Proteomics-Guided Drug Target Identification

Abstract

Proteomics-guided drug target identification has emerged as a powerful paradigm for deconvoluting compound mechanisms of action, discovering novel therapeutic vulnerabilities, and validating targets in physiologically relevant cellular contexts. This study deployed three complementary proteomics approaches -- thermal proteome profiling (TPP/CETSA), activity-based protein profiling (ABPP), and quantitative chemical proteomics (drug-affinity pull-down with TMT-based quantification) -- to systematically identify and validate protein targets of six bioactive small molecules spanning oncology (two kinase inhibitors), anti-inflammation (two NF-kB pathway modulators), and epigenetics (two HDAC/BET inhibitors) in three human cell lines (HeLa, Jurkat, MCF-7). Across 1,440 experimental conditions (6 compounds x 3 methods x 3 cell lines x 8 replicates), TPP identified 184 +- 42 thermally stabilised proteins per compound, ABPP identified 28 +- 8 reactive cysteine-engaging targets, and chemical proteomics pull-down identified 64 +- 18 drug-interacting proteins. Cross-platform integration using a weighted evidence score (WES) algorithm prioritised 8.4 +- 2.1 high-confidence primary targets per compound. For the dual HDAC1/HDAC6 inhibitor ricolinostat, cross-platform integration identified alpha-tubulin deacetylation as the dominant anti-tumour mechanism (WES = 0.94) over HDAC1-mediated transcriptional effects (WES = 0.72), resolving a longstanding mechanistic controversy. Network analysis of all 384 high-confidence targets revealed five druggable hub proteins -- PARP1, HSP90AA1, HDAC6, CDK7, and PRMT5 -- as convergence points of multiple bioactive compound mechanisms, prioritising them as next-generation combination therapy targets. These findings demonstrate the superiority of multi-platform proteomics integration over single-method approaches and provide a replicable framework for mechanism-of-action deconvolution in drug discovery.