Breast cancer represents a diverse group of diseases, which can be broadly categorized into five main subtypes based on the gene expression patterns of tumor cells. The pioneering work by Perou and Sørlie and colleagues^1,2 has demonstrated the association of these subtypes with prognosis and treatment responses. The stratification of breast cancer into molecular subtypes has been instrumental in guiding therapeutic decisions and enhancing our understanding of tumor biology.

While the study of gene expression has been vital, an equally important aspect is the investigation of proteins, the functional products of genes and central to cellular structure, function, and regulation. Analyzing the global protein expression levels, commonly referred to as ‘expression proteomics’, is crucial for understanding the cellular mechanisms at play. Since proteins are involved in complex interaction networks, they provide a more dynamic and functional perspective compared to mRNA analysis alone.

Beyond the cancer cells themselves, the tumor microenvironment is a key component in the progression of breast cancer. The tumor microenvironment comprises various cellular and non-cellular elements surrounding the tumor, such as vascular cells, fibroblasts, immune cells, extracellular matrix, and signaling molecules, which collectively can either impede or promote tumor progression³.

Hypoxia is an important driver of cancer progression⁴. Rapidly growing tumors often develop hypoxic regions as they outgrow their blood supply which pushes the tumor to adapt by altering its metabolism, stimulating the formation of new blood vessels, and even helping cancer cells to break free from the primary tumor site.

Our approaches have been to 1) Explore the breast cancer cell secretome to study messages to the tumor microenvironment, and how this changes in a hypoxic environment, and 2) Characterize the global proteome of the tumor microenvironment in low-grade (luminal-like) and high-grade (basal-like) breast cancers. Combining these two approaches may provide insights into how the tumor cells and the microenvironment interact to promote breast cancer progression. This insight may unlock potential new strategies for treatment.

Luminal-like and basal-like breast cancer cells respond differently to hypoxia.

Extensive work has been performed in the field of hypoxia, especially related to the hypoxia-inducible factors (HIFs). However, the hypoxia response is not well characterized at the global proteome level. In our study, we have investigated the secreted proteins from breast cancer cell lines by mass spectrometry-based proteomics analysis to explore how the cancer cells can affect the microenvironment under hypoxic conditions. We found that luminal-like and basal-like breast cancer cells respond differently to a hypoxic environment. The luminal-like cells underwent metabolic adaptations and increased the secretion of pro-angiogenic proteins. In contrast, the basal-like cells increased the secretion of proteins that are associated with developmental processes and inflammation. Interestingly, some of the features that were increased in the luminal-like hypoxic secretomes were already higher in the basal-like cells prior to hypoxia exposure. This included vascular endothelial growth factor A (VEGFA), a key player in angiogenesis. We were surprised by the large differences between the two breast cancer subtypes (luminal-like and basal-like), and while this needs additional follow-up studies for verification, our data indicates that the basal-like cells are in an activated angiogenic-like state at baseline.

An integrated approach identifies a tumor microenvironment hypoxia signature.

The tumor microenvironment has received less attention compared to the cancerous epithelial cells. While the cells of the microenvironment themselves are not cancerous, they can be influenced by factors secreted by nearby cancer cells to become pro-tumorigenic. Proteins that are secreted and released from tumor cells are part of the tumor microenvironment. To study the expression of secreted proteins, and other proteins, in this compartment, we separated the tumor microenvironment and tumor epithelium by laser-capture microdissection and analyzed the extracted proteins by mass spectrometry-based proteomics. Next, we extracted the proteins that were upregulated in response to hypoxia in the cell line experiments and the proteins that were exclusively differentially abundant (between basal-like and luminal-like tumors) only in the tumor microenvironment (not tumor epithelium). This resulted in 33 proteins, which we considered a microenvironment-based hypoxia signature.

The 33-protein (33P) signature was correlated with signatures of aggressive cancer features, including hypoxia, angiogenesis, epithelial-to-mesenchymal transition (EMT), and stemness, in the retrospective observational METABRIC-Discovery cohort that contains clinical outcome data. We found that a high 33P score (by upper quartile) was associated with large tumor size, high histologic grade, lymph node metastases, ER negative tumors, and a basal-like phenotype. Notably, a high 33P score was independently associated with decreased breast cancer specific survival when adjusting for the basic prognostic factors tumor diameter, histologic grade and lymph node status, in both the luminal-like and basal-like breast cancer subtypes.

To explore the potential interaction between 33P and various treatments, we applied the METABRIC-Discovery cohort with information on endocrine treatment, chemotherapy, and radiation therapy. Interestingly, in the group that received radiotherapy, we found different survival patterns for the patients that had a high 33P-score compared to the patients that had a low score. In the group that did not receive radiotherapy, there was no difference in survival. Of note, these analyses were performed on retrospective observational data; a randomized clinical trial is needed to confirm these observations.

In conclusion, in this study we have combined in vitro hypoxia secretome experiments with compartment specific breast cancer tissue proteomics. We identified a 33-protein microenvironment-based hypoxia signature that can identify patients with ‘low-grade’ breast cancer and assumed good prognosis, that may have a ‘high grade’ microenvironment and poorer prognosis than reflected by the epithelial cell-based breast cancer subtypes. The importance of 33P in relation to radiation therapy should be examined in follow-up studies. Moreover, the 33-protein signature may be a potential marker for deciding patients that benefit from radiotherapy.

References

1. Perou, C. M. et al. Molecular portraits of human breast tumours. Nature 406, 747-752 (2000). https://doi.org:10.1038/35021093
2. Sorlie, T. et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 98, 10869-10874 (2001). https://doi.org:10.1073/pnas.191367098
3. Hanahan, D. & Coussens, L. M. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell 21, 309-322 (2012). https://doi.org:10.1016/j.ccr.2012.02.022
4. Wicks, E. E. & Semenza, G. L. Hypoxia-inducible factors: cancer progression and clinical translation. J Clin Invest 132 (2022). https://doi.org:10.1172/JCI159839