背景和目的：食管鳞状细胞癌是常见且高度致命的消化系统恶性肿瘤之一。以5-氟尿嘧啶（5-Fluorouracil, 5-FU）和顺铂（cisplatin, CDDP）为基础的化疗是局部中晚期食管鳞癌的主要治疗手段，然而只有部分患者能从化疗中获益。化疗抵抗是导致食管鳞癌患者临床治疗失败的主要原因，但是化疗抵抗的机制尚不清楚。为了解析食管鳞癌化疗抵抗的生物学机制，本研究建立了食管鳞癌患者来源的肿瘤异种移植模型（patient-derived xenografts，PDX）并开展了体内化疗敏感性实验，在临床前环境下探究了与化疗反应性相关的分子特征。研究结果对预测食管鳞癌化疗抵抗程度、提高化疗疗效具有重要意义。

方法：将食管鳞癌患者原发肿瘤组织移植至免疫缺陷小鼠皮下，以建立食管鳞癌PDX模型。对已建立的PDX模型进行5-FU和CDDP联合化疗处理，根据化疗后小鼠肿瘤体积变化将PDX分为化疗应答组和无应答组。分别对患者原发肿瘤和PDX肿瘤组织进行转录组测序，使用Disambigate软件过滤PDX中鼠源测序序列。通过基因富集分析鉴定与化疗反应性相关的信号通路，差异表达分析确定与化疗抵抗相关的特征基因。使用免疫组化、马松三色染色和多重免疫荧光染色对肿瘤细胞和癌症相关成纤维细胞（cancer-associated fibroblast, CAF）进行蛋白表达定量和空间定位分析，使用免疫印记实验检测成纤维细胞活化相关蛋白与凋亡通路相关蛋白的表达水平。体外共培养实验检测肿瘤细胞与成纤维细胞共培养后肿瘤细胞表型变化和成纤维细胞的趋化能力。利用酶联免疫吸附实验（enzyme-linked immuno sorbent assay, ELISA）检测患者血浆中S100A8的含量。通过染色质免疫沉淀实验确定调控S100A8表达的转录因子。以logistic回归分析计算S100A8表达与化疗反应性的比值比（odds ratio, OR）及95%置信区间（confidence interval, CI）。采用Cox比例风险模型计算风险比(hazard ratio, HR)和95%置信区间，分析揭示S100A8基因表达水平与食管鳞癌和其他肿瘤化疗患者的总生存期之间的相关性。使用Student's t检验比较两组实验结果之间的差异是否显著。

结果：我们发现化疗无应答组显著高表达肌成纤维细胞样CAF特征基因和细胞外基质相关基因。免疫组化和马松三色染色证实无应答组成纤维细胞的活化程度更高、肿瘤间质沉积的胶原蛋白更多。对PDX肿瘤的转录组测序数据分析发现肿瘤细胞中S100A8基因的表达水平与CAF的激活水平和化疗抵抗程度显著正相关。S100A8是警报素家族成员之一，分泌到细胞外后可通过与成纤维细胞表面的CD147受体结合，激活细胞内RhoA-ROCK-MLC2信号通路，诱导转录因子MRTF-A核易位，从而上调细胞外基质相关基因的表达。活化的成纤维细胞激活肿瘤细胞中FAK-SRC和ERK通路，促进肿瘤细胞在化疗后存活。抑制肿瘤细胞中S100A8的表达或使用药物阻断S100A8-CD147通路显著提高异种移植肿瘤的化疗敏感性。ELISA分析显示化疗抵抗的食管鳞癌患者外周血中S100A8的含量更高。多因素logistic回归分析揭示外周血S100A8的水平是评估食管鳞癌患者化疗反应性的独立危险因素。Cox比例风险模型分析揭示S100A8表达水平与食管鳞癌和其他肿瘤化疗患者的较短生存期相关。

结论：本研究确定了肿瘤细胞产生的S100A8是决定食管鳞癌化疗敏感性的重要因素之一。机制上，我们发现了S100A8的一种新功能，即S100A8可通过CD147-RhoA-ROCK-MLC2信号通路激活成纤维细胞并诱导细胞外基质重塑，导致化疗抵抗，揭示了S100A8在调控肿瘤基质微环境中的重要作用。治疗上，我们证明阻断S100A8-CD147通路可抑制成纤维细胞的激活，减少肿瘤微环境中的胶原蛋白的沉积，从而增加食管鳞癌的化疗敏感性。预后上，我们证实外周血和肿瘤组织中的S100A8水平可以作为预测化疗反应性的指标，S100A8是评估食管鳞癌患者化疗反应性的独立危险因素。总之，我们的研究系统解析了S100A8介导的食管鳞癌化疗抵抗的生物学机制，并强调了警报素S100A8在食管鳞癌患者临床管理中的潜在价值。

Background & Aims: Esophageal squamous-cell carcinoma (ESCC) is one of the common and highly lethal malignant tumors of the digestive system. Fluorouracil (5-FU)/cisplatin (CDDP)-based chemotherapy is a standard treatment for locally advanced or metastatic ESCC. However, only a subset of patients benefit from chemotherapy. Chemoresistance is the primary factor contributing to treatment failure in patients with ESCC; however, the mechanisms underlying chemoresistance remain to be fully understood. To elucidate the biological mechanisms of chemoresistance in ESCC, in this study we established patient-derived xenograft (PDX) models from ESCC patients and conducted in vivo chemosensitivity experiments to explore the molecular characteristics associated with chemotherapy responsiveness in a preclinical setting. The findings of this study have important implications for predicting the extent of chemoresistance in ESCC and improving chemotherapy efficacy.

Methods: ESCC PDX models were established by transplanting primary tumor tissues from ESCC patients into immunodeficient mice subcutaneously. The established PDX models were treated with a combination of 5-FU and CDDP chemotherapy, and based on the changes in tumor volume post-chemotherapy, the PDX models were classified into responders and non-responders. Transcriptional profiling was conducted on both patients’ primary tumors and PDXs’ tumor tissues, with mouse-derived sequencing sequences filtered out using the Disambiguate software. Gene enrichment analysis was employed to identify signaling pathways associated with chemotherapy responsiveness, and differential expression analysis was used to identify characteristic genes related to chemoresistance. Immunohistochemistry, Masson’s trichrome staining, and multiplex immunofluorescence staining were utilized to quantitatively analyze protein expression and spatial localization of tumor cells and cancer-associated fibroblast (CAF). Immunoblotting experiments were carried out to determine the expression levels of proteins related to CAF activation and apoptosis pathways. Co-culture experiments were performed to assess changes in tumor cell phenotype and the chemotactic ability of fibroblasts. Enzyme-linked immuno sorbent assay (ELISA) was utilized to measure the levels of S100A8 in patients’ plasma. Chromatin immunoprecipitation experiments were employed to identify transcription factors regulating S100A8 expression. Logistic regression analysis was used to calculate the odds ratio (OR) and 95% confidence interval (CI) for S100A8 levels and chemotherapy responsiveness. Cox proportional risk models calculated hazard ratio (HR) and 95% CI analyses revealed associations between S100A8 gene expression levels and overall survival in patients treated with chemotherapy for ESCC and other tumors. Student’s t test was conducted to evaluate the significance of differences between the experimental results of the two groups.

Results: We found that the chemotherapy non-responsive group exhibited significantly elevated expression of myofibroblastic CAF characteristic genes and extracellular matrix-related genes. Immunohistochemistry and Masson’s trichrome staining confirmed higher levels of CAF activation and increased collagen deposition in the tumor stromal in the non-responsive group. Transcriptomic analysis of PDX tumor tissues revealed a significant positive correlation between the expression levels of the S100A8 gene in tumor cells and the levels of CAF activation and chemoresistance. S100A8 is a member of the alarmin family. After secreted into the extracellular, S100A8 binds to the CD147 receptor of fibroblasts, activating the intracellular RhoA-ROCK-MLC2 signaling pathway, inducing nuclear translocation of the transcription factor MRTF-A, thereby upregulating the expression of extracellular matrix-related genes. Activated CAF stimulate the FAK-SRC and ERK pathways in tumor cells, promoting tumor cells survival after chemotherapy. Inhibiting S100A8 expression in tumor cells or blocking the S100A8-CD147 pathway significantly enhances chemosensitivity in xenograft tumors. ELISA analysis revealed higher levels of S100A8 in peripheral blood of ESCC patients resistant to chemotherapy. Multivariable logistic regression analysis revealed that the levels of S100A8 in peripheral blood is an independent risk factor for assessing chemotherapy responsiveness. Multivariable Cox regression analysis revealed that the levels of S100A8 were associated with worse overall survival (OS) time of ESCC and other cancer patients recieving chemotherapy.

Conclusion: This study identified the tumor cells-derived S100A8 as one of the critical factors determining the sensitivity of ESCC to chemotherapy. Mechanistically, we discovered a novel function of S100A8 whereby it can activate fibroblasts through the CD147-RhoA-ROCK-MLC2 signaling pathway, inducing extracellular matrix remodeling and consequently leading to chemoresistance, revealing the significant role of S100A8 in regulating the tumor stromal microenvironment. Therapeutically, we demonstrated that blocking the S100A8-CD147 pathway can suppress CAF activation, reduce collagen deposition in the tumor microenvironment, thereby enhancing chemosensitivity in ESCC. Prognostically, we found the S100A8 levels in peripheral blood can serve as an indicator of chemotherapy responsiveness. In conclusion, our research systematically elucidated the molecular mechanisms of S100A8-mediated chemoresistance in ESCC and underscored the potential value of alarmin S100A8 in the clinical management of these patients.