This study reports a comprehensive single-cell spatial transcriptomic atlas of metastatic breast cancer. The atlas comprises 126 tumor specimens from 44 patients and captures transcriptional profiles at single-cell resolution across primary and metastatic lesions. In total, the resource contains 520,850 cells, including 171,766 annotated cancer cells and 349,084 cells from the surrounding tumor microenvironment. The authors frame the work around the clinical challenge that metastatic breast cancer remains a leading cause of cancer mortality in women, with limited rationale-based therapeutic options and frequently poor outcomes.
The investigators used a cost-effective spatial transcriptomics approach capable of resolving transcriptomes at single-cell level across many samples. The resulting atlas integrates cellular identity and spatial context to study intra-tumor heterogeneity and tumor–stromal interactions. The data span both tumor and microenvironmental compartments, facilitating comparisons of cancer cell states, immune cell populations, and stromal features within their spatial niches.
For most patients in the cohort, multiple time points were collected during the course of metastatic disease. This longitudinal sampling enabled the authors to detect treatment-induced transcriptional changes specifically within the tumor cell compartment. Those changes were linked to the emergence of therapy resistance, allowing the atlas to capture dynamic shifts in cancer cell programs that occur under systemic therapy across the course of disease progression.
Analysis of the tumor microenvironment revealed that distinct immune cell subpopulations became enriched at progression following particular therapeutic regimens. The authors report that these therapy-associated immune changes were predictive of hormone receptor loss, indicating a relationship between immune contexture at progression and alterations in tumor receptor status. The atlas therefore connects specific immune shifts to clinically relevant phenotypic changes in tumor cells.
Using gene expression-derived signals, the study inferred DNA copy number alterations (CNAs) from the single-cell transcriptomic data to reconstruct genomic features at single-cell resolution. By coupling inferred CNAs to longitudinal sampling, the authors tracked subclonal tumor cell selection during progression. This approach allowed identification of clonal dynamics, including the selection and expansion of subclones under therapeutic pressure across time.
Combining spatial single-cell transcriptomes with inferred CNAs and longitudinal data, the authors identified transcriptomic features that characterize subclones that outcompete others and display future metastatic potential. These transcriptomic signatures reflect the properties of clones that expand during progression and may underlie metastatic fitness. The atlas therefore provides candidate molecular correlates of clonal advantage in a treated metastatic setting.
All raw and processed single-cell spatial transcriptomics and spectral flow cytometry data are deposited at the BioImage Archive (accession S-BIAD2706) and are accessible at the Single Cell Portal (accession SCP3851). The analysis code used in the study is publicly available on GitHub at the repository listed by the authors. Ethical oversight was provided by the local medical ethics committee of the Netherlands Cancer Institute and the Portuguese Oncology Institute of Porto (IRB reference CFMPB411), and all patients gave informed consent for translational research. Two authors are employees of Xilis B.V., which owns the MOS technology; other authors declared no competing interests.
The resource provides a spatial, single-cell perspective on metastatic breast cancer development and progression across a heterogeneous systemic treatment landscape. By integrating tumor cell states, immune composition, inferred genomic alterations, and longitudinal sampling, the atlas is positioned to support hypothesis generation around mechanisms of therapy resistance, immune involvement in progression, and clonal selection that enables future metastasis. The study emphasizes the complexity of metastatic disease and supplies a large, publicly accessible dataset and analytic code to facilitate further research.
(Links and accession numbers reported by the authors: BioImage Archive S-BIAD2706; Single Cell Portal SCP3851; code repository on GitHub. IRB reference: CFMPB411.)