Imatinib inhibits the kinase activity of Bcr-Abl and is currently the most effective drug for treatment of chronic myeloid leukemia (CML). Imatinib also blocks c-Abl, a physiological tyrosine kinase activated by a variety of stress signals including damaged DNA. We investigated the effect of pharmacological inhibition of c-Abl on the processing of irradiation-induced DNA damage in Bcr-Abl-negative cells. Cell lines and peripheral blood mononuclear cells (PBMCs) from healthy volunteers were treated with imatinib or dasatinib before γ-irradiation. Inhibition of c-Abl caused an enhanced irradiation-induced mutation frequency and slowdown of DNA repair, whereas imatinib was ineffective in cells expressing a T315I variant of c-Abl. Mutation frequency and repair kinetics were also studied in c-Abl−/− murine embryonic fibroblasts (MEFs) retransfected with wild-type c-Abl (wt-Abl) or a kinase-defect variant of Abl (KD-Abl). Enhanced mutation frequency as well as delayed DNA repair was observed in cells expressing KD-Abl. These data indicate that pharmacological inhibition of c-Abl compromises DNA-damage response.
The development of imatinib mesylate (STI571, Gleevec) for the treatment of chronic myelogenous leukemia (CML) is one of the most important advances in malignant disease management during recent years (Buchdunger et al., 1996; Druker et al., 1996). Imatinib directly binds to the kinase domain of Bcr-Abl blocking its activity. In addition, it targets a limited number of other tyrosine kinases such as c-Abl, c-KIT and platelet-derived growth factor receptor (Carroll et al., 1997; Heinrich et al., 2000). Most patients with newly diagnosed CML achieve durable responses upon imatinib therapy (Druker et al., 2006). However, even in patients with complete molecular remission, the Bcr-Abl-positive clone quickly reappears after discontinuation of imatinib (Michor et al., 2005). Therefore, prolonged, maybe lifelong therapy is required to control CML with imatinib. As a consequence, it is of great importance to understand the biological effects of imatinib on normal cells. The observation of an unexpected cardiac toxicity related to c-Abl inhibition (Kerkelä et al., 2006) demonstrates that pharmacological targeting of physiological c-Abl is of clinical relevance.
c-Abl is a tightly regulated cellular tyrosine kinase involved in several stress signaling pathways. It interacts with many proteins involved in processing and repair of DNA damage such as p53 (Wei et al., 2005), p73 (Yuan et al., 1999), Mdm2 (Goldberg et al., 2002), Rad51 (Yuan et al., 1998), DNA-PK (Kharbanda et al., 1997), WRN (Cheng et al., 2003), CSB (Imam et al., 2007) and BRCA1 (Foray et al., 2002). In addition, several studies have shown that the activity of c-Abl is upregulated following genotoxic agents or γ-irradiation (Shaul and Ben-Yehoyada, 2005). Therefore, it seems likely that c-Abl plays an important role for the maintenance of genetic stability. This prompted us to investigate whether long-term inhibition of the c-Abl activity affects genetic stability in Bcr-Abl-negative cells. We analyzed the frequency of hypoxanthine-guanine phosphoribosyltransferase (HPRT) loss-of-function mutations in cell lines exposed to imatinib and in murine embryonic fibroblasts (MEFs) expressing either a kinase defect variant of c-Abl (KD-Abl) or wild type c-Abl (wt-Abl). As shown in Figure 1a, significantly more HPRT mutant colonies were observed in a γ-irradiated Bcr-Abl-negative cell line in the presence of imatinib. This contrasts the effect of imatinib observed in Bcr-Abl-positive cells. We and others found that expression of the active Bcr-Abl kinase in CML cells results in genetic instability (Canitrot et al., 1999; Salloukh and Laneuville, 2000; van der Kuip et al., 2004). This can be partially reversed by imatinib in Bcr-Abl-positive cells resulting in mutation frequencies comparable to those seen in imatinib-treated Bcr-Abl-negative cells (Figure 1b). Thus, inhibition of c-Abl causes an increase of the mutation frequency, whereas blocking of Bcr-Abl activity reduces this frequency. We then asked whether c-Abl inhibition might contribute to genetic instability also in Bcr-Abl-positive cells. Therefore, we analyzed mutation frequency in cells transfected with an imatinib-resistant bcr-abl mutant (Bcr-AblT315I). As expected, in cells expressing imatinib-resistant Bcr-Abl and imatinib-sensitive c-Abl, the HPRT mutation frequency is significantly enhanced in the presence of imatinib (Figure 1b). To verify that inhibition of c-Abl is the responsible mechanism for the enhanced mutation frequency, we examined MEFs derived from c-Abl−/− mice retransfected with either KD-Abl or wt-Abl. As shown in Figure 1c, HPRT mutation frequency was consistently elevated in cells lacking c-Abl kinase activity. We conclude that inhibition of c-Abl compromises mechanisms responsible for the maintenance of genomic integrity.
It has been shown that activation of c-Abl by many sources of DNA damage is associated with inhibition of cell growth and induction of apoptosis (Kharbanda et al., 1998; Wang, 2000). Therefore, we examined whether the effect of c-Abl inhibition on genetic stability is due to differences in cell proliferation and induction of apoptosis. Treatment with imatinib neither affected the total number of cell doublings over a 10-day cultivation period nor long-term clonogenic survival in 32D and BaF3 cells (Figure 2a). In addition, induction of cell death following irradiation was not significantly altered by imatinib in 32D or BaF3 cells (Figure 2b).
Proteins involved in DNA repair are among the well-characterized targets of the c-Abl kinase. To address the question whether imatinib may interfere with the kinetics of γ-irradiation-induced DNA strand break repair, we performed comet assays in murine cell lines and primary human peripheral blood mononuclear cells (PBMCs) from healthy volunteers. The upper panel of Figure 3a shows the kinetics of DNA strand break repair in Bcr-Abl-negative cell lines (left panel) and in PBMCs (right panel) in the presence or absence of imatinib. Imatinib had no effect on the peak amount of irradiation-induced DNA strand breaks (Figure 3a, time 0) or on constitutive breaks in non-irradiated controls (not shown). However, pretreatment with imatinib significantly impaired the efficiency of DNA repair (Figure 3a, upper panel). This phenomenon was also observed in BaF3 cells and with different doses of irradiation (3 and 5 Gy; not shown). We next investigated if other c-Abl inhibitors such as dasatinib also affect DNA repair. As shown in Figure 3a (lower panel), a comparable delay of DNA repair was observed with dasatinib in all cells tested including PBMCs from healthy volunteers (not shown for BaF3 cells). To further verify that this phenomenon is caused by inhibition of c-Abl, we examined the repair kinetics in MEFs expressing either KD-Abl or wt-Abl. Cells expressing the kinase-defect form again showed a significant delay in DNA repair compared to cells transfected with wt-Abl (Figure 3b). To exclude imatinib effects unrelated to c-Abl inhibition, we generated a cell line transfected with an imatinib-resistant c-Abl construct (c-AblT315I). Expression of this c-Abl variant was sufficient to completely abolish the effect of imatinib on the kinetics of DNA repair (Figure 3c).
In conclusion, our experiments demonstrate that inhibition of c-Abl both affects the velocity of DNA repair and the maintenance of genetic stability. These phenomenons are caused by deficient c-Abl kinase activity irrespective whether genetic or pharmacological methods were applied. c-Abl activity plays a role in DNA repair processes, proliferation and induction of apoptosis (Kharbanda et al., 1998). The ultimate Abl-dependent response to DNA damage may depend on the cell type as well as on the type of DNA damage induced. Our results suggest that in hematopoietic cells, c-Abl inhibition alters DNA repair processes rather than survival after irradiation. In accordance with this, Uemura and Griffin (2000) found no effects of imatinib on survival of Bcr-Abl-negative hematopoietic cells after ionizing radiation.
The question remains whether our findings are of importance for the further development of c-Abl kinase inhibitors. So far, clinical observations do not support a direct mutagenic effect of imatinib. The increase of chromosomal aberrations in Philadelphia-negative cells reported during early clinical studies of imatinib (Andersen et al., 2002; Terre et al., 2004) is not observed in patients treated upfront with imatinib. However, agents with mutagenic capacity such as cigarette smoking cause malignant disease not before decades of exposure. Therefore, a carcinogenic risk of imatinib may become evident only after long-term treatment. Consequently, patients treated for many years with imatinib should be closely monitored to timely identify unwanted adverse events.
We are grateful to Tabea Peußer and Kerstin Willecke for technical assistance. The MEFs were a kind gift from Dr JY Wang (UCSD, San Diego, CA, USA).
Research Grants and other Financial Support: This work was financially supported by the Robert Bosch foundation project O2A and the Deutsche Forschungsgemeinschaft project KU 2338/1-1.