Designing novel inhibitors of transcription factor BRD4

Utilizing Receptor.AI’s hit identification workflow to target member of a family of 4 similar proteins

3.2M

compounds
screened

100

hit candidates
selected

17

compounds with high
affinity to BRD4

3

potent hits
identified

*Project workflow

01/ Background

  • The target is Human Bromodomain-containing protein 4 (BRD4) — transcription factor with an important role in cancer development.
  • BRD4 is a member of a family of at least four proteins sharing the same structural motif.

02/ Methodology

  • 3.2M compounds were prefiltered from 8M stock library based on
16 psychem/druglike filters.
  • Using target’s protein data, compounds were subjected to
    semi-ligand DTI model scoring.
  • Evaluated the most relevant ADMET endpoints.
  • Performed proteome-wide selectivity scoring.
  • ~50K ligands subjected to ArtiDock AI docking and AI rescoring.
  • 100 compounds selected for in vitro validation from 1000 best using AI-guided hit candidates selection.
  • Experimental validation was performed using a
    differential scanning fluorimetry assay.

03/ Results

  • 17 compounds demonstrated a significant binding effect.
  • 8 of them belong to known classes of BRD4 inhibitors.
  • 9 hits are entirely new to BRD4, representing 8 chemical classes.
*Difference between BRD4-control and
BRD4-ligand complex melting temperatures
*ADMET-PK profile

04/ Best Compound

  • Previously unknown class of BRD4 inhibitors.
  • Has shown the highest potency in experiments (IC50 = 470 nM).