PO.TB10.07 · 肿瘤生物学
肿瘤与免疫细胞结构的空间分区界定了前列腺癌中相反的预后作用
Spatial compartmentalization of tumor and immune cellular architecture defines opposing prognostic roles in prostate cancer
该海报暂无可下载的资料
AACR 官方页面
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
背景:
理解免疫细胞在前列腺癌(PCa)中的作用仍然具有挑战性,关于免疫浸润预后价值的报道相互矛盾。我们假设免疫细胞与临床结局的关联由其空间定位而非绝对整体丰度所驱动。为验证这一假设,我们构建了一个PCa患者临床队列,并整合了对肿瘤和周围基质中空间分析的免疫细胞的多重IHC(n=41)。
方法:
利用IHC多重检测评估细胞角蛋白阳性(CKpos;上皮巢)和细胞角蛋白阴性(CKneg;基质)区室内的CD3+、CD8+、CD20+、FoxP3+和CD68+细胞。将细胞计数、密度和比值与转移时间(TTM)、生化复发(BCR)、BMI、肿瘤负荷、肿瘤内异质性、PSA以及TMPRSS2:ERG/ETV4融合进行相关分析。在数据收集验证和整理后,连续变量间相关性的统计分析使用Spearman或Pearson相关,而分类变量使用卡方检验。
结果:
我们的关联分析揭示了这样一种模式:CKpos肿瘤巢(癌细胞)内的免疫浸润与临床结局呈现一致的有利关联(r=0.36-0.52,P<0.001),其中CD8+ T细胞与延长的TTM和推迟的BCR相关性最强。相比之下,CKneg区室(基质)内相同的免疫群体则表现出相反的相关性(r=-0.31至-0.47,P<0.05),其中基质FoxP3+ T细胞与加速转移的关联最强。有趣的是,尽管存在强烈的跨区室相关性(r=0.66-0.79,P<0.001),相同的免疫亚群在同一患者样本内却呈现相反的预后方向,提示存在协调募集但区室特异性作用。较高的BMI仅与基质(非保护性)淋巴细胞浸润增加相关(r=0.39-0.59,P<0.05)。肿瘤负荷与瘤内T细胞和基质巨噬细胞比值呈负相关。TMPRSS2:ERG和ETV4/TMPRSS2融合与基质CD8+浸润升高显著相关,而较高的治疗前PSA与基质CD68+密度相关,将肿瘤生物学与巨噬细胞募集联系起来。
结论:
这项工作表明,免疫细胞群体的空间位置决定了它们与PCa结局的关联。值得注意的是,我们的数据表明高负荷侵袭性肿瘤可能与保护性瘤内免疫的抑制相关。我们当前的工作旨在于更大队列中验证这些观察结果。这些发现凸显了向空间分辨的肿瘤微环境和免疫生物标志物转变的必要性,以用于风险分层和免疫治疗的患者选择。
查看英文原文 English abstract
Background:
Understanding the role of immune cells in prostate cancer (PCa) remains challenging, with conflicting reports on immune infiltration's prognostic value. We hypothesized that spatial localization, rather than immune cells absolute bulk abundance, drives their association with clinical outcomes. To test this, we generated a clinical cohort of PCa patients integrated with multiplex IHC from spatially profiled immune cells in both the tumor and surrounding stroma (n=41).
Methods:
IHC multiplex assay was utilized to assess CD3+, CD8+, CD20+, FoxP3+, and CD68+ cells within cytokeratin-positive (CKpos; epithelial nests) and cytokeratin-negative (CKneg; stroma) compartments. Cell counts, densities, and ratios were correlated with time to metastasis (TTM), biochemical recurrence (BCR), BMI, tumor burden, intratumoral heterogeneity, PSA, and TMPRSS2:ERG/ETV4 fusions. After data collection validation and curation, statistical analysis of correlations between continuous variables was done using Spearman's or Pearson's correlation, while Chi square test was used for categorical variables.
Results:
Our association analysis revealed a pattern where immune infiltration within CKpos tumor nests (cancer cells) showed consistently favorable associations with clinical outcome (r=0.36-0.52, P < 0.001), with CD8+ T-cells showing the strongest correlation with prolonged TTM and delayed BCR. In contrast, the same immune populations in CKneg compartment (stroma) demonstrated opposite correlations (r=-0.31 to -0.47, P < 0.05), with stromal FoxP3+ T-cells most strongly associated with accelerated metastasis. Interestingly, identical immune subsets displayed opposite prognostic directions within the same patient samples, despite strong cross-compartment correlations (r=0.66-0.79, P < 0.001), indicating coordinated recruitment but compartment-specific roles. Higher BMI correlated exclusively with increased stromal (non-protective) lymphocytic infiltration (r=0.39-0.59, P < 0.05). Tumor burden inversely correlated with intratumoral T-cell and stromal macrophage ratios. TMPRSS2:ERG and ETV4/TMPRSS2 fusions were significantly associated with elevated stromal CD8+ infiltration, and higher pre-treatment PSA correlated with stromal CD68+ density, linking tumor biology with macrophage recruitment.
Conclusion:
This work demonstrates that the spatial location of immune cell populations determines their association with outcomes in PCa. Notably, our data indicate that high-burden aggressive tumors might be associated with suppression in protective intratumoral immunity. Our current work is aimed at validating these observations in a larger cohort. These findings highlight the need for a shift towards spatially-resolved tumor microenvironment and immune biomarkers for risk stratification and patient selection for immunotherapies.
利益披露 Disclosure
A. Elsehemy, None..
X. Ci, None..
M. Lam, None..
M. Ramotar, None..
A. Berlin, None.