Chronic myeloid leukemia (CML) is characterized by the presence of the BCR-ABL oncogene which encodes an activated tyrosine kinase. Despite the success of tyrosine kinase inhibitors (TKI), such as imatinib (IM), in CML treatment, a considerable percentage (12–50%) of patients develops resistance to TKIs. This can result in progression to blast crisis (BC), at which stage durable response to any TKI therapy is minimal. It is still unclear how additional genetic lesions, believed to play a major role in the transformation to BC, are generated, but it is likely that the BCR-ABL induction of DNA damage and low-fidelity DNA repair, added to inhibition of apoptosis contributes to this process.
The Growth Arrest and DNA Damage 45 (GADD45A, B and G) proteins are tumor suppressors which coordinate cell cycle arrest, DNA repair and apoptosis in response to genotoxic and oncogenic stress. Failure to activate GADD45 expression, e.g. due to the presence of leukemic oncogenes or its promoter methylation, is associated with an attenuated DNA-damage response and impaired drug response.
A microarray study showed that GADD45G is in the top 10% of genes downregulated in CML patients in BC compared to those in chronic phase (CP). Given the critical role of GADD45 proteins as key regulators of the cellular stress response, our hypothesis is that GADD45G reduced expression in CML-BC is mediated by BCR-ABL and contributes to the accumulation of mutations seen in BC. Therefore the aims of this work are to determine the mechanism of GADD45G downregulation by BCR-ABL and its functional effects on CML cells.
We have confirmed the reduced GADD45G levels in CD34+ cells from BC patients versus CP and normal controls using quantitative RT-PCR. Furthermore, ectopic expression of BCR-ABL in a myeloid cell line resulted in decreased expression of GADD45G and, conversely, BCR-ABL inhibition with IM in several CML cell lines resulted in induction of GADD45G expression. These results confirm GADD45G as a down-modulated target of BCR-ABL. To determine the effect of GADD45G expression on CML cells, a CML cell line was transduced with GADD45G and showed significantly reduced proliferation, G1 cell cycle arrest and decrease in cell viability due to increased apoptosis.
Once the GADD45G tumor suppressor effect in CML was confirmed, we investigated the mechanism by which BCR-ABL inhibits its expression. Silencing of GADD45G expression in solid tumors can occur by promoter methylation; therefore, we analyzed the GADD45G promoter methylation status in the presence or absence of BCR-ABL. Bisulfite genomic sequencing of BCR-ABL-positive and –negative hematopoietic cell lines and on peripheral blood mononuclear cells from CML-CP and BC patients, as well as of normal controls, showed no correlation between the presence of BCR-ABL and promoter methylation.
Another mechanism possibly employed by BCR-ABL to regulate GADD45G expression is interference with transcription factor (TF) binding to the promoter. Bioinformatics analysis revealed, among several other TF binding sites, the presence of two highly conserved RUNX1/AML1 binding sites on the promoter. The TF RUNX1 is a key controller of hematopoiesis and heterodimerises with CBF-β to regulate gene expression. RUNX1 mutations in AML result in lower GADD45A expression and increase in DNA damage. In CML, RUNX1 was reported as mutated in >30% of patients in BC, and was suggested to cooperate with Bcr-Abl in the induction of AML in mice. It has also been shown to be a target of inhibitory phosphorylation by SRC kinases, which display increased activity in CML-BC. We therefore co-transfected a CML cell line with a GADD45G promoter luciferase reporter and both RUNX1 and CBF-β constructs. This resulted in a 10-fold increase in luciferase activity, indicating that RUNX1 is involved in the regulation of GADD45G expression in CML cells.
Overall, our results suggest that GADD45G has a tumor suppressor role in CML and that its down-regulation by BCR-ABL could, therefore, contribute to the accumulation of DNA damage and progression to BC. In addition, our data identify RUNX1 as an inducer of GADD45G expression and, together with data from the literature, suggest that RUNX1 mutations or protein inactivation in CML could result in the observed reduced GADD45G expression. We are currently investigating this possibility.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.