| First Author | Hirabayashi Y | Year | 2009 |
| Journal | Exp Hematol | Volume | 37 |
| Issue | 2 | Pages | 195-205 |
| PubMed ID | 19070417 | Mgi Jnum | J:146574 |
| Mgi Id | MGI:3837924 | Doi | 10.1016/j.exphem.2008.10.006 |
| Citation | Hirabayashi Y, et al. (2009) Comparison of murine gene expression profiles between spontaneous and radiation-induced myelogenous leukemias: stochastic and probabilistic expression variances in the former vs radiation-specific expression commonalities in the latter. Exp Hematol 37(2):195-205 |
| abstractText | OBJECTIVE: To elucidate the common characteristics of murine radiation-induced myelogenous leukemias, global gene-chip expression profiles were compared with age-matched steady-state bone marrow tissue profiles and spontaneous myelogenous leukemia profiles. MATERIALS AND METHODS: Six each of C3H/He mice-derived radiation-induced and spontaneously developed myelogenous leukemias were analyzed. Bone marrow cells from five each of 2- and 21-month-old mice were used to subtract nonleukemic information in the analysis. mRNAs from individual mice were analyzed separately using 45,101 gene chips followed by computational biological analysis. RESULTS: First, principal component analysis (PCA) was performed to discriminate the gene expression profiles of individual mice with radiation-induced myelogenous leukemia from those of bone marrow cells from 2- or 21-month-old mice. Discriminant union genes for individual leukemias were then selected, which finally yielded 242 genes, among which six are radiation-related genes including Hus-1, Edf1a2, andVegf-c; 16 are apoptosis/cell-death-related genes, 13 are cell-cycle/cell-growth-related genes, and 50 are suppressor/promoter genes. PCA of these 242 genes consistently enabled the discrimination of the radiation-induced leukemias from the spontaneous leukemias. Second, the other components of the same PCA provided four different eigenvector clusters in an unsupervised manner representing four histopathological findings, with which the differential diagnosis in molecular taxonomy was significant as determined by analysis of variance of the global gene expression profiles. CONCLUSION: Discriminant union genes in radiation-induced myelogenous leukemias against spontaneous myelogenous leukemias and age-matched nonleukemic bone marrow profilings generated by unsupervised computational analysis essentially represent probabilistic biomarkers for radiation-induced myelogenous leukemias, which may contribute to developing a model for risk of secondary carcinogenesis in patients treated by whole-body irradiation. |
