ESMO Congress 2023 | OncologyPRO

Abstract 2278P

Background

The heterogeneity of the tumor microenvironment is a major barrier for effective cancer therapies. Similar to the immune tumor microenvironment the stromal environment, composed of cancer-associated fibroblasts (CAFs), pericytes, adipocytes and endothelial cells, is also heterogeneous. Multiple studies showed that CAFs contain heterogenous cell populations with opposing pro and anti-tumoral function. The heterogeneity of endothelial cells (ECs) is less well described largely due to the relative low frequencies in tumor tissues compare to CAFs. We set out to study the heterogeneity of the tumor microenvironment using truly single-cell spatial transcriptomic by the 10x Genomics Xenium platform.

Methods

We analyzed 20 human breast tumors by Xenium single-cell analyzes using fluorescent in situ sequencing (FISSEQ). We used the available breast panel that contains 280 genes, with 8 probes for each gene. The analysis of the data was preformed using the Seurat pipeline in R (version 4.9.9.9039).

Results

We could readily identify blood vessel structures that surrounded PECAM1/VWF positive cells. After multiple clustering we identified five types of endothelial cells with divergent gene expression profiles. We defined the following clusters of ECs: 1) Proliferating ECs, 2) CXCL12+ ECs, 3) POSTN^Hi/VWF^Hi ECs, 4) TPD52^Hi ECs and POSTN^Low/VWF^Hi ECs. The endothelial cells also showed spatial variations in distribution. Data on EC subtypes will correlation and other cell types and cell neighborhoods will be presented.

Conclusions

ECs in breast cancer tumors are highly heterogeneous and can be further classified into functionally divergent subtypes, potentially contributing to the heterogenous responses observed to treatments.