Erythropoietin Promotes Resistance Against the Abl Tyrosine Kinase Inhibitor Imatinib (STI571) in K562 Human Leukemia Cells

970 Vol. 1, 970–980, November 2003 Molecular Cancer Research

Karin M. Kirschner and Kurt Baltensperger

Institute of Pharmacology, University of Bern, Bern, Switzerland

Abstract malignant transformation of cells. Recent advances in the Chronic myeloid leukemia is characterized by the treatment of CML have largely relied on imatinib mesylate Philadelphia chromosome translocation that causes (STI571, Gleevec), a rationally designed inhibitor of the Abl expression of Bcr-Abl, a deregulated tyrosine kinase. tyrosine kinase (3). Imatinib potently suppresses the tyrosine Imatinib mesylate (STI571, Gleevec), a therapeutically kinase activity of the Bcr-Abl fusion protein and induces used inhibitor of Bcr-Abl, causes apoptosis of Bcr- apoptosis in Bcr-Abl-positive cells (4–9). The success of this Abl-positive cells. In the leukemia cell line K562, we novel drug is based on Bcr-Abl’s central place as the single observed spontaneous resistance to imatinib at very low most important component to maintain factor independence of frequencies when cells were exposed to the drug (1 MM) CML cells. The efficacy of imatinib in the treatment of chronic for more than 4 weeks. Surprisingly, in the presence of phase CML is well documented (8). Resistance resulting in erythropoietin (Epo), K562 cells were temporarily able to failure of treatment at initiation of therapy seems to be rare. sustain proliferation in the presence of imatinib, and However, several recent reports indicate that patients in the imatinib-resistant clones could be isolated with high accelerated phase and blast crisis that first responded well but frequencies. From such imatinib-resistant, Epo- relapsed during or after treatment had acquired resistance to dependent clones, sublines could be established that imatinib (10, 11). These observations prompted a number of were resistant to imatinib in the absence of Epo. investigations into potential mechanisms enabling leukemic Mitogen-activated protein (MAP) kinase activity was cells to evade the drug regimen. inhibited by imatinib treatment but could be partially Recent reports suggest the existence of at least five types of restored by Epo. Inhibition of MAP kinase or phospha- basic mechanisms for the development of imatinib resistance: tidylinositol 3-kinase blocked the protective effect of gene mutations in Bcr-Abl interfering with imatinib binding Epo. The data suggest that K562 cells acquire factor without abrogating ATP binding (10–13), elevated levels of the dependency under imatinib/Epo treatment, allowing Bcr-Abl gene product (10, 14), activation of alternate signaling them to escape from imatinib-induced, immediate cell pathways capable of ensuring cell survival (15), increased death. This pool of cells provides the basis for the levels of the multidrug resistance P-glycoprotein (16), and outgrowth of imatinib-resistant clones of unlimited elevated levels of plasma components that may bind imatinib proliferative capacity. Thus, Epo, an endogenous regu- and therefore lower its effective concentration (17). While most lator of hematopoiesis, promotes the development of publications focused on the nature of the primary internal resistance to imatinib. resistance mechanisms (18, 19), the potential role of endogenous or therapeutic cytokines as external promoters of Introduction resistance development was not investigated. Cytokines play The fusion protein Bcr-Abl, resulting from the Philadelphia an important role during the development of CML cells, mainly (Ph) chromosome translocation, represents a constitutively in the chronic phase of the disease. There is evidence for an active tyrosine kinase, which has been identified as the autocrine interleukin-3 (IL-3) loop in CML (20), and survival of primary CML cells in vitro is stimulated by cytokines (21). causative principle for most of the chronic myeloid leukemias + (CML) and in Ph+ acute lymphoblastic leukemias (1, 2). Effects of cytokines on imatinib-treated Ph cells were Bcr-Abl expression leads to cytokine independence and investigated in primary CML cells (22) and in cell lines transduced with viral vectors containing Bcr-Abl (4, 23, 24). In both cell systems, increased short-term survival rates of the cells were detected. Received 6/25/03; accepted 9/12/03. The costs of publication of this article were defrayed in part by the payment of For long-term analysis of cytokine effects on imatinib- page charges. This article must therefore be hereby marked advertisement in treated cells, primary cells are not suitable and cell lines derived accordance with 18 U.S.C. Section 1734 solely to indicate this fact. from CML patients are required. One of the most widely used Grant support: Swiss National Science Foundation grant 31-059124 and Swiss + Cancer League grant SKL 778-2-1999. and well-established Ph models is the K562 cell line, which Note: Current address of Dr. Karin M. Kirschner is at Johannes-Mu¨ller-Institut fu¨r was isolated from peripheral blood of a CML patient in blast Physiologie, Charite´- Universita¨tsmedizin Berlin, Campus Charite´-Mitte, Tuchol- skystrasse 2, 10117 Berlin, Germany. crisis (25). K562 cells are highly sensitive to imatinib treatment, Requests for reprints: Kurt Baltensperger, Pharmakologisches Institut der which induces partial differentiation followed by apoptosis. Universita¨t Bern, Friedbu¨hlstrasse 49, Postfach 51, CH-3010 Bern, Switzerland. Several laboratories reported the generation of imatinib- Phone: +41-31-632-3290; Fax: +41-31-632-4992. E-mail: [email protected] resistant sublines of K562 cells using a complex drug treatment Copyright D 2003 American Association for Cancer Research. protocol with gradually increasing drug concentrations over

several months (14, 15, 26). Development of resistant clones significantly lower when compared with a 4.6 F 0.4-fold showing molecular changes similar to those found in primary increase in cultures that received Epo simultaneously with CML cells demonstrates the general potential and suitability of imatinib (data not shown). These data suggest that in the K562 cells for analysis of resistance mechanisms (10, 14, 26). absence of Bcr-Abl signaling, K562 cells acquire factor K562 cells express the p210 form of the Bcr-Abl gene dependency. Imatinib itself does not initiate an immediate and fusion, which is the predominant form found in CML. Imatinib irreversible apoptotic response, and the loss of Bcr-Abl treatment inhibits tyrosine kinase activity of Bcr-Abl and, as a signaling may be compensated by alternative signaling consequence, abrogates phosphorylation of target proteins. through the Epo receptor. Thus, the status of target protein tyrosine phosphorylation has We hypothesized that such conditions may provide an been used as an indicator of Bcr-Abl activity and for the control environment for development of imatinib resistance because of drug efficacy in primary as well as in K562 and other cell cycling occurs in the presence of imatinib. Therefore, leukemia cell lines. Various downstream signaling components long-term cultures were initiated to assess whether cultures of the Bcr-Abl tyrosine kinase that are found in primary cells under imatinib and Epo would eventually cease proliferation are also present in K562 cells (2, 27, 28), allowing for the or whether subpopulations would survive the treatment and functional analysis of highly relevant signal transducers. In form continuous resistant cell cultures. Typically, cultures particular, the Raf/mitogen-activated protein (MAP) kinase and showed initial proliferation (phase I) when maintained in the phosphatidylinositol 3 (PI 3)-kinase pathways are activated by presence of imatinib and Epo but stopped proliferating after Bcr-Abl along with signal transducer and activator of 2–3 weeks (Fig. 1B). After a lag phase, net cell growth transcription 5 (STAT5). resumed after about 4 weeks of culture (phase II). Under these Examining the kinetics of cell death in imatinib-treated conditions, cell proliferation was sustained for at least cells, we noticed that erythropoietin (Epo) caused a marked 4 months. delay of apoptosis. We hypothesized that continuous cell Epo may rescue cells from imatinib-induced cell death by cycling in the presence of the Bcr-Abl inhibitor may promote simply inhibiting apoptosis while cells slowly cycle. Alterna- the development of imatinib resistance. Here, we describe that tively, cell cycling may be normal but cell death results in low Epo markedly enhanced the frequency of imatinib-resistant net growth rates of the entire culture. To distinguish these two K562 cells detected in long-term cultures. Imatinib treatment possibilities, proliferation and cell death were monitored in drives K562 cells into factor dependency that can be met by imatinib-treated K562 cells in the presence or absence of Epo. Epo, thereby ensuring their escape from imatinib-induced cell Propidium iodide-labeled cells were analyzed by flow cytom- death. Most of the Epo-dependent cultures ultimately give rise etry to assess the amount of cells with fragmented nuclei, to imatinib-resistant sublines that are able to sustain growth in indicating dead cells (Fig. 1C, gate 1). A large proportion (45%) the absence of Epo. Our data suggest that cytokines may raise of imatinib-treated cells died within the first 3 days. Most cells the frequency by which leukemic cells develop resistance had stopped proliferating by day 2 (indicated by the lack of the to imatinib. G2 peak in gate 2). The presence of Epo resulted in continued proliferation and much lower proportions of dead cells (5% at day 3). Prolonged exposure to Epo/imatinib resulted in a Results gradual increase of dead cells (Fig. 1C). This indicates that Epo Epo Rescues K562 Cells From Imatinib-Associated Cell delayed cell death in a large portion of the culture and Death and Supports Their Proliferation prevented cell death of a small subpopulation (indicated by the f K562 cells are highly sensitive to imatinib (EC50 0.1 AM; presence of the G2 peak). This subpopulation is sufficient to 6). They stop proliferation and undergo apoptosis within 2–3 sustain long-term cell growth. days of treatment with inhibitory concentrations of the drug In K562 cells, Bcr-Abl causes tyrosine phosphorylation of (1 AM), resulting in a rapid decline in numbers of viable cells multiple target proteins, which is reduced by imatinib treatment in culture (Fig. 1A). In several cell models, including primary (26). As expected, the level of overall tyrosine phosphorylation CML cells, cytokines were reported to prevent this apoptotic declined and remained low in Epo/imatinib-treated cells during response. We found that in K562 cells, Epo effectively phase I (Fig. 1D). Cells treated with imatinib alone showed only suppressed immediate imatinib-induced cell death (Fig. 1A), a partial reduction of overall tyrosine phosphorylation at day 2 although Epo without imatinib did not enhance proliferation (Fig. 1D, lane 3) yet stopped proliferating (lack of G2 peak) and of K562 cells (data not shown). Surprisingly, Epo not only underwent apoptosis by day 3. By contrast, more pronounced protected from cell death but also allowed cells to proliferate inhibition of overall tyrosine phosphorylation in the long-term for up to 3 weeks. Proliferation rates were reduced when cultures is still compatible with cell growth when Epo is compared with K562 cells maintained in the absence of present. CrkL, a major target of Bcr-Abl, is tyrosine imatinib (Fig. 1A). Withdrawal of Epo from imatinib/Epo- phosphorylated in K562 cells but remains predominantly treated cultures in this initial phase of treatment led to rapid dephosphorylated under imatinib and imatinib/Epo. Therefore, cell death, resulting in a sharp decline in cell numbers the effect of Epo appears not to be mediated through a (exemplified in the figure by withdrawal of Epo at day 15). restoration of Bcr-Abl-dependent tyrosine phosphorylation of Epo also proved to be effective in supporting cell growth target proteins. when added after imatinib: Epo exposure 20 h after initiation In summary, in phase I, Epo supports limited cell cycling of of imatinib treatment resulted in a 3.8 F 0.4-fold increase in imatinib-treated K562 cells and delays but does not prevent cell cell number after 72 h, which was slightly but not death. Under these conditions, cell cycling appears to be

FIGURE 1. Survival and resistant growth of imatinib-treated K562 cells in the presence of Epo. A. Growth of untreated K562 cells (squares), imatinib- treated cells (circles), imatinib/Epo-treated cells (closed triangles), or imatinib/Epo-treated cells after withdrawal of Epo (open triangles) was plotted as cumulative changes in cell numbers over 20 days. Epo withdrawal from imatinib/Epo-treated cells at day 15 is indicated by an arrow. The assay was performed thrice with similar results. B. Growth of imatinib/Epo-treated cells over 40 days. The experiment was performed as described in A. C. K562 cells and cells exposed to Epo, imatinib, or imatinib/Epo were collected at the indicated time points, stained with propidium iodide, and analyzed for DNA content by flow cytometry. Frequency histograms of propidium iodide intensities are represented. Nuclei were gated into fragmented (gate 1) and intact (gate 2) populations. Populations with intact nuclei represent cells in all phases of the cell cycle. Peaks representing G1 and G2 phase nuclei are indicated by arrows. The number given for gate 1 indicates the percentage DNA in fragmented nuclei. The experiment was performed six times with similar results. D. Immunoblotting of lysates of untreated cells, Epo-treated cells, and cells treated with imatinib or imatinib/Epo for the indicated times. Total cell lysates were analyzed by Western blot and probed with anti-phosphotyrosine (P-Tyr), anti-CrkL antibody, or anti-h-actin antibodies (actin) to ensure equal protein loading. Tyrosine-phosphorylated CrkL (P-CrkL) migrates slower than the unphosphorylated form (CrkL). The table indicates the percentages of phosphorylated and nonphosphorylated CrkL. Data represent one experiment out of three showing similar results.

independent of Bcr-Abl tyrosine kinase activity. In phase II, wells to 114 of 192 corresponding to an f100-fold increase in imatinib-resistant but Epo-dependent cells outgrow from phase I the frequency of imatinib-resistant cultures (estimated by cultures and establish cultures with long-term proliferative limiting dilution analysis). Withdrawal of Epo in resistant capacity. Table 1. Imatinib-Resistant Clones From K562 Cells Cul- Epo Promotes the Development of Resistance Against tured in the Presence or Absence of Epo Imatinib in K562 Cells To examine whether Epo promoted the development of Wells With Proliferating Imatinib-Resistant Cells (per Dish of 96 Wells) resistant cells from K562 cultures, the frequencies of imatinib-

resistant cells that could be obtained from K562 cultures in the Cells/well seeded Imatinib Imatinib/Epo presence or absence of Epo were analyzed. The ratio of Epo- independent imatinib resistance in K562 cells was investigated 3000 0 44 by application of inhibitory concentrations (1 AM) of imatinib 047 6000 1 54 over a period of 4 weeks. Only 2 of 192 wells showed viable 160 cells, when 6000 cells/well were seeded initially (Table 1). Frequency of resistance (Â10À6) f1.7 159 F 9 Inclusion of Epo in the assay raised the number of positive

clones obtained under Epo/imatinib treatment resulted in cell phosphobands in some Epo-independent, imatinib-resistant death over a period of 3 weeks, demonstrating that resistance at sublines (Fig. 2, C and D, respectively). As expected, CrkL this point required the continuous presence of Epo (data not was mostly phosphorylated in untreated K562 cells (P-CrkL; shown). The increase in resistance frequencies by 2 orders of Fig. 2, C and D) but was dephosphorylated in phase I cells as magnitude indicates that reacquired factor dependency facili- demonstrated by its increased mobility on the Western blot. In tates the outgrowth of imatinib-resistant clones in K562 cells phase III cells, no restoration of CrkL phosphorylation above with high efficiency. levels in phase II cells was detectable (Fig. 2D). Bcr-Abl/CrkL signaling therefore appears to be minor compared with wild- type K562 cells, indicating that reactivation of Bcr-Abl is Characterization of Epo-Dependent and Epo- unlikely to be responsible for resistance in these clones. Independent Imatinib-Resistant K562 Cells Consistently, phase III cells still showed high ratios of Imatinib-resistant K562 cells that were isolated in the hemoglobin-positive cells, arguing against reactivation of Bcr- limiting dilution assay (Table 1) were further characterized. The Abl (Fig. 2B). two clones that emerged from cells treated with imatinib alone The three phases of imatinib-resistant growth of K562 cells as well as six randomly selected Epo-dependent, imatinib- in the presence of Epo are schematically represented in Fig. 3. resistant clones from the imatinib/Epo-treated cultures (seeded The K562/Epo/imatinib cell model provides a new paradigm at the lower density of 3000 cells/well) were expanded. Neither for the generation of imatinib-resistant cells in vitro and for the of the two clones that were isolated from cultures and exposed analysis of their properties in the different phases of resistance to imatinib alone was capable of sustaining unlimited growth in development. imatinib. The six Epo-dependent, imatinib-resistant clones showed long-term cell proliferation. Clones 1–5 were Cytokines Other Than Epo Do Not Support Resistance examined for their ability to sustain growth in imatinib/Epo in K562 Cells at concentrations of up to 5 AM (Fig. 2A). Clones in imatinib/ Modulatory effects of cytokines on the efficacy of Epo could be maintained in continuous cultures for at least 4 imatinib-dependent cell death have been described for months without any indication of loss of resistance (data not granulocyte macrophage colony-stimulating factor (GM-CSF) shown). From these clones, resistant sublines (clones 1a–5a) and IL-3 (4, 6). Therefore, we tested whether imatinib-treated could be isolated that did no longer require Epo to sustain K562 cells would respond to cytokines other than Epo by proliferation in the presence of 1 AM imatinib (Fig. 2A). A sixth restoration of cell growth in the presence of imatinib. The Epo-dependent clone failed to establish Epo-independent influence of the cytokines stem cell factor (SCF), granulocyte growth and was therefore not included in the figure. colony-stimulating factor (G-CSF), GM-CSF, or IL-3 on the Apparently, Epo-dependent, imatinib-resistant cultures from proliferation of imatinib-treated K562 cells was monitored. phase II were capable of founding Epo-independent sublines None of the four factors was capable of stimulating the (denoted phase III) with high efficiency (over 80%). In the growth of imatinib-treated K562 cells (Fig. 4). Combinations absence of Epo, four of these sublines (clones 1a and 3a–5a) of two or of all four factors were equally ineffective (data not showed resistant growth up to a concentration of 2.5 AM of shown). These data indicate that proliferation of K562 cells imatinib and one (clone 2a) even proliferated at 5 AM of the treated with imatinib cannot be experimentally influenced by drug. Blocking Bcr-Abl activity by imatinib results in SCF, G-CSF, GM-CSF, or IL-3, the four major cytokines progression of K562 cells into erythroid differentiation, involved in normal hematopoietic progenitor cell develop- indicated by the production of hemoglobin (4). While Epo- ment. As a consequence, further analysis of cytokine-induced dependent, imatinib-resistant cells (phase II) showed very high effects on imatinib resistance in K562 cells focused on the ratios of hemoglobin-positive cells as detected by benzidine role of Epo. staining, ratios in imatinib-resistant sublines (phase III) were consistently reduced but still remained above background ratios of K562 cells that were treated with Epo alone (Fig. 2B). Epo-Sustained Growth of K562 Cells in Imatinib Depends Development of resistance appears to shift cells into a less on PI 3-Kinase and MAP Kinase Activities differentiated stage. PI 3-kinase and the MAP kinase cascade are essential and Several clinical studies on the efficacy of imatinib in the sufficient for erythroid differentiation and proliferation (29). To treatment of CML reported restoration of Bcr-Abl signaling in evaluate whether blocking of Epo receptor-dependent cellular imatinib-resistant cells on relapse of the disease. Therefore, signal transduction could influence Epo-induced proliferation of overall tyrosine phosphorylation was examined in Epo- imatinib-treated K562 cells in phase I of resistant growth, dependent and Epo-independent resistant clones by Western inhibitors of PI 3-kinase (LY294002) and the MAP kinase blotting (Fig. 2, C and D, respectively). While in untreated cascade (PD98059) were used. As expected (30), exposure to K562 cells numerous tyrosine-phosphorylated proteins were LY294002 resulted in strongly reduced growth rates of K562 detected, naive cells that were exposed to Epo/imatinib for 14 cells (Fig. 5). Doubling times of f7 days compared with 19 h in days (phase I) showed only few bands (Fig. 2, C and D). untreated cells were estimated. In cultures that also received However, Epo-dependent and Epo-independent resistant clones imatinib or imatinib/Epo, proliferation was completely blocked. (phases II and III, respectively) showed partial restoration of This result indicates that proliferation of imatinib/Epo-treated, tyrosine phosphorylation levels with significant differences nonresistant cells is strongly dependent on the activity of PI 3- between the individual clones, including the emergence of new kinase. Furthermore, PI 3-kinase appears to be required for

FIGURE 2. Analysis of imatinib-resistant clones. A. Growth of imatinib-resistant clones in the presence of 1 AM (squares), 2.5 AM (triangles), or 5 AM (circles) (circles) imatinib was plotted as cumulative changes in cell numbers. The measurement was performed by trypan blue exclusion cell counts. Epo-dependent clones (clones 1, 2, 3, 4, and 5) were cultured in the presence of Epo and imatinib (imatinib + Epo). Epo-independent clones (clones 1a, 2a, 3a, 4a, and 5a; imatinib) that originated from the Epo-dependent clones were cultured in 1 AM imatinib but in the absence of Epo for 7 weeks before the presented time course. Points, mean values of three independent measurements. B. Detection of erythroid differentiation by benzidine staining of hemoglobin-producing cells. Cells treated with Epo, imatinib, or imatinib/Epo for 2 days (black bars). Imatinib-resistant, Epo-dependent clones (1, 2, 3, 4, and 5; imatinib + Epo; dark gray bars). Imatinib-resistant, Epo-independent clones (1a, 2a, 3a, 4a, and 5a; imatinib; light gray bars) cultured for 7 weeks in imatinib alone before the experiment. Columns, mean values of six measurements. C and D. Immunoblotting of lysates of untreated cells, cells treated with imatinib/Epo for 14 days, Epo-dependent clones 1–5 (C; imatinib + Epo), and Epo-independent clones 1a – 5a (D; imatinib). Epo-independent clones (1a, 2a, 3a, 4a, and 5a) were treated with imatinib alone for at least 7 weeks before the experiment. Samples were probed with anti-phosphotyrosine (P-Tyr), anti-CrkL antibody, or anti-h-actin antibodies (actin) to ensure equal protein loading. Tyrosine-phosphorylated CrkL (P-CrkL) and unphosphorylated form (CrkL) are indicated by arrows. The tables indicate the percentages of phosphorylated and unphosphorylated CrkL. Both experiments were performed thrice with similar results.

FIGURE 3. Schematic representation summarizing the growth phases involved in Epo-promoted development of resistance in K562 cells. In the presence of Epo, K562 cells are capable of surviving imatinib treatment and proliferating for a limited time (phase I: survival). After 2 – 3 weeks, most cells die and imatinib-resistant clones emerge that are still Epo dependent (phase II: Epo-dependent resistant growth). Epo withdrawal (indicated by arrows) in phase I and early in phase II results in rapid loss of cell viability and collapse of the entire culture. Epo withdrawal later in phase II is again accompanied by massive cell death, but sublines of resistant clones emerge that are capable of sustaining cell proliferation in the absence of Epo (phase III: Epo-independent resistant growth). Overall levels of tyrosine phosphorylation (P-Tyr) during phase I growth gradually decreases and then increases again as imatinib-resistant cultures form (phases II and III). CrkL phosphorylation (P-CrkL), an indicator of Bcr-Abl activity, is reduced under imatinib treatment but does not rise again during development of resistance.

Epo-dependent cell proliferation under imatinib treatment. totic Bcl-xL (32, 33), was examined using an antibody Inactivation of the MAP kinase cascade by the MAP kinase directed against phosphorylated STAT5 (Tyr694). Several kinase 1 (MEK1) inhibitor PD98059 reduced growth rates of reports demonstrated constitutive activation of STAT5 in imatinib/Epo-treated K562 cells, and a sharp decline of cell K562 cells as a result of Bcr-Abl activity (27, 34). On viability was detected after prolonged exposure (>10 days). This addition of imatinib, STAT5 phosphorylation was reduced, but indicates that the activation of the MAP kinase cascade seems to a significant proportion of STAT5 phosphorylation remained be important for phase I survival of imatinib/Epo-treated cells that were rescued from immediate cell death by Epo. Activation of MAP kinase by Epo treatment of imatinib- exposed cells was confirmed in immunoblotting experiments with an antibody to activation-specific phosphorylation sites of MAP kinase (Fig. 6). Untreated K562 cells showed a high level of phosphorylated MAP kinase (lane 1), which is known to be due to activation by Bcr-Abl (31). In cells treated with imatinib, no phosphorylation of MAP kinase was detectable (lane 2). However, under these experimental conditions, Epo was capable of partially activating MAP kinase (lane 3). In cells treated with Epo alone, phosphorylation levels of MAP kinase were similar as in untreated controls (cf. lanes 1 and 4). The data shown in Fig. 6 represent samples collected after 24 h of treatment. Additional samples were preincubated with imatinib for 1 h and then exposed to Epo for time points ranging from 10 min to 2 h after Epo administration. These samples showed no Epo-dependent MAP kinase phosphorylation, indicating that activation of the MAP kinase signaling cascade by Epo occurs only after prolonged exposure (data not shown). Taken together, these data demonstrate that in Bcr-Abl-inhibited K562 cells sustained Epo signaling partially activates MAP kinase and that survival and proliferation of FIGURE 4. SCF, GM-CSF, G-CSF, and IL-3 do not enhance proliferation of imatinib-treated K562 cells. Proliferation rates of untreated cells in phase I depends on activation of both MAP kinase and control cells or cells exposed to imatinib + Epo, imatinib + SCF (25 ng/ml), PI 3-kinase. imatinib + GM-CSF (5 ng/ml), G-GSF (10 ng/ml), imatinib + IL-3 (3 U/ml), or imatinib alone were measured. The plot shows cumulative changes in Activation of STAT5, another direct target of the Epo cell numbers. Points, mean values of three independent measurements; receptor and a known modulator of expression of antiapop- bars, SE (smaller than symbols).

FIGURE 5. Proliferation of K562 cells treated with imatinib and Epo is sensitive to MEK1 and PI 3-kinase inhibitors. Growth of untreated, imatinib/Epo- treated, or imatinib-treated cells in the absence (control) or in the presence of the PI 3-kinase inhibitor LY294002 or the MEK1 inhibitor PD98059 is plotted as cumulative change in cell numbers.

detectable (Fig. 6, lane 2). Contrary to MAP kinase, the proliferation or whether it might require the simultaneous original phosphorylation level of STAT5 was not restored by presence of factors in fetal bovine serum that was used as cell the addition of Epo for 24 h (lane 3). These results do not culture supplement. Therefore, K562 cells were cultured in a exclude a possible role of STAT5 in the survival of imatinib/ serum-free defined medium for 3 days before the experiment Epo-treated K562 cells but demonstrate that the rescuing and then incubated with imatinib in the presence or absence of effect of Epo is not mediated by additional Epo-dependent Epo (Fig. 7C). Even under these conditions, Epo was capable STAT5 phosphorylation. of sustaining cell proliferation in the presence of imatinib, excluding the possibility that an unknown serum factor might Physiological Concentrations of Epo Induce Proliferation be essential to support the action of Epo. We conclude that of Imatinib-Treated K562 Cells Epo at physiological concentrations can support survival of In all experiments described thus far, Epo concentrations of imatinib-treated K562 cells. 4 units/ml were used to stimulate cell growth. Epo concentrations in the plasma of healthy subjects are much lower, Discussion ranging from 5 to 36 mU/ml (35). To test whether Epo at such Accumulation of mutations in CML from chronic phase to low concentrations was still effective, cell growth in the blast crisis due to high genetic instability is well documented presence of imatinib (1 AM) was determined with Epo (36). Patients that relapse during imatinib treatment frequently concentrations ranging from 0.05 mU/ml to 10 units/ml (Fig. 7A). Maximum stimulation of growth was achieved at Epo concentrations equal or higher than 1 unit/ml. The EC50 was calculated as f50 mU/ml. From this experiment, we conclude that physiological plasma concentrations of Epo are sufficient to enhance proliferation of imatinib-treated K562 cells. To confirm that proliferation of imatinib-treated K562 cells was due to Epo and not to some other component in the Epo formulation, a neutralizing anti-Epo antibody was used to block binding of Epo to its receptor. As shown in Fig. 7B, the antibody alone had no growth effect on imatinib-treated cells. Exposure of cells to Epo at concentrations of 0.1 and 1 unit/ ml, led to proliferation. At Epo concentrations of 0.1 unit/ml, the antibody abolished proliferation of K562 cells completely, which is consistent with Epo being the active principle. At the maximum effective concentration of Epo (1 unit/ml), the antibody in the concentration used (5 Ag/ml) was no longer able to block proliferation apparently due to substoichiometric levels with respect to Epo. This latter result demonstrates that FIGURE 6. Epo activates MAP kinase in imatinib-treated K562 cells. the buffer solution of the antibody preparation itself did not Immunoblotting of lysates of untreated, imatinib-treated, imatinib/Epo- interfere with proliferation, because the inhibitory effect could treated, or Epo-treated cells. Samples were probed with antibodies to phosphorylated p44/p42 MAP kinase (Thr202/Tyr204), MAP kinase, be overcome by higher Epo concentrations. Finally, we phosphorylated STAT5 (Tyr694), or STAT5. Western blots of one of two determined whether Epo alone acted to stimulate cell experiments with similar results.

show specific changes in Bcr-Abl itself, leading to reduced drug binding (37). Less intensely studied is the question whether genetic changes are already present before imatinib treatment or whether resistant clones may develop during treatment. Genetic changes evolving during imatinib treatment are not viewed as a potential reason for imatinib resistance (38). Such changes are expected to happen only in rare cases because Bcr-Abl-positive cells treated with imatinib immediately stop proliferation in vitro (4). With the lack of time for additional cell divisions, it appears unlikely that resistance could occur through de novo genetic changes after initiation of imatinib treatment. However, this assumption may prove invalid in situations that would allow for limited proliferation of Bcr-Abl-positive cells under the selective pressure of an inhibitory agent. The present study suggests a link between a cytokine- dependent rescue mechanism protecting Bcr-Abl-positive cells from imatinib-associated cell death and the development of imatinib resistance. Antiapoptotic effects of cytokines on Bcr- Abl-positive cell lines during imatinib treatment are well known. Mouse cell lines like 32D, Ba/F3, FDC-P1, and the human cell line MO7e become factor independent when transduced with Bcr-Abl. In the absence of cytokines, imatinib treatment results in cell death that can be prevented by IL-3 (4, 23, 24), demonstrating that factor dependency was reacquired through inhibition of the Bcr-Abl tyrosine kinase. However, these studies did not address the question of whether the combined treatment of cells with imatinib and appropriate cytokines could result in imatinib resistance. Our studies provide the first evidence that cells that are capable of reestablishing cytokine dependency may escape effective drug treatment in the short term and acquire imatinib resistance in the long term. Because the protective effect could even be detected at physiologically low concentrations of Epo, the concept of cytokine-assisted development of resistance may also be of clinical relevance. This conclusion relies on the use of a CML cell line and not on primary CML cells. Therefore, experiments to verify the cytokine effect in primary cells would be necessary. However, long-term cultures of primary leukemia cells that would be required for detecting the outgrowth of resistant clones are not yet feasible. Signaling pathways activated by the Epo receptor supported survival and substituted for Bcr-Abl signaling for a limited time. MAP kinase and PI 3-kinase, but not STAT5, a transcription factor activated through Janus kinase 2 binding to the Epo receptor (32), appear to be critically involved in phase I. Our findings are consistent with the observation that activation of the PI 3-kinase and the MAP kinase cascades are FIGURE 7. Epo in physiological concentrations induces proliferation in sufficient for proliferation of Epo-dependent primary cells (29) imatinib-treated cells that can be blocked by a neutralizing antibody to and suggest that similar mechanisms may regulate proliferation Epo. A. Dose-response curve for the stimulatory effect of Epo on imatinib- treated K562 cells. Cell numbers were determined after 3 days of in normal primary cells and Epo-dependent proliferation in the treatment. Starting cell numbers were 5 Â 104. Points, mean values of four imatinib-treated K562 cell line. We conclude that the independent cultures; bars, SE. The curve represents a nonlinear least- mechanism that protects K562 cells from imatinib-associated square fit to the data points. B. Growth of imatinib-treated K562 cells in the presence of an inhibitory anti-Epo antibody and increasing concentrations cell death involves PI 3-kinase and MAP kinase but is of Epo. K562 cells (1 Â 105) were seeded, and final cell numbers were independent of STAT5 activation. The K562 cell line may measured after 3 days of Epo treatment in the presence or absence of serve as a model system to further analyze the roles of these antibody. Columns, mean of three independent cultures. Bars, SE. *, Significant difference between the two values (analyzed by Student’s t test, signaling cascades in the promotion of resistance development. P < 0.01). C. Growth of untreated, imatinib/Epo-treated, Epo-treated, or Beside Epo, no other factors were effective in K562 cells imatinib-treated cells in medium containing fetal bovine serum (serum)or BIT, a defined serum substitute (serum free), was determined. The to protect from imatinib-induced cell death. This does not cumulative changes in cell numbers are plotted. mean that other cytokines may not be effective in other cell

lines or primary CML cells. Published data on the effects of 10% fetal bovine serum, 2 mM Glutamax I (Life Technologies), cytokines on K562 cells are not conclusive. In particular, the 1mM sodium pyruvate, 50 units/ml penicillin, and 50 Ag/ml presence of functional receptors for IL-3, G-CSF, GM-CSF, and streptomycin. Serum-free cultures of K562 cells were main- SCF has not been established. However, expression of the Epo tained in Iscove medium (Iscove’s modification of Dulbecco’s receptor in K562 cells is well documented (39). K562 cells may MEM) supplemented with 20% BIT-9500 serum substitute thus present a model system for further studies on the protective containing bovine serum albumin, insulin, and transferrin effect of Epo as a representative for other cytokines. (StemCell Technologies, Vancouver, British Columbia, Can- Investigations whether Epo or other cytokines are able to ada), 2 mM Glutamax I, 1 mM sodium pyruvate, 50 units/ml promote resistance in primary CML cells are limited by the fact penicillin, and 50 Ag/ml streptomycin. Cells were kept at 37jC that these CML cells only show short-term survival in vitro. in a humidified atmosphere of 95% air and 5% CO2. The advantage of the K562/Epo/imatinib cell model for Proliferating cells were passaged every 2–3 days and seeded analysis of imatinib resistance lies in the clear distinction and at a density of 105 cells/ml. Slowly proliferating or non- accessibility of the different phases during resistance develop- proliferating cultures were kept at a density of 5 Â 105 cells/ml ment. This provides a basis for better understanding the and medium was partially (50%) replaced every 2–3 days. individual steps involved in cytokine-promoted development of Hemoglobin-producing cells were detected by benzidine resistance. Further analysis of the involvement of the MAP staining of cellular suspensions (40). kinase and PI 3-kinase pathways in phase I should provide information on the mechanisms that protect cells from imatinib- Cell Proliferation Assay induced apoptosis. Analysis of cells in phase II should give Proliferation of cell cultures was monitored by counting insights into the kinetics of resistance development. Finally, viable cells based on trypan blue exclusion. Cumulative cell analysis of phase III cells is likely to identify in this model numbers were calculated by multiplying the number of cells per system some of the known molecular correlates to clinically culture by the splitting factor. observed resistance mechanisms, such as gene amplification or point mutations in Bcr-Abl. In this model, the critical cellular Cell Cycle and Viability Assay changes allowing for the continuous proliferation under the Viability was analyzed by DNA flow cytometry (41). In selective pressure of imatinib may now be examined. brief, cells were resuspended in hypotonic fluorochrome In conclusion, in the K562/Epo/imatinib model system, a solution containing 50 Ag/ml propidium iodide, 0.1% sodium new modality of progression toward resistance was identified. citrate, and 0.1% (v/v) Triton-X 100 and incubated at 4jCfor Endogenous cytokines may assume an unexpected role in the 6 h before analysis. development of imatinib resistance by supporting factor- dependent growth. Antibodies The following commercially available monoclonal antibodies were used as recommended by the manufacturer: anti- Materials and Methods MAP kinase 1/2 (Upstate Biotechnology, Lake Placid, NY), Reagents anti-phospho-p44/p42 MAP kinase (Thr202/Tyr204) E10, anti- Cell culture media and supplements were purchased from phospho-STAT5 (Tyr694), anti-CrkL (32H4), anti-phosphotyr- Life Technologies, Inc. (Basel, Switzerland). Epo, GM-CSF, osine (P-Tyr-100; Cell Signaling, Beverly, MA), anti-STAT and IL-3 were generous gifts from Cilag AG (Schaffhausen, (sc-835; Santa Cruz Biotechnology Inc., Santa Cruz, CA). Switzerland), Werthenstein-Chemie AG (Schachen, Switzer- Antibodies to h-actin and neutralizing antibodies to human Epo land), and Novartis Pharma AG (Basel, Switzerland), respec- were purchased from Sigma-Aldrich, Inc. (St. Louis, MO) as tively. SCF and G-CSF were purchased from PeproTech well as the horseradish peroxidase-conjugated secondary anti- (London, United Kingdom). Epo was used at concentrations bodies to mouse IgG (whole molecule) and rabbit IgG (whole of 4 units/ml if not otherwise indicated. Imatinib was kindly molecule). provided by Dr. Elisabeth Buchdunger from Novartis Pharma and was prepared as a 10-mM stock solution in DMSO. Imatinib was used at a final concentration of 1 AM. Working solutions Immunoblotting were diluted in cell culture media and added directly to cells Cells were washed thrice with Dulbecco’s PBS (Sigma). A with no more than 0.1% of final DMSO concentration. sample was removed before the third washing step for protein LY294002 and PD98059 were purchased from Sigma (Buchs, determination using the bicinchoninic acid method (Pierce Switzerland) and Alexis Biochemicals (Lausen, Switzerland) Chemical Co., Rockford, IL). The cell pellet was resuspended 6 and used as recommended at concentrations of 50 and 20 AM, in an appropriate amount (100 Al/10 cells) of sample buffer respectively. Unless otherwise mentioned, all chemicals (62.5 mM Tris-HCl [pH 7.5], 2% SDS, 10% (v/v) glycerol, (analytical grade) were from Sigma, Fluka (Buchs, Switzerland) 50 mM DTT, and 0.01% bromophenol blue) and lysed by or Merck AG (Dietikon, Switzerland). sonication for 10–15 s. Samples were denatured for 5 min at 95jC. Fifteen micrograms of protein/sample were resolved on 12% SDS polyacrylamide gels and electrophoretically trans- Cell Culture ferred to nitrocellulose filters. The membranes were blocked K562 cells were purchased from the American Type Culture with 5% nonfat dry milk in TBS-Tween (50 mM Tris-HCl Collection (Rockville, MD) and maintained in RPMI 1640, [pH 7.5], 150 mM NaCl, 1% Tween 20), incubated with the

indicated antibodies for immunostaining, and visualized by 7. Druker, B. J., Sawyers, C. L., Kantarjian, H., Resta, D. J., Reese, S. F., Ford, J. M., Capdeville, R., and Talpaz, M. Activity of a specific inhibitor of the BCR- chemiluminescence and exposure to Hyperfilm (ECL Plus ABL tyrosine kinase in the blast crisis of chronic myeloid leukemia and acute Western Blotting Detection System, Amersham, Buckingham- lymphoblastic leukemia with the Philadelphia chromosome. N. Engl. J. Med., shire, United Kingdom). As molecular size markers, a 344: 1038 – 1042, 2001. prestained SDS-PAGE low range standard (Bio-Rad Laborato- 8. Druker, B. J., Talpaz, M., Resta, D. J., Peng, B., Buchdunger, E., Ford, J. M., Lydon, N. B., Kantarjian, H., Capdeville, R., Ohno-Jones, S., and Sawyers, C. L. ries AG, Reinach, Switzerland) was used. Intensities of P-CrkL Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in and CrkL bands on Hyperfilm sheets were determined using a chronic myeloid leukemia. N. Engl. J. Med., 344: 1031 – 1037, 2001. optical digitizing system and the QuantityOne software package 9. Mauro, M. J. and Druker, B. J. STI571: targeting BCR-ABL as therapy for (Geldoc, Bio-Rad Laboratories). CML. Oncologist. 6: 233 – 238, 2001. 10. Gorre, M. E., Mohammed, M., Ellwood, K., Hsu, N., Paquette, R., Rao P. N., and Sawyers, C. L. Clinical resistance to STI-571 cancer therapy caused by BCR- Analysis of Resistance Frequencies and Selection of ABL gene mutation or amplification. Science, 293: 876 – 880, 2001. 11. von Bubnoff, N., Schneller, F., Peschel, C., and Duyster, J. BCR-ABL gene Resistant Cultures mutations in relation to clinical resistance of Philadelphia-chromosome-positive Cells were seeded into 96-well dishes (3000 or 6000 cells/ leukemia to STI571: a prospective study. Lancet, 359: 487 – 491, 2002. well). The culture medium was supplemented with imatinib 12. Ricci, C., Scappini, B., Divoky, V., Gatto, S., Onida, F., Verstovsek, S., (1 AM) and Epo (4 units/ml) or imatinib alone. Fifty percent of Kantarjian, H. M., and Beran, M. Mutation in the ATP-binding pocket of the ABL kinase domain in an STI571-resistant BCR/ABL-positive cell line. Cancer Res., the medium was changed every third day without removing any 62: 5995 – 5998, 2002. cells. Four weeks after initial seeding, wells with proliferating 13. Roumiantsev, S., Shah, N. P., Gorre, M. E., Nicoll, J., Brasher, B. B., cells were identified using a light microscope. The frequency of Sawyers, C. L., and Van Etten, R. A. Clinical resistance to the kinase resistant clones was calculated using the limiting dilution inhibitor STI-571 in chronic myeloid leukemia by mutation of Tyr-253 in the Abl kinase domain P-loop. Proc. Natl. Acad. Sci. USA, 99: 10700 – 10705, estimate (42). Briefly, the logarithm of the percentage of wells 2002. with no cell growth was plotted against the number of cells 14. le Coutre, P., Tassi, E., Varella-Garcia, M., Barni, R., Mologni, L., Cabrita, seeded per well. A linear regression line was fitted to the two G., Marchesi, E., Supino, R., and Gambacorti-Passerini, C. Induction of resistance to the Abelson inhibitor STI571 in human leukemic cells through gene data points (3000 and 6000 cells seeded) originating at the a amplification. Blood, 95: 1758 – 1766, 2000. priori value for no cells seeded (100%). The limiting dilution 15. Donato, N. J., Wu, J. Y., Stapley, J., Garllick, G., Lin, H., Arlinghaus, R., and estimate was derived from the intercept with the abscissa at the Talpaz, M. BCR-ABL independence and LYN kinase overexpression in chronic ordinate value of 37%, which represents the percentage of myelogenous leukemia cells selected for resistance to STI571. Blood, 101: 690 – 698, 2003. empty wells expected at limiting dilution. For further analysis 16. Mahon, F. X., Deininger, M. W., Schultheis, B., Chabrol, J., Reiffers, of resistant cultures, cells from individual wells were expanded J., Goldman, J. M., and Melo, J. V. Selection and characterization of BCR- in imatinib and Epo (1 AM and 4 units/ml, respectively) or ABL positive cell lines with differential sensitivity to the tyrosine kinase imatinib alone after removing Epo by two wash cycles. inhibitor STI571: diverse mechanisms of resistance. Blood, 96: 1070 – 1079, 2000. 17. Gambacorti-Passerini, C., Barni, R., le Coutre, P., Zucchetti, M., Cabrita, G., Cleris, L., Rossi, F., Gianazza, E., Brueggen, J., Cozens, R., Pioltelli, P., Pogliani, Data Analysis E., Corneo, G., Formelli, F., and D’Incalci, M. Role of a1 acid glycoprotein in the Statistical analyses were performed using the software in vivo resistance of human BCR-ABL(+) leukemic cells to the abl inhibitor package GraphPad Prism, version 2.0c (GraphPad Software, STI571. J. Natl. Cancer Inst., 92: 1641 – 1650, 2000. Inc., San Diego, CA). Error values refer to the SE. 18. Barthe, C., Cony-Makhoul, P., Melo, J. V., and Mahon, J. R. Roots of clinical resistance to STI-571 cancer therapy. Science, 293: 2163, 2001. 19. Anastasiadou, E. and Schwaller, J. Role of constitutively activated protein Acknowledgments tyrosine kinases in malignant myeloproliferative disorders: an update. Curr. Opin. We thank Dr. E. Buchdunger from Novartis Pharma AG (Basel, Switzerland) for Hematol., 10: 40 – 48, 2003. providing imatinib and handling the manuscript preview process as part of the 20. Jiang, X., Lopez, A., Holyoake, T., Eaves, A., and Eaves, C. Autocrine Material Transfer Agreement and Dr. H. Porzig for critically reading the production and action of IL-3 and granulocyte colony-stimulation factor in chronic manuscript and helpful suggestions. myeloid leukemia. Proc. Natl. Acad. Sci. USA, 96: 12804 – 12809, 1999. 21. Luna-Bautista, F., Sanchez-Valle, E., Ayala-Sanchez, M., Morales-Polanco, M., Meillon-Garcia, L., Benitez-Bribiesca, L., and Mayani, H. Kinetics of References hematopoiesis in bone marrow cultures from patients with chronic myeloid 1. Sawyers, C. L. Chronic myeloid leukemia. N. Engl. J. Med., 340: 1330 – leukemia: effect of recombinant cytokines in dexter-type long-term cultures. 1340, 1999. Hematology, 8 (3): 155 – 163, 2003. 2. Deininger, M. W., Goldman, J. M., and Melo, J. V. The molecular biology of 22. Graham, S. M., Jo¨rgensen, H. G., Allan, E., Pearson, C., Alcorn, M. J., chronic myeloid leukemia. Blood, 96: 3343 – 3356, 2000. Richmond, L., and Holyoake, T. L. Primitive, quiescent, Philadelphia-positive 3. Buchdunger, E., Zimmermann, J., Mett, H., Meyer, T., Mu¨ller, M., Druker, stem cells from patients with chronic myeloid leukemia are insensitive to STI571 B. J, and Lydon, N. B. Inhibition of the Abl protein-tyrosine kinase in vitro and in in vitro. Blood, 99: 319 – 325, 2002. vivo by a 2-phenylaminopyrimidine derivative. Cancer Res., 56: 100 – 104, 1996. 23. Dorsey, J. F., Cunnick, J. M, Lanehart, R., Huang, M., Kraker, A. J., Bhalla, 4. Druker, B. J, Tamura, S., Buchdunger, E., Ohno, S., Segal, G. M., Fanning, S., K. N., Jove, R., and Wu, J. Interleukin-3 protects Bcr-Abl transformed Zimmermann, J., and Lydon, N. B. Effects of a selective inhibitor of the Abl hematopoietic progenitor cells from apoptosis induced by Bcr-Abl tyrosine tyrosine kinase on the growth of Bcr-Abl positive cells. Nat. Med., 5: 561 – 566, kinase inhibitors. Leukemia, 16: 1589 – 1595, 2002. 1996. 24. Sun, X., Layton, J. E., Elefanty, A., and Lieschke, G. J. Comparison of 5. Deininger, M. W., Goldman, J. M., Lydon, N., and Melo, J. V. The tyrosine effects of the tyrosine kinase inhibitors AG957, AG490, and STI571 on BCR- kinase inhibitor CGP57148B selectively inhibits the growth of BCR-ABL- ABL-expressing cells, demonstrating synergy between AG490 and STI571. positive cells. Blood, 90: 3691 – 3698, 1997. Blood, 97: 2008 – 2015, 2001. 6. Gambacorti-Passerini, C., le Coutre, P., Mologni, L., Fanelli, M., Bertazzoli, 25. Lozzio, C. B. and Lozzio, B. B. Human chronic myelogenous leukemia cell- C., Marchesi, E., DiNicola, M., Biondi, A., Corneo, G. M., Belotti, D., Pogliani, line with positive Philadelphia chromosome. Blood, 45: 321 – 334, 1975. E., and Lydon, N. B. Inhibition of the ABL kinase activity blocks the proliferation 26. Weisberg, E. and Griffin, J. D. Mechanism of resistance to the ABL tyrosine of BCR/ABL+ leukemic cells and induces apoptosis. Blood Cells Mol. Dis., 23: kinase inhibitor STI571 in BCR/ABL-transformed hematopoietic cell lines. 380 – 394, 1997. Blood, 95: 3498 – 3505, 2000.

27. de Groot, R. P., Raaijmakers, J. A., Lammers, J. W., Jove, R., and 34. Ilaria, R. L., Jr. and Van Etten, R. A. P210 and P190 (BCR/ABL) Koenderman, L. STAT5 activation by BCR-ABL contributes to transformation of induce the tyrosine phosphorylation and DNA binding activity of K562 Leukemia cells. Blood, 94: 1108 – 1112, 1999. multiple specific STAT family members. J. Biol. Chem., 271: 31704 – 28. Sonoyama, J., Matsumura, I., Ezoe, S., Satoh, Y., Zhang, X., Kataoka, Y., 31710, 1996. Takai, E., Mizuki, M., Machii, T., Wakao, H., and Kanakura, Y. Functional 35. N. W. Tietz, E. L. Pruden, R. A. McPherson, and S. A. Fuhrman (eds.), cooperation among Ras, STAT5, and phosphatidylinositol 3-kinase is required for Clinical Guide to Laboratory Tests. Philadelphia, PA: Saunders; 1995. full oncogenic activities of BCR/ABL in K562 cells. J. Biol. Chem., 277: 8076 – 36. Skorski, T. BCR/ABL regulates response to DNA damage: the role in 8082, 2002. resistance to genotoxic treatment and in genomic instability. Oncogene, 21: 29. Klingmu¨ller, U., Wu, H., Hsiao, J. G., Toker, A., Duckworth, B. C., Cantley, 8591 – 8604, 2002. L. C., and Lodish, H. F. Identification of a novel pathway important for 37. Gorre, M. E. and Sawyers, C. L. Molecular mechanisms of resistance to proliferation and differentiation of primary erythroid progenitors. Proc. Natl. STI571 in chronic myeloid leukemia. Curr. Opin. Hematol., 9: 303 – 307, 2002. Acad. Sci. USA, 94: 3016 – 3021, 1997. 38. Luzzatto, L. and Melo, J. V. Acquired resistance to imatinib mesylate: 30. Skorski, T., Kanakarja, P., Nieborowska-Skorska, M., Ratajczak, M. Z., Wen, selection for pre-existing mutant cells. Blood, 100: 1105, 2002. S. C., Zon, G., Gewirtz, A. M., Perussia, B., and Calabretta, B. Phosphatidy- linositol-3 kinase activity is regulated by BCR/ABL and is required for growth of 39. Fraser, J. K., Lin, F. K., and Berridge, M. V. Expression and modulation of Philadelphia chromosome-positive cells. Blood, 86: 726 – 736, 1995. specific, high affinity binding sites for erythropoietin on the human erythroleukemic cell line K562. Blood, 71 (1): 104 – 109, 1988 Jan. 31. Kang, C. D., Do, I. R., Kim, K. W., Ahn, B. K., Kim, S. H., Chung, B. S., Jhun, B. H., and Yoo, M. A. Role of Ras/ERK-dependent pathway in the 40. Orkin, S. H., Harosi, F. I., and Leder, P. Differentiation in erythroleukemic erythroid differentiation of K562 cells. Exp. Mol. Med., 31: 76 – 82, 1999. cells and their somatic hybrids. Proc. Natl. Acad. Sci. USA, 72: 98 – 102, 1975. 32. Damen, J. E., Wakao, H., Miyajima, A., Krosl, J., Humphries, R. K., Cutler, 41. Nicoletti, I., Migliorati, G., Pagliacci, M. C., Grignani, F., and Riccardi, R. L., and Krystal, G. Tyrosine 343 in the erythropoietin receptor positively C. A rapid and simple method for measuring thymocyte apoptosis by regulates erythropoietin-induced cell proliferation and Stat5 activation. EMBO J., propidium iodide staining and flow cytometry. J. Immunol. Methods, 139: 14: 5557 – 5568, 1995. 271 – 279, 1991. 33. Socolovsky, M., Fallon, A. E., Wang, S., Brugnara, C., and Lodish, H. F. 42. Henry, C., Marbrook, J., Vann, D. C., Kodlin, D., and Wofsy, G. In:B.B. Fetal anemia and apoptosis of red cell progenitors in Stat5aÀ/À5bÀ/À mice: a Mishell and S. M. Shiigi (eds.), Selected Methods in Immunology, pp. 138 – 152. direct role for Stat5 in Bcl-X(L) induction. Cell, 98: 181 – 191, 1999. San Francisco, CA: Freeman, 1981.

Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2003 American Association for Cancer Research. Erythropoietin Promotes Resistance Against the Abl Tyrosine Kinase Inhibitor Imatinib (STI571) in K562 Human Leukemia Cells11Swiss National Science Foundation grant 31-059124 and Swiss Cancer League grant SKL 778-2-1999. Note: Current address of Dr. Karin M. Kirschner is at Johannes-Müller Institut fu?r Physiologie, Charité - Universitätsmedizin Berlin, Campus Charité-Mitte, Tucholskystrasse 2, 10117 Berlin, Germany.

Karin M. Kirschner and Kurt Baltensperger

Mol Cancer Res 2003;1:970-980.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/1/13/970

Cited articles This article cites 38 articles, 23 of which you can access for free at: http://mcr.aacrjournals.org/content/1/13/970.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/1/13/970.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/1/13/970. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.