A computer approach for mapping the architecture of human tissues in unprecedented depth has been created by researchers. Their method promises to speed up research on organ-scale cellular interactions and might lead to strong new diagnostic tools for a variety of disorders.
The approach, which was published in Nature Methods on October 31st, sprang from the scientists’ discomfort with the gap between traditional microscopy and contemporary single-cell molecular research. “When you look at tissues under the microscope, you find a number of cells that are clustered together spatially — you see that organisation in pictures pretty quickly,” said lead author Junbum Kim, a Weill Cornell Medicine graduate student in physiology and biophysics.
“Now, cell biologists have the capacity to investigate individual cells in amazing detail, down to which genes each cell expresses, so they’re concentrating on the cells rather than the tissue structure,” he explained. “It’s critical for researchers to learn more about the details of tissue structure; fundamental changes in the relationships between cells within a tissue drive both healthy and diseased organ function,” said senior author Dr. Olivier Elemento, director of the Englander Institute for Precision Medicine and professor of physiology and biophysics, as well as computational genomics in computational biomedicine, at Weill Cornell Medicine.
However, manually merging single cell data with tissue structure maps is time-consuming and inefficient.
Machine learning algorithms have showed some promise for automating the procedure, but their capabilities are limited by the data needed to train them. To address this, Kim and his colleagues devised an unsupervised computational technique that defines structural areas inside a tissue by combining single-cell gene expression patterns with cell locations.
Dr. Andre Rendeiro, who was a postdoctoral fellow at Weill Cornell Medicine during the study and is now a primary investigator at the Austrian Academy of Sciences’ Research Center for Molecular Medicine in Vienna, Austria, compares the new technique to mapping a metropolis like New York: “One approach would be to go to every crossroads and count each type of building: is it residential, commercial, or a shop or restaurant?” Putting all of the data into one matrix and the positions of the buildings into another, one might combine the two matrices and seek for patterns.
“Essentially, we could begin to make broad statements about where different neighbourhoods are and where their borders are based on the abundance of, say, residential versus commercial buildings – just as anyone walking through the Upper East Side, Midtown, or Downtown would do based on their observations,” Dr. Rendeiro explained.
The researchers utilised the novel technology to create precise maps of several types of tissues, detecting and measuring new elements of microanatomy – the patterns that arise at a microscopic scale when cells interact and govern tissue function. They also proved that their approach could create delicate shades of differentiation between different disease states in collaboration with a colleague at the University of North Carolina in Chapel Hill who studies lung illness.
While cancer and other chronic diseases can produce significant changes in tissue structure, precise microanatomy may also aid in the diagnosis and treatment of more acute ailments. Rendeiro cites severe COVID-19 as an example, stating that “a lot of immune cells go into the neighbourhood, and there’s pretty significant alteration in the lung tissue.” The researchers are now using their innovative method to a variety of tissues to better understand how changes in tissue architecture underpin tissue function in health and malfunction in illness.