Identification and Optimization of RNA-Splicing Modulators as Huntingtin Protein-Lowering Agents for the Treatment of Huntington's Disease.

Huntington's disease (HD) is caused by an expanded CAG trinucleotide repeat in exon 1 of the huntingtin (HTT) gene. We report the design of a series of HTT pre-mRNA splicing modulators that lower huntingtin (HTT) protein, including the toxic mutant huntingtin (mHTT), by promoting insertion of a pseudoexon containing a premature termination codon at the exon 49-50 junction. The resulting transcript undergoes nonsense-mediated decay, leading to a reduction of HTT mRNA transcripts and protein levels. The starting benzamide core was modified to pyrazine amide and further optimized to give a potent, CNS-penetrant, and orally bioavailable HTT-splicing modulator 27. This compound reduced canonical splicing of the HTT RNA exon 49-50 and demonstrated significant HTT-lowering in both human HD stem cells and mouse BACHD models. Compound 27 is a structurally diverse HTT-splicing modulator that may help understand the mechanism of adverse effects such as peripheral neuropathy associated with branaplam.

Discovery of Spiro-azaindoline Inhibitors of Hematopoietic Progenitor Kinase 1 (HPK1).

Hematopoietic progenitor kinase 1 (HPK1) is implicated as a negative regulator of T-cell receptor-induced T-cell activation. Studies using HPK1 kinase-dead knock-in animals have demonstrated the loss of HPK1 kinase activity resulted in an increase in T-cell function and tumor growth inhibition in glioma models. Herein, we describe the discovery of a series of small molecule inhibitors of HPK1. Using a structure-based drug design approach, the kinase selectivity of the molecules was significantly improved by inducing and stabilizing an unusual P-loop folded binding mode. The metabolic liabilities of the initial 7-azaindole high-throughput screening hit were mitigated by addressing a key metabolic soft spot along with physicochemical property-based optimization. The resulting spiro-azaindoline HPK1 inhibitors demonstrated improved in vitro ADME properties and the ability to induce cytokine production in primary human T-cells.

Discovery of LRRK2 inhibitors by using an ensemble of virtual screening methods.

In this paper, we present the results of a ligand- and structure-based virtual screen targeting LRRK2, a kinase that has been implicated in Parkinson's disease. For the ligand-based virtual screen, the structures of 12 competitor compounds were used as queries for a variety of 2D and 3D searches. The structure-based virtual screen relied on homology models of LRRK2, as no X-ray structure is currently available in the public domain. From the virtual screening, 662 compounds were purchased, of which 35 showed IC50 values below 10µM in wild-type and/or mutant LRRK2 (a hit rate of 5.3%). Of these 35 hits, four were deemed to have potential for medicinal chemistry follow-up.

The Rational Design of Selective Benzoxazepin Inhibitors of the α-Isoform of Phosphoinositide 3-Kinase Culminating in the Identification of (S)-2-((2-(1-Isopropyl-1H-1,2,4-triazol-5-yl)-5,6-dihydrobenzo[f]imidazo[1,2-d][1,4]oxazepin-9-yl)oxy)propanamide (GDC-0326).

Inhibitors of the class I phosphoinositide 3-kinase (PI3K) isoform PI3Kα have received substantial attention for their potential use in cancer therapy. Despite the particular attraction of targeting PI3Kα, achieving selectivity for the inhibition of this isoform has proved challenging. Herein we report the discovery of inhibitors of PI3Kα that have selectivity over the other class I isoforms and all other kinases tested. In GDC-0032 (3, taselisib), we previously minimized inhibition of PI3Kß relative to the other class I insoforms. Subsequently, we extended our efforts to identify PI3Kα-specific inhibitors using PI3Kα crystal structures to inform the design of benzoxazepin inhibitors with selectivity for PI3Kα through interactions with a nonconserved residue. Several molecules selective for PI3Kα relative to the other class I isoforms, as well as other kinases, were identified. Optimization of properties related to drug metabolism then culminated in the identification of the clinical candidate GDC-0326 (4).

Mitigation of Acetylcholine Esterase Activity in the 1,7-Diazacarbazole Series of Inhibitors of Checkpoint Kinase 1.

Checkpoint kinase 1 (ChK1) plays a key role in the DNA damage response, facilitating cell-cycle arrest to provide sufficient time for lesion repair. This leads to the hypothesis that inhibition of ChK1 might enhance the effectiveness of DNA-damaging therapies in the treatment of cancer. Lead compound 1 (GNE-783), the prototype of the 1,7-diazacarbazole class of ChK1 inhibitors, was found to be a highly potent inhibitor of acetylcholine esterase (AChE) and unsuitable for development. A campaign of analogue synthesis established SAR delineating ChK1 and AChE activities and allowing identification of new leads with improved profiles. In silico docking using a model of AChE permitted rationalization of the observed SAR. Compounds 19 (GNE-900) and 30 (GNE-145) were identified as selective, orally bioavailable ChK1 inhibitors offering excellent in vitro potency with significantly reduced AChE activity. In combination with gemcitabine, these compounds demonstrate an in vivo pharmacodynamic effect and are efficacious in a mouse p53 mutant xenograft model.

A reversed sulfonamide series of selective RORc inverse agonists.

The identification of a new series of RORc inverse agonists is described. Comprehensive structure-activity relationship studies of this reversed sulfonamide series identified potent RORc inverse agonists in biochemical and cellular assays which were also selective against a panel of nuclear receptors. Our work has contributed a compound that may serve as a useful in vitro tool to delineate the complex biological pathways involved in signalling through RORc. An X-ray co-crystal structure of an analogue with RORc has also provided useful insights into the binding interactions of the new series.

Discovery of novel allosteric mitogen-activated protein kinase kinase (MEK) 1,2 inhibitors possessing bidentate Ser212 interactions.

Using structure-based design, two novel series of highly potent biaryl amine mitogen-activated protein kinase kinase (MEK) inhibitors have been discovered. These series contain an H-bond acceptor, in a shifted position compared with previously disclosed compounds, and an adjacent H-bond donor, resulting in a bidentate interaction with the Ser212 residue of MEK1. The most potent compound identified, 1 (G-894), is orally active in in vivo pharmacodynamic and tumor xenograft models.

8-Fluoroimidazo[1,2-a]pyridine: synthesis, physicochemical properties and evaluation as a bioisosteric replacement for imidazo[1,2-a]pyrimidine in an allosteric modulator ligand of the GABA A receptor.

8-Fluoroimidazo[1,2-a]pyridine has been established as a physicochemical mimic of imidazo[1,2-a]pyrimidine, using both in silico and traditional techniques. Furthermore, a novel synthesis of a 3,7-disubstituted-8-fluoroimidazopyridine 3 has been developed and the utility of the physicochemical mimicry has been demonstrated in an in vitro system. Here, the 8-fluoroimidazopyridine ring contained in ligand 3 acts as a bioisosteric replacement for imidazopyrimidine in the GABA(A) receptor modulator 2.

Mycobacterium tuberculosis chaperonin 10 is secreted in the macrophage phagosome: is secretion due to dissociation and adoption of a partially helical structure at the membrane?

To confirm that Mycobacterium tuberculosis chaperonin 10 (Cpn10) is secreted outside the live bacillus, infected macrophages were examined by electron microscopy. This revealed that the mycobacterial protein accumulates both in the wall of the bacterium and in the matrix of the phagosomes in which ingested mycobacteria survive within infected macrophages. To understand the structural implications underlying this secretion, a structural study of M. tuberculosis Cpn10 was performed under conditions that are generally believed to mimic the membrane environment. It was found that in buffer-organic solvent mixtures, the mycobacterial protein forms two main species, namely, a partially helical monomer that prevails in dilute solutions at room temperature and a dimer that folds into a beta-sheet-dominated structure and prevails in either concentrated protein solutions at room temperature or in dilute solutions at low temperature. A partially helical monomer was also found and was completely associated with negatively charged detergents in a micelle-bound state. Remarkably, zwitterionic lipids had no effect on the protein structure. By using N- and C-truncated forms of the protein, the C- and N-terminal sequences were identified as possessing an amphiphilic helical character and as selectively associating with acidic detergent micelles. When the study was extended to other chaperonins, it was found that human Cpn10 is also monomeric and partially helical in dilute organic solvent-buffer mixtures. In contrast, Escherichia coli Cpn10 is mostly dimeric and predominately beta-sheet in both dilute and concentrated solutions. Interestingly, human Cpn10 also crosses biological membranes, whereas the E. coli homologue is strictly cytosolic. These results suggest that dissociation to partially helical monomers and interaction with acidic lipids may be two important steps in the mechanism of secretion of M. tuberculosis Cpn10 to the external environment.

A consensus neural network-based technique for discriminating soluble and poorly soluble compounds.

BCUT [Burden, CAS, and University of Texas] descriptors, defined as eigenvalues of modified connectivity matrices, have traditionally been applied to drug design tasks such as defining receptor relevant subspaces to assist in compound selections. In this paper we present studies of consensus neural networks trained on BCUTs to discriminate compounds with poor aqueous solubility from those with reasonable solubility. This level was set at 0.1 mg/mL on advice from drug formulation and drug discovery scientists. By applying strict criteria to the insolubility predictions, approximately 95% of compounds are classified correctly. For compounds whose predictions have a lower level of confidence, further parameters are examined in order to flag those considered to possess unsuitable biopharmaceutical and physicochemical properties. This approach is not designed to be applied in isolation but is intended to be used as a filter in the selection of screening candidates, compound purchases, and the application of synthetic priorities to combinatorial libraries.

Selecting screening candidates for kinase and G protein-coupled receptor targets using neural networks.

A series of neural networks has been trained, using consensus methods, to recognize compounds that act at biological targets belonging to specific gene families. The MDDR database was used to provide compounds targeted against gene families and sets of randomly selected molecules. BCUT parameters were employed as input descriptors that encode structural properties and information relevant to ligand-receptor interactions. In each case, the networks identified over 80% of the compounds targeting a gene family. The technique was applied to purchasing compounds from external suppliers, and results from screening against one gene family demonstrated impressive abilities to predict the activity of the majority of known hit compounds.