Our Four Pillars of Drug Discovery
High probability target selection requires the integrated analysis of many factors. This begins with a patient unmet need in mind, and then includes assessments of normal and disease pathophysiology, experimental results, genomics, and pattern recognition. Target dependency for a given disease is essential for it to be the focus of a drug discovery program. Essential and orthogonal lines of evidence—across data sets, model systems, and independent labs—enhance our conviction. When considering a target, we also give equal weight to the question of therapeutic window. Will it be possible to avoid toxicity at conservative (i.e., high) estimates of drug exposure? Most targets in the published literature do not survive scrutiny when the above criteria are both applied.
In the building of our pipeline, we have chosen to keep an open mind about disease area. While our efforts are concentrated in oncology, we are also working in hematology, immunology, and rare disease. Since few targets survive rigorous vetting, we thought it dangerous to build a company around a technology or thematic biology thesis. It can be tempting to compromise (just a little) on target conviction for the sake of a tighter elevator pitch. We don’t feel that is a good tradeoff. A program cannot survive the wrong choice of targets, but scientists and colleagues can overcome druggability challenges with innovation, hard work and a little bit of luck.
Medicinal chemistry is a relatively new field and continues to evolve rapidly as a scientific discipline. At its core, medicinal chemistry involves the design, synthesis, and optimization of new compounds for therapeutic application. While the field has shifted over time – from empiric in vivo screening, to high throughput in vitro screening, and now to computer-aided drug design – medicinal chemistry remains foundational to drug discovery. We believe that so-called “small molecule” approaches are here to stay, even as new modalities such as complex biologics and cell and gene therapies prove their wonderful value in the clinic. Our chemists leverage traditional approaches and are using new tools including allostery, covalency, molecular glue and heterobifunctional degraders to expand the druggable universe for small molecules.
Computer-aided drug design has changed the conceptual approach to discovery chemistry. In silico simulations can identify chemical scaffolds and functional groups that bind proteins with affinity and selectivity. It is now possible to use cloud computing to perform massive simulations to enable screening of ultra-large chemical spaces, and to explore the intricacies of protein motion and visualize the docking of small molecules into binding pockets. Advances in artificial intelligence and machine learning enable us to better identify synthetically accessible chemical structures, predict their properties, optimize the many parameters important for a medicine, and uncover patterns that otherwise would go unnoticed. Our highly experienced medicine design team uses deep expertise in software engineering to integrate these advances in computation with high quality, real-world experimental data generated by colleagues in adjacent fields, especially structural biology. Core to the Treeline thesis is that the human-machine interaction will be critical to our success, so we place strong emphasis on connecting and leveraging these diverse areas of expertise across our team.
Structural Biology & Protein Science
Our drug targets are proteins, thus it is critical to understand their structures and motions, and to have the capability to measure their functions and biophysical properties. Our team synthesizes, purifies and images proteins at atomic resolution. We exploit various subdisciplines including SPR, mass spec, crystallography, in-silico molecular dynamics, and cryo-EM, to illuminate novel protein targets and complexes. With these experimental data in hand, computational chemists can then build truly predictive in silico models. In turn, medicinal chemists can iterate within a lead series of compounds to optimize biologic effect, target specificity, or physiochemical properties.