Researchers at the Kennedy Institute have developed a new computational framework that allows simultaneous analysis of gene expression and mechanical forces within cells and tissues, uncovering insights into how the interplay between transcriptional and mechanical signals guides processes such as cell fate decisions or the formation of spatially distinct tissue compartments.

Representation of the spatial mechano-transcriptomics methodology, which allows spatial profiling of gene expression and mechanical forces at single cell resolution.

Published in Nature Methods the new methodology integrates image-based spatial transcriptomics, which maps gene expression at single-cell resolution, with computational inference of the mechanical forces acting on individual cells across the the tissue. This allows the researchers to identify connections between the molecular programs controlling cell fate and the physical forces shaping tissues.

Lead researcher Adrien Hallou, Group Leader in Tissue Biology at the Kennedy Institute, said: ‘Spatial profiling technologies provide insights into how molecular programs are influenced by local signalling and environmental cues. However, cell fate specification and tissue patterning involve the interplay of biochemical and mechanical feedback. Therefore, to test for associations between them new methods were needed.

‘Our approach provides a window into the reciprocal relationship between gene regulation and mechanics in tissues, and by considering these factors together, we can gain a more holistic understanding of the mechanisms driving tissue morphogenesis during development, tissue maintenance at homeostasis or tissue architecture and function dysregulation in diseases.’

The research team developed a three-step computational pipeline applied to sections of the developing mouse embryo.

1. Data Processing and Image Analysis: Starting with images of tissue samples where cell membranes were marked for easy identification, this step involved segmenting the images using AI-based algorithms to pinpoint individual cells and measure their shapes very accurately.
2. Mechanical Force Inference: An algorithm based on a mathematical model of tissue mechanics and geometry was used to infer the mechanical forces acting on each individual  cell in the embryo section. By understanding the tensions at cell junctions and the internal pressures within cells, it is possible to quantify how these mechanical forces influence cell behaviour.
3. Statistical Analysis: The mechanical measurements were combined with gene expression data to identify relationships between the two at single-cell level. Advanced spatial modelling techniques were used to account for spatial variations, ensuring the findings were robust and reliable.

A spatial mechano-transcriptomics framework applied to a E8.5 mouse embryo seqFISH dataset

a, An overview of the spatial mechano-transcriptomics pipeline, showing immunostained membrane images and single-cell RNA-sequencing references as inputs, followed by deep learning segmentation, tension and pressure inference, gene expression imputation and final integrated mechanical-transcriptomic output maps. b, A schematic (top left) and images of the two different embryo sagittal sections considered in this study with close-ups of the three different brain regions studied in more details thereafter. Dataset 1: FMH and NC regions of embryo 1 brain; dataset 2: CM and FHM regions of embryo 2 brain; and dataset 3: MHB region of embryo 2 brain. 

The researchers focused their analysis on three key regions in the head of the developing E8.5 mouse embryo. In each region, they found that the boundaries between tissue compartments were characterised by elevated mechanical tension at cell-cell junctions, compared to the surrounding tissue compartments. This mechanical signature aligned with distinct spatial gene expression profiles, suggesting that both biochemical and physical cues contribute to the establishment and maintenance of tissue compartments.

They also showed that some genes always show expression patterns that correlate with the mechanical state of the cells, for example being up- or down-regulated when cellular pressure is high or low, or when cells are under mechanical tension or compression from their neighbours.

‘Our findings demonstrate how computational integration of spatial transcriptomics and mechanical profiling at single-cell resolution  can uncover important and novel relationships between gene regulation and physical forces, understanding how tissues form and function,’ said Adrien. ‘Our framework could be applied to different types of tissues and conditions, paving the way for a better understanding of the role of mechanical forces in a variety of pathologies, such as inflammatory diseases or cancer, and for new diagnostic and therapeutic strategies and advances in regenerative medicine.’

Source – Kennedy Institute of Rheumatology