Immune subtyping of ER-positive, PR-negative and HER2-negative breast cancers by immunohistochemistry: Molecular basis and clinical relevance.

Abstract

e12553 Background: The diversity and plasticity behind ER+/PR−/HER2− breast cancer (BC) remain underexplored. In our previous work, an immune subtype classification for ER+/PR-/HER2- BC with RNA-seq data was established to distinguish between immune-hot and -cold tumors to provide guidance for immune checkpoint inhibitor (ICI) therapy. However, considering the economic costs, the clinical application of sequencing methods is limited. This study aimed to identify biomarkers of the immune subtype classification and develop a diagnostic procedure for ER+/PR-/HER2- BC immune subtyping based on the immunohistochemistry (IHC) technique, thereby identifying candidates of ICI therapy and refining treatment strategies for ER+/PR-/HER2- BC. Methods: RNA-seq data for 411 patients diagnosed with ER+/PR-/HER2- BC were obtained from the METABRIC database (METABRIC cohort). The immune subtype classification was established by consensus clustering with 1480 immune-related genes. Biomarkers for each immune subtype were identified from differentially expressed genes and validated using logistic regression and ROC curve analysis. Validation of the immune classification was conducted using RNA-seq data from the GEO and TCGA databases. Transcriptomic gene chip sequencing was performed on surgical specimens from 15 ER+/PR-/HER2- BC patients in our center (Chip cohort) to classify the immune subtypes. Biomarkers identified were validated and a diagnostic procedure was designed using IHC techniques. Results: In the METABRIC cohort, immune subtypes including Clusters 1-5 were established. Five biomarkers including DCN, SPP1, CXCL9, STC2, and SRGAP1 were identified, with the AUC values of ROC curves being 0.90, 0.82, 0.99, 0.70, and 0.89, respectively. Notably, Cluster 3 exhibited higher infiltration of CD8+ T cells, B cells, and M1 macrophages, with a greater association with immune activation pathways, thus being defined as the immune-hot subtype. Cluster 1, also showing a moderate immune cell infiltration, was characterized by a more pronounced stromal infiltration. In the Chip cohort, IHC revealed that Cluster 3 exhibited the highest levels of CD8+ T cell and B cell infiltration as well as PD-L1 expression, suggesting a potential benefit from ICI therapy. In contrast, Cluster 1 demonstrated the most extensive infiltration of endothelial and fibroblast cells. The other subtypes were characterized by a suppressed immune microenvironment, especially in Cluster 2, where an infiltration of SPP1+ macrophages was observed. A diagnostic procedure was designed with the biomarkers using IHC techniques. Conclusions: This study validated the immune subtype classification for ER+/PR-/HER2- BC. An easy-to-use IHC-based diagnostic procedure was developed, facilitating the identification of potential ICI therapy candidates in clinical practice at a lower cost.