Journal: medRxiv
Article Title: Development and first-in-human CAR T therapy against the pathognomonic MiT-fusion driven protein GPNMB
doi: 10.1101/2025.02.26.24319604
Figure Lengend Snippet: a , Hematoxylin and eosin (H&E; left panel) of lung tumour biopsy sample taken 35 days post-GCAR1 treatment, with (remaining panels) IHC staining for the CAR target GPNMB; TFE3 to mark tumor cells; CD3 to identify all T cells; MYC to identify GCAR1 cells. Scale bars = 100 μm. b , Spatial plot of GPNMB gene expression across three samples profiled with Visium HD, including the ASPS-3 patient primary tumor and 2 research core biopsies. c , H&E plots of the spatially profiled samples in (b), with spatial plots of gene expression program usage values, and top 10 scoring genes per program (heatmaps). d , Barplots of the proportion of 24 μm bins per sample that have a minimum usage of each program (usage >0.05). Primary (grey); biopsy (green). Biopsy samples are merged together. e , Pearson correlation between spatial GPNMB expression in each sample type (rows) and spatial program usage (columns). f , Heatmap plot of the inverse rank of selected genes (rows) in each program (columns). Values indicate the relative importance of these genes to defining a given program. g , Pearson correlation of gene expression (rows) and program usage (columns) values in the biopsy samples. Checkpoint receptor genes with high correlation to Program 9 (CD8 T cells) are boxed in black; ligands are boxed in red when involving tumor or vascular programs. h , Dot plot of T cell program co-usage within T cell spatial niches in the biopsy samples. Niches are T cell positive bins stratified based on the additional presence of fibroblasts (niche 1), pericytes and endothelial cells (niche 2), or the absence of those stromal components (niche 3; i.e. tumor). Within each T cell niche, program co-usage is calculated for each of 15 programs, setting bins with usage to 1, otherwise 0. The proportion of 1 vs 0 bins are summarized for each niche. i , Schematic of the strategy for T cell-proximal gene expression analysis performed in panel (j). Each 24 μm bin positive for a T cell (based on expression of any T cell gene, including (CD3 components, CD4, CD8A, or GCAR) was considered to be within the region of interest, while the remaining bins were considered outside. The radius around each T cell was increased by 1, to encompass up to 3 nearest neighbours. j , Proportion of bins within and outside T cell regions in each sample (primary tumor, biopsy cores, xenograft), positive for expression of PD1 (top row) and PDL1 (bottom row). k, Heatmap of Student’s t-test (two-sided) p-values among pairs of regions (n=4 values per region) from panel (j).
Article Snippet: The Visium HD Spatial Gene Expression Kit (10x Genomics, human probe panel) was used to profile two ASPS biopsy samples, one region of the primary tumor, and a xenograft sample.
Techniques: Immunohistochemistry, Gene Expression, Expressing
