β-Catenin Suppression Targets Imatinib Resistant Leukemia Stem Cells In Mice with BCR-ABL Induced Myeloproliferative Disease

Abstract 93

CML is driven by the oncogenic fusion kinase BCR-ABL which is capable of corrupting and transforming hematopoietic stem cells (HSC) and thus inititating myeloproliferative disease (MPD). While imatinib (IM) has become an effective treatment for CML, the leukemia stem cells (LSC) which reside in the bone marrow (BM) remain largely unaffected by imatinib treatment. As LSC rely on their ability to self-renew, several genes/pathways responsible for self-renewal properties have been investigated as potential targets of LSCs in myeloid neoplasia. Recent studies clearly indicate that canonical Wnt/β-catenin-signaling plays a pivotal role in initiation and development of BCR-ABL induced MPD (Zhao, Cancer Cell 2009; Hu, Leukemia 2009). However, the impact of β-catenin (Ctnnb1) on the maintenance of LSC in established disease, which more accurately models the clinical scenario, has not been extensively studied.

To address the impact of targeting Ctnnb1 in CML-LSC, we used a conditional Ctnnb1 knockout mouse with an estrogen-receptor driven Cre-recombinase as BM-donors. CML was induced by retroviral transduction of either Ctnnb1^flox/flox ER-Cre+, Ctnnb1^flox/wt ER-Cre+, or Ctnnb1^wt/wt ER-Cre+ murine bone marrow with a virus encoding BCR-ABL (MSCV-BCR-ABLp210-GFP), and injection of transduced cells into wildtype syngeneic recipient mice. Administration of tamoxifen to recipient mice with CML led to excision of Ctnnb1 in the Ctnnb1^flox/flox homozygous or Ctnnb1^flox/wt heterozygous cells.

Interestingly genetic inactivation of Ctnnb1 after onset of CML in primary recipient mice resulted only in minor prolongation of survival. The animals died due to organ infiltration (primarily lung) by differentiated progenitor cells that were presumably not affected by Ctnnb1 deletion. However, in the BM and peripheral blood, a clear reduction of GFP+ cells was evident in heterozygous - and even more pronounced - in homozygous animals, indicating a potential effect on more immature cells and perhaps LSC. To determine if LSC were compromised and to investigate the efficacy of a combination treatment with IM, we used a serial BM-transplantation assay. Consistent with previously published data, we found that leukemia cells in the Sca-1/GFP+ fraction are able to re-establish CML in secondary recipient mice and thus contain LSC. IM treatment alone led to a relative increase of Sca-1/GFP+ LSC over time. However, we found a significant reduction of Sca-1/GFP+ bone marrow cells after inactivation of Ctnnb1 during IM-treatment. Mice transplanted with Ctnnb1^flox/flox ER-Cre+ CML cells had on average 0.025% Sca-1/GFP+ (total of 7,375 cells/mouse) cells in the bone marrow after treatment of the recipient mice with IM and tamoxifen, mice that received Ctnnb1^flox/wt ER-Cre+ CML had 0.036% Sca-1/GFP+ cells (total of 10,584 cells/mouse) and mice that received Ctnnb1^wt/wt ER-Cre+ CML had 0.075% Sca-1/GFP+ cells (total of 24,375 cells/mouse). As indomethacin has recently been shown to inhibit Ctnnb1 in hematopoietic and AML stem cells (Goessling, Cell 2009; Wang, Science 2010), we aimed to investigate the efficacy of indomethacin combined with IM. We were able to demonstrate a decrease in Ctnnb1 protein after indomethacin treatment of human cell lines and transformed murine BM cells by either immunoblot, flow-cytometry or immunofluorescence. Moreover, we were able to show synergy between indomethacin and IM co-treatment which produced almost complete abrogation of Ctnnb1 protein. IM/Indomethacin co-treatment of mice with CML decreased the Kit+Sca1⁺ fraction of the Lin-GFP+ bone marrow population to 0.92% (1799 cells/mouse) whereas treatment with IM alone was similar to untreated mice with 2.8% of the Lin-GFP+ population being Kit+Sca1+ (22,601 cells/mouse). Importantly, transplantation of the bone marrow from treated mice into tertiary recipients showed significantly prolonged survival of tertiary recipient mice that received bone marrow from donors treated with the combination therapy (IM/INDO) compared to IM/DMSO controls (p=0.0038*).

In summary, we demonstrate that inhibiting Ctnnb1 by genetic inactivation or drug treatment is an effective combination therapy with imatinib mesylate and targets CML leukemia stem cells. This provides strong evidence for the development and clinical use of β-catenin-targeted therapies in CML.