PO.TB02.02 · 肿瘤生物学

单细胞和空间分析识别原发性乳腺血管肉瘤的肿瘤特征和可干预的表面靶点

Single-cell and spatial profiling identifies tumor features and actionable surface targets in primary breast angiosarcoma

海报缩略图:单细胞和空间分析识别原发性乳腺血管肉瘤的肿瘤特征和可干预的表面靶点
编号 711 展板 1 时间 4/19 02:00–05:00 区域 Section 29 主讲 Asumi Iesato, MD;PhD
分会场 Molecular Pathology
查看 PDF 下载 PDF 🔒 查看 / 下载完整 PDF 需登录并开通下载套餐 · 查看套餐 / 开通 AACR 官方页面

作者与单位 Authors & Affiliations

Asumi Iesato, Masashi Akiya, Kazutaka Otsuji, Sumito Saeki, Natsue Uehiro, Tetsuo Noda, Tomo Osako, Takayuki Ueno, Reo Maruyama

Japanese Foundation for Cancer Research, Tokyo, Japan

摘要 Abstract

中文摘要
背景:原发性乳腺血管肉瘤(PBA)是一种极为罕见且侵袭性强的血管恶性肿瘤,治疗选择有限。基因组研究已识别出复发性改变,如KDR、PIK3CA、TP53和FLT4,但基因表达模式鲜有报道,其空间背景也仍不明确。目前仅发表了两项空间转录组学研究,均来自单一机构,且主要聚焦于免疫细胞分布,同时将乳腺和非乳腺血管肉瘤合并分析。这两项研究基于内皮标志物(ERG/CD31)识别肿瘤区域;然而,这些标志物在去分化血管肉瘤中可能减弱,可靠的肿瘤特异性标志物仍不可得。PBA是否与其他解剖部位的血管肉瘤共有分子特征也仍不清楚。目的:描绘原发性和转移性PBA的转录程序,并使用整合空间分析对其进行验证,同时识别具有治疗适用性的候选肿瘤细胞标志物。 方法:我们整合了来自配对PBA标本的单细胞RNA测序(scRNA-seq)和空间转录组学。使用BD Rhapsody scRNA-seq分析新鲜肿瘤(361T、256T),并与正常乳腺组织(388D)整合。通过InferCNV推断CNV。对两例的FFPE样本进行Visium空间转录组学,其中一例(256T)有卵巢和肝脏转移灶可用。在3例PBA病例中通过免疫组化评估CD276表达。 结果:来自三个样本的整合scRNA-seq识别出三个内皮细胞簇:c5(正常内皮细胞)、c7(肿瘤内皮细胞)和c11(淋巴管内皮细胞)。簇7表达PECAM1/CDH5以及血管生成/淋巴管生成调控因子,InferCNV确认了其肿瘤身份。c7与c5之间的差异表达产生了一个肿瘤特异性基因集(AS_signature),富集于细胞外基质组织、受体酪氨酸激酶信号传导和基底膜重塑。空间InferCNV在Visium肿瘤区域中再现了肿瘤相关的CNV。对256T卵巢和肝脏转移灶的联合分析显示完全的UMAP分离和不同的CNV结构;原发肿瘤亚簇优先定位于卵巢或肝脏病灶,提示器官选择性的亚克隆播散。识别出12个膜定位候选物,其中CD276(B7-H3)显示出强烈的膜性染色。 结论:整合的单细胞和空间分析描绘了PBA中肿瘤来源的内皮特征,并验证了其空间和CNV背景。AS_signature稳健地将肿瘤内皮与正常内皮区分开来。不同的转移性CNV特征提示存在具有器官特异性趋向的亚克隆演化。CD276成为一个具有临床可干预性的膜抗原,可用于ADC开发。
查看英文原文 English abstract
Background: Primary breast angiosarcoma (PBA) is an extremely rare and aggressive vascular malignancy with limited therapeutic options. Genomic studies have identified recurrent alterations such as KDR, PIK3CA, TP53, and FLT4, but gene-expression patterns have been scarcely reported, and their spatial context remains poorly defined. Only two spatial transcriptomic studies have been published, both from a single institution and focused mainly on immune-cell distribution while analyzing breast and non-breast angiosarcomas together. These two studies identified tumor regions based on endothelial markers (ERG/CD31); yet these markers can be diminished in dedifferentiated angiosarcoma, and reliable tumor-specific markers remain unavailable. Whether PBA shares molecular features with angiosarcomas from other anatomical sites also remains unclear. Objective : To delineate transcriptional programs of primary and metastatic PBA and validate them using integrated spatial analysis, and to identify candidate tumor cell markers with therapeutic applicability. Methods: We integrated single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics from matched PBA specimens. Fresh tumors (361T, 256T) were profiled using BD Rhapsody scRNA-seq and integrated with normal breast tissue (388D). CNVs were inferred by InferCNV. Visium spatial transcriptomics were performed on FFPE samples from two cases, with ovarian and liver metastases available from one (256T). CD276 expression was evaluated by immunohistochemistry in 3 PBA cases. Results: Integrated scRNA-seq from the three samples identified three endothelial clusters: c5 (normal ECs), c7 (tumor ECs), and c11 (lymphatic ECs). Cluster 7 expressed PECAM1/CDH5 and angiogenic/lymphangiogenic regulators, and InferCNV confirmed its tumor identity. Differential expression between c7 and c5 yielded a tumor-specific gene set (AS_signature) enriched for extracellular-matrix organization, receptor tyrosine-kinase signaling, and basement-membrane remodeling. Spatial InferCNV reproduced tumor-associated CNVs in Visium tumor regions. Joint analysis of ovarian and liver metastases of 256T showed complete UMAP separation and divergent CNV architectures; primary-tumor subclusters localized preferentially to ovarian or hepatic lesions, suggesting organ-selective subclonal dissemination. Twelve membrane-localized candidates were identified, and CD276 (B7-H3) showed strong membranous staining. Conclusions: Integrated single-cell and spatial profiling delineated tumor-derived endothelial features in PBA and validated their spatial and CNV context. The AS_signature robustly distinguished tumor from normal endothelium. Divergent metastatic CNV profiles suggest subclonal evolution with organ-specific tropism. CD276 emerged as a clinically actionable membrane antigen for ADC development.
利益披露 Disclosure
A. Iesato, Japanese Foundation for Multidisciplinary Treatment of Cancer ). Kowa Life Science Foundation ). Chugai Foundation for Innovative Drug Discovery Science ). M. Akiya, None.. K. Otsuji, None.. S. Saeki, None.. N. Uehiro, None.. T. Noda, None.. T. Osako, None.. T. Ueno, None.. R. Maruyama, None.

← 返回 AACR 2026 检索