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

Hematopoietic neoplastic diseases develop in C3H/He and C57BL/6 mice after benzene exposure: strain differences in bone marrow tissue responses observed using microarrays

In this study, Trp53-deficient and wild-type mice of both C57BL/6 and C3H/He strains were exposed to benzene (33, 100, and 300 ppm; 6h/day, 5 days/week for 26 weeks) and then observed for lifetime. As results, first, the incidence of nonthymic lymphomas in C57BL/6 mice and acute myeloid leukemias (AMLs) in C3H/He mice showed linear responses at the lower exposure level in Trp53-deficient mice; second, the incidence of thymic lymphomas in C57BL/6 mice and nonthymic lymphomas in C3H/He mice increased without a plateau-like ceiling; thus, the former equivocal induction of hematopoietic neoplasms (HPNs) in the case of low-dose benzene exposure was assumed to be based on the DNA repair potential in wild-type mice, and the latter limited increase in HPNs in the case of high-dose benzene exposure was considered to be due to excessive apoptosis in wild-type mice. Concerning the incidence of AMLs, though a dose of 300 ppm benzene inhalation induced 9% AMLs in wild-type C3H/He mice-AML-prone, it induced AMLs in 38% of Trp53-deficient C3H/He mice. Because AMLs were also observed in Trp53-deficient mice, including in the C57BL/6 mice, benzene exposure may also be a potent inducer of AMLs in mice with some strain differences. In the present study, to elucidate the hematopoietic stem cell-specific, aryl hydrocarbon-receptor-related low-dose adverse effect, global gene expression in the bone marrow was analyzed at 28 days after 2-week-intermittent exposure to 150 mg/kg b.w. benzene, by gavage, i.e., equivalent to the above inhalation protocol with 300 ppm. We observed two conceptually different gene expression profiles; "common gene profiles" (CGPs) shared among mice in each group, and "stochastic gene profiles" (SGPs), i.e., unique union genes from one individual mouse to another. The CGPs of the experimental group and the SGPs of each individual mouse were separately characterized by individual assay. Concerning the CGPs, reciprocal strain differences between C3H/He and C57BL/6 mice in expression gene profiles, both plausible for leukemogenesis, were identified; namely, dominant downmodulations of Sltm and Cryl1, related to suppression of apoptosis and genomic instability in C3H/He mice, respectively, and dominant downmodulations of Atrx/rad54 and Kdm2a, related to a decrease in DNA repair and genomic instability, respectively, in C57BL/6 mice. These findings imply that these reciprocal gene expression differences induced by benzene exposure may lead each strain to undergo different hematopoietic neoplastic pathways. In contrast, each individual mouse often shows a unique SGP. SGPs often include transcription factors, which regulate reciprocal signaling pathways including further SGPs. Among them, apoptosis-related genes expressed in C57BL/6 mice and those in C3H/He mice were attributable to different combinations of SGPs. Such stochastic case-by-case gene expression may be in good agreement with the individual and strain differences observed following benzene exposure. Because gene chip microarray techniques can elucidate stochastic changes in gene expression profiles, possible stochastic toxicology and its future role are discussed.