Colorectal cancer (CRC) is a heterogeneous disease. More insight into the biological diversity of CRC is needed to improve therapeutic outcomes. Established CRC cell lines are frequently used and were shown to be representative models of the main subtypes of CRC at the genomic and transcriptomic level. In the present work, we established stable, luciferase expressing derivatives from 10 well-established CRC cell lines, generated spheroids and subcutaneous xenograft tumors in nude mice, and performed comparative characterization of these model systems. Transcriptomic analyses revealed the close relation of cell lines with their derived spheroids and xenograft tumors. The preclinical model systems clustered with patient tumor samples when compared to normal tissue thereby confirming that cell-line-based tumor models retain specific characteristics of primary tumors. Xenografts showed different differentiation patterns and bioluminescence imaging revealed metastatic spread to the lungs. In addition, the models were classified according to the CMS classification system, with further sub-classification according to the recently identified two intrinsic epithelial tumor cell states of CRC, iCMS2 and iCMS3. The combined data showed that regarding primary tumor characteristics, 3D-spheroid cultures resemble xenografts more closely than 2D-cultured cells do. Furthermore, we set up a bioluminescence-based spheroid cytotoxicity assay in order to be able to perform dose–response relationship studies in analogy to typical monolayer assays. Applying the established assay, we studied the efficacy of oxaliplatin. Seven of the ten used cell lines showed a significant reduction in the response to oxaliplatin in the 3D-spheroid model compared to the 2D-monolayer model. Therapy studies in selected xenograft models confirmed the response or lack of response to oxaliplatin treatment. Analyses of differentially expressed genes in these models identified CAV1 as a possible marker of oxaliplatin resistance. In conclusion, we established a combined 2D/3D, in vitro/in vivo model system representing the heterogeneity of CRC, which can be used in preclinical research applications.
结直肠癌是一种异质性疾病。为改善治疗效果，需进一步了解其生物学多样性。已建立的结直肠癌细胞系被广泛使用，并在基因组和转录组水平上被证明能代表结直肠癌主要亚型。本研究选取10种成熟的结直肠癌细胞系，建立稳定表达荧光素酶的衍生株，构建球体培养模型及裸鼠皮下移植瘤模型，并对这些模型系统进行比较表征。转录组分析显示细胞系与其衍生的球体及移植瘤具有高度相关性。与正常组织相比，这些临床前模型系统与患者肿瘤样本聚类，证实基于细胞系的肿瘤模型保留了原发肿瘤的特定特征。移植瘤呈现不同的分化模式，生物发光成像显示其向肺部转移扩散。此外，这些模型根据CMS分类系统进行分类，并依据近期发现的结直肠癌两种固有上皮肿瘤细胞状态（iCMS2和iCMS3）进行亚类划分。综合数据表明，在原发肿瘤特征方面，3D球体培养模型比2D培养细胞更接近移植瘤模型。同时，我们建立了基于生物发光的球体细胞毒性检测方法，使其能够像典型单层培养实验那样进行剂量-反应关系研究。应用该检测方法，我们研究了奥沙利铂的疗效。在十种细胞系中，有七种在3D球体模型中对奥沙利铂的反应显著低于2D单层模型。在选定移植瘤模型中进行的治疗研究证实了其对奥沙利铂治疗的反应或无反应。通过对这些模型中差异表达基因的分析，发现CAV1可能是奥沙利铂耐药性的潜在标志物。综上所述，我们建立了一个结合2D/3D、体外/体内实验的模型系统，能够体现结直肠癌的异质性，可用于临床前研究应用。
肿瘤(癌症)患者之家