Assessing Immune Biomarkers of Response to Anthracyclines in Breast Cancer

Background: Pathologists have long recognized that the interaction between immune and tumour cells is critical in the development and progression of breast cancer. Studies have demonstrated the presence of tumour-infiltrating lymphocytes (TILs) correlates with improved clinical outcome in breast cancer especially in the triple negative and HER-2 positive subtypes. TILs predict for improved response to certain therapies including chemotherapy and trastuzumab. The predictive value of TILs in ER positive tumours is less clear. It has been demonstrated the higher presence of the immune microenvironment is associated with a better prognosis and as a result a higher likelihood of benefit from chemotherapy and possibly from immunotherapy, whereas cold immune microenvironment carries greater risk of relapse and lower benefit from chemotherapy and possibly immuno-therapies. In this study we examined the immune landscape in early breast cancer patients prior to anthracycline treatment.

Methods: NanoString’s DSP platform was used to analysis 256 patient samples from the BR9601 clinical trial. A panel of 56 antibodies incorporating the Immune Profile Core; 22 targets, IO Drug Target, Immune Activation Status, Immune Cell Typing, and Pan Tumour were examined and 2 fluorescent markers (pan-cytokeratin (pan-CK) and CD45) were used to select regions of interest (ROI).

Results: Proteomic analysis was successfully performed on 256 cores from the BR9601 cohort. Multivariate analysis demonstrated 27 proteins were associated with distant relapse free survival. A T-cell score was generated using the average expression of CD3, CD4 and CD8. The T-cell score was examined in both the tumour and TME. Using the T cell score it was apparent that the cohort had a range of immune “hot” and immune “cold” tumours. It was demonstrated that immune “hot” TME doesn’t not always correlate with immune “hot” tumour expression. Proteins that were most associated with T-cell exclusion (p<0.01) in the TME were Fibronectin, B7-H3, PTEN, ER-α, TGFB1, FAPα and CD34. This would indicate that these proteins are causal inhibitors of T-cell invasion.

Conclusion: In conclusion, this study highlights the significance of assessing the entire tumour since TILs, tumour and stromal cells collectively engage in a complex interplay that contributes to disease development and progression. NanoString’s GeoMx DSP is a promising technology for multiplexed analysis. TILs, whether measured using automated software, or estimation by protein profiling, are predictive of chemotherapy benefit.