Table 1_A multi-omics approach elucidates the link between artificial food colorings and common cancers.docx

BackgroundArtificial food colorings (AFCs) are widely used, yet their potential links to cancer remain unclear. We investigated associations between commonly used AFCs and cancer-related molecular networks and prognosis.
MethodsAFCs-related targets were collected from CTD, ChEMBL, SEA, and TargetNet, and cancer-related targets from GeneCards, OMIM, and CTD. Overlapping targets were subjected to STRING-based PPI analysis and Cytoscape visualization, followed by GO/KEGG enrichment. Core targets were evaluated for differential expression in GEO datasets of non-small cell lung cancer (NSCLC), colon adenocarcinoma (COAD), gastric cancer (GC), and breast cancer (BRCA), with GSEA for pathway characterization. Expression patterns were examined using GEPIA2. TCGA transcriptomic and clinical data were used to construct prognostic models via univariate Cox regression, LASSO selection, and multivariate Cox regression. Key genes were assessed using the Human Protein Atlas (HPA) and qPCR, and in vivo experiments evaluated tumor growth under AFCs exposure.
ResultsFour high-exposure AFCs were analyzed. We identified 108 shared AFCs–cancer targets and prioritized 50 core targets. Enrichment analyses highlighted cancer-relevant functional themes, including cell-cycle regulation (cyclin-dependent protein kinase holoenzyme complex) and oncogenic signaling (PI3K–Akt pathway). Multiple core targets were dysregulated in GEO tumor datasets, and GSEA identified consistently enriched pathways across cancer types. TCGA-derived signatures stratified patients into distinct risk groups with significantly different overall survival. HPA supported protein-level differences for selected targets, qPCR indicated that Allura Red AC or Tartrazine modulated prognostic gene expression in cancer cell lines, and AFCs exposure was associated with accelerated LLC tumor growth in mice.
ConclusionThis integrative analysis suggests that commonly used AFCs may be associated with cancer-related molecular networks and adverse prognosis in NSCLC, COAD, GC, and BRCA, informing future safety evaluation and regulation.

研究背景：人工食用色素（Artificial Food Colorings, AFCs）应用广泛，但其与癌症的潜在关联仍不明确。本研究探讨了常用人工食用色素与癌症相关分子网络及预后之间的关联。
研究方法：从CTD、ChEMBL、SEA及TargetNet数据库收集人工食用色素相关靶点，从GeneCards、OMIM及CTD数据库收集癌症相关靶点。取二者的重叠靶点进行基于STRING的蛋白质相互作用（Protein-Protein Interaction, PPI）分析，并通过Cytoscape进行可视化，随后开展GO/KEGG富集分析。选取核心靶点，在非小细胞肺癌（Non-Small Cell Lung Cancer, NSCLC）、结肠腺癌（Colon Adenocarcinoma, COAD）、胃癌（Gastric Cancer, GC）及乳腺癌（Breast Cancer, BRCA）的GEO数据集内评估其差异表达情况，并通过基因集富集分析（Gene Set Enrichment Analysis, GSEA）对通路特征进行解析。利用GEPIA2平台检测基因表达模式。采用TCGA转录组与临床数据，通过单因素Cox回归、LASSO筛选及多因素Cox回归构建预后模型。借助人类蛋白质图谱（Human Protein Atlas, HPA）与实时定量聚合酶链式反应（quantitative Polymerase Chain Reaction, qPCR）对关键基因进行验证，并通过体内实验评估人工食用色素暴露下的肿瘤生长情况。
研究结果：本研究分析了4种高暴露量人工食用色素。共鉴定出108个人工食用色素-癌症共享靶点，并筛选得到50个核心靶点。富集分析揭示了与癌症相关的核心功能主题，包括细胞周期调控（细胞周期蛋白依赖性激酶全酶复合物）及致癌信号通路（PI3K-Akt信号通路）。多个核心靶点在GEO肿瘤数据集内呈现表达失调，且GSEA分析在多种癌症类型中均鉴定到持续富集的通路。基于TCGA构建的特征评分可将患者划分为不同风险组，两组患者的总生存期存在显著差异。HPA验证了部分靶点的蛋白水平表达差异，qPCR实验显示诱惑红AC（Allura Red AC）或柠檬黄（Tartrazine）可调控癌细胞系内的预后相关基因表达，而人工食用色素暴露与小鼠体内LLC肿瘤生长加速显著相关。
研究结论：本整合分析表明，常用人工食用色素可能与非小细胞肺癌、结肠腺癌、胃癌及乳腺癌中的癌症相关分子网络及不良预后存在关联，可为未来的安全性评价与监管提供参考依据。