Acknowledgment This work was supported
This work was supported by the Swedish Research Council, the Knut and Alice Wallenberg Foundation, the Göran Gustafsson Foundation, the Swedish Pain Relief Foundation, the Torsten Söderbergs Stiftelse, the Swedish Children\'s Cancer Foundation, and the Swedish Cancer Society.
The human dCTPase belongs to a class of nucleoside triphosphate pyrophosphatase enzymes heavily involved in deoxynucleotide triphosphate (dNTP) pool metabolism, and a growing body of evidence supports their role in human pathology and utility as therapeutic targets., , , , , , , dCTPase shows substrate preference towards 5-modified dNTPs, and has been found to be overexpressed in multiple carcinomas, associated with cancer stemness,, and to be responsible for a decreased response to some cytidine analogues, through drug efflux and degradation mechanisms., Only a handful of dCTPase inhibitors have been reported to date (–, )., , , , As a continuation of our previous letter on novel back-up series of dCTPase inhibitors, here we report on new chemical space associated with the dCTPase enzyme. Two screening campaigns against the full-length recombinant human dCTPase protein, using a microtiter plate-adapted Malachite Green assay,, , have yielded a number of promising dCTPase inhibitors, including benzimidazoles, triazolothiadiazoles, and piperazin-1-ylpyridazines. The Chembridge DiverSET (50,000 compounds), a commercial screening library often used in drug discovery, and a subset of the CBCS screening set (5500 compounds), were screened at 10μM Latrunculin B concentration. After inspection for potential pan assay interference compounds (PAINS), and confirmation of hit identity and purity by LC–MS, we focused on chemotypes with available SAR within the screening sets, and with developability precedents. Upon structural review, pyrimidinone, pyrazolopyrimidinone, imidazopurinone, thiadiazolopyrimidinone, thiazolopyrimidinone, isoxazolopyridinone, isoxazolopyrimidinone, and furopyridinone hits were clustered as pyridone-/pyrimidinone-derived chemotypes. Some of these chemical scaffolds have been previously described in drug discovery projects against unrelated biologic targets, such as P2X7-R, PASK, EphB4, NMDA-R and Ca-R ()., , , , With the help of DataWarrior software, we analysed hit ligand efficiency in respect to molecular weight (Ligand Efficiency, LE) and lipophilicity (Lipophilic ligand Efficiency, LLE). Guided by lean efficiency scores used in fragment based drug discovery (LE ≥0.3 and LLE ≥3), hit optimisation can be biased towards smaller and more specific ligands ()., , , Given the polyaromatic systems present in this cluster, and their low calculated LogP, we relied on LLE and sp3 carbon fraction (Fsp3) to estimate aqueous solubility, instead of calculated LogS. The non-fused pyrimidin-4(3)-one hits followed the general formula presented in . The tri-substituted morpholino derivative (IC=0.92μM) and the tetra-substituted bromophenyl derivative (IC=0.37μM) were the most active compounds in this cluster, and were among the most efficient binders, with displaying an LE=0.4, and an LLE>5 (a). When the -bromine atom in the R2 benzene ring was replaced with - hydroxy or methyl groups, enzymatic inhibition decreased ( and . ), and replacement of the R2 aromatic ring with a polar ethanol chain completely abolished activity (. ). Interestingly, compound displayed the best balance between LE, LLE and Fsp3, highlighting desirable physicochemical properties and ligand efficiency for further optimization (f). The pyrazolo[1,5-]pyrimidin-7(4)-one hits displayed a tetra-substituted bicyclic core (), with an aromatic ring at the R2 position, -substituted or unsubstituted for sub-micromolar potency (– and ), and a small methyl group at the R1 position, which could not be replaced with a methoxymethyl substituent (. ). The most active compounds in this series are reminiscent of benzimidazole derivatives recently described as dCTPase inhibitors. The R3 position in this scaffold was usually substituted with a methyl group, and tolerated a larger phenyl group, when R4 was unsubstituted (). A short methyl ester chain at R3 abolished enzymatic inhibition (), but moving the same methyl ester to R4 retained low micromolar potency (), as so did fusing R3 and R4 into a lipophilic cyclohexane ring (). Hits in the pyrazolopyrimidinone cluster presented good to very good LE (0.3–0.5), and moderate LLE (2–3) (b), although Fsp3 fraction was generally low, indicating potential low aqueous solubility (f).