Combination with an Antisense Oligonucleotide Synergistically Improves the Antileukemic Efficacy of Erucylphospho-N,N,N -trimethylpropylammonium in Chronic Myeloid Leukemia Cell Lines

CML⁴ is a malignant disease of the human hematopoietic stem cell, which is characterized by the reciprocal translocation t(9,22)(q34;q11) and in most cases is associated with the Philadelphia chromosome. This genetic rearrangement leads to the expression of a BCR-ABL fusion protein with elevated and dysregulated tyrosine kinase activity (1). The oncoprotein is necessary and sufficient for transformation and growth factor-independent growth of hematopoietic and nonhematopoietic cell lines (2, 3). Hydroxyurea and other cytotoxic agents can control the clinical symptoms of CML such as myeloid hyperplasia, leukocytosis, and organomegaly but have little or no effect on progression into blast crisis or myelofibrosis with subsequent bone marrow failure. An IFN-based therapy is currently the standard regimen for CML, but the tyrosinase kinase inhibitor imatinib (Glivec, STI-571) has been introduced recently and shows promising therapeutic results (4–6). At the molecular level, CML is well characterized (1) and, because of its distinct genetic alteration, is suited for an antisense therapy by which short spans of nucleic acids are used to disrupt the expression of disease-related genes (7).

APCs are a new group of antineoplastic agents (8–10) that selectively damage leukemia cells without affecting the normal bone marrow (11). Recently, erucylphosphocholine and its congener ErPC₃ were described as the first i.v.-injectable APCs that lack hemolytic properties and show activity in vitro as well as an increased therapeutic ratio in vivo (12). In contrast with most antitumor agents, this class can even stimulate the production of hematopoietic progenitor cells (13, 14). An exception to the sensitivity of leukemia cells toward APCs is the erythroleukemic cell line K-562 (11), which expresses high levels of BCR-ABL (15). BCR-ABL can inhibit or at least delay apoptosis in leukemia cells, and its expression level influences the antileukemic efficacy of APCs (14). Therefore, we speculated that a reduced expression of BCR-ABL would render CML cells sensitive toward APCs.

In the present study, we investigated the monoeffects and combination effects of ErPC₃ and an ASO directed against the start codon of bcr (ASO-bcr) that was to reduce the expression of the BCR-ABL protein. From a panel of five leukemia cell lines with the bcr-abl rearrangement that was used in this study, the monoeffect and combination effects on colony formation were determined in the most resistant CML cell line (K-562) and in normal murine long-term bone marrow cell culture. The latter results were correlated with the induction of chromosomal aberrations of ErPC₃ on murine bone marrow in vivo.

ErPC₃ was synthesized as published elsewhere (16). Phosphorothioate oligonucleotides were synthesized by the Central Group for Oligonucleotide Synthesis and Sequencing of the Deutsches Krebsforschungszentrum according to the sequences reported by McGahon et al. (17). The structures are as follows: antisense (ASO-bcr), 5′-GCC CAC CGG GTC CAC CAT-3′, nonsense (NSO), 5′-CGC GCC TCG TCC CAA GCA-3′; junction site antisense (ASO-b2a2), 5′-CGCTGAAGGGCTTCTTCCTTATTGAT-3′; junction site antisense (ASO-b3a2), 5′-CGCTGAAGGGCTTTGAACTCTGCTT-3′; and TAT (mismatch), 5′-CTCTTTCCTTAT-3′.

The following BCR-ABL-expressing cell lines were used: K-562, LAMA-84, AR-230, CML-T1, and BV-173. In addition, HL-60, HL-60^dox, and SKW-3 cells were used as leukemic cell lines without BCR-ABL expression. All cells were grown as suspension cultures (RPMI 1640 supplemented with 10% FCS and 2 mm l-glutamine) at 37°C in an incubator with humidified atmosphere and 5% CO₂ and were passaged two or three times/week to keep them in log phase.

The efficiency of electrotransfection was determined by a fluorescein-labeled randomized sequence phosphorothioate oligonucleotide of 20-bp length (FITC-control). Single or double impulses of 2 ms duration ranging from 300 to 700 V/cm were used to transfect this oligonucleotide into K-562 cells. After cytospin centrifugation onto a slide, the treated cells were fixed in methanol, and the ratio of fluorescing cells was determined by a fluorescence microscope. For therapy, cells were transfected with specific phosphorothioate ASOs or NSOs by electroporation (two rectangle pulses at 200 V for 2 ms in 0.4-cm gap electroporation chambers, Eppendorf, Hamburg, Germany; corresponding to 500V/cm) using an electroporation impulse generator (EPI2500; Dr. L. Fischer, Heidelberg, Germany) with the capacitance set at 1200 μF) in RPMI 1640 supplemented with 10% FCS without phenol red as indicated below. After incubation for 24 or more hours in the presence of the oligonucleotides or ErPC₃, respectively, 2,500 (K-562) or 10,000 (AR-230, BV-173) cells were transferred into 5 ml of semisolid medium containing 0.8% RPMI-methylcellulose and 30% FCS. Cells were plated into 3.5-cm Petri dishes (1 ml/dish). Usual cell culture conditions (37°C, 5% CO₂ in humidified air) were applied. Colony formation (clusters of 20 or more cells) was scored by an inverted microscope after 7 days. At least three dishes/treatment protocol were used.

Colony counts were presented as mean with corresponding SD. For comparison between treatment groups, the Kruskal-Wallis test was applied (18). P ≥ 0.05 was considered significant. Predicted theoretical values were calculated according to the equation: C = a × b/100, where a and b are colony counts in percentage of untreated control obtained in response to single agents. For each concentration, applied theoretical values were calculated and compared with the real value of the combination. Two-factorial ANOVA was applied independently to test for synergistic interaction between the two agents (19). In addition, isobologram and multiple drug effect analysis was used to examine the nature of the interaction observed (20, 21). To that purpose, the dose-effect curve for each drug alone was determined based on the experimental observations using the median-effect principle; the CombI for each combination was then calculated according to the following equation:

where (D)₁ and (D)₂ are the doses of drug 1 and drug 2 that have X effect when used in combination, and (D_x)₁ and (D_x)₂ are the doses of drug 1 and drug 2 that have the same X effect when used alone. When CombI = 1, this equation represents the conservation isobologram and indicates additive effects. CombI values <1.0 indicate a more than expected additive effect (synergism).

Transfected cells were incubated, harvested, and washed in PBS. Thereafter, cell pellets (2 × 10⁶ cells) were lysed in buffer [0.1 m NaCl, 25 mm Tris-Cl (pH 8.0), 0.001m EDTA (pH 8), 25% glycerol, and 10% β-mercaptoethanol] containing 1 mm phenylmethylsulfonyl fluoride (Sigma), 50 μg/ml aprotinin (Sigma), 1 mm Na₃VO₄, and 2 mg/ml leupeptin (Sigma). After vigorously vortexing, 6× SDS loading buffer was added. Lysates were boiled for 5 min in a water bath at 100°C and centrifuged at 13,000 rpm for 10 min at room temperature.

Cell lysates corresponding to 300,000 cells were loaded on a 6% SDS polyacrylamide gel, separated by electrophoresis, and transferred to a polyvinylidene difluoride membrane, blocked with 5% skimmed milk in PBS, and incubated with a mouse monoclonal antibody against ABL or a rabbit polyclonal antibody against BCR (Santa Cruz Biotechnology, Santa Cruz, CA) for 2 h in 1× PBS, 0.5% milk, and 0.5% Triton X-100. After washing in PBS containing Tween 20 (0.05%), an horseradish peroxidase-conjugated antimouse or antirabbit IgG secondary antibody (Santa Cruz Biotechnology) and ECL (Santa Cruz Biotechnology) were used to detect both BCR-ABL and c-ABL or BCR constitutive proteins.

LTBMCCs were obtained as described previously (11). Briefly, female C57BL6 mice (Charles River, Wiga, Germany) were used as donors. After euthanizing the animals by cervical dislocation, femurs were removed and the marrow was flushed aseptically with a 21-gauge needle fitted to a syringe into LTBMCC medium (10% FCS, 10% horse serum, 10⁻⁶ m hydrocortisone, 0.017 mg/ml transferrin, 0.007 μl/ml β-mercaptoethanol, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 2.5 μg/ml amphotericin B in MEM-α). A single cell suspension was prepared by repeated gentle aspiration of the marrow plug. Bone marrow cells (1 × 10⁷) were plated in 25-cm² culture flasks and incubated at 37°C in 5% CO₂ in humidified air. Weekly, half of the medium in each flask was removed, and an equal volume of fresh medium was added. After an adherent layer had been established, the cultures were refed with 1 × 10⁷ fresh bone marrow cells/flask. Two weeks later, ErPC₃ was added, and electrotransfection of the nonadherent cells was performed. After incubation, cells were collected by trypsinization, and 250,000 cells were plated in 5 ml of semisolid medium (0.8% RPMI-methylcellulose, 30% horse serum, and 0.1 ng/ml recombinant mouse granulocyte/macrophage-colony stimulating factor). Aliquots of 1 ml/dish (three dishes/treatment group) were seeded in 3.5-cm Petri dishes and cultivated for 10 days. Colonies (clusters of 20 or more cells) were counted using an inverted microscope.

The cytogenetic investigation of bone marrow cells was performed as described previously with some modifications (11). Briefly, male and female C57BL6 mice were treated i.p. with ErPC₃ as indicated in Table 1. At the end of the treatment schedule, animals were injected with colchicine at a dose of 0.4 mg/kg body weight. Thereafter, mice were euthanized, and bone marrow cells were flushed from their femurs and hypotonized in 0.075 m KCl solution at 37°C for 20 min. After fixation in methanol:acetic acid (3:1), slides were made, air dried, and stained with 5% of Giemsa solution. To score chromosomal aberrations, at least 50 well-spread metaphases/animal were analyzed. In addition, the mitotic index was calculated by counting the number of dividing cells.

Transfection with impulses of 2-ms duration ranging from 300 to 700 V/cm was associated with a voltage-dependent increase in fluorescing cells (Fig. 1). The ratio of strongly fluorescing cells increased from 20 to nearly 40% (1 × 700 V/cm) and that of weakly fluorescing cells from 40 to 60% (2 × 500 V/cm), thus adding up to 60–90% of total cells. A pulse strength over 500 V/cm did not contribute to a further increase in the ratio of fluorescing cells. Except for 500 V/cm, there was no significant difference between using one or two impulses with regard to oligomer uptake. In the absence of electrotransfection, there was no discernible uptake of the oligonucleotide (<5% fluorescing cells after 1 h of incubation).

A dose-response study in K-562 cells with the bcr-directed ASO and a posttreatment period of 72 h after electrotransfection showed a clear concentration dependency in colony growth inhibition ranging from 20% (2.5 μm) to 40% (10 μm). The control oligonucleotides ASO-b2a2 and NSO-bcr showed no significant inhibition of colony formation (Fig. 2A). Repeated electrotransfection with two posttreatment periods of 72 h each did not cause a further decrease in the clonal growth of K-562 cells (Fig. 2B).

To broaden the spectrum of cell lines, the inhibitory effect of the ASO-bcr was also investigated in BV-173 cells. In accord with their low BCR-ABL level (see below), BV-173 cells showed a >80% inhibition of colony formation after a single round of electrotransfection (Fig. 2C). The inappropriate junction site ASO-b3a2, however, was completely devoid of an inhibitory effect.

Subsequent studies focused on the combination effect of the ASO-bcr with the APC ErPC₃ (Figs. 3 and 4). To that purpose, cells of the more resistant cell line K-562 were treated with the ASO-bcr, NSO, ErPC₃, or their combinations (Fig. 3). After 24 h of posttreatment incubation, the colony number was decreased by 45% in response to the ASO-bcr, or ASO-bcr plus ErPC₃ (P < 0.05; Fig. 3A). Neither ErPC₃ alone nor the NSO effected a significant decrease in colony numbers. After 48 h, the latter two agents caused a colony count reduction of 30%, whereas the ASO-bcr alone induced a 70% reduction in clonal growth (P < 0.05). Most strikingly, the combination of the ASO-bcr with ErPC₃ was synergistically active, resulting in 97% reduction of colonies as opposed to 80% that would have been expected from additive interaction (Fig. 3B).

This synergism prompted us to perform a dose-response study with the ASO-bcr and ErPC₃. On the basis of the optimal posttreatment period of 48 h, ErPC₃ reduced the clonal growth of K-562 cells in a concentration-dependent manner; no effect was observed at 10 μm, but 35% colony growth inhibition was found at 20 μm and 65% (P < 0.05) at 30 μm. The ASO-bcr caused a growth inhibition of 30 and 35% at 5 and 10 μm concentrations (Fig. 4A). The sequential treatment with ErPC₃ and the ASO-bcr resulted in synergistic growth inhibition at each dose level tested (Fig. 4B). The combination of 10 μm ErPC₃ and 5 and 10 μm ASO-bcr resulted in 57 and 68% colony inhibition (hatched columns) instead of 29 and 33% (open columns) that would have been expected from additive interaction. Isobologram analysis demonstrated CombI values of 0.40 and 0.36, respectively. Likewise, combination of 20 μm ErPC₃ with 5 and 10 μm ASO-bcr led to 72 and 75% inhibition as opposed to 55 and 58% that would have been expected, and isobologram analysis gave CombI values of 0.69 and 0.67, respectively. Finally, the combination of 30 μm ErPC₃ with 5 and 10 μm ASO-bcr caused 85 and 92% clonal growth inhibition instead of 77 and 79%, and isobologram analysis yielded CombI values of 0.89 and 0.78, respectively. All observed colony formation in response to the sequential combination was >30% below the respective expected (additive) effects and can therefore be considered significantly reduced. This was confirmed by the presence of a significant (P < 0.0001) interaction between the two agents in a two-factorial ANOVA as well as by CombI values, which were <1 and thus indicate synergism.

Using an anti-ABL antibody, the expression levels of the 210-kDa oncoprotein were compared in five CML human cell lines (K-562, LAMA-84, CML-T1, BV-173, and AR-230) and in AML-derived HL-60 cells that were used as a negative control for BCR-ABL (Fig. 5). K-562 cells were found to have the highest amount of the 210-kDa fusion protein, followed by LAMA-84 cells; CML-T1 and BV-173 cells showed a comparable, lower expression of this BCR-ABL protein than the other two cell lines. AR-230 cells express a large amount of the 230-kDa BCR-ABL protein (Fig. 5A, Lane 7). The level of this protein is comparable or even greater than that of K-562 cells. There was no band corresponding to the normal c-ABL protein in LAMA-84, BV-173, and AR-230 cells. (Fig. 5A). All CML cell lines under investigation express the normal 145-kDa BCR protein with the exception of CML-T1 cells (Fig. 5B).

To investigate whether electrotransfection with the bcr-directed ASO would result in inhibition of BCR-ABL expression, the CML cell lines K-562, LAMA-84, CML-T1, and BV-173 were treated either with ASO-bcr, NSO, or ErPC₃. The respective BCR-ABL levels were assessed by Western blot.

K-562 cells showed no difference in BCR-ABL expression after a 48-h incubation period with the ASO-bcr or NSO, regardless of whether the monoclonal anti-ABL (Fig. 6a) or the polyclonal anti-BCR antibodies were used (Fig. 6b). The constitutive proteins c-ABL or BCR served as internal controls. To overcome this resistance, the electroporation treatment was repeated once (total treatment period, 144 h), and by this the corresponding BCR-ABL level was distinctly reduced in response to the ASO-bcr but not the NSO (Fig. 6,c, Lanes 2 and 3). Densitometry of the BCR-ABL bands indicated 100 and 27% of the respective control for NSO and ASO treatments. For compensating any differences in loading, these percentages were normalized to the amount of ABL detected concomitantly that was used as internal control. Thus, the relative level of BCR-ABL was 83% (NSO) and 31% (ASO) of the control (Fig. 6c). In contrast to the K-562 resistance, LAMA-84, CML-T1, and BV-173 cells showed reduced BCR-ABL expression already in response to a single electrotransfection with the ASO-bcr but not the NSO (data not shown).

A single round of electrotransfection with the ASO-bcr and subsequent posttreatment period of 72 h was not effective in changing BCR-ABL expression of AR-230 cells (differences from control <35%, data not shown). However, as for K-562 cells, two rounds of this treatment reduced the BCR-ABL level significantly, as evidenced by Western blot (Fig. 7A). Densitometry analyses revealed 50% (TAT), 35% (ASO-b3a2), 46% (NSO), and 23% (ASO-bcr) of control BCR-ABL. After normalization to actin, the figures were 52% (TAT), 59% (ASO-b3a2), 96% (NSO), and 27% (ASO-bcr) of control BCR-ABL. The same treatment schedule resulted in 80% reduction of AR-230 colony growth (Fig. 7B). This effect contrasts with the distinctly lower inhibition of BCR-ABL level and clonogenicity of AR-230 cells after electrotransfection with the oligonucleotides NSO, TAT (mismatch), and the junction site-directed ASO-b3a2.

Treatment of murine LTBMCCs with different concentrations of ErPC₃ (5–40 μm) did not inhibit the growth of CFU-GM. In contrast, 40 μm ErPC₃ induced a significant stimulation of colony growth (Fig. 8A). To control for a possible influence of the antisense therapy on normal hematopoietic cells, nonadherent LTBMCC cells were exposed to 10 μm ASO-bcr administered by electroporation. Remarkably, at 24 h after the electroporation, there was no reduction in CFU-GM colony formation (Fig. 8B).

Previous investigations have shown that alkylphosphocholines have very low if any clastogenic activity in vivo. To verify this expectation for ErPC₃, we performed cytogenetic analyses on bone marrow metaphase plates of ErPC₃-treated female and male mice. The results are summarized in Table 1.

The dose of 124 mg/kg (corresponding to 100 mg/kg of hexadecylphosphocholine) was well tolerated within the observation period of 48 h as compared with 24 h, chosen for the positive control, mitomycin C. Even after this prolonged observation period, there was no significant increase in the percentage of bone marrow cells with chromosomal aberrations nor in the mitotic index. In contrast, mitomycin C increased the number of cells with aberrations 11-fold and decreased the mitotic index by a factor of almost four.

APCs are candidates for antileukemic therapy because they combine properties such as cytotoxic activity, induction of apoptosis, and stimulation of normal hematopoietic cells (11, 13, 14, 22, 23). In the present study, we focused on ErPC₃, a derivative closely related to erucylphosphocholine, which is the prototype of i.v.-injectable APCs (12). This subgroup of APCs is appropriate for systemic i.v. therapy because it lacks significant hemolytic activity. Comparing the two agents, ErPC₃ forms micellar and water-clear solutions, whereas erucylphosphocholine tends to arrange in gel-like lamellar structures at temperatures <37°C. Both compounds are equally efficient against human leukemia cells and combine cytotoxic properties as well as those restoring cellular adherence (11, 14). However, we have selected ErPC₃ for a more detailed analysis for its better solubility and increased bone marrow stimulatory activity. The experimental data obtained indicate that ErPC₃ can significantly reduce the growth of K-562 cells that have been described as relatively resistant to ether lipids and APCs (11, 22). However, the inhibitory effect of ErPC₃ is not sufficient to completely inhibit the colony growth of this BCR-ABL-expressing cell line. Our assumption was that specific inhibition of the BCR-ABL protein would help to further reduce the K-562 colony growth. Our panel of human CML-derived cell lines consisted of three cell lines with myeloid (K-562, LAMA-84, and AR-230) and two cell lines of lymphoid phenotype (CML-T1 and BV-173). They express the typical CML oncoprotein, although they differ by the type of fusion site: CML-T1 and BV-173 cells contain a b2a2 fusion, K-562 and LAMA-84 are characterized by a b3a2 fusion (15). AR-230 cells express a fusion protein with a molecular mass of 230 kDa because of the insertion of 180 amino acids encoded by 3′ exons of bcr (B4 to C3; Ref. 24). Their resistance against APCs (14) correlates with the respective level of the BCR-ABL expression as evidenced by Western blot (Fig. 5). K-562 and AR-230 cells showed the highest expression levels among the five cell lines investigated. Single exposure to an ASO directed against the transcription start of bcr caused reduced BCR-ABL expression in those three cell lines with genuinely low BCR-ABL levels, but AR-230 and K-562 cells responded only after repeated exposure. Recently, it has been shown that increased levels of BCR-ABL mediated by gene amplification are associated with and probably determine the resistance and growth factor independence of Ph+ cell lines (25, 26); this finding explains why K-562 cells were among the most resistant in our panel. Their resistance to various antileukemic drugs is furthermore based on blockade of mitochondrial release of cytochrome c and activation of caspase-3, which results in an antiapoptotic effect (27). K-562 cells have been shown to carry multiple functioning bcr-abl copies (15) and were included into extended studies under the assumption that eradication of this cell line would predict also sensitivity of the other CML cell lines used. In partial contrast to the observation that BCR-ABL levels were reduced only after repeated electrotransfection, a single ASO-bcr transfection had already a significant inhibitory effect on K-562 colony numbers. For this discrepancy, we have developed the following explanation. Previous experiments have shown that electroporation treatment with the ASO-bcr has no overt toxic or growth-inhibitory effect on cell growth in suspension culture, as determined by 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfonyl)-2H-tetrazolium assay and cell count (data not shown). The reason for this inefficacy probably is that autocrine stimulation resulting from growth factor production overcomes an increasing growth factor dependency that is caused by the ASO-bcr-induced reduction of BCR-ABL levels. This situation in liquid culture contrasts with colony growth in semisolid medium, where cells that are capable of clonal growth have no benefit from paracrine growth factor production and need to rely on BCR-ABL sustained growth factor independence. Therefore, a slight reduction in BRC-ABL level, which was not discovered at 48 h after electrotransfection, supposedly resulted in the observed considerable reduction of K-562 colony growth. To further increase this effect, we combined ErPC₃ treatment with transfection of the antisense oligonucleotide. The combination treatment resulted in a nearly complete inhibition of the colonial growth of K-562 cells after 96 h of incubation because of a strong synergistic effect (Fig. 3B). This effect is detectable in a broad concentration window of both agents (Fig. 4B). The mechanistic basis of this synergy is largely unknown to date. We speculate that both agents increase the capability of the cell to undergo either apoptosis or differentiation. Experiments along these lines are planned to be performed in the near future. A similarly strong inhibitory effect was observed in CML blasts after combination treatment with mafosfamide and an ASO directed against the b2a2 junction site (28). CML blasts, however, generally contain lower amounts of BCR-ABL than K-562 and AR-230 cells, and therefore, the present combination appears to be more powerful and is expected to be more effective in eradicating CML blasts. Interestingly, the inhibitory activity depended on the time after electrotransfection because an incubation period of 48 h (24 h for ErPC₃ and the ASO, respectively; Fig. 3A) was much less effective than 96 h, i.e., 48 h, respectively (Figs. 3B and 4B). It was optimal to perform the antisense therapy after the APC treatment. Furthermore, the electrotransfection was of importance, because the bcr-directed ASO that has been known since 1994 (17) gained its remarkable efficacy because of the rectangular impulses with controlled duration used. The conditions of electrotransfection used in this study were in the same range as described as optimal for K-562 cells (29). The advantages of electrotransfection over other transfection techniques have been described (30, 31). This corresponds to our preliminary experiments with transfecting agents such as DOTAP and Lipofectamine that resulted in unsatisfactory transfection using DOTAP and in unspecific inhibition of colony formation after Lipofectamine (data not shown). It is conceivable that the simple phosphorothioate backbone of the ASO-bcr used can be chemically optimized. Such a development seems promising because its target, i.e., the transcription start of bcr, makes it universally active against BCR-ABL expression, even in unusual fusion types such as that of AR-230 cells (Fig. 7).

To check whether this therapeutic schedule can affect normal hemopoietic cells, ErPC₃ and ASO-bcr effects were studied in normal LTBMCCs as well. As expected, ErPC₃ did not reduce the number of CFU-GM-forming cells and even stimulated the colony number significantly at a concentration of 40 μm (Fig. 8A). Similar to other APCs investigated previously (11), ErPC₃ did not show any significant clastogenic or antimitotic effect on mouse bone marrow cells in vivo. The other combination therapy partner, the anti-bcr ASO administered by electrotransfection, also did not show any toxicity on hematopoietic progenitor cells (nonadherent murine LTBMCCs; Fig. 8B). Because transfection with the anti-bcr ASO did not significantly inhibit CFU-GM colony formation, it is derived that inhibition of BCR synthesis does not impair cellular functions of normal hematopoietic progenitor cells that are required for their proliferation capacity as assessed by a CFU assay.

The combination of an optimized anti-bcr ASO with ErPC₃ appears to be a promising treatment for purging of bone marrow from contaminating CML blasts, because the ex vivo situation allows us to apply the necessary electrotransfection. The treatment could also be useful for patients who have developed resistance against the tyrosine kinase inhibitor imatinib (Glivec, STI571), as has been described recently (32–34).

Taken together, our results indicate that the combination treatment of ErPC₃ and the ASO-bcr is synergistically active in CML cells and well tolerated by normal hematopoietic cells. The BCR-ABL expression level is responsible at least in part for the therapeutic resistance of CML-derived leukemic blasts. The combined application of the agents might be a useful tool for purging the bone marrow from contaminating leukemic cells before an autologous bone marrow transplantation.

We are indebted to Dr. Robert Owen for stylistic improvements of the manuscript.