Science

Heterobifunctionals

We are leaders in heterobifunctional degrader design. Our integrated, end-to-end capabilities represent a step change by reducing the typically empirical optimization of this new modality through ternary complex prediction, thus fully unlocking the potential of delivering transformative new medicines.

Heterobifunctional degraders have rapidly moved from a mere scientific curiosity to a highly promising new modality with vast potential to modulate diseases with small molecule-like chemistry. While the field has harvested the lowest hanging fruits, we are now entering a phase where ligands for a protein of interest may not be available or may require significant optimization.

At Roivant, we are challenging convention by accelerating ligand and heterobifunctional optimization through the integration of physics- and Machine Learning-based predictive sciences, continuously reinforced with experimental data across all drug discovery disciplines.

Cornerstones

Our end-to-end differentiated capabilities to prosecute targets across therapeutic areas via heterobifunctionals rest on four cornerstones, which we are applying beyond degraders:

01

Degradability Demonstration and E3 Ligase Fitness

We determine key protein parameters experimentally and demonstrate degradability and E3-ligase fitness. Our in-house proteomics platform delivers a deep understanding of degradation in the disease context. Using disease-relevant cell lines we determine the expression levels of the protein of interest (POI) and of E3 ligases, as well as their half-life and post translational modifications. The integration of these data generates a protein and ligase profile which informs the suitability of the POI and E3 ligase for degradation.

We are utilizing our Biodegrader platform to elucidate the degradability of a protein or the fitness of an E3 ligase, in cases where a ligand for the POI or E3 ligase may not exist. These cell systems interrogate the ability of a protein-reporter fusion to degrade the protein of interest. Excitingly, our approach allows the screening of multiple E3 ligases, thus inferring which one is the most suited for a given protein of interest.


02

New Proprietary Ligands

The combination of integrated hit finding with our leading QUAISAR platform enables the rapid discovery of proprietary ligands.

We are combining integrated hit finding screening technologies with proprietary compound collections to identify high quality proprietary ligands for both the protein of interest and novel, and in some cases tissue selective, E3 ligases. This covers virtual-ligand screening (VLS) leveraging our computational platform, high throughput screening (HTS), DNA-encoded libraries (DEL), fragment and covalent and knowledge base libraries. In addition, our physics-based platform, combined with our integrated experimental capabilities serves to accelerate hit and lead optimization approaches.


03

Predictive Heterobifunctional Assembly

We apply Dynamite, our predictive heterobifunctional assembly, into our bespoke degrader Design-Predict-Make-Test cycle. Dynamite is a set of predictive models of ternary complex assembly that allow for the rational design of heterobifunctional molecules that drive high degradation efficiency, with spatiotemporal control, while maintaining drug-like properties fit for therapeutic purpose. Our models focus on three essential steps for recruited ubiquitin-proteasome system degradation:

  • formation of the ternary complex induced by a degrader molecule
  • conformational heterogeneity of the ternary complex,
  • degradation efficiency via the ubiquitin ligase macromolecular assembly

To apply our method to a protein of interest, we combine experimental biophysical data with molecular dynamics (MD) simulations to accurately predict ternary complex structures at atomic resolution. We integrate hydrogen-deuterium exchange mass spectrometry (HDX-MS, which measures the solvent exposure of protein residues) directly into advanced MD simulation algorithms to improve the efficiency and accuracy of the ternary structure predictions ( Atomic-Resolution Prediction of Degrader-mediated Ternary Complex Structures by Combining Molecular Simulations with Hydrogen Deuterium Exchange).


04

Enhanced Delivery

To broaden the therapeutic potential of heterobifunctional degraders, we are applying our expertise in formulation enhanced delivery to identify fit for therapeutic routes of administration.

Achieving oral bioavailability for non-cereblon recruited heterobifunctional degraders remains a significant challenge. In order to systematically enable oral delivery beyond cereblon ligands, we are driving our molecular design: by integrating in silico assays focusing on the identification of low clearance, good solubility, maximum permeability, and high catalytic protein turnover early in our design process. Thoughtful and disciplined design maximizes the probability of our heterobifunctional degraders being amenable to oral absorption. Oral delivery is then further enabled through advanced formulation and delivery, for which we are investing in the development of formulation and delivery techniques intended to increase bioavailability.


video section decoration

Atomic-Resolution Prediction of TPD Complex Structures by Combining MD with Experimental HDX-MS


Expansion to Novel Modes-of-Action with Heterobifunctionals

Heterobifunctional degraders are a demonstration of how bivalent molecules can bring proteins together in a non-canonical way, and thus rewire biology. This chemically-induced proximity to hijack biological processes and thus modulate cellular dysfunctions to treat diseases is therefore a vast opportunity to enable novel modes-of-actions, particularly for function up-regulation, which is typically a hard drug discovery challenge. All our heterobifunctional capabilities can and are being translated into other mechanisms.

Keep reading

02

Covalency

We develop Covalent Therapeutics using our proprietary suite of proteome-wide datasets, assays, and covalent library.

03

Designing Best-in-Class Small Molecules

Our computation-first approach is used to design molecules that rapidly reach the desired TPP for disease areas where effective therapeutic solutions are lacking.