Geospatial genetic evolution and phenotypic plasticity in triple-negative breast cancer

The interplay between genotype and phenotype in shaping spatial heterogeneity and metastatic evolution in triple-negative breast cancer (TNBC) remains poorly understood.

Here, we developed a pathology-guided geospatial genomic and transcriptomic sequencing workflow (pGeo-G&T-seq) and applied it to TNBC specimens from primary tumors and matched metastatic sites. Using laser capture microdissection (LCM) we isolated mRNA and gDNA for whole-exome sequencing, whole-genome sequencing and whole-transcriptome sequencing to analyze spatial genomic heterogeneity and transcriptomic plasticity. Findings were further evaluated using independent bulk transcriptomic, single-cell immunotherapy, and spatial transcriptomic TNBC datasets.

The analysis unveiled spatial intratumoral heterogeneity at genomic and transcriptomic levels characterized by distinct spatial patterns of "decay" or "uniformity", demonstrating regional clonal expansion and genetic mosaicism, including heterogeneity in driver alterations and molecular subtypes. Notably, we identified a previously uncharacterized oxidative phosphorylation (OXPHOS)-associated transcriptional state associated with poor prognosis and features of limited immunotherapy responsiveness in single-cell cohorts. Intriguingly, genetic ancestry was associated with relatively stable transcriptional programs in a subset of cases. Furthermore, phylogenetically closely related clones from primary tumors and lymph node metastases shared convergent biological processes, suggesting that phenotypic traits may be partially retained during metastatic colonization.

These findings support a model in which genetic evolution and phenotypic retention contribute to the spatial heterogeneity and metastatic progression of TNBC. More broadly, our pathology-guided spatial multi-omics framework enables the dissection of genotype-phenotype coupling in intact tissue, while the identification of an OXPHOS-associated transcriptional program may inform biologically guided patient stratification and generate hypotheses for future therapeutic studies.