The NH₂-Terminal Coiled-Coil Domain and Tyrosine 177 Play Important Roles in Induction of a Myeloproliferative Disease in Mice by Bcr-Abl

Abstract

Bcr-Abl, a fusion protein generated by t(9;22)(q34;q11) translocation, plays a critical role in the pathogenesis of chronic myelogenous leukemia (CML). It has been shown that Bcr-Abl contains multiple functional domains and motifs and can disrupt regulation of many signaling pathways and cellular functions. However, the role of specific domains and motifs of Bcr-Abl or of specific signaling pathways in the complex in vivo pathogenesis of CML is not completely known. We have previously shown that expression of Bcr-Abl in bone marrow cells by retroviral transduction efficiently induces a myeloproliferative disorder (MPD) in mice resembling human CML. We have also shown that the Abl kinase activity within Bcr-Abl is essential for Bcr-Abl leukemogenesis, yet activation of the Abl kinase without Bcr sequences is not sufficient to induce MPD in mice. In this study we investigated the role of Bcr sequences within Bcr-Abl in inducing MPD using this murine model for CML. We found that the NH₂-terminal coiled-coil (CC) domain was both essential and sufficient, even though not efficient, to activate Abl to induce an MPD in mice. Interestingly, deletion of the Src homology 3 domain complemented the deficiencies of the CC-deleted Bcr-Abl in inducing MPD in mice. We further demonstrated that the Grb2 binding site at Y177 played an important role in efficient induction of MPD. These studies directly demonstrated the important roles of Bcr sequences in induction of MPD by Bcr-Abl.

The bcr-abl oncogene, produced from the t(9;22)(q34;q11) chromosomal translocation known as the Philadelphia chromosome (Ph), is associated with 95% of the cases of chronic myelogenous leukemia (CML) and also with 20% of the adult and 5% of the pediatric cases of acute lymphoblastic leukemia (ALL) (23, 33). Depending on the nature of the translocation and exactly how the bcr and abl sequences become spliced into a final bcr-abl mRNA, various Bcr-Abl fusion proteins, including p185, p210, and p230, can be generated that show a preferential association with different types of leukemia (33). Clinical and laboratory studies indicate that Bcr-Abl plays an essential role in initiation of the chronic phase of CML and also plays a critical role in the maintenance and progression of the disease (10, 20, 50).

Bcr-Abl contains multiple functional domains and motifs. Abl-derived sequences in Bcr-Abl contain Src homology 3 (SH3), SH2, and tyrosine kinase domains in their N-terminal half, as well as a DNA binding domain, an actin binding domain, nuclear localization signals, and SH3 binding sites in their C-terminal region (41). The Bcr region (in the major p210 form) contains a coiled-coil (CC) oligomerization domain, a serine/threonine kinase domain, a Pleckstrin homology (PH) domain, a Dbl guanine-nucleotide exchange factor homology domain, and binding sites for the Abl SH2 domain and Grb2, Grb10, and 14-3-3 proteins (2, 25, 41, 49). The multiple domains of Bcr-Abl work cooperatively to activate intracellular signaling pathways commonly used in hematopoietic growth factor receptor signaling. These signaling pathways involve Ras, Raf, phosphatidylinositol 3 (PI3) kinase, Akt, JUN NH₂-terminal kinase (JNK), signal transducer and activator of transcription (STAT), Rac, Myc, and Bcl-2 or Bcl-x_L (43). Bcr-Abl also disrupts the adhesion pathway in CML cells (46).

Despite the advances in biochemical studies of Bcr-Abl, until recently there has not been a good experimental system that would allow the direct determination of the effect of specific domains and motifs of Bcr-Abl or of specific signaling pathways on the complex in vivo pathogenesis of the CML disease phenotype. Recently we and others have shown that expression of Bcr-Abl in mouse bone marrow cells by retroviral transduction efficiently induces a myeloproliferative disorder (MPD) resembling the chronic phase of human CML (24, 37, 50). This murine model for CML provides an effective in vivo experimental system to study the roles and relative importance of domains of Bcr-Abl and of signaling events affected by Bcr-Abl in leukemogenesis. It has been shown previously that the protein tyrosine kinase activity of Bcr-Abl is essential for its leukemogenic potential in vivo (50). However, c-Abl activated by an SH3 deletion did not induce MPD, although an SH3-deleted Bcr-Abl still induced a fatal MPD (14). These results indicate that activation of the Abl kinase alone is not sufficient for induction of MPD and that Bcr sequences in Bcr-Abl play an important role in inducing MPD.

Among Bcr domains and motifs, it has been shown that the NH₂-terminal CC domain plays an essential role in Bcr-Abl transformation (30). Deletion of the CC domain completely abolishes the transforming ability of Bcr-Abl in fibroblast cell lines, hematopoietic cell lines, and fresh bone marrow cells, although the CC domain-deleted Bcr-Abl retains detectable kinase activity, is autophosphorylated, and can activate Ras in cells (28, 30, 35, 44). The importance of the CC domain is also strengthened by the discovery of Tel-Abl (36). The Tel-Abl protein is a fusion of c-Abl and Tel, a member of the Ets family of transcription factors. A common property found between Bcr and Tel, two otherwise seemingly unrelated proteins, is that both can form oligomers. The helix-loop-helix domain of Tel, like the CC domain of Bcr, mediates the oligomerization of Tel-Abl and is required for Abl kinase activation, enhanced association with actin fibers, and transforming ability (13). These findings suggest the possibility that the only role in leukemogenesis of the Bcr and Tel portions of Bcr-Abl and Tel-Abl is to contribute an oligomerization domain to activate Abl. However, fusion of the Bcr CC domain alone was not sufficient to activate Abl to transform fibroblast cells (32). It has been shown that the adapter protein Grb2 binding site Y177, one of the motifs of Bcr-Abl that are important for the activation of Ras, is also important for Bcr-Abl function (39, 40). Mutation of Y177 diminishes Bcr-Abl-induced Ras activation and transformation in fibroblast cells (1, 39). However, the Grb2 binding site is not required to induce growth-factor independence in growth-factor-dependent hematopoietic cell lines and to transform bone marrow cells in vitro (5, 12). These results indicate that the importance of Y177 for in vitro transformation is dependent on the type of cells used, so the role of specific Bcr domains and motifs in leukemogenesis remains unclear.

In this study, we examined the role of the CC domain and Y177 of Bcr-Abl in the murine CML model. We found that Bcr-Abl without the NH₂-terminal CC domain failed to induce MPD in mice but still induced a T-cell leukemia and/or lymphoma with a greatly extended latency. Deletion of the Abl SH3 domain can restore the ability of the CC domain-deleted Bcr-Abl to induce MPD. We also demonstrated that the NH₂-terminal CC domain alone is sufficient, yet not efficient, to activate Abl to induce MPD and that Y177 is required for efficient induction of MPD.

MATERIALS AND METHODS

DNA constructs.

The p210 form of Bcr-Abl expression construct MSCV-bcr-abl/p210-IRES-gfp used in this study was described previously (50). To make ΔCC (Fig. 1A), the 5′ EcoRI-XhoI fragment (fragment A) of bcr-abl was subcloned into the EcoRI and XhoI sites of pBluescript II SK. The fragment between the SpeI site in the multiple cloning site of the vector and first MscI site in fragment A, cleaved by partial MscI digestion, was replaced by the SpeI and MscI linker: NT6 (5′ CTAGTTTGCTGG3′) and NT7 (5′CCAGCAAA3′). The fragment of SpeI and XhoI was excised out from pBluescript II SK and used with a SnaBI-XbaI adapter containing Kozak consensus sequences NT4 (5′GTAACCATGGCCT3′) and NT5 (5′CTAGAGGCCATGGTTAC3′) to replace the corresponding SnaBI-XhoI fragment of pBabe–bcr-abl. The SnaBI-EcoRI fragment containing the ΔCC DNA was then cloned into the HpaI and EcoRI sites of MSCV-IRES-gfp vector (Fig. 1). In the resulting ΔCC, the first 61 amino acids of Bcr-Abl were replaced by four amino acids—MASS. To make CC-Abl, first two overlapping fragments were amplified by PCR from bcr-abl with 5′ primer NT10 (5′ CTCCCTTTATCCAGCCCTCAC3′) and 3′ primer NT155 (5′GAAGGGCTTTAAAGCCCCATCGCTGCCGGTC3′) for fragment A and 5′ primer NT156 (5′GATGGGGCTTTAAAGCCCTTCAGCGGCCAGTAG3′) and 3′ primer NT125 (5′CCATCAGAAGCAGTGTTGATC3′) for fragment B. Then both fragments A and B were purified with the QIAquick PCR purification kit (Qiagen Inc., Chatsworth, Calif.) and mixed together as a template to generate PCR fragment C with 5′ primer NT10 and 3′ primer NT125. The fragment C was digested with EcoRI and HincII, and this EcoRI-HincII fragment together with the 3′ HincII-EcoRI fragment containing most of the abl sequences from bcr-abl were ligated into the EcoRI site of MSCV-IRES-gfp (Fig. 1A). To make ΔCC-ΔSH3, the SalI-EcoRI fragment of ΔCC in MSCV-ΔCC-IRES-gfp was replaced by the corresponding SalI-EcoRI fragment of ΔSH3 Bcr-Abl (14). The MSCV-Y177F-IRES-gfp construct was generated by swapping the first 660 bases (XhoI site) of the bcr-abl coding sequence between wild-type (wt) bcr-abl and a Y177F mutant (39). All sequences that were produced by PCR amplification were verified to be correct by sequencing.

FIG. 1.

(A) Schematic diagram of the retrovirus expression vector and Bcr-Abl proteins used in this study. Abl-derived sequences are shown in gray. Domains and motifs in Bcr-Abl include the following: the NH₂-terminal CC domain of Bcr (CC), the Grb2 SH2 binding site (Y177), the SH3 domain (3), the SH2 domain (2), and the Abl tyrosine kinase domain (K). Amino acid positions in Bcr-Abl where changes were made are indicated. Abbreviations for restriction enzymes: H, HpaI; RI, EcoRI. (B) Mass populations of infected 32D cells (GFP positive) were starved of IL-3 for 24 h. Equal amounts of cell lysates of the different 32D cell populations, as indicated, were separated on an SDS–6 to 15% polyacrylamide gradient gel and analyzed by immunoblotting with Ab3. (C) The same lysates as in panel B were analyzed with antiphosphotyrosine. The molecular mass standards are shown in kilodaltons in B and C. (D) The same lysates as in panel B were analyzed with anti-phospho-STAT5, anti-phospho-Akt, or anti-Dynamin antibodies as indicated. The filters were stripped and reblotted with anti-STAT5, anti-Akt, or anti-Actin antibodies, respectively. (E) Coimmunoprecipitation of Grb2 with Bcr-Abl proteins. Different bcr-abl constructs or vector, as indicated, were transfected into BOSC23 cells and anti-Abl IP was carried out 48 h later. Immunoprecipitates were immunoblotted with Ab3 (top panel) and an anti-Grb2 antibody (middle panel). Whole lysates were also blotted with the anti-Grb2 antibody (lower panel).

Cell culture and retrovirus preparation.

The NIH 3T3 mouse fibroblasts and BOSC23 cells were maintained as previously described (14, 50). The 32D clone 3 (32D) cells were maintained in interleukin-3 (IL-3)-containing medium (Dulbecco modified Eagle medium [DMEM] containing 10% fetal bovine serum [FBS], 100 U of penicillin/ml, 100 μg of streptomycin/ml, and 10% WEHI 3B-conditioned medium [WEHI-CM] as the source of IL-3). Helper-free retroviruses were generated by transiently transfecting retroviral constructs (Fig. 1A) into BOSC23 cells as described previously (38) and were wrapped in aluminum foil and stored at 4°C for up to 4 days without significant change in virus titers. BOSC23-conditioned medium was made just like making retrovirus except that there was no DNA in the transfection mix. Infection and virus titering was performed as previously described (14). All viruses were normalized to equivalent titers with BOSC23-conditioned medium just before infection of bone marrow cells or other cell lines.

In vitro transformation assay.

NIH 3T3 cells, plated at a density of 1.5 × 10⁵ per 60-mm plate 24 h prior to infection, were infected for 4 h with 2 ml of infection mix (50% viral supernatant, 10% donor calf serum, 100 U of penicillin/ml, 100 μg of streptomycin/ml, and 8 μg of Polybrene/ml in DMEM). Two days later, 10⁵, 10⁴, or 10³ cells from each plate were plated into six-well plates in 0.3% Bacto agar (Becton Dickinson, Sparks, Md.), 20% FBS, 200 U of penicillin/ml, and 200 μg of streptomycin/ml in DMEM. Wells were refed 2 weeks later. Colonies were counted under a microscope 5 weeks after plating. For bone marrow colony assays, the retroviral transduced bone marrow cells (see “Bone marrow transduction and transplantation and pathological diagnosis” below) were washed twice in excess phosphate-buffered saline (PBS) (GIBCO BRL, Grand Island, N.Y.) and were plated into six-well plates at 10⁵ cells per well (0.3% Bacto agar, 20% FBS, 200 U of penicillin/ml, 200 μg of streptomycin/ml, 50 μM β-mercaptoethanol in DMEM). For infection of 32D cells, 0.5 × 10⁶ cells/ml were incubated in DMEM containing 10% FBS, 10% WEHI-CM, 4 μg of Polybrene/ml, 50% retroviral supernatant, 100 U of penicillin/ml, and 100 μg of streptomycin/ml for 24 h. After 24 h, the cells were washed with PBS once and maintained in IL-3 medium. The green fluorescent protein (GFP)-positive 32D cells were sorted out the next day, expanded in IL-3-containing medium for a week, and resorted to a purity greater than 99.8% for all transduced 32D cells. These twice-sorted GFP⁺ 32D cells were maintained in IL-3 medium until further use without change in the percentage of GFP-positive cells in all samples.

32D cell proliferation and survival analysis.

The twice-sorted 32D cells freshly grown in IL-3 medium were washed three times in PBS, and viable cells were determined by fluorescence-activated cell sorter (FACS) analysis with propidium iodide staining. Equal numbers of viable cells from different samples were resuspended in medium with or without IL-3. The percentage of viable cells in culture was determined by FACS analysis, and the total number of cells was counted on a Coulter Counter (Model Z1; Coulter Particle Characterization, Hialeah, Fla.) each day for 4 days. The cell concentration of the culture was maintained below 2 × 10⁶ cells/ml during the whole experiment by diluting the culture in fresh medium. The total number of viable cells at each time point was calculated by multiplying the percentage of viable cells with the total number of cells and the dilution factor when it applied.

Bone marrow transduction and transplantation and pathological diagnosis.

Bone marrow cell infection and transplantation and pathological diagnosis were performed as previously described (50). Total blood cells and white blood cells (WBCs) were counted on a Coulter Counter (see above).

Flow cytometry and Southern blotting.

Flow cytometry analyses, cell sorting, genomic DNA preparation, and Southern blot analyses were performed as described previously (50).

Immunokinase assay.

Expression constructs were transfected into BOSC23 cells as described in “Cell culture and retrovirus preparation” above. Two days later, the cells were lysed in lysis buffer (50 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid [HEPES] [pH 7.4], 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EGTA, 1.5 mM MgCl₂, 1 mM dithiothreitol [DTT], 10 mM NaF, 1 mM sodium orthovanadate, 1 mM freshly made phenylmenthylsulfonyl fluoride, 1× complete protease inhibitor cocktail [Boehringer Mannheim, Indianapolis, Ind.]). Cell lysates were quantified with the Coomassie protein assay reagent (PIERCE, Rockford, Ill.), adjusted to equal concentration with the above lysis buffer. One milligram of total proteins (in 500 μl) was immunoprecipitated with the anti-Abl antibody (Ab3) at 4°C for 2 h, followed by the addition of 50 μl of UltraLink immobilized protein G beads (PIERCE) that had been washed three times with the lysis buffer. After 30 min of incubation at 4°C, the immunoprecipitates were collected by centrifugation, washed three times in lysis buffer and twice in kinase buffer (10 mM MgCl₂, 1 mM DTT, 50 mM HEPES), and were aliquoted equally into two Eppendorf tubes: one for immunoblotting (to determine how much of each Abl protein was used in kinase reactions) and one for the in vitro kinase assay. The kinase assay was performed in a total volume of 30 μl containing 1× kinase buffer, 1 μg of glutathione S-transferase (GST)–Crk_1–225 as a substrate, 0.5 mM ATP, and 5 μCi of γ[³²P]ATP for 30 min at room temperature, mixing the reaction by tapping the bottom of the tube every 10 min. Then 30 μl of 2× sodium dodecyl sulfate (SDS) sample buffer was added to stop the reaction, and the mixture was heated for 10 min at 100°C. The UltraLink beads were brought down by centrifugation, and an equal amount of supernatant was separated on an 6 to 15% SDS–polyacrylamide gradient gel and was exposed to X-ray film. The phosphorylation of the substrate was measured using a PhosphorImager and was normalized by the amount of Abl protein in each reaction.

Cell lysates and immunoblotting.

The twice-sorted 32D cells (described above) freshly grown in IL-3 medium were washed twice in excess PBS and starved in DMEM containing 10% FBS for 24 h. The live cells were counted by exclusion of trypan blue dye (GIBCO BRL), washed once, and resuspended in PBS at a concentration of 2 × 10⁷ live cells/ml; then an equal volume of 2× SDS sample buffer was added, samples were heated at 100°C for 5 min, and the debris was cleared by centrifugation. Immunoblotting was performed as previously described (14). Antibodies used in this study, except anti-Grb2 and anti-STAT5 (both purchased from Pharmingen/Transduction Laboratories, San Diego, Calif.), were the same as previously described (14).

IP.

Immunoprecipation (IP) was performed as previously described (42) with modifications. BOSC23 cells were transfected with different constructs (Fig. 1A). Two days later, cells from three 60-mm plates were collected; washed in ice-cold PBS; lysed in 500 μl of lysis buffer containing 1% Nonidet P-40 (NP-40), 20 mM Tris (pH 8.0), 50 mM NaCl, and 10 mM EDTA as well as 1 mM phenylmethylsulfonyl fluoride, 10 μg of aprotinin/ml, 10 mM NaF, 2 mM sodium orthovanadate, and 2× complete protease inhibitor cocktail (from a 50× stock); and incubated on ice for 25 min. Thereafter insoluble material was removed by centrifugation at 15,000 × g for 15 min. BOSC23 cell lysates (150 μl each) were diluted by the addition of 450 μl of incubation buffer (same as lysis buffer except that it does not contain NP-40). Ab3 was added to each sample, which were then incubated at 4°C on a rotating plate. After 3 h of incubation, 60 μl of UltraLink immobilized protein G beads was added to each sample. Following an additional 2 h of incubation at 4°C on a rotating plate, the beads were collected and washed three times with freshly made ice-cold IP wash buffer (same as lysis buffer except that the concentration of NP-40 was 0.1 instead of 1%) and were subsequently boiled with 50 μl of 2× SDS sample buffer before loading on SDS-polyacrylamide gels.

RESULTS

CC domain is necessary for Bcr-Abl to induce MPD in mice.

To examine the role of the NH₂-terminal CC domain of Bcr-Abl in the induction of MPD in vivo, we made a ΔCC mutant of Bcr-Abl in which the first 61 amino acid residues of Bcr-Abl containing the CC domain (30) were deleted (Fig. 1A). We first characterized the ΔCC mutant of Bcr-Abl, as well as other Bcr-Abl mutants shown in Fig. 1A, in a 32D clone 3 murine immature myeloid cell line. The 32D cells were infected with retrovirus containing an enhanced GFP gene (gfp) (Vector) or with retroviruses containing wt bcr-abl plus gfp or bcr-abl mutants plus gfp as shown in Fig. 1A. To avoid biased selection of certain cell clones that might occur during the outgrowth of factor-independent 32D populations due to the different oncogenic potential of Bcr-Abl and its mutants, we maintained the 32D cells in the presence of 10% WEHI-CM as a source of IL-3 during and after retrovirus infection. Then, mass populations of infected 32D cells (GFP positive) were isolated by FACS sorting (see Materials and Methods) and were used to characterize the expression of Bcr-Abl proteins, tyrosine phosphorylation patterns of intracellular proteins, and activation of the signaling proteins STAT5 and Akt 24 h after IL-3 withdrawal. The growth rate and viability of these cell populations were also examined in the presence or absence of 10% WEHI-CM.

As shown in Fig. 1B, ΔCC protein was expressed at a level similar to that of wt Bcr-Abl. However, consistent with the previous report (30), the kinase activity of ΔCC was drastically reduced compared to that of wt Bcr-Abl. The autophosphorylation of ΔCC in 32D cells was reduced 10-fold (calculated as the ratio of the amount of phosphorylated wt Bcr-Abl to phosphorylated ΔCC, denominated by the corresponding expression levels of wt Bcr-Abl and CC-Abl) compared to that of wt Bcr-Abl (Fig. 1C). Tyrosine phosphorylation of certain cellular proteins, such as p120 and p62 (likely to be Bcr-Abl substrates c-Cbl and p62dok, respectively), in ΔCC-expressing 32D cells was also decreased compared to 32D cells expressing wt Bcr-Abl (Fig. 1C). An in vitro immunoprecipitation-kinase assay, using GST-Crk_1–225 as a substrate, also revealed that deletion of the CC domain decreased the kinase activity of Bcr-Abl by 2.5-fold (Table 1). The smaller reduction in kinase activity of ΔCC in vitro may be due to the insensitivity of the in vitro kinase assay in distinguishing various forms of Abl, as previously shown (29).

TABLE 1.

In vitro kinase activity and transforming potential of Bcr-Abl and Bcr-Abl mutants

Construct	In vitro kinase activitya	NIH 3T3 coloniesbc	Bone marrow cell coloniesbd
GFP vector	NA	<0.3	<0.3
c-Abl	1.0	ND	ND
Bcr-Abl	5.5	309 ± 26	100 ± 3.7
ΔCC	2.2	<0.3	0.7 ± 0.0
ΔCC-ΔSH3	1.7	<0.3	33.6 ± 4.1
CC-Abl	2.1	<0.3	10.6 ± 2.6

^aThe kinase activity was measured by in vitro IP-kinase assay using GST-Crk_1–225 as a substrate, normalized by the amount of Abl proteins. The numbers represent the relative kinase activity of the Abl proteins. NA, not applicable. 

^bTiter-matched viruses were used for infection. The average number from three wells is given as the mean ± the standard deviation. The entry <0.3 indicates that no colonies were seen in any of the three wells. ND, not done. 

^cWells were refed after 2 weeks. Five weeks after plating, colonies were counted. 

^dNo exogenous cytokines were included in this assay. Colonies were counted 10 days after plating. 

As reported previously (3), we found that wt Bcr-Abl activated STAT5 in 32D cells in the absence of IL-3 (Fig. 1D, lane 2). However, the level of tyrosine-phosphorylated STAT5 in ΔCC-expressing cells was significantly decreased compared to that in Bcr-Abl-expressing 32D cells (Fig. 1D, pSTAT5, lane 5 versus lane 2), indicating that ΔCC has a reduced ability to activate STAT5. The level of activated Akt (pAkt) was also decreased slightly in ΔCC-expressing cells compared to that in Bcr-Abl-expressing 32D cells (Fig. 1D, pAkt, lane 5 versus lane 2).

It has been shown that Grb2 can bind Bcr-Abl through its phosphorylated Y177 (39, 40). Since ΔCC retains some ability to activate the Abl kinase and contains Y177, we examined whether ΔCC can bind to Grb2 by coimmunoprecipitation using BOSC23 cells transiently expressing Bcr-Abl variants. As shown in Fig. 1E, Grb2 was brought down from cells transfected with Bcr-Abl by the anti-Abl anti-body Ab3 (lane 1). A small fraction of Grb2 had a slower migration rate (Fig. 1E, lane 1 and bottom panel). This may reflect that some Grb2 was phosphorylated by Bcr-Abl. Indeed, it has been found that Bcr-Abl can phosphorylate Grb2 in vitro, which caused a slower migration of Grb2 (Subrahmanyam and Ren, unpublished data). A trace amount of Grb2 was also brought down from control cells containing vector alone (Fig. 1E, lane 2), kinase-deficient Bcr-Abl (lane 3), and the Y177F mutant of Bcr-Abl (lane 7). This weak binding may reflect the Grb2-Abl interaction through the Grb2 SH3 domains and the SH3 binding sites in the carboxyl-terminal region of Abl (42). Interestingly, Grb2 binds to ΔCC (Fig. 1E, lane 5) as strongly as it binds to wt Bcr-Abl, indicating that Y177 can still be phosphorylated in ΔCC. However, consistent with ΔCC's weak kinase activity, much less slow-migrating Grb2 was present in ΔCC cells (Fig. 1E, lane 5).

In vitro transformation assays showed that deletion of the CC domain abolished Bcr-Abl's transforming ability in the NIH 3T3 fibroblast cells (Table 1). This result is consistent with the previous report (30). However, in contrast to the previous results (28, 30, 44), ΔCC retained some ability to confer growth-factor independence in 32D cells (Fig. 2). Although the growth rate of ΔCC-expressing cells was greatly reduced (Fig. 2A), ΔCC had an ability to maintain the viability of 32D cells upon IL-3 withdrawal similar to that of wt Bcr-Abl (Fig. 2B). We also found that ΔCC could confer IL-3 independence in BaF3 cells (data not shown). The difference in the ability of CC domain-deleted Bcr-Abl in conferring growth-factor independence to hematopoietic cell lines found in this study versus the earlier reports (28, 30) could be due to different cell lines being used in different laboratories.

FIG. 2.

Effects of Bcr-Abl mutants versus those of wt Bcr-Abl on proliferation and survival of 32D cells. Various sorted GFP-positive 32D cell populations (as indicated) cultured in the presence of 10% WEHI-CM as a source of IL-3 were washed three times in PBS and resuspended in medium with or without 10% WEHI-CM. The total number of cells was counted on a Coulter Counter, the percentage of viable cells was determined by propidium iodide staining and FACS analysis, and the total number of live cells was calculated. Shown are the growth rate (A) (y axis is in log scale) and viability (B) of the 32D cells in the presence (lower panel) or absence (upper panel) of 10% WEHI-CM.

We next examined the leukemogenicity of ΔCC in mice compared to wt Bcr-Abl using the conditions of the mouse CML model (50). As shown before (50), mice receiving bcr-abl-infected bone marrow cells (Bcr-Abl mice) rapidly developed a fatal MPD (Fig. 3 and see Fig. 5). Interestingly, mice receiving ΔCC-infected bone marrow cells (ΔCC mice) also developed a fatal disease. However, the disease in ΔCC mice was drastically different from the wt Bcr-Abl-induced disease. First, the disease in ΔCC mice developed after a much longer latency period (the median latency was 115 days, compared to a median latency of 20 days for wt Bcr-Abl-induced disease) (Fig. 3 and Table 2). Second, while all diseased Bcr-Abl mice had very high peripheral WBC counts (usually >200,000 cells per μl) (Fig. 4A), a high WBC count (>100,000 cells per μl) was detected in only 35% (8 of 23 in three independent experiments) of ΔCC mice during the whole course of the experiment (Fig. 4B). Finally, the most important difference was that the types of cells involved in wt Bcr-Abl and ΔCC diseases were different. In diseased wt Bcr-Abl mice, a large number of myeloid (Mac-1⁺) cells accumulate in peripheral blood, spleen, liver, and bone marrow (Fig. 5B and data not shown). In contrast, all ΔCC mice developed T-cell leukemia and/or lymphoma, manifesting thymic lymphoma, lymphadenopathy, and pleural effusion (Fig. 5C and data not shown). Some ΔCC mice had a large number of GFP⁺ T lymphoblastic cells in peripheral blood while others had many fewer tumor cells in their peripheral blood. In all cases examined, the T lymphoid tumor cell phenotype was either Thy 1.2⁺ CD4⁺ CD8⁺ or Thy 1.2⁺ CD4^−to+ CD8⁺ and stained negative for Mac-1, CD19, and Ter119 (an erythroid cell-surface marker) (data not shown). GFP⁻ myeloid cells were also elevated in most of these mice, suggesting that ΔCC-induced T-cell tumor induces a reactive myeloproliferation (Fig. 5C). Western blot analysis demonstrated that ΔCC and Bcr-Abl were expressed at a similar level in tumor cells from ΔCC mice and Bcr-Abl mice, respectively (data not shown).

FIG. 3.

Survival of mice receiving bone marrow cells transduced with retroviruses carrying bcr-abl or bcr-abl mutants, as indicated. The curves were generated by Kaplan-Meier survival analysis using data collected from one representative experiment for each retrovirus. The number of mice in each group is indicated. Asterisks indicate the ΔCC-ΔSH3 mice that died with T-cell leukemia or lymphoma or anemia after a period of MPD.

FIG. 5.

Immunophenotypes of peripheral blood WBCs from a vector control mouse (A), Bcr-Abl mouse (B), ΔCC mouse (C), ΔCC-ΔSH3 mouse (D) and CC-Abl mouse (E). WBC counts and the time (days post-BMT) when the data were collected are shown for each mouse.

TABLE 2.

Summary of bone marrow transduction and transplantation experiments

Recipient mice	No. of mice (no. of expts)a	Median latency in days (range)	Percentage of mice that developed MPDb	Phenotypes at autopsyc
GFP vector	8 (2)	NAd	0	NAd
Bcr-Abl	16 (3)	20 (18–25)	100	16 MPD
ΔCC	23 (3)	115 (75–143)	0	23 TLL
ΔCC-ΔSH3	21 (2)	48 (28–180)	100	15 MPD, 3 TLL, 2 anemia, 1 unknown
CC-Abl	38 (3)	138 (45–222)	76	10 MPD, 18 TLL, 10 mixed TLL and MPD
Y177F	21 (2)	101 (70–214)	80e	7 mixed T-lymphomae and MPD, 2 T-lymphomae, 2 anemiae, 4 unknowne

^aTotal number of mice analyzed for each construct in x number of independent experiments. 

^bMPD: at any time during the experiment, WBC counts were >80,000 per μl with GFP⁺ granulocyte predominance. 

^cMPD phenotype at biopsy: mice had high WBC counts; most GFP⁺ cells were granulocytic cell series just before they were sacrificed or died; no massive T-cell involvement at autopsy. TLL, any one of thymic lymphoma, lymphadenopathy, T-cell leukemia, pleural effusion, or a combination of these tumors. Mixed TLL and MPD mean mice had both MPD and TLL. Anemia, whole blood cell counts were <2.5 × 10⁶ per μl. 

^dNot applicable. 

^eData obtained from 15 mice in which disease was analyzed. The other six mice were used for measuring survival rate only. T-lymphoma, abdominal T-cell lymphoma (also with pleural effusion in most cases). One of the mice that died of anemia had MPD at an earlier stage. Four mice that had an early MPD phase died before diagnosis could be made (unknown). 

FIG. 4.

Comparison of WBC counts among mice receiving bone marrow cells transduced with retroviruses carrying bcr-abl or bcr-abl mutants. WBC counts of wt Bcr-Abl mice (A), ΔCC mice (B), ΔCC-ΔSH3 mice (C), and CC-Abl mice (D) were plotted versus the time (days) post-BMT. Note that some graphs use different scales for the y axis.

It has previously been shown that expression of Bcr-Abl in freshly isolated bone marrow cells from 5-FU-treated mice (5-FU BM) could promote the formation of colonies of myeloid origin in soft agar without any exogenous cytokines (14, 15). Using this assay, we found that deletion of the CC domain virtually abolished the ability of Bcr-Abl to induce colony formation (Table 1). The above in vivo (summarized in Table 2) and in vitro experiments clearly demonstrated that the CC oligomerization domain of Bcr-Abl is essential for induction of myeloproliferation.

Deletion of the Abl SH3 domain rescues the ability of ΔCC to induce MPD in mice.

The NH₂-terminal CC domain of Bcr-Abl has been shown to play an important role in activation of the Abl kinase (30, 31). To examine whether the inability of ΔCC to induce CML-like disease is due to a lesser activation of the Abl kinase, we introduced a deletion mutation of the Abl SH3 domain in ΔCC. Mutations of the SH3 domain in c-Abl have been shown to activate the Abl kinase activity and its oncogenic potential (11, 21). It has been shown previously that SH3-deleted c-Abl is not capable of inducing MPD in mice, but a naturally occurring Bcr-Abl variant, Bcr-Abl/b3a3, in which a large portion of the SH3 domain is deleted, induces the same MPD as the major form of Bcr-Abl (b3a2) (14). We therefore constructed a ΔCC mutant of Bcr-Abl/b3a3, termed ΔCC-ΔSH3 (Fig. 1A), and examined the leukemogenicity of this mutant.

As shown in Fig. 1C, deletion of the Abl SH3 domain increased the tyrosine kinase activity of ΔCC for both autophosphorylation (by about sevenfold) and tyrosine phosphorylation of intracellular proteins in 32D cells. Interestingly, the in vitro kinase assay was unable to reveal the increase of the Abl kinase activity in ΔCC-ΔSH3 compared to ΔCC (Table 1). These results are consistent with the effect of SH3 deletion in c-Abl, where elevated kinase activity of SH3-deleted c-Abl can be detected in cells but not in vitro (29). In addition, deletion of the Abl SH3 domain also restored ΔCC's ability to activate STAT5, although the pAkt level in ΔCC-ΔSH3-expressing 32D cells remained slightly lower than that in wt Bcr-Abl-expressing 32D cells (Fig. 1D). A coimmunoprecipitation experiment showed that ΔCC-ΔSH3 could bind Grb2 strongly (Fig. 1E, lane 6). Consistent with the increased kinase activity of ΔCC-ΔSH3, more slow-migrating Grb2 was present in ΔCC-ΔSH3 cells than in ΔCC cells (Fig. 1E, lane 6 versus lane 5).

In vitro transformation assays showed that ΔCC-ΔSH3 was incapable of transforming NIH 3T3 cells (Table 1). It was reported that a similar ΔCC-ΔSH3, Δ1-40–ΔSH3, can transform Rat1 cells (28). This difference could be due to the different fibroblast cell lines used. However, deletion of the SH3 domain in ΔCC did increase the ability of ΔCC to confer growth-factor independence in 32D cells (Fig. 2A) and partially restored the ability of ΔCC to stimulate growth of primary 5-FU BM in vitro (Table 1).

When ΔCC-ΔSH3 was introduced into bone marrow cells under the conditions of the mouse CML model, it also induced a fatal disease (Fig. 3). However, in contrast to ΔCC mice, all ΔCC-ΔSH3 mice developed an MPD (Fig. 5D) and had high WBC counts (Fig. 4C), demonstrating that deletion of the SH3 domain rescues the ability of ΔCC to induce MPD in mice. However, the rescue was not complete compared to wt Bcr-Abl. First, although all ΔCC-ΔSH3 mice developed MPD, only 71% (15 of a total 21 in two independent experiments) died during this MPD stage (Fig. 3 and Table 2). The ΔCC-ΔSH3 mice that died with MPD displayed hepatomegaly, splenomegaly, and pulmonary hemorrhages, the same general features seen in wt Bcr-Abl mice. Three of the ΔCC-ΔSH3 mice that survived a period of myeloproliferative syndrome developed T-cell tumors (thymic lymphoma and pleural effusion), two died of anemia, and one died before diagnosis could be made. Second, ΔCC-ΔSH3 induced diseases with a longer latency; the median latency of the ΔCC-ΔSH3 disease was 48 days, and the average disease latency in ΔCC-ΔSH3 mice was similar to that of ΔCC mice (P = 0.575). Even though ΔCC-ΔSH3 mice that died of MPD (Fig. 3) had a significantly shorter latency than observed for ΔCC mice (P = 0.0002), the latency was still significantly longer than that of Bcr-Abl mice (P < 0.0001). These results suggest that the CC domain may have other functions than just to overcome the inhibitory function of the SH3 domain and/or that the Abl SH3 domain may have a positive role in MPD induction, which overlaps with Bcr sequences. Nevertheless, the results presented here indicate that the main function of the CC domain of Bcr-Abl required in induction of MPD can be largely rescued by deletion of the Abl SH3 domain.

CC domain of Bcr-Abl is sufficient to activate Abl for inducing an MPD in mice.

After showing that the NH₂-terminal CC domain of Bcr-Abl is essential for induction of MPD in mice, we examined whether the CC domain alone is sufficient to activate Abl to induce MPD. The NH₂-terminal CC domain of Bcr is predicted by sequence analysis to consist of amino acids from 28 to 68 in the NH₂ terminus of Bcr-Abl, which was confirmed by the finding that the first 71 amino acids of Bcr can form oligomers in vitro (30). It was also shown that fusion of the first 63 amino acids of Bcr to c-Abl was sufficient to activate the Abl kinase activity, actin association, and transformation of hematopoietic cell lines (30). However, the SEG program, a computer software that predicts globular domains based on amino acid composition (48), indicated that the first 77 amino acid residues of Bcr may form a globular structure. It is possible that the first 63 amino acid residues of Bcr are enough to form a functional CC domain, but the extra 14 amino acid residues may help its folding within Bcr. We therefore fused a DNA fragment encoding the first 77 amino acid residues of Bcr directly to c-abl starting at its second exon to generate cc-abl (Fig. 1A).

Consistent with previous reports (30), fusion of the CC domain activated the Abl kinase (Fig. 1C and Table 1). The reduction of autophosphorylation of CC-Abl (by about three- fold compared to wt Bcr-Abl) in 32D cells may be, at least in part, due to loss of tyrosine phosphorylation sites in the Bcr region (25, 26, 39, 40, 49). In addition, CC-Abl-expressing 32D cells had an amount of phospho-STAT5 similar to that of wt Bcr-Abl-expressing cells, although the pAkt level was slightly lower in CC-Abl-expressing 32D cells than in wt Bcr-Abl-expressing 32D cells. Consistent with the lack of the Grb2 SH2 binding site Y177, CC-Abl had a greatly reduced ability to bind Grb2 (Fig. 1E, lane 4).

Fusion of Bcr's NH₂-terminal CC domain alone did not activate c-Abl's oncogenic potential in NIH 3T3 cells (Table 1), as shown previously in Rat1 cells (30). However, fusion of the NH₂-terminal CC domain of Bcr to c-Abl did confer IL-3 independence in 32D cells (Fig. 2) and induced colony formation of 5-FU BM, with less efficiency than wt Bcr-Abl (Table 1).

When CC-Abl was introduced into mice by bone marrow transduction and transplantation, it induced an MPD (displaying high WBC counts with granulocyte predominance and hepatosplenomegaly) (Fig. 4D and 5E and data not shown) in the majority of recipient mice (Table 2), and some of the CC-Abl mice died of the MPD (Fig. 4D). Southern blot analysis showed multiple proviral integrations with distinct intensity in peripheral blood WBCs from most CC-Abl mice, indicating that CC-Abl induced a polyclonal MPD (see Fig. 7A). All these peripheral blood samples were collected during the MPD phase (with the WBC count between 150,000 and 240,000/μl) between days 66 and 73 post-bone marrow transplantation (BMT). These results indicate that the MPD induced by CC-Abl, just like that induced by Bcr-Abl, is polyclonal and that CC-Abl itself, like Bcr-Abl, is sufficient to induce an MPD. However, CC-Abl was not efficient in inducing the fatal MPD compared to wt Bcr-Abl. First, CC-Abl mice survived much longer than Bcr-Abl mice (P < 0.0001) (Fig. 3). Second, only 76% (29 of 38 in three independent experiments) of CC-Abl mice developed MPD, and among these 29 CC-Abl mice, only 10 died during the MPD phase. Third, CC-Abl did not induce pulmonary hemorrhages, which may explain why CC-Abl mice with MPD lived much longer than wt Bcr-Abl mice (all of which had pulmonary hemorrhages and died rapidly). Finally, there were significantly more B and T lymphoid cells (both GFP⁺ and GFP⁻) in the peripheral blood of the majority of CC-Abl mice throughout the whole course of disease development, including MPD phase, than seen in wt Bcr-Abl mice with MPD (Fig. 5B and E). These results suggest that CC-Abl has the ability to promote the proliferation of both myeloid and lymphoid cells.

FIG. 7.

Proviral integrations in cells isolated from CC-Abl mice. Genomic DNA was isolated from peripheral blood WBCs of five CC-Abl mice during an MPD phase (with WBC count between 150,000 and 240,000/μl) between days 66 and 73 post-BMT (A) and from indicated tissues of CC-Abl mice BMT20.19 (B) and BMT17.19 (C). The DNA was then digested with EcoRI (A, B1, and C1) or Bg1II (B2) and subjected to Southern blot analysis using a ³²P-labeled IRES-gfp fragment as a probe. The filter in panel C1 was stripped and reprobed with a fragment from the 3′ end of bcr-abl cDNA to detect the full-length cc-abl cDNA (4.5 kb) (C2). The peripheral blood of mouse BMT20.19 (B) had a WBC count of 259,000/μl and contained 74.9% GFP⁺MAC-1⁺, 2.5% GFP⁺CD19⁺, and 1.1% GFP⁺Thy 1.2⁺ cells (M74.9B2.5T1.1) at day 68; a WBC count of 139,000/μl and M44.6B5.3T8.5 at day 165; and a WBC count of 115,000/μl and M45.4B4.5T2.9 at day 214 post-BMT. The sorted spleen T cells (B) had a 98.6% purity. The peripheral blood of mouse BMT17.19 (C) had a WBC count of 156,000/μl and contained M15.1B0.8T32.6 when it was sacrificed at day 128 post-BMT. The sorted spleen myeloid and T cells from this mouse had 88.8 and 94.5% purity, respectively. Th.L., thymic lymphoma; Sp.M., sorted spleen Mac-1⁺ cells; Sp.T., sorted spleen Thy 1.2⁺ cells; Pl. eff., pleural effusion. Molecular size markers are shown on the right.

To further illustrate the course of disease development in those CC-Abl mice that survived the MPD phase, we performed FACS analyses of peripheral blood WBCs from several CC-Abl mice at different time points during disease development and show the data from one representative mouse, BMT17.15, in Fig. 6. When mouse BMT17.15 first developed a high WBC count, the majority of peripheral WBCs were myeloid cells (Fig. 6B). The MPD syndrome was sustained for more than a month in this mouse (Fig. 6A, B, and C). Then the WBC counts decreased and remained low for 4 months (Fig. 6A). Later, when the WBC counts increased again (Fig. 6A), there were many fewer GFP⁺ myeloid cells but more GFP⁺ T lymphoid cells (Fig. 6D). As the WBC counts continued to increase, there were progressively more GFP⁺ T lymphoid cells and fewer GFP⁺ myeloid cells present in the peripheral blood of this mouse (Fig. 6D and E). When mouse BMT17.15 died, pathological examination revealed that it had developed thymic lymphoma, splenomegaly, and pleural effusion (the majority of cells in the pleural effusion were T lymphoblastic cells [data not shown]). Some CC-Abl mice that survived an MPD phase did not develop high WBC counts at later time points after BMT and died of T-cell lymphoma (BMT17.17 is an example [Fig. 4D]). Those CC-Abl mice who failed to develop MPD also developed a T-cell malignancy. In summary, among 38 CC-Abl mice, 10 developed and died of MPD, 9 developed and died primarily of T-cell malignancy, and 19 developed an MPD in an early phase and later died of either T-cell malignancy (9 of 19 mice) or mixed MPD and T-cell tumors (10 of 19 mice) (Table 2).

FIG. 6.

The course of disease development in a representative CC-Abl mouse that developed a CML-like syndrome during the early phase of disease and a T-cell leukemia and lymphoma in the later phase of disease. Peripheral blood WBCs of mouse BMT17.15 (a CC-Abl mouse that was also shown in Fig. 4D) were counted and were analyzed for the presence of myeloid (Mac-1⁺) cells and B (CD19⁺) and T (Thy 1.2⁺) lymphoid cells by FACS analysis at different days post-BMT. The WBC counts versus the time plot (A) for mouse BMT17.15 are taken from Fig. 4D. WBC counts and FACS profiles are shown for day 43 (B), day 71 (C), day 195 (D), and day 215 (E), as indicated.

The fact that some CC-Abl mice developed T-cell malignancies after a very long MPD phase suggests that CC-Abl may have been targeted into hematopoietic stem cells and induced an MPD that subsequently transformed to T-cell malignancy. Alternatively, the MPD and T-cell malignancy induced by CC-Abl could have different cell origins and therefore represent different transforming events. To distinguish between these possibilities, we examined the proviral integration pattern in cells isolated during the MPD phase versus the T-cell leukemia and/or lymphoma phase (Fig. 7B and C). Mouse BMT20.19 developed predominantly an MPD at an early stage, and the MPD was sustained throughout the whole experiment (see WBC counts and percentages of GFP-positive myeloid cells at days 68, 165, and 214 post-BMT in the legend of Fig. 7). When this mouse was sacrificed at day 214 post-BMT due to a moribund condition, we found that it also had thymic lymphoma. Southern blot analysis showed that peripheral blood cells isolated in the MPD phase shared a common proviral integration site (approximately 4.5 kb) with cells isolated from the thymic lymphoma and sorted T cells from the spleen at the terminal stage of the disease (Fig. 7B1), indicating that the T-cell tumor was derived from a clone that also contributed to the MPD during the early phase. A second restriction enzyme digestion and Southern blot analysis (Fig. 7B2) confirmed this conclusion. It is notable that not all clones in the MPD phase developed a T-cell tumor and that predominant MPD clones differed at different time points of the disease development (Fig. 7B1). The latter phenomenon may reflect that progenitor cells at different developmental stages were targeted by CC-Abl. Analysis of a mouse with mixed MPD and T-cell leukemia (mouse BMT17.19) also showed that sorted myeloid cells shared common proviral integration sites with sorted T lymphoid cells (Fig. 7C1).

Interestingly, two CC-Abl mice (BMT20.17 and BMT20.19) also developed solid myeloid tumors that contained a large number of myeloid blast cells (data not shown). Although it is possible that these myeloblast tumors represent myeloid blast transformation of MPD, this is hard to study due to its rare occurrence.

Grb2-binding site Y177 of Bcr-Abl is required for efficient induction of MPD in mice.

The inefficiency of CC-Abl in inducing MPD suggests that Bcr sequences besides the CC domain play an important role in efficient induction of the disease. One of the important motifs in the Bcr region besides the CC oligomerization domain is the Grb2 SH2 binding site, which contains the phosphorylated tyrosine 177. To examine the role of Y177 in the induction of MPD in mice, we introduced the Y177F mutant of Bcr-Abl (Fig. 1) into mice using the conditions of the mouse CML model. We found that Y177F induced a fatal disease in most recipient mice with a significantly longer latency than wt Bcr-Abl (Fig. 8A). Among the 15 Y177F mice in which the disease was examined, 12 developed an MPD at an early stage, characterized by high WBC count (ranging from 84,000 to 360,000 cells/μl) with granulocyte predominance (Fig. 8B and data not shown). None of these mice died in the MPD phase. Seven of these 12 mice then developed a fatal T-cell malignancy at a later stage (Fig. 8C). One of these 12 mice died of anemia, possibly due to failure of long-term bone marrow reconstitution (Fig. 8A), and the 4 others died before diagnosis could be made. Among the three mice that did not develop an early MPD, two developed a fatal T-cell malignancy and one died of anemia (Fig. 8A). The T-cell tumors in all Y177F mice examined contained CD4^−to+ CD8⁺ cells (Fig. 8D), as was seen in CC-Abl mice (data not shown). However, the locations of Y177F-induced T-cell tumors were different from those of CC-Abl-induced T-cell tumors. While CC-Abl mice with T-cell malignancies manifested T-cell leukemia, thymic lymphoma, pleural effusion, and occasional lymphomas in the neck of mice, Y177F mice developed massive abdominal lymphomas in all mice. Pleural effusion was also found in six of nine mice examined, while thymic lymphoma was detected in only one of these nine mice. Despite this difference in the T-cell malignancies induced by CC-Abl versus the Y177F mutant of Bcr-Abl, the results described above indicate that the Grb2-binding site Y177 is important for Bcr-Abl to efficiently induce MPD.

FIG. 8.

(A) Survival of mice receiving bone marrow cells transduced by Y177F- and wt bcr-abl-containing retroviruses. The curves were generated by Kaplan-Meier survival analysis using data collected from one representative experiment. The two Y177F mice that died of anemia are marked with ★. One of these two mice had MPD at an earlier phase. (B) FACS profiles of the peripheral blood WBCs of a representative vector control mouse and diseased wt Bcr-Abl and Y177F mice. Also shown are the time (days post-BMT) and the peripheral blood WBC counts at the time of analysis. (C) FACS profiles of the abdominal tumor and spleen from a representative Y177F mouse with mixed MPD and T-cell malignancy. (D) FACS analysis characterizing the phenotype of the T-cell tumor from the same mouse as in panel C. The CD4 versus CD8 analysis was done on gated GFP-positive cells.

DISCUSSION

It has been shown previously that the Abl kinase activity is essential for Bcr-Abl leukemogenesis, yet activation of the Abl kinase without Bcr sequences is not sufficient to induce MPD in mice (14, 50). In this study, we found that the NH₂-terminal CC domain of Bcr is both essential and sufficient, albeit not fully efficient, for Bcr to activate Abl to induce MPD in our murine model for CML. We also found that the Grb2 SH2 binding site at Y177 played an important role for Bcr-Abl to induce efficiently the MPD, although the ability of Bcr-Abl to bind Grb2 directly was neither essential (since CC-Abl lacked the ability to bind Grb2 yet it induced an MPD in mice) nor sufficient (since ΔCC retained the ability to bind Grb2 yet it failed to induce MPD in mice) for Bcr to enable Abl to induce MPD.

The fusion of Bcr sequences to c-Abl generates an active protein tyrosine kinase in a manner similar to the activation of receptor tyrosine kinases (RTKs). RTKs are activated through ligand-induced dimerization, and the subsequent transautophosphorylation of receptor cytoplasmic tails creates binding sites for downstream signaling molecules. The fusion of Bcr sequences to c-Abl results in a constitutive oligomerization of the fusion protein and activation of the Abl kinase (30). Bcr also contains other functional motifs that link Bcr-Abl to downstream signaling molecules. Bcr sequences, therefore, change the activation, localization, and signaling properties of c-Abl. Consistent with this scenario, the NH₂-terminal CC oligomerization domain of Bcr was shown here to be both necessary and sufficient, although not fully efficient by itself, to activate c-Abl to induce MPD. Also consistent with this notion, the tyrosine phosphorylation site Y177, which is known to bind Grb2, was shown here to play an important role in the efficient induction of MPD by Bcr-Abl.

The fact that activation of the Abl kinase by deletion of the Abl SH3 domain could rescue the ability of ΔCC to induce MPD is consistent with the notion that a main function of the CC domain of Bcr is to activate the Abl kinase activity. However, oligomerization through the NH₂-terminal CC domain of Bcr appears not to be the sole mechanism for Bcr to activate the Abl kinase. In this study we found that ΔCC was still capable of inducing a T-cell malignancy in mice, albeit with a long disease latency. We also found for the first time that ΔCC had the same antiapoptotic activity as wt Bcr-Abl in 32D cells, although it had a much lower proliferation-stimulating capacity than wt Bcr-Abl in 32D cells (Fig. 2). It is possible that ΔCC maintains survival of hematopoietic cells in mice and that further neoplastic transformation may be due to the acquisition of additional genetic abnormalities. The development of T-cell- specific malignancy after a long latency may be a bias in the murine model, since it is a common end result of expression of several oncogenes or their mutants, which are not associated with T-cell malignancies in humans (4, 14, 19, 45). The oncogenic activity of ΔCC correlates with the fact that it retains some kinase activity (for both autophosphorylation and phosphorylation of intracellular proteins [Fig. 1B]), retains the ability to bind Grb2 (Fig. 1E), and retains some ability to activate STAT5 (Fig. 1C) and Ras in cells (28, 30, 35, 44). These results indicate that the rest of the Bcr sequences have some ability to activate the Abl kinase. Further identifying motif(s) in the Bcr region that are responsible for activation of the Abl kinase and further elucidating the mechanisms by which such motif(s) activate Abl will be important for understanding the regulation of the Abl kinase and for intervening in Bcr-Abl leukemogenesis.

The fact that both CC-Abl and ΔCC-ΔSH3 can induce MPD suggests that one of the key functions of the NH₂-terminal CC domain is to activate the Abl kinase. However, it has been shown previously that c-Abl activated by SH3 deletion is not sufficient to induce MPD (14). Together these data suggest that the fusion of the NH₂-terminal CC domain of Bcr to c-Abl has a function(s) in addition to activating the Abl kinase, and this additional function of the CC domain has an effect similar to that of fusing Bcr amino acids 64 to 927 to an activated c-Abl (ΔCC-ΔSH3). The CC domain of Bcr-Abl has been shown not only to play an important role in activation of the Abl kinase domain but also to enhance the association of Bcr-Abl with actin filaments (30, 31). The CC domain may also enhance functions of other domains of Abl, such as increasing the binding of Bcr-Abl to ligands of the Abl SH2 domain, SH3 domain, and other motifs in Abl, and it may also mediate interactions between Bcr-Abl and c-Bcr or other CC domain-containing proteins. The fusion of the CC domain or CC domain-deleted Bcr sequences to c-Abl also leads to the removal of the myristoylation site of c-Abl. Myristoylation may affect the cellular localization of the Abl oncoproteins, which, in turn, may affect signaling pathways important for induction of MPD. Further studies will be conducted to address these possibilities.

In this study we found that ΔCC-ΔSH3 was incapable of transforming NIH 3T3 cells (Table 1). The different effect of the deletion of the SH3 domain on c-Abl and ΔCC is probably due to the absence of the myristoylation site in ΔCC-ΔSH3, since the myristoylation site was shown to be required for the SH3-deleted c-Abl to transform fibroblast cells (7). The wt Bcr-Abl, which also lacks the myristoylation site yet has the ability to transform our NIH 3T3 cell line, may have an ability to associate with the plasma membrane in cells. Such ability may be somehow disrupted in ΔCC-ΔSH3.

The Y177F mutation in Bcr-Abl has been shown to block transformation of fibroblast cells by Bcr-Abl but not to affect the ability of Bcr-Abl to induce factor independence in hematopoietic cell lines and to transform primary bone marrow cells (5, 12, 39). These previous results indicate that the role of Y177 in Bcr-Abl transformation is cellular context dependent. In this study we found that Y177 in Bcr-Abl is required for efficient induction of MPD in mice, even under the condition of overexpression of the mutant. This experiment further demonstrates the importance of using an in vivo experimental system to study the roles and relative importance of domains and motifs of Bcr-Abl and of signaling events affected by Bcr-Abl in leukemogenesis.

One of the consequences of the Grb2–Bcr-Abl interaction appears to be the phosphorylation of Grb2. We found that a fraction of Grb2 in BOSC23 cells overexpressing Bcr-Abl and ΔCC-ΔSH3 migrated more slowly (Fig. 1E), suggesting that Grb2 was phosphorylated. Consistent with this notion we have found that Bcr-Abl could phosphorylate Grb2 in vitro (Subrahmanyam and Ren, unpublished data). Since very little slow-migrating Grb2 was present in BOSC23 cells overexpressing CC-Abl and the Y177F mutant, which contain an activated Abl kinase yet fail to bind Grb2 (Fig. 1E and data not shown), the phosphorylation of Grb2 was likely Bcr-Abl-binding dependent. Indeed, we found that the phosphorylation of Grb2 by the Y177F mutant was reduced compared to that by wt Bcr-Abl in vitro (Subrahmanyam and Ren, unpublished data). Similar to our finding, it has been shown that Grb2 was tyrosine phosphorylated in 293 cells overexpressing both platelet-derived growth factor receptor (PDGFR) and Grb2 (27). Tyrosine phosphorylation of Grb2 may regulate its function. Further studies are needed to examine whether Grb2 can be tyrosine phosphorylated under physiological conditions.

Although both CC-Abl and the Y177F Bcr-Abl mutant had a reduced ability to induce MPD and both induced T-cell malignancies with extended disease latency, the anatomical distribution of Y177F Bcr-Abl mutant- and CC-Abl-induced T-cell tumors was drastically different. While CC-Abl tumor cells spread mostly into peripheral blood, thymus, and to a much lesser extent, lymph nodes, Y177F Bcr-Abl mutant tumor cells had a strong preference to locate in abdominal mesenteric lymph nodes. This result suggests that signaling by Bcr sequences other than the CC domain and Y177 governs the types and/or homing of the malignant T lymphoid cells. Further determination of the effect of specific domains and motifs of Bcr-Abl and specific signaling pathways on the complex disease phenotypes in vivo using the murine CML model will help in the design of additional rational therapeutic interventions for CML. Insights gained from in vivo experiments will also contribute to understanding mechanisms of leukemogenesis in general and may also shed light on the molecular mechanisms of normal hematopoiesis.

ADDENDUM

It was reported, after submission of our results, that Y177 was required for the induction of MPD in mice by Bcr-Abl (34). Similar to our results, Y177F was shown to induce a massive abdominal T-cell malignancy. However, in that report only a few Y177F mice had moderately increased neutrophils in peripheral blood, bone marrow, and spleen, and some Y177F mice developed B-lymphoid leukemia (34). The differences between our results and those reported (34) may be due to different retroviral titers and/or different retroviral transduction methods used. Bcr-Abl has been shown to induce a variety of hematological neoplasms, including acute B-lymphocytic leukemia; pre-B-cell lymphoma; T-cell leukemia and/or lymphoma; macrophage, erythroid, and mast cell tumors; myelomonocytic leukemia; and myeloproliferative disease (6, 8, 9, 16–18, 20, 22, 24, 37, 47, 50). The nature of the hematological neoplasms induced by Bcr-Abl has been shown to be influenced by the genetic background of mouse strains, by treatment of bone marrow cells (5-FU treated versus non-5-FU treated), as well as by infection conditions or the promoters used in transgenic animals (8, 9, 16–18, 20, 24, 47), which may ultimately affect what cells Bcr-Abl expression is targeted into and, therefore, what disease Bcr-Abl induces.