Novel Oncogenic Mutations of CBL in Human Acute Myeloid Leukemia That Activate Growth and Survival Pathways Depend on Increased Metabolism*

Acute myeloid leukemia (AML) is characterized by multiple mutagenic events that affect proliferation, survival, as well as differentiation. Recently, gain-of-function mutations in the α helical structure within the linker sequence of the E3 ubiquitin ligase CBL have been associated with AML. We identified four novel CBL mutations, including a point mutation (Y371H) and a putative splice site mutation in AML specimens. Characterization of these two CBL mutants revealed that coexpression with the receptor tyrosine kinases FLT3 (Fms-like tyrosine kinase 3) or KIT-induced ligand independent growth or ligand hyperresponsiveness, respectively. Growth of cells expressing mutant CBL required expression and kinase activity of FLT3. In addition to the CBL-dependent phosphorylation of FLT3 and CBL itself, transformation was associated with activation of Akt and STAT5 and required functional expression of the small GTPases Rho, Rac, and Cdc42. Furthermore, the mutations led to constitutively elevated intracellular reactive oxygen species levels, which is commonly linked to increased glucose metabolism in cancer cells. Inhibition of hexokinase with 2-deoxyglucose blocked the transforming activity of CBL mutants and reduced activation of signaling mechanisms. Overall, our data demonstrate that mutations of CBL alter cellular biology at multiple levels and require not only the activation of receptor proximal signaling events but also an increase in cellular glucose metabolism. Pathways that are activated by CBL gain-of-function mutations can be efficiently targeted by small molecule drugs.

Acute myeloid leukemia (AML) is characterized by multiple mutagenic events that affect proliferation, survival, as well as differentiation. Recently, gain-of-function mutations in the ␣ helical structure within the linker sequence of the E3 ubiquitin ligase CBL have been associated with AML. We identified four novel CBL mutations, including a point mutation (Y371H) and a putative splice site mutation in AML specimens. Characterization of these two CBL mutants revealed that coexpression with the receptor tyrosine kinases FLT3 (Fms-like tyrosine kinase 3) or KIT-induced ligand independent growth or ligand hyperresponsiveness, respectively. Growth of cells expressing mutant CBL required expression and kinase activity of FLT3. In addition to the CBL-dependent phosphorylation of FLT3 and CBL itself, transformation was associated with activation of Akt and STAT5 and required functional expression of the small GTPases Rho, Rac, and Cdc42. Furthermore, the mutations led to constitutively elevated intracellular reactive oxygen species levels, which is commonly linked to increased glucose metabolism in cancer cells. Inhibition of hexokinase with 2-deoxyglucose blocked the transforming activity of CBL mutants and reduced activation of signaling mechanisms. Overall, our data demonstrate that mutations of CBL alter cellular biology at multiple levels and require not only the activation of receptor proximal signaling events but also an increase in cellular glucose metabolism. Pathways that are activated by CBL gain-of-function mutations can be efficiently targeted by small molecule drugs.
Molecular mechanisms that directly contribute to myeloid transformation in acute myeloid leukemia (AML) 2 are not well understood, making it difficult to devise targeted therapies. Mutational activation of receptor-tyrosine kinases (RTKs) is a common event in AML and is thought to significantly contribute to the disease phenotype. Activating mutations in the FLT3 and KIT RTKs are found in about one-third or one-tenth of AML cases, respectively (1). Deregulation of RTK pathways is likely a late event in leukemia development and may contribute to the severity of the disease rather than as a disease-initiating event (2). The role of cellular forms of FLT3, KIT, or other RTKs in AML is not known. Normally, RTKs are transiently activated through ligand binding, which can trigger signaling pathways that stimulate, for example, growth and survival. Ligand-induced activation of these RTKs will also trigger negative regulatory mechanisms that lead to the termination of the signaling mechanism. In cells expressing oncogenic RTKs, this process is likely disrupted, leading to prolonged or aberrant activation.
RTKs are negatively regulated in vivo in part by CBL (3), which belongs to a family of related proteins that also include CBL-b and CBL-c (4,5). All CBL family members contain ubiquitin ligase (E3) activity (3). CBL shares the closest homology to CBL-b and both proteins contain a tyrosine kinase-binding (TKB) domain, which in turn is composed of a four-helix bundle, an EF-hand, and an SH2-like phosphotyrosine-binding domain; a RING finger domain; a ubiquitin-associated threehelix bundle domain (UBA); and tyrosine residues that can be inducibly phosphorylated as well as a proline-rich domain known to be involved in protein-protein interactions (3). CBL-b appears to be transcriptionally regulated and its expression is increased during myeloid differentiation (5). Despite their similar structure and function, CBL and CBL-b can be associated with distinct signaling pathways in the same cellular context (6). CBL-c is about half the size of CBL and does not contain the C-terminal proline-rich domain, the UBA domain, and lacks several phosphorylation sites as well (4). The TKB domain in CBL proteins enables them to recognize phosphorylated target proteins, typically tyrosine kinases. Also, ubiquitin-conjugating enzymes (E2s) are recruited through the RING finger domain for mono-or polyubiquitination of CBLassociated partners, thus effectively regulating their stability or function (3). In addition, the C-terminal regions of CBL and CBL-b are also postulated to participate in RTK endocytosis through the interaction with CIN85 and related proteins (7)(8)(9).
Recently, several gain-of-function mutations of CBL have been identified in myeloid malignancies, including juvenile and chronic myelomonocytic leukemias, AML, myelodysplastic syndromes, and myeloproliferative neoplasms (10 -17). These mutations are frequently found in the linker region between RING finger and TKB domains and can be associated with acquired 11q uniparental disomy (12)(13)(14)(15). Mutations in this region lead to a loss of the E3 activity of CBL (11). However, this altered function has not been directly linked to the transforming activity of CBL and it has not been excluded whether it constitutes a potential epiphenomenon. In mice, disruption of the Cbl gene has not been reported to lead to myeloid malignancies but Cbl-regulated pathways are hyperresponsive to stimuli, leading to phenotypes including lymphoid hyperplasia and primary splenic extramedullary hematopoiesis (18,19). Thus, the molecular mechanisms that lead to transformation by oncogenic Cbl may either affect pathways different from its E3 activity or complement the loss of its enzymatic activity. We sought to identify CBL mutations in AML and identify receptor proximal events that regulate their transforming activity.
In this study, we have identified four novel mutations in the CBL sequence from AML patient specimens. Two mutations in the linker region of CBL (hereafter referred to as mutant CBL) between the TKB and RING finger domains were further characterized. Expression of mutant CBL in FLT3 expressing BaF3 cells was found to lead to factor-independent growth, which correlated with the activation of growth and survival pathways, involving STAT5, Akt, and Rho family GTPases. In addition, CBL-transformed cells showed elevated levels of intracellular ROS and increased glucose uptake. Inhibiting the active glucose metabolism in mutant CBL-transformed cells reduced dependent growth and activation of signaling pathways, suggesting potential targets for therapeutic approaches.

EXPERIMENTAL PROCEDURES
Nucleotide Sequence Analysis of CBL-To identify novel CBL mutations, cDNAs from 43 AML patients and 10 healthy individuals were used to amplify the linker sequence. For the analysis of genomic mutations, exons 2 to 16 of CBL were amplified by PCR using Qiagen PCR reagent, as per the manufacturer's suggested protocol (Qiagen, Valencia, CA) and primers listed in supplemental Table S1. The mutational status was identified by sequencing using a standard fluorescent dye method. Sequencing results were analyzed for mutations using Mutation Surveyor version 2.61 (Softgenetics, State College, PA) and con-firmed by sequencing of the reverse coding strand. cDNA samples were obtained from the University of Mannheim, and patient specimens for the preparation of genomic DNA from the University of Pennsylvania and the University of Chicago with informed consent and approval by the respective Institutional Review Boards.
Measurement of Intracellular Reactive Oxygen Species-The relative intracellular levels of ROS were determined by flow cytometry using the cell-permeable redox-sensitive fluorochrome DCF-DA (2Ј,7Ј-dichlorofluorescein diacetate; Calbiochem) similar to previously described methods (23). Briefly, cells (1 ϫ 10 6 ) were incubated with DCF-DA (20 M) in phosphate-buffered saline (PBS, Mediatech, Manassas, VA) for 5 min at 37°C, and subsequently washed twice in cold PBS. Cells were analyzed by flow cytometry using a FACSCanto II flow cytometer and the FACSDiva cytometry analysis software (BD Biosciences, San Jose, CA).
Cell Cycle Analysis-BaF3.FLT3 cells expressing CBL wildtype were growth factor-deprived for 24 h and compared with cells expressing mutant forms of CBL maintained in the absence of growth factors. Cells were fixed with 70% (v/v) eth-anol in PBS, incubated on ice for 30 min, and subsequently treated with RNase (10 g of RNase/ml; Qiagen) for 20 min at 37°C and stained with propidium iodide (5 g/ml; Sigma) at 4°C for 10 min. Cell cycle distribution was determined by flow cytometry and analyzed using the ModFit LT (Verity Software House, Topsham, ME) software.
Apoptosis Assays-The induction of apoptosis was determined in growth factor-deprived cells using the Annexin-V-Fluos staining kit (Roche Diagnostics, Indianapolis, IN) according to the manufacturer's directions. Annexin V-positive staining was determined by flow cytometry with the FACSDiva analysis software. Activity of the pro-apoptotic caspases 3/7 was determined in growth factor-deprived cells using the Caspase-Glo 3/7 assay kit (Promega, Madison, WI) and quantified on a Luminoskan Ascent luminometer (Thermo Labsystems, Franklin, MA).
Targeted Knockdown Using siRNA-BaF3.FLT3 cells expressing CBL mutants were transiently transfected with 5 g of SmartPool siRNA against Rac1, RhoA, or Cdc42 (Dharmacon, Lafayette, CO) and compared with cells transfected with scrambled control siRNA (Dharmacon). The siRNA was transfected using the Nucleofector device with 100 l of Nucleofector solution V (Amaxa, Gaithersburg, MD) according to the manufacturer's instructions. Protein knockdown was confirmed by immunoblotting 24 h after transfection and cell growth was determined by trypan blue exclusion 48 h after transfection.

Identification of Novel CBL Mutations in AML Patient
Specimens-Gain-of-function mutations in the linker sequence of CBL have been first described in AML patient specimens (10,11). In an effort to identify novel CBL mutations we performed DNA sequencing on AML patient specimens. An initial screen identified two novel CBL mutations in the linker sequence out of 43 AML cDNA samples, suggesting a mutation rate of 5% (2/43). The first mutation included a putative heterozygous splice site mutation (intron 7/exon 8) resulting in a 6-bp insertion (TCAAAG), predicting the insertion of serine and lysine (SK) at positions 366/367 of the coding sequence (Ins(SK366)). The second mutation was a homozygous single base substitution, resulting in a tyrosine to histidine substitution at position 371 of the coding sequence (Y371H) (Fig. 1A). Both mutations were located in the ␣ helical structure within the linker sequence of CBL, upstream of the RING finger, which is required for its E3 function (Fig. 1B). These variations were not found in samples from normal donors or in GenBank TM and other searchable databases. Next, genomic DNA sequencing was performed to identify potential mutations outside of the linker sequence in 37 AML specimens, an approach that does not allow the identification of putative splice site mutations. Two additional single base mutations were identified, suggesting a mutation rate of 5% (2/37). One mutation was a homozygous K287R substitution (exon 5) in close proximity to the phosphotyrosine-binding pocket of the SH2-like domain and the other a heterozygous R499L mutation (exon 10) adjacent to the GRB2 SH3 domain binding site ( Fig.  1B and supplemental Fig. S1).
CBL Mutants Are Transforming in BaF3 Cells Expressing FLT3 and KIT-Activating mutations in the linker sequence of CBL have been associated with factor-independent growth in cell line models, when co-expressed with RTKs (10, 11). Therefore, the transforming activity of mutant CBL was tested using the IL-3-dependent BaF3 cell line, engineered to express the functional FLT3 receptor. BaF3.FLT3 cells expressing CBL mutants were compared with wild-type (WT) CBL-expressing cells. As a positive control, BaF3.FLT3 cells with a transforming mutant of CBL, harboring a deletion of Tyr 371 in the linker sequence (CBL.⌬Y371), were used (24). The growth of these cells was determined in a 4-day culture in the absence of IL-3. Consistent with this model, we did not find factor-independent growth of cells expressing the CBL mutants in the absence of FLT3 in BaF3 cells (not shown). Whereas BaF3.FLT3 cells expressing WT CBL did not grow, the cell number increased 26-fold for CBL.Ins(SK366) and 17-fold for the CBL.Y371Hexpressing cells. This was comparable with the 22-fold increase of the CBL.⌬Y371-expressing BaF3.FLT3 cells ( Fig. 2A, left  panel).
To further confirm whether FLT3 kinase activity is required for the transforming activity of mutant CBL, we treated cells with the FLT3 kinase inhibitor midostaurin (25). This small molecule drug led to a dose-dependent reduction of cell growth in a 3-day culture of BaF3 cells expressing the constitutively active FLT3 with an internal tandem duplication (FLT3-ITD), with an IC 50 of 50 nM (not shown). Midostaurin (50 nM) also led to a reduction in cell growth in BaF3.FLT3 cells expressing CBL.Ins(SK366) (46.1 Ϯ 10.0%), CBL.Y371H (45.1 Ϯ 8.5%), or CBL.⌬Y371 (54.1 Ϯ 6.7%), relative to untreated cells. The reduction in cell growth could be significantly alleviated in the presence of IL-3, and was comparable with growth of cells expressing WT CBL in the presence of midostaurin (Fig. 2B).
To determine whether the transforming activity of mutant CBL can be induced by RTKs other than FLT3, we also expressed mutant CBL with the KIT RTK in BaF3 cells. Unlike BaF3 cells co-expressing CBL mutants with FLT3, the coexpression of CBL mutants with KIT did not result in a growth factor-independent phenotype. However, in the presence of the KIT ligand, stem cell factor (SCF), mutant CBL-expressing cells displayed hyper-responsiveness to the growth factor in a dosedependent manner, when compared with WT CBL-expressing cells (Fig. 2C, left panel). At 10 ng/ml of SCF the growth of CBL.Ins(SK366) and CBL.Y371H cells was increased 6.8-and 5.0-fold, respectively, when compared with cells expressing WT CBL. Cells expressing the ⌬Y371 mutant displayed hyperresponsiveness as well, with a 5.4-fold increase in cell growth compared with WT CBL cells in response to 10 ng/ml of SCF. In control experiments the protein expression of KIT, CBL, HAtagged CBL, and actin was determined and compared with untransfected BaF3 cells (Fig. 2, A and C, right panels).
CBL Mutants Prevent Apoptosis, Promote Cell Cycle Progression, and Increase ROS-To determine the effect of IL-3 withdrawal on apoptosis in mutant CBL-expressing cells, we measured the total percentage of Annexin V-positive cells in response to growth factor withdrawal. BaF3 cells co-expressing FLT3 and CBL mutants were grown in the absence of IL-3 and compared with WT CBL-expressing cells that were IL-3 deprived for 24 h. Whereas in WT CBL-expressing BaF3.FLT3 cells the levels of Annexin V-positive cells were 5.7 Ϯ 1.1% of the total population, the percentage of apoptotic cells decreased in the presence of CBL.Ins(SK366) (1.4 Ϯ 0.4%) and CBL.Y371H (1.5 Ϯ 0.5%) as well as CBL.⌬Y371 (1.3 Ϯ 0.2%) (Fig. 3A, left panel). We further confirmed reduced apoptosis by measuring relative activities of caspases 3/7, which are involved in the apoptotic pathway. Similar to the above data, reduced caspase 3/7 activity (fold-change relative to WT CBL-expressing cells) was observed in the CBL.Ins(SK366) (0.12 Ϯ 0.01) and CBL.Y371H (0.20 Ϯ 0.02) as well as CBL.⌬Y371 (0.14 Ϯ 0.02) expressing cells (Fig. 3A, right panel).
Next, cell cycle distribution was determined in BaF3.FLT3 cells in response to mutant CBL. Cells expressing WT CBL were compared with cells expressing CBL mutants in the absence of IL-3. The majority of cells expressing WT CBL accumulated in the G 1 phase (69.2 Ϯ 9.3%), consistent with a G 1 cell cycle arrest. The proportion of cells in the G 1 phase was reduced in CBL.Ins(SK366) (52.4 Ϯ 1.9%), CBL.Y371H (53.9 Ϯ 2.3%), or CBL.⌬Y371 (50.2 Ϯ 5.9%) cell lines, and consequently the relative number of cells in S increased (Fig. 3B). Also, an additional 5.0 Ϯ 2.8% of WT CBL-expressing cells were in the sub-G 1 phase, which is consistent with increased apoptosis, compared with mutant CBL-expressing cells that had a lower percentage of additional cells in sub-G 1 (Ͻ0.5%). These data demonstrate that the mutated forms of CBL examined here  OCTOBER 15, 2010 • VOLUME 285 • NUMBER 42 promote cell growth through reduced apoptosis and increased cell cycle progression.

Novel CBL Mutations in AML and Effects on Glucose Metabolism
We and others have previously shown that transforming tyrosine kinase activity leads to an increase of ROS in hematopoietic cells and this effect was shown to be crucial for cell growth and tyrosine kinase activity itself. We therefore sought to determine whether mutant CBL could increase ROS in BaF3.FLT3 cells (Fig. 3C). A significant increase in ROS was found in CBL.Ins(SK366) (59.9 Ϯ 23.9%), CBL.Y371H (71.6 Ϯ 13.2%), and CBL.⌬Y371 (67.6 Ϯ 11.5%) expressing cells compared with WT CBL-expressing cells (p Ͻ 0.05, n ϭ 3). Interestingly, comparable changes were found in cells transformed by the FLT3 oncogene with internal tandem duplication (not shown). Overall, these results are consistent with a role of ROS in transformation of hematopoietic cells.
CBL Mutants Alter Akt and STAT5 Phosphorylation-Theabovedataimplied that FLT3 kinase activity is required for transforming activity of mutant forms of CBL. We therefore initially asked whether the FLT3 receptor is phosphorylated at its activation site (Tyr 842 ) and other phosphorylation sites (Tyr 955 and Tyr 969 ). We also determined the activation of common growth (STAT5) and viability (Akt) pathways. Cells expressing CBL WT were compared with mutant CBL-expressing cells. We found that FLT3 (Tyr 842 , Tyr 955 , and Tyr 969 ), Akt (Ser 473 ), and STAT5 (Tyr 694 ) were phosphorylated in mutant, but not in the WT CBL cells (Fig. 4A, left panels). Next, the FLT3 inhibitor midostaurin was used to determine the requirements of the FLT3 kinase activity for phosphorylation. We compared BaF3.FLT3 cells expressing CBL mutants to those expressing the transforming CBL⌬Y371 mutant. Consistent with an important role of FLT3 in mutant CBL transformation in BaF3.FLT3 cells, phosphorylation of the RTK was dependent on the FLT3 kinase activity itself. FLT3 kinase activity was found to be required for phosphorylation at the activation sites of Akt and STAT5 (Fig. 4A, right  panels). Next, the effect of the antioxidant NAC on Akt and STAT5 phosphorylation in mutant CBL-expressing cells was determined (Fig.  4B). Cells were left untreated or treated with 20 mM NAC, and protein phosphorylation was measured 2 days later. The phosphorylation of Akt and STAT5 was abolished in the presence of NAC. Interestingly, the same NAC concentration is also sufficient to reduce cell growth and expression of the STAT5 target Pim-1 in cells containing oncogenic JAK2-V617F (22,23). The results suggest that a high oxidative state is important to maintain phosphorylation of Akt and STAT5 in cells containing mutant CBL. Because CBL is also a common substrate of tyro- sine kinases, we asked whether FLT3 kinase activity regulates CBL phosphorylation. CBL was only marginally phosphorylated (Tyr 674 , Tyr 700 , and Tyr 774 ) in WT CBL cells compared with mutant CBL cells (Fig. 4C, left panels). Similar to the above experiments, CBL phosphorylation was reduced in response to midostaurin in cells expressing mutant CBL (Fig. 4C, right panels).
The Small GTP-binding Proteins Rho, Rac, and Cdc42 Are Required for Transformation by Mutant CBL-Phosphorylation on Tyr 700 and Tyr 774 on CBL allows for the recruitment of nucleotide exchange factors. Vav1, a guanine nucleotide exchange factor for Rac and Cdc42, can bind directly to Tyr 700 through its SH2 domain (26) and the SH2 domains of the Crkl and Crk adapter proteins can bind to Tyr 700 and Tyr 774 (27,28). The Crk family proteins can further recruit nucleotide exchange factors, including C3G, SOS, DOCK, and others (29). Small GTPases are target proteins of these nucleotide exchange factors and have been implicated in a variety of cellular functions that contribute among others to cell growth (30). We asked whether small molecule drugs targeting Rac (NSC23766), Rho (Y27632), and Cdc42 (Secramine A) would alter cell growth in CBL-transformed cells. Our experimental data in Figs. 2A and 3A already suggest that WT CBL expressing cells are not transformed to growth factor independence. Therefore, cells expressing mutant CBL were compared with those expressing the transforming tyrosine kinase oncoproteins FLT3-ITD, BCR-ABL, and JAK2-V617F. Consistent with a role of these GTPases in transformation by mutant CBL, treatment of mutant CBL-expressing BaF3.FLT3 cells led to a dose-dependent reduction of cell growth (Fig. 5A), which was comparable with effects on cells expressing oncogenic tyrosine kinases (Fig. 5B). The IC 50 for inhibition of cell growth was between 100 and 300 M for NSC23766, 10 and 30 M for Y27632, and 1 and 3 M for Secramine A for mutant CBL-expressing cells. As it is true for most, if not all small molecule drugs, there are likely additional targets that are inhibited at the effective concentrations. To further verify the involvement of these small GTPases, we performed targeted knockdown of Rac1, RhoA, and Cdc42 in mutant CBL-transformed cells. Transient transfection with pools of siRNA led to a measurable reduction of targeted proteins (Fig. 5C, left panel). The knockdown of the GTPases was similar in the three mutant CBL-expressing BaF3.FLT3 cell lines and correlated with reduced cell growth for targeting of Rac (51.0 -63.1% of control), Rho (43.6 -56.2% of control), and Cdc42 (36.9 -63.8% of control) (Fig. 5C, right panel). Thus, these results confirm the findings obtained from cells treated with small molecule drugs, implicating Rho GTPases in the regulation of cell growth in mutant CBL-transformed cells.
Overall these data suggest a mechanism whereby mutant CBL is associated with activation of the FLT3 kinase activity and downstream signaling targets, including STAT5, Akt, and other signaling molecules related to FLT3 signaling, including the phosphorylation of CBL itself.
Glucose Metabolism Is Required for Transformation by CBL Mutants-Increased glucose metabolism is one of the characteristics of proliferating cancer cells. To evaluate changes in glucose uptake, we compared mutant CBL to WT CBL-expressing BaF3.FLT3 cells using the fluorescent glucose analog 2-NBD-glucose, which is transported into the cell by glucose transporters. Transformation by mutant CBL was found to significantly correlate with increased glucose uptake compared with WT CBL-expressing cells (24.3-26.3% increase, n ϭ 3, p Ͻ 0.05) (Fig. 6A). Similarly, we observed an increased uptake in 2-NBD-glucose in BaF3.FLT3-ITD compared with the parental cell line (not shown). To determine the significance of elevated glucose uptake, we measured cell growth in response to the hexokinase inhibitor 2-deoxyglucose, which inhibits glucose from entering the glycolytic pathway. Treatment of BaF3.FLT3 cells transformed by mutant CBL with this small molecule drug was found to inhibit cell growth in a dose-dependent manner and was maximal at around 2 g/liter of 2-deoxyglucose, which is also the concentration of glucose in culture medium (Fig. 6B). Aerobic glucose metabolism is linked to mitochondrial electron transport, which is thought to lead to the production of ROS (31). This chain of events was also observed in BaF3.FLT3 cells expressing mutant CBL in response to hexokinase inhibition by 2-deoxyglucose (2 g/liter). Treatment with 2-deoxyglucose abolished the above observed (Fig. 3C) increase in ROS in cells transformed by mutant CBL. The relative ROS levels in cells containing CBL.Ins(SK366) (93.1 Ϯ 11.3%), CBL.Y371H (93.7 Ϯ 12.2%), and CBL.⌬Y371 (90.9 Ϯ 11.9%) were compara-  OCTOBER 15, 2010 • VOLUME 285 • NUMBER 42 ble with the WT CBL control (Fig. 6C). ROS and additional signaling events regulate the balance between altered tyrosine kinases and protein-tyrosine phosphatase activities in hematopoietic cells, thus effectively modulating phosphotyrosine-dependent signaling (31). We found that 2-deoxyglucose treatment was sufficient to lead to reduced FLT3 phosphorylation in mutant CBL-transformed cells (Fig. 6D). Thus mutant CBL is likely to increase cellular metabolism and FLT3 signaling through an interdependent mechanism.

DISCUSSION
Mutations of CBL in the linker region are associated with transformation in a subset of myeloid malignancies (10 -17).
We have further defined molecular mechanisms involved in transformation by CBL.(SK366) and CBL.Y371H and identified crucial signaling pathways that are required for CBL-driven factor independent growth in cell line models. We found FLT3 to be phosphorylated on Tyr 842 (activation loop), Tyr 955 and Tyr 969 (both in the carboxylterminal region) only in the presence of mutant CBL, suggesting that mutant CBL is involved in the ligand independent activation of the RTK. It should be emphasized that our data do not demonstrate phosphorylation of FLT3 by mutant CBL through direct interaction, but this would be an intriguing possibility. Although the presence of mutant CBL correlates with increased FLT3 phosphorylation, the detailed molecular mechanism involved in this process remains unclear and warrants further investigation. This putative mechanism is unlikely to be simply a result of altered CBL E3 activity, because ligand independent RTK activation is not known to be associated with Cbl gene disruption (18,19). Activation of FLT3 by CBL led to signaling through growth (STAT5) and survival (Akt) pathways, required expression and function of the Rho-family small GTPases (RhoA, Rac1, and Cdc42) and in turn increased CBL phosphorylation. Even though our data suggest that the small molecule FLT3 inhibitor midostaurin (25) can block the increased phosphorylation of AKT, STAT5, and CBL in mutant CBL-expressing BaF3.FLT3 cells, this does not indicate that FLT3 is directly involved in the phosphorylation process, given that this inhibitor may have additional targets. It is thus possible that other kinases participate in the phosphorylation of these proteins, particularly because none of these signaling proteins are known to be common substrates of FLT3. The functional role of CBL phosphorylation in the regulation of these proteins will require further evaluation. Nevertheless, there is considerable overlap with ligand-activated signaling through the FLT3 receptor (32), hinting at growth factor receptors as primary targets of mutant CBL. A model consistent with our findings would involve the ligand independent or hyperresponsive stimulation of growth factor receptors by mutant CBL followed by the activation of growth and viability pathways, similar to active forms of FLT3 and KIT in AML. Constitutive activation of FLT3 and KIT receptor tyrosine kinases is already associated with transformation in AML (1) and mutant CBL would therefore provide an alternative mechanism to stimulate these pathways.
Similar to previously published CBL linker region mutations, we found that mutant CBL does not lead to factor-independent growth when expressed in a factor-dependent cell line in the absence of an RTK (11,14). The transforming activity of CBL did not only require the expression of the FLT3 RTK but also its enzymatic activity. Midostaurin specifically reduced cell growth of BaF3 cells expressing both CBL.(SK366) and CBL.Y371H mutants with an IC 50 of 50 nM. Induction of cell growth in response to mutant CBL correlated with reduced apoptosis and an increase of cells in S phase with reduced G 1 phase. G 1 /S phase transition in factor-dependent hematopoietic cells is regulated in part by D cyclins (33) and it would be interesting to determine how mutant CBL regulates this path-way in FLT3 containing cells. Unlike FLT3, the KIT RTK did not cause complete factor independence in response to mutant CBL, but rather displayed hyper-responsiveness to ligand stimulation. In contrast, recent data by Sanada et al. (13) suggest that loss of WT CBL may be required for the transforming potential of mutant CBL in vivo. It was shown that in the presence of WT CBL, linker region mutant forms of CBL failed to fully transform murine hematopoietic (lineage Ϫ stem-cell antigen 1 ϩ KIT ϩ ) stem cells (HSC). Nevertheless, in the absence of CBL, cells were hyperproliferative in response to SCF, thrombopoietin, IL-3, and FLT3 ligand. It is not known whether these results are also applicable to human HSC but there are intrinsic differences to murine HSC. Most notably, FLT3 expression in murine HSC appears to be a determinant for the loss of self-renewal capacity in the myeloid lineage, which is different from their human counterparts (34,35). Additionally, oncogenic v-Cbl, a Gag-Cbl fusion protein that lacks 60% of the C terminus of c-Cbl, causes pre-B cell lymphomas and some myeloid leukemias in mice (36 -38). The fact that v-Cbl transforms in the presence of endogenous c-Cbl would argue against a strict requirement for loss of CBL levels to develop tumorigenic potential for mutant CBL. It is likely that mutant CBL also signals through other RTKs, but there is accumulating evidence that FLT3 has a functional role in this pathway and thus data obtained with murine HSC have to be interpreted cautiously.
Overall, there appears to be a causal relationship between expression of mutant CBL and phosphorylation of FLT3 on its activation site as well as regulation of FLT3-related signaling pathways. It cannot be excluded that another RTK can fully substitute for active FLT3 in transformation by mutant CBL in hematopoietic cells. However, our results suggest that in the case of FLT3, mutant CBL not only activates the RTK but in turn FLT3-induced metabolism is required for mutant CBLinduced FLT3 activation. Our data do not define the direct inter-relationship between cellular metabolism and biological activities. It is possible that the observed increase in cellular glucose metabolism may be one of the consequences of increased growth and viability due to CBL mutation. Inhibition of glucose metabolism may therefore exert a more generalized FIGURE 6. Glucose metabolism is required for transformation by CBL mutants. Growth factor-deprived BaF3.FLT3 cells were used in combination with CBL wild-type (WT) or mutant forms of CBL, as indicated. A, 2-NBD-glucose uptake was measured by flow cytometry in CBL-transformed cells and compared with cells expressing WT CBL. B, changes in cell growth in response to 2-deoxyglucose were determined after 24 h (means Ϯ S.E., n ϭ 4). C, ROS levels were measured by DCF-DA staining in response to 2-deoxyglucose (means Ϯ S.E., n ϭ 3). D, FLT3 as well as tyrosine-phosphorylated forms were detected by immunoblotting in response to 2-deoxyglucose (DOG), as indicated. OCTOBER 15, 2010 • VOLUME 285 • NUMBER 42 effect on cell physiology. On the other hand, it is also possible that increased metabolism fuels increased biological activities. Nevertheless, our data show that mutant CBL is required for increased glucose metabolism, cell growth, and viability.

Novel CBL Mutations in AML and Effects on Glucose Metabolism
Metabolic pathway genes have only recently been recognized as early targets for transformation in cancer stem cells associated with AML (39). We have shown that mutant CBL led to increased 2-NBD-glucose uptake, which is an indication for increased glucose uptake and glucose metabolism. This is also consistent with elevated levels of ROS and activation of the STAT5 and PI3K/Akt pathways in CBL-transformed cells. In hematopoietic cells, increased ROS are required for growth and survival (31). ROS can contribute to cell cycle progression by themselves but may not be sufficient to sustain cell growth (40). Our previous data link aberrant kinase activity in myeloid malignancies to STAT5 and PI3K pathway-dependent ROS activation (41). In cells transformed by the BCR-ABL tyrosine kinase, increased ROS depend on PI3K activity as well as a hyperactive glucose metabolism and are mainly of mitochondrial origin (41). Also, excess production of ROS is thought to cause oxidative DNA lesions that have the potential to create disease promoting mutagenic events (42). Thus, in addition to activating growth signals, CBL has a potential role in causing genomic instability and therefore drug resistance. Inhibition of glucose metabolism with the hexokinase inhibitor 2-deoxyglucose decreased cell growth and intracellular ROS. Importantly, blocking cellular glucose metabolism prevented the phosphorylation of FLT3 at its activation site by mutant CBL. Therefore, limiting the activity of the glycolytic pathway may target a crucial mechanism for CBL-dependent transformation.
In summary, identification of the mutational status of CBL in myeloid malignancies will be of important diagnostic and potential therapeutic value. Using siRNA approaches and small molecule drugs, we have further characterized major pathways involved in transformation through the CBL/FLT3 pathway. Because mutant CBL has the potential to interact with different RTKs and non-kinase receptors, inhibiting a single tyrosine kinase may have limited success in vivo for the treatment of mutant CBL-driven malignancies. It is therefore important to further characterize the mechanisms that lead to transformation by the gain-of-function mutations in CBL and identify targeted approaches that synergize or supplement traditional therapies.