Targeting multiple kinase pathways in leukemic progenitors and stem cells is essential for improved treatment of Ph⁺ leukemia in mice

SRC Kinases Remain Active After Imatinib Inhibition of BCR-ABL Kinase Activity.

We tested the hypothesis that imatinib may not inactivate SRC kinases activated by BCR-ABL using a BCR-ABL-expressing pre-B cell line (ENU) (5). The cells were treated with or without imatinib. Compared with cells bearing the empty vector, Western blot analysis showed that SRC kinases were activated in cells expressing one of two forms of BCR-ABL (P190 and P210, which differ in molecular weight and are expressed predominantly in Ph⁺ ALL and CML, respectively), and imatinib treatment markedly inhibited BCR-ABL kinase activity but did not result in a decrease in SRC activation (Fig. 1A). These results indicate that, although imatinib was very effective in inhibiting BCR-ABL phosphorylation, it was unable to affect BCR-ABL-stimulated phosphorylation of SRC kinases. To demonstrate this finding further, we used the P190 or P210 form of BCR-ABL to transform mouse bone marrow (BM) cells. These cells were then treated with imatinib. Imatinib inhibited BCR-ABL phosphorylation, resulting in decreased phosphorylation of downstream signaling molecule CrkL, but did not affect BCR-ABL-stimulated phosphorylation of SRC kinases (Fig. 1B). These observations indicate that, in imatinib-treated BCR-ABL-expressing cells, SRC kinases are still active and may participate in cellular transformation by BCR-ABL.

Key Role of SRC Kinases in Malignant Transformation of B-Lymphoid Cells.

To investigate the role of SRC kinase activation in transformation of B-lymphoid cells and in the survival and proliferation of leukemic cells, we showed that v-SRC, an active form of SRC kinase, directly transformed B-lymphoid cells in vitro (Fig. 2A), suggesting that activated SRC alone is sufficient to stimulate aberrant proliferation of hematopoietic precursors. To examine whether inhibition of SRC kinases attenuates transformation of mouse BM cells by BCR-ABL, we used a dual SRC/BCR-ABL inhibitor dasatinib (Fig. 2B). We transformed mouse BM cells with the BCR-ABL-T315I mutant that is resistant to inhibition of BCR-ABL kinase activity by both imatinib (4, 12, 13) and dasatinib (7). This allowed us to dissociate the inhibitory effects on the BCR-ABL kinase vs. the effects on SRC kinases. Dasatinib reduced survival (Fig. 2C) and induced apoptosis (data not shown) of the leukemic cells, demonstrating that BCR-ABL-activated SRC kinases in imatinib-treated cells play a critical role in BCR-ABL-mediated transformation of B-lymphoid cells.

We further investigated the role of SRC kinase in B-ALL development using a mouse model of B-ALL (14). We treated mice with B-ALL induced by BCR-ABL-T315I with imatinib or dasatinib. Imatinib showed no therapeutic effect, whereas dasatinib significantly prolonged survival of the mice (P < 0.01) (Fig. 2D). Dasatinib inhibited SRC kinase activity in vivo (Fig. 2E). These results indicate that SRC kinases play a critical role in B-ALL development. However, targeting SRC kinases alone did not cure the disease (Fig. 2D), which may be due to the incomplete inhibition of SRC kinase activity in vivo at the dose of dasatinib used (Fig. 2E); a higher dose of dasatinib may further improve survival of the mice. To further support the role of SRC kinases in B-ALL development, we compared growth potential of BCR-ABL-transduced wild-type and Lyn^−/−Hck^−/−Fgr^−/− BM cells. We monitored the levels of pre-B leukemic cells expressing BCR-ABL (represented by GFP expression) over a 4-week time period in peripheral blood of mice receiving BCR-ABL-transduced wild-type or Lyn^−/−Hck^−/−Fgr^−/− BM cells. Levels of GFP⁺B220⁺ B-lymphoid leukemic cells were significantly lower in mice receiving BCR-ABL-transduced Lyn^−/−Hck^−/−Fgr^−/− BM cells than in those receiving BCR-ABL-transduced wild-type BM cells at all time points measured, although there was an initial increase in leukemic cells in mice receiving the transduced Lyn^−/−Hck^−/−Fgr^−/− BM cells (Fig. 2F). Strikingly, in mice receiving the transduced Lyn^−/−Hck^−/−Fgr^−/− BM cells, leukemic cells almost disappeared 5 weeks following B-ALL induction, whereas in the mice receiving the transduced wild-type BM cells, ≈45% of leukemic cells persisted (Fig. 2G).

Inhibition Solely of BCR-ABL Kinase Activity Without SRC Kinase Inhibition Is Insufficient for B-ALL Treatment.

Because SRC kinases are still active when BCR-ABL phosphorylation is inhibited by imatinib (Fig. 1), we examined whether inhibition of BCR-ABL kinase activity alone, with SRC kinase still active, is sufficient to control B-ALL. We treated mice with BCR-ABL-induced B-ALL with imatinib (which inhibits only BCR-ABL kinase activity) or with dasatinib (which inhibits both BCR-ABL and SRC kinase activity). Imatinib had a weak therapeutic effect on B-ALL (Fig. 3A), suggesting that inhibition solely of BCR-ABL kinase activity is insufficient to control the disease. By contrast, dasatinib maintained long-term survival of the mice with B-ALL induced by P190 or P210 BCR-ABL (Fig. 3A), indicating that both BCR-ABL kinase activity and SRC pathway must be targeted for treating this disease. The therapeutic effects of these two drugs on B-ALL correlated with reduced levels of GFP⁺ leukemic cells in peripheral blood of the treated B-ALL mice (Fig. 3B). The weak therapeutic effect of imatinib (Fig. 3A) cannot be attributed to an inability to inhibit BCR-ABL kinase activity in vivo, because imatinib significantly inhibited BCR-ABL phosphorylation to a similar extent compared with dasatinib in leukemic cells from pleural effusion of the treated B-ALL mice (Fig. 3C). To exclude the possibility that the better therapeutic effect of dasatinib over imatinib (Fig. 3A) could be attributed to the potency difference of these drugs on BCR-ABL kinase activity (7) but not to the additional anti-SRC effect of dasatinib (Fig. 2B), we treated BCR-ABL-transformed BM cells under Whitlock–Witte conditions with a SRC kinase inhibitor PP2 alone (15) (which did not inhibit BCR-ABL kinase activity at the concentrations used in this study), with imatinib alone, and with both PP2 and imatinib. Either drug alone inhibited proliferation of the cells, but both drugs together had a much stronger inhibitory effect (Fig. 3D). These results support the critical role of SRC kinases in B-ALL development.

Progression to Lymphoid Blast Crisis CML Requires Activation of SRC Kinases.

Chronic-phase CML advances to blastic phase. We genetically tested whether SRC kinases play a role in CML transition to lymphoid blast crisis using a serial transplantation assay (16). Mice were transplanted with BCR-ABL-transduced BM cells from either wild-type or Lyn^−/−Hck^−/−Fgr^−/− mice to induce CML, and BM cells from the CML mice were subsequently transferred into recipient mice. Mice receiving wild-type CML BM cells developed B-ALL, shown by GFP⁺/B220⁺ leukemic cells in peripheral blood, whereas none of the mice receiving Lyn^−/−Hck^−/−Fgr^−/− CML BM cells developed this disease (Fig. 3E). These results indicate that CML transition to lymphoid blast crisis requires SRC kinases. CML progression is associated with additional genetic changes, including mutations in the tumor suppressor genes INK4^a, pRB, and p53 (17–19). A recent study showed that Arf gene loss enhances the oncogenicity of imatinib and limits imatinib response to BCR-ABL-induced B-ALL in mice (20). To test whether SRC kinases are effective targets for B-ALL when tumor suppressor gene function is defective, BCR-ABL-transduced BM cells from p53-deficient mice were transplanted into lethally irradiated wild-type recipient mice followed by treatment with imatinib or dasatinib. Dasatinib was more effective than imatinib in suppressing p53-deficient pre-B leukemic cells, but these mice eventually died (Fig. 3F). These results suggest that loss of p53 function causes reduction of dasatinib response to BCR-ABL-induced B-ALL, although a significant degree of response remained.

Pro-B Leukemic Cells Are Identified as B-ALL Stem Cells, and Continuous Treatment with Dasatinib May Prevent Them from Developing into B-ALL.

We tested whether dasatinib could completely eradicate leukemic cells in B-ALL mice, leading to cures. Although dasatinib remarkably prolonged survival of B-ALL mice (Fig. 3A), a small percentage of GFP⁺ cells (<1%) remained in the peripheral blood of these mice even after 3 months of treatment (Fig. 4A). After treatment was stopped, B-ALL reappeared in most mice (Fig. 4A) within 1 month for P190BCR-ABL-induced and within 2 months for P210BCR-ABL-induced B-ALL. The relapsed B-ALL mice were treated again with dasatinib, and the percentage of GFP⁺ cells dropped again to <1% (Fig. 4A) and remained at this level during continuous drug treatment for 2 months (data not shown). After two rounds of treatment discontinuations, relapses, and retreatment, the mice remained sensitive to the next round of dasatinib therapy (data not shown). Still, a low level (<1%) of GFP⁺ cells persisted in the BM of the treated B-ALL mice, and these cells were capable of transferring the same disease to secondary recipient mice (data not shown). These results indicate that continuous administration of dasatinib could prevent these residual cells from developing into fatal B-ALL, although this compound did not completely kill the residual leukemic cells (Fig. 4A).

We identified the cell types of these residual GFP⁺ cells as B220⁺/CD43⁺ and B220⁺/CD43⁻ pro-/pre-B cells (Fig. 4B), and these progenitor leukemic cells may have acquired self-renewal capacity and function as B-ALL stem cells. To test this hypothesis, we sorted by FACS the B220⁺/CD19⁺/GFP⁺ cells from the BM of B-ALL mice, followed by transplantation of the cells into recipient mice. These mice developed B-ALL after 2 months, and leukemic cells in peripheral blood were CD19⁺/CD43⁺ pro-B cells (Fig. 4C). We conclude that CD19⁺/B220⁺/CD43⁺ pro-B cells expressing BCR-ABL can function as B-ALL stem cells. To support this conclusion, we transferred purified BCR-ABL-expressing CD19⁺/B220⁺/CD43⁺ pro-B cells into recipient mice; these cells induced leukemia and had potential to differentiate (data not shown).

Dasatinib Significantly Prolongs Survival of CML Mice but Does Not Eradicate CML Stem Cells.

Dasatinib is very effective in controlling B-ALL (Fig. 3A). We tested whether dasatinib also is effective at treating CML in mice. CML mice treated with dasatinib lived significantly longer than those treated with imatinib (Fig. 5A), which correlated with significantly lower numbers of BCR-ABL-expressing leukemic cells in peripheral blood (Fig. 5B) compared with placebo- or imatinib-treated mice. However, all dasatinib-treated CML mice eventually died of this disease (Fig. 5A), indicating that, like imatinib (21), this drug may not eradicate leukemic stem cells in CML mice. Because CML in mice originates from multilineage repopulating cells (14), we tested whether dasatinib kills BCR-ABL-expressing hematopoietic stem cells (HSCs) in vivo. We used BALB/c mice to induce CML because we used this strain in our therapeutic experiments in this study (Figs. 3A and 5A). We treated CML mice with a placebo, imatinib, or dasatinib for 14 days, starting from day 8 after CML induction by BCR-ABL, and found that BCR-ABL-expressing HSCs (GFP⁺CD34⁻c-Kit⁺Hoe⁻) existed in the side population (22) of BM cells from the imatinib- or dasatinib-treated CML mice (Fig. 5C). This observation indicates that neither imatinib nor dasatinib completely eradicates BCR-ABL-expressing HSCs, suggesting that neither drug will cure CML and that targeting at least one additional component of BCR-ABL-expressing HSCs is required for curing the disease. Because analysis of HSCs in side populations for the existence of the stem cells is not quantitative, we further analyzed BCR-ABL-expressing HSCs in dasatinib-treated CML mice (in B6 background) by identifying the GFP⁺Lin⁻c-kit⁺Sca-1⁺ population. Compared with placebo-treated mice, dasatinib reduced the numbers of BCR-ABL-expressing HSCs but failed to eradicate these cells completely in CML mice (data not shown), consistent with the finding using BALB/c mice (Fig. 5C). This biological effect on BCR-ABL-expressing HSCs does not support the possibility that inability of dasatinib to completely eradicate BCR-ABL-expressing HSCs may be attributed to the failure of dasatinib to access the stem cells, because we detected inhibition of intracellular BCR-ABL phosphorylation by dasatinib in the stem cells (Fig. 5D). The inability of dasatinib to cure CML mice is not attributed to the appearance of BCR-ABL-T315I clone in the mice, because CML mice treated with dasatinib for ≈3 months contained >40% of GFP⁺Gr-1⁺ cells, among which there were large numbers of stem cells (Fig. 5E), and sequencing analysis of isolated genomic DNA from BM cells of these mice did not show T315I mutation in the BCR-ABL kinase domain (data not shown). The failure of imatinib to eradicate BCR-ABL-expressing HSCs is not related to c-kit function, because both imatinib and dasatinib inhibit c-kit (23). These results suggest that inhibition of BCR-ABL kinase activity alone is insufficient to eradicate CML stem cells.

To identify CML stem cells, we tested whether BCR-ABL-expressing HSCs function as the stem cells. We first sorted C57BL/6 (B6) BM cells transduced with BCR-ABL retrovirus into two separate populations: Sca-1⁻ and Sca-1⁺. These two populations of cells were transferred, respectively, into B6 mice. Only the mice receiving BCR-ABL-transduced Sca-1⁺ cells developed and died of CML, diagnosed by detecting GFP⁺ myeloid cells (Gr-1⁺) in the peripheral blood of the mice (Fig. 6A). This result suggests that early BM progenitors contain CML stem cells. To narrow down the specific cell lineages that function as CML stem cells, HSCs (Lin⁻c-kit⁺Sca-1⁺) were sorted out from BM cells transduced with BCR-ABL retrovirus, followed by transfer into recipient mice. The mice developed and died of CML (data not shown). To confirm definitively that BCR-ABL-expressing HSCs are CML stem cells, we isolated BM cells from primary CML mice and sorted out the BCR-ABL-expressing HSCs (GFP⁺Lin⁻c-Kit⁺Sca-1⁺) by FACS. The sorted cells were transferred into recipient mice, and the mice developed and died of CML (Fig. 6B), indicating that BCR-ABL expressing HSCs function as CML stem cells.

Cell Lines.

The BaF/3 pre-B and ENU cell lines were grown in RPMI medium 1640 containing 10% FCS, 10% WEHI medium, and 50 μM 2-mercaptoethanol.

Whitlock–Witte Culture.

BM cells were transduced with the BCR-ABL retrovirus and cultured as described previously (5).

Antibodies and Western Blot Analysis.

Antibodies against phosphotyrosine, c-ABL, CrkL, β-actin, and the SRC kinases were purchased from Santa Cruz Biotechnology (Santa Cruz, CA); antibodies against c-Abl-Y412, phospho-CrkL, and SRC-Y416 were purchased from Cell Signaling Technology (Danvers, MA). Protein lysates were prepared by lysing cells in RIPA buffer, and immunoprecipitation and Western blotting were carried out as described previously (27).

BM Transduction/Transplantation.

The retroviral vector MSCV-IRES-eGFP carrying the BCR-ABL cDNA was used to make virus stock as described previously (14). Four- to 10-week-old wild-type BABL/c or C57BL/6 (The Jackson Laboratory) and homozygous SRC triple gene knockout (Lyn^−/−Hck^−/−Fgr^−/−) mice (5) were used for leukemogenesis experiments (14, 28).

Flow Cytometry.

Hematopoietic cells were collected from the diseased mice and analyzed by FACS analysis as described previously (5).

Drug Treatment.

Dasatinib was dissolved in 80 mM citric acid (pH 2.1) to make 10 mg/ml stock solution and then diluted to 1 mg/ml with 80 mM citric acid (pH 3.1) for use. Imatinib was dissolved in water directly at a concentration of 10 mg/ml. The drugs were given orally in a volume of <0.5 ml by gavage twice a day, at 10 mg per kilogram of body weight per dose for dasatinib and 100 mg per kilogram of body weight for imatinib, beginning at 8 days after BM transplantation and continuing until the morbidity or death of the leukemic mice.