Sensitivity to imatinib therapy may be predicted by testing Wilms tumor gene expression and colony growth after a short in vitro incubation

Authors


Abstract

BACKGROUND

The objective of the current study was to verify the ability to predict response to imatinib therapy using in vitro assays to evaluate the inhibition of Wilms tumor gene (WT1) expression and colony growth after samples obtained from patients with chronic myelogenous leukemia (CML) before the start of treatment were subjected to short-term incubation with imatinib.

METHODS

WT1 transcript levels and colony growth in bone marrow (BM) samples from 23 patients with CML that was later identified as being responsive to imatinib and from 13 patients with CML that was later identified as not being responsive to imatinib were evaluated after incubation of these samples with imatinib at a concentration of 1 μM for 18 hours. In addition, real-time quantitative polymerase chain reaction (RQ-PCR) analysis of WT1 expression was performed during follow-up, and the results were analyzed for associations with cytogenetic response and with BCR/ABL transcript levels as determined using RQ-PCR analysis.

RESULTS

Before treatment, it was found that WT1 expression was elevated in BM samples obtained from all patients with CML. WT1 expression and colony growth were reduced significantly after an 18-hour incubation with imatinib in samples obtained from patients who were later identified as responders to treatment, but not in samples obtained from patients who did not experience responses to treatment. Inhibition of WT1 expression in vitro was associated with inhibition of imatinib-induced BCR-ABL tyrosine kinase activity, a finding that also has been made in studies involving certain Philadelphia chromosome (Ph)-positive and Ph-negative cell lines.

CONCLUSIONS

Inhibition of WT1 transcript levels after a short period of in vitro exposure of pretherapy BM samples to imatinib was correlated with inhibition of colony growth and may represent the basis for an easy test that is capable of predicting the sensitivity of CML to treatment with imatinib for individual patients. Cancer 2004. © 2004 American Cancer Society.

The molecular result of the t(9;22)(q34;q11) translocation, which gives rise to the Philadelphia chromosome (Ph) and characterizes chronic myelogenous leukemia (CML), is the fusion of the BCR gene and the ABL gene to form a chimeric gene, leading to the expression of fusion proteins with deregulated tyrosine kinase (TK) activity.1, 2 It has been recognized that this defect plays a key role in the pathogenesis of CML.3 TKs are enzymes that transfer phosphate groups from adenosine triphosphate to tyrosine residues on substrate proteins that, in turn, regulate cellular processes such as proliferation, differentiation, and survival.4 This function has made TKs attractive therapeutic targets for selective inhibition.

Imatinib (Gleevec, formerly STI571; Novartis, Basel, Switzerland) is a potent and selective competitive inhibitor of the BCR-ABL protein TK (PTK).5, 6 In recent years, this agent has demonstrated remarkable efficacy in inducing cytogenetic remission in patients with CML for whom previous treatment with interferon-alpha (IFN-α) was ineffective or intolerable, as well as in patients to whom it is administered at diagnosis. Nonetheless, there is a constant (but different) percentage of patients in each of these groups for whom treatment with imatinib is not effective.7–10 Several mechanisms of resistance to imatinib have been identified.11 In a substantial proportion of patients, the basis for resistance is a genetic change in the BCR-ABL gene itself (specifically, point mutations within the PTK or within other protein domains).12, 13 Other mechanisms include amplification or increased expression of the BCR-ABL gene12 and clonal evolution or overexpression of the Pgp protein.14 Resistance to imatinib can be classified into distinct clinical scenarios in which the mechanisms of resistance are likely to be quite different. These scenarios include 1) upfront resistance to imatinib and 2) resistance heralded by disease recurrence while the patient is receiving imatinib after an initial response. All of the mechanisms mentioned above appear to be related preferentially to the second type of resistance.

Reasons for the lack of conversion from a hematologic response to a cytogenetic response in patients with CML remains largely unknown. One possible explanation is that BCR-ABL-independent signals are responsible for the survival of Ph-positive clones in spite of effective targeting of the BCR-ABL signal transduction pathway.15, 16

At present, many investigators are attempting to identify clinical parameters that may be useful in predicting whether a cytogenetic response will be induced by imatinib, thereby allowing clinicians to manage patients with CML optimally. In the current report, we describe a system that measures in vitro inhibition of Wilms tumor gene (WT1) expression by imatinib to identify patients with cytogenetically resistant disease before the start of imatinib therapy.

WT1 is a tumor suppressor gene encoding a zinc finger transcription factor located at 11p13, which originally was identified for its involvement in Wilms tumor pathogenesis.17, 18 In normal hematopoietic cells, according to real-time, quantitative-polymerase chain reaction (RQ-PCR) analysis, WT1 is expressed at very low levels.19 In contrast, WT1 expression is elevated in several types of acute and chronic leukemia, including CML.20–22

Despite these findings, the significance of WT1 expression in leukemia awaits clarification, and regarding the role of this gene in leukemogenesis, we believe that its expression may represent the final endpoint of a number of different transforming pathways that are activated in leukemia cells. Apart from the biologic significance of WT1 overexpression in human leukemias, it is now clear that WT1 may represent a useful marker of leukemic hematopoiesis that not only is suitable for diagnostic and prognostic purposes19–22 but also eventually may be useful in the development of novel therapeutic approaches.20

MATERIALS AND METHODS

Patients and Control Individuals

After informed consent was obtained, 36 patients with CML who experienced failure following IFN-α treatment and who were enrolled in one of the imatinib protocols of the Italian Cooperative Study Group on CML were included in the study. Twenty-five patients had disease that was resistant to IFN-α (7 had hematologically resistant disease, 14 had cytogenetically resistant disease, and 4 had disease that progressed to an accelerated phase during IFN-α therapy), and 11 were unable to tolerate IFN-α treatment. The chronic and accelerated phases of CML were defined according to published criteria.23 Hematologic failure was defined as either hematologic resistance (failure to achieve a complete hematologic response after ≥ 6 months of IFN-α treatment) or disease recurrence after a complete hematologic response had been achieved. Cytogenetic failure was defined as either cytogenetic resistance (i.e., ≥ 65% of cells in metaphase were Ph positive after at least 1 year of IFN-α therapy) or disease recurrence after a major cytogenetic response had been achieved.

Inability to tolerate IFN-α was defined by the presence of any Grade ≥ 3 nonhematologic toxicity (as defined by the National Cancer Institute Common Toxicity Criteria).24 Cytogenetic responders were defined as patients who had a complete cytogenetic response (CCR; i.e., no cells in metaphase were Ph positive) or a major cytogenetic response (i.e., < 35% of cells in metaphase were Ph positive) after 1 year of treatment.

Twenty-three of the 36 patients in the current study had cytogenetic responses to imatinib during the first year of treatment; 16 experienced complete cytogenetic remission, and 7 experienced a major cytogenetic remission. The remaining 13 patients had CML that exhibited upfront cytogenetic resistance.

Before the start of imatinib therapy and during follow-up, all patients were evaluated by cytogenetic methods and by RQ-PCR analysis for expression of WT1 and BCR-ABL in both peripheral blood (PB) and bone marrow (BM). BM samples that were collected before imatinib therapy were available for in vitro incubation with imatinib.

Cytogenetic Analysis

For all patients who were included in the current study, cytogenetic analysis was performed after a 1-day culture of unstimulated BM cells was grown. G-banding karyotypes were analyzed and classified according to the International System for Human Cytogenetic Nomenclature.

Cell Lines

Imatinib-sensitive cell lines K562s and KCLs; their resistant counterparts, K562r and KCLr; and the Ph-negative control cell lines HL60 and Me-1 were used to investigate the effect on WT1 expression of inhibition of BCR/ABL TK activity. Imatinib was diluted in dimethyl sulfoxide (DMSO) stocked at a concentration of 1 mM at −20 °C. K562s, K562r, KCLs, KCLr, HL60, and Me-1 cells were incubated in RPMI-1640 medium at a concentration of 106 cells/mL for 6–12 hours and for 18 hours in a humidified atmosphere (37 °C, 5% CO2) with increasing doses of imatinib (0.5 μM, 1.0 μM, 5.0 μM, 10.0 μM, and 25.0 μM). Along with each experimental sample, an appropriate control sample treated with vehicle alone (DMSO) was tested. At the end of the incubations, cells were washed twice with phosphate-buffered saline (PBS) and resuspended in RPMI-1640 for RNA extraction, assessment of apoptosis, and Western blot analysis. In addition, to gain a better understanding of the correlation between the deregulated TK activity induced by BCR/ABL and WT1 overexpression, the COS cell line, represented by normal fibroblasts, was transfected with a BCR/ABL-expressing plasmid and evaluated for WT1 expression before and after BCR/ABL transfection and after incubation of transfected cells with 1 μM imatinib for 18 hours.

Transient Transfection

One day before transfection, the COS cell line was seeded in 6-well cell culture plates at a density of 1.5 × 105 cells per well. Transient transfection was performed using LipofectAMINE reagent (Life Technologies, Bethesda, MD) according to the manufacturer's protocol along with 1 μg BCR/ABL pcDNA3 plasmid (donated by Dr. Rocco Piazza, Italian National Cancer Institute, Milan, Italy).

Cell Separation Procedures and Incubation with Imatinib

After informed consent was obtained, BM cells were collected from patients with CML who were candidates for imatinib treatment before they started therapy. Mononuclear cells (MNCs) were separated by centrifugation on a Ficoll-Hypaque gradient and resuspended in RPMI-1640 (GIBCO, Grand Island, NY) at a concentration of 106 cells/mL. MNCs were incubated with imatinib at a concentration of 1 μM for 18 hours (37 °C, 5% CO2) in a humidified atmosphere. At the end of the incubation period, cells were washed twice. The incubated cells subsequently were used for RNA extraction, assessment of apoptosis, assessment of colony growth, and Western blot analysis. In addition, all patients who were included in the study had BM samples collected every 3 months during follow-up and analyzed for BCR/ABL and WT1 expression levels using RQ-PCR and cytogenetic methods.

Apoptosis Assay

Cell lines or aliquots of BM cells that were incubated with 1 μM imatinib for 18 hours and control samples were labeled with fluorescein isothiocyanate–conjugated annexin V (annexin V–FITC) and with 7-aminoactinomycin D (7-AAD). In brief, cells were washed once in PBS and once in 1X binding buffer, and 5 μL annexin V–FITC and 5 μL 7-AAD were added to the cells. Cells were incubated at room temperature for 15 minutes, after which 300 μL 1X binding buffer was added and cells were analyzed by flow cytometry. Apoptotic cells were defined as being annexin V positive and 7-AAD negative.

Multilineage Colony-Forming Unit, Erythroid Burst, and Granulocyte-Macrophage–Colony-Forming Unit Assays

Aliquots of BM cells that were incubated with 1 μM imatinib for 18 hours and control samples were tested for their ability to give rise to colony growth in semisolid culture. Assays for multilineage colony-forming units (CFU-Mix), erythroid bursts (BFU-E), and granulocyte-macrophage–colony-forming units (CFU-GM) were carried out as described elsewhere.25 Progenitor cell growth was evaluated according to previously published criteria.26

Western Blot Analysis

Aliquots of BM cells that were incubated with 1 μM imatinib for 18 hours were assessed via Western blot analysis to demonstrate the blocking of BCR-ABL phosphorylation. In addition, cell lines (K562s, K562r, KCLs, KCLr, HL60, and Me-1) that had been incubated with increasing doses of imatinib (0.5 μM, 1.0 μM, and 5.0 μM) were analyzed using the same procedure. Western blot analysis was carried out as described elsewhere27 using the monoclonal antibody anti-pTyr (PY-99; Santa Cruz Biotechnology, Santa Cruz, CA) and the monoclonal antibody anti-β-actin (A-5316; Sigma, St. Louis, MO).

RQ-PCR Analysis of WT1, BCR-ABL, GUS, and β-2-Microglobulin

At the end of the incubation period, total RNA was extracted from imatinib-treated BM cells and cell lines according to standard procedures. The reverse transcriptase step was adapted from the BIOMED 1 protocol.28 RQ-PCR reactions and fluorescence measurements were made using the ABI PRISM 7700 Sequence Detection System (PE Applied Biosystems, Foster City, CA) according to previously published procedures.19

The following RQ-PCR primers and probe were used for WT1: exon 7, 5′-CAGGCTGCAATAAGAGATATTTTAAGCT-3′; exon 8, 5′-GAAGTCACACTGGTATGGTTTCTCA-3′; and exon 7, Fam-CTTACAGATGCACAGCAGGAAGCACACTG-Tamra. The following RQ-PCR primers and probe were used for GUS: exon 11, 5′-GAAAATATGTGGTTGGAGAGCTCATT-3′; exon 12, 5 ′-CCGAGTGAAGATCCCCTTTTTA-3′; and exon 12, Fam-CCAGCACTCTCGTCGGTGACTGTTCA-Tamra. The following RQ-PCR primers and probe were used for β-2-microglobulin: exon 2, 5′-GAGTATGCCTGCCGTGTG-3′; exon 4, 5′-AATCCAAATGCGGCATCT-3′; and exon 3, Fam-CCTCCATGATGCTGCTTACATGTCTC-Tamra. Finally, the following RQ-PCR primers and probe were used for BCR/ABL: BCR, 5′-TCCGCTGACCATCAATAAGGA-3′; ABL, 5′-CACTCAGACCCTGAGGCTCAA-3′; and ABL, Fam:5′-CCCTTCAGCGGCCAGTAGCATCTGA-Tamra.

The WT1 and BCR/ABL counts obtained by RQ-PCR analysis were normalized with respect to the number of GUS transcripts and are expressed as the number of WT1 copies per 104 copies of GUS. The WT1 counts obtained in COS cells were normalized with respect to the number of β-2-microglobulin transcripts and are expressed as the number of WT1 copies per 106 β-2-microglobulin copies.

RESULTS

WT1 Expression in CML

The analysis of WT1 expression in 72 healthy volunteers and the reanalysis using GUS (rather than ABL) as the control gene confirmed our previous observation of low WT1 transcript levels in normal BM and PB cells.19 The median number of WT1 copies per 104 GUS copies in normal BM cells was 19 (range, 0–82 WT1 copies per 104 GUS copies), and the corresponding median value in PB cells was 1.5 (range, 0–14 WT1 copies per 104 GUS copies).

The mean number of WT1 copies per 104 GUS copies in patients with chronic-phase CML was 3262 in BM samples (n = 78; range, 171–54,171 WT1 copies per 104 GUS copies) and 416 in PB samples (n = 69; range, 40–3193 WT1 copies per 104GUS copies), with a large degree of individual-to-individual variability in observed expression levels. These data are similar to those obtained using ABL in place of GUS as the control gene, as has been reported previously.21 The 36 patients who were included in the current study had WT1 values in BM samples that were similar to those observed in our enlarged series (mean, 3690 WT1 copies per 104GUS copies; range, 268–10,987 WT1 copies per 104GUS copies). Figure 1 shows that the mean number of WT1 copies before treatment in patients who later were identified as cytogenetic responders was 2012 WT1 copies per 104GUS copies; in contrast, in patients with resistant CML, the corresponding mean value was 6103 WT1 copies per 104GUS copies (P = 0.0006 [t test]). In addition, during follow-up, WT1 transcript levels corresponded to the degree of cytogenetic response and strictly paralleled the behavior of BCR-ABL transcript levels as evaluated by RQ-PCR (data not shown). In patients who had cytogenetic responses, WT1 levels decreased progressively according to the degree of cytogenetic response and returned to values more similar to those detected in BM samples from healthy patients. The mean number of WT1 copies detected at first observation of a CCR was 61, whereas the corresponding mean value for patients who achieved only a major cytogenetic response was 149. Among patients with resistant disease, high levels of the WT1 transcript, like those observed before treatment, were correlated with the absence of cytogenetic response.

Figure 1.

WT1 expression was evaluated at diagnosis (time point A) in bone marrow cells collected from patients who were identified as responders (red dots) or nonresponders (black dots) to imatinib therapy. Time point B represents the time at which complete cytogenetic remission or the best observed cytogenetic response was achieved in responders, and it represents 1 year of treatment in nonresponders.

Effect of Imatinib on the K562, KCL, HL60 and Me-1 Cell Lines

To test the influence on WT1 expression exerted by inhibition of the TK activity of BCR-ABL, we incubated K562s, K562r, KCLs, and KCLr cells and the Ph-negative cell lines HL60 and Me-1 with increasing doses of imatinib (0.5 μM, 1.0 μM, 5.0 μM, and 10 μM) for 18 hours. The optimal concentration of imatinib that inhibited tyrosine phosphorylation was determined by Western blot analysis. Figure 2 shows that after 18 hours of incubation with imatinib at a concentration of 0.5 μM, the level of tyrosine phosphorylation began to decrease and became almost undetectable at concentrations of 1 and 5 μM. Based on these data, we selected a concentration of 1 μM for the in vitro assay of primary CML cells.

Figure 2.

Western blot identification of the optimal concentration of imatinib for inhibiting tyrosine phosphorylation (p-Tyr). After 18 hours of incubation of K562 cells with 0.5 μM imatinib, tyrosine phosphorylation levels began to decrease, and these levels became nearly undetectable at concentrations of 1 μM and 5 μM. CTRL: control.

Apoptotic cell counts were evaluated in control samples and in samples that were incubated with 1 μM imatinib using flow cytometry for the detection of annexin V–positive cells. We noted only a slight increase in the percentage of apoptotic cells in treated K562s samples compared with untreated samples (13% vs. 7%). Similarly, for the K562r cell line, the proportion of apoptotic cells was 11% in treated samples and 10% in control samples. Similar results were observed in KCLs and KCLr cells (KCLs: 16% vs. 12%; KCLr: 15% vs. 14%).

The treated and control samples were tested for WT1 and BCR-ABL expression levels using RQ-PCR analysis. Figure 3 shows that the amount of WT1 transcript in incubated cells was inhibited strongly and that the reduction was more pronounced at higher doses. A 3-fold reduction was observed at a dose of 0.5 μM, a 9-fold reduction was observed at a dose of 1.0 μM, and an 11-fold reduction was observed at a dose of 5.0 μM in K562s cells. Similar results were obtained by incubating KCL cells with imatinib: a 1.5-fold reduction was observed at a dose of 0.5 μM, a 7.7-fold reduction was observed at a dose of 1.0 μM, and an 8.5-fold reduction was observed at a dose of 5.0 μM. In contrast, no significant inhibition of WT1 expression was detected in imatinib-resistant cell lines (K56r cells: 30,025 WT1 copies per 104GUS copies [treated] vs. 30,560 WT1 copies per 104GUS copies [untreated]; KCLr cells: 23,260 WT1 copies per 104GUS copies [treated] vs. 24,160 WT1 copies per 104GUS copies [untreated]). Under these culture conditions, imatinib incubation did not affect expression levels of BCR-ABL or GUS.

Figure 3.

Effects of imatinib incubation on WT1 transcript levels in K562s cells, K562r cells, KCLs cells, and KCLr cells, as well as in the Philadelphia chromosome-negative cell lines HL60 and Me-1. (The letters s and r in the names of these cell lines indicate sensitivity and resistance to imatinib, respectively.) After 18 hours of incubation, WT1 expression was strongly down-regulated in K562s cells and KCLs cells, but not in K562r or KCLr cells. This down-regulation also was observed in HL60 and Me-1 cells.

The same experiments were performed using the HL60 and Me-1 cell lines. The apoptosis assay revealed that 10% of cells in the control HL60 sample were apoptotic, compared with 13% in the treated HL60 sample, and that 11% of cells in the control Me-1 sample were apoptotic, compared with 14% in the treated sample. Imatinib incubation did not affect WT1 and ABL transcript levels in either HL60 cells (24,496 WT1 copies per 104GUS copies [treated] vs. 23,476 WT1 copies per 104GUS copies [untreated]) or Me-1 cells (16,543 WT1 copies per 104GUS copies [treated] vs. 16,654 WT1 copies per 104GUS copies [untreated]).

WT1 Expression in the COS Cell Line and in a COS BCR/ABL-Transfected Cell Line before and after Imatinib Incubation

In another analysis aimed at determining whether WT1 expression was linked directly to BCR-ABL TK activity, we evaluated WT1 expression in the COS cell line before and after transfection with BCR-ABL. Therefore, WT1 expression was determined in COS cells, in COS cells that were transfected with BCR/ABL, and in the same cells after incubation with 1 μM imatinib for 18 hours (Fig. 4).

Figure 4.

WT1 expression in COS cells before and after transfection with BCR/ABL and after incubation of transfected cells with 1 μM imatinib for 18 hours.

Untransfected COS cells expressed very low levels of the WT1 transcript (13 WT1 copies per 106 β-2-microglobulin copies). In contrast, after transfection with BCR/ABL, these cells expressed very high levels of WT1 (14,230 WT1 copies per 106β-2-microglobulin copies). Incubation with 1 μM imatinib for 18 hours decreased WT1 transcript levels significantly, to 4890 WT1 copies per 106β-2-microglobulin copies. Again, BCR/ABL transcript levels did not decrease significantly after incubation with imatinib (data not shown).

WT1 Expression in BM Samples after In Vitro Incubation with Imatinib

Thirty-six BM samples that were collected before imatinib therapy from patients with CML were incubated with 1 μM imatinib under the same conditions that were used for the treatment of cell lines. After 1 year of follow-up, 13 of 36 patients proved to have disease that was cytogenetically resistant to imatinib, whereas the remaining 23 patients experienced responses to treatment (including 16 CCRs and 7 major cytogenetic responses).

Figure 5 shows that in all samples collected at diagnosis from patients who subsequently had a cytogenetic response to imatinib, in vitro incubation with imatinib induced marked and significant inhibition of WT1 transcript levels (P = 4.85 × 10−6 [Student t test for paired data]), with the mean change in these levels being a 16.6-fold reduction. The mean number of WT1 copies in control samples was 2555 per 104GUS copies, compared with 245 per 104GUS copies in treated cells. In contrast, in vitro imatinib treatment did not result in significant inhibition of WT1 expression in patients who later were identified as having imatinib-resistant disease. In this latter group of patients, the mean number of WT1 copies was 5904 per 104GUS copies in control samples, compared with 5713 per 104GUS copies in treated samples (P = 0.57) (Fig. 6).

Figure 5.

WT1 transcript levels evaluated in control samples and in imatinib-incubated cells collected before imatinib treatment from 13 patients who subsequently were identified as having imatinib-resistant disease (black dots) and from 23 patients who subsequently had complete or major cytogenetic responses to imatinib therapy (red dots).

Figure 6.

Mean WT1 copy numbers in control (CTRL) samples (black columns) from 23 responders and from 13 patients with imatinib-resistant disease and in accompanying samples that were treated in vitro with imatinib for 18 hours at a concentration of 1 μM (gray columns).

Under these culture conditions, BCR-ABL transcript levels were weakly attenuated by imatinib treatment in both patients with responsive disease and patients with resistant disease, and there was no significant difference between the two groups (P = 0.36). Flow cytometric analyses for the determination of annexin V–positive cell counts were performed for all treated and untreated cells and demonstrated that the percentage of apoptotic cells present after imatinib incubation was not significantly greater than the corresponding percentage in untreated control samples (P = 0.06).

Colony Growth Assay

Figure 7 shows that when BM samples obtained from patients who experienced cytogenetic responses were incubated with 1 μM imatinib for 18 hours, CFU-Mix (5.7 colonies per 50,000 mononuclear cells [untreated] vs. 0.7 colonies per 50,000 mononuclear cells [treated]; P = 1.0 × 10−11), BFU-E (19.0 colonies per 50,000 mononuclear cells [untreated] vs. 3.6 colonies per 50,000 mononuclear cells [treated]; P = 2.3 × 10−9), and CFU-GM (50.0 colonies per 50,000 mononuclear cells [untreated] vs. 7.0 colonies per 50,000 mononuclear cells [treated]; P = 2.5 × 10−11) assays revealed marked inhibition of colony growth. In contrast, no significant inhibition of colony growth was detected after incubation of BM cells obtained from patients with cytogenetically resistant disease; the mean colony counts per 50,000 mononuclear cells were 5.0 [untreated] and 4.6 [treated] in the CFU-Mix assay (P = 0.15); 16.9 [untreated] and 15.7 [treated] in the BFU-E assay (P = 0.11); and 49.9 [untreated] and 48.7 [treated] in the CFU-GM assay (P = 0.15). Table 1 and Figure 8 show that a strict correlation between suppression of colony growth and inhibition of WT1 expression was observed in both responders and patients with resistant disease (r = 0.97). Moreover, as shown in Figure 8, for a given patient, the inhibition of WT1 expression observed after in vitro incubation mirrored the decrease in WT1 expression observed in vivo after therapy.

Figure 7.

Mean number of colonies obtained from control and imatinib-treated cells collected from patients who subsequently were identified as having imatinib-resistant disease and from patients who subsequently had complete or major cytogenetic responses. CTRL: control; MNCs: mononuclear cells; CFU-Mix: multilineage colony-forming units; BFU-E: erythroid bursts; CFU-GM: granulocyte-macrophage–colony-forming units.

Table 1. Number of WT1 Copies per 104GUS Copies and Number of Hematopoietic Colonies in Control Samples and in Samples Incubated with Imatinib That Were Collected from Patients Later Identified as Having Disease That Was Responsive or Resistant to Imatinib Therapy, with Corresponding Data on Percent Inhibition
ResponseWT1 expressionPercent inhibitionNo. of coloniesPercent inhibition
ControlImatinib-treatedControlImatinib-treated
  1. CCR: complete cytogenetic response; MCR: major cytogenetic response; CR: cytogenetic response.

Responders
 CCR924593190.08511.087.1
 CCR423537391.28910.088.8
 CCR295012395.8806.092.5
 CCR149514790.010214.585.8
 CCR375612396.8323.090.7
 CCR4568182.37918.077.3
 CCR10159091.28910.088.8
 CCR254034586.59921.078.8
 CCR215611095.010812.088.9
 CCR358541088.612728.078.0
 CCR43404399.0493.093.8
 CCR489721995.5698.088.4
 CCR1021981.4287.075.0
 CCR11105095.5485.089.6
 CCR3284586.310113.087.1
 CCR143635475.32810.062.5
 MCR201231568.99019.078.9
 MCR290031989.0829.089.0
 MCR197132483.6529.082.7
 MCR9347092.59310.089.2
 MCR140021384.86210.083.9
 MCR18115197.29012.086.7
 MCR410190078.13612.066.7
Nonresponders
 No CR967594622.24844.09.4
 No CR420042160.06062.00.0
 No CR1675138817120112.06.7
 No CR546753462.26062.00.0
 No CR6145457325.58981.09.0
 No CR623365470.05954.08.5
 No CR1498124317.08680.07.0
 No CR611562850.03830.021.1
 No CR547077750.09699.00.0
 No CR454347860.08684.02.4
 No CR9867687530.34648.00.0
 No CR978697680.29198.00.0
 No CR608760120.05450.07.4
Figure 8.

(A) Results of regression analysis of percent WT1 inhibition in vivo after imatinib therapy versus percent WT1 transcript inhibition after in vitro incubation with imatinib in both nonresponders and responders (r = 0.965). (B) Results of regression analysis of percent WT1 transcript inhibition versus percent colony growth inhibition after in vitro incubation with imatinib in bone marrow cells collected from responders and nonresponders before the start of therapy (r = 0.97).

DISCUSSION

On the basis of results obtained in a number of experiments performed on cell lines and on primary CML cells from BM samples, it appears that in most patients with CML, increased WT1 expression is triggered, at least in part, by the deregulation of BCR-ABL TK activity.29 The signal transduction pathway that is activated by BCR-ABL certainly is not the only pathway responsible for increased WT1 expression, as WT1 expression is activated in almost all patients with acute myelogenous leukemia as well as in most patients with ALL, including those without any apparent TK activity; this finding is in agreement with previously reported observations.19

The hypothesis that WT1 is overexpressed in CML as a consequence of the presence of increased BCR-ABL TK activity is supported by our transfection experiments. The parental COS cell line, under normal conditions, expresses very low levels of the WT1 transcript. In contrast, the COS cells that were transfected with BCR/ABL exhibited a marked increase in WT1 expression; however, after incubation with imatinib, inhibition of BCR/ABL TK activity resulted in significant down-regulation of WT1 expression.

Currently, the search for a common checkpoint at which a number of transforming events convene to activate WT1 expression has been unsuccessful. We observed that short-term in vitro incubation (18 hours) of K562 and KCL cells with increasing concentrations of imatinib led to a significant reduction in WT1 expression levels, a phenomenon that was not observed in the BCR-ABL-negative cell lines (HL60 and Me-1) that were used as controls. This effect cannot be ascribed to apoptosis, because the numbers of apoptotic cells did not increase significantly after incubation. In addition, treatment of the K562 cell line with other compounds, such as IFN-α or cytotoxic agents, did not result in WT1 down-regulation (data not shown). These data point to the inhibition of BCR-ABL TK activity as the event that is responsible for WT1 down-regulation.

This same approach can be translated into an in vitro assay that has the potential to predict the sensitivity to imatinib of primary Ph-positive samples obtained from patients with CML. We found that when BM cells obtained from patients who subsequently were identified as cytogenetic responders to treatment were incubated with imatinib, a sudden and marked down-regulation of WT1 transcript levels was observed. In contrast, this phenomenon was not observed in samples that were obtained from patients who subsequently were identified as having imatinib-resistant disease. A possible explanation for this phenomenon is that in patients with resistant disease, due to point mutations or due to BCR-ABL overexpression, imatinib is not able to inhibit BCR-ABL TK activity to a sufficient degree. Alternatively, in agreement with the mechanisms that have been proposed to explain primary resistance, additional genetic lesions may occur in Ph-positive cells from patients with resistant disease, with these lesions making BCR-ABL TK activity redundant and unnecessary for maintaining cells' proliferative and transformed status. These pathways also may be responsible for the maintenance of increased WT1 transcription levels despite the imatinib-induced inhibition of BCR-ABL TK activity. On the basis of these data, our assay of WT1 down-regulation after in vitro exposure to imatinib may become a useful and straightforward tool with which to predict the sensitivity to imatinib of Ph-positive samples obtained from patients at the time of diagnosis and, thus, the sensitivity of these patients' malignancies to treatment with imatinib. The validity of this test is supported by the finding of a strict correlation between the results of the WT1 assay and the results of assays evaluating the inhibition of colony growth formation after in vitro exposure to imatinib. (The latter type of assay has recently been proposed as a method for predicting responsiveness to imatinib.30)

Two major findings emerged from the in vivo data on WT1 expression in patients with CML. First, patients with major or CCRs induced by imatinib had proportional decreases in WT1 transcript levels in their in vitro samples; however, the extent to which this finding is attributable either to direct inhibition of WT1 expression as a result of the inhibition of BCR-ABL TK activity by imatinib or to the dilution of Ph-positive cells by increasingly prevalent Ph-negative cells remains difficult to determine at present. Second, significantly higher pretreatment WT1 transcript levels were observed in patients who subsequently were identified as having imatinib-resistant disease compared with patients who subsequently experienced responses to imatinib. This finding suggests that WT1 levels at diagnosis may represent a prognostic marker for patients with CML, although the value of this marker must be confirmed in a wider and more uniform series of patients.

Acknowledgements

The authors thank Dr. Rocco Piazza (National Cancer Institute, Milan, Italy) for constructing the BCR/ABL plasmid.

Ancillary