spotlight - Nature

Leukemia (2005) 19, 2159–2165 & 2005 Nature Publishing Group All rights reserved 0887-6924/05 $30.00 www.nature.com/leu

Selecting and deselecting imatinib-resistant clones: observations made by longitudinal, quantitative monitoring of mutated BCR-ABL ˚ nonli3, MA Sovershaev2, M Olsen1, T Gedde-Dahl4, H Hjort-Hansen5 and B Skogen1 FXE Gruber1, T Lamark2, A A

Resistance to imatinib during the treatment of chronic myeloid leukaemia (CML) is frequently associated with point mutations in the ABL gene encoding the ATP binding region likely to cause disease relapse. Early diagnosis and monitoring of these mutations may be important in order to prevent rapid expansion of resistant clones. We describe a quantitative mutationspecific PCR assay based on the readily available Taqman platform. Selectivity for the mutated target is conferred by mutation-specific primers destabilised by additional mismatches. The assay can be carried out in parallel to standard BCR-ABL quantification and is therefore more quickly compared to standard sequencing procedures. The sensitivity of the assay reaches 0.1%. It also allows for quantitative assessment of mutated clones. By analysing sequential samples of resistant subjects, we show how mutated clones were selected, maintained or deselected depending on the individual treatment setting. The high sensitivity and practical merits of this method makes it a good candidate for prospective molecular surveillance of patients at high risk for imatinib resistance. Leukemia (2005) 19, 2159–2165. doi:10.1038/sj.leu.2403983; published online 13 October 2005 Keywords: CML; imatinib; mutations; resistance

PCR products have been sequenced. Nevertheless, this requires sequencing of high numbers of clones to reach definitive conclusion.7,11 PCR products can also be analysed for mutations using denaturing high-performance liquid chromatography (D-HPLC).12,13 This method determines the presence of a mutation, but unlike sequencing, does not identify the actual mutation.13 Assays targeting known mutations are more sensitive, but depend either on existence of restriction sites or allele-specific primer annealing.6,14 Based on the amplification refractory mutation system,15 we have developed a reverse transcriptase, quantitative PCR assay for detection of mutated clones (qPCRMUT), a sensitive and practical PCR-based technique for detection and monitoring of selected key mutations associated with imatinib resistance. We propose that targeted screening and quantification of BCR-ABL mutants should be a part of standard molecular surveillance of high-risk CML patients.

Materials and methods

Patients and preparation of samples Introduction Chronic myeloid leukaemia (CML) is a proliferative stem cell disorder characterised by the presence of the Philadelphia (Ph) chromosome, which contains the chimaeric BCR-ABL fusion gene encoding a constitutively active tyrosine kinase.1 Targeted inhibition of BCR-ABL protein by imatinib, a tyrosine kinase inhibitor, induces cytogenetic remission in substantially more patients than previous therapies, such as interferon alpha.2 Thus, imatinib is now the preferred drug for all phases of CML. However, point mutations in the kinase domain of the BCR-ABL gene have been shown to induce resistance to imatinib treatment3–8 and therefore are regarded as adverse prognostic factors.4,6,7 In particular, leukaemic cells with mutations in the region of the ABL gene encoding for the P-Loop are linked to rapid evolution of blast crisis when patients are treated with imatinib.4,7 The standard strategy to identify patients at risk of developing resistance is to monitor blood BCR-ABL transcript levels.9,10 In addition, identification and monitoring of mutated clones could be directly beneficial, as their existence is associated with the development of drug resistance.4 Standard procedures like sequencing of the ABL kinase domain are cumbersome and insensitive.7 To increase sensitivity, subcloned Correspondence: Dr FXE Gruber, Department of Immunology and Transfusion Medicine, University Hospital of Northern Norway, Breivika, Tromsø 9038, Norway; Fax: þ 47 776 26304; E-mail: [email protected] Received 17 June 2005; accepted 14 September 2005; published online 13 October 2005

The regional committee for medical research ethics approved the study and all patients signed an informed consent. BCR-ABLpositive blood samples sent to our laboratory for BCR-ABL mRNA quantification were examined for the existence of imatinib resistance-associated mutations. For the standard mutation analysis, total RNA was isolated from 107 peripheral blood mononuclear cells (PBMC) using the Qiagen RNeasy kit (Qiagen GmbH, Hilden, Germany). RNA (1 mg) was routinely transcribed into cDNA using polyT primers and Superscript II reverse transcriptase following the manufacturer’s instruction (Invitrogen, Carlsbad, USA). The samples were assessed for quality by analysing the ABL copy count. A total of 64 samples from 39 subjects, amplifying ABL before cycle 30 in the real-time PCR assay, were included in the study. For negative controls, RNA isolated from PBMC of healthy blood donors was used.

Design of primers and probes used in the study The assay is based on detection of mutated clones in RNA templates transcribed to cDNA. Mutated clones were analysed by real-time PCR using the Taqman chemistry (Applied Biosystems, Foster City, CA, USA), and the selectivity for detection of mutated clones was conferred by the amplification refractory mutational system (ARMS).15 We denote the assay as qPCRMUT. Clinically important point mutations are clustered around the codons for the P-Loop and the amino acids 315 and 351.7 Gene-specific forward primers and probes complementary to sequences immediately upstream of these three

SPOTLIGHT

1 Department of Immunology and Transfusion Medicine, University Hospital of Northern Norway, University of Tromsø, Tromsø, Norway; 2Biochemistry Department, Institute of Medical Biology, University of Tromsø, Tromsø, Norway; 3Department of Pharmacology, Institute of Pharmacy, University of Tromsø, Tromsø, Norway; 4Section of Hematology, Medical Department, Rikshospitalet University Hospital, Oslo, Norway; and 5Section of Hematology, Medical Department, St Olavs University Hospital, Trondheim, Norway

Quantification of mutated Bcr-Abl clones FXE Gruber et al

2160 Table 1 Mut. cluster P-Loop P-Loop P-Loop P-Loop P-Loop P-Loop P-Loop P-Loop P-Loop 315 351

Primers used for quantitative assessment of mutated BCR-ABL clones Point mut. 1106C4G 1113G4A 1119A4G 1120G4C 1120G4T 1121T4C 1122A4T 1127G4A 1128A4T 1308C4T 1416T4C

Amino acid L248V G250E Q252R Q252Hc Q252Ht Y253H Y253F E255K E255V T315I M351T

Forward primer 50 –30 GGTGTGTCCCCCAACTACGA

TGCAGTCATGAAAGAGATCAAA GAACGCCGTGGTGCTGCT

Mutation-specific primer 50 –30 a

GTACTGGCCCCGGCCCAC ACCTCCCCGTACTAGCCCT TCGTACACCTCCCCATACC CTCGTACACCTCCCTGTAG CTCGTACACCTCCCCGTAA CCTCGTACACCTCCCTGTG CCTCGTACACCTCCCCGA ACACGCCCTCGTACATCTT CACACGCCCTCGTATACCA TCCCGTAGGTCATGAATTCAA TTTCTTCTCCAGGTGCTCCG

Size (bp) 81 89 95 96 96 97 97 103 104 104 67

The following probes were used in combination with the primers denoted above (50 -FAMyTAMRA-30 ): AAGTGGGAGATGGAACGCACGGAC for the P-Loop cluster, ACCCTAACCTAGTGCAGCTCCTT for the 315 cluster and TACATGGCCACTCAGATCTCGTCA for the 351 cluster. a Underlined nucleotides denote mismatches introduced in order to increase selectivity for the amplification of mutated alleles.

SPOTLIGHT

mutational clusters were designed (Table 1). In all, 11 allelespecific reverse primers with a 30 -end complementation to the mutated nucleotide were designed, targeting 11 different point mutations in the ABL gene. Additionally, we destabilised the mutation-specific primers by introducing noncomplementary nucleotides near the 30 -end in order to minimise background signals that originated from misannealing of the mutationspecific primer to the wild-type template (background amplification).

Optimisation of the qPCRMUT assay The different primer/probe combinations were tested by amplifying cDNA samples of patients containing mutated clones found by sequencing. To evaluate the degree of background amplification, cDNA from healthy blood donors was analysed using the same assay. The cDNA quality of these controls was in the same range as the patient samples (data not shown). The primer/probe combination resulting in the greatest difference in CT values (cycle threshold, DCT) between wild-type and mutated genes, as determined by real time PCR, was chosen for the assay (Supplemental Figure 1 illustrates the development of the Y253H assay using samples from a Y253H-positive patient and healthy blood donors). For mutations, where patient samples were not available, we analysed a 10-fold dilution of mutated plasmids in wild-type plasmids (106 copies/ml) and chose the mutation-specific primers with maximal DCT. In our experience, T–G or A–C mismatches in position n-3 to n-5 performed best regarding selectivity and sensitivity for the mutated template.

Determination of the cutoff value For the purpose of cutoff determination, we analysed cDNA samples of five individual blood donors using the mutationspecific assay and calculated a cutoff threshold on the level of CT values. The curve with lowest CT value, as analysed in these five samples minus a margin of one CT, was used as a threshold for the determination of positive samples on the CT level (cutoff CT). Since the degree of amplification of normal cDNA by the mutational assay also depends on the samples’ overall cDNA quality, we also calculated a cutoff ratio by transferring CT values from the analysis of the background amplification and the housekeeping gene of normal cDNA into copy number. Then we calculated a ratio consisting of copy number ‘background amplification’ divided by copy number housekeeping gene, Leukemia

multiplied by 100. In some mutation-specific assays, the background amplification rate was quite modest (amplification curves beyond cycles 40). In these cases, a copy number corresponding to cycles 40 was used in the calculation of the cutoff ratio. Results from the analysis of unknown samples were compared to the established CT cutoff as well as to amplification curves derived from analysis of normal cDNA on each plate. In addition, ratios from analysis of unknown samples were related to the cutoff ratios. In cases where results on the CT and ratio level were close to the established cutoff, new samples of the same patient were analysed after 1–2 months. If the new analysis distinguished the amplification curve/ratio more clearly from the background, confirming a selection of the mutated clone on imatinib treatment, the samples were considered positive.

Absolute quantification of BCR-ABL and ABL copy count BCR-ABL and ABL mRNA transcripts were analysed quantitatively by reverse transcriptase real-time PCR using Taqman probes and primers as described in Gabert et al16 and BCR-ABL fragments containing plasmid dilutions. Results were expressed as a ratio of BCR-ABL copy count divided by ABL copy count in order to normalise for variations in the reverse transcription process.

Absolute quantification of mutated templates A 1293 bp large cDNA fragment containing BCR and ABL sequences was amplified by PCR using primers listed in Supplemental Table 1 and cloned into a TOPO TA vector (Invitrogen). In all, 11 different mutants were produced by sitedirected mutagenesis following the manufacturer’s instructions (Stratagene, La Jolla, CA, USA). Copy numbers of linearized plasmids were calculated after measurement of the OD at 260 nm. Dilutions (10-fold) containing 106 to 101 copies/ml were analysed using the qPCRMUT assay. The resulting CT values were plotted against log plasmid copy number. Slope and y-intercept values from these standard curves were used for the calculation of absolute copy number as follows: copy number (mutated clone)/40 ng RNA ¼ (CT unknown sampley-interceptstandard curve/ slopestandard curve). Analogous to the BCR-ABL/ABL ratio, we normalised the result for variations in the reverse transcription step by dividing copy number of mutated BCR-ABL by copy number of ABL.


qPCRMUT conditions

Evaluation of inter- and intra-assay variability

A measure of 2 ml of patient cDNA equivalent to 40 ng of RNA were analysed using the mutation-specific primers. The qPCR assay was run on an Abi Prism Sequence detection system 7900 using standard Taqman PCR conditions (501C–2 min, 951C–10 min, 45 cycles at 951C–15 s and 601C–1 min) and the Universal Taqman PCR master mix (Applied Biosystems, Foster City, CA, USA).

A cDNA sample of a patient known to contain a G250E mutated clone coexisting with the wild-type BCR-ABL was analysed in five replicates on three occasions for testing of the reproducibility of the mutation-specific real-time assay. The intra-assay variability of the G250E detection on the CT level was 0.64% and the interassay variability was 1.27%. The ABL housekeeping gene analysis was run on the same plates; calculating variation based on ratios resulted in 14% intra-assay and 25% inter-assay variation.

The BCR-ABL kinase domain was isolated by PCR using BCR-ABL-specific primers (Supplemental Table 1 for primer sequences). In a subsequent PCR step, an aliquot (1:1000 diluted template from the first PCR reaction) was subjected to amplification of the cDNA sequence encoding the kinase domain, as described by Hochhaus et al.6 The resulting template including the Abl kinase domain was sequenced in both directions using Big Dye chemistry (Applied Biosystems).

Results

Characterisation of the mutation-specific PCR assays An example of setting up a qPCRMUT analysis is presented in Supplemental Figure 1 illustrating how the real-time platform can be used for rapid evaluation of ARMS PCR assays. Primers amplifying almost exclusively the targeted mutated clone could be designed for all of the 11 mutations. The amount of background amplification was in most of the mutations restricted beyond cycles 40. A higher background amplification and lower sensitivity, probably due to the extraordinary CG richness in this region, was accepted for some P-Loop mutations. However, using the ARMS PCR rules, mutationspecific assays of reasonable sensitivity could be designed for the P-Loop mutations so as to selectively amplify mutated BCRABL over a 3 log range (sensitivity 0.1%, Table 2). As illustrated in Supplemental Figure 2 for the analysis of the Y253H-mutated clone, analysis of 10-fold diluted linearised plasmid standards were used to characterise the individual assays. A linear amplification was observed over six orders of magnitude from 106 to 101 copies in all assays. The CT value corresponding to amplification of one copy (Y-intercept) was in all cases within a reproducible range of real-time PCR analysis (Table 2). All assays were close to 100% effective, as evaluated by slope values derived from the analysis of plasmid standards (Table 2).

Table 2

Evaluation of sensitivity The sensitivity of the assays for the 11 different mutations was determined. First, we calculated a ratio consisting of copy number background amplification and copy number housekeeping gene as described above for each of the 11 mutations. This threshold defines the lowest copy number of RNA in mutated cells normalised for cDNA quality that can be distinguished from background amplification (cutoff ratio). We also serially diluted mutated plasmids in a buffer containing 106 wild-type copies/ml in a second experiment testing the ability of the system to discriminate between mutated and wild-type BCRABL. The qPCRMUT approach was able to discriminate 103 to 104 mutated plasmids/ml diluted in a buffer containing 106 wildtype plasmids/ml (sensitivity 1–0.1%, Table 2 and Supplemental Figure 3). In a third experiment, we diluted patient samples containing a mutated clone 10-fold in wild-type BCR-ABL cDNA from a patient never treated with imatinib (1:1 until 1:1000). The dilutions were analysed both by sequencing and qPCRMUT allowing for a direct comparison of these methods. Three different samples containing mutations 1113G4A (G250E), 1121T4C (Y253H) and 1416T4C (M351T) were tested. By sequencing, the mutated clones could be detected in dilutions 1:10, 1:5 and 1:2, respectively, resulting in a sensitivity of 10% (G250E), 20% (Y253H) and 50% (M351T). Yet, using the qPCR method the mutations could be detected at 1:1000 dilution for all three variants (Table 3), confirming a reproducible sensitivity of 0.1–1% for this approach in patient samples.

Clinical validation of the assay For the purpose of clinical validation, the described qPCRMUT assay was used for analysis of CML and Ph þ ALL patients, who showed characteristics of suboptimal imatinib treatment. At the same time, the ATP binding region was sequenced in all

SPOTLIGHT

Sequence analysis of the BCR-ABL kinase domain

2161

Efficiency, cutoff determination and sensitivity of qPCRMUT for detection of 11 different mutations associated with imatinib resistance

Mutation

Slope

Y-intercept

E

Cutoff CT

Cutoff ratio

Sensitivity (plasmid dilution) (%)

L248V G250E Q252R Q252Hc Q252Ht Y253H Y253F E255K E255V T315I M351T

3.39 3.53 3.44 3.38 3.38 3.37 3.47 3.61 3.47 3.54 3.97

39.32 38.05 38.44 34.80 34.80 34.15 35.99 33.47 37.33 33.55 35.06

1.97 1.91 1.95 1.98 1.98 1.98 1.94 1.89 1.94 1.92 1.79

39.00 33.98 39.00 39.00 37.65 33.45 33.79 31.87 36.48 39.00 36.71

0.07 0.04 0.04 0.02 0.004 0.06 0.2 0.1 0.08 0.001 0.006

0.1 1 1 0.1 0.1 0.1 0.1 0.1 1 1 0.1

Leukemia


CP ¼ chronic phase; AP ¼ accelerated phase; BC ¼ blast crisis. a Position according to GenBank no. M14752.

48 15 1 12 35 17 120 (m) (m) (f) (m) (f) (f) (m) 77 42 35 24 55 75 44

4 17 62 41 (f) (f) (m) (f) 60 52 71 70

CP CP CP 1. Relapse 1. BC 1. AP CP CML CML CML ALL CML CML CML #5 #6 #7 #8 #9 #10 #11

20 36 6 1.5 8 15 12

CP HR 1. BC 2. Relapse 2. BC 2. AP CP

Alive Alive Dead Dead Dead Dead Alive

T1121C T1121C G1113A T1416C T1121C T1416C T1416C G1113A G1120C T1121C C1106G G1120C Dead Dead Alive Dead Relapse Bc AP BC 2. 2. 2. 1. ALL CML CML CML #1 #2 #3 #4

1. Relapse 1. BC 1. AP CP

Disease Patient

In five subjects, we were able to analyse sequential samples. Mutated clones discovered in the screening analysis were followed by sequencing and qPCRMUT in samples taken before, during and after imatinib resistance developed. A total of 30 samples were analysed. In eight of these 30 samples mutations were found by qPCRMUT, but not by sequencing (samples marked by ! in Figure 1a–e). In these samples, the sensitivity of the single-step, targeted procedure was superior to the nested PCR-based sequencing approach. In one subject of a Ph þ ALL patient, we could even detect a Y253H-mutated clone before the onset of imatinib treatment (Figure 1a). Figure 1 illustrates the kinetic development of six individual mutated BCR-ABL clones (G250E 1,Y253H 3 and M351T 2) vs total BCRABL during the disease course of one ALL (Figure 1a) and four CML patients (Figure 1b–e). The observation period varied between the cases from 7 (case A) to 50 months (case C). Table 4 gives an overview on relevant clinical data and disease features. Quantitative analysis in cases A, B and E demonstrates a rapid selection of the mutated clones upon imatinib treatment followed by a parallel increase in total tumour burden as shown by an increase in the BCR-ABL ratio. This observation points out that these clones are the actual cause for disease relapse. The M351T-mutated clone presented in Figure 1e increases in a similar way when exposed to imatinib 400-mg daily. However, after dose increase to 600 mg, ratio values for the mutated clone and total BCR-ABL declined slightly, indicating that the M351T clone is sensitive to increased doses of imatinib. Cases A and B progressed into a terminal disease phase at the end of the observation period, whereas case E still was in haematological remission at that time (Table 4). The corresponding sequencing data confirm that the mutated clone takes over total BCR-ABL during the selection by imatinib, but this technique is generally less sensitive than the qPCRMUT and not quantitative. In Figure

Summary of clinical features for all patients with imatinib resistance-associated mutations

Sequential monitoring of mutated BCR-ABL clones

2.5 15 15 17

Point mutation detecteda Clinical status at last follow-up Clinical status at the time of mutation detection

samples. The cohort included imatinib-treated, Ph þ patients, who had signed informed consent and met one of the following requirements: (i) suboptimal response to treatment in terms of BCR-ABL reduction of o3 log after treatment duration of minimum 6 months or (ii) any increase in BCR-ABL ratio or (iii) haematologic relapse. In total, 39 subjects were screened. As listed in Table 4, 12 individual mutated clones were discovered in 11 subjects by the targeted qPCRMUT assay. In one individual, the assay detected two mutated clones (M351T and Y253H), whereas sequencing only detected the M351T mutation. The finding of two independent clones was confirmed by cloning of the Bcr-Abl kinase domain and sequencing of individual clones (data not shown). Sequencing did not uncover further mutations within the ATP binding region, pointing out that the test panel contained the clinically most important mutations.

Duration of imatinib treatment (months)

0.1 0.1 0.1

Clinical status at onset of imatinib treatment (months)

10 20 50

Table 4

SPOTLIGHT Leukemia

qPCRMUT Patient sample diluted in BCR-ABL cDNA (%)

Time from diagnosis to start imatinib therapy (months)

G250E Y253H M351T

Sequencing (%)

Age at diagnosis (years) (sex)

Amino-acid substitution

Amino-acid substitution

Table 3 Sensitivity of mutation-specific ARMS analysis vs sequencing as evaluated by dilution of patient samples

Y253H Y253H G250E M351T Y253H M351T M351T G250E Q252Hc Y253H L248V Q252Hc

2162


Figure 1 (a–e) Longitudinal, quantitative monitoring of total BCR-ABL and mutated BCR-ABL clones in five subjects presenting features of imatinib resistance (case A: ALL (denoted as case #1 in Table 4), cases (b–e): CML (cases #2–#5 in Table 4)). Filled circles denote MRD (ratio BCRABL/ABL (%)), and open circles denote the corresponding ratios for the mutated clones. Sequencing data are expressed as: for not detected and þ for detected. The ! symbol shows samples where mutations were detected by the qPCRMUT, but not by sequencing. Treatment modalities are described by horizontal arrows in the lower part of the figures. The stippled line below denotes the detection limit for mutational analysis (qPCRMUT).

1c and d, we present quantitative data on the development of a G250E, Y253H and M351T clone after cessation of imatinib treatment. Both patients received busulphane after resistance to imatinib was diagnosed. In case D, one sample was further analysed for BCR-ABL and the mutated clones before the patient died in a complication as a result of sepsis, whereas case C was repeatedly monitored over a period of 50 months. Measurements of total BCR-ABL and sequencing of the kinase domain in case D demonstrates conflicting data, which were difficult to understand. At the time of resistance (first sampling), we detected two individual mutated BCR-ABL clones in addition to wild-type BCR-ABL, a dominating M351T and small amounts of Y253H. At 6 months after withdrawal of imatinib and during

SPOTLIGHT

2163

treatment with busulphane, the P-Loop clone had taken over the majority of BCR-ABL-positive cells, whereas the M351T was almost 2 log factors reduced and no longer detectable by sequencing. Branford et al4 recently showed that P-Loopmutated clones preferentially caused relapse in the first year of imatinib treatment, whereas other mutations, for example, M351T, relapsed in the second and third year of treatment. Thus, in our case, the M351T mutation may have existed at the onset of imatinib treatment. The delayed selection of the Y253H mutation suggests that this clone occurred during imatinib treatment. These data point out that busulphane has an impact on the mutated clones. In particular, the overall proportion of the Y253H clone in relation to tumour burden increased upon Leukemia


2164 busulphane treatment. A similar observation was made in case C where the detected G250E-mutated clone fluctuated over 2 log factors upon repetitive withdrawal and reinitiation of busulphane therapy, whereas total tumour burden was rather stable in that period. We interpret the data illustrated in Figure 1c and d as a shift between unmutated and mutated BCR-ABL cells dependent on the treatment setting. It seems that busulphane selects for the G250E, whereas treatment-free periods deselect the mutated clone in favour of unmutated BCR-ABL.

Discussion

SPOTLIGHT Leukemia

This paper describes the qPCRMUT method, which is a new analytical approach for sensitive detection and quantitative monitoring of BCR-ABL mutations associated with resistance to imatinib. The assay targets the most important mutations representing more than 70% of all described patient cases.7 Moreover, the main structure of the assay easily allows for the inclusion of other targets. The results demonstrate that the qPCRMUT real-time assay in this study is 50- to 100-fold more sensitive than the sequencing strategy. Extensive analysis of both diluted patient samples as well as plasmid mixtures containing variable concentrations of mutated plasmids confirmed a reproducible sensitivity of 0.1–1%. Moreover, we demonstrate by monitoring sequential patient samples that the assay is capable of detecting mutated clones earlier than by sequencing. Assays of similar sensitivity for detection of imatinib resistanceassociated mutations have been described before.6,14,17,18 Using the ARMS principle, however, mutation-specific assays can be designed for every available template independent of restriction sites. In addition, the assay consists of one PCR step reducing the risk for contamination, and it works on the widely used Taqman platform in parallel with BCR-ABL quantification. However, more than 30 different mutations have been described in association with imatinib resistance. A targeted screening for all these mutations could be difficult to perform. Mutational analysis, as described here, should therefore be aimed at the sensitive detection of key residues associated with imatinib resistance and quantitative monitoring of known mutated clones. Future directions as combination therapy and use of new generation tyrosine kinase inhibitors19 may reduce the number of mutations causing resistance and then targeted detection will certainly replace techniques like sequencing or D-HPLC. Mutations in the ATP binding region have been clearly associated with resistance to imatinib.3–5,20 However, recent reports point out that mutated clones can be present without being selected upon imatinib treatment.21 Taking advantage of our quantitative approach for mutational monitoring, we show that the presented mutations actually induce an increase in tumour burden, which terminally leads to disease relapse. In one case of an M351T mutation, dose escalation led to a decline in tumour burden (Figure 1e). This observation is consistent with preclinical data, which demonstrate that an M351T-mutated clone could be overcome by dose increase.22 Since quantitative monitoring of BCR-ABL mutations has not been commonly used, there are little data on the development of mutated clones after cessation of imatinib. Discontinuation of imatinib was reported to result in a spontaneous reversion from BC to CP in two case descriptions.22–24 Mutational analysis was carried out in one of these patients, and demonstrated a significant reduction of a Y253H mutation after withdrawal of imatinib. The mechanism behind the observed drop of mutated BCR-ABL upon imatinib interruption might be based on a suppression of the mutated clone by the reapparent wild type.25

These cases are of importance for the clinical management of PLoop-mutated and imatinib-resistant patients.4 The interaction between wild-type and mutated clone could be exploited as a therapeutic option. In this respect, our quantitative measurements of a G250E clone dropping almost 2 log factors during the period of withdrawal at two occasions are highly consistent with the above presented mechanism of a direct interaction between mutated and unmutated BCR-ABL clones. Surprisingly, we find that busulphane treatment maintains the imatinib-selected G250E clone on a high level after cessation of imatinib. And, furthermore, after a treatment pause, the reinitiated busulphane drug is even able to select for the G250E clone (Figure 1c). In light of this observation, we want to stress that a recently published study26 presented clinical data on patients with imatinib-resistant, P-Loop-mutated BCR-ABL. Six of the eight patients die, but most of them progress after a latency of 8–12 months after withdrawal of imatinib. The majority of the patients were, during this period, according to the authors, treated with hydroxyurea. When considering our observation of increased proliferate capacity of P-Loop clones during treatment with busulphane, it seems hydroxyurea might have a similar impact on the mutated clones. In summary, we present data showing that qPCRMUT might produce significant information that can both be useful for as well in monitoring as in treatment guidance of CML and ALL patients under treatment with imatinib. In particular, the quantitative aspect of the assay seems to be highly relevant for monitoring patients exhibiting treatment resistance. Our data imply that mutant clones are the actual cause for disease relapse, and that resistance might be overcome by dose increase for M351T or treatment discontinuation for P-Loop mutations. According to our observations also nontargeted treatment like busulphane might lead to a proliferation of P-Loop-mutated BCR-ABL clones. If confirmed in larger sample groups, data based on quantitative monitoring of mutated clones could be used to optimise the treatment of imatinib-resistant patients. Further studies on the clinical relevance of sensitive detection and quantification of mutated clones are underway as the method has been implemented as part of several therapy programs for CML.

Acknowledgements This work was supported by scientific grants from the Northern Norwegian Health Authorities and the Norwegian Cancer Association. We thank Geir Bjørkøy, Terje Johansen and Ole Peter Rekvig for critical discussion of the project, Per Arne Standal, Malvin Sjo, Alf K Haaland, Aleksandra Silje and Nina Gulbrandsen for sending patient samples and Labforum for sequencing the samples. Special thanks to Kimmo Porkka for discussing molecular features of CML and Rasmus Goll for help with statistics.

Supplementary Information Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu).

References 1 Sawyers CL. Chronic myeloid leukemia. N Engl J Med 1999; 340: 1330–1340. 2 O’Brien SG, Guilhot F, Larson RA, Gathmann I, Baccarani M, Cervantes F et al. Imatinib compared with interferon and low-dose


3

4

5 6 7

8

9

10

11

12

13

14

15

16

17

18 19

20 21

22

23

24 25 26

to STI571, and they can pre-exist to the onset of treatment. Blood 2002; 100: 1014–1018. Newton CR, Graham A, Heptinstall LE, Powell SJ, Summers C, Kalsheker N et al. Analysis of any point mutation in DNA. The amplification refractory mutation system (ARMS). Nucleic Acids Res 1989; 17: 2503–2516. Gabert J, Beillard E, van dVV, Bi W, Grimwade D, Pallisgaard N et al. Standardization and quality control studies of ‘real-time’ quantitative reverse transcriptase polymerase chain reaction of fusion gene transcripts for residual disease detection in leukemia – a Europe Against Cancer program. Leukemia 2003; 17: 2318–2357. Kreuzer KA, le Coutre P, Landt O, Na IK, Schwarz M, Schultheis K et al. Preexistence and evolution of imatinib mesylate-resistant clones in chronic myelogenous leukemia detected by a PNA-based PCR clamping technique. Ann Hematol 2003; 82: 284–289. Liu WH, Makrigiorgos GM. Sensitive and quantitative detection of mutations associated with clinical resistance to STI-571. Leukemia Res 2003; 27: 979–982. Burgess MR, Skaggs BJ, Shah NP, Lee FY, Sawyers CL. Comparative analysis of two clinically active BCR-ABL kinase inhibitors reveals the role of conformation-specific binding in resistance. Proc Natl Acad Sci USA 2005; 102: 3395–3400. Shah NP, Sawyers CL. Mechanisms of resistance to STI571 in Philadelphia chromosome-associated leukemias. Oncogene 2003; 22: 7389–7395. Willis S, Lange T, Demehri S, Otto S, Crossman L, Niederwieser D et al. High sensitivity detection of BCR-ABL kinase domain mutations in imatinib-naive patients: correlation with clonal cytogenetic evolution but not response to therapy. Blood 2005; 106: 2128–2137. Corbin AS, La Rosee P, Stoffregen EP, Druker BJ, Deininger MW. Several Bcr-Abl kinase domain mutants associated with imatinib mesylate resistance remain sensitive to imatinib. Blood 2003; 101: 4611–4614. Liu NS, O’Brien S. Spontaneous reversion from blast to chronic phase after withdrawal of imatinib mesylate in a patient with chronic myelogenous leukemia. Leukemia Lymphoma 2002; 43: 2413–2415. Muller MC, Lahaye T, Hochhaus A. Resistenz auf tumorspezifische Therapie mit Imatinib duch klonale Selektion mutierter Zellen. Dtsch Med Wochenschr 2002; 127: 2205–2207. Hochhaus A, La Rosee P. Imatinib therapy in chronic myelogenous leukemia: strategies to avoid and overcome resistance. Leukemia 2004; 18: 1321–1331. Soverini S, Martinelli G, Rosti G, Bassi S, Amabile M, Poerio A et al. ABL mutations in late chronic phase chronic myeloid leukemia patients with up-front cytogenetic resistance to imatinib are associated with a greater likelihood of progression to blast crisis and shorter survival: a study by the GIMEMA Working Party on chronic myeloid leukemia. J Clin Oncol 2005; 23: 4100–4109.

SPOTLIGHT

2165 cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med 2003; 348: 994–1004. Branford S, Rudzki Z, Walsh S, Grigg A, Arthur C, Taylor K et al. High frequency of point mutations clustered within the adenosine triphosphate-binding region of BCR/ABL in patients with chronic myeloid leukemia or Ph-positive acute lymphoblastic leukemia who develop imatinib (STI571) resistance. Blood 2002; 99: 3472–3475. Branford S, Rudzki Z, Walsh S, Parkinson I, Grigg A, Szer J et al. Detection of BCR-ABL mutations in patients with CML treated with imatinib is virtually always accompanied by clinical resistance, and mutations in the ATP phosphate-binding loop (P-Loop) are associated with a poor prognosis. Blood 2003; 102: 276–283. Gorre ME, Mohammed M, Ellwood K, Hsu N, Paquette R, Rao PN et al. Clinical resistance to STI-571 cancer therapy caused by BCRABL gene mutation or amplification. Science 2001; 293: 876–880. Hochhaus A, Kreil S, Corbin AS, La Rosee P, Muller MC, Lahaye T et al. Molecular and chromosomal mechanisms of resistance to imatinib (STI571) therapy. Leukemia 2002; 16: 2190–2196. Shah NP, Nicoll JM, Nagar B, Gorre ME, Paquette RL, Kuriyan J et al. Multiple BCR-ABL kinase domain mutations confer polyclonal resistance to the tyrosine kinase inhibitor imatinib (STI571) in chronic phase and blast crisis chronic myeloid leukemia. Cancer Cell 2002; 2: 117–125. von Bubnoff N, Schneller F, Peschel C, Duyster J. BCR-ABL gene mutations in relation to clinical resistance of Philadelphiachromosome-positive leukaemia to STI571: a prospective study. Lancet 2002; 359: 487–491. Branford S, Rudzki Z, Parkinson I, Grigg A, Taylor K, Seymour JF et al. Real-time quantitative PCR analysis can be used as a primary screen to identify patients with CML treated with imatinib who have BCR-ABL kinase domain mutations. Blood 2004; 104: 2926–2932. Hughes TP, Kaeda J, Branford S, Rudzki Z, Hochhaus A, Hensley ML et al. Frequency of major molecular responses to imatinib or interferon alfa plus cytarabine in newly diagnosed chronic myeloid leukemia. N Engl J Med 2003; 349: 1423–1432. Hofmann WK, Komor M, Wassmann B, Jones LC, Gschaidmeier H, Hoelzer D et al. Presence of the BCR-ABL mutation Glu255Lys prior to STI571 (imatinib) treatment in patients with Ph+ acute lymphoblastic leukemia. Blood 2003; 102: 659–661. Deininger MW, McGreevey L, Willis S, Bainbridge TM, Druker BJ, Heinrich MC. Detection of ABL kinase domain mutations with denaturing high-performance liquid chromatography. Leukemia 2004; 18: 864–871. Soverini S, Martinelli G, Amabile M, Poerio A, Bianchini M, Rosti G et al. Denaturing-HPLC-based assay for detection of ABL mutations in chronic myeloid leukemia patients resistant to Imatinib. Clin Chem 2004; 50: 1205–1213. Roche-Lestienne C, Soenen-Cornu V, Grardel-Duflos N, Lai JL, Philippe N, Facon T et al. Several types of mutations of the Abl gene can be found in chronic myeloid leukemia patients resistant

Leukemia