Frontiers in Spatial Genomics and Transcriptomics

Spatial genomics and transcriptomics are powerful tools that allow scientists to connect their knowledge of cells, biomarkers, and disease onset, providing novel and enriching answers and perspectives. This includes the spatial location of key cellular components and their various associations, including mRNA and proteins, and their role in influencing intracellular and extracellular dynamics. Spatial transcriptomics uses data from cells’ transcriptomes, or their whole mRNA content, to determine their roles and functions. This article reviews some of the technologies powering advances in the field, including current trends and innovations, and highlights future challenges and possibilities—some of which were unthinkable a few years ago.

Creating atlases

One major development transforming the field is “atlasing,” which involves creating an exhaustive map of cells within a specific area of the human body. This could help scientists better understand the nature of different cell types within a tissue, as well as the connections between these cells, observes Jacob Stern, Director of Product Management at 10x Genomics. This could benefit several areas, including research and development on chronic diseases, from cancer through to Alzheimer’s. It could even help shed light on limb regeneration in axolotl amphibians. Previously, the interplay of histology and spatial analysis has only been possible through the outputs of specialized researchers. “Imaging data and genomic data are orthogonal and complementary, and there are different communities of researchers that have developed expertise in analyzing each. Combining these data could yield greater insights than the sum of their parts,” Stern comments.

Researchers such as Michela Noseda, Senior Lecturer at the National Heart and Lung Institute at Imperial College London, routinely carry out critical spatial transcriptomics-based atlasing work. She leads a team focused on the causes of heart failure and is a co-principal investigator of a project aiming to create a human heart cell atlas. Her group’s first task requires a combination of different approaches. “We are using cardiac cellular and molecular biology in connection with several transcriptomic experiments to understand disease from the perspective of a single nucleus or a single cell, combining spatial transcriptomics and multiomics with gene expression as well as protein analysis.”

This has been made possible by studying human tissues from organ donors and patients affected by heart diseases. Her research has, at the same time, opened up new challenges, for example, how to extend the application of these techniques to tissues from female and adolescent patients, as getting the necessary access to tissues and ensuing data in these situations can present more difficulties.

“The problem with these two categories is similar and different. In younger patients, it is difficult to access relevant tissues because organs from potential donors are more likely to be prioritized for organ transplantation than for research,” she explains. “With women, we generally have a 50% representation for tissue donors. However, for some diseases, there are differences in the proportion of men and women that are studied, as they affect proportionally more men than women. Increasing cohort sizes to better reflect variation in the overall population would be a major improvement. It could help us account for crucial factors like age, as premenopausal and postmenopausal women are known to have different risk factors. Similarly, including patients from different ancestries, and studying younger individuals from birth onward, could help us build a more complete reference atlas,” she concludes.

Newer technologies are enabling Noseda and her team to derive a fuller picture from the data available. Her group works with different tools to conduct their research, for example, it is involved in a project integrating single-cell and single-nuclei data with spatial transcriptomic data obtained in partnership with NanoString Technologies, a company creating solutions in spatial biology to visualize molecular interactions in a three-dimensional setting. It is also collaborating with Resolve Biosciences, which is working with multi-analyte and highly-multiplex spatial analysis to address different diseases.

Another spatial platform used by Noseda’s team is 10x Genomics’ Visium, one of the company’s key platforms. As Stern explains: “Visium enables discovery research by allowing whole-transcriptome analysis of whole tissue sections via a next-generation sequencing readout. As such, Visium provides an ideal tool for unbiased spatial discovery.” The tool is designed to work in tandem with other technologies produced by the company, like Xenium, developed for use after an initial assessment has been carried out with Visium.

“Xenium enables subcellular mapping of hundreds of RNA targets—directly in tissue, without sequencing. This generates high-quality data with high sensitivity and specificity, revealing new insights into cellular structure and function,” explains Stern.

Visualization

Julia Kennedy-Darling, Vice President of Innovation R&D at Akoya Biosciences, notes that one powerful application of spatial transcriptomics is its ability to map out reciprocal connections between cancer and immune cells. This could be used to better track the effectiveness of the immune system’s response by providing spatial data, instead of, or in addition to, relying mainly on the overall quantification of mRNA, proteins, and other key biomarkers.

According to Kennedy-Darling, the variety of high-dimensional spatial data also provides opportunities for improving patient stratification and, as a result, the potential for improved treatment. Constant progress along different but connected tracks will put such achievements within arm’s reach. “First, multiomics brought us the technical feasibility to detect both protein and RNA in the same sample. For example, RNA markers can help us detect signaling cascades and map the functional state of a tissue. Protein markers like PD-L1 and Lag3 can combine this with cell typing and key target detection, leading to a more complete picture of the tissue,” she explains.

“Second, new algorithms and data analysis methods are enabling correlations of non-spatial with spatial datasets, enriching the value of spatial information. Finally, there’s been an emerging recognition of the trade-off between ‘-plex’ and ‘throughput’ and how this pertains to patient cohort analysis. The ability to analyze large cohorts of patient samples spatially will likely involve discovering a signature or combination of targets that are meaningful in upstream discovery, and then applying this to large volumes of samples in a high-throughput manner,” she adds.

For Akoya Biosciences, visualization is the main focus of its technology. The integration of RNA technology with a protein-detecting high-plex assay enables the company to analyze patient cohort data via imaging entire sample slides, using its PhenoCycler-Fusion. “Our technology makes it possible to look at dozens of different patient samples, at once, including multiple samples from each patient. In this way, we can start to understand what information correlates with the biological question of interest. Once investigators find a signature of interest, a smaller subset of markers can be run on the Phenomalger HT Platform, a complementary high-throughput system that can routinely run six-plex whole slide analyses in minutes,” explains Kennedy-Darling.

Additional tools have been developed to simplify the workflow for novice users. “The proteomic and transcriptomic panels, branded as PhenoCode, that run on these systems can be used to answer key questions about the tumor microenvironment, which is especially useful for cancer immunotherapy research,” Kennedy-Darling adds.

Another way of visualizing spatial information is using oligonucleotides, as is offered by Twist Bioscience. The company can produce large numbers of oligonucleotides for different applications. “We can manufacture up to 1 million highly uniform oligos at a time. Our oligos have been used in multiplexed error-robust fluorescence in situ hybridization (MERFISH) [a technology that creates images for spatially resolved transcriptomics], to visualize gene expression using a combinatorial barcoding scheme. This technique requires highly accurate DNA synthesis. Our oligos have also been used in OligoPaint, which can help scientists visualize the genome,” says Emily Leproust, CEO and Co-founder of Twist Bioscience.

This approach has been enabled by the company's proprietary technology, which has miniaturized the chemical reaction required to synthesize DNA. This miniaturization has led to a substantial improvement in production volumes, for example Twist can manufacture one million oligos in a single run compared to approaches that use a 96-well plate to make 96 different oligos at once, Leproust states.

Democratizing the technology

An interesting question is whether small laboratories and CROs will soon be able to create large-scale spatial genomics processes in-house, scaling up and effectively democratizing spatial transcriptomic techniques. Kennedy-Darling outlines how this will depend on what will bring more value: a higher throughput of tissues, versus fewer tissues at a higher resolution. It could also depend on the ability of technology and procedures to improve speed, while at the same time ensuring simple workflows.

Stern underlines one major potential obstacle as being the need to handle, process, and analyze very large datasets, which could increase the need for multi-disciplinary collaboration between different institutions. Asking the right questions, bearing in mind the instruments available, is another important consideration. Noseda identifies two main hurdles from her perspective as being cost, as well as the ability to analyze data accurately and feasibly.

In conclusion, forming multi-disciplinary collaborations and partnerships to manage and analyze data could be a promising way to reduce expenses and the need for resources when applying spatial transcriptomics, including sourcing the right expertise to achieve meaningful results. Such solutions could ultimately help scale up the application of spatial genomics and transcriptomics across the industry and improve the feasibility of larger-scale programs.