Molecular studies in chronic myeloid leukemia patients treated with tyrosine kinase inhibitors

Molecular Studies in Chronic Myeloid Leukemia Patients Treated With Tyrosine Kinase Inhibitors Chdes L. Sawyen The tyrosine kinase inhibitor imatinib mesylate (Gleevec TM, Novartis Pharmaceuticals Corp, East Hanover, NJ) (formerly STl571) blocks the constitutively activated Bcr-Abl tyrosine kinase that is characteristic of chronic myeloid leukemia (CML). Molecular analysis for the presence of residual Bcr-Abl-positive cells in patients with a cytogenetic response following treatment with imatinib mesylate reveals that some patients have undetectable disease using quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) assays capable of detecting 1 in 10" Philadelphia chromosome-positive (Ph+) cells. To examine whether the leukemia is still Bcr-Abl-dependent in patients who have responded to imatinib mesylate but have relapsed, a quantitative assay that directly measures enzymatic activity of Bcr-Abl toward one of its major signaling substrates has been developed. This assay allows monitoring both of the imatinib mesylate sensitivity of patient cells in vitro, and of the endogenous inhibition of Bcr-Abl kinase activity during imatinib mesylate treatment and relapse. Studies show that imatinib mesylate resistance is associated with restored activation of the Bcr-Abl signal transduction pathway in the majority of cases, indicating that Bcr-Abl remains a valid target for therapeutic intervention. Understanding resistance mechanisms of Phf leukemia to imatinib mesylate will allow design of therapies to overcome such barriers to efficacy. Semin Hematol38(suppl8):1521. Copyright 0 2001 by W.B. Saunders Company,

T

HE CAUSATIVE AGENT of chronic myeloid leukemia (CML) is the Philadelphia (Ph) chromosome, a reciprocal genetic recombination that creates the activated tyrosine kinase fusion gene and protein, Bcr-Ab1.l Activity of the normal cellular proto-oncogene c-a61 resides in the kinase domain that is kept under tight regulatory control via well-coordinated signals from cytokines and other environmental stimuli. Loss of this control in the Ph chromosome-the first exon of the dcr locus replaces the normal 5’ regulatory abl sequencesresults in a deregulated and constitutively activated kinase domain. The phenotype of chronic phase CML is the ultimate result-proliferation, decreased adherence to bone marrow stroma, and reduced apoptosis of Bcr-Abl-positive cells.i-3 Nearly all patients with CML express the Bcr-Abl protein, rendering it an ideal target for drug therapy that is designed to eliminate the leukemic cell or revert its phenotype to normal. Phase I clinical trials of the Abl kinase inhibitor imatinib mesylate (GleevecTM, Novartis Pharmaceuticals Corp, East Hanover, NJ) (formerly ST1571) have validated the concept of targeted signal transduction inhibition. Nearly all patients in chronic phase CML in these trials obtained a hematologic response (HR) and the Seminars

in Hematology,

majority attained a cytogenetic response (CR) that appears to be durable at 1 year.* As CML progresses from chronic phase to blast crisis, most affected cells acquire secondary genetic and molecular abnormalities.* Imatinib mesylate treatment induced HR and CR in patients with CML in blast crisis as well, even in the presence of secondary genetic abnormalities. However, despite initial response in advanced phase CML, relapse occurs frequently.> Thus, understanding the mechanisms of response and resistance to imatinib mesylate is necessary in order to overcome barriers to efficacy. This presentation summarizes recent studies at the molecular level, which provide insight into imatinib mesylate response and resistance in patients with CML.

From the Department of Medicine, Division of Hematology and Oncology, Molecular Biology Institute, University of California Los Angeles School of Medicine, Los Angeles, CA. Address f”eprint requests to Charles Sawyers, MD, UCLA School of Medicine, Factor Building, Room 11-394, 10833 LeConte Ave, Los Angeles, CA 90095. Copyright 0 2001 by W.B. Saunders Company 0037-1963/01/3803-8003$35.00/0 doi:l0.1053/shem.2001.27078

Vol 38, No 3, Suppl8

(July),

2001:

pp 15-21

15

16

Chdrles L. Sawyers

Molecular Studies to Monitor Efficacy of Imatinib Mesylate Therapy There are three means to monitor drug efficacy in CML. HR, historically the oldest criterion, is defined as normalization of the white blood cell (WBC) and platelet counts and bone marrow morphology. CR (which can be major or minor) is defined as absence of the Ph chromosome in blood and bone marrow. A major CR comprises both complete CR (no Phf chromosomes by conventional cytogenetic techniques) and partial CR (1% to 35% Ph+ chromosomes). Recently developed techniques can monitor the remission or progression of disease at the molecular level. Fluorescence in situ hybridization (FISH) is a “molecular cytogenetic” technique that detects the bw-abl fusion gene in the DNA of interphase or hypermetaphase nuclei. Reverse-transcriptase polymerase chain reaction (RT-PCR), often qualified aseither quantitative and/or real-time PCR, detects bcr-abl mRNA transcripts in lysates of CML cells. At the protein level, a CRKL substrate phosphorylation assaymeasuresfunctional Bcr-Abl tyrosine kinase activity. The clinical value of molecular testing is currently under scrutiny and remains to be fully defined.l Results of the imatinib mesylate phase I clinical trial illustrate the utility of FISH and RT-PCR analyses in chronic-phase CML (Table 1).*s6After treatment with imatinib mesylate, a significant percentage of patients with CML in chronic phase become cytogenetically negative, defined as O/20 Ph+ metaphases. Of these evaluable Ph- patients treated at our center, 86% were also negative by FISH, defined as O/250 interphase nuclei bcr-dbl-positive. Thus, FISH correlates well with traditional cytogenetic analysis and is somewhat more sensitive, detecting residual bcr-abl in some Ph- patients. Upon further analysis, 50% of evaluable Ph‘ Table 1. PCR-Negative lmatinib

Remission Mesylate

in Patients at UCLA

Treated

With

No. of Patients

Cytogenetically FISH-negative PCR-negative

negative (O/20 (O/250 nuclei) (l/lo5 cells)

metaphases)

29 (86%) 9/18 (50%)

24/28

patients were also bcr-ubl-negative by quantitative RT-PCR. It should be noted that since the sensitivity of this assayis l/lo5 cells, PCR negativity implies a profound level of remission. Whether the increased sensitivity of PCR over cytogenetic analysis gives it superior clinical value remains to be seen. PCR provides a single “snapshot” in time of quantitative tumor burden and can be used to monitor bcr-abl transcripts over a course of therapy, analogous to human immunodeficiency virus (HIV) detection. Other studies also demonstrate the value of molecular monitoring to detect minimal residual diseaseduring therapy. A large percentage of patients enrolled in the imatinib mesylate multicenter phase II clinical trial achieved a complete CR. However, asshown by Kantarjian and colleagues, all patients showed evidence of minimal residual diseaseby real-time quantitative RT-PCR.’ Other phase II analyses of patients using quantitative RT-PCR detected only extremely low bcr-abl transcript levels in patients in complete CR and correctly identified those with a partial or minor response. These studies all suggest that long-term follow-up is needed to determine whether imatinib mesylate induces molecular remission as well as complete CR. As the sensitivity of PCR can vary using different methods, it will be important to compare results among centers using uniform technologies.

Drug Resistance Issues During Imatinib Mesylate Therapy in CML The problem of drug resistance, defined as absenceof a significant HR or CR, or failure to maintain a response, is illustrated by results of the imatinib mesylate phase I clinical trial in advanced-phase CML. 5 Nearly six in 10 patients in myeloid blast crisis had a response (defined as <15% blasts in bone marrow) and one third had a complete response(<5% blasts in marrow). Responserates in patients in lymphoid blast crisis were even higher. However, despite initially more promising responses,all patients in lymphoid blast crisis eventually relapsed. Durable responses were seen only rarely among patients in myeloid blast crisis who had an initial complete response to ima-

17

Molecuhr Studies of Kinase Inhibition in CML

tinib mesylate. Relapses in patients with advanced CML occurred within 2 to 4 months in lymphoid blast crisis and slightly slower in myeloid blast crisis. One third of patients in myeloid blast crisis survived at 1 year, but the durability of the response in myeloid blast crisis is not yet known. Several hypotheses can be put forth to explain failure to achieve a cytogenetic response in chronic-phase patients. The effectiveness of imatinib mesylate may vary according to the differentiation stage of the target cell (Bcr-Abl may be more readily inhibited in myeloid progenitors and neutrophils compared with less mature cells). Relatively resistant stem cells would have a self-renewal advantage and come to dominate hematopoiesis over time. The percentage of Bcr-Abl-positive stem cells within a particular patient might greatly influence drug efficacy, and this variable could explain why some patients, those presumably carrying a larger tumor burden of Ph+ stem cells, do not respond initially. It is hoped that given enough time and optimal drug dosage, these patients may eventually convert to Ph negativity. Hence, it may be important to be able to identify, with molecular and antibody markers, and phenotype the Bcr-Abl status of the primitive stem cell population. This information might allow therapy to be customized so that patients with a higher percentage of BcrAbl-positive stem cells would be candidates for more intensive imatinib mesylate therapy (Fig 1). Other hypotheses can explain why patients with CML in advanced stage develop resistance. Secondary genetic changes arising within the Ph+ clone could provide oncogenic signals that replace Bcr-Abl as the sole determinant of transformation. Bcr-Abl would become irrelevant and relapse would be due to Bcr-Ablindependent cell growth. Alternatively, BcrAbl might be reactivated or overcome a previously effective dose of imatinib mesylate. Possible mechanisms of reactivation include gene amplification of bcr-abl or overexpression of the multidrug resistance gene (MDR-1). Overexpression of MDR would lead to increased levels of P-glycoprotein (Pgp), a cell membrane efflux pump that lowers net intra-

Stem Cells i Bone Marrow Progenitors 1 Peripheral Blood Neutrophils

0

Ph-positive Ph-negative

.

Duration of chronic phase

Figure 1. As hematopoiesis proceeds from immature stem cells to mature neutrophils in CML, the number of Ph+ cells expands tremendously. During progression of chronic phase, Ph+ cells also represent a larger fraction of total stem cells. A method to quantify Phf cells in the stem cell pool would provide insight about which patients might respond more quickly to treatment.

cellular drug concentration. Finally, the drug could be functionally inactivated in vivo through enzyme modification or biotransformation (for example, by the liver cytochrome P4>o system), or functionally sequestered by the acute phase reactive protein a,-acid glycoprotein (a,-AGP), which binds to imatinib mesylate in the circulation and thereby lowers its effective tissue concentration8

Preclinical Studies Addressing Resistance Mechanisms Consistent with the observation that resistance is generally not a problem in chronic-phase CML, preclinical studies showed that imatinib mesylate significantly inhibits the survival of chronic phase CML-derived myeloid progenitors.9 Inhibition was specific with little effect on normal marrow-derived progenitors and Ph-, non-CML hematopoietic cell lines. Conversely, several blast crisis-derived Bcr-Abl-positive CML cell lines were refractory to imatinib mesylate. From these cell lines a series of sensitive and resistant sublines were generated by in vitro culture over 4 to 5 months in low doses of the drug. l”,ll For these studies, resistance was defined as the ability to proliferate/ survive indefinitely in the continuous presence of 1 pmol/L imatinib mesylate. Two of four resistant CML cell lines upregulated Bcr-Abl at the mRNA (1.6 to 5.5-fold by Northern blots) and protein levels (6- to 12-

18

Charles L. Sawym

fold by Western blots probed with anti-Abl or anti-Bcr). Bcr-Abl upregulation was quite dramatic in murine cell lines (up to 34-fold); however, expression in some of these cell line models is driven by a transgene and may not apply to regulation of the endogenous fusion gene in CML. Importantly, resistance was always relative-cell lines refractory at a given imatinib mesylate concentration were still sensitive at higher doses. In some cell lines upregulation of Bcr-Abl protein was attributed to gene amplification that was readily demonstrated by FISH. In resistant sublines, hybridization signals were multiplied 3- to 6-fold over the baseline signals present in parental, drugsensitive cells already possessing multiple copies of bcr-ab1. Furthermore, in one resistant line both MDR and bcr-ubl were upregulated, indicating that multiple mechanisms can operate simultaneously within a cell. By contrast, in other CML cell lines resistance and increased Bcr-Abl expression were not associated with gene amplification.10~12 Additional experimental phenomena with these cell lines are worth noting. A small number of resistant CML cell lines were able to regain imatinib mesylate sensitivity after drug withdrawal for 2 to 3 months, concurrent with reduced Bcr-Abl protein expression.9 The mechanism(s) underlying altered protein levels was not identified but the ability to revert to drug sensitivity may have clinical implications for targeted therapy. Other investigators have examined so-called “host-mediated” resistance in an experimental animal model8 In these studies, immunocompromised nude mice inoculated with CML cells developed elevated plasma levels of AGP. In vitro culture of CML blasts with AGP lowers intracellular imatinib mesylate levels up to lofold, an effect reversed by coculture with erythromycin, which competes with imatinib mesylate for ar,-AGP binding. As a result, in vivo resistance to imatinib mesylate that develops in CML tumor-bearing mice was overcome by coadministration of erythromycin-plasma levels of ctlAGP-bound imatinib mesylate decreased and levels in critical target organs (tumor, kidney, liver, and spleen) increased significantly. Mutations in the Bcr-Abl kinase domain

under selective pressure by imatinib mesylate comprise another class of potential resistance mechanisms. Particular amino acid residues have been implicated, but Bcr-Abl point mutations were ruled out by sequencing the entire kinase domain of several resistant CML cell lines.rO Compensating mutations in other genes also have been postulated to explain resistance. Possibilities include genes involved in target protein or drug metabolism or degradation, tyrosine dephosphorylation, and pro- or antiapoptotic pathways. Whether resistance mechanisms hypothesized or demonstrated in preclinical studies also operate in imatinib mesylate-treated patients with CML will be the subject of intense research as the clinical development of this drug proceeds.

Molecular Studies to Monitor Relapse During Treatment A critical question is whether relapsing leukemia is still Bcr-Abl-mediated-that is, theoretically still imatinib mesylate-inhibitablesince continued therapy will not likely succeed if secondary oncogenic changes have rendered cells Bcr-Abl independent. Directly measuring Bcr-Abl activity in patient samples, primarily via Bcr-Abl autophosphorylation or total cellular tyrosine phosphorylation, is difficult due to limiting cell numbers. In order to monitor imatinib mesylate resistance we have developed a surrogate immunoblot assay for Bcr-Abl activity to help determine if in vivo resistance mirrors in vitro mechanisms. This immunoblot assay measures the phosphorylation status of a physiologic Bcr-Abl substrate, CRKL, in patient blood or bone marrow. Specifically, it detects the band shift of phosphorylated CRKL that migrates above unmodified CRKL on a sodium dodecyl sulfate-polyacrylamide gel electropheresis (SDS-PAGE) gel. We believe this assay offers several experimental advantages: (1) it is sensitive, since CRKL is highly phosphorylated by Bcr-Abl on at least two tyrosine residues; (2) it is objective, quantitative, and reproducible-the intensity ratio of the two migrating forms of CRKL is measured and the readout is percentage CRKL phosphorylation; (3) it is accurate, since normal

Molecular

Studies of Kinase

individuals and patients with Ph. CML show less than 20% phosphorylation; and (4) finally, it is a valid surrogate for Bcr-Abl activity since reduced phosphorylated CRKL levels parallel reduced levels of phosphorylated Bcr-Abl and c-Abl (Fig 2). In patient cells treated ex vivo with increasing concentrations of imatinib mesylate there is a dose-dependent reduction in phosphorylation. Of greater clinical relevance, endogenous (in vivo) inhibition of Bcr-Abl activity can also be measured. Prior to treatment, CRKL phosphorylation levels range from 60% to 80%, and in responding patients these levels fall to 30% to 50% after 5 to 26 days of imatinib mesylate treatment. But after relapse, phosphorylated CRKL levels approach or exceed pretreatment levels, with 18 of 18 advanced-stage relapsed CML patients showing reactivation of functional Bcr-Abl (Fig 3). Using this assay we have been able to establish that resistance is cell-autonomous in the leukemic clone. When patient cells were removed from the leukemic milieu during relapse, they retained altered ex vivo biochemical sensitivity to imatinib mesylate. A dose-response and 50% inhibitory concentration (IC,,) for imatinib mesylate was determined by plotting percentage of CRKL phosphorylation over a range of drug concentrations in vitro. A shift in the dose-response to imatinib mesylate (up to 0

@clcu + CRKL’

0.1

0.3

1.0

5.0

10.0

PM imatinib

mesylate

dXKL

Figure 2. Measuring Bcr-Abl signal transduction using substrates. CRKL is a natural substrate of BcrAbl and when phosphorylated (P-CRKL) appears as a protein of slightly higher molecular weight than unmodified CRKL on an SDSPAGE gel. In the absence of imatinib mesylate, CRKL is phosphorylated while in the presence of 10 pmol/L imatinib mesylate CRKL is nearly entirely in its unmodified form. CRKL can be phosphorylated by Bcr-Abl on tyrosines 198 and 207. Adapted with permission.13 Copyright 2001 American Association for the Advancement of Science.

Inhibition

19

in CML

days on imatinib

pre-TX

mesylate

TX

Figure 3. Reactivation of Bcr-Abl signaling in relapsed patient cells. At the beginning of treatment (transplantation, day 0) the level of CRKL phosphorylation ranges from 60% to 80%. After 5 days of treatment, the level falls to between 40% to 60%, but by day 22 the patient has relapsed and the level of CRKL phosphorylation return to pretreatment (pretransplantation) levels. Adapted with permission.13 Copyright 2001 American Association for the Advancement of Science.

lo-fold higher IC,,) is seen in patients with “relative” resistance (Fig 4). These results are consistent with studies by other investigators on functional reactivation of Bcr-Abl. Gambacorti-Passerini et al and Mahon et al measured endogenous Bcr-Abl phosphorylation and demonstrated reactivation of kinase activity in vivo in patients with accelerated or blast crisis phase CML.14,r5 The relapsing leukemic cells remained partially sensitive to imatinib mesylate in vitro, when assayed by Bcr-Abl phosphorylation and inhibition of proliferation. Reactivation of Bcr-Abl activity attributable to gene amplification has been documented in three of 11 patients examined.r6 At treatment onset, patients had a single Ph chromosome by cytogenetics and a single bcr-abl copy by FISH. Several months into imatinib mesylate therapy copy number was increased, up to 20-fold, by FISH. Gene amplification was also manifested as a cytogenetically observable marker. Three relapsed patients on imatinib mesylate (two with advanced-phase CML, one with Ph+ acute lymphocytic leukemia) were observed to acquire a duplicated, inverted Ph chromosome.

20

Patient 1

Patient 2

100 d 8

100 d

80

imatinib

mesylate

(PM)

8 v f

80

1 $ 0 f s

40

80

20 0 imatinib

mesylate

(PM)

Figure 4. Altered ex vivo biochemical sensitivity to imatinib mesylate in relapsed patients. Patients’ cells are exposed to increasing concentration of imatinib mesylate ex vivo. Prior to imatinib mesylate treatment, CRKL phosphorylation is reduced at drug concentrations 2 0.1 pmol/L. In cells from patients who have relapsed, the dose of imatinib mesylate necessary to reduce CRKL phosphorylation is much higher (1 to 10 pmol/L. Adapted with permission. I3 Copyright 2001 American Association for the Advancement of Science.

This marker chromosome was unstable, varying in copy number in different cells, and forming multiple ring chromosomes. The increasing fraction of cells bearing the marker chromosome suggested selective pressure by imatinib mesylate on this clone.”

Conclusion: Molecular

Early Lessons From Analyses in CML

The overall conclusion from molecular studies to date is that the mechanism of imatinib mesylate relapse is generally reactivation of Bcr-Abl signaling. The persistence of functional Bcr-Abl tyrosine kinase activity suggests that relapse is unlikely to be explained by secondary oncogenic changes. Therefore, the other genetic and molecular events that are observed concurrently in advanced phase CML cells may indeed be “secondary” to the primary translocation event and not sufficient to support leukemic cell growth on their own. Targeted inhibition of Bcr-Abl by imatinib mesylate presumably exerts selective pressure on leukemic cells, such that clones with sufficient genomic instability can overcome the drug. Gene amplification is one such mechanism. Taken together, the results of preclinical studies offer optimism with respect to Bcr-Abltargeted therapy. While often multifactorial, resistance is relative, not absolute. Therefore, it may be worthwhile to treat patients with CML

with higher doses of imatinib mesylate or concomitantly with other targeted inhibitors.

Note Added in Proof A recent publication has established that point mutations in the kinase domain of BcrAbl can occur in patients with blast crisis CML who relapse after an initial response to ST1571.13

References 1. Sawyers CL: Chronic myeloid leukemia. N Engl J Med 340:1330-1340, 1999 2. Fader1 S, Talpaz M, Estrov Z, et al: The biology of chronic myeloid leukemia. N Engl J Med 341: 164172, 1999 3. Sawyers CL: Signal transduction pathways involved in BCR-A3L transformation. Baillieres Clin Haemato1 10:223-231, 1997 4. Druker BJ, Talpaz M, Rester DJ, et al: Efficacy and safety of a specific inhibitor of the Bcr-Abl tyrosine kinase in chronic myeloid leukemia. N Engl J Med 344:1031-1037, 2001 5. Druker BJ, Sawyers CL, Kantarjian H, et al: Activity of a specific inhibitor of the Bcr-Abl tyrosine kinase in the blast crisis of chronic myeloid leukemia and acute lymphoblastic leukemia with the Philadelphia Chromosome. N Engl J Med 344:1038-1042, 2001 6. Shah NP, Snyder DS, Nicoll JM, et al: PCR-negative molecular remissions in chronic-, accelerated-, and blast crisis-phase CML patients treated with ST1571, an Abl-specific kinase inhibitor. Blood 96:471a, 2000 (suppl 1, abstr)

21

Moleczdar Studies of Kinase Inhibition in CML

7. Quackenbush RC, Talpaz M, Guo J, et al: Monitoring minimal residual disease by Bcr-Abl Western and quantitative competitive RT-PCR in Philadelphia chromosome (Ph)-positive chronic myelogenous leukemia (CMR) on ST1571 therapy. Blood 96:736a, 2000 (suppl 1, abstr) 8. Gambacorti-Passerini C, Barni R, le Coutre P, et al: Role of (pi acid glycoprotein in the in vivo resistance of human Bcr-Abl+ leukemic cells to the Abl inhibitor STI571. J Nat1 Cancer Inst 92:1641-1650, 2000 9. Deininger MW, Goldman JM, Lydon N, et al: The tyrosine kinase inhibitor CGP57148B selectively inhibits the growth of BCR-ABL-positive cells. Blood 90:3691-3698, 1997 10. lMahon FX, Deininger MW, Schultheis B, et al: Selection and characterization of BCR-ABL positive cell lines with differential sensitivity to the tyrosine kinase inhibitor ST1571: Diverse mechanisms of resistance. Blood 96:1070-1079, 2000 11. le Coutre P, Tassi E, Varella-Garcia M, et al: Induction of resistance to the Abelson inhibitor ST1571 in human leukemic cells through gene amplification. Blood 95:1758-1766, 2000 12. Weisberg E, Griffin JD: Mechanism of resistance to the ABL tyrosine kinase inhibitor ST1571 in BCRi

ABL-transformed

hematopoietic

cell

lines.

Blood

95:3498-3505,2000 13.

Gorre ME, Mohammed M, Ellwood K, et al: Clinical resistance to STI-571 cancer therapy caused by BCRABL gene mutation or amplification. Science 1062538, published online 21 June 2001 14. Gambacorti-Passerini C, Verga M, Rossi F, et al: ST1571 administered to chronic myeloid leukemia (CML) patients inhibits cell proliferation, induces

15.

16.

17.

apoptosis and initially causes partial inhibition of BcriAbl autophosphorylation. Blood 96:345a, 2000 (suppl 1, abstr) Mahon F, Belloc F, Chalet C, et al: Detection of resistance to ST1571 in patients with Bcr-Abl positive acute leukemia and chronic myeloid leukemia (CML) blast crisis. Blood 96:47Ia, 2000 (suppl 1, abstr) Gorre ME, Banks K, Hsu NC, et al: Relapse in Ph+ leukemia patients treated with an Abl-specific kinase inhibitor is associated with reactivation of Bcr-Abl. Blood 96:470a, 2000 (suppl 1, abstr) Mohammed M, Shin S, Deng S, et al: BCRIABL gene amplification: A possible mechanism of drug resistance in patients inhibitor. Blood

treated 96:344a,

with an Abl-specific kinase 2000 (suppl 1, abstr)