a Department of Radiation Oncology, Technical University of Munich, Munich, Germany.
b Bundeswehr Institute of Radiobiology, Munich, Germany.
Radiat Res. 2018 Sep;190(3):226-235. doi: 10.1667/RR15013.1. Epub 2018 Jun 20.
In previous studies we determined a gene expression signature in baboons for predicting the severity of hematological acute radiation syndrome. We subsequently validated a set of eight of these genes in leukemia patients undergoing total-body irradiation. In the current study, we addressed the effect of intra-individual variability on the basal level of expression of those eight radiation-responsive genes identified previously, by examining baseline levels in 200 unexposed healthy human donors (122 males and 88 females with an average age of 46 years) using real-time PCR. In addition to the eight candidate genes ( DAGLA, WNT3, CD177, PLA2G16, WLS, POU2AF1, STAT4 and PRF1), we examined two more genes ( FDXR and DDB2) widely used in ex vivo whole blood experiments. Although significant sex- (seven genes) and age-dependent (two genes) differences in expression were found, the fold changes ranged only between 1.1-1.6. These were well within the twofold differences in gene expression generally considered to represent control values. Age and sex contributed less than 20-30% to the complete inter-individual variance, which is calculated as the fold change between the lowest (reference) and the highest Ct value minimum-maximum fold change (min-max FC). Min-max FCs ranging between 10-17 were observed for most genes; however, for three genes, min-max FCs of complete inter-individual variance were found to be 37.1 ( WNT3), 51.4 ( WLS) and 1,627.8 ( CD177). In addition, to determine whether discrimination between healthy and diseased baboons might be altered by replacing the published gene expression data of the 18 healthy baboons with that of the 200 healthy humans, we employed logistic regression analysis and calculated the area under the receiver operating characteristic (ROC) curve. The additional inter-individual variance of the human data set had either no impact or marginal impact on the ROC area, since up to 32-fold change gene expression differences between healthy and diseased baboons were observed.
在之前的研究中，我们确定了一个用于预测狒狒血液急性辐射综合征严重程度的基因表达特征。随后，我们在接受全身照射的白血病患者中验证了其中 8 个基因的一组。在当前的研究中，我们通过使用实时 PCR 检查 200 名未暴露的健康人类供体（122 名男性和 88 名女性，平均年龄为 46 岁）的基线水平，研究了个体内变异性对先前确定的 8 个辐射反应基因的基础表达水平的影响。除了 8 个候选基因（DAGLA、WNT3、CD177、PLA2G16、WLS、POU2AF1、STAT4 和 PRF1）之外，我们还检查了另外两个在体外全血实验中广泛使用的基因（FDXR 和 DDB2）。尽管发现了表达上显著的性别（7 个基因）和年龄依赖性（2 个基因）差异，但折叠变化仅在 1.1-1.6 之间。这些都在通常认为代表对照值的基因表达两倍差异范围内。年龄和性别对个体间差异的贡献小于 20-30%，这是计算最低（参考）和最高 Ct 值之间的倍数变化（最小-最大倍数变化（min-max FC）。大多数基因的 min-max FC 在 10-17 之间；然而，对于三个基因，个体间差异的 min-max FC 分别为 37.1（WNT3）、51.4（WLS）和 1627.8（CD177）。此外，为了确定用 200 名健康人类的基因表达数据替代 18 名健康狒狒的已发表基因表达数据是否会改变健康和患病狒狒之间的区分，我们采用逻辑回归分析并计算了接收器操作特征（ROC）曲线下的面积。人类数据集的额外个体间差异对 ROC 面积要么没有影响，要么只有微小的影响，因为在健康和患病狒狒之间观察到高达 32 倍的基因表达差异。
