High content phenotypic and AI-enabled discovery of a new pipeline of innovative therapeutics tailored towards targeting Glioblastoma heterogeneity

Prof Neil Carragher n.carragher@ed.ac.uk

Funding information (students eligible to apply): UK/International students

Project Description

Glioblastoma (GBM) is the most common and aggressive primary brain tumour in adults. Since the introduction of post-surgical, concomitant radiotherapy and adjuvant temozolomide in 2005 (the ‘Stupp protocol’) and despite substantial drug discovery investments there have been no new effective treatments introduced for GBM in the past 20 years. As a consequence, prognosis is poor and 5 year survival rates from time of diagnosis remain at less than 5% ¹. Significant challenges in the treatment of GBM include effective drug delivery and overcoming intra-patient tumour heterogeneity.

Recent advances in high content screening technologies integrating with Artificial Intelligence/Machine Learning (AI/ML) methodologies are transforming in vitro phenotypic screening of drug libraries in cell-based models of disease². While phenotypic screening of relatively small libraries of target annotated and approved drug sets may help identify some opportunities for onward translation, the potential to discover truly novel therapeutic strategies for GBM is limited by the restricted chemical and target diversity of these libraries. Further limitations include use of industry standard median well-level aggregated phenotypic measurements which fail to capture the phenotypic diversity and heterogeneity present in more relevant patient-derived GBM stem cell lines. This PhD project proposal aims to fully exploit the latest advances high content imaging and AI/ML methodologies to identify novel treatment strategies which overcome GBM heterogeneity through three complementary research workpackages:

1. Screen a bespoke 20k CNS-penetrant small molecule library  that we have specifically designed using novel chemoinformatics approaches and AI/ML to ensure low toxicity and high predictability for crossing the blood-brain-barrier and targeting mammalian proteins. Primary screening will be performed using the high-throughput ‘Cell Painting’ microscopy assay already established in our group across a panel heterogeneous patient-derived GBM stem cell lines. We will prioritize hit compounds that are active across >2 genetically diverse GBM stem cells with minimal activity on normal neuronal stem cells.
2. In collaboration with colleagues in Edinburgh School of Informatics we will apply novel AI/ML methods to interrogate multi-parametric phenotypic signatures and link phenotypic, structural and physicochemical compound fingerprints to identify further hits, with high predictivity of targeting specific GBM phenotypes (e.g. senescence; mitochondrial stress and invasive capacity) through AI-enabled virtual screening³.
3. We will feed hit compound structures into in silico target prediction pipelines, triaging compound hits for further validation and development based on biological target and ‘med-chem’ tractability. We will validate prioritized hits through follow up dose-response investigation in both 2D and 3D in vitro GBM stem cell models.

Expected training outcomes and career development opportunities:

• High throughput drug screening including; compound library handling and laboratory automation (liquid and plate handling robotics)
• Cell culture: 2D and 3D spheroid patient-derived glioblastoma models
• Cell Pharmacology
• Automated quantitative Image analysis
• Implementation of Artificial Intelligence/Machine Learning workflows
• Intellectual property review and freedom-to-operate search
• Communication skills
• Presentation at national and international meetings

References

1. Johanssen T, McVeigh L, Erridge S, Higgins G, Straehla J, Frame M, Aittokallio T, Carragher NO, Ebner D. Glioblastoma and the search for non-hypothesis driven combination therapeutics in academia. Front Oncol. 2023 Jan 17;12:1075559. doi: 10.3389/fonc.2022.1075559.
2. Way GP, Sailem H, Shave S, Kasprowicz R, Carragher NO. Evolution and impact of high content imaging. SLAS Discov. 2023 Sep 3:S2472-5552(23)00066-7. doi: 10.1016/j.slasd.2023.08.009.
3. Smer-Barreto V, Elliott R, Dawson J, Lorente-Macías A, Unciti-Broceta A, Oyarzún D, Carragher NO. Identification of drug candidates against glioblastoma with machine learning and high-throughput screening of heterogeneous cellular models. bioRxiv 2025.03.06.641926; doi: https://doi.org/10.1101/2025.03.06.641926

Categories

Biological Sciences

Bioinformatics; Cancer Biology; Cell Biology; 

Chemistry

Computational Chemistry; Pharmaceutical Chemistry; 

Computer Science

Artificial Intelligence; Machine Learning; Networks

Mathematics

Data Analysis

Medicine

Pharmacology