Innovation is leading to a comprehensive ‘map’ of tumours

For many decades, our understanding of the spatial biology of tumours came only from microscopy, for example, using immunohistochemistry to identify specific proteins within a tumour section. However, advances in spatial transcriptomics technology mean that we can now pinpoint which genes are active within different parts of the tumour and its microenvironment.

In my research group, we first used spatial transcriptomics to investigate prostate cancer and revealed an unexplored landscape of heterogeneity (Nat Commun. 2018;9:2419). Using novel computational procedures, we were able to elicit spatial, transcriptome-wide expression patterns enabling deconvolution of molecular events in the cancer and associated microenvironment. We investigated different tumours within an individual patient and from there were able to establish an evolutionary tree, piecing together the different genetic steps that led to the formation of the multifocal tumours (Nature. 2022;608:360–367). Unexpectedly, we also noted that apparently normal epithelial cells near the tumour carried some genetic alterations found within the tumour, a finding that was not evident by microscopy.

One of the most exciting clinical applications of these findings and this type of research is the potential to discover commonalities in precursor lesions that could be useful for early diagnostics, circumventing the heterogeneity that characterises more advanced disease. One could envisage that genetic findings could eventually be translated into plasma protein biomarkers that are sufficiently specific to replace crude measures such as prostate-specific antigen.

Building on our work with primary prostate tumours, we went on to analyse metastases for the first time, genetically evaluating the clones that are able to seed elsewhere. Identifying markers that could help predict metastasis before the event is another interesting opportunity and something that we are currently investigating. By profiling multiple whole prostate organs, we hope to identify common denominators for metastasis, and with that information, analyse core biopsies to see if any changes can be detected in the earlier stages of disease.

In addition to the individual evolution of tumours, there is patient-specific adaptation of the immune system. We used transcriptomics of B-cell and T-cell receptors to explore lymphocyte clonal dynamics within tissues (Science. 2023;382:eadf8486). This approach led us to describe niches of immune cells, thereby providing important information on how our immune system recognises different parts of the tumour. With more extensive analysis, we may be able to gain insights into novel tumour antigens, and thus identify new actionable targets, and learn to predict how an individual’s immune system will react to their tumour.

With analyses of breast cancer, we have reported previously unidentified relationships (Nat Commun. 2021;12:6012; Nat Genet. 2021;53:1334–1347). In addition to neoplastic cell heterogeneity, we observed mesenchymal cells displaying diverse functions and cell-surface protein expression through differentiation within three major lineages. Using single-cell signatures, we deconvoluted large breast cancer cohorts to stratify them into nine clusters, termed ‘ecotypes’, and found that each was associated with unique cellular compositions and clinical outcomes. In more recent work, we have moved away from gene expression to study spatial metabolomics related to the release of metabolites (Nat Biotechnol. 2024;42:1046–1050). These studies are made possible by advances in spatial mass spectrometry imaging that enable assessment of diverse biomolecules from intact tissues at high resolution, which could in turn lead to new therapeutic possibilities.

In contrast to bulk analysis, spatial transcriptomics provides an ‘atlas’ – a comprehensive map of each tumour in each experiment. With state-of-the-art technologies, we can gain an even greater understanding of the complexity of all the different compartments and their interactions to open up new avenues of research that we could never have imagined before.