Mapping the breast tumor microenvironment: proximity analysis reveals spatial relationships between macrophage subtypes and metastasis-initiating cancer cells

Abstract

Metastasis is responsible for the majority of cancer-related fatalities. We previously identified specific cancer cell populations responsible for metastatic events which are cytokeratin-14 (CK14) and E-cadherin positive in luminal tumors, and E-cadherin and vimentin positive in triple-negative tumors. Since cancer cells evolve within a complex ecosystem comprised of immune cells and stromal cells, we sought to decipher the spatial interactions of these aggressive cancer cell populations within the tumor microenvironment (TME). We used imaging mass cytometry to detect 36 proteins in tumor microarrays containing paired primary and metastatic lesions from luminal or triple-negative breast cancers (TNBC), resulting in a dataset of 1,477,337 annotated cells. Focusing on metastasis-initiating cell populations, we observed close proximity to specific fibroblast and macrophage subtypes, a relationship maintained between primary and metastatic tumors. Notably, high CK14 in luminal cancer cells and high vimentin in TNBC cells correlated with close proximity to specific macrophage subtypes (CD163^intCD206^intPDL1^intHLA-DR⁺ or PDL1^highARG1^high). Our in-depth spatial analysis demonstrates that metastasis-initiating cancer cells consistently colocalizes with distinct cell populations within the TME, suggesting a role for these cell-cell interactions in promoting metastasis.

Fig. 1. Imaging mass cytometry (IMC) for luminal breast cancers (Lum BC) and triple-negative breast cancers (TNBC).

A Biospecimens are incorporated into a tissue microarray to be stained together with a cocktail of metal-conjugated antibodies. Tissues are ablated to quantify ion content on a per-pixel (1 μm²) basis and rendered into images for single-cell segmentation and analysis. Created with BioRender.com. B Panel of antibodies used. Stars indicate the markers of metastasis-initiating cancer cells. C Representative IMC results: markers of ductal epithelial cells (CK8 and CK14), macrophages (CD68), and T cells (CD3) are shown for primary and metastatic Lum BC and TNBC. Matching primary and metastatic sites for four unique patients are shown. Bn brain, Lg lung, Lr liver, LUM BC luminal breast cancer, TNBC triple-negative breast cancer.

Fig. 2. Single-cell profiling with IMC reveals LUM and TNBC sub-clusters based on CK8 and CK14 expression and tumor microenvironment (TME) compositions.

A Results for FlowSOM clustering of IMC data are shown. Scaled expression profile of all markers for each cluster is reflected as heatmap. B Scaled expression profiles of selected markers for CK⁺ clusters for Lum BC and TNBC. C Differential plot of abundances favoring TNBC (negative blue values) or Lum BC (positive orange values) for all CK⁺ clusters. D Correlation heatmap of markers relevant to phenotyping CK⁺ clusters. E Stacked bar plots of abundances of each cluster (% of total annotated cells) faceted by organ site and cancer type from each image. B B cells, C collagen^+, CK cytokeratin, DC dendritic cells, E e-cadherin, GI bowel, Gran granulocytes, LG lung, LN lymph node, lo low expression, LV liver; LUM BC luminal breast cancer, Mac macrophage, NA not assigned, NK natural killer cells, OV ovary, PC pancreas, PL pleura, PS PDPN⁺SMA^+, Str stromal cells, SP spine, Tc cytotoxic T cells, ThEM effector memory helper T cells, ThN naive helper T cells, TNBC triple-negative breast cancer, Treg, regulatory T cells, V vimentin.

Fig. 3. Profiles of stromal and macrophage phenotypes within the TME.

A Stacked bar plots of cluster abundances (% of total annotated cells) of stromal clusters faceted by organ site and cancer type from each image for every case (SPC). Every stacked bar is sorted left-to-right for greatest-to-least abundances of stromal clusters. B Scaled expression profiles of select markers for stromal clusters for LUM and TNBC. C Correlation heatmap of markers relevant to phenotyping stromal clusters. D Stacked bar plots of abundances (% of total annotated cells) for macrophage clusters faceted by organ site and cancer type from each image for every case (SPC). E Scaled expression profiles of select markers for macrophage clusters. F Correlation heatmap of markers relevant to phenotyping macrophage clusters. BC breast cancer; GI bowel, LG lung LN lymph node, LUM BC, luminal breast cancer, LV liver, Mac macrophage OV ovary, PC pancreas, PL pleura, SP spine, SPC specimen, TNBC triple-negative breast cancer. V or VIM vimentin.

Fig. 4. Spatial relationships of CK⁺ and TME cell types.

Network visualization plots show distance relationships among clusters based on average shortest distances between cell types for TNBC primary (A), TNBC metastases (C), LUM primary (E) and LUM metastases (G). Node sizes reflect relative abundance of the cell type. Edge thickness positively correlates with shorter distances. Heatmaps of mean distances (μm) between CK+ and TME cell types within each image and then averaged across the entire dataset are shown for TNBC primary (B), TNBC metastases (D), LUM primary (F) and LUM metastases (H). B B cells, C collagen^+, CK cytokeratin, DC dendritic cells, E e-cadherin, Gran granulocytes, lo low expression, Lym lymphoid, LUM BC luminal breast cancer, Mac macrophage, Myl myeloid, NA not assigned, NK natural killer cells, PS PDPN⁺SMA^+, Str stromal cells, Tc cytotoxic T cells, ThEM effector memory helper T cells, ThN naive helper T cells TNBC triple-negative breast cancer, Treg regulatory T cells, V vimentin.

Fig. 5. CK+ cells in close proximity with specific macrophage clusters express higher levels of markers associated with metastasis-initiating cancer cells.

A 3D visualizations of CK⁺ cells in TNBC rendered into peaks for proximity to macrophages (z-axis and scaled color gradient) against their expression of VIM (x-axis) and ECAD (y-axis). B Similar 3D visualizations of CK⁺ cells in LUM for proximity to macrophages (z-axis) against their expression of CK14 (x-axis) and CK8 (y-axis). C Similar 3D visualizations of CK⁺ cells in LUM against CK14 (x-axis) and VIM (y-axis). D Heatmaps for fold-differences in the expression of (top) vimentin and (bottom) Ki-67 by CK⁺ cells that are “close” (<50 μm) over “far” (≥50 μm) with respect to each macrophage cluster. Expression of vimentin and Ki-67 were compared between close and far CK+ cells using Wilcoxon matched-pairs rank sum test. **p < 0.01, ***p < 0.005. ECAD e-cadherin, LUM BC luminal breast cancer; SPC specimen/case number, TNBC triple-negative breast cancer, VIM vimentin.