Spatial single cell analysis of clinical breast cancer brain metastases

Abstract

Background
Brain metastases are a common complication in cancer, particularly in patients with breast cancer, lung cancer, and melanoma, and are associated with poor prognosis. Understanding the spatial organization of tumor and immune cells is critical for uncovering the mechanisms of metastasis. Spatial transcriptomics technologies now allow gene expression profiling while preserving tissue architecture, providing new insights into tumor heterogeneity and microenvironmental interactions.

Aim
This study investigates tumor cell heterogeneity, spatial organization, and tumor–immune interactions in brain metastases by analyzing spatial transcriptomics data from breast cancer patients.

Methods
Spatial transcriptomics data from brain metastases of seven breast cancer patients were analyzed. Cell typing was performed using a combined reference profile derived from two NanoString-provided cell profiles representing immuno-oncology and brain-resident cell populations. Tumor heterogeneity was further investigated through subclustering and pathway enrichment analysis. To study the spatial organization, spatial niches were identified, and intercellular communication networks were inferred by performing ligand–receptor analysis.

Results
The spatial transcriptomics profiling of the brain metastases revealed substantial inter-patient heterogeneity, particularly within tumor cell populations, while non-malignant cells showed more conserved clustering patterns. Cell typing identified distinct functional subtypes within macrophages (e.g., CD163+ M2-like, SPP1+, ISG15+) and fibroblasts (e.g., myofibroblast-like ACTA2+, antigen-presenting CD74+), sometimes exhibiting patient-specific enrichment. Tumor cell clusters were largely patient-specific, displaying variable epithelial marker expression (PanCK) suggestive of epithelial-to-mesenchymal transition, and expressing key drivers like HER2 (matching clinical data), VEGFA, MMP7, and EGFR across different clusters. Spatial analysis defined recurring microenvironmental niches shared across patients, notably a stromal and a myeloid rich niche, often spatially adjacent. Spatially informed ligand-receptor modelling highlighted endothelial cells as major signal senders and astrocytes as notable receivers, particularly interacting with tumor cells via the MIF-(CD74/CD44) pathway and with extracellular matrix components via the CD44 receptor.

Conclusion
This study reveals substantial intratumoral heterogeneity within stromal and immune compartments of breast cancer brain metastases and identifies conserved spatial niche architectures dominated by either stromal or myeloid cells. These findings highlight the functional complexity of the metastatic microenvironment and support the therapeutic potential of targeting non-malignant cell populations—such as cancer-associated fibroblasts and macrophages—to disrupt tumor-supportive niches and improve treatment outcomes.

Popular Abstract

Tumor cells behavior and interactions in brain metastases

When cancer spreads to the brain, it leads to a serious and often life-threatening condition known as brain metastasis. Tumor cells can travel through the bloodstream after detaching from their primary site and may eventually reach the brain. Some of these cells manage to survive and adapt to this new environment—but what allows them to do so? And how do they interact with the healthy cells already residing in the brain?

These questions formed the basis of my thesis project. To investigate them, we studied samples of brain metastases originating from breast cancer. Our aim was to better understand how tumor cells behave once they reach the brain and how they interact with their new surroundings.

One important aspect of brain metastases is their complexity. Even within a single tumor, there can be a great deal of diversity among the cancer cells. Some may be more aggressive, others more resistant to treatment, and some may be found close to blood vessels or immune cells. This variation is also influenced by the presence and behavior of non-cancerous cells in the surrounding tissue, including which genes are active or inactive in specific cell types.

In this project, we examined the different types of cells within the tumor microenvironment and explored how these cells are arranged in space. We also looked at how cells might interact with one another—both tumor cells and the non-malignant cells around them. By studying these interactions, we hope to learn more about how cancer adapts to the brain and what role the surrounding environment plays in that process.

Understanding these dynamics could help researchers identify new strategies for treating brain metastases in the future. Exploring the supportive roles that non-cancerous cells might play, and how cancer cells communicate with them, is an important step toward developing therapies that can target not only the tumor itself but also the environment that enables its growth.

Master’s Degree Project in Bioinformatics 30 credits 2025
Department of Biology, Lund University

Advisor: Catharina Hagerling
Advisors Department: Lund University Cancer Center

• author: Fabbri, Alberto
• supervisor: Catharina Hagerling
• organization: Degree Projects in Bioinformatics
• course: BINP50 20242
• year: 2025
• type: H2 - Master's Degree (Two Years)
• subject: Biology and Life Sciences
• language: English
• id: 9212582
• date added to LUP: 2025-09-17 12:35:28
• date last changed: 2025-09-17 12:35:28