Considering the important role that cyclin dependent kinases play in regulating the cell cycle, these proteins are attractive targets for drug development. Selective CDK4/6 inhibitors such as palbociclib, abemaciclib, and ribociclib have become mainstays for use in HR+ breast cancer. Building upon the success of CDK4/6 inhibitors, attention has recently turned to the development of selective CDK2 inhibitors. Recent clinical evidence has implicated increased CDK2 activity in resistance to CDK4/6 inhibitors. However, the broadly essential kinase CDK1 has high homology to CDK2 and is a key off-target challenge in the development of CDK2 inhibitors. We describe here our efforts in utilizing a novel molecular dynamics based strategy for the identification of selectivity via a residue conserved in both CDK2 and CDK1, despite nearly homologous ATP binding sites. This strategy led to a dramatic enrichment in compounds with >100x selectivity over CDK1 as well as tumor growth inhibition activity in efficacy studies. In the course of advancing these selective and potent compounds, we encountered unique clearance mechanisms in rodent that were not predicted from in vitro studies and did not apply to higher species. We will describe the additional characterization that was undertaken through detailed mechanistic in vivo studies and the relevance to human dose prediction. These overall learnings have implications for medicinal chemistry programs where selectivity and species-specific PK challenges are encountered.

Protein Arginine Methyl Transferase 5 (PRMT5) is a type II arginine methyltransferase that catalyzes the post-translational symmetric dimethylation of multiple protein substrates. This enzyme plays a critical role in regulating numerous biological processes including transcription, cell cycle progression, spliceosome assembly, and DNA repair. As such, dysregulation of PRMT5 activity is implicated in the development and progression of multiple cancer types, and is therefore a clinical target of growing interest within the pharmaceutical industry. This walk will describe the modeling and structure-based drug design, and robust synthetic efforts, that enabled the identification of a novel 5,5-fused bicyclic nucleoside-derived class of potent and selective PRMT5 inhibitors. Specifically, the utilization of compound docking and strain energy calculations to inspire novel designs will be addressed, along with the development of flexible synthetic approaches which allowed us to access structurally complex chemotypes (5 contiguous stereocenters). Specific efforts to overcome suboptimal solubility, bioavailability, and Cyp TDI will also be addressed. Ultimately, the team was able to rapidly prioritize and identify several diverse lead compounds with overall favorable properties, including promising in vivo activity and a low QD human dose projection.

The concept of a pharmacophore is of critical importance in the field of drug discovery. A pharmacophore relates the arrangement of chemical features in space to an observed endpoint, and there are various equally valid interpretations - any pharmacophore hypothesis is as valid as any other. Their usage is not only limited to help understand the interactions between ligands and macromolecules, but can also be applied to guide virtual screening efforts and fragment-based drug designs.

In the current work we explore the automated creation of pharmacophores using a variety of different computational strategies, and compare them to pharmacophore hypotheses created by expert users. The problem of pharmacophore comparison is non-trivial and foundational in its nature, some methods to enable such comparisons will be disclosed.

We will describe which data inform which approach to employ, and how to optimize pharmacophore hypotheses using complementary facets of each method. The insights shared will be directly actionable by any medicinal chemistry team looking to use pharmacophores in their work.

This computational study employs Molecular Mechanics Generalized Born (MMGB) analysis to evaluate the binding energetics of methylthioadenosine (MTA)-cooperative PRMT5 inhibitors using X-ray co-crystal structures. These inhibitors selectively target the PRMT5/MTA complex formed in MTAP-deficient cancers, leveraging MTA accumulation to enhance binding specificity. Structural analyses reveal critical interactions stabilizing inhibitor-PRMT5/MTA ternary complexes, including halogen bonds, hydrogen bonds with backbone carbonyl groups, and van der Waals contacts with MTA’s methyl group. MMGB calculations highlight energetic contributions from cooperative MTA binding, demonstrating 6–280-fold selectivity for PRMT5/MTA over apo-PRMT5. The simulations further rationalize potency differences between inhibitor scaffolds, showing how substitutions modulate steric complementarity while maintaining key interactions. This energetic profiling provides a mechanistic basis for the observed synthetic lethality in MTAP-null tumors and informs structure-guided optimization of brain-penetrant inhibitors with improved therapeutic indices.

Toll-like receptors (TLR) 7 and 8 play critical roles in modulating adaptive and innate immune responses, making them promising therapeutic targets in the field of immuno-oncology. Despite the potential of TLR7/8 agonists to enhance anti-tumor immunity, systemic administration often leads to poor tolerance, highlighting the need for better delivery methods. Recent advancements and clinical successes with antibody-drug conjugates (ADCs) offer a promising targeted delivery strategy to mitigate the tolerability issues linked to systemic TLR7/8 agonist use. This presentation outlines the development of a novel series of potent, pyrazolopyrimidine-based TLR7/8 agonists designed as ADC payloads. Through structure-based design, we generated a class of agonists with tunable receptor selectivity and desired payload properties. Representative ADCs made from this series of payloads demonstrate robust efficacy in both in vitro and in vivo preclinical models.