Spatial Biology in Drug Discovery

Today’s spatial proteomics and transcriptomics platforms allow researchers to probe gene expression in tissue sections with subcellular resolution. Profiling molecular phenotypes while retaining spatial information of expression is not only suggesting new relationships among signaling molecules, but is also deepening our understanding of disease mechanisms. This article looks at advances in spatial biology and how they are helping to advance disease research and drug discoveries.

Iterative, whole-slide spatial proteomics

Grappling with the underpinnings of any human disease is made easier when you know the cell types involved, and the important molecules they express. The specific architecture of cancerous tumors is an excellent example. “We are increasingly learning how important the specific tumor microenvironment is in determining the progression of a cancer and its response to therapies, both traditional and emerging,” says Prachi Bogetto, Senior Manager, Translational Research at Leica Microsystems.

The Cell DIVE platform from Leica Microsystems supports whole-slide imaging of protein biomarkers with single-cell resolution. “The sheer number of biomarkers that can be visualized using Cell DIVE can drive precise phenotyping of many different cell types, [and allow] for critical biological pathways to be explored in concert, rather than in isolation,” says Bogetto. Cell DIVE labels tissues using an iterative protocol that gives researchers the flexibility to change the design of a study as needed, depending upon initial results. “The labeling protocol is available to use with any antibody panel that answers the question and is tissue preserving,” she adds. “This allows re-probing of the same tissue as the questions evolve in this nascent but rapidly evolving area.”

A new slide carrier for Cell DIVE, ClickWell, removes the potentially tissue-damaging steps of adding and removing coverslips, and makes the bench workflow compatible with automation methods. “This allows researchers to increase the ease and scale of large multiplexed imaging studies, which will help accelerate understanding of disease biology and drug discovery,” says Bogetto.

Correlating Cell DIVE’s proteomics data with additional information further expands our understanding of cell states and disease. In a recent study headed by Maryam Pourmaleki at the Memorial Sloan Kettering Cancer Center, researchers combined single-cell proteomics using Cell DIVE with transcriptomics and genomics data to investigate melanoma tumor progression. “This correlated data could inform the efficacy of IL-2 treatment in patients based on the tumor heterogeneity that is revealed,” says Bogetto.

Spatial transcriptomics and cellular neighborhoods

Like people, cells don’t live in isolation; scientists are learning that cellular neighborhoods are composed of different cell types that influence each other. Vizgen’s recently released MERSCOPE™ platform uses a massively parallel imaging technique called MERFISH (multiplexed error-robust fluorescence in situ hybridization) to measure the copy number and spatial distribution of 500 different genes with subcellular resolution.

Vizgen is working to profile such neighborhoods with Nir Hacohen, Director of the Cell Circuits Program at the Broad Institute of MIT and Harvard, and Director of the Center for Cancer Immunology at Massachusetts General Hospital. “These [neighborhoods] are niches within the tumor microenvironment, groups of cells that are all interacting together to create functional hubs within the colorectal tumor,” says Emanuel. “Being able to profile these tumors and all these interactions at the neighborhood level has a lot of promise, and can help to understand how to perturb them to make the tumor less viable.”

Vizgen also collaborated with Compugen, a company interested in discovering unknown drug targets and expanding use of cancer immunotherapies, to map different cell types and their potential interactions in colorectal tumors. “In particular, they found activated dendritic cells secreting CXCL10 adjacent to CXCR3+ CD8 T cells, suggesting there’s some sort of cell-cell communication happening between these cell types,” says Emanuel. Using spatial profiling to interrogate cell types and signals involved in organized, cellular neighborhoods such as the tumor microenvironment is an approach that is only just beginning to bear fruit.

Vizgen’s platform will soon expand to accept formalin-fixed, paraffin-embedded (FFPE) tissue, a common sample type that can be especially helpful when studying infectious diseases. “With fresh frozen tissues, an infectious agent is still active, so it requires certain biohazard considerations,” says George Emanuel, Scientific Co-founder and Senior Director of Technology and Partnerships at Vizgen. “But if you can use FFPE tissue where everything’s deactivated early on, it makes it much easier to use infectious disease samples.”

Further expansions are on the horizon. Because proteins can mark information distinct from that marked by RNA, Vizgen will also soon add the option of up to 5 custom protein biomarkers along with a gene panel. “Proteins can help to trace some cellular or subcellular morphologies, for example the projections coming out of microglia or astrocytes,” says Emanuel. Protein biomarkers, in addition to increasing the profiling panel from 500 to 1000 genes, will give Vizgen users an even broader range of spatial biology tools.

Spatial proteomics in translational studies

Spatial biology platforms have the potential to accelerate the translational work of drug discovery. According to Cliff Hoyt, VP of Translational and Scientific Affairs at Akoya Biosciences, Akoya’s spatial proteomics platform is ideal for the spatial phenotyping used in translational studies in drug discovery, which involves analyzing the entire sample (perhaps 10s to 1000s of tissue slides) with great sensitivity and reproducibility. Researchers start with a drug candidate and look for signatures of biomarkers within its effects (whether genomic, transcriptomic, or proteomic) that help to understand the drug’s mechanism of action, among other things. “The problem becomes not just looking for co-expression and spatial biology, but also trying to understand that information in the context of the significant heterogeneity between patients and within patients,” he says. Typically, around 3 to 6 biomarkers are needed to capture a drug’s biological signature. While more could be included, using lower-plex panels facilitates scanning of entire slides.

In translational work, even low levels of protein expression can affect drug interactions and responses. Using a spatial phenotyping platform such as Akoya’s, which has an amplified workflow to boost sensitivity, is helpful. “In a translational platform, there’s as much or more emphasis on the analytical performance as the markers themselves,” says Hoyt. “That’s another area that we are really focusing on—creating a sample analysis workflow that is analytically very robust.”

Because translational studies scan whole tissues, detecting and characterizing millions of cells, data-mining tools are vital for finding valuable spatial biology signatures. Akoya has partnered with Johns Hopkins University scientists who developed AstroPath, a platform that uses star-mapping astrophysics principles in order to predict immunotherapy responses in some types of tumors. Akoya has also partnered with a computational biology group at AstraZeneca to mine data for actionable measurements. Hoyt is optimistic about the future of using predictive biomarkers in drug combination trials. “In some trials, people are throwing everything at the wall to see what sticks,” he says. “Biomarkers are going to help to understand why a patient responds, and what is the actual biological mechanism of drug interaction.”

Hoyt is glad that multiplex spatial biology platforms are now able to handle the increasing complexity of today’s research questions. “It’s an exciting time for multiplex spatial biology platforms, because there’s such a need for them,” he says. “Especially when it's no longer a drug directly targeting a tumor cell—it’s now about affecting how cells interact with one another.”