Dataset for: PROGNOSTIC AND FUNCTIONAL ROLE OF SUBTYPE-SPECIFIC TUMOR-STROMA INTERACTION IN BREAST CANCER

None of the clinically relevant gene expression signatures available for breast cancer were specifically developed to capture the influence of the microenvironment on tumor cells. Here, we attempted to build subtype-specific signatures derived from an in vitro model reproducing tumor cell modifications after interaction with activated or normal stromal cells. Gene expression signatures derived from HER2+, luminal and basal breast cancer cell lines (treated by normal fibroblasts or cancer-associated fibroblasts conditioned media) were evaluated in clinical tumors by in silico analysis on published gene expression profiles (GEPs). Patients were classified as microenvironment-positive (µENV+ve), i.e with tumors showing molecular profiles suggesting activation by the stroma, or negative (µENV-ve) based on correlation of their tumors’ GEP with the respective subtype-specific signature. Patients with estrogen receptor alpha (ER)+/HER2-/µENV+ve tumors were characterized by 2.5-fold higher risk of developing distant metastases (HR= 2.546; 95% Cl: 1.751-3.701, P=9.84E-07), while µENV status did not affect, or only suggested the risk of distant metastases, in women with HER2+ (HR=1.541; 95% Cl: 0.788-3.012, P=0.206) or ER-/HER2- tumors (HR=1.894; 95% Cl: 0.938-3.824; P=0.0747), respectively. In ER+/HER2- tumors, the µENV status remained significantly associated with metastatic progression (HR=2.098; CI: 1.214-3.624; P=0.00791) in multivariable analysis including size, age and genomic grade index. Validity of our in vitro model was also supported by in vitro biological endpoints such as cell growth (MTT-assay) and migration/invasion (Transwell-assay). In vitro-derived gene signatures tracing the bidirectional interaction with cancer activated fibroblasts is subtype-specific and adds independent prognostic information to classical prognostic variables in women with ER+/HER2- tumors.

目前已有的乳腺癌临床相关基因表达特征（gene expression signatures），均未专门用于捕捉肿瘤微环境对肿瘤细胞的影响。本研究旨在构建源自体外模型（in vitro model）的亚型特异性基因表达特征，该模型可复现肿瘤细胞与活化或正常基质细胞相互作用后的基因表达改变。我们通过生物信息学分析（in silico analysis），对已发表的基因表达谱（gene expression profiles, GEPs）进行分析，在临床肿瘤样本中评估了源自HER2阳性、管腔型及基底型乳腺癌细胞系（经正常成纤维细胞或癌症相关成纤维细胞条件培养基处理）的基因表达特征。根据肿瘤基因表达谱与对应亚型特异性特征的相关性，将患者分为微环境阳性（µENV+ve）组与微环境阴性（µENV-ve）组，其中微环境阳性组指肿瘤分子谱提示存在基质活化的患者。雌激素受体α（estrogen receptor alpha, ER）阳性/HER2阴性/µENV阳性的肿瘤患者，发生远处转移的风险升高2.5倍（风险比（hazard ratio, HR）=2.546；95%置信区间（CI）：1.751~3.701，P=9.84×10^-7）；而µENV状态对HER2阳性患者（风险比（HR）=1.541；95%置信区间（CI）：0.788~3.012，P=0.206）或三阴性乳腺癌（ER-/HER2-）患者（风险比（HR）=1.894；95%置信区间（CI）：0.938~3.824，P=0.0747）的远处转移风险无显著影响，仅提示潜在风险。在ER阳性/HER2阴性的肿瘤患者中，在校正肿瘤大小、年龄及基因组分级指数的多变量分析中，µENV状态仍与转移进展显著相关（风险比（HR）=2.098；95%置信区间（CI）：1.214~3.624；P=0.00791）。本体外模型的有效性还得到了体外生物学终点的支持，包括细胞增殖实验（MTT-assay）及迁移/侵袭实验（Transwell-assay）。本研究中源自体外模型的基因表达特征，可特异性反映肿瘤细胞与癌症活化成纤维细胞的双向相互作用，且能为ER阳性/HER2阴性乳腺癌患者的经典预后变量提供独立的预后信息。