Ensemble Simulations and Experimental Free Energy Distributions: Evaluation and Characterization of Isoxazole Amides as SMYD3 Inhibitors.

Optimization of binding affinities for ligands to their target protein is a primary objective in rational drug discovery. Herein, we report on a collaborative study that evaluates various compounds designed to bind to the SET and MYND domain-containing protein 3 (SMYD3). SMYD3 is a histone methyltransferase and plays an important role in transcriptional regulation in cell proliferation, cell cycle, and human carcinogenesis. Experimental measurements using the scintillation proximity assay show that the distributions of binding free energies from a large number of independent measurements exhibit non-normal properties. We use ESMACS (enhanced sampling of molecular dynamics with approximation of continuum solvent) and TIES (thermodynamic integration with enhanced sampling) protocols to predict the binding free energies and to provide a detailed chemical insight into the nature of ligand-protein binding. Our results show that the 1-trajectory ESMACS protocol works well for the set of ligands studied here. Although one unexplained outlier exists, we obtain excellent statistical ranking across the set of compounds from the ESMACS protocol and good agreement between calculations and experiments for the relative binding free energies from the TIES protocol. ESMACS and TIES are again found to be powerful protocols for the accurate comparison of the binding free energies.

GRAM: A True Null Model for Relative Binding Affinity Predictions.

Relative binding affinity prediction is a critical component in computer aided drug design. A significant amount of effort has been dedicated to developing rapid and reliable in silico methods. However, robust assessment of their performance is still a complicated issue, as it requires a performance measure applicable in the prospective setting and more importantly a true null model that defines the expected performance of being random in an objective manner. Although many performance metrics, such as the Pearson correlation coefficient (r), mean unsigned error (MUE), and root-mean-square error (RMSE), are frequently used in the literature, a true and nontrivial null model has yet been identified. To address this problem, here we introduce an interval estimate as an additional measure, namely, the prediction interval (PI), which can be estimated from the error distribution of the predictions. The benefits of using the interval estimate are (1) it provides the uncertainty range in the predicted activities, which is important in prospective applications, and (2) a true null model with well-defined PI can be established. We provide one such example termed the Gaussian Random Affinity Model (GRAM), which is based on the empirical observation that the affinity change in a typical lead optimization effort has the tendency to distribute normally N (0, σ). Having an analytically defined PI that only depends on the variation in the activities, GRAM should, in principle, allow us to compare the performance of relative binding affinity prediction methods in a standard way, ultimately critical to measuring the progress made in algorithm development.

Discovery of a crystalline sulforaphane analog with good solid-state stability and engagement of the Nrf2 pathway in vitro and in vivo.

The antioxidant natural product sulforaphane (SFN) is an oil with poor aqueous and thermal stability. Recent work with SFN has sought to optimize methods of formulation for oral and topical administration. Herein we report the design of new analogs of SFN with the goal of improving stability and drug-like properties. Lead compounds were selected based on potency in a cellular screen and physicochemical properties. Among these, 12 had good aqueous solubility, permeability and long-term solid-state stability at 23 °C. Compound 12 also displayed comparable or better efficacy in cellular assays relative to SFN and had in vivo activity in a mouse cigarette smoke challenge model of acute oxidative stress.

EZH2 inhibition as a therapeutic strategy for lymphoma with EZH2-activating mutations.

In eukaryotes, post-translational modification of histones is critical for regulation of chromatin structure and gene expression. EZH2 is the catalytic subunit of the polycomb repressive complex 2 (PRC2) and is involved in repressing gene expression through methylation of histone H3 on lysine 27 (H3K27). EZH2 overexpression is implicated in tumorigenesis and correlates with poor prognosis in several tumour types. Additionally, somatic heterozygous mutations of Y641 and A677 residues within the catalytic SET domain of EZH2 occur in diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma. The Y641 residue is the most frequently mutated residue, with up to 22% of germinal centre B-cell DLBCL and follicular lymphoma harbouring mutations at this site. These lymphomas have increased H3K27 tri-methylation (H3K27me3) owing to altered substrate preferences of the mutant enzymes. However, it is unknown whether specific, direct inhibition of EZH2 methyltransferase activity will be effective in treating EZH2 mutant lymphomas. Here we demonstrate that GSK126, a potent, highly selective, S-adenosyl-methionine-competitive, small-molecule inhibitor of EZH2 methyltransferase activity, decreases global H3K27me3 levels and reactivates silenced PRC2 target genes. GSK126 effectively inhibits the proliferation of EZH2 mutant DLBCL cell lines and markedly inhibits the growth of EZH2 mutant DLBCL xenografts in mice. Together, these data demonstrate that pharmacological inhibition of EZH2 activity may provide a promising treatment for EZH2 mutant lymphoma.

GSK114: A selective inhibitor for elucidating the biological role of TNNI3K.

A series of selective TNNI3K inhibitors were developed by modifying the hinge-binding heterocycle of a previously reported dual TNNI3K/B-Raf inhibitor. The resulting quinazoline-containing compounds exhibit a large preference (up to 250-fold) for binding to TNNI3K versus B-Raf, are useful probes for elucidating the biological pathways associated with TNNI3K, and are leads for discovering novel cardiac medicines. GSK114 emerged as a leading inhibitor, displaying significant bias (40-fold) for TNNI3K over B-Raf, exceptional broad spectrum kinase selectivity, and adequate oral exposure to enable its use in cellular and in vivo studies.

Mutation of A677 in histone methyltransferase EZH2 in human B-cell lymphoma promotes hypertrimethylation of histone H3 on lysine 27 (H3K27).

Trimethylation of histone H3 on lysine 27 (H3K27me3) is a repressive posttranslational modification mediated by the histone methyltransferase EZH2. EZH2 is a component of the polycomb repressive complex 2 and is overexpressed in many cancers. In B-cell lymphomas, its substrate preference is frequently altered through somatic mutation of the EZH2 Y641 residue. Herein, we identify mutation of EZH2 A677 to a glycine (A677G) among lymphoma cell lines and primary tumor specimens. Similar to Y641 mutant cell lines, an A677G mutant cell line revealed aberrantly elevated H3K27me3 and decreased monomethylated H3K27 (H3K27me1) and dimethylated H3K27 (H3K27me2). A677G EZH2 possessed catalytic activity with a substrate specificity that was distinct from those of both WT EZH2 and Y641 mutants. Whereas WT EZH2 displayed a preference for substrates with less methylation [unmethylated H3K27 (H3K27me0):me1:me2 k(cat)/K(m) ratio = 9:6:1] and Y641 mutants preferred substrates with greater methylation (H3K27me0:me1:me2 k(cat)/K(m) ratio = 1:2:13), the A677G EZH2 demonstrated nearly equal efficiency for all three substrates (H3K27me0:me1:me2 k(cat)/K(m) ratio = 1.1:0.6:1). When transiently expressed in cells, A677G EZH2, but not WT EZH2, increased global H3K27me3 and decreased H3K27me2. Structural modeling of WT and mutant EZH2 suggested that the A677G mutation acquires the ability to methylate H3K27me2 through enlargement of the lysine tunnel while preserving activity with H3K27me0/me1 substrates through retention of the Y641 residue that is crucial for orientation of these smaller substrates. This mutation highlights the interplay between Y641 and A677 residues in the substrate specificity of EZH2 and identifies another lymphoma patient population that harbors an activating mutation of EZH2.

Discovery of Proline-Based p300/CBP Inhibitors Using DNA-Encoded Library Technology in Combination with High-Throughput Screening.

E1A binding protein (p300) and CREB binding protein (CBP) are two highly homologous and multidomain histone acetyltransferases. These two proteins are involved in many cellular processes by acting as coactivators of a large number of transcription factors. Dysregulation of p300/CBP has been found in a variety of cancers and other diseases, and inhibition has been shown to decrease Myc expression. Herein, we report the identification of a series of highly potent, proline-based small-molecule p300/CBP histone acetyltransferase (HAT) inhibitors using DNA-encoded library technology in combination with high-throughput screening. The strategy of reducing ChromlogD and fluorination of metabolic soft spots was explored to improve the pharmacokinetic properties of potent p300 inhibitors. Fluorination of both cyclobutyl and proline rings of 22 led to not only reduced clearance but also improved cMyc cellular potency.

Identification of Diarylurea Inhibitors of the Cardiac-Specific Kinase TNNI3K by Designing Selectivity Against VEGFR2, p38α, and B-Raf.

A series of diarylurea inhibitors of the cardiac-specific kinase TNNI3K were developed to elucidate the biological function of TNNI3K and evaluate TNNI3K as a therapeutic target for the treatment of cardiovascular diseases. Utilizing a structure-based design, enhancements in kinase selectivity were engineered into the series, capitalizing on the established X-ray crystal structures of TNNI3K, VEGFR2, p38α, and B-Raf. Our efforts culminated in the discovery of an in vivo tool compound 47 (GSK329), which exhibited desirable TNNI3K potency and rat pharmacokinetic properties as well as promising kinase selectivity against VEGFR2 (40-fold), p38α (80-fold), and B-Raf (>200-fold). Compound 47 demonstrated positive cardioprotective outcomes in a mouse model of ischemia/reperfusion cardiac injury, indicating that optimized exemplars from this series, such as 47, are favorable leads for discovering novel medicines for cardiac diseases.

Discovery of a first-in-class reversible DNMT1-selective inhibitor with improved tolerability and efficacy in acute myeloid leukemia.

DNA methylation, a key epigenetic driver of transcriptional silencing, is universally dysregulated in cancer. Reversal of DNA methylation by hypomethylating agents, such as the cytidine analogs decitabine or azacytidine, has demonstrated clinical benefit in hematologic malignancies. These nucleoside analogs are incorporated into replicating DNA where they inhibit DNA cytosine methyltransferases DNMT1, DNMT3A and DNMT3B through irreversible covalent interactions. These agents induce notable toxicity to normal blood cells thus limiting their clinical doses. Herein we report the discovery of GSK3685032, a potent first-in-class DNMT1-selective inhibitor that was shown via crystallographic studies to compete with the active-site loop of DNMT1 for penetration into hemi-methylated DNA between two CpG base pairs. GSK3685032 induces robust loss of DNA methylation, transcriptional activation and cancer cell growth inhibition in vitro. Due to improved in vivo tolerability compared with decitabine, GSK3685032 yields superior tumor regression and survival mouse models of acute myeloid leukemia.

Discovery of Isoxazole Amides as Potent and Selective SMYD3 Inhibitors.

We report herein the discovery of isoxazole amides as potent and selective SET and MYND Domain-Containing Protein 3 (SMYD3) inhibitors. Elucidation of the structure-activity relationship of the high-throughput screening (HTS) lead compound 1 provided potent and selective SMYD3 inhibitors. The SAR optimization, cocrystal structures of small molecules with SMYD3, and mode of inhibition (MOI) characterization of compounds are described. The synthesis and biological and pharmacokinetic profiles of compounds are also presented.

Discovery of GSK8612, a Highly Selective and Potent TBK1 Inhibitor.

The serine/threonine protein kinase TBK1 (Tank-binding Kinase-1) is a noncanonical member of the IkB kinase (IKK) family. This kinase regulates signaling pathways in innate immunity, oncogenesis, energy homeostasis, autophagy, and neuroinflammation. Herein, we report the discovery and characterization of a novel potent and highly selective TBK1 inhibitor, GSK8612. In cellular assays, this small molecule inhibited toll-like receptor (TLR)3-induced interferon regulatory factor (IRF)3 phosphorylation in Ramos cells and type I interferon (IFN) secretion in primary human mononuclear cells. In THP1 cells, GSK8612 was able to inhibit secretion of interferon beta (IFNß) in response to dsDNA and cGAMP, the natural ligand for STING. GSK8612 is a TBK1 small molecule inhibitor displaying an excellent selectivity profile and therefore represents an ideal probe to further dissect the biology of TBK1 in models of immunity, neuroinflammation, obesity, or cancer.

Predicting absolute ligand binding free energies to a simple model site.

A central challenge in structure-based ligand design is the accurate prediction of binding free energies. Here we apply alchemical free energy calculations in explicit solvent to predict ligand binding in a model cavity in T4 lysozyme. Even in this simple site, there are challenges. We made systematic improvements, beginning with single poses from docking, then including multiple poses, additional protein conformational changes, and using an improved charge model. Computed absolute binding free energies had an RMS error of 1.9 kcal/mol relative to previously determined experimental values. In blind prospective tests, the methods correctly discriminated between several true ligands and decoys in a set of putative binders identified by docking. In these prospective tests, the RMS error in predicted binding free energies relative to those subsequently determined experimentally was only 0.6 kcal/mol. X-ray crystal structures of the new ligands bound in the cavity corresponded closely to predictions from the free energy calculations, but sometimes differed from those predicted by docking. Finally, we examined the impact of holding the protein rigid, as in docking, with a view to learning how approximations made in docking affect accuracy and how they may be improved.

4,6-Diaminopyrimidines as Highly Preferred Troponin I-Interacting Kinase (TNNI3K) Inhibitors.

Structure-guided progression of a purine-derived series of TNNI3K inhibitors directed design efforts that produced a novel series of 4,6-diaminopyrimidine inhibitors, an emerging kinase binding motif. Herein, we report a detailed understanding of the intrinsic conformational preferences of the scaffold, which impart high specificity for TNNI3K. Further manipulation of the template based on the conformational analysis and additional structure-activity relationship studies provided enhancements in kinase selectivity and pharmacokinetics that furnished an advanced series of potent inhibitors. The optimized compounds (e.g., GSK854) are suitable leads for identifying new cardiac medicines and have been employed as in vivo tools in investigational studies aimed at defining the role of TNNI3K within heart failure.

A High-Throughput Dose-Response Cellular Thermal Shift Assay for Rapid Screening of Drug Target Engagement in Living Cells, Exemplified Using SMYD3 and IDO1.

A persistent problem in early small-molecule drug discovery is the frequent lack of rank-order correlation between biochemical potencies derived from initial screens using purified proteins and the diminished potency and efficacy observed in subsequent disease-relevant cellular phenotypic assays. The introduction of the cellular thermal shift assay (CETSA) has bridged this gap by enabling assessment of drug target engagement directly in live cells based on ligand-induced changes in protein thermal stability. Initial success in applying CETSA across multiple drug target classes motivated our investigation into replacing the low-throughput, manually intensive Western blot readout with a quantitative, automated higher-throughput assay that would provide sufficient capacity to use CETSA as a primary hit qualification strategy. We introduce a high-throughput dose-response cellular thermal shift assay (HTDR-CETSA), a single-pot homogenous assay adapted for high-density microtiter plate format. The assay features titratable BacMam expression of full-length target proteins fused to the DiscoverX 42 amino acid ePL tag in HeLa suspension cells, facilitating enzyme fragment complementation-based chemiluminescent quantification of ligand-stabilized soluble protein. This simplified format can accommodate determination of full-dose CETSA curves for hundreds of individual compounds/analyst/day in replicates. HTDR-CETSA data generated for substrate site and alternate binding mode inhibitors of the histone-lysine N-methyltransferase SMYD3 in HeLa suspension cells demonstrate excellent correlation with rank-order potencies observed in cellular mechanistic assays and direct translation to target engagement of endogenous Smyd3 in cancer-relevant cell lines. We envision this workflow to be generically applicable to HTDR-CETSA screening spanning a wide variety of soluble intracellular protein target classes.

A Perspective on Water Site Prediction Methods for Structure Based Drug Design.

Over the last decade, a number of computational methods have been developed, which attempt to evaluate the thermodynamic properties of individual water molecules at the solute-solvent interface, in order to assess contributions to protein-ligand binding. In some cases, these tools tell us what we already know, e.g. that hydrophobic pockets prefer lipophilic substituents, and in other cases the methods only seem to add clarity when retrospectively applied. Hence we have grappled with how to utilize such approaches to understand non-intuitive results and to generate chemistry ideas that otherwise would not have been developed. Here we provide our perspective on these methods and describe how results have been interpreted and applied. We include examples from GSK and elsewhere that highlight how water methods have been (1) utilized retrospectively to explain non-intuitive structure- activity relationships and (2) applied prospectively for chemistry design. Finally, we discuss where this field of study could lead to maximal impact in drug discovery research.

Substituent Effects on Drug-Receptor H-bond Interactions: Correlations Useful for the Design of Kinase Inhibitors.

Investigation of troponin I-interacting kinase (TNNI3K) as a potential target for the treatment of heart failure has produced a series of substituted N-methyl-3-(pyrimidin-4-ylamino)benzenesulfonamide inhibitors that display excellent potency and selectivity against a broad spectrum of protein kinases. Crystal structures of prototypical members bound to the ATP-binding site of TNNI3K reveal two anchoring hydrogen bond contacts: (1) from the hinge region amide N-H to the pyrimidine nitrogen and (2) from the sulfonamide N-H to the gatekeeper threonine. Evaluation of various para-substituted benzenesulfonamides defined a substituent effect on binding affinity resulting from modulation of the sulfonamide H-bond donor strength. An opposite electronic effect emerged for the hinge NH-pyrimidine H-bond interaction, which is further illuminated in the correlation of calculated H-bond acceptor strength and TNNI3K affinity for a variety of hinge binding heterocycles. These fundamental correlations on drug-receptor H-bond interactions may be generally useful tools for the optimization of potency and selectivity in the design of kinase inhibitors.

Structure-Based Design of a Novel SMYD3 Inhibitor that Bridges the SAM-and MEKK2-Binding Pockets.

SMYD3 is a lysine methyltransferase overexpressed in colorectal, breast, prostate, and hepatocellular tumors, and has been implicated as an oncogene in human malignancies. Methylation of MEKK2 by SMYD3 is important for regulation of the MEK/ERK pathway, suggesting the possibility of selectively targeting SMYD3 in RAS-driven cancers. Structural and kinetic characterization of SMYD3 was undertaken leading to a co-crystal structure of SMYD3 with a MEKK2-peptide substrate bound, and the observation that SMYD3 follows a partially processive mechanism. These insights allowed for the design of GSK2807, a potent and selective, SAM-competitive inhibitor of SMYD3 (Ki = 14 nM). A high-resolution crystal structure reveals that GSK2807 bridges the gap between the SAM-binding pocket and the substrate lysine tunnel of SMYD3. Taken together, our data demonstrate that small-molecule inhibitors of SMYD3 can be designed to prevent methylation of MEKK2 and these could have potential use as anticancer therapeutics.

HLA-B*57:01 Confers Susceptibility to Pazopanib-Associated Liver Injury in Patients with Cancer.

PURPOSE: Pazopanib is an effective treatment for advanced renal cell carcinoma and soft-tissue sarcoma. Transaminase elevations have been commonly observed in pazopanib-treated patients. We conducted pharmacogenetic analyses to explore mechanistic insight into pazopanib-induced liver injury. EXPERIMENTAL DESIGN: The discovery analysis tested association between four-digit HLA alleles and alanine aminotransferase (ALT) elevation in pazopanib-treated patients with cancer from eight clinical trials (N = 1,188). We conducted confirmatory analysis using an independent dataset of pazopanib-treated patients from 23 additional trials (N = 1,002). Genome-wide association study (GWAS) for transaminase elevations was also conducted. RESULTS: The discovery study identified an association between HLA-B*57:01 carriage and ALT elevation [P = 5.0 × 10(-5) for maximum on-treatment ALT (MaxALT); P = 4.8 × 10(-4) for time to ALT > 3× upper limit of normal (ULN) event; P = 4.1 × 10(-5) for time to ALT > 5× ULN event] that is significant after adjustment for number of HLA alleles tested. We confirmed these associations with time to ALT elevation event (P = 8.1 × 10(-4) for ALT > 3× ULN, P = 9.8 × 10(-3) for ALT > 5× ULN) in an independent dataset. In the combined data, HLA-B*57:01 carriage was associated with ALT elevation (P = 4.3 × 10(-5) for MaxALT, P = 5.1 × 10(-6) for time to ALT > 3×ULN event, P = 5.8 × 10(-6) for time to ALT > 5× ULN event). In HLA-B*57:01 carriers and noncarriers, frequency of ALT > 3× ULN was 31% and 19%, respectively, and frequency of ALT > 5× ULN was 18% and 10%, respectively. GWAS revealed a possible borderline association, which requires further evaluation. CONCLUSIONS: These data indicate that HLA-B*57:01 carriage confers higher risk of ALT elevation in patients receiving pazopanib and provide novel insight implicating an immune-mediated mechanism for pazopanib-associated hepatotoxicity in some patients.

Decoys for docking.

Molecular docking is widely used to predict novel lead compounds for drug discovery. Success depends on the quality of the docking scoring function, among other factors. An imperfect scoring function can mislead by predicting incorrect ligand geometries or by selecting nonbinding molecules over true ligands. These false-positive hits may be considered "decoys". Although these decoys are frustrating, they potentially provide important tests for a docking algorithm; the more subtle the decoy, the more rigorous the test. Indeed, decoy databases have been used to improve protein structure prediction algorithms and protein-protein docking algorithms. Here, we describe 20 geometric decoys in five enzymes and 166 "hit list" decoys-i.e., molecules predicted to bind by our docking program that were tested and found not to do so-for beta-lactamase and two cavity sites in lysozyme. Especially in the cavity sites, which are very simple, these decoys highlight particular weaknesses in our scoring function. We also consider the performance of five other widely used docking scoring functions against our geometric and hit list decoys. Intriguingly, whereas many of these other scoring functions performed better on the geometric decoys, they typically performed worse on the hit list decoys, often highly ranking molecules that seemed to poorly complement the model sites. Several of these "hits"from the other scoring functions were tested experimentally and found, in fact, to be decoys. Collectively, these decoys provide a tool for the development and improvement of molecular docking scoring functions. Such improvements may, in turn, be rapidly tested experimentally against these and related experimental systems, which are well-behaved in assays and for structure determination.

Identification of Purines and 7-Deazapurines as Potent and Selective Type I Inhibitors of Troponin I-Interacting Kinase (TNNI3K).

A series of cardiac troponin I-interacting kinase (TNNI3K) inhibitors arising from 3-((9H-purin-6-yl)amino)-N-methyl-benzenesulfonamide (1) is disclosed along with fundamental structure-function relationships that delineate the role of each element of 1 for TNNI3K recognition. An X-ray structure of 1 bound to TNNI3K confirmed its Type I binding mode and is used to rationalize the structure-activity relationship and employed to design potent, selective, and orally bioavailable TNNI3K inhibitors. Identification of the 7-deazapurine heterocycle as a superior template (vs purine) and its elaboration by introduction of C4-benzenesulfonamide and C7- and C8-7-deazapurine substituents produced compounds with substantial improvements in potency (>1000-fold), general kinase selectivity (10-fold improvement), and pharmacokinetic properties (>10-fold increase in poDNAUC). Optimal members of the series have properties suitable for use in in vitro and in vivo experiments aimed at elucidating the role of TNNI3K in cardiac biology and serve as leads for developing novel heart failure medicines.