Gaining Content and Context with Spatial Biology

Spatial biology, whether its spatial genomics, proteomics, or transcriptomics, involves assigning a morphological context to cellular function and observed physiological changes. Biocompare recently hosted a Bench Tips webinar where scientists shared their experiences using various spatial biology techniques and best practices related to sample preparation, experimental design, and data analysis. Here are three key questions that they addressed during the webinar discussion.

Does spatial context really matter?

“Our work is at the intersection of tumorigenesis, inflammation, and embryogenesis and we identify the cells that are involved in oncofetal reprogramming,” explained Jennifer Currenti, Ph.D., a Postdoctoral Fellow in the Laboratory of Dr. Ankur Sharma at the Harry Perkins Institute of Medical Research and Curtin University. She is studying similarities between fetal and tumor microenvironments with respect to T cells in hepatocellular carcinoma using single-cell and spatial technologies. “Single-cell analysis does not tell you how these different cells are interacting with each other,” Currenti noted. “For that you need spatial biology.” Spatial techniques help her identify the location of cells involved and how they were interacting with each other. She is planning to use this spatial data to assess the impact of oncofetal reprogramming within the tumor ecosystem on immune checkpoint inhibitor outcomes, which could, in turn, help tailor the immunotherapy for patients.

Yered Pita-Juárez, Ph.D., is a Senior Postdoctoral Researcher in the Laboratory of Dr. Ioannis Vlachos in the Department of Pathology at Beth Israel Deaconess Medical Center (BIDMC), Harvard Medical School, and a Postdoctoral Research Scholar at the Broad Institute of MIT and Harvard. He is looking at the impact of SARS-CoV-2 infection in the liver taken from autopsy samples of Covid 19 patients. “We wanted to see if the SARS-CoV-2 virus was present in the liver and how Covid 19 impacted the liver,” he explained. Using viral load counts, in situ hybridization and single-nucleus RNA sequencing they showed that SARS-CoV-2 was indeed present in the liver. Then using probes specific for viral and human genes and spatial transcriptomics they were able to locate the regions in the liver samples where the virus was present.

Similarly, Jimmy Lee, Ph.D., a Postdoctoral Fellow in the laboratory of Dr. Omer Bayraktar in the Department of Cellular Genetics at the Wellcome Sanger Institute, is using spatial biology to study the pathophysiology of Covid 19 in post-mortem lung tissue. “We have used conventional histological and pathological knowledge along with spatial transcriptomics to map the extent of lung damage by Covid 19 in time and space and to identify targets for therapeutic intervention,” Lee reported. The spatial mapping of cell neighborhoods along the stages of disease progression showed that, even though the cell composition in the early and late stages of disease is mostly similar, the infiltration of new cells such as immune cells into the lungs, contributes to the morphology changes in the late stage of the disease.

Is it important to study the whole tissue or certain targeted regions?

It’s always a trade-off. Digital spatial profiling is a technique that can be used to study RNA or protein within specific regions of interest, using formalin-fixed paraffin-embedded (FFPE) or fresh, frozen tissue samples. “The regions of interest that need to be studied in greater detail can be selected based on their location or certain morphological features in the tissue,” said Currenti. “The regions selected can be of any geometrical shape, contoured, gridded, segmented, or cell-type specific.” The selected regions are typically up to 600 microns in diameter and the resolution tends to be high, which makes single-cell analysis possible.

On the other hand, whole tissue profiling can also be used to study RNA or protein in FFPE or fresh, frozen tissue but the resolution is lower. Currenti is working with a chip-based whole tissue section profiling technology called STOmics that is not yet commercially available but can be used for studying RNA in single cells from freshly frozen tissue. “The tissue section is placed on the chip that has spatially barcoded probes,” explained Currenti. “After mRNA capture and cDNA synthesis, the cDNA library is constructed, followed by sequencing.” The limitation is that the single-cell resolution reduces the number of genes that can be identified and currently this technology cannot be used in FFPE samples or for studying proteins.

Lee uses whole-genome tissue profiling to map the cellular landscape of lung pathogenesis. “Our collaborators had shown that different extents of alveoli damage could be found in different regions of the same lung section. Hence, we needed to maximize the flexibility to see different regions within the same sample, as opposed to only in the regions of interest.” To overcome the lack of single-cell resolution in their transcriptome profiling studies, the lab developed a principled Bayesian model called Cell2Location to identify the cell types in a certain region of interest using the single-cell lung atlas as reference. “The beauty of this model is that we were able to map the cell type that contributes to the transcriptome data in each region and also estimate the abundance of each cell type in the region to get a more quantitative analysis,” Lee said.

How best to deal with challenging samples?

“Spatial profiling is finicky and every small detail matters,” Currenti explained. The quality of the tissue matters, and particularly if the sample is precious or in small quantity it is advisable to start out with a pilot experiment to iron out all the kinks. Sample preparation at every step can be challenging, from getting the right tissue sections to library prep and sequencing.

“The liver can be a challenging tissue to work with, and we had to optimize the steps for mRNA capture and cDNA synthesis,” Currenti explained. Permeabilization time is critical to ensure that you have enough mRNA captured to be able see the various regions of interest and to minimize RNA diffusion out of the cell. “It’s a fine balance between maintaining the outline of the cell while seeing the structural elements inside. So, we had to find the correct microscope to help us determine the right permeabilization time.” Similarly, determining the reverse transcription (RT) time is also important. “For our experiment, the RT is done in an incubator and not in a thermocycler and it can vary by sample,” says Currenti. “So, we had to make sure that there was enough cDNA to make a library.”

Although the liver sample is challenging, according to Pita-Juárez, the liver is a prime candidate for spatial transcriptomics as it is a highly structured organ. Using the gene expression data and pathway scores they were able to recapitulate the structure of the liver looking at the regions of interest. “This allowed us to integrate the data from the single nucleus RNA sequencing and spatial transcriptomic analysis to better under the effects of Covid 19 in the liver.” Lee explained that the organized structure of the lung also helped him when combining single-cell and spatial genomics to understand cellular patterns in different regions of the lung. “We used computational tools to resolve and characterize the microenvironments.”

A lot of decisions in spatial biology experiments are driven by the type of sample and the biological question that needs to be addressed. However, a lot also depends on experience, planning, optimization, and the willingness to remain flexible and creative.