Chronic myeloid leukemia (CML) is characterized by a reciprocal chromosomal translocation t (9; 22) (q34; q11) resulting in the BCR-ABL oncogenic fusion gene, which encodes the cytoplasmic BCR-ABL protein. The BCR-ABL protein has constitutive tyrosine kinase activity that mediates cellular transformation and leukemogenic effects [1].

Imatinib Mesylate (IM) the selective inhibitor of the BCR-ABL tyrosine kinase, competitively targets the adenosine 5’-triphosphate (ATP) binding site of the AKD and blocks downstream signal transduction pathways leading to growth arrest and apoptosis [2]. Because of its excellent safety and important therapeutic benefit for patients with CML, IM has become the standard of care for the treatment of CML. However, despite high rates of hematologic, cytogenetic and molecular responses, refractoriness or acquired resistance after initial response to IM is observed in a significant proportion of patients [3].

Point mutations within the AKD are emerging as the most frequent mechanism for reactivation of BCR-ABL kinase activity [4]. These mutations affect amino acids involved in IM binding or in regulatory regions of the AKD and result in decreased sensitivity to the drug [5]. It is postulated that mutations within the AKD can prevent IM from binding by changing the conformation of the BCR-ABL protein [6].

Molecular monitoring of BCR-ABL transcript levels for CML patients treated with IM has proved effective in defining patient response and provides important guide for clinical management and has prognostic relevance, in addition that it can indicate timing of mutation testing. Knowing the AKD mutational status is clinically useful, since it may prompt a change in therapy [7].

Biochemical and cellular assays have demonstrated that different AKD mutations result in varying levels of resistance [8,9]. Clinical studies have shown that the site of mutation within the kinase domain may affect disease outcome. The different mutations may therefore require different strategies to overcome resistance such as dose escalation for those that confer moderate resistance, or shifting to second generation TKIs or even transplantation for more resistant mutations [10].

The aim of the current study was to perform mutation analysis in a series of IM treated CML Egyptian patients of different response category in light of serial BCR-ABL measurements by RQ-PCR. We report on their type, frequency, relation to IM response and impact on clinical outcome of the disease.

Study group

Between March 2005 and April 2011, 175 newly diagnosed adult patients with Philadelphia-positive (Ph⁺) CPCML were enrolled in this study. Seventy two patients; 44 males and 28 females were selected for ABL kinase domain mutational screening representing all resistant patients (28/175, 16%) in this study with >2 fold rising BCR-ABL level in addition to all suboptimal responders (18/175, 10%) and a fraction of patients representing optimal responders (n:30) with stable/decreasing transcript level. The study was conducted in accordance with the Declaration of Helsinki and was approved by the National Cancer Institute (NCI) Institutional Review Board. Median age of patients at diagnosis was 42 years (range 18-72 years). Median Follow up period was 41.6 month (7.5-49.5).

Patient eligibility

Inclusion criteria were (1) morphologic and cytogenetic evidence of Ph⁺ CML in early chronic phase (defined as less than 12 months from diagnosis); (2) age 18 years or older; (3) normal renal and hepatic functions (4) normal cardiac function. Women at childbearing age were required to have a negative pregnancy test before starting IM, and to use contraception during therapy. Exclusion criteria included previous treatment for CML (Busulfan, IFN-α, or Ara-C). Exceptions included hydroxyurea and anagrelide for the treatment of elevated WBC (>50×10⁹/L) and platelet count (>700×10⁹/L), respectively; usage was limited to 4 weeks before starting IM therapy.

Treatment

Patients received 400 mg IM orally once a day. The dose was reduced for any ≥ Grade 3 drug-related hematologic toxicity. No dose adjustments were made for Grade 1 or 2 hematologic toxicity. No dose reductions below 300 mg/day were allowed. Any toxicity had to be resolved within 28 days.

Quantitation of BCR-ABL mRNA transcript levels by RQPCR

Patients were monitored by RQ-PCR for BCR-ABL transcript levels at diagnosis and at 3-month intervals during IM therapy. The ABL control gene and the BCR-ABL target gene transcript levels were quantitated using TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, California, USA), the BCR-ABL Mbcr FusionQuant Kit (Ipsogen, Luminy Biotech Entreprises, Marseille, France) and the Smart Cycler detection system (Cepheid, USA). A set of reference RNA was provided from Ipsogen to convert results into international scale (IS).

Definition of treatment responses

Hematologic, molecular and cytogenetic responses were defined according to European Leukemia Net criteria. MMR was defined as a 3-log reduction of BCR-ABL transcripts level, corresponding to ≤ 0.1% on international scale (IS) [11]. PFS was defined as loss of hematologic or cytogenetic response, death, or development of advanced CML; OS defined by the absence of death from any reason [12].

Abl Kinases Domain (AKD) mutations screening

Mutation screening was performed using allele specific oligonucleotide - polymerase chain reaction (ASO-PCR). Genomic DNA was extracted from peripheral blood using Gentra Puregene blood kit (QIAGEN, Hilden, Germany). ASO-PCR assay was established for the detection of 16 known mutations which were selected according to their frequency in IM-resistant CML patients. Mutation panel selected included M244V, L248V, G250E, Q252H (a, b) showing 2 different nucleotides substitution, Y253H, Y253F, E255K, E255V, V299L, F311L, T315I, F317L, M343T, M351T, E355G and F359V mutations. Mutated or wild-type sequences were specifically amplified in a PCR reaction to analyze the most frequently identified mutations in the AKD (amino acids 220 to 498). The strong specificity of the assay was demonstrated for each mutation by detection of wild ABL sequence in DNA from clinically-declared healthy controls. Sequence of forward and reverse primers for both wild and mutant types used for ASO-PCR was adapted from previous reports [13,14]. The amplified products were detected by electrophoresis on 2% ethidium bromide-stained agarose gel.

Statistical analysis

The chi-square test was used to determine the significance between variables. Kaplan-Meier survival curves were used to assess outcome of patients with and without mutations [15]. Log rank test was used to compare between groups. PFS was calculated from start of treatment until the first reported appearance of AP or ABC. OS was calculated from start of treatment until death. All analyses were performed using the statistical package for the social sciences (SPSS software 17; SPSS Inc., Chicago, USA) [16].

RQ-PCR results and mutational status

Twenty eight resistant patients (18 with 1ry resistance and 10 with 2ry resistance) and all 18 suboptimal responders in addition to a group of 30 patients with optimal response to IM randomly selected (15 in CMolR and 15 in MMolR) were submitted for mutation screening. Sixteen resistant patients (16/28, 57%) showed a mutation event while none of the optimal (0/30, 0%) or suboptimal responders (0/18, 0%) with stable or decreasing transcript level showed any detectable mutation (p<0.001). Baseline characteristics, mutational status and disease outcome of the 28 IM-resistant patients are listed in table 1.

Frequency of AKD mutations in IM-resistant CML patients

Nine different AKD mutations were found in 16/28 (57%) patients. Mutations of the P-loop (amino acids 248-255) were detected in 9/16 (56%) patients while non-P loop mutations were detected in 7/16 (44%) patients. Q252H was detected in 3/16 patients (19%), Y253H in 2/16 patients (12%), Y253F in 2/17 (12%) and E255K in 2/17 (12%) patients. From seven non–P-loop mutations, 1/16 patient (6%) showed V299L, 4/17 patients (25%) showed M351T, and 2/16 patients (12%) showed F359V (Figure 1). One patient had both Y253H and E255K mutations (more than one mutant clone per patient). T315I was detected in 1/16 (6%) patients only.

Mutation site, frequency, and disease phase

Mutations were detected in 6/17 (35%) patients in CP and in 3/3 (100%) patients in AP and in 7/8 (87%) patients in ABC at time of rise of BCR-ABL level (Table 2). Frequency of mutations was higher in advanced phases (AP and ABC) than in CP-CML (Table 3, p=0.01). Among 6 mutations found in CP-CML patients; 5 mutations (83%) were non–P-loop and 1 mutation (17%) was located in the P-loop region. Two/3 mutations (67%) of AP were in the P-loop region. Among 7 detected mutations in ABC patients, 5 mutations (71%) were located in the P-loop region, 1 mutation (14%) non–P-loop and 1 mutation (14%) was located in gatekeeper region T315I.

Impact of mutational status on OS and PFS

Median OS of the 56 patients without detectable mutations was 42.6 months (13.5- 49.5) versus of 37.5 months (7.5-45) in the mutation positive group. The estimated OS and PFS at 49 months in the mutation positive group were 56.3% and 37.5%, respectively versus 98.2% in non-mutation group (p<0.001) (Figures 3a and 3b). Median PFS of the mutation positive group was 13.5 months (3-45) versus 42.6 months (13.5- 49.5) in the non mutation group (p<0.001). Seven/9 (78%) patients with P-loop mutation (Q252Ha, Q252Hb, Y253F, Y253H, E255K) and one patient with T315I mutation died as a result of disease progression to either AP or ABC. Median OS of P-loop/ T315I group was 14 months (7.5-39) versus 42 months (40-45) in patients with non P-loop mutations (Table 1; Figure 2). No disease related death occurred in patients with non P-loop mutations. The estimated rates of OS (Figure 3a) and PFS (Figure 3b) at 48 months for patients with P-loop/T315I mutations were 22.2% and 11% versus 100% and 71% in patients with non-P loop mutations, respectively (P = 0.003, p=0.017) (Figures 3c and 3d). Estimated OS of resistant patients with mutations detected in CP was 100% versus 30% for patients with mutations detected in AP/ABC (p=0.001) (Figure 3e).

Measurement of BCR-ABL transcript levels by RQ-PCR proved to be the gold standard of identifying patients at risk of IM failure [17,18]. More recently studies focused on other predictive factors of response to IM therapy as the 10% BCR-ABL^IS transcript level at 3 months and the pretreatment transcript level [19-21]. However, point mutations within the AKD have been reported as the most common mechanism for IM resistance or progression in CML patients. Identification of molecular basis of IM resistance is valuable as it can provide prognostic information and contribute to determining appropriate therapy to prevent or overcome resistance [3]. It was suggested that a 2-fold increase in transcript level can be an indication of AKD mutation testing [18]. However, it is still difficult to define the extent of transcript increase that justifies mutation screening [22-24].

The choice of therapy should be guided by multiple factors including mutational analysis, disease phase, patients’ characteristics and safety profile [25]. The ELN recommendations consider any rise in transcripts level a warning element requiring more stringent and careful monitoring [11] whereas fluctuations in PCR results in MMolR patients may be due to sampling effect. Since the ELN included confirmed loss of MMolR among events defining a suboptimal response to IM, in patients showing an increase in BCR-ABL transcript level, a confirmed loss of MMolR should trigger a mutation analysis as this could be a quite reproducible predictor of loss of CCyR [26].

In the present study among 28 refractory patients showing >2 fold BCR-ABL transcripts rise, 16 (57.1%) showed detectable mutations whereas no mutation was detected in optimal responders with stable or decreasing transcript level which add proof to Soverini et al. [26]. Strikingly, mutations were also undetected in suboptimal responders which raise the issue of other mechanisms of resistance as low OCT1 activity resulting in an inadequate intracellular concentration of IM [27]. All refractory patients did not achieve MMolR which explains the high mutation incidence at the time of rise of BCR-ABL. Mutation rate in this cohort was 16/28 (57%) which is relatively higher than that observed by others [28,29]. One study reported AKD mutations at a lower frequency (2 of 44 patients with relapsed or refractory disease) [30]. Differences in the frequency of mutation detection may be attributed to several factors; differences in the sensitivities of the techniques, in the time point of analysis, in phase of the disease and eventually due to differences in the genetic make-up of patient populations [31]. The frequency of BCR-ABL mutations in the present study was attributed to advanced phase of the disease at which mutations were detected [28]. Ten out of 16 patients with detectable mutation were in ABC (n: 7) and in AP (n:3). Seven/10 patients with detectable mutations in advanced phase died. Mutations were also detected in 6 CP patients, suggesting that this event is not restricted to patients in AP or ABC. However, none of them died due to shift to 2^nd generation TKI, as early intervention that prevented progression of the disease. However, frequency of mutations was higher in advanced phases (AP and ABC) (10/11, 90.9%) compared to CP (6/17, 35.3%) indicating a cause – effect relation between disease progression and development of mutation.

OS and PFS were significantly decreased in patients harboring AKD mutations, a finding that has been stressed by others [18]. More importantly was the impact of the site of mutation on prognosis since different mutations have been associated with different degrees of resistance to IM as 7/8 patients with P loop and T315I mutations experienced disease progression. The increased kinase activity and transformation potency of several ATP binding loop mutations explain their association with poor prognosis leading to enhanced transforming capacity and associated genomic instability [30,32,33].

P-loop mutations conferred a true resistant phenotype and were not amenable to IM dose increase and implied alternative therapeutic strategies as second generation TKIs [34]. Two patients with P-loop mutations (Q252H and Y253H) were shifted to nilotinib and achieved CCyR in 1 year. T315I was detected in only one patient in the studied cohort who had no HLA identical donor for allogeneic stem cell transplantation and died in ABC. This rate was somewhat lower than that reported by Jabbour et al. [28] but higher than reported in two European studies [30].

Within our series there was also one patient harboring two known drug-resistant mutations, conferring increased oncogenic potency. On the other hand, patients with mutations that confer partial IM resistance as F359V regained clinical response with higher IM doses if the mutation was the sole abnormality [8,24]. In the present study, 2 patients with F359V mutation regained MMolR by increasing IM dose to 600 mg. The 4 patients harboring M351T mutation were shifted to dasatinib and 3 of them achieved MMolR within 12 months while one patient achieved CCyR. In this issue a stress on the choice of intervention with respect to the site and type of mutation should be considered [10].

In conclusion, more than 2 fold rise in BCR-ABL transcripts ratio in resistant CML patients who demonstrate molecular refractoriness to IM triggers mutation analysis as an early predictor of progression. At this time, an early shift of therapy to appropriate 2^nd generation TKI is beneficial when a positive mutation is detected in order to prevent disease acceleration. However, screening MMR patients with stable or decreasing BCR-ABL level for mutations is not warranted. ASO PCR is a specific and sensitive technique for mutation screening. M351T seems to be the commonest mutation in this CML Egyptian cohort. P-loop mutations are strongly associated with poor prognosis and mainly detected in advanced phases.