Fig. 1

Identification of thio-xanthine/methylxanthine derivatives as PARG inhibitors. a PARG expression is upregulated in the majority of human cancers. Analysis of PARG expression in TCGA tumors and matched normal tissues. Log p-values were from Wilcoxon tests with greater values indicating stronger differences; PARG expression was higher in tumors than in matched controls in all tumor types (green) except in kidney malignancies (KIRC and KICH, red). Datasets with >10 normal samples and significant p-values were included. Box plots display the interquartile range (IQR) from Q1 to Q3 (25–75% percentiles), median (centre line), whiskers extending to the minimum (Q1–1.5*IQR) and maximum (Q3 + 1.5*IQR) and outliers (dots). b the raw data of HTS for JA2-3. The high reproducibility of the TR-FRET assay for PARG activity is reflected by a Z-factor of 0.6–0.8 (Supplementary Fig. 2), this kinetic high-throughput screen was performed against the NCI Diversity Set II library (1990 compounds in seven 384-well plates) at 6.7 μM compound concentrations and two PARG concentrations (12.5 nM and 25 nM). c Thio-xanthine/methylxanthine derivatives are potent PARG inhibitors. De-PARylation of PARP1C by PARG (25 nM) in the presence of the designated compound at 6.7 μM (left panel); the structure of the chemicals (right panel), among five chemotypes identified from the HTS (Supplementary Fig. 3). The JA2 xanthine/methylxanthine series was selected as the lead pharmacophore, based on its structural similarities to adenine, potency in vitro, and favorable drug-like characteristics. Dose-dependent inhibition of PARG activity by the JA2 series compounds was quantitatively analyzed using a gel-based PARG activity assay (Supplementary Fig. 4). d Three representative HTS hits of the JA2 chemotype (JA2-3, JA2-4, and JA2-5) show a potent PARG inhibition with the sub-micromolar range of IC₅₀ values. Source Data are provided as a Source Data file.