Validation of the Predictive Toxicogenomics Space with cell culture data

Cell culture for in vitro cytotoxicity predictions. To validate the predictive performance of the PTGS, a set of CMap instances for 38 compounds that were not included in the NCI-60 data set were assessed. MCF7 (ATCC® HTB22™), PC­3 (ATCC® CRL­1435™) and HL­60 (ATCC® CCL­240™) cell lines were obtained directly from American Type Culture Collection (LGC Promochem AB) and maintained at 37 °C with 5 % CO2 in a humidified incubator according to provider’s instructions. Cell number was titrated to ensure that cell proliferation remained in a linear-exponential phase throughout the experiment (1000-2000 cells per well were plated). Each experiment was performed from unique assay ready cells (same passage). Data quality and assay comparability were first verified by replicating the measurements for 36 instances for 16 different compounds already measured in NCI-60. Measurements were carried out at the Institute for Molecular Medicine Finland, FIMM. The authors want to thank Ida Lindenschmidt and the High Throughput Biomedicine unit at FIMM for technical support to cellular high-throughput screening assays.Table 1. Raw data as % of cell viability after compound treatment. Molar concentrations from -8 to -4 on the log10 scale (5 doses) were employed. Columns: Compound, Cell Line, Cell.viability(%) (-8 to -4, log10.conc.).Table 2. Calculated GI50 values for control treatments, 16 compounds which have corresponding cytotoxicity data in the NCI-60 DTP database. Columns: Chemical, CellLine, CMapDose (concentration at which CMap profile was measured), GI50.NCI60 (GI50 in the NCI60 DTP database), Batch.NCI60 (batch in the NCI60 database), Batch.FIMM (batch in FIMM dataset), GI50.FIMM (GI50 value in the FIMM dataset).Table 3. Calculated GI50 values for test treatments, 38 compounds which have gene expression data in the Connectivity Map database (used to calculate PTGS cytotoxicity virtual GI50 scores). Columns: Chemical, CellLine, CMapDose (concentration at which CMap profile was measured), Batch.FIMM (batch in FIMM dataset), GI50.FIMM (GI50 value in the FIMM dataset).用于体外细胞毒性预测的细胞培养实验。为验证PTGS的预测性能，本研究针对38种未纳入NCI-60数据集的化合物，开展了连接图谱（Connectivity Map，简称CMap）实例集的评估工作。MCF7（ATCC® HTB22™）、PC-3（ATCC® CRL-1435™）与HL-60（ATCC® CCL-240™）细胞系均直接从美国模式培养物集存库（现隶属于LGC Promochem AB）获取，严格遵循供应商的操作指南，于37℃、5% CO₂的湿润培养箱中进行传代培养。为确保实验全程细胞增殖维持在线性指数生长期，我们对细胞接种密度进行了滴定优化，最终每孔接种1000~2000个细胞。所有实验均采用同一传代批次的即用型检测细胞开展。首先通过重复测定16种已在NCI-60数据库中完成检测的化合物的36个实例，对数据质量与实验可比性进行验证。本研究的所有检测工作均在芬兰分子医学研究所（FIMM）完成。作者谨向Ida Lindenschmidt及FIMM高通量生物医学单元致以诚挚谢意，感谢其为细胞高通量筛选实验提供的技术支持。表1 化合物处理后细胞活力百分比原始数据。实验采用log₁₀浓度区间为-8至-4的5个梯度剂量。列名：Compound（化合物）、Cell Line（细胞系）、Cell.viability(%) (-8 to -4, log10.conc.)（细胞活力百分比（对应log₁₀浓度-8至-4））。表2 对照处理组的GI₅₀计算值：共纳入16种化合物，其NCI-60 DTP数据库中存有对应细胞毒性数据。列名：Chemical（化合物）、CellLine（细胞系）、CMapDose（CMap谱图测定时的浓度）、GI50.NCI60（NCI60 DTP数据库中的GI₅₀值）、Batch.NCI60（NCI60数据库中的批次信息）、Batch.FIMM（FIMM数据集的批次信息）、GI50.FIMM（FIMM数据集中的GI₅₀值）。表3 受试处理组的GI₅₀计算值：共纳入38种化合物，其连接图谱（Connectivity Map）数据库中存有基因表达数据，用于计算PTGS细胞毒性虚拟GI₅₀评分。列名：Chemical（化合物）、CellLine（细胞系）、CMapDose（CMap谱图测定时的浓度）、Batch.FIMM（FIMM数据集的批次信息）、GI50.FIMM（FIMM数据集中的GI₅₀值）。