Exploring the Key Molecular drivers of lymphovascular invasion in Invasive Breast Cancer

Abstract

ABSTRACT BACKGROUND: Lymphovascular invasion (LVI) plays a crucial role in breast cancer (BC) metastasis and presents as a reliable indicator of poor outcome in terms of distant metastasis and shorter survival. Therefore, accurate and reproducible assessment of LVI and an understanding of its molecular mechanisms can help guide management decision-making. However, molecular mechanisms of LVI are complex and overlap with other biological processes, such as metastasis and BC progression in general. Therefore, deciphering the molecular mechanisms of LVI and identifying its driver genes remain challenging. This study aimed to utilise several well-defined BC datasets, evaluating whether the LVI mechanisms in BC are controlled by distinct transcriptome and proteomic profiles, and to assess the roles of potential target genes using various molecular techniques. METHODS: In an attempt to discover new targets associated with LVI status, bioinformatics analysis was applied on multiple large cohorts of patients with BC, including the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) cohort (n = 1,980) and The Cancer Genome Atlas Breast Carcinoma (TCGA-BRCA; n = 854) as discovery datasets and the Breast Cancer Gene-Expression Miner (bc-GenExMiner) dataset (Affymetrix; n = 4,904) cohorts as a validation dataset. The differential mRNA expression was performed to identify key genes with respect to LVI status, which was previously defined based on morphology and immunohistochemistry (IHC) for endothelial markers. In addition, mass spectrometry of a well-defined subset of BC with well-characterised LVI status was utilised to identify target genes at the protein level. The targeted genes that were strongly associated with LVI at the transcriptomic levels (HSP90AA1, SLC9A3R1, NOP56 and HMGB3) were further investigated to confirm their prognostic value and association with LVI at the protein level utilising IHC and tissue microarrays of a large annotated early stage BC with long-term follow-up Nottingham cohort (n = 2,497). The protein expression of (HSP90AA1, SLC9A3R1 and NOP56) was correlated with clinicopathological parameters, a large panel of relevant biomarkers, and patient outcomes to identify clinically reliable LVI-associated genes. Further functional experiments followed this to explore the mechanistic insights of the most likely target gene (HMGB3) for LVI in BC to identify novel BC prognostic and predictive value towards personalised therapy in patients with early stage BC.