Predicting Breast Cancer Metastasis with Proteomics

ER+, HER2- breast cancer metastases remain a persistent clinical challenge, and the search for molecular tumor markers that can predict or detect metastases is currently underway. Recent research by Mosquim Junior and colleagues presents a comprehensive multiomic analysis that moves beyond genomics and transcriptomics to directly interrogate tumor biology at the protein and phosphoprotein level.

Study Design

This study is an observational, retrospective analysis conducted within the prospective Sweden Cancerome Analysis Network-Breast (SCAN-B) cohort and includes 182 ER+, HER2- primary breast tumors, comprising 120 no special type (NST) and 62 invasive lobular carcinoma (ILC) histological subtypes.

Tumors were stratified according to lymph node status and the occurrence of distant recurrence, with a clinical follow-up of at least five years. Survival outcomes assessed included recurrence-free survival (RFS) and distant recurrence-free survival (DRFS).

Uncovering Immune-Defined Tumor Subtypes

Prior to investigating metastatic progression, the authors validated the robustness of their multiomic dataset by analyzing well-established biological differences between ILC and NST tumors. Consistent with known ILC biology, loss of E-cadherin and adherens junction-associated proteins was robustly detected across transcriptomic, proteomic, and phosphoproteomic layers, which provided confidence in the quality of the data and the validity of subsequent analyses.

Unsupervised consensus clustering of the proteomic data revealed five molecular subtypes with distinct biology and outcomes. Cluster 2 stood out as “immune-hot” as it was enriched with interferon signaling, antigen presentation, and infiltrating CD4+ and CD8+ T cells and macrophages. Notably, this immune-enriched cluster trended toward worse RFS (hazard ratio [HR] 1.77, p = 0.09), suggesting that immune activation in ER+ disease may reflect immune exhaustion rather than effective tumor control.

In contrast, cluster 3 showed significantly improved RFS (HR 0.34, p = 0.04).

Markers Linked to Lymph Node Metastasis

To identify features associated with lymphatic spread, the authors applied multi-omics factor analysis (MOFA), which is a method that captures shared patterns across data types. Among hundreds of candidates, a single phosphorylation site on ES8L2 phosphopeptide emerged as independently prognostic. Lower abundance of this phosphosite was observed in LN+ tumors, and higher levels were associated with improved RFS (HR 0.61; multivariable p = 0.05). EPS8L2 participates in actin cytoskeleton remodeling and Ras signaling, processes central to cell motility and supporting a plausible role in limiting metastatic spread.

Signals of Distant Metastatic Risk

When focusing specifically on distant recurrence among LN- NST tumors, proteins belonging to the heat shock protein 90 (HSP90) family stood out. Elevated HSP90 abundance correlated with significantly worse DRFS (HR 2.10; multivariable p = 0.0058). HSP90 functions as a molecular chaperone stabilizing numerous oncogenic client proteins, and its association with aggressive behavior aligns with prior therapeutic interest in HSP90 inhibition.

Why This Matters

This study demonstrates that integrating proteomics and phosphoproteomics with transcriptomics can reveal metastatic risk signals invisible to gene expression alone. The identification of EPS8L2 phosphorylation and HSP90 abundance as markers of lymphatic and distant spread, respectively, highlights actionable biology already intersecting with drug development efforts.

Reference:
Mosquim Junior S, Zamore M, Vallon-Christersson J, Rydén L, Levander F. Multiomic profiling of ER-positive HER2-negative breast cancer reveals markers associated with metastatic spread. Breast Cancer Res. Published online January 12, 2026. doi:10.1186/s13058-025-02173-9