The molecular imaging decoding human tissue in health and disease

Summary: In order to bring new medicines to patients faster and more effectively, we have to improve our understanding of the diseases we want to treat. Using advanced molecular imaging – propelled most recently by mass spectrometry – our scientists are able to probe and analyse tissue samples in a depth and detail that was previously impossible. By harnessing the incredible powers of AI and machine learning, the full molecular complexity is finally becoming decipherable and is already revealing insights that have the potential to fundamentally transform future drug discovery and development.

How molecular understanding of diseases is helping evaluate drug safety

Integral to modern drug discovery is the ability to understand – in as much detail as possible – the effects of our compounds on human tissues at a cellular level. In other words, exactly what our candidate drugs are doing in a patient’s body. For decades, scientists have used the traditional techniques of histology and histopathology, staining tissue samples and looking for particular cellular morphologies, markers and signals on microscope slides. To work out where drug molecules go within the body (biodistribution), previously the only available option was autoradiography, an expensive and laborious radio-labelled approach with limited ability to only image distribution of a single target.

Now, a suite of innovative technologies which make up our advanced molecular imaging capabilities have enabled a quantum leap in our ability to understand disease processes and evaluate drug efficacy and safety. These technologies let us probe every tissue sample – whether from patient biopsies, animal models or advanced cell cultures – in unprecedented depth. Combining these remarkable powers of detection with the incredible analysis and interpretation capabilities of artificial intelligence (AI) and machine learning means we can explore uncharted territory with open minds in our search for the unknown and unexpected.

Navigating the human body’s molecular complexity with mass spectrometry imaging

In this field, advances in mass spectrometry imaging (MSI) have been the real game changer. We are now able to simultaneously measure the individual masses of molecules using a mass spectrometer, to visualise their spatial distributions – whether peptides, proteins, lipids, endogenous metabolites or drug molecules – within the microenvironment of a tissue, providing vital clues to their inter-relationships and allowing us to better assess the safety and efficacy of compounds.

Mass spectrometry itself is of course not new. Utilised in many areas of research and development, it relies on the process of ionisation: turning tissue samples into gaseous form. What is special about MSI is that the tissue sample is not homogenised (mixed-up) prior to ionisation. Instead, it is snap-frozen and ionised directly from its intact surface, so that each molecule’s original position is known. The whole sample is scanned a few microns at a time with each discrete location analysed forming one pixel in the images we ultimately generate. This provides a wealth of digital information.

Creating a ‘Google Earth’ view

Today’s advanced molecular imaging techniques – and MSI in particular – mean we can now create the most detailed molecular maps ever. Much like Google Earth software enables a satellite view of the planet down to 3D views of individual streets and buildings, our new imaging capabilities allow us to zoom in and out from the micro to the macro level and back. Our mass spectrometers even enable us to do the equivalent of looking through the window of a house to see where the sofa is.

Every molecule we detect has its own map, pieced together by tens of thousands of images from different perspectives. We can “see” drug molecules, biomarkers and the tissue microenvironment simultaneously and examine the picture from the genomic and molecular viewpoint up to the cellular, tissue, organ and patient level. Vast datasets are generated from healthy, diseased and drug-treated tissue samples which can then be mined by AI and machine learning techniques to spot patterns, connections and relationships at a level of complexity far greater than ever before, turning information into insights and insights into knowledge. 

Armed with this knowledge we’ll be more equipped to design safe and effective drugs, develop optimal drug delivery methods, work out appropriate dosing and monitor disease progression. Using these approaches, we are already uncovering new insights such as:

• The first mechanistic description of how metabolites generated by the gut microbiome might play a role in neurological conditions like Parkinson’s¹
• The mechanism of nutrient sensing and utilisation in lung metastasis² and colorectal cancer³
• A novel 3D lung adenocarcinoma model that can mimic the in vivo tissue and the tumour microenvironmennt⁴
• Preclinical imaging to explore the penetration of oncology drug candidates across the blood brain barrier and their impact on brain tumour growth⁵
• High-resolution 3D visualisation of nanomedicine distribution providing valuable insights into the influence of the tumour vasculature and microenvironment on nanomedicine localisation⁶

The Cancer Research UK ‘Grand Challenge’

AstraZeneca is part of an innovative five-year multi-centre, multidisciplinary collaboration that aims to plot the most detailed map ever seen of the molecular landscape of malignant tumours.

Funded by Cancer Research UK (CRUK), this ‘Grand Challenge’ is led by Professor Josephine Bunch from the National Physics Laboratory and the consortium includes leading scientists from the Francis Crick Institute, the Beatson Institute in Glasgow, the University of Cambridge and many others.

The team’s database will be made available to researchers worldwide and its approaches used to create standardised, best-practice guidance for deployment of the latest technologies.

Partnering with academia is key for us to push the boundaries of science and drive disease discovery and understanding to help us create the next generation of therapeutics.

References

1. Microbiome-derived carnitine mimics as previously unknown mediators of gut-brain axis communication - 2020 Mar 11;6(11):eaax6328. doi: 10.1126/sciadv.aax6328. 
2. ILC2-driven innate immune checkpoint mechanism antagonizes NK cell antimetastatic function in the lung - 2020 Sep;21(9):998-1009. doi: 10.1038/s41590-020-0745-y
3. The amino acid transporter SLC7A5 is required for efficient growth of KRAS-mutant colorectal cancer. Nat Genet 53, 16–26 (2021). http://doi.org/10.1038/s41588-020-00753-3
4.  Characterization of an Aggregated Three-Dimensional Cell Culture Model by Multimodal Mass Spectrometry Imaging - 2020 Sep 15;92(18):12538-12547. doi: 10.1021/acs.analchem.0c02389. 
5. Preclinical comparison of the blood brain barrier (BBB) permeability of osimertinib with other EGFR TKIs - 2020 Oct 7;clincanres.1871.2019. doi: 10.1158/1078-0432.CCR-19-1871. 
6. High-resolution 3D visualization of nanomedicine distribution in tumors - 2020 Jan 1;10(2):880-897. doi: 10.7150/thno.37178.