BCR/ABL-mediated Increased Expression of Multiple Known and Novel Genes That May Contribute to the Pathogenesis of Chronic Myelogenous Leukemia¹

CML³ is a clonal myeloproliferative disorder of the multipotent hematopoietic stem cell (1, 2). At the molecular level, CML is characterized by the Philadelphia chromosome resulting from a balanced translocation between chromosome 9 and 22 (3). This translocation leads to the formation of the BCR/ABL fusion gene, encoding the p210^BCR/ABL oncoprotein (4–6). Compared with endogenous p160^Abl, the tyrosine kinase activity of p210^BCR/ABL is increased significantly (7). The involvement and absolute requirement of the p210^BCR/ABL protein in the pathophysiology of CML has been demonstrated in cell line models (8–11) and animal transplantation models (12–14). At the cellular level, presence of p210^BCR/ABL is associated with increased proliferation (15–18), increased resistance to apoptosis (19–21), and alterations in adhesion properties (22, 23). p210^BCR/ABL activates several downstream effector molecules involved in the control of proliferation, like RAS, RAF, MYC, and STAT. The ability of p210^BCR/ABL to block apoptosis has been related to RAS activation (24, 25), BCL-2 expression (26), Stat-5-mediated activation of BCL-XL (27), and activation of phosphatidylinositol 3′-kinase and Akt (28, 29). p210^BCR/ABL also phosphorylates and/or activates paxillin (30), the small GTP-binding protein Rac (31), and FAK/PYK2 (32), proteins critically important in cell adhesion and motility. However, the exact mechanism(s) underlying p210^BCR/ABL-mediated transformation remain not fully understood.

We hypothesized that presence of p210^BCR/ABL may up or down-regulate expression of known and yet to be identified genes, and that these alterations in gene expression contribute to the pathophysiology of p210^BCR/ABL-mediated transformation. Few studies have addressed this hypothesis, and most studies evaluated p210^BCR/ABL-dependent changes in expressed gene profile in BCR/ABL-transduced cell lines, which, by definition, are transformed before introduction of BCR/ABL (33). Among the genes differentially expressed in p210^BCR/ABL+ compared with p210^BCR/ABL− cell lines, are BCL-XL (27), the RAS-related KIR gene (34), the melanoma-related antigen PRAME (35), and the gene Ian4 (36).

To define downstream targets of p210^BCR/ABL, we studied alterations in the gene expression profile in a model system in which the BCR/ABL cDNA was retrovirally transduced in primary UCB CD34⁺ cells. We have demonstrated recently that BCR/ABL-transduced CD34⁺ UCB cells exhibit similar defects in proliferation, apoptosis, adhesion, and migration as primary CML CD34⁺ cells (37). Subtractive hybridization was used to isolate transcripts expressed differentially between CD34⁺ cells with or without p210^BCR/ABL, and results were confirmed in CD34⁺ cells from patients with newly diagnosed CP CML.

K562 cells were obtained from American Type Culture Collection (CCL-243) and maintained in Iscove’s modified Dulbecco’s medium (Invitrogen, Carlsbad, CA) supplemented with 10% FCS (HyClone, Logan, UT) and 1% l-glutamine (Invitrogen). The 293 cell line was obtained from American Type Culture Collection (CCL-86) and maintained in DMEM high glucose (Invitrogen) with 10% FCS.

Abs used in this study were directed against RAN (mouse monoclonal IgG2a; BD Biosciences, Palo Alto, CA), NME1 (Anti-Nm23 Ab, mouse monoclonal IgG2b; BD Biosciences), HSPC150 (Anti-E2 Ab, rabbit polyclonal IgG; PharMingen, Palo Alto, CA), ABL (monoclonal IgG1; Santa Cruz, Santa Cruz, CA), phosphotyrosine (mouse monoclonal IgG1; UBI, Lake Placid, NY), and CRKL (mouse monoclonal IgG1; UBI). The anti-NUP98-Ab was kindly provided by Dr. Van Deursen (Mayo Clinic, Rochester, MN). STI571 was a generous gift from Novartis (Basel, Switzerland).

BM was obtained from 12 patients with early CP BCR/ABL-positive CML and 7 normal (NL) healthy volunteers. In addition, we obtained 20 UCB samples from healthy term deliveries. All of the samples were obtained after informed consent using guidelines approved by the Committee on the Use of Human Subjects at the University of Minnesota. CD34⁺ cell-enriched populations were selected from mononuclear cells using immunomagnetic columns (Miltenyi Biotec, Sunnyvale, CA). Purity of CD34⁺ cells after two passes over the immunomagnetic columns was >90%.

After fluorescence-activated cell sorter sorting and before lysis for RNA or protein extraction, CD34⁺ cells were maintained for 12–16 h in serum-free medium with low-dose cytokines consisting of BIT-9500 (Stem Cell Technologies, Vancouver, British Columbia, Canada) supplemented with Iscove’s modified Dulbecco’s medium with 50 μm 2-mercaptoethanol (Sigma, St. Louis, MO), 40 μg/ml low-density lipoprotein (Sigma), 250 pg/ml granulocyte colony-stimulating factor (Amgen, Thousand Oaks, CA), 10 pg/ml granulocyte macrophage colony-stimulating factor (Immunex, Seattle, WA), 1 ng/ml IL-6 (R&D Systems, Minneapolis, MN), 50 pg/ml leukemia inhibitory factor (R&D Systems), 200 pg/ml macrophage inflammatory factor 1α (R&D Systems), and 200 pg/ml stem cell factor (Amgen). For STI571 studies, CD34⁺ cells were maintained for 48 h in serum-free medium with low-dose cytokines in the presence of 1 μm STI571.

UCB CD34⁺ cells were transduced with an MSCV-based retroviral vector containing either eGFP cDNA alone (M-eGFP) or the BCR/ABL-cDNA upstream from an IRES-eGFP sequence (M-p210-eGFP) as described previously (37).

Total RNA from UCB eGFP⁺CD34⁺ and BCR/ABL⁺ eGFP⁺CD34⁺ cells was isolated using the RNeasy procedure (Qiagen, Valencia, CA), and polyadenylated RNA was purified with the μMACS mRNA isolation kit (Miltenyi Biotec). Double-stranded cDNA was synthesized using the SMART cDNA synthesis kit (Clontech, Palo Alto, CA). Subtractive hybridization was carried out using the PCR-Select cDNA subtraction kit (Clontech) per the manufacturer’s protocol. To enrich for transcripts present in BCR/ABL⁺eGFP⁺ cells, subtraction was done using excess cDNA from eGFP⁺CD34⁺ cells (FORWARD). The subtraction was also performed in REVERSE orientation by subtracting excess cDNA from BCR/ABL⁺eGFP⁺ cells from eGFP⁺cDNAs. The subtracted cDNA(s) were then cloned using the Advantage PCR cloning kit (Clontech).

Differential screening was performed by high-throughput dot blot analysis using the PCR-Select differential screening kit according to manufacturer’s protocol (Clontech). Clones (1000) from the FORWARD subtracted library (present in BCR/ABL⁺eGFP⁺ cells but not eGFP⁺ cells) were blotted simultaneously on four nylon membranes (Hybond N+; Amersham, Arlington Heights, IL), hybridized with ³²P-labeled forward (present in BCR/ABL⁺eGFP⁺ cells but not eGFP⁺ cells) and reverse (present in eGFP⁺ cells but not BCR/ABL⁺eGFP⁺cells) subtracted cDNA probes, and unsubtracted cDNA from eGFP⁺ and BCR/ABL⁺eGFP⁺ cells. Only those clones that hybridized with the FORWARD subtracted probe (and cDNA from BCR/ABL⁺eGFP⁺ cells) but not the reverse subtracted probe (or cDNA from eGFP⁺ cells) were selected for sequencing.

Total RNA from UCB eGFP⁺CD34⁺ and BCR/ABL⁺eGFP⁺CD34⁺ cells was isolated using the RNeasy procedure (Qiagen), and polyadenylated RNA was purified with the μMACS mRNA isolation kit. Double-stranded cDNA was synthesized using the SMART cDNA synthesis kit (Clontech). Two μg of double-stranded cDNA was separated on a 1.2% agarose gel and transferred to a nylon membrane (Roche, Mannheim, Germany). Hybridizations were carried out at 42°C using DIG Easy Hyb (Roche). Probes were generated by PCR amplification of the subtracted cDNA clones using flanking primers (Nested PCR primer 1 and Nested primer PCR 2; Clontech) and purified by gel extraction (QIAquick Gel Extraction kit; Qiagen). The purified probes were labeled by random primed DNA labeling with the DIG DNA labeling kit (Roche) and purified with the High Pure PCR product purification kit (Roche). As a control for RNA loading, the blots were probed with β-actin cDNA. Detection was done using CSPD ready to use, according to the manufacturer’s protocol (Boehringer Mannheim, Mannheim, Germany).

cDNA clones that were confirmed overexpressed in BCR/ABL⁺eGFP⁺ cells in the high through-put dot blot screen were sequenced using the dye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA). Sequences were characterized by BLAST algorithm to SwissProt, GenBank protein and nucleotide collections, murine and human EST contigs, and on-line available stem cell databases. The sequences were considered part of known genes if they shared ≥95% homology over at least a 100-bp DNA sequence by BLAST search.

Total RNA from UCB CD34⁺ cells transduced with either eGFP or BCR/ABL-eGFP, or from BM CD34⁺ cells from CP CML patients and NL donors was isolated using the RNeasy procedure (Qiagen). Contaminating chromosomal DNA was digested with DNase I according to the manufacturer’s instructions (Invitrogen). cDNA was synthesized by reverse transcription of total RNA for 50′ at 42°C using oligodeoxythymidylic acid primers (Invitrogen) and Superscript II reverse transcription (Invitrogen). Real-time quantitative PCR of the known and unknown sequences was performed using the SYBR Green I dye (SYBR Green PCR Master Mix; Applied Biosystems). Sequence-specific primers were designed (Table 1), and to prevent nonspecific amplification, primer concentration optimization was performed and verified by gel analysis. The amplification was performed in an ABI Prism 7700 Sequence Detector System (Applied Biosystems) using 40 cycles of a two-step PCR (15 s at 95°C and 60 s at 60°C) after initial denaturation (95°C for 10 min). Amplification of β-actin mRNA as an endogenous control was used to standardize the amount of sample added to the reactions. To compare the relative amount of target gene in the different NL BM and CP CML samples, we designated arbitrarily one of the NL BM samples as reference in all of the Q-RT-PCR experiments (calibrator) and expressed the averaged sample value as ln of the percentage of the calibrator value. Real-time quantitative PCR for BCR/ABL mRNA was done using a fluorogenic probe designed using the Primer Express software version 1.0 (Applied Biosystems). Amplification of BCR/ABL and ABL (as internal house keeping gene) was performed with 1× TaqMan Universal mix (Applied Biosystems), 900 nm of each primer, and 225 nm of TaqMan probe for BCR/ABL amplification, or 800 nm of each primer and 200 nm of TaqMan probe for ABL amplification. Both amplifications were performed in an ABI Prism 7700 Sequence Detector System (Applied Biosystems) as described above. The BCR/ABL ratio was calculated in reference to the BCR/ABL RNA levels in of the K562 cell line.

Five × 10⁴ eGFP⁺ and BCR/ABL-eGFP⁺-transduced UCB CD34⁺ cells, and CP CML or NL CD34⁺ cells were resuspended in lysis buffer [IPWB; 50 nm Tris (pH 7.4), 250 mm NaCl, 2 mm EDTA, 1% NP40, 50 mm NaF, 2 mm NaVO₄, 1 mm NaPO₄, and Complete protease inhibitor mixture (Roche)]. Equal amounts of protein were mixed with sample buffer [2% SDS, 10% glycerol, 0.96 m 2-mercaptoethanol, 0.3 m Tris (pH 6.8), and 0.02% bromphenol blue], boiled for 5 min, separated by SDS-PAGE, and then transferred to Immuno-Blot polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA). Membranes were blocked using 5% nonfat dry milk in PBS-T (pH 7.6; 0.05% Tween 20) and incubated for 75 min with specific Abs. Immunoreactive bands were visualized using secondary horseradish peroxidase-conjugated Abs and chemiluminescence (Amersham).

The significance of differences between mRNA levels in eGFP⁺BCR/ABL⁺CD34⁺ and eGFP⁺CD34⁺ samples was determined using the two-sample paired t test. The significance of the difference between mRNA levels in BM CD34⁺ cells from CP CML patients and NL donors was analyzed by two-sample t test with unequal variance after ln transformation of the values.

Expression of BCR/ABL mRNA transcripts in M-p210-eGFP transduced eGFP⁺CD34⁺ cells and CP CML CD34⁺ cells was quantified by Q-RT-PCR. Transcript levels in M-p210-eGFP transduced UCB CD34⁺ cells were 0.9–1.7-fold the levels seen in the erythroleukemia cell line K562 cells, whereas BCR/ABL mRNA levels in CP CML CD34⁺ cells ranged between 0.4- and 5.6-fold those in K562 cells.

Subtractive hybridization was done using the Clontech PCR-Select System to isolate transcripts expressed differentially between eGFP⁺CD34⁺ and eGFP⁺BCR/ABL⁺CD34⁺ UCB cells. To minimize false-positive clones before sequencing, we performed a differential screen by high-throughput dot blot analysis of sequences selected by the initial subtractive hybridization step using the PCR-Select differential screening kit (Clontech). Approximately 12% of the 1000 initially identified clones isolated from the FORWARD subtraction were confirmed to be expressed higher in BCR/ABL than in eGFP-transduced cells. These results were additionally confirmed by virtual Northern blot (Fig. 1). Of the 70 candidate genes evaluated by virtual Northern, 34 were confirmed to be overexpressed in BCR/ABL compared with eGFP-transduced cells. These cDNA clones were sequenced, and sequences were characterized by comparison to available genomic, RNA, and protein databases. Twenty of 34 represent known sequences, and 14 of 34 represent unknown sequences (Table 2). Fig. 2 shows the sequence distribution between known and novel sequences, and the classification of the known sequences by function.

Sequence-specific primers were designed for 14 of 34 overexpressed genes, and Q-RT-PCR was performed to obtain independent confirmation of their overexpression in 3–4 additional UCB samples that were independently transduced under identical conditions. As shown in Fig. 3, all 14 of the genes overexpressed in BCR/ABL-transduced cells used for the subtractive approach were also overexpressed in additional BCR/ABL-transduced UCB samples. The 8 overexpressed genes with known function (Fig. 3a) were TOPK, CLC, NME1, IL1RL1, HSPC150, RAN, NUP98, and PBX3 (overexpression ranged between a mean 2.7- and 14.9-fold). Overexpression of the 6 novel sequences Seq1, Seq11, Seq15, Seq27, Seq60, and Seq62 ranged between a mean of 2.0- and 4.6-fold (Fig. 3b).

Although the CML model system isolates the direct effect of BCR/ABL on the expressed gene profile, results found in the model system may still differ from primary CML. Therefore, we also tested the expression level of the 8 genes with known function and the 6 novel sequences in CD34⁺ cells obtained from CP CML and NL BM. Fig. 4 depicts the expression of the known genes (Fig. 4a) and novel sequences (Fig. 4b) in CD34⁺ cells from 8 early CP CML samples compared with 3–4 NL donors. Results are expressed as the ln of the percentage of expression of the genes in each CP CML or NL CD34⁺ population compared with the level in one of the NL donors used as reference in all of the Q-RT-PCR experiments (calibrator, described in “Materials and Methods”). Despite patient-to-patient and NL donor-donor variability, all of the genes differentially expressed in the eGFP- versus BCR/ABL-transduced UCB CD34⁺ cells were also differentially expressed in CP CML versus NL marrow cells. Overexpression in CML versus NL CD34⁺ cells for the different genes ranged between a mean of 2.5- to 7.4-fold. We also confirmed that the 6 novel sequences were overexpressed in CML CD34⁺ cells with a mean overexpression of 2.8–5.5-fold.

The expression pattern of NUP98, RAN, HSPC150, and NME1 was then studied at the protein level by Western blot analysis of BCR/ABL versus eGFP-transduced UCB CD34⁺ cells (Fig. 5a) and CP CML versus NL CD34⁺ cells (Fig. 5b). Presence of the p210^BCR/ABL protein was confirmed in BCR/ABL-transduced UCB CD34⁺ cells. Consistent with the Q-RT-PCR results, we found that expression of NUP98, HSPC150, RAN, and NME1 was increased in BCR/ABL-transduced compared with eGFP-transduced UCB CD34⁺. We also compared protein levels for the known overexpressed genes in primary CP CML and NL CD34⁺ cells. RAN protein levels were higher in CP CML than in NL CD34⁺ cells. Likewise, HSPC150 was overexpressed in CML compared with NL CD34⁺ cells. NUP98 protein could not be detected in CD34⁺ cells from 2 NL donors, whereas significant levels of NUP98 were measured in the 4 CML samples. The only protein that was not overexpressed consistently in CP CML CD34⁺ cells compared with NL CD34⁺ cells, despite elevated levels of mRNA was NME1, which was overexpressed in only 1 of the 4 CML samples tested.

To determine whether the effect of p210^BCR/ABL on gene expression was caused by the BCR/ABL tyrosine kinase activity, we treated BCR/ABL and eGFP-transduced CD34⁺ UCB cells, or CP CML and NL BM CD34⁺ cells with 1 μm STI571 for 48 h. STI571 did not suppress BCR/ABL mRNA levels in BCR/ABL⁺ cells and did not modify expression of candidate genes in eGFP-transduced UCB and NL BM CD34⁺ cells (data not shown). Treatment of BCR/ABL⁺ cells with STI571 reversed the elevated mRNA levels of some of the sequences, whereas others stayed unchanged. In the presence of STI571, mRNA levels of NUP98, HSPC150, TOPK, RAN, and NME1 decreased by 2.2–22.6-fold in p210-transduced UCB CD34⁺ cells (Fig. 6a), and 1.4–3.9-fold in CML CD34⁺ cells (Fig. 6b), reaching levels similar to eGFP-transduced UCB or NL BM CD34⁺ cells. Levels of CLC, IL1RL1, and PBX3 mRNA remained unchanged in BCR/ABL-transduced UCB and CP CML CD34⁺ cells in the presence of STI571. Levels of CLC and PBX3 mRNA were actually increased after treatment of CP CML CD34⁺ cells with STI571.

As shown on Fig. 7,a, treatment of CML CD34⁺ cells with STI571 decreased phosphorylation of the p210^BCR/ABL and CRKL, but no change was seen in control NL CD34⁺ cells. As was seen for mRNA levels, treatment of CML CD34⁺ cells with STI571 significantly decreased protein levels of HSPC150 and NUP98 reaching levels seen in primary normal CD34⁺ cells (Fig. 7b).

Through subtractive hybridization, we have identified for the first time that introduction of the BCR/ABL oncogene in primary CD34⁺ progenitors causes increased expression of 14 genes (8 previously known and 6 novel genes), of which the overexpression was confirmed in CP CML CD34⁺ cells. Some genes have been implicated previously in cellular processes thought to be disturbed in CML, whereas other genes may be involved in new cellular pathways responsible for the BCR/ABL-mediated leukemic process.

Despite patient-to-patient and donor-to-donor variability, all of the genes showed a consistent pattern of up-regulation in BCR/ABL-transduced UCB and CP CML CD34⁺ cells. This demonstrates that the BCR/ABL-transduced UCB CD34⁺ cell model is suitable to study effects of BCR/ABL. We elected to initially study BCR/ABL-transduced UCB CD34⁺ cells rather than primary CML cells, because this allows us to determine immediate BCR/ABL-mediated changes in expressed gene profile, without possible confounding effects of additional genetic changes induced by prolonged presence of the P210^BCR/ABL oncoprotein even in newly diagnosed CML CD34⁺ cells.

Treatment of BCR/ABL⁺ cells with the Abl-specific tyrosine kinase inhibitor STI571 reversed increased expression of most genes. This was seen for the NUP98, HSPC150, RAN, TOPK, and NME1 genes. However, treatment with STI571 did not affect mRNA levels of the IL1RL1, CLC, and PBX3 genes. This suggests that two different categories of effectors exist, in some of which the overexpression is directly dependent on BCR/ABL tyrosine kinase activity and could be considered downstream of BCR/ABL, and others of which the overexpression is regulated by BCR/ABL but independent of its tyrosine kinase activity. Studies are ongoing to determine which BCR/ABL domain may be involved in the deregulated expression of the latter group of genes.

One of the genes that was overexpressed is CLC, present in the acidic granules of eosinophils and basophils. Although there is not likely a cause-effect relationship between this protein and CML, overexpression of this gene is consistent with the eosinophilia/basophilia seen in CML.

We also found increased expression of NUP98, and RAN mRNA and protein. NUP98 and RAN are both involved in transport of proteins and nucleic acids between the nucleus and cytoplasm through the NPC. NUP98 is a nucleoporin localized at the nucleoplasmic side of the NPC basket, and contains a number of FG repeats in its NH₂-terminal portion that act as docking sites in import and export pathways (38, 39). RAN is a small GTPase that cycles between a GDP-bound form (RanGDP) and a GTP-bound form (Ran-GTP), which plays an important role in both import and export (40, 41). NUP98 is a site for GDP/GTP exchange on RAN during the nuclear import mediated by Karyopherin β2, establishing a direct link between these two proteins in the complex nucleocytoplasmic transport machinery (42).

We hypothesize that deregulated expression of these two proteins may lead to pleiotropic effects associated with BCR/ABL expression, as many cellular processes depend on NPC function. For example, the NPC transports large amounts of mRNA and ribosomes important in protein synthesis. NUP98 also participates in nuclear export of proteins, such as HIV type 1 Rev (43). Furthermore, Vigneri and Wang (44) showed recently that in contrast to P160 ABL, p210^BCR/ABL is not found in the nucleus. However, when BCR/ABL-containing cells are treated with STI and nuclear export is blocked, cell death through apoptosis occurs. Therefore, we speculate that BCR/ABL up-regulates expression of NUP98 and RAN, which may export p210^BCR/ABL from the nucleus, to protect against apoptosis. If this hypothesis can be confirmed, it may be possible to target NUP98-mediated transport of p210^BCR/ABL and induce increased cell apoptosis.

Nucleocytoplasmic transport also plays important regulatory roles in processes such as cell cycle control (45). For example, subcellular compartmentalization regulates the function and/or stability of cyclin B1 and the cyclin-dependent kinase inhibitor p27^Kip1. Yeast two hybrid studies have shown that the murine nucleoporin Nup50 interacts directly with p27^Kip1 (46). We have shown recently that p27^Kip is found mainly in the cytosol of primary CML CD34⁺ progenitors, although p27^Kip is structurally normal. Whether increased expression of NUP98 and/or RAN is responsible for the excess export of p27^Kip into the cytosol is currently being examined. If confirmed, targeting of NUP98 may translocate p27^Kip back to the nucleus and increase its effects on cell proliferation inhibition.

A fourth gene found overexpressed in BCR/ABL-containing cells is PBX3, which is a member of the highly conserved PBX family of TALE homeobox genes along with PBX1 and PBX2 (47, 48). PBX proteins function as part of large complexes with other homeodomain-containing proteins to regulate gene expression during developmental and/or differentiation processes. Although significant sequence homology exists between PBX3 and PBX1 or PBX2, and PBX3 is ubiquitously expressed, little is known about its functional role. Because PBX associates with heterologous homeodomain proteins known to be implicated in hematopoiesis and leukemia (49), it is possible that PBX3 overexpression may play a role in BCR/ABL-mediated myeloid differentiation and transformation. Of note, two novel isoforms of PBX3 (PBX3C and D) were described recently in the BCR/ABL-positive cell line, K562, in addition to the canonical forms PBX3A and B (50). We demonstrated the presence of the four PBX3 isoforms in BCR/ABL-transduced UCB and early CP CML CD34⁺ cells, and identified two additional isoforms PBX3E and F (51). Because one of the new PBX3 isoforms misses the homeobox domain, it may affect PBX-mediated transcriptional regulation by acting as a dominant-negative isoform for PBX3. Of note, PBX3 overexpression was not reversed by treating BCR/ABL-positive cells with STI571. Studies aimed at understanding the role of the canonical PBX3 isoforms as well as the alternatively spliced forms seen in BCR/ABL-transduced cells will be important, because persistent overexpression in the presence of STI571 may contribute to escape of the disease from treatment.

A fifth gene and protein overexpressed in BCR/ABL-positive cells is HSPC150, an ubiquitin-conjugating enzyme E2. After activation of ubiquitin by the enzyme E1, E2 transfers ubiquitin to the ubiquitin-protein ligase E3, to which the substrate protein is specifically bound (52). Ubiquitin-dependent proteolysis is a critical component of diverse biological processes including cell cycle progression, the immune response, protein transport, and apoptosis (53–55). Increasing evidence is accumulating that ubiquitin-dependent proteolysis plays a role in leukemic development. Proteins that function as growth inhibitors/tumor suppressors are degraded through ubiquitin-dependent proteolysis, including p53, p27^Kip, and the Abi proteins (56–58). The finding that Abi protein expression is lost in cells after expression of BCR/ABL, together with the discovery that Abi proteins are absent in cells lines and BM cells isolated from patients with aggressive BCR/ABL positive leukemia, suggest that degradation of Abi proteins by the ubiquitin-proteasome pathways may be a component in the progression of CML. It is possible that overexpression of an ubiquitin-conjugating enzyme E2 in BCR/ABL-positive cells is responsible for this finding and also degrades other proteins important for maintaining “normal” hematopoietic cell biology.

Three other genes, TOPK, IL1RL1, and NME1 that are overexpressed in BCR/ABL⁺ cells are thought to play a role in cell proliferation and differentiation. As such, they may be responsible for the increased proliferation and skewed differentiation seen in this malignancy. TOPK (T-LAK cell-originated protein kinase), is a novel member of the MAPK kinase 3/6-related MAPKK family, which phosphorylates p38 MAPK (59). TOPK was thought to be specifically expressed in T lymphocytes, and, therefore, play an important role as a regulator in immune cells. We found that TOPK is the most overexpressed protein in BCR/ABL⁺ UCB CD34⁺ cells, which do not contain T cells, suggesting a novel function of TOPK in CML CD34⁺ cells.

IL1RL1 or ST2 is a member of the IL-1 receptor family and is expressed in a wide variety of human cells, including hematopoietic cells (60, 61). IL1RL1 has been identified as a primary response gene in cell growth control. Moreover, expression of IL1RL1 is greatly diminished when the human leukemic cell line UT-7 is induced to differentiate, suggesting an antidifferentiation role of IL1RL1 (62). Overexpression of IL1RL1 by BCR/ABL may play a role in the deranged differentiation seen in CML.

The NME1 or NM23A protein is involved in regulation of tumor metastasis, and inhibits differentiation of mouse and human leukemia cells (63). NME1 is overexpressed in AML and CML in blast crisis, and may have prognostic value in these malignancies (64). Moreover, a correlation between NME1 expression and the proliferation state of hematopoietic cells has been reported (65). Overexpression of NME1 mRNA in BCR/ABL⁺ UCB and CP CML CD34⁺ cells might contribute to the abnormal proliferation and aberrant differentiation of CML CD34⁺ cells. However, overexpression of NME1 at the protein level shown in BCR/ABL-transduced UCB was only seen in 1 CP CML patient among 4 tested. The reason for this is currently not clear, but it is possible that additional genetic changes that have occurred in primary CML CD34⁺ cells, even in newly diagnosed patients, might alter NME1 expression.

We reported elsewhere the increased expression of genes involved in mRNA splicing (SRPK1, HNRPA2B1, DDX21, DDX10, and SF3B3) in BCR/ABL-transduced UCB and CP CML CD34⁺ cells (51). Overexpression of these genes correlated with the identification of alternatively spliced forms of the genes PBX3 and PYK2 in BCR/ABL-positive CD34⁺ cells. These findings established for the first time that p210^BCR/ABL-induced alteration of mRNA splicing could be a potential mechanism of CML pathogenesis.

In summary, using subtractive hybridization we identified known and novel genes differentially expressed in NL and CML CD34⁺ hematopoietic progenitor cells. Effects on differentiation, proliferation, and adhesion of the constitutive overexpression by retroviral transduction of some of these known and unknown genes in normal cells are currently ongoing. Such studies should yield important new information on genetic changes that occur downstream from BCR/ABL that may in the future lead to novel targeted therapies. Future subtraction experiments between BCR/ABL-transduced UCB cell model and cells from CML patients at different stages may also yield additional novel insights in genetic changes involved in CML but not attributable to BCR/ABL.

We thank Prof. Kathryn Chaloner for the statistical analysis, and Huilin Qi, Todd Lenvik, Mo Dao, Sonya Melikova, Scott Dylla, and Stem Cell Institute Lab for technical assistance. Additional thanks to Dr. Van Deursen for reagents critical to these studies.