Closed-loop deep learning in early-stage drug discovery - cloud platform for targeted protein degradation

Projektinformationen

CelerisTx - Celeris One Platform

ID Finanzhilfevereinbarung: 190126641

DOI

10.3030/190126641

Projekt abgeschlossen

EK-Unterschriftsdatum 14 Dezember 2022

Startdatum 1 Juli 2022

Enddatum 29 Februar 2024

Finanziert unter

• The European Innovation Council (EIC)

Gesamtkosten

€ 4 860 501,25

EU-Beitrag

€ 2 500 000,00

Koordiniert durch

CELERIS THERAPEUTICS GMBH
Austria

Periodic Reporting for period 2 - CelerisTx - Celeris One Platform (Closed-loop deep learning in early-stage drug discovery - cloud platform for targeted protein degradation)

Berichtszeitraum: 2023-05-01 bis 2024-02-29

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Although we, as humans, already know a lot about medicine and biology and have significant achievements, we still struggle with incurable diseases. Parkinson's disease, Alzheimer's disease, various forms of cancer, and other diseases continue to cost precious lifetimes and cause grief and suffering.

One of the reasons for this is that current therapeutic modalities cannot adequately treat all the disease-causing molecules in the human body. These are also referred to as "undruggable targets". There is a new class of therapeutic molecules that use cellular disposal systems to degrade disease-causing proteins. This class is called degraders and includes PROTACs, LYTACs, molecular glues, and other examples.

As promising as this class is for making incurable diseases curable, it is difficult to develop. Therefore, we are developing novel computational methods to reduce the cost of developing this much-needed medicine and dramatically increase the probability of success.

In 2021 and 2022, we researched and developed our computational methods and are now translating these "in silico" efforts into real biology and chemistry in our new 800sqm wet lab in Graz, Austria. Through laboratory automation and scaling of our platform, we enable the research and development of new medicine.

Our AI-enabled platform predicts biomolecular interactions and leverages generative AI to design novel fragments of small molecule degrades. With the platform we develop in house drug discovery programs and collaborate with big pharma companies.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

We raised preseed and seed funding from venture capital firms focusing on deep tech and life sciences from around the globe.

We finished our biosafety level 2 wet lab facility in Graz, Austria. The building was an old printing facility, and we amended it to meet the legal requirements for such a lab and our in-house requirements regarding biology and chemistry. We now have a working tissue culture, microbiology, general assay division, and high-content imaging in-house, and soon the synthesis chemistry and an operative team of about 22 FTEs to date.

We developed novel graph-based algorithms to predict the formation of ternary complexes, which are required to induce targeted protein degradation with degraders. A ternary complex is a molecular complex consisting of three entities. In our context, it's
1. the protein to be degraded
2. an enzyme that triggers degradation, and
3. a molecule (the degrader) that connects these two biomolecules.

This project's most recent and recognizable success is the positive biological validation of one of our novel compounds for our Oncology I program as hit. We're further validating the hit and so far, have western blots that show degradation and SPR data that confirm the assumption. Soon, we'll start an additional Series based on the results to optimize that compound further.

Forming a ternary complex requires multiple steps, which we divide into DTI (drug-target interactions) and PPI (protein-protein interactions). Additionally, since both fragments binding to the biomolecules are mostly given, the linking fragment (linker) is of significant relevance for degraders. For this reason, generating relevant linkers is an important part of the computer-based pipeline.

The modules of our platform can be hosted in various computational environments. We successfully hosted it so far on Microsoft Azure, Amazon AWS, Google Cloud, and in-house infrastructure.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

We're working on much-needed medicines against currently incurable diseases such as various types of cancers. With our AI-enabled platform, we come up with new ways to develop novel drugs exploiting a novel therapeutic modality - targeted protein degradation with the ubiquitin-proteasome system.

So far, we have published multiple articles in peer-reviewed journals demonstrating the leaps that we exploit in our own for three in-house oncology programs and two collaborations with the two big pharma companies, Boehringer Ingelheim and Merck KGaA, Darmstadt, Germany.

Cell Culture

Chemistry

Biology

Interface between Chemistry and Biology

Microbiology

CelerisTx R&D Center in Graz, Austria