3D spatial transcriptomics and ECM imaging in the tumor microenvironment

This Learning Wednesday paper note highlights “Combining spatial transcriptomics and ECM imaging in 3D for mapping cellular interactions in the tumor microenvironment.” The study is relevant to AI-enabled 3D spatial transcriptomics, focusing on how three-dimensional tissue context can alter what researchers observe and measure.

Selected notes from the paper

“Single-cell spatial transcriptomic (ST) maps receptor-ligand interactions but usually remains limited to 2D tissue sections and lacks ECM readouts. Here, we integrate 3D ST with ECM imaging in serial sections from one clinical lung carcinoma to systematically quantify molecular states, cell-cell interactions, and ECM remodeling in CN.” ​

“Unsupervised clustering of gene expression in segmented cells identified 18 cell types, annotated based on the expression of canonical marker genes.” ​

“To reconstruct 3D cellular neighborhoods, we leveraged STIM, which employs state-of-the-art computer vision techniques, to computationally align ST data and generate a 3D molecular map of the TME at single-cell resolution.” ​

“Unbiased clustering of 3D neighborhoods revealed 10 multicellular niches sharing a specific neighborhood composition.” ​

“As cells live and interact in 3D tissues, we hypothesized that analyzing 3D cellular neighborhoods would improve our ability to study multicellular niches.” ​

“By design, 3D neighborhoods included cells from the sections immediately above and below the z plane, comprising a 2.28-fold larger area than their 2D counterparts.” ​

“As expected, 2D neighborhoods featured a lower number of neighbors and lower cell type diversity than their 3D counterparts (median of 71 cells from 9 cell types/neighborhood in 3D vs. 32 cells from 7 cell types in 2D, p < 0.005).” ​

“2D niches failed to identify ‘dendritic cell niches,’ which were validated by IF staining.” ​

“3D neighborhoods enabled the systematic study of receptor-ligand interactions between physically proximal cells and revealed which ligand activities were spatially organized within multicellular niches.” ​

“We identified 3 ECM compartments using k-means clustering: one elastin-rich ‘homeostatic,’ one elastin and collagen poor ‘degraded,’ and one collagen rich ‘desmoplastic’ ECM compartment.” ​

“Fibroblasts displayed six transcriptomic states, spatially linked to specific ECM compartments and multicellular niches.” ​

“EMT is activated progressively from the tumor core to the desmoplastic stroma.” ​

“This proof-of-principle study highlights the potential of in-depth CN profiling in routine clinical samples to inform microenvironment-directed therapies.​"

From an Alpenglow perspective, this paper is useful because it connects AI-enabled 3D spatial transcriptomics with a broader need in 3D spatial biology, measuring tissue architecture across depth while preserving context for quantitative analysis.