Divergence of mutational signatures in association with breast cancer subtype

Abnormal molecular processes occurring throughout the genome leave distinct somatic mutational patterns termed mutational signatures. Exploring the associations between mutational signatures and clinicopathological features can unravel potential mechanisms driving tumorigenic processes. We analyzed whole genome sequencing (WGS) data of tumor and peripheral blood samples from 37 primary breast cancer (BC) patients receiving neoadjuvant chemotherapy. Comprehensive clinico-pathologic features were correlated with genomic profiles and mutational signatures. Somatic mutational landscapes were highly concordant with known BC data sets. Remarkably, we observed a divergence of dominant mutational signatures in association with BC subtype. Signature 5 was overrepresented in hormone receptor positive (HR+) patients, whereas triple-negative tumors mostly lacked Signature 5, but expectedly overrepresented Signature 3. We validated these findings in a large WGS data set of BC, demonstrating dominance of Signature 5 in HR+ patients, mostly in luminal A subtype. We further investigated the association between Signature 5 and gene expression signatures, and identified potential networks, likely related to estrogen regulation. Our results suggest that the yet elusive Signature 5 represents an alternative mechanism for mutation accumulation in HR+ BC, independent of the homologous recombination repair machinery related to Signature 3. This study provides theoretical basis for further elucidating the processes promoting hormonal breast carcinogenesis.