Spatial transcriptomics is revolutionizing our understanding of cellular dynamics and interactions in the tumor microenvironment (TME). By mapping distinct cellular populations and their interactions, we can gain valuable insights into tumor behavior. To investigate these dynamics in breast cancer, we conducted Xenium In-situ spatial transcriptomics on twenty formalin-fixed paraffin-embedded (FFPE) samples taken from eight patients undergoing neoadjuvant chemotherapy. These samples were taken pre-treatment, mid-treatment and post-treatment from four cases with pathological complete response (pCR), and four cases with residual disease (RD).
Our analyses found distinct cellular composition changes between pCR and RD cases. Patients with RD demonstrated increases in endothelial, perivascular-like (PVL) and cancer-associated fibroblasts (CAFs) from pre- to mid-treatment. This was accompanied by decreases in myeloid and B cells at mid-treatment. In contrast, patients with pCR demonstrated increases in CAFs, myeloid and B cells, with reductions in plasmablasts and cancer epithelial cells at mid-treatment. This suggests that an increase in myeloid and B cells observed between pre-treatment and mid-treatment samples may be predictive.
We also performed spatially constrained ligand-receptor (LR) analyses for 61 LR pairs. In RD cases, post-treatment samples exhibited an increase in signals from PVL and myeloid cells compared to mid-treatment. PCR cases exhibited more signaling from endothelial and B cells than RD cases at mid-treatment. Several interactions were observed to be specific to treatment response and timepoint. For example, interactions such as LGALS9/CD44 between CAFs and T cells, GDF15/TGFBR2 between cancer cells and CAFs, and DHH/PTCH1 between endothelial and CAFs, were only significant at mid-treatment of RD cases, suggesting a potential association with chemotherapy resistance.
Our findings indicate that the dynamic changes in cellular composition and interactions within the TME differs between responders and non-responders. To explore these differences, we plan on performing cellular niche analyses, which may identify biomarkers for informing novel therapeutic strategies.