Identification of new prognostic biomarkers for triple negative breast cancer

Abstract

Breast cancer is a complex, heterogeneous disease encompassing many different tumour entities.Management decisions are based on variables such as tumour size, histological grade, expression ofhormone (oestrogen and progesterone) receptors (ER/PR), and expression of human epidermalgrowth factor receptor-2 (HER2). Clinical outcomes are highly variable amongst patients with‘triple-negative’ breast cancer (TNBC; ER/PR/HER2-negative), but we currently cannot accuratelypredict prognosis in this group, thus inadequate disease control and over-treatment are challenges inthe clinic. Furthermore, new therapeutic options are needed for TNBC, as there are currently nomolecular-targeted therapies available.Published data suggest that expression of the SRY (sex determining region Y)-box 10 (SOX10)transcription factor is associated with TNBC, particularly those exhibiting molecular similarity tothe basal/myo-epithelium of the normal breast (‘basal-like’ breast cancer; BLBC). SOX10 belongsto the SOX family of transcription factors. It is an important regulator of early neural crestdevelopment and of melanocyte differentiation, in which expression is highest in melanocyteprogenitors and is silenced in mature melanocytes. SOX10 is expressed in 87-97% of melanomas,and is used as a diagnostic marker clinicallyThe aims of the current study were to:1. Investigate SOX10 mRNA levels and mechanisms controlling its expression in breast cancer;2. Investigate the prognostic significance of SOX10 in breast cancer;3. Evaluate SOX10 protein expression in normal and tumour-adjacent normal breast;4. Use gene expression profiling and in vitro and in vivo functional assays to investigate thepossible functional roles of SOX10 in breast cancer; and5. Investigate the prognostic significance of two cancer-testis antigens (MAGE-A and NY-ESO-1) in TNBC.Analysis of published breast cancer genomic datasets showed that SOX10 mRNA expression wasinversely correlated with gene methylation (r2 0.567; p<00001), suggesting epigenetic regulation isimportant in breast cancer. A subgroup of BLBCs (~65%) demonstrated high expression andsignificantly lower SOX10 methylation than non-BLBCs (χ2 p<0.0001). SOX10 copy-number gainswere also more frequent in BLBCs (χ2p=0.0002). SOX10 mRNA expression was not associatedwith the clinical outcomes tested, including response to chemotherapy or overall survival, thoughimmunohistochemical (IHC) analysis of breast cancer tissue microarrays showed that SOX10protein expression was strongly associated with markers of poor outcome (e.g. histological grade, p<0.0001) and predicted poor survival (hazard ratio (HR): 3.66; 95% CI: 1.78-7.51; p<0.0001).Critically, SOX10 protein expression also predicted poor survival in basal-like TNBCs (HR 3.20;95% CI: 1.28-8.03; p=0.0168; n=52).Little is known about the roles of SOX10 in either normal or neoplastic breast tissue. To investigatethe roles of SOX10 in the human mammary epithelium, we analysed its expression in situ inreduction mammoplasty specimens (n=18) and tumour-associated normal breast tissue (n=19).SOX10 was consistently expressed in the myoepithelium and heterogeneously expressed in luminalcells. In luminal epithelia, there was a strong correlation between expression of SOX10 and theluminal progenitor marker c-kit, and an inverse association with ER and cytokeratin 8/18. In silicoanalysis of published gene expression data from mammary epithelial compartments showed SOX10was most highly expressed in luminal progenitor cells, and the least in mature luminal cells (χ2p<0.0029). In light of its role in differentiation in other tissues (e.g. melanocytes), these datasuggest that SOX10 may regulate lineage differentiation in the mammary gland.In order to investigate possible roles of SOX10 in breast cancer, we employed in vitro and in vivobreast cancer models. In-house screening experiments demonstrated infrequent expression ofSOX10 in breast cancer cell lines. Only 2 of 39 basal-like lines (~5%) had detectable levels ofSOX10 mRNA. Short hairpin RNA (shRNA)-mediated knockdown of SOX10 in basal-like MDAMB-435breast cancer cells did not alter their proliferation rate, migratory behaviour or sphereforming capacity in vitro, but did reduce colony formation on plastic (p=0.028), consistent with theidea that it regulates progenitor activity. In vivo tumourigenicity experiments showed that SOX10knockdown reduced tumour formation in the mouse mammary fat pad. Differential gene expressionmicroarray analysis of these cells coupled with gene ontology analysis of SOX10-high and –lowBLBCs showed a significant association between SOX10 and immune response genes.In light of its roles in coordinating differentiation in the neural crest and the melanocyte lineage, wehypothesise that SOX10 is involved in regulating mammary epithelial lineage differentiation andthat its expression is inappropriately activated and/or maintained in sporadic basal-like TNBCs, andthat the involvement of SOX10 contributes to poor outcome. Future studies are required to fullyunderstand the functional role of SOX10 in breast cancer.