Spatial Proteomics of the Normal Breast Collagen Stroma: Links to Density and Body Mass Index

Collagen breast stroma can become a breast cancer risk factor, yet proteomic regulation of normal breast stroma remains poorly defined. This study evaluates the spatial regulation of the collagen proteome from normal breast tissue. Normal breast tissue sections from the Susan G. Komen tissue bank were used (n = 40), with data including genetic ancestry (n = 20 African ancestry; n = 20 European ancestry), body-mass-index (BMI), age, and mammogram density by the Breast Imaging Reporting and Data System (BI-RADS). 10-plex cell marker staining showed CD44 and COL1A1 markers modulated with BMI. Collagen fiber widths by second harmonic generation microscopy contrasted in BMI categories by genetic ancestry. Targeted extracellular matrix proteomics mass spectrometry imaging showed the collagen alpha-1(I) chain proteome was spatially heterogeneous across the normal breast microenvironment with site-specific post-translational modification of proline hydroxylation. Signatures computationally extracted from stroma-rich regions reported that 47 collagen peptides distinguished BI-RADS categories (area under the receiver operating curve >0.7; p-value >0.05). Multivariate modeling of collagen peptides, fiber metrics, and clinical features supported a strong positive association with BMI as a determinant of collagen alterations in the normal breast. This study provides a foundation for larger studies investigating the clinical value of spatial collagen proteome alterations in human breast.乳腺胶原间质（Collagen breast stroma）可成为乳腺癌风险因子，但目前针对正常乳腺间质的蛋白质组调控机制仍未得到充分阐明。本研究针对正常乳腺组织的胶原蛋白质组空间调控特征展开评估。研究使用了苏珊·G·科门组织库（Susan G. Komen tissue bank）提供的正常乳腺组织样本（n=40），配套数据涵盖遗传祖先信息（20例为非洲血统、20例为欧洲血统）、体重指数（BMI）、年龄，以及基于乳腺影像报告与数据系统（BI-RADS）分级的乳腺钼靶密度。10重细胞标志物染色结果显示，CD44与COL1A1的表达水平随BMI发生调控变化。通过二次谐波生成显微镜（second harmonic generation microscopy）测得的胶原纤维宽度，在不同遗传祖先群体中随BMI类别呈现出显著差异。靶向细胞外基质蛋白质组学质谱成像分析显示，胶原α1(I)链（collagen alpha-1(I) chain）蛋白质组在正常乳腺微环境中呈现空间异质性，且脯氨酸羟化（proline hydroxylation）存在位点特异性的翻译后修饰。通过计算从富间质区域提取的分子特征显示，47种胶原肽可有效区分BI-RADS分级（受试者工作特征曲线下面积>0.7；p值>0.05）。针对胶原肽、纤维形态指标及临床特征的多变量建模结果证实，BMI作为正常乳腺胶原改变的关键影响因子，与其存在显著正相关关联。本研究为后续开展更大规模的、探究人体乳腺中空间胶原蛋白质组改变临床价值的相关研究奠定了基础。