BCR-ABL1 tyrosine kinase sustained MECOM expression in chronic myeloid leukaemia


Correspondence: Professor Chris Bartholomew, Department of Life Sciences, City Campus, Glasgow Caledonian University, Cowcaddens Road, Glasgow G4 OBA, UK. E-mail: c.bartholomew@gcu.ac.uk


MECOM oncogene expression correlates with chronic myeloid leukaemia (CML) progression. Here we show that the knockdown of MECOM (E) and MECOM (ME) isoforms reduces cell division at low cell density, inhibits colony-forming cells by 34% and moderately reduces BCR-ABL1 mRNA and protein expression but not tyrosine kinase catalytic activity in K562 cells. We also show that both E and ME are expressed in CD34+ selected cells of both CML chronic phase (CML-CP), and non-CML (normal) origin. Furthermore, MECOM mRNA and protein expression were repressed by imatinib mesylate treatment of CML-CP CD34+ cells, K562 and KY01 cell lines whereas imatinib had no effect in non-CML BCR-ABL1 −ve CD34+ cells. Together these results suggest that BCR-ABL1 tyrosine kinase catalytic activity regulates MECOM gene expression in CML-CP progenitor cells and that the BCR-ABL1 oncoprotein partially mediates its biological activity through MECOM. MECOM gene expression in CML-CP progenitor cells would provide an in vivo selective advantage, contributing to CML pathogenesis.

Chronic Myeloid Leukaemia (CML) is a disorder of haemopoietic stem cells (HSC) (Hamilton et al, 2010), characterized by the Philadelphia (Ph) chromosome (Kurzrock et al, 1988). The balanced translocation t(9;22) creates a novel fusion gene BCR-ABL1 (Ben-Neriah et al, 1986) which encodes a spatially (Wetzler et al, 1993) and functionally (Konopka et al, 1984) de-regulated tyrosine kinase, BCR-ABL1. BCR-ABL1 inappropriately activates the MAPK, PI3K and JAK-STAT signal transduction pathways (Pendergast et al, 1993; Skorski et al, 1997; Carlesso et al, 1996) contributing to abnormal myeloid cell proliferation, differentiation, transformation and survival (Smith et al, 2003).

CML usually progresses through three stages, which are designated as chronic (CP), accelerated (AP) and terminal blast crisis (BC) phases (Irvine et al, 2010). CML-CP and CML-BC resemble myeloproliferative disorder and acute leukaemia respectively; the transition between stages is unpredictable, non-time limited and inevitable unless treated (Elrick et al, 2005). Imatinib mesylate (IM; Glivec®; Novartis Pharma, Frimley, UK), a rationally designed tyrosine kinase inhibitor (TKI) that selectively inhibits BCR-ABL1 tyrosine kinase catalytic activity, is the currently favoured therapeutic agent that successfully manages the majority of patients with CML-CP (Valent, 2010), although it is much less effective when administered at advanced stages (CML-AP/BC) (O'Hare et al, 2006).

The t(9:22) translocation is a CML disease-initiating progenitor cell genetic change, the mutation being present in cells at all stages of the disease (Melo & Barnes, 2007). Disease progression requires the acquisition of new genetic abnormalities and various genes have been implicated in the majority of cases (reviewed in Melo & Barnes, 2007). Indeed, enhanced expression of MECOM (MDS1 and EVI1 complex locus, also known as EVI1, MDS1), a proto-oncogene located on chromosome 3q26, is frequently observed in CML-BC (Russell et al, 1993; Carapeti et al, 1996; Ogawa et al, 1996; De Weer et al, 2008). The MECOM gene encodes a zinc finger transcription factor with important roles both in normal development and leukemogenesis (Wieser, 2007). MECOM belongs to the positive regulatory (PR) domain family and is expressed as multiple naturally occurring alternatively spliced variants (Huang, 1999; Alzuherri et al, 2006). One form, designated MECOM (E), encodes the originally described protein (Morishita et al, 1990a) whereas another results from splicing of the coding region of the MDS1 gene with exon 2 of the EVI1 gene (encoding the PR domain) translating MECOM (ME) protein (Fears et al, 1996). Both proteins contain two domains of 7 (ZF1) and 3 (ZF2) repeats of the zinc finger motif (Morishita et al, 1988), function as DNA binding transcription factors (Wieser, 2007) and contribute to the progression of acute leukaemia (Morishita et al, 1992). Enhanced expression of MECOM in CML-BC implicates this transcription factor in disease progression (Ogawa et al, 1996).

Previous studies have shown that normal human CD34+ haemopoietic cells express MECOM (Gerhardt et al, 1997). Furthermore, this gene has been shown to have a role in self-renewal, proliferation and the repopulating capacity of murine HSC in Mecom null mice (Yuasa et al, 2005). However, MECOM's role in CML has not been fully determined. MECOM over-expression in CML-BC has been found both in the presence and absence of chromosome 3q26 abnormalities (Morishita et al, 1990b). Some chromosome 3q26 translocations generate enhanced expression of intact MECOM (E) including t(3:9;17;22), t(3;7), t(2;3), inv(3) and t(3;8) (De Weer et al, 2008; Henzan et al, 2004; Stevens-Kroef et al, 2004; Suzukawa et al, 1997; Lin et al, 2009) whereas others create novel fusion proteins involving ETV6 t(3;12) (Nakamura et al, 2002) or RUNX1 t(3;21) (Mitani et al, 1994). Many of the same genetic changes are also present in poor prognosis acute myeloid leukaemia.

Between 60–70% of CML-BC cells express MECOM in the absence of detectable gross cytogenetic abnormalities but in these cases it is unclear if expression is a marker or a driver of disease (Ogawa et al, 1996). These and other studies suggest that MECOM is not expressed in CML-CP mononuclear cells from bone marrow or peripheral blood, but CML-CP CD34+ cells have not been previously examined. This study investigated MECOM gene expression, the effect of IM treatment and the biological activity of this gene in primary CML-CP CD34+ progenitor cells as well as CML-derived cell lines.

Materials and methods

Cell culture

K562 and human embryonic kidney (HEK) 293T cells were cultured at 37°C in 5% CO2 in complete medium (CM) comprising RPMI 1640 medium or Dulbecco'e modified Eagle medium (DMEM) respectively, supplemented with 10% fetal calf serum (FCS), 2·5 mmol/l glutamine, 50 mg/ml penicillin, 50 units/ml streptomycin (all sourced from Lonza Group Ltd, Basel, Switzerland), and 200 mg/ml Geneticin® (for HEK293T cells only; Invitrogen, Paisley, UK). Lentivirus-infected cells were selected and maintained in CM and 2 μg/ml puromycin (Sigma-Aldrich, St Louis, MO, USA). TKI-treated cells were incubated with CM supplemented with 5 μmol/l IM (LC labs, Woburn, MA, USA). Hydroxycarbamide (HC) treated cells were incubated in CM supplemented with 400 μmol/l HC (Sigma-Aldrich). For colony-forming cell (CFC) assays, 1000 cells were plated in 1·5 ml Methocult® (StemCell Technologies SARL, Grenoble, France) in 30 mm petri dishes and cultured at 37°C, 5% CO2, 12 d. For IM-treated CFC assays, Methocult® was supplemented with 5 μmol/l IM. Colonies were counted using an inverted microscope (CK2; Olympus UK Ltd, Southend-on Sea, UK).

Preparation of total cellular RNA, cDNA synthesis and real time quantitative reverse transcription polymerase chain reaction (qRT-PCR)

RNA was prepared from cells at exponential growth phase semi-confluent cultures by the Trizol method (Invitrogen). 500 ng of total cellular RNA was used to synthesize cDNA with the Superscript® III 1st strand synthesis supermix for qPCR according to the manufacturer's instructions (Invitrogen). 5% of the cDNA reaction was used for qRT-PCR using ABsolute Blue QPCR mix (ABgene, Epsom, UK), gene-specific oligonucleotide primers and dual labelled probes, 95°C, 15 min followed by 40 cycles 95°C, 15 s, 60°C, 1 min in an OPTICON 2 DNA engine (MJ Research INC, Waltham, MA, U.S.A).

The efficiencies of the qRT-PCR reactions were calculated by using the formula Efficiency = −1 + 10(−1/slope) against the standard curve of each assay over a gradient of template concentration with each gene. The efficiency for MECOM (E), MECOM (ME) and GAPDH primers/probes were 112%, 115% and 110%. Relative expression levels of MECOM (E) and GAPDH or MECOM (ME) and GAPDH were determined using the arithmetic comparative math formula method (Livak & Schmittgen, 2001) and were determined relative to the calibrators MECOM (E) or MECOM (ME) respectively in K562 or CD34+ cells as described.


Gene specific oligonucleotides were synthesized and supplied by Integrated DNA Technologies (Leuven, Belgium).

  • 5′ FAM, 3′ TAMRA Human MECOM (E) and MECOM (ME) probe:

Preparation of plasmid DNA

pLKO.1 plasmids (Sigma-Aldrich) HB11 (MECOM shRNA, CCGGGCACTACGTCTTCCTTAAATACTCGAGTATTTAAGAAGACGTAGTCTTTTT), HB14 (MECOM shRNA, CCGGTGCAGGGTCACTCATCTAAAGCTCGAGCTTTAGATGAGTGACCCTGCATTTTT) and NT (MISSION® Non-target shRNA control vector) were prepared by affinity chromatography using Nucleobond® PC500EF gravity flow columns according to manufacturer's instruction (Macherey-Nagal GmbH & Co. Kg, Düren, Germany).

Production of lentivirus and infection of K562 cells

4–5 × 106 HEK293T cells were cultured in CM supplemented with 10% tetracycline-free FCS (Clontech Laboratories Inc., Mountain View, CA, USA) as described, and 3 μg of pLKO.1 recombinant Lentivirus plasmid DNA, Lenti-X HT packaging system (Clontech Laboratories Inc.) plasmid DNA transfected using the Lentiphos HT system (Clontech Laboratories Inc.) according to the manufacturer's instructions. Virus-containing cell supernatants were passed through a 0·45 μm cellulose acetate filter (Nalgene Company, Rochester, NY, USA) and viral titres determined using the Lenti-X qRT-PCR Titration Kit (Clontech Laboratories Inc.). 2 × 105 K562 cells were transduced at a multiplicity of infection of 40:1 with recombinant Lentivirus in CM supplemented with 4 μg/ml polybrene in six-well plates. Plates were centrifuged (Allegra X-22R; Beckman Coulter, Inc. Brea, CA, USA) at 1200 g, 20°C for 60 min and cultured at 37°C, 5% CO2 for 24 h. Transduced cells were selected in CM supplemented with 2 μg/ml puromycin (Sigma-Aldrich).

Western blotting

Protein extracts, sodium dodecyl sulphate polyacrylamide gel electrophoresis and Western blotting were performed as described previously (Bartholomew et al, 1997) with either α-MECOM (C50E12; Cell Signaling Technology, New England Biolabs, Hitchin, UK) or α-GAPDH (CA5; Fitzgerald Industries, North Acton, MA, USA) and diluted 1/1000 or 1/5000. α-c-ABL1 (2862), α-Phospho-CrKL (Tyr207) and α-CrKL (32H4) were each obtained from Cell Signaling Technology and diluted 1/1000. Anti-phosphotyrosine (4G10) was obtained from Millipore (Temecula, CA, USA) and diluted 1/1000. Appropriate horseradish peroxidase-conjugated anti-rabbit or anti-mouse (Sigma-Aldrich) IgG secondary antibodies were used at 1/5000 dilutions and detection was performed by enhanced chemiluminescence (Pierce, Rockford, IL, USA). Relative protein quantification was determined by densitometric analysis using Image Lab Software v3·0 (Bio-Rad laboratories Ltd, Hemel Hempstead, UK).

Single cell proliferation assay

1 × 106 cultured cells were incubated with 1/10 diluted fluorescein isothiocyanate (FITC)-labelled CD45 antibody (BD Biosciences, San Jose, CA, USA) for 15 min, washed with phosphate-buffered saline (PBS) (supplemented with 2% FCS). 50 μg/ml propidium iodide (PI; Sigma-Aldrich) was added and the cells immediately washed with PBS/2% FCS. Washed cells were re-suspended in 500 μl of PBS and single FITC+/PI negative cells isolated by cell sorting (BD FACS Aria; BD Biosciences) were dispensed into each well of a 96 well tissue culture plate by the automatic cell dispensing unit. Cells were monitored by visual inspection using an inverted light microscope (CKX41; Olympus UK Ltd) to confirm presence of single cells and daily for 4 days for evidence of cell division. All cells that divided at least once were classified as proliferating.

Preparation of primary CD34+ cells

Leukapheresis samples were obtained with informed consent as part of the routine assessment of untreated, newly diagnosed patients with CML-CP. Non-CML leukapheresis collections were processed as Ph-negative controls. Samples were enriched to >90% CD34+ progenitors by positive selection (CliniMACS®; Miltenyi Biotec, Bergisch Gladbach, Germany) and cryopreserved. CD34+ cells were cultured at 37°C in 5% CO2 in Iscove's modified Dulbecco's medium (Invitrogen) supplemented with serum substitute (bovine serum albumin, insulin and transferrin: ‘BIT’; StemCell Technologies), glutamine (Lonza Group Ltd), penicillin/streptomycin (Lonza Group Ltd) and five growth factors (IL-3, IL-6, flt3-L, and SCF from StemCell Technologies; G-CSF from Chugai Pharma Europe Ltd, London, UK) as previously described (Jørgensen et al, 2005).


MECOM knock down in K562 cells

Lentiviral vectors encoding non-target (NT) control shRNA and shRNAs targeting MECOM (HB11 and HB14), were used to create lentivirus particles by transient transfection of HEK293T cells as described in Materials and methods, and generated virus titres of 6 × 108/ml (NT), 1·2 × 109/ml (HB11) and 3 × 109/ml (HB14). HB11 is complimentary to a section of exon 7 of MECOM that is alternatively spliced to generate MECOM Δ324 and therefore should not knockdown (KD) production of this isoform (Fig 1A). HB14 is complimentary to sequences present in all naturally occurring transcripts (3′ untranslated region) and therefore should KD all isoforms (Fig 1A).

Figure 1.

MECOM knockdown in K562 cells. (A) Schematic representation of the four indicated natural MECOM isoforms. The position of shRNA homology for HB11 and HB14 are indicated by thick horizontal lines showing that HB11 is complimentary to sequences located in exon 7 present in all transcripts except MECOM (Δ324), and HB14 is complimentary to sequences located in the 3′ untranslated region of all transcripts. The coding exons 3–16 are numbered and indicated by boxes as described previously (Alzuherri et al, 2006). The MECOM (E) exon 2 and MECOM (MDS1) are also boxed and indicated for MECOM (ME). Splice variant-deleted exons are shown as black boxes. Both introns and 3′ untranslated regions are shown by horizontal lines that connect exons. (B) Histogram of MECOM (E) (white bars) or MECOM (ME) (black bars) mRNA levels normalized for GAPDH mRNA for the indicated cell populations relative to the K562 calibrator for MECOM (E) [GAPDH-normalized MECOM (E)] or MECOM (ME) [GAPDH-normalized MECOM (ME)] respectively, as determined by qRT-PCR. Columns are the mean of an experiment performed in quadruplicate and the error bars the standard deviation. (C) Western blot analysis of the indicated whole cell protein extracts using α-MECOM and α–GAPDH antibodies. The position of the 145 kDa MECOM (E), 88 kDa MECOM (Δ324) and 35 kDa GAPDH proteins are indicated by arrows.

Total cellular RNA and proteins derived from puromycin-selected cell populations from independent cultures of K562 cells infected with each lentivirus were examined by qRT-PCR and Western blot analysis to investigate the success of MECOM KD. The results showed 60–70% KD of MECOM gene expression (E and ME transcripts) in K562 cells with HB11, HB14 or a combination of HB11 and HB14, relative to NT control cells (Fig 1B). Western blot analysis showed that production of both 145 kDa MECOM (E) and 88 kDa MECOM (Δ324) isoforms were significantly repressed (80–90%) by HB14 or the HB11/HB14 combination whereas only the MECOM E protein was repressed by HB11 (70–80%) (Fig 1C), consistent with the qRT-PCR data and the anticipated specificity of the two shRNAs. The MECOM (ME) protein was not detected. Western blot analysis for GAPDH detected similar levels of the 35 kDa protein showing equal loading of total cell protein lysates in each case (Fig 1C).

MECOM KD has no significant impact on BCR-ABL1 kinase activity

To characterize MECOM KD K562 cells, we investigated if there was an effect on BCR-ABL1 at the message, protein or catalytic activity (via CrKL phosphorylation) levels. qRT-PCR results showed a slight reduction of BCR-ABL1 gene expression in MECOM KD K562 with HB11, HB14 and HB11/14 infected cells but not in NT control cells (Fig 2A). The same trend was seen by Western blot analysis of BCR-ABL1 protein levels with α-c-ABL1 antibodies (Fig 2B). However, BCR-ABL1 tyrosine kinase catalytic activity was unaffected by MECOM KD, as shown by uniform phosphorylation levels of the BCR-ABL1 substrate CrKL in all cells examined (Fig 2B). The abundance of total CrKL protein was similar in all cells and equal levels of GAPDH confirmed similar protein levels in the cell extracts examined (Fig 2B).

Figure 2.

BCR-ABL1 expression and catalytic activity in K562, NT, HB11, HB14 and HB(11 + 14) cells. (A) Histogram of BCR-ABL1 mRNA levels normalized for GAPDH mRNA relative to GAPDH-normalized BCR-ABL1 mRNA in K562 cells (calibrator) for the indicated cell populations, determined by qRT-PCR. Columns are the mean of an experiment performed in quadruplicate and error bars indicate standard deviation. (B) Western blot analysis of the indicated whole cell protein extracts using α-c-ABL1, α-Phospho-CrKL (Tyr207), α-CrKL and α–GAPDH antibodies. The position of the 210 kDa BCR-ABL1, 39 kDa p-CrKL and CrKL and 35 kDa GAPDH proteins are indicated by arrows.

MECOM KD reduced K562 CFC and single cell proliferation

To investigate the impact of MECOM KD on K562 cells, single parental K562, NT control and HB14-infected cells were sorted by fluorescence-activated cell sorting (FACS) into the wells of a 96-well cell culture plate and cell division was monitored. The results showed that a significantly reduced proportion of the MECOM KD K562 single cells were able to proliferate (Fig 3) and if they did divide, they turned over less frequently than parental K562 or NT control cells (data not shown). These data suggest the proliferation capacity of MECOM KD K562 is reduced.

Figure 3.

K562 and HB14 KD single cell proliferation assay. The histogram shows the percentage of undivided single cells at 0–4 d of culture. At day 0 all single cells in 96-well plates are undivided, giving a value of 100%. Black bars, K562 cells; grey bars, NT cells; white bars, HB14. All K562 and NT cells had divided at least once by day 4 compared to only 50% of HB14 cells. The results shown are the mean of an experiment performed twice.

We next investigated the functional effect of MECOM KD by examining the number of CFC in semi-solid culture media. The results showed a significant reduction in CFC (34%; P < 0·0001) in all MECOM KD cell populations (Fig 4A) relative to NT control cell populations (Fig 4A). Interestingly, all the colonies produced by the KD cell populations were not only fewer in number but also reduced in size (Fig 4C, D). These data suggest that MECOM KD significantly reduces K562 CFC activity.

Figure 4.

Colony-forming activity of K562 and derivative MECOM KD cells. (A) The number of CFC per 1000 cells plated for the indicated cell populations is shown. Each CFC assay was performed in triplicate for the indicated cell population and the results shown are a typical example of an experiment performed four times. ***< 0·0001. (B) The same as (A) for K562 cells with or without 5 μmol/l IM. (C) Images of typical colonies formed by the indicated cell populations. (D) Histogram showing mean colony diameter of indicated cell population as a percentage of mean K562 colony diameter. Error bars indicate the standard deviation. n = 20 for the total number of each cell colony type analysed.

MECOM expression in CD34+ cells

Previous studies have shown detectable MECOM gene transcripts in CML-BC but not CML-CP patient cells (Ogawa et al, 1996). To see if MECOM expression was detectable in CML-CP, we examined CD34+ selected cells derived from the peripheral blood of three newly diagnosed patients (demographics shown in Table 1) and in one allogeneic normal donor CD34+ as well as CD34+ cells from two non-CML patients (collectively designated ‘non-CML’) as controls. MECOM (E) and MECOM (ME) expression were normalized to GAPDH, then further scaled to GAPDH normalized MECOM (E) or MECOM (ME) calibrator respectively in normal CD34+ cells (sample 010). Both MECOM (E) and MECOM (ME) isoforms were readily detected in CD34+ cells, of both CML and non-CML origin. The relative abundance of both transcripts was similar in CML-CP CD34+ cells compared to the non-CML CD34+ cells (Fig 5A, B). These data show that MECOM (E) and MECOM (ME) gene transcripts are readily detected in primitive CML-CP CD34+ cells.

Figure 5.

MECOM (E) (white bars) and MECOM (ME) (black bars) gene expression in the indicated donor Ph− CD34+ non-CML cells (donors 011, 012) and Ph+ CD34+CML-CP cells (donors 249, 255, 257). Histograms indicate MECOM (E) and MECOM (ME) mRNA levels normalized for GAPDH mRNA relative to GAPDH normalized MECOM (E) or MECOM (ME) mRNA respectively, of calibrator cells (Ph− CD34+ normal cells 010), determined by qRT-PCR. Columns are the mean of an experiment performed in quadruplicate and the error bars the standard deviation.

Table 1. Characteristics of donor samples.
Patient IDSexAge (years)Diagnosis
  1. CML, chronic myeloid leukaemia; N/A, not available.

  2. Shows sex, age and diagnosis of donor CD34+ cell samples used for qRT-PCR analysis.

010Male45Normal donor
011Female51Multiple myeloma
012Female61Mantle cell lymphoma
019Male62Mantle cell lymphoma

MECOM expression in CML-CP cells and CML cell lines is repressed by IM

IM treatment has been shown to dramatically reduce the survival of p210 BCR-ABL1+ cells in vitro. As MECOM is a putative survival factor, we were interested to know if switching off the kinase survival pathway with IM would have any impact on the MECOM expression in CML-CP CD34+ cells. Interestingly, the results showed a time-dependent reduction of both MECOM (E) and MECOM (ME) gene expression in IM-treated cells relative to untreated cells (Fig 6A, B). The results also showed that MECOM (E) and MECOM (ME) expression was reduced independently of IM when CML-CP CD34+ cells (donor 289) were cultured for 12 h or more (Fig 6A, B). However, there was a statistically very significant reduction of MECOM (E) and MECOM (ME) in IM-treated cells, relative to cells cultured in growth medium alone, at both 12 (MECOM (E) P < 0·0008, MECOM (ME) P < 0·009) and 24 h (MECOM (E) P < 0·0001, MECOM (ME) P < 0·0009) (Fig 6A, B). MECOM (E) and MECOM (ME) expression was also examined in non-CML CD34+ cells (donor 019) + and − IM. IM treatment has no effect on MECOM expression in these BCR-ABL1 −ve CD34+ cells, although, once again expression was slightly decreased following 12 h or more of cells in culture in the presence or absence of the TKI (Fig 6C, D).

Figure 6.

MECOM (E) (white bars) and MECOM (ME) (black bars) gene expression in donor 289 (Ph+ CML-CP) cells and Ph− donor 019 cells exposed to 5 μmol/l IM for various times. Histograms indicate MECOM (E) (A, C) and MECOM (ME) (B, D) mRNA levels normalized for GAPDH mRNA relative to GAPDH normalized MECOM (E) (A, C) or MECOM (ME) (B, D) mRNA in calibrator cells [0 h, untreated donor 289 (A, B) or 019 (C, D)] respectively, in the presence (+) or absence (−) of 5 μmol/l IM for 6, 12 and 24 h determined by qRT-PCR. Columns are the mean of an experiment performed in quadruplicate and error bars indicate the standard deviation.

To confirm the effect of IM on MECOM gene and protein expression we examined its impact in the CML-derived cell lines K562 and KY01. Both MECOM (E) and MECOM (ME) isoforms were dramatically reduced after 6 h and continued to be repressed further for the time points shown (>90% after 24 h, Fig 7A) in K562. The same IM-mediated repression of MECOM (E) and MECOM (ME) isoforms was also observed in another CML-derived cell line, KYO1 (Fig 7B). Western blot analysis detected the 145 kDa MECOM (E) protein in both cell lines (Fig 7C), The abundance of the 145 kDa MECOM (E) protein rapidly diminished within 6 h of IM treatment and further reduced after 24 h (Fig 7C, 70–90% reduction after 24 h), consistent with the qRT-PCR data. IM treatment for 24 h inhibited phosphorylation of BCR-ABL1 (Fig 7D). Furthermore, inhibition of BCR-ABL1 catalytic activity was observed after 6 h IM treatment and was sustained for at least 24 h, as observed by dramatically-reduced phosphorylation of the CrKL substrate protein (Fig 7E). Counting of trypan blue (Sigma-Aldrich)-treated cells demonstrated that 5 μmol/l IM treatment inhibited K562 proliferation but had no effect on cell viability over the 24 h period examined (data not shown).

Figure 7.

MECOM (E) and MECOM (ME) mRNA levels in K562 (A) and KY01 (B) cell lines treated with 5 μmol/l IM for the time indicated. Analysis was as described in the legend to Fig 1 with untreated 0 h K562 (A) or KY01 (B) as calibrator. (C) Western blot analysis of whole cell protein extracts derived from cells treated with either 5 μmol/l IM (K562 and KY01) or 400 μmol/l HC for the indicated times using α-MECOM and α–GAPDH antibodies. Also shown are Western blot analysis of whole cell protein extracts derived from IM-treated (+) or untreated (−) cells for the indicated times with α-phosphotyrosine, α-c-ABL1 and α-GAPDH (D) and α-pCrKL, α-CrKL and α-GAPDH (E) antibodies.

To determine whether the impact of IM on MECOM is specific, or a more general consequence of inhibiting cell proliferation, we examined the effect of hydroxycarbamide (HC) in K562 cells. Western blot analysis showed no major change in the 145 kDa MECOM (E) protein following treatment of K562 cells with HC for up to 24 h (Fig 7C).

Our data show that IM rapidly inhibited MECOM gene expression in primary CML-CP CD34+ cells and CML-derived cell lines. Furthermore, MECOM KD significantly reduced K562 CFC activity (Fig 4A); this compared with an 80% reduction in K562 CFC achieved following IM (5 μmol/l) treatment (Fig 4B). Together these data suggest that MECOM repression could partially mediate the cellular response to IM.


This study showed that MECOM has a role in cell proliferation in Ph+ cells. KD of MECOM in K562 cells reduced their proliferative capacity in CFC as well as single cell proliferation assays. These observations were the same for KD either of all MECOM isoforms (with HB14-treated cells) or when MECOM (Δ324) expression only was retained (HB11-treated cells), showing this latter truncated protein, which lacks transforming activity (Kilbey & Bartholomew, 1998) cannot compensate for reduced MECOM (E) and MECOM (ME) isoforms. It is possible that our studies underestimate the impact of MECOM repression due to partial KD only because other studies including Mecom KO mice showed a more severe effect on HSC numbers (Yuasa et al, 2005) and an almost complete cell cycle arrest was reported in MECOM KD K562 cells (Lugthart et al, 2011).

The MECOM (ME) 185 kDa protein was not detected in K562 or KY01 cells using α-MECOM antibody, despite qRT-PCR data demonstrating RNA expression of this isoform. It is currently unclear if the MECOM (ME) protein is not produced in K562 cells or is at levels below the detection limits of our assays. This study might have underestimated the abundance of MECOM (E) encoding transcripts, as there are multiple transcription initiation sites (1a, 1b, 1c and 3L) that are not detected by the specific primers and probe set used here (Aytekin et al, 2005; Lugthart et al, 2008). Expression of MECOM (E) alone is associated with poor prognosis acute leukaemia (Barjesteh van Waalwijk van Doorn-Khosrovani et al, 2003). Therefore higher levels of MECOM (E) relative to MECOM (ME) would be likely to contribute to disease progression. A similar observation has been made for another PR family gene, PRDM2, where the shorter PR domain-deleted forms are always found in chromosome 1p36-linked malignancies, consistent with the view that the long PR domain-containing forms have tumour suppressor activity (Huang, 1999).

We also showed here that MECOM expression is not only seen in K562 and KY01 cell lines but also in primary CML-CP cells, whereas previous reports have suggested it is only observed in CML-BC (Ogawa et al, 1996). It is likely that the discrepancy between our results and previous data are due to increased sensitivity of detection resulting from the analysis of a CD34+ subpopulation of peripheral blood mononuclear cells, which express the highest levels of MECOM. Both CD34+ non-CML cells and CD34+ CML cells expressed similar levels of both MECOM (E) and MECOM (ME) gene transcripts. Previous studies show both isoforms are present in normal tissues (Fears et al, 1996; Wimmer et al, 1998). Only one of the samples in this study, donor 010, was normal but the relative abundance of both transcripts was similar here to that of the other non-CML and CML samples examined. Therefore, these data show: (i) that CD34+ cells express high levels of MECOM and (ii) that both transcripts are present in CML cells at similar levels to normal and non-CML cells.

In this study we observed that MECOM KD had no effect on BCR-ABL1 kinase activity, so the oncogenic catalytic kinase was then switched-off by treatment with the TKI, IM. IM mediated inhibition of BCR-ABL1 resulted in a rapid repression of both MECOM (ME) and MECOM (E) gene expression in CML CD34+, K562 and KY01 cells. This effect is likely to be specific, as IM treatment had no effect on MECOM expression in BCR-ABL1-ve non-CML CD34+ cells (Fig 6C, D). Western blot analysis showed a dramatic reduction of the 145 kDa MECOM (E) protein in IM-treated K562 and KY01 cells. This repression was not a consequence of general inhibition of cell proliferation but was BCR-ABL1 specific, as treatment with the anti-metabolite, HC did not alter MECOM (E) protein levels. Furthermore, this was not the result of general cell toxicity, as trypan blue exclusion studies showed the cells retained viability for at least 24 h of IM treatment and there was no impact on cellular levels of numerous other proteins examined, including CrKL, GAPDH, STAT5, ERK1/2, AKT and BCR-ABL1 (data not shown).

As was seen for K562 cell line, IM-mediated inhibition of BCR-ABL1 results in a rapid decline in MECOM gene expression at both the mRNA and protein level in primary CML-CP cells. MECOM expression also declined, by 50%, following culture of primary CML-CP cells for 12–24 h. This reduction is not as great as in the presence of IM and might reflect maturation of cells or ex-vivo culture not adequately replicating in vivo conditions. A similar partial decline in MECOM expression was also observed in non-CML CD34+ cells cultured for 12 h or more.

Inhibition of MECOM expression by IM demonstrates for the first time that its expression is regulated in BCR-ABL1+ cells by the catalytic activity of this aberrant kinase. Regulation of mRNA expression is rapid, suggesting it is a direct response to inhibition of tyrosine kinase-mediated signalling from the BCR-ABL1 protein. Several studies have previously identified many potential BCR-ABL1 target genes but MECOM was not described (Håkansson et al, 2008; Nunoda et al, 2007; Bianchini et al, 2007). This study represents the first report of a signal transduction pathway that regulates MECOM gene expression. Although BCR-ABL1 transmits an aberrant signal, these results suggest MECOM expression is modulated by one or more of the three major pathways, JAK-STAT, MAPK or PI3K, activated by this promiscuous kinase (Pendergast et al, 1993; Skorski et al, 1997; Carlesso et al, 1996). The particular pathway involved is currently under investigation. The molecular basis of BCR-ABL1-mediated MECOM regulation should facilitate how expression of this developmentally important gene is controlled in HSC and other tissues where it is normally expressed (Yuasa et al, 2005; Goyama et al, 2008; Hoyt et al, 1997).

These data establish a link between BCR-ABL1 kinase catalytic activity and MECOM gene expression. We propose that BCR-ABL1 positively regulates MECOM gene expression. The level of MECOM gene expression is not elevated by BCR-ABL1 kinase relative to expression levels observed in non-CML and normal primitive haemopoetic cells. However, it is deregulated, because the mechanism regulating MECOM gene expression in normal CD34+ cells is distinct from that in CML-CP CD34+ cells, given that, in the latter, expression is dependent upon BCR-ABL1 catalytic activity. The BCR-ABL1 kinase might activate a pathway that normally regulates MECOM production in primitive haemopoietic cells. Since BCR-ABL1 kinase is constitutively active, it will continuously stimulate the pathway leading to sustained de-regulated MECOM gene expression. It is unlikely that MECOM expression is repressed by inhibition of other receptors (PDGFRΑ, PDGFRΒ or KIT) that are known to be inactivated by IM (Fabbro et al, 1999) because IM has no effect on BCR-ABL1 −ve CD34+ cells (Fig 6C, D).

Maintenance of MECOM expression in primitive HSC is probably a selective advantage. Retroviral tagging studies showed that proviral insertions are frequently seen in the MECOM locus in dominant non-malignant HSC clones retrieved from transplant recipients (Kustikova et al, 2005) and primary myeloid CD34+ex-vivo cultures are enriched for viral insertions in this gene (Sellers et al, 2010). Furthermore, MECOM is required for the survival and proliferation of HSC (Yuasa et al, 2005; Goyama et al, 2008). The MECOM KD studies in K562 cells described above support this notion, as these cells show a reduced proliferative capacity in CFC and single cell proliferation assays. Our results suggest that BCR-ABL1 tyrosine kinase catalytic activity sustains MECOM gene expression in CML-CP CD34+ cells and that this gives cells a selective advantage in a manner analogous to retroviral insertion. It is possible the impact of MECOM KD in primary CML-CP cells would be even greater than seen with K562, which is an immortal cell line with a number of additional genetic abnormalities. We are currently investigating this.

Sustained expression of MECOM could be one of the mechanisms by which BCR-ABL1 contributes a selective advantage to primitive haemopoietic cells in CML, resulting in an increased production of mature cells in peripheral blood. In this case, gross chromosome abnormalities are not required as the BCR-ABL1 mutation causes de-regulation of MECOM gene expression. Interestingly, MECOM translocations are frequently observed in CML patients treated with TKI inhibitors that progress to blast crisis (Paquette et al, 2011) and enhanced MECOM expression is a predictor of poor prognosis in TKI-resistant CML-CP (Daghistani et al, 2010). This suggests that inhibition of BCR-ABL1 kinase may select for cells that de-regulate MECOM expression by alternative mechanisms. Mutations causing elevated levels of MECOM (E) relative to MECOM (ME), or its fusion proteins that can occur in CML-BC might be necessary for other MECOM-mediated biological activities including inhibition of terminal cell differentiation. Indeed, previous studies show that MECOM mediated inhibition of granulocyte differentiation is dependent on the level of expression (Khanna-Gupta et al, 1996). Our results suggest BCR-ABL1-mediated MECOM gene expression represents a novel mechanism of de-regulating this gene in leukaemia.


We would like to thank Aubrey Thompson and Jennifer Havens (RNA Interference Technology Resource, Mayo Clinic, Jacksonville, FL, USA) for providing pLKO.1 recombinant plasmids. We would also like to thank Linda Scobie, Glasgow Caledonian University (GCU) for providing HEK293T cells. This work was supported by a GCU PhD scholarship (SR) and Leukaemia & Lymphoma Research project grant 08018 (CB, GS, TLH, JVM). This study was also supported by the Glasgow Experimental Cancer Medicine Centre (ECMC), which is funded by Cancer Research UK and by the Chief Scientist's Office (Scotland). SR, PR and MAEB performed the research. CB, HJ, TH, GS and JVM designed the study and wrote the paper.

Conflict of Interest

The authors declare no conflict of interest.