Written by:
Senior Director, Imaging & Data Analytics, Clinical Pharmacology & Safety Sciences, AstraZeneca
Executive Director, Bioscience, Early Oncology, AstraZeneca
To advance our understanding of disease biology, we are using the power of multi-omics to gain novel insights into the molecular changes that underpin cancer, and response to treatment. By combining different ‘omics’ – such as genomic, metabolomic and proteomic information – our ambition is to improve patient selection in clinical trials, develop targeted therapies and advance precision medicine approaches.
What is multi-omics data?
To understand a complex, heterogeneous disease like cancer, we need to build a full view of its biology. Multi-omics is a powerful predictive and diagnostic approach, which is helping us achieve just this by bringing together multiple complex ‘omics’ datasets, which machine learning and AI can translate into meaningful biological insights.1,2
In Oncology R&D at AstraZeneca, we use multi-omic imaging in pre-clinical discovery to influence clinical decision making and inform patient selection. It is already helping our scientists understand drug responses and will allow physicians to select patients most likely to respond to treatment.
For this reason, we are in the process of embedding our multi-omic platform across our AstraZeneca pipeline, to gain new insights into how our medicines work, and ultimately bring new precision medicines to patients.
Importantly, our approach allows us to continue to layer in more complexity from new analytical sources so we can delve even deeper into the underlying mechanisms of cancer and other disease biology.
Capturing multi-omics data with imaging
Whereas conventional methods use processed whole tissue for analysis, new advanced molecular imaging technologies can analyse individual cells within intact tissue, using so called spatial multi-omic analysis.3 In this way, high resolution multi-omic imaging data can be captured, pixel by pixel across tissue, and added into a combined dataset to create a richer map of disease biology.3
To achieve this deeper insight, we are pioneering multi-omic imaging by combining multiple datasets to analyse each tissue sample. For example, in our collaborative study published in Nature Metabolism, mass spectrometry and spatial transcriptomics revealed functional changes in breast cancer tissue driven by the myc gene, a known tumour driver.4,5
By analysing the proteome and transcriptome within intact tissue sections, spatial multi-omics revealed that depriving human and mouse mammary tumours of vitamin B5 reduces their growth – an insight that may be useful in developing cancer therapies in the future.4
Pioneering metabolic tissue imaging in cancer
Over the last decade, novel tissue imaging methods such as mass spectrometry imaging (MSI) have been developed, with which we are now able to monitor drug distribution, metabolism and delivery as they happen6, a leap forward from traditional methods that can only capture the final response to treatment.
Using MSI, we can image thousands of metabolites in preclinical and clinical studies across each tissue, together with drug concentration and distribution. These data can be combined with other multi-omic datasets to generate new insights into the relationship between tumour drivers, tumour metabolism, and treatment response, and also to identify potential new disease biology.
For example, in a second study we published in Nature Metabolism, multimodal mass spectrometry-based metabolomics and imaging mapped the impact of common genetic drivers of colorectal cancer on intestinal metabolism.7
By capturing the differences between healthy cells and cancer cells, we were able to detect genotype-dependent metabolic changes and identify that targeting a key metabolic pathway has potential future therapeutic value for colorectal cancer.7
Metabolic imaging technologies can also be used to characterise metabolic pathways in patients, complementing insights from other omics approaches.8 Our collaborative study in Proceedings of the National Academy of Sciences represents the first-time spatial metabolomics, transcriptomics and immunohistochemistry were used to predict, and mechanistically explain, the risk of disease reoccurrence in patients who have had surgery for prostate cancer.9
This is a new insight into prostate cancer progression, and could become a future non-invasive approach to rapidly assess patient tumours.9
The future of multi-omics
In addition to oncology research, we are beginning to apply multi-omic imaging technology across other disease areas at AstraZeneca, including cardiovascular, renal and metabolic diseases, respiratory and immune-related diseases, neuroscience and rare diseases.
As our understanding of the molecular changes that underpin disease and response to treatment improves, so will our ability to develop precision medicines that target the right medicine, to the right patient, at the right time.
Multi-omics technologies are advancing rapidly, and we are excited for their potential to transform drug development, and help bring new, targeted medicines to patients.
