Cellular Response to Cytarabine Is Modulated by the DNA Mismatch Repair Pathway: Implications for Treatment of Acute Myeloid Leukemia

Abstract 1819

The DNA mismatch repair (MMR) pathway is responsible for repair of spontaneous errors arising during DNA replication, thus maintaining the integrity of the genome. DNA MMR is frequently dysregulated in some forms of leukemia. We and others have shown that microsatellite instability, the hallmark of dysfunctional DNA MMR, is present in up to 90% of therapy-related myeloid leukemia, 50% of relapsed myeloid leukemia, but is rarely seen in de novo leukemia. Paradoxically, functional MMR mediates the cytotoxicity of certain chemotherapeutic agents, particularly methylating agents and the nucleoside analogue 6-thioguanine (6-TG), and dysregulation of the MMR pathway confers tolerance to these agents. In the present study, using cell lines harboring defects in MMR components, we show that MMR status also modulates response to the nucleoside analogue cytarabine (Ara-C) and to other therapeutic nucleoside analogues commonly used in the treatment of leukemia. We initially determined gene and protein expression levels of the major MMR components (MSH2, MSH3, MSH6, MLH1 and PMS2) in two psuedo-isogenic cell line pairs, and investigated the ability of cell extracts to bind to defined mismatches in electrophoretic mobility shift assays. In cytotoxicity assays, the cell lines HL-60R (which demonstrates 200-fold overexpression of MSH3) and MT-1 (which lacks functional MSH6 due to bi-allelic gene mutation) were tolerant to the cytotoxic effects of a methylating agent, methylnitrosourea (MNU), and 6-TG, relative to their respective parental cell lines. We also generated a panel of fully isogenic MMR-defective cell lines in which either MSH2, MSH3 or MSH6 protein was reduced to almost negligible levels using short hairpin RNA-mediated gene knockdown. Knockdown of either MSH2 or MSH6 conferred tolerance to the killing effects of MNU and 6-TG by virtue of loss of MutSα activity (a heterodimer of MSH2 and MSH6 responsible for recognition of base:base mispairs), whereas knockdown of MSH3 did not affect cellular response to these agents. Consistent with a role for MMR in affecting cellular response to other nucleoside analogues used to treat leukemia, the cell lines also displayed differential toxicity to the killing effects of Ara-C, clofarabine, cladribine and fludarabine compared to their MMR-proficient parental counterparts, however the exact response was dependent on the specific nature of the MMR defect. Cell lines with a reduction in MSH2 protein demonstrated hypersensitivity to cytotoxicity induced by these nucleoside analogues. Conversely, MSH3 knockdown conferred resistance to the cytotoxic effects of these agents. These data suggest that DNA MMR can affect response to nucleoside analogues via multiple mechanisms, and may also involve interaction of DNA MMR components with other DNA repair pathways. One possibility is that these agents induce base lesions in DNA recognized by DNA MMR components. Consistent with this model, we have shown that Ara-C induces DNA polymerase slippage in vitro, generating a substrate potentially recognized by MutSβ (a heterodimer of MSH2 and MSH3 responsible for recognition of small insertions and extrahelical loops). Furthermore, we have also shown that Ara-C is mutagenic at the thymidine kinase and hypoxanthine-guanine phosphoribosyltransferase loci in the TK6 cell line. Taken together, these data suggest that cellular MMR status affects response to nucleoside analogues and furthermore, the specific nature of the defect is important in determining the exact response. These findings have implications for the use of nucleoside analogues in the treatment of cancers where MMR dysfunction has been identified to occur with high frequency, such as therapy-related and relapsed acute myeloid leukemia.