Polyclonality and Metabolic Heterogeneity in a Colorectal Tumor Model

The monoclonal origin of cancer is widely accepted, although numerous studies suggest that some are of polyclonal origin. Loss-of checkpoints in transformed cells gives rise to carcinomas comprising a wide diversity of cell types that fulfill distinct, even complementary, metabolic functions, contrasting with a hypothetical monoclonal origin. Here, using a Drosophila intestinal tumor model, we show that, despite an identical genetic background, these tumors 1), comprise a conserved set of different metabolic-specialized clusters; 2), are always polyclonal and derive from several clones characterized by distinct metabolic specificity; 3) depend on motility of the founder clones for their growth; 4) share metabolic needs similar to those of human cancers. In summary, our study indicates that, in this model, tumor formation always requires assembly between founder clones potentially providing distinct cellular functions, as visualized by their metabolic heterogeneity. Thus, this polyclonal assembly would constitute a critical step of tumor progression. Overall design: To investigate metabolism in tumors, single and group of 6 tumors were dissected and sequenced. Tumorous flies were heat-shocked at 37°C for 2-hrs 2-3 days after adult eclosion. At least 21-dpi flies were selected based on GFP fluorescence through the abdomens. Tumors were dissected on ice in Ringer buffer, rinsed twice ice-cold Ringer baths and the extracellular matrix was digested by gentle pipetting for 3-5 minutes with 20µL 5X Trypsin (PAN BIOTECH P10-022100) twice diluted. Efficient digestion was promptly checked under fluorescent dissecting microscope. Consequently, for each sample, 1X Ringer buffer was added up to 1.5mL and each tube was filtered through 30µm filters (Sysmex CellTrics). Tubes were centrifugated for 4 minutes at 1000G, at 4°C; 1mL of supernatant was discarded and the same volume of fresh 1X Ringer buffer was added. A second centrifugation (4 minutes, 1000G, 4°C) was done and after carefully removing all supernatant, pellets were gently resuspended in 30µL 1X Ringer buffer for sequencing. The whole sample preparation procedure did not exceed 20 minutes. Cell suspension were loaded into the Chromium Controller using a Chromium Next GEM Single Cell 3' v3.1 kit (10x Genomics) to generate a droplet emulsion. cDNA were purified and libraries were prepared, according to the manufacturer recommendations. Single cell libraries were sequenced on a NextSeq 2000 Intrument (Illumina) using a P2-100 cycles kit. FastQ files were analyzed using the Cell Ranger software (10X Genomics, version 6.0.1), including alignment, filtering and quantitation of reads on the Drosophila genome r6.42 and generation of feature-barcode matrices. We used the Seurat R pipeline 10 for sample quality control and analysis. The R script used to evaluate sample quality, generate clusters, umaps, heatmaps and GO analysis is available at: https://github.com/Pdelam/Delamotte-et-al.-2024.git

肿瘤的单克隆起源已被广泛接受，尽管诸多研究表明部分肿瘤为多克隆起源。转化细胞的检查点缺失会导致癌组织包含多种细胞类型，这些细胞执行不同甚至互补的代谢功能，这与假说中的单克隆起源相悖。本研究借助果蝇肠道肿瘤模型（Drosophila intestinal tumor model）开展实验，结果显示，即便遗传背景完全一致，此类肿瘤仍具备以下特征：1）包含一组保守的、具有不同代谢特化功能的细胞簇；2）始终为多克隆起源，由多个具有独特代谢特异性的克隆演化而来；3）生长依赖于创始克隆的运动能力；4）与人类癌症具有相似的代谢需求。
综上，本研究表明，在该模型中，肿瘤发生始终需要具有潜在不同细胞功能的创始克隆之间发生组装，这一点可通过其代谢异质性得以体现。因此，这种多克隆组装可构成肿瘤进展的关键步骤。
实验整体设计：为探究肿瘤中的代谢特征，我们解剖了单份及6份成组的肿瘤组织并进行测序。成体果蝇羽化后2-3天，于37℃热激2小时。在肿瘤诱导后至少21天，通过腹部绿色荧光蛋白（GFP，Green Fluorescent Protein）荧光筛选携带肿瘤的果蝇。肿瘤在冰上的林格氏液（Ringer buffer）中解剖，随后用预冷的林格氏液漂洗两次；将肿瘤组织置于20μL经两倍稀释的5×胰蛋白酶（Trypsin，PAN BIOTECH P10-022100）中，轻柔吹打3-5分钟以消化细胞外基质。在荧光体式显微镜下及时确认消化效果。随后，向每份样本中加入1×林格氏液至总体积1.5mL，使用30μm滤器（Sysmex CellTrics）过滤样本。将样本管置于4℃、1000G条件下离心4分钟，弃去1mL上清液，加入等体积新鲜1×林格氏液。再次进行4℃、1000G离心4分钟，小心弃去全部上清液后，用30μL 1×林格氏液轻柔重悬沉淀，用于后续测序。整个样本制备流程耗时不超过20分钟。
将细胞悬液上样至Chromium Controller系统，使用Chromium Next GEM单细胞3'端v3.1试剂盒（10x Genomics）制备液滴微乳液。按照制造商的操作指南纯化cDNA并构建测序文库。使用P2-100循环试剂盒，在NextSeq 2000测序仪（Illumina）上完成单细胞文库测序。
使用Cell Ranger软件（10X Genomics，版本6.0.1）分析FastQ文件，包括比对、过滤、定量果蝇参考基因组r6.42的测序reads，并生成特征-条形码矩阵。我们采用Seurat R分析流程进行样本质控与数据分析。用于评估样本质量、生成细胞簇、统一流形逼近与投影（UMAP）图、热图及基因本体（GO，Gene Ontology）富集分析的R脚本可在以下网址获取：https://github.com/Pdelam/Delamotte-et-al.-2024.git