Induction of γ-Globin Gene Expression Via the Nrf2/Antioxidant Response Element Signaling Pathway.

Abstract 975

Induction of fetal hemoglobin (HbF) has proven therapeutic potential to treat sickle cell disease and β-thalassemia. However, agents known to be effective in humans, including HU, DNMT inhibitors and butyrate derivatives are not ideal due to suppression of hematopoeisis and the possibility of long-term side effects including those related to DNA mutation and epigenetic changes. Recently, two natural compounds, angelicin (Lampronti, et al, Eur J Hematol 2003) and resveratrol (Rodrigue, et al, Hematology 2001) have been found to induce γ-globin gene expression in K562 cells. These agents may be important lead compounds as they are generally non-cytotoxic and are being evaluated in ongoing human trials as cancer chemoprevention agents. These and several other agents are thought to work by activating antioxidant response pathway genes. The products of these genes are enzymes involved in antioxidant and detoxification activities and include NADPH-quinone oxireductase 1 (NQO1), glutamate-cysteine ligase (GCL) and glutathione S-transferase (GST). The activation of these genes is mediated by the transcription factor NF-E2 related factor 2 (Nrf2) which binds to a specific antioxidant response element (ARE) sequence (TGACnnnGCA) in target gene promoters. The proximal γ-globin promoter contains an ARE sequence between two CAAT boxes, suggesting that it too may be activated by Nrf2. This led us to hypothesize that drugs which activate the ARE/Nrf2 pathway may provide a less toxic approach to HbF induction. Several compounds that activate this pathway are already approved for human use or are under investigation in human trials. To test this hypothesis, we treated K562 cells with various Nrf2 pathway activators. First, we determined the maximum concentration of each compound that did not inhibit proliferation of K562 cells. Using these doses, we performed time course experiments by treating K562 cells and then measuring steady-state mRNA levels of Nrf2 target genes and γ-globin using quantitative real-time PCR. We initially tested six compounds that are known to induce antioxidant response genes. We found the most pronounced γ-globin induction followed treatment with tert-butylhydroquinone (tBHQ) (2.8 fold), 3H-1,2 dithiole-3-thione (D3T) (2.4 fold) and curcumin (2.1 fold). We next tested tBHQ in erythroid precursors isolated from normal human bone marrow. In these primary cells it also significantly increased expression of γ-globin mRNA and of Nrf2 target genes NQO1, GCLM and GSTP1. To determine if Nrf2 was necessary for tBHQ induction of γ-globin mRNA we used siRNA to knockdown Nrf2 expression in K562 cells. Our results show that transiently silencing Nrf2 transcription prevented γ-globin and NQO1 gene induction by tBHQ compared to samples transfected with scrambled siRNA (p < 0.01). To determine if this requirement for Nrf2 is due to Nrf2 binding at the ARE consensus sequence in the γ-globin promoter, ChIP analysis was performed. This demonstrated that Nrf2 binding at both the γ-globin and NQO1 promoters was progressively increased at 4 and 6 hours after tBHQ treatment compared to the negative control necdin promoter where there was no change. Taken together, these results suggest that the antioxidant tBHQ induces γ-globin mRNA expression through the Nrf2/ARE pathway and that this may be a less toxic strategy for γ-globin gene induction in people with hemoglobinopathies.