Backing into Cancer: Effects of Arsenic on Cell Differentiation

Although the carcinogenicity of arsenic in humans has been unequivocally demonstrated (IARC, 1980) and limited evidence of animal carcinogenicity is available (Kitchin, 2001), an understanding of the mechanism(s) responsible for these effects remain elusive. Among the numerous studies performed to gain a better understanding of mechanisms of arsenic carcinogenicity, a few have involved analyses of potential contributions from intracellular metabolism (including reduction and methylation of arsenic), induction of oxidative stress, depletion of glutathione, and the generation of reactive oxygen species. In contrast, many more studies have focused primarily on genetic mechanisms. While arsenic fails to induce mutations in bacteria or Chinese hamster cells (Rossman et al., 1980), it does cause gene amplification (Lee et al., 1988) and sister chromatid exchange (Lerda, 1994). It has been suggested that these clastogenic and gene expression dysregulating effects might be due to modulation of normal DNA methylation patterns (Mass and Wang, 1997; Zhao et al., 1997; Zhong and Mass, 2001) arising, in part, from altered repair of methylated bases (Lee-Chen et al., 1993) or increased methylation of cytosine (Yamanaka et al., 1997). Though not a direct mutagen, arsenic is also known to act as a comutagen with select agents (i.e., UV radiation, X-rays, alkylating agents) (Jha et al., 1992). This comutagenic effect is apparently mediated by inhibition of DNA repair (Abernathy et al., 1999) arising from effects of arsenic on DNA ligases II and I (Li and Rossman, 1989). Clearly, damage to DNA and/or effects upon RNA transcription/translation present an acceptable grouping of mechanisms that explain, in part, how environmental carcinogens act to bring about their ultimate pathologies. However, for an atypical carcinogen like arsenic that is neither a classic initiator nor a promoter, many investigators have concluded that alternative approaches to study cell transformation and eventually cancer needed to be postulated. One alternative viewpoint that evolved was that in order to explain the carcinogenicity of arsenic, a basic examination of macrocellular processes that might be perturbed by exposure, including regulation of cell proliferation or responsiveness to exogenous signals, needed to be undertaken in appropriate in vitro systems or using cells from arsenic-exposed hosts. One tenet that became a basis for studies in support of these macrocellular pathways was that cancer could be viewed as an imbalance that arises in the cellular homeostatic control over two antagonistic processes—proliferation and differentiation. The relationship between these processes is complex. For example, any cell that is differentially competent will normally remain undifferentiated in the presence of proliferative stimuli; however, these same cells will undergo differentiation and arrest of proliferation in the presence/absence of appropriate biologic signals. Since proliferation is halted during cell differentiation, it is reasonable to conclude that exiting from the cell cycle is absolutely essential; in fact, for cells to undergo differentiation, cells must exit the cycle at G0/G1 and enter into mitogenically quiescent states (Scott et al., 1982; Wier and Scott, 1986). Thus, as noted in the highlighted article and in previous publications (Trouba et al., 2000a,b), if an agent is able to uncouple these antagonistic processes, block differentiation outright, and/or inhibit differentiation by preventing arrest in the required stage of the normal cell cycle, these cells could then be maintained in mitogenically responsive states. Under this scenario, the cells would display increased proliferative activities, a characteristic evident in the earliest stages of most cancers. It is therefore not surprising that the effect of arsenic on both processes has recently been receiving considerable attention. Normally, inhibition of proliferation involves inhibition (by transcriptional targets of p53, including Gadd45 and p21) of Cdc2, a cyclin-dependent kinase whose activation is required for a cell to enter mitosis. Under acute exposure scenarios, significant induction of growth inhibitory protein p53 (and to a lesser extent p21) was noted in different cell types following arsenite treatment (Salazar et al., 1997; Vogt and Rossman, 1 To whom all correspondence should be addressed. Fax: (845) 351–5472. E-mail: cohenm@charlotte.med.nyu.edu. TOXICOLOGICAL SCIENCES 65, 161–163 (2002) Copyright © 2002 by the Society of Toxicology