AI-driven discovery platform enables new way to find novel pathways as well as progress towards identifying patients who may better respond to drugs in the clinic
Preclinical data highlights the potential benefits of Exscientia’s AI-driven design to rapidly discover molecules with optimised properties, including GTAEXS-617
Exscientia (Nasdaq: EXAI) today announced the acceptance of three abstracts for poster presentation at the upcoming American Association for Cancer Research (AACR) Annual Meeting 2022, being held April 8-13, 2022, at the Ernest N. Morial Convention Center in New Orleans, LA.
“The abstracts highlighted at AACR demonstrate the potential of our functional precision oncology tools to improve clinical and patient outcomes by guiding treatment and patient selection using a combination of AI and disease-relevant models,” said Andrew Hopkins, DPhil., CEO and founder of Exscientia. “We believe these data demonstrate the exciting potential of our pipeline and further validate our translational research capabilities and AI-driven drug discovery platform as we continue in our efforts to deliver better molecules and identify promising therapeutic assets that have the best chance of clinical success.”
Abstracts Accepted for Poster Presentation:
Title: Enriching for adenosine antagonist patient responses through deep learning
Session Title: Immunomodulatory Agents and Interventions
Abstract Number: #4150
Date/Time: Wednesday, April 13 / 9:00 AM – 12:30 PM CT
Translation of preclinical data to the clinical setting has been a persistent gap in successful drug discovery. In this study, researchers leveraged Exscientia’s AI-driven platform to develop patient gene signatures that could guide and better inform clinical study of new medicine candidates. By using deep learning driven image analysis, researchers are working towards identifying an adenosine-induced, tumour protective, immunosuppression biomarker to potentially improve the likelihood of clinical success for A2aR targeted therapies. Further, by leveraging patient material as well as baseline and treatment condition transcriptomics, Exscientia was able to model and functionally validate patient gene signatures to map the association of anti-cancer immune activity with the inhibition of adenosine signaling by EXS-21546, Exscientia’s clinical stage A2a antagonist, in development for the treatment of high adenosine signature cancers. These encouraging data suggest that stratification of patient gene signatures could be implemented in future studies of EXS-21546 to identify patients that may respond optimally to A2aR targeted therapies.
Title: AI-driven discovery and profiling of GTAEXS-617, a selective and highly potent inhibitor of CDK7
Session Title: Emerging New Anticancer Agents
Abstract Number: #3930
Date/Time: Wednesday, April 13 / 9:00 AM – 12:30 PM CT
Historically, CDK7 inhibition, a validated target that has been shown to severely limit the ability of cancer cells to proliferate in vitro and in vivo, has been challenging to address due to side effect profiles from development candidates, possibly due to covalent binding mechanism of action or poor oral absorption. By leveraging AI models and active learning, Exscientia was able to design an orally bioavailable, highly potent and selective small-molecule antagonist of CDK7, GTAEXS-617, currently in IND-enabling studies as a potential treatment for transcriptionally addicted cancers, including ovarian and breast cancer. Preclinical data show that ‘617 has potent anti-proliferative activity in in vitro models of high-grade serous ovarian cancer (HGSOC) and triple negative breast cancer (TNBC), and potent anti-tumour activity in HGSOC and TNBC xenograft tumour-bearing mice, resulting in complete tumour regression. By leveraging Exscientia’s precision oncology platform, researchers examined the impact of ‘617 on primary patient ovarian cancer samples. On the platform, two response groups of the patient samples began to form, and researchers are continuing to investigate this phenomenon with the aim to define a patient selection biomarker to enrich patients more likely to respond to CDK7 inhibition. The AI-driven platform was able to improve upon historic design concerns with CDK7 inhibitors, including efflux and GI tract toxicity.
Title: Deep learning supported high content analysis of primary patient samples identifies ALK inhibition as a novel mechanism of action in a subset of ovarian cancers
Session Title: New Technologies for Drug Discovery
Abstract Number: #1893
Date/Time: Monday, April 11 / 1:30 PM – 5:00 PM CT
Targeted therapies are needed for patients suffering from a myriad of diseases, but preclinical drug discovery is often performed in murine models which are not human disease relevant and lack the microenvironment and heterogeneity of human biology. This study highlights the potential of Exscientia’s precision medicine platform to identify novel targets and targetable pathways using human disease relevant patient tissue models, which could have the potential to improve patient outcomes by uncovering clinical relevance at the target discovery stage. By evaluating malignant pleural effusion and ascites from 20 patients with ovarian cancer against greater than 80 small molecules using high content microscopy, researchers were able to identify a pathway containing anaplastic lymphoma kinase (ALK) as a potential novel target in a subset of ovarian cancer patient samples. These encouraging data support further research of patient-focused drug development using human disease relevant models and deep learning to better understand the target landscape for ovarian cancer and the potential for the development of novel therapeutic approaches.
For more such updates and perspectives around Digital Innovation, IoT, Data Infrastructure, AI & Cybersecurity, go to AI-Techpark.com.
