PROTEIN |
| PDB Structures | 1B6C, 1IAS, 1PY5, 1RW8 (SB-431542 co-crystal), 2WOT, 2WOU, 3GXL, 3HMM, 3KCF, 4X2F, 5E8S/T/U/W/X/Z (inhibitor series), 5QIK/L/M, 6B8Y, 8YHF, 9B9F, 9F6X, 9FK5 |
| Total IC50 records | 2,484 |
| Records with pChEMBL ≥7.0 | 50+ (sub-100 nM potency) |
| Records with pChEMBL ≥8.0 | 15+ (sub-10 nM potency) |
| Curated drug mechanisms | 2 (galunisertib, vactosertib) |
| Rank | ChEMBL ID | Name | IC50 (nM) | pChEMBL | Assay Type |
|---|---|---|---|---|---|
| 1 | CHEMBL341280 | Bromo-pyrazole-quinoline | 1.2 | 8.92 | Functional (NIH 3T3) |
| 2 | CHEMBL129677 | Pyrazole-quinoline | 2.9 | 8.54 | Functional (mink lung) |
| 3 | CHEMBL185238 | GW786460X | 4.0 | 8.40 | Binding (Sf9) |
| 4 | CHEMBL3260567 | Vactosertib | 7.0 | 8.15 | Binding (Sf9 kinase) |
| 5 | CHEMBL361237 | GW785804X | 10.0 | 8.00 | Binding (Sf9) |
| 6 | CHEMBL365663 | GW780159X | 10.0 | 8.00 | Binding + Functional |
| 7 | CHEMBL183354 | Pyrazole-naphthyridine | 10.0 | 8.00 | Binding (Sf9) |
| 8 | CHEMBL2364611 | Galunisertib | 56.0 | 7.25 | Binding |
| 9 | CHEMBL440084 | SB-431542 | 94.0 | 7.03 | Binding |
| Property | Value |
|---|---|
| IUPAC stem | -sertib (serine/threonine kinase inhibitors) |
| MW | 399.43 Da |
| ALogP | 4.24 |
| PSA | 83.79 Ų |
| HBD / HBA | 2 / 6 |
| Ro5 violations | 0 |
| QED | 0.46 |
| Max Phase | Phase 2 |
| ALK5 IC50 range | 7.0 — 41.0 nM (median ~12 nM) |
| Clinical trial (MDS) | NCT03074006 — Phase 1 in low/intermediate MDS, COMPLETED, n=9 |
| Clinical trial (heme) | NCT05400122 — Phase 1b vactosertib + NK cells in hematologic malignancies (SUSPENDED) |
| Sponsor | MedPacto, Inc. |
| SMILES | Cc1cccc(-c2[nH]c(CNc3ccccc3F)nc2-c2ccc3ncnn3c2)n1 |
Assessment: Top translational candidate. Already tested in MDS patients (Phase 1 completed). Potent ALK5 inhibitor with clean Ro5 profile. The completed NCT03074006 trial (IPSS-R very low/low/intermediate) directly overlaps with our target population.
| Property | Value |
|---|---|
| MW | 369.43 Da |
| ALogP | 3.51 |
| PSA | 86.69 Ų |
| HBD / HBA | 1 / 5 |
| Ro5 violations | 0 |
| QED | 0.60 (best among candidates) |
| Max Phase | Phase 2 |
| ALK5 IC50 | 56 nM (moderate) |
| Clinical trial (MDS) | NCT02008318 — Phase 2/3 in very low/low/intermediate MDS, COMPLETED, n=43, Eli Lilly |
| SMILES | Cc1cccc(-c2nn3c(c2-c2ccnc4ccc(C(N)=O)cc24)CCC3)n1 |
Assessment: The most clinically advanced ALK5 inhibitor in MDS. Phase 2/3 data available (Santini V et al., Clin Cancer Res 2019). Less potent than vactosertib but highest QED and most extensive clinical safety data. The published results showed modest hematologic improvement in a subset of patients, supporting the biological rationale but also highlighting the need for combination or biomarker-driven strategies.
| Property | Value |
|---|---|
| MW | 384.40 Da |
| ALogP | 3.63 |
| PSA | 103.12 Ų |
| Ro5 violations | 0 |
| ALK5 IC50 | 94 nM |
| Co-crystal PDB | 1RW8 (reference for docking) |
| Status | Research-grade only |
Assessment: Widely used reference compound. PDB 1RW8 co-crystal structure serves as the primary docking reference. Should be included as benchmark in all experimental assays.
| NCT ID | Drug | Phase | Condition | Status | N | Sponsor |
|---|---|---|---|---|---|---|
| NCT03074006 | Vactosertib | Phase 1 | Low/Int MDS | COMPLETED | 9 | MedPacto |
| NCT02008318 | Galunisertib | Phase 2/3 | VLow/Low/Int MDS | COMPLETED | 43 | Eli Lilly |
| NCT05400122 | Vactosertib + NK cells | Phase 1b | Heme malignancies incl. MDS | SUSPENDED | 12 | D. Wald |
NCT03074006 inclusion criteria directly match our target: - IPSS-R Very Low, Low, or Intermediate risk - Hb ≤ 9.0 g/dL OR RBC transfusion-dependent (≥2 units/month) - Must have failed or not be candidates for standard treatments - Primary endpoint: Hematologic improvement (IWG 2006)
This validates our experimental design — we’re modeling exactly this clinical scenario in vitro.
STRING_query_genes.txt (30
genes)Target structure: PDB 1RW8 (ALK5 + SB-431542) — use this as reference Alternative: PDB 5E8S/5E8T (newer structures with different inhibitors) Positive controls: Vactosertib, Galunisertib, SB-431542 Tool: AutoDock Vina or CB-Dock2 (https://cadd.labshare.cn/cb-dock2/) Binding box: Center on SB-431542 position in 1RW8, grid 25×25×25 Å, exhaustiveness=32
Relevant datasets to query: - GSE19429 (MDS CD34+ vs healthy, n=183) - GSE58831 (MDS subtypes) - GSE4619 (MDS bone marrow)
GSEA gene sets: HALLMARK_TGF_BETA_SIGNALING, KEGG_TGF_BETA_SIGNALING_PATHWAY
Based on currently available data (ChEMBL + ClinicalTrials.gov):
| Molecule | Network (×3) | DGIdb (×2) | Docking (×2) | ChEMBL (×2) | CMap (×1) | Clinical (×2) | Transcr (×1) | TOTAL |
|---|---|---|---|---|---|---|---|---|
| Vactosertib | TBD | TBD | REF | 4 (pChEMBL 8.15) | TBD | 4 (Phase 1 MDS) | TBD | ≥8 partial |
| Galunisertib | TBD | TBD | TBD | 3 (pChEMBL 7.25) | TBD | 4 (Phase 2/3 MDS) | TBD | ≥7 partial |
| SB-431542 | TBD | TBD | REF | 3 (pChEMBL 7.03) | TBD | 0 (research only) | TBD | ≥3 partial |
| GW786460X | TBD | TBD | TBD | 4 (pChEMBL 8.40) | TBD | 0 | TBD | ≥4 partial |
Note: Network proximity scores (NedRex), DGIdb matches, docking scores, CMap connectivity, and transcriptomic validation require desktop Cytoscape and web database queries. These will fill the remaining matrix cells.
The discovery that vactosertib has already completed a Phase 1 trial specifically in low/intermediate MDS (NCT03074006, MedPacto) is a significant finding for our proposal. This means:
| File | Description | Use |
|---|---|---|
| seed_gene_list.tsv | 30 annotated seed genes with IDs | Master reference |
| STRING_query_genes.txt | Gene symbols for STRING paste | STRING database input |
| compound_analysis.tsv | 9 candidate compounds with full profiles | Scoring matrix input |
| This report (findings_report.md) | Comprehensive Phase I findings | Documentation |
Report generated: March 2026 Next steps: Complete STRING/Cytoscape/NedRex/Docking steps on desktop, then finalize scoring matrix