The Multi-Objective Computational Design and Prioritization of Novel TSPO PET Radioligands for Improved Inflammation Imaging after Traumatic Brain Injury
Clinicians employ TSPO PET imaging to quantify the extent of TBI-associated neuroinflammation through radioligands that bind to the central pocket of the human translocator protein. However, many radioligands underperform due to pocket dissociation and low blood-brain barrier permeability. I hypothesize that a targeted cheminformatic pipeline can accelerate the pre-synthetic design of novel TSPO ligands with improved physicochemistry. Eight second-generation ligands were benchmarked using preclinical physicochemical and PET metrics. Emapunil was selected as the modifiable scaffold. Haloalkylation was curated across two functional groups, establishing a 1922-analog dataset. Following prioritization based on molecular descriptors and docking energy, 15 lead candidates (assigned IDs of A#) proceeded to PBPK modeling. Leads were z-scored, and the top three candidates advanced to plasma-input Logan analysis. Consensus ADMET prioritized leads for short-timescale molecular dynamics. The top four analogs were evaluated through a five-replicate 10-ns MD panel along with the reference, and retrosynthetic routes were devised through ASKCOS. A7 showed a significantly improved static and dynamic profile over emapunil. A7 displayed a 5.4-fold increase in partition coefficient and fourfold increase in brain-plasma concentration ratio in the PBPK simulations versus the reference. Logan VT slope estimates were stable for A3, A6, and A7. MD showed A7 retained 28.25±5.94 mean fixed-pocket contacts and ≥10 contacts in 97.96±2.04% of frames. A7’s contact advantage over AC-5216 was statistically significant after Holm correction (p=0.0317). ASKCOS predicted 200 complete synthetic routes for A3, A8, A15, and emapunil with buyable terminals, while A7 displayed an unresolved N-difluoromethylation terminal. Pareto analysis identified A7 as the only nondominated analog. A7 emerged as the strongest radioligand candidate and demonstrates the effectiveness of a cohesive cheminformatic system in PET radioligand development.
Authors
- Dhruv Subramanian (ORCID: https://orcid.org/0009-0008-1852-1249)
Publication Details
- Journal
- ChemRxiv
- Published
- 2026-11-16
- DOI
- https://doi.org/10.26434/chemrxiv.15003660/v2
- Primary Topic
- Computational Drug Discovery Methods
- Type
- preprint