Management of imatinib-resistant patients with chronic myeloid leukemia

Abstract

Since its approval in 2001 for frontline management of chronic myelogenous leukemia (CML), imatinib has proven to be very effective in achieving high remission rates and improving prognosis. However, up to 33% of patients will not achieve optimal response. This has led researchers to develop new second- and third-generation tyrosine kinase inhibitors. In this article, we review the mechanisms of resistance, recommendations for monitoring, assessment of milestones, and management options for patients with CML who are resistant to imatinib therapy. We further explain the potential pitfalls that can lead to unnecessary discontinuation, the prognosis of patients whose condition fails to respond to treatment, and the upcoming therapies.

Introduction

Imatinib mesylate (Gleevec, Novartis Pharma, New Jersey, USA) is a first-generation tyrosine kinase inhibitor (TKI) that was approved for frontline therapy in patients with chronic myeloid leukemia (CML) by the US Food and Drug Administration (FDA) in 2002. It is dosed at 400 mg daily in patients with CML in the chronic phase (CML-CP) and 600 mg daily in patients with CML in the accelerated phase (CML-AP) and for those in the blast phase (CML-BP). Imatinib mesylate is a BCR-ABL-targeted therapy and considered the standard of care in CML management. It has shown to produce superior results in terms of response rates, prognosis, and side-effect profile compared with the previously accepted standard, combination therapy with interferon and cytarabine.

The pathogenesis of CML involves a characteristic genetic abnormality: the fusion of the Abelson murine leukemia (ABL) gene in chromosome 9 with the breakpoint cluster region (BCR) gene in chromosome 22. This balanced translocation between chromosomes 9 and 22 [t(9;22)(q34;q11)], termed the Philadelphia (Ph) chromosome, is the hallmark of CML. The fused BCR-ABL gene encodes the 210 kDa BCR-ABL fusion oncoprotein, which contains the activated tyrosine kinase region of ABL. This p210BCR-ABL is a deregulated, constitutively active tyrosine kinase that promotes growth and replication by causing aberrancies in proliferation, apoptosis resistance and adhesion through interference with β-integrin signaling and with multiple downstream pathways, such as Janus kinase signal transducer and activator of transcription (Jak-STAT), Phosphatidylinositol 3-kinases – protein kinase B (PI3K/Akt), JUN kinase, MYC, Wnt-β-catenin and Ras-Raf-MAPK signaling routes.

Imatinib works through competitive inhibition at the adenosine triphosphate (ATP) binding site of the BCR-ABL protein, which results in the inhibition of phosphorylation of proteins involved in BCR-ABL signal transduction. It shows specificity for BCR-ABL but also the receptor for platelet-derived growth factor (PDGF), and c-kit tyrosine kinases [Druker and Lydon, 2000]. The BCR-ABL inhibition results in apoptosis of the hematopoietic cells that express BCR-ABL without affecting the normal cells [Deininger et al. 1997; Druker and Lydon, 2000].

Despite the positive results obtained in previous studies, approximately 33% of patients with CML treated with imatinib do not achieve a complete cytogenetic response (CCyR), while others have drug resistance or cannot tolerate drug-related toxicities [Bixby and Talpaz, 2009; Hochhaus et al. 2009b; Kantarjian et al 2011.]

Causes for imatinib treatment failure

Mechanisms of resistance

There are two categories of resistance: primary resistance is the failure to achieve any of the landmark responses established by the European LeukemiaNet (ELN) or National Comprehensive Cancer Network (NCCN) guidelines. Primary resistance can be further divided into primary hematologic resistance, which occurs in 2–4% of cases who fail to normalize peripheral counts within 3–6 months of initiation of treatment; or primary cytogenetic resistance, which is more common, and occurs in approximately 15–25% of patients who fail to achieve any level of cytogenetic response (CyR) at 6 months, a major CyR (MCyR) at 12 months or a CCyR at 18 months [Shah, 2007]. Secondary resistance occurs in those who have previously achieved and subsequently lost their response in accordance with those guidelines. The mechanisms of resistance to imatinib can be either BCR-ABL dependent (gene amplification or point mutations) or BCR-ABL independent.

BCR-ABL-dependent mutations

Point mutations in the BCR-ABL kinase domain (KD) can lead to imatinib resistance, particularly secondary resistance, and are responsible for treatment failure in many cases [Bixby and Talpaz, 2009; Branford et al. 2003; Hochhaus et al. 2002; Jabbour et al. 2006; Lee et al. 2008; O’Hare et al. 2005]. Numerous mutations have been characterized throughout the ABL sequence, including the ATP phosphate-binding loop (P-loop), and substrate-binding site mutations. Point mutations can change the conformation of the BCR-ABL oncoprotein to the active form, causing a conformational change in the imatinib binding site, prohibiting imatinib binding [Lee et al. 2008], and can also eliminate critical molecules required for bonding, thus mitigating its efficacy [Bixby and Talpaz, 2009; Deininger et al. 2005; O’Hare et al. 2005; Shah, 2005].

Over 100 different point mutations have been identified so far and important examples include T315I, Y253H and F255K, among others. T315I and certain mutations occurring in the P-loop are the most frequently identified mutations [Ravandi, 2011]. The T315I mutation (also known as the gatekeeper mutation) is seen in 4–15% of patients with imatinib resistance, and results when a single C to T nucleotide substitution occurs at position 944 of the ABL gene, resulting in a threonine to isoleucine substitution at amino acid 315 (Th315 to Ile315). This eliminates a critical oxygen molecule needed for hydrogen bonding between imatinib and the ABL kinase. It confers resistance not only to imatinib but also to nilotinib (Tasigna, Novartis Pharma) and dasatinib (Sprycel, Bristol-Myers Squibb Pharma, Princeton, NJ, USA) and has been shown to impact long-term outcome [Jabbour et al. 2006; Nicolini et al. 2006]. The clinical significance of other mutations, such as the P-loop mutations, is still controversial [Jabbour et al. 2006; Khorashad et al. 2008].

Amplification of the ABL kinase oncogene has been observed in several studies [Bixby and Talpaz, 2009; Gorre et al. 2001; Le Coutre et al. 2000; Mahon et al. 2000; Weisberg and Griffin, 2000]. However, in the majority of patients, it was not shown to be a primary mode of treatment failure. Reactivation of BCR-ABL signal transduction is another mechanism of resistance and has been associated with both BCR-ABL point mutations and gene amplification [Gorre et al. 2001].

To validate the ELN guidelines regarding when to perform mutational analysis, Soverini and colleagues, from GIMEMA (Gruppo Italiano Malattie EMatologiche dell’Adulto), performed mutational analyses on their database of 1301 patients. These data were presented at the American Society of Hematology (ASH) 2011 annual meeting [Soverini et al. 2011]. Imatinib-resistant mutations were detected in zero and two imatinib-naïve patients with CML-CP and CML-BP at initial diagnosis respectively. Among those treated with imatinib who had suboptimal response and those with imatinib failure, only 11/233 (4.7%) and 45/166 (27.1%) were found to be positive for one or more BCR-ABL KD mutations respectively. Among those who achieved CCyR but were molecular suboptimal responders, 0/52 with less than major molecular response (MMR) and 4/95(4.2%) who lost MMR were found positive for KD mutations. Interestingly, newly acquired mutations were detected in 93/131 (71%) patients who lost a previously achieved hematologic response or CyR. The authors concluded that patients with KD mutations are more likely to have suboptimal CyR than suboptimal molecular responders and any BCR-ABL transcript increase that is not associated with MMR loss should not trigger a mutation analysis. This later statement actually supported the report by Kantarjian and colleagues [Kantarjian et al. 2009c].

BCR-ABL-independent mechanisms

BCR-ABL-independent mechanisms of resistance to imatinib include increased efflux of the drug by increased expression of P-glycoprotein efflux pumps [Bixby and Talpaz, 2009; Che et al. 2002; Jabbour et al. 2011b; Kotaki et al. 2003; Rumpold et al. 2005], overexpression of the P-170 glycoprotein that enhances drug efflux and reduces intracellular drug accumulation [Chu and DeVita, 2010], decreased drug uptake secondary to decreased expression of the drug uptake transporter human organic cation transporter 1 (hOCT1) [Bixby and Talpaz, 2009; Hughes et al. 2006; Thomas et al. 2004; Wang et al. 2008; White et al. 2006], sequestration of imatinib by increased serum protein α1 acid glycoprotein, which binds imatinib and impairs subsequent binding to ABL kinase [Gambacorti-Passerini et al. 2000; Jabbour et al. 2011b; Widmer et al. 2006], low serum drug concentration, and alternative signaling pathway activation through Ras/Raf/MEK kinase, STAT, Erk2, or SFK phosphorylation of BCR-ABL [Bixby and Talpaz, 2009]. Elevated transcript levels of prostaglandin-endoperoxide synthase 1/cyclooxygenase 1, which encodes an enzyme that metabolizes imatinib, has also been associated with primary resistance [Zhang et al. 2009].

In a recent study published by Ng and associates a novel mechanism was identified among East Asians that explains the primary resistance and suboptimal response to TKIs compared with other ethnic groups [Ng et al. 2012]. The TKI resistance observed was secondary to a germline deletion polymorphism of the BH3 domain of BIM (BCL2L11) that is needed for TKIs to induce apoptosis in malignancies driven by kinases. This allows the researchers to develop BH3 mimetics that can be used conjunction with TKIs to overcome the resistance and thus to personalize the therapy.

Patient compliance

Rates of imatinib adherence have been estimated to range from 75% to 90%, and lower adherence rates correlate with worse outcome [Darkow et al. 2007; Marin et al. 2010; Noens et al. 2009]. In a study of 87 patients with CML-CP treated with imatinib 400 mg daily, adherence rates of 90% or less resulted in MMR rates of 28.4% compared with 94.5% in patients with greater than 90% adherence rates (p < 0.001) [Marin et al. 2010]. Complete molecular response (CMR) rates were 0% versus 43.8% respectively (p = 0.002), and no molecular responses were observed when adherence rates were 80% or lower. Lower adherence rates have been described in younger patients, those who experience adverse events (AEs) related to therapy, and those who require dose escalations.

In conclusion, patient compliance should be evaluated in patients with TKI resistance. Mutation analysis should be considered in someone who has a suboptimal response or has had a loss of cytogenetic or molecular response at any time [O’Brien S et al. 2011].

Measuring response in chronic myeloid leukemia in the chronic phase

Important components of treatment response

Based on available evidence from several independent series [Alvarado et al. 2009; Kantarjian et al. 2009b] and the outcomes of four expert consensus conferences, the ELN has published management recommendations, including definitions for treatment failure, suboptimal response and optimal response, in patients with CML-CP treated with imatinib. The criteria defining treatment response are summarized in Table 1.

Table 1.

Monitoring treatment response in chronic myelogenous leukemia (CML).

Complete hematologic remission	Cytogenetic remission				Molecular response
1. Normalization of peripheral counts and differential*	Karyotype analysis of metaphases from bone marrow				Quantitative PCR (RT-PCR)
2. Disappearance of all signs and symptoms of CML$		Major cytogenetic response (0–34% Ph+)			Complete molecular response	Major molecular response
Optimal response (3 months)	Complete cytogenetic response	Partial cytogenetic response	Minor cytogenetic response	No cytogenetic response	Undetectable BCR–ABL transcripts	BCR–ABL/ABL ratio of <0.1% (international scale)
	(0% Ph+)	1–35% Ph+	36–95% Ph+	>95% Ph+
Optimal response-3m	12 months	6 months	3 months	–		18 months

^*Normalization of peripheral counts is defined through platelet count less than 450 × 10⁹, white blood cell count less than 10 × 10⁹, the absence of immature granulocytes and less than 5% basophils on differential.

^$Includes resolution of splenomegaly.

Ph, Philadelphia; RT-PCR, real-time polymerase chain reaction.

Defining imatinib treatment failure

Criteria for failure (Table 2) with imatinib therapy are judged relative to the duration of the therapy. This includes the lack of a complete hematologic response (CHR) at 3 months, lack of a CyR at 6 months (Ph > 95%), lack of MCyR at 12 months (Ph > 35%) irrespective of hematologic response, and lack of CCyR at 18 months of therapy. It is also considered failure if there is a loss of response to imatinib therapy at any time, including cytogenetic or hematologic relapse [Baccarani and Dreyling, 2009]. The risk of loss of MCyR or CHR, or progression to accelerated or blast phase, is highest in the first 2–3 years of therapy, with decreases in failure rates as therapy continues [Alvarado et al. 2009; Baccarani et al. 2009; Deininger et al. 2009; Druker et al. 2006]. In the second revision of 2013 guidelines update by NCCN, hematologic response is no longer considered to make treatment response/failure decision at 3 months. Instead, quantitative real-time polymerase chain reaction (RT-PCR) using the international scale to monitor the BCR-ABL transcript levels is used. A BCR-ABL transcript level of up to 10% by quantitative RT-PCR is considered treatment response and 10% and over is considered failure. If a quantitative RT-PCR is unavailable, a partial CyR (PCyR) by bone marrow cytogenetics at 3 months is considered treatment response and anything less than PCyR is considered failure. Also, the evaluation of treatment response at 6 months was removed in this update. Instead, these guidelines suggest a 3-monthly quantitative RT-PCR for 3 years and every 3–6 months thereafter if a response is achieved at 3 months. If no response is achieved in the first 3 months, patient compliance should be evaluated and mutational analysis performed. At 12 months, if a patient has not achieved a CCyR, it is considered as treatment failure.

Table 2.

Treatment response criteria.

Time on therapy (months)	Treatment failure	Suboptimal response	Optimal response
3*	No CHR	No CG response	CHR with <95% Ph+
6*	>95% Ph+ (no CyR)	36–95% Ph+ (less partial CyR)	≤35% Ph+ (partial CyR)
12	≥35% Ph+	1–35% Ph+ (partial CyR)	0% Ph+
18	≥1% Ph+	<MMR	MMR
Any	Loss of CHR	Loss of MMR	Stable or improving MMR
	Loss of CCyR	Imatinib-sensitive mutation
	Mutation, CE

European LeukemiaNet’s (ELN) criteria for patient failure and suboptimal response to imatinib therapy.

^*These criteria were modified in the 2013 update of the National Comprehensive Cancer Network guidelines.

CCyR, complete cytogenetic response; CE, clonal evolution; CG, cytogenetic; CHR, complete hematologic response; MMR, major molecular response; Ph, Philadelphia chromosome.

Importance of suboptimal response

Suboptimal responses are those that do not meet criteria for response, or those for failure, but represent a slow or inadequate response. Suboptimal response early in therapy is more prognostic than at later time points. Patients classified as suboptimal responders at 6 months have similar outcomes in terms of overall survival (OS), progression-free survival (PFS) and event-free survival (EFS) as those whose condition fails to respond to therapy. In contrast, those with suboptimal response at 18 months of therapy have outcomes not statistically different than those classified as having an optimal response [Alvarado et al. 2009; Hughes et al. 2003; Marin et al. 2008]. Patients with suboptimal responses have greater risk of disease progression compared with optimal responders [Druker et al. 2006; Hochhaus et al. 2008b]. For this reason, NCCN guidelines recommend therapy change as if in treatment failure.

The primary goal of therapy for patients with CML is still the achievement of CCyR. Those who achieve this goal, have a low probability of eventually progressing. Also, achieving a MMR early is desirable, as it further improves the long-term outcome. Thus, evaluating response by quantitative real-time PCR (RT-PCR) for this molecular response at critical time points, as stated by the ELN, is important. In the International Randomized study of Interferon vs STI571 (IRIS) study, patients who achieved MMR by 12 months had estimated EFS of 99% at 7 years and MMR by 18 months had 100% freedom from progression to AP/BP and 95% EFS at 7 years compared with those with CCyR without MMR by 18 months [Hughes et al. 2010]. However, according to the study published by Kantarjian and colleagues, an increase in RT-PCR among patients in CCyR is not a criterion for treatment failure. Most patients with increases in RT-PCR remain in CCyR. Patients who lose a MMR or never achieve a MMR and have more than 1 log increase of RT-PCR should be monitored more closely, and may be evaluated for mutations of the BCR-ABL KD [Kantarjian et al. 2009c].

As the IRIS trial and several independent retrospective reviews have confirmed [Deininger et al. 2009; Druker et al. 2006; Hughes et al. 2003; O’Brien et al. 2003] CyR is the gold standard for assessing optimal response and predicting long-term outcome. Thus, the bone marrow must be reassessed at 3/6 months and at 12 months if CCyR was not achieved. As per international guidelines, NCCN and ELN, fluorescence in situ hybridization (FISH) and RT-PCR serve as important monitoring tools after CCyR is achieved. FISH is highly sensitive in assessing CyR, particularly when no analyzable metaphases are obtained from bone marrow. Thus, molecular response should be assessed by RT-PCR every 3 months in the peripheral blood until a MMR is achieved and then every 3–6 months [Alvarado et al. 2009; Baccarani et al. 2006; Marin et al. 2008; O’Brien et al. 2011]. The NCCN guidelines state that patients with a stable MMR can be monitored every 6 months [O’Brien et al. 2011].

It is worth mentioning that the endpoints to assess treatment response proposed by ELN are based on clinical trials with imatinib. Also, it is important to note that the CyR and survival rates obtained in the IRIS trial are not completely reproducible in the community and imatinib was shown to be less effective in obtaining the same results outside a clinical trial. Lucas and colleagues showed that, by 24 months, in 49% of patients, their condition failed to respond to imatinib treatment [Lucas et al. 2008]. Furthermore, these same endpoints may not be applicable for patients who are treated upfront with dasatinib or nilotinib. We have recently reported that among patients with newly diagnosed CML-CP treated with second-generation TKIs, the achievement of an early CCyR (e.g. at 3 or 6 months) correlated with an optimal outcome.

Potential pitfalls

Imatinib intolerance

Intolerance to imatinib therapy is one of the prime reasons for discontinuation of therapy. In general, a patient should be categorized as being intolerant to TKI therapy if the patient meets one or more of the following criteria: any life-threatening grade 4 nonhematologic toxicity; any grade 3 or 4 nonhematologic toxicity that has recurred despite dose reduction and treatment of symptoms; any grade 2 nonhematologic toxicity that persists for more than a month despite optimal supportive measures; a grade 3 or 4 hematologic toxicity that does not respond to supportive measures and requires dose reductions below the accepted minimal effective dose [Jabbour et al. 2011a]. Common side effects of imatinib in order of incidence include neutropenia, thrombocytopenia, anemia, elevated liver enzymes, edema, nausea, muscle cramps, musculoskeletal pains, rash, fatigue, diarrhea and headache [O’Brien et al. 2003]. Common toxicity criteria may be used to grade AEs and identify acute toxicities, and likewise, may help determine TKI intolerance. With long-term therapies such as TKIs, a patient’s quality of life may be a better tool to gauge therapy intolerance [Pinilla-Ibarz et al. 2011]. This is a particularly salient factor in CML since patient adherence is critical to successful long-term disease management.

Rising BCR-ABL transcript levels: when to respond

A small increase in BCR-ABL transcript levels does not necessarily indicate treatment failure or loss of response. This is supported by the 3-year follow up of the IRIS trial in which CCyR, regardless of MMR, was associated with improved OS relative to interferon and cytarabine [Kantarjian et al. 2006b]. In the setting of a previously achieved MMR, a new 1 log, or fivefold, increase in BCR-ABL transcripts by RT-PCR should be evaluated by repeating the RT-PCR in 1–3 months, and if the increase is confirmed, conventional cytogenetics on a bone marrow sample should be obtained to evaluate for loss of CyR [Branford et al. 2003; Ross et al. 2006]. Furthermore, the limitation of using molecular response as an endpoint is the lack of standardization in laboratories that perform this test [Hughes et al. 2006]. Variability among centers can be introduced because of methodological differences. Molecular studies including a combination of FISH and RT-PCR may be required to ensure concordance and high-quality stability of response. Importantly, both tests can return false-positive or false-negative results. Bone marrow cytogenetic studies are warranted every 2–3 years or more if abnormalities are found in Ph-negative diploid cells (e.g. abnormalities in chromosomes 5 or 7).

As per ELN recommendations, mutation analysis to identify any KD mutations is beneficial in cases of imatinib failure, in cases of increased BCR-ABL transcript levels leading to MMR loss, in any other case of suboptimal response (for imatinib patients), and in cases of hematologic or cytogenetic failure during second-line dasatinib or nilotinib therapy. Furthermore it was also suggested that performing regular mutational analysis is worthless given the stable nature of CML-CP on imatinib.

Chromosomal abnormalities in normal metaphases

Chromosomal abnormality (CA) in Ph-negative cells occurs in 5% of cases of CML responding to imatinib. Most disappear spontaneously with continued therapy. In a study conducted in 258 patients with newly diagnosed CML-CP at MD Anderson Cancer Center treated with imatinib, CA occurred in a small subset and was transient with no significant impact on outcome. Thus, no changes in treatment strategy are usually warranted in the absence of additional complications. The progression of CML to myelodysplastic syndrome (MDS) or acute myelogenous leukemia (AML) in those with cytogenetic abnormalities with normal peripheral counts or morphology is low. Cytogenetic analysis should suffice to monitor these patients with a ‘wait and watch’ approach, as was reiterated by Loriaux and colleagues in their 2004 study [Loriaux and Deininger, 2004]. Additional studies in larger cohorts are needed to identify the best treatment strategy for these patients [Jabbour et al. 2007].

Management of imatinib failure

Imatinib dose escalation

In a retrospective review of patients who underwent imatinib dose escalations as part of the IRIS trial according to IRIS protocol guidelines from a standard dose of 400 mg daily to doses of 600 or 800 mg daily, statistically significant improvements in responses were noted. In those who had not achieved a CHR by 3 months, 86% achieved a CHR 3 months after dose escalation and 29% went on to achieve a CCyR by 12 months. In patients who had not achieved a minor CyR by 12 months, 25% achieved a MCyR by 12 months after imatinib dose escalation and 50% by 24 months. Of these patients, 50% went on to achieve a CCyR by 48 months. In patients with a loss of MCyR, 50% regained their MCyR within 12.5 months, 33% of whom went on to achieve a CCyR. In patients who exhibited progression of disease, 67% experienced normalization of their white blood cell count. Of note, dose escalations were not attempted for those who lost CCyR [Kantarjian et al. 2009a].

In another study assessing 84 patients over 61 months whose condition failed to respond to standard-dose imatinib, doses were escalated to 600–800 mg daily. In 21 patients with hematologic failure, 48% achieved a CHR, with only 14% achieving a CyR. In 63 patients with cytogenetic failure, 75% responded with CCyR; 2- and 3-year EFS rates were 57% and 47% respectively, with OS rates of 84% and 76% respectively [Jabbour et al. 2009a]. Interestingly, among those with cytogenetic failures, a second CCyR was achieved in 52% with dose escalation and these responses were durable in 88% after 2 years. In contrast, less than 5% of those who had lost their hematologic response (25%) achieved a CyR, and even then, it was transient [Jabbour et al. 2009a].

Thus, while dose escalation after failure of standard-dose imatinib is an important and practical treatment option, it is likely to be effective only in the subset of patients with previous CyRs and is not indicated for patients with intolerance to the drug, particularly at the higher doses. Clinical consideration should be given to second-generation TKIs in this setting [Kantarjian et al. 2009b]. Current treatment guidelines recommend dose escalation of imatinib to 600 or 800 mg/day in cases of suboptimal response [O’Hare et al. 2005]. However, it should be acknowledged that there are minimal data available regarding the effectiveness of this approach [Jabbour et al. 2009a; Kantarjian et al. 2009a]. For patients with clear failure to imatinib therapy, dose escalation from an initial dose of 400 mg is no longer recommended in the 2013 update of the NCCN guidelines. The current approach is to change therapy to a second-generation TKI, unless the patient is intolerant to dasatinib or nilotinib, although allogeneic hematopoeitic stem cell transplantation (allo-HSCT) is also an option following treatment failure [Deininger, 2008; Kantarjian et al. 2006a].

Second-generation tyrosine kinase inhibitors

In an international, multicenter, phase II trial assessing 387 patients with CML-CP who had demonstrated imatinib intolerance or resistance (START-C trial), patients were treated with dasatinib 70 mg twice daily. After 2 years, data showed rates of CHR of 91%, CCyR of 53%, and MMR of 47%. Two-year PFS and OS were 80% and 94% respectively [Mauro et al. 2009]. These data were supported by another study in which dasatinib produced MCyR rates of 45% and CCyR rates of 33% in patients with CML-CP who were imatinib resistant or intolerant [Brave et al. 2008].

The dose optimization study CA180034 randomized 622 patients with imatinib-resistant or -intolerant CML-CP to four dasatinib treatment arms, including 100 mg daily, 50 mg twice daily, 140 mg daily and 70 mg twice daily. Dasatinib dosed at 100 mg daily produced similar CyR and PFS rates as other dosing schedules, with significantly fewer occurrences of grade 3–4 neutropenia, thrombocytopenia, anemia and pleural effusions. There were also fewer treatment interruptions, reductions and discontinuations at 100 mg daily compared with the other treatment groups.

Nilotinib has also been investigated as a therapeutic strategy in patients with CML-CP who have exhibited imatinib failure. One phase II study included 321 patients receiving nilotinib 400 mg twice daily after previous imatinib failure. The median length of therapy was 18.7 months, with 62% of patients receiving treatment for at least 12 months, and 42% receiving treatment for 24 months or more. The median dose delivered was 788.5 mg/day. In this population, CCyR was achieved in 46%, and of these, 56% achieved a MMR. The overall MMR rate was 28% for the entire study population. The 24-month rates of PFS and OS were 64% and 87% respectively [Kantarjian et al. 2010]. Unlike imatinib, nilotinib activity is not affected by hOCT1 [White et al. 2006].

Bosutinib was recently approved by FDA on 4 September 2012 for this indication. A phase I/II study evaluating the efficacy and safety of bosutinib (500 mg once daily) in 288 patients with imatinib-resistant or -intolerant CML-CP demonstrated that after a median follow up of 24.2 months, 86% of patients achieved CHR, 53% had a MCyR (41% with a CCyR). At 2 years, PFS was 79% and OS was 92%. Responses were seen across BCR-ABL mutants, except T315I [Gambacorti-Passerini, 2010]. These data suggest that bosutinib is effective with good tolerability in patients with imatinib-resistant or -intolerant CML-CP. Table 3 summarizes the response to second-generation TKIs dasatinib, nilotinib and bosutinib in patients who are imatinib resistant or intolerant in CML-CP, -AP and -BP.

Table 3.

Response to second-generation tyrosine kinase inhibitors (dasatinib, nilotinib and bosutinib) in patients whose condition is imatinib resistant or who are intolerant in chronic myelogenous leukemia CP, AP and BP.

	Dasatinib				Nilotinib				Bosutinib
	CP n = 387	AP n = 174	MyBP n = 109	LyBP n = 48	CP n = 321	AP n = 137	MyBP n = 105	LyBP n = 31	CP n = 146	AP n = 51	BP n = 38
Median follow up (months)	15	14	12+	12+	24	9	3	3	7	6	3
Resistant to imatinib (%)	74	93	91	88	70	80	82	82	69	NR	NR
Hematologic response (%)	–	79	50	40	94	56	22	19	85	54	36
CHR (%)	91	45	27	29	76	31	11	13	81	54	36
NEL (%)	–	19	7	6	–	12	1	0	–	0	0
Cytogenetic response (%)	NR	44	36	52	NR	NR	NR	NR	–	NR	NR
Complete (%)	49	32	26	46	46	20	29	32	34	27	35
Partial (%)	11	7	7	6	15	12	10	16	13	20	18
Survival (%) (at 12 months)	96 (15)	82 (12)	50 (12)	50 (5)	87 (24)	67 (24)	42 (12)	42 (12)	98 (12)	60 (12)	50 (10)

[Brave et al. 2008; Hochhaus et al. 2009a].

AP, accelerated phase; CHR, complete hematologic response; CP, chronic phase; LyBP, lymphoid blast phase; MyBP, myeloid blast phase; NEL, no evidence of leukemia; NR, not reported.

Choosing the right second-generation tyrosine kinase inhibitor

Safety and tolerability are important considerations in choosing a TKI. The potential impact of the drug’s AE profile on any of the patient’s preexisting conditions should be considered in choosing between second-generation BCR-ABL inhibitors. Pleural effusion is more common for patients receiving dasatinib therapy. Therefore patients with risk factors for pleural effusion such as a prior cardiac history, chronic obstructive pulmonary disease and hypertension are at greater risk for developing these complications on dasatinib therapy [Quintas-Cardama et al. 2007]. Risk factors for developing pleural effusions while taking dasatinib also include disease stage (BP > AP > CP) and previous lung problems, such as smoking or infections. The AE profile for nilotinib includes an increased risk of pancreatitis. Therefore, in patients with a prior history of severe pancreatitis, nilotinib should be used with caution and patients should be monitored closely for recurrence.

Furthermore, mutational data can help in choosing the right second-generation TKI, as certain mutations are resistant to specific drugs [Branford et al. 2009]. A study of specific BCR-ABL KD mutations identified after TKI therapy by Hochhaus and colleagues reported that failure of dasatinib therapy was more commonly associated with mutations at V299 and F317, while nilotinib resistance was associated with mutation in the P-loop, especially at Y253 and E255, or at the F311 or F359 residue [Hochhaus et al. 2008a].

Defining adequate response to second-generation tyrosine kinase inhibitors after imatinib failure

To better define adequate response to secondary TKI therapy, Jabbour and colleagues analyzed outcomes in 113 patients receiving nilotinib (N = 43) or dasatinib (N = 70) after imatinib failure. They reported that after 12 months of therapy, patients achieving a MCyR had a significant projected 1-year survival advantage of 97% versus 84% in those with a minor CyR or CHR (p = 0.02) [Tam et al. 2008]. Furthermore, in patients with CML being treated with second-generation TKIs after imatinib failure, the absence of CyR to previous imatinib therapy and a performance status of 1 or higher were found to be predictive of poor response to second-generation TKI therapy, and poor long-term outcome, warranting an alternative approach such as allo-HSCT [Jabbour et al. 2009b].

When to choose allogeneic hematopoeitic stem cell transplantation

The number of patients undergoing allo-HSCT for CML-CP has dropped significantly since TKIs were introduced. However, many patients may eventually require transplantation due to the development of TKI resistance associated with very resistant mutations or when patients otherwise progress and evolve into AP/BP. Allo-HSCT remains an important therapeutic option for CML-CP when patients are harboring the T315I mutation, in those whose condition fails to respond to second-generation TKIs [Baccarani et al. 2009], and in young patients with CML-CP who have closely matched donors available, especially if they harbor high half maximal inhibitory concentration (IC50) mutations and have not achieved a MCyR after 12 months of therapy with a second-generation TKI. Patients who do not exhibit these factors could reasonably continue TKI therapy until failure. Age may also be a factor in considering allo-HSCT. Older patients (those over the age of 70) or those unable to find a well matched stem cell donor may forgo curative transplantation for several years of controlled CML disease.

Imatinib is less effective in advanced stage chronic myelogenous leukemia

Though imatinib has revolutionized the management of CML-CP, it is less effective in the advanced phases of CML. As mentioned above, mutations, which occur more frequently in the advanced phases of disease, decrease the response rate to TKI-based therapies, leading to therapy failure [Soverini et al. 2006]. In patients with imatinib failure in CML-AP/CML-BP, a second-generation TKI with or without combination chemotherapy can be used as a bridge therapy, after which an allo-HSCT should be performed. Similarly, evidence of the T315I mutation in any phase of CML should warrant the use of a specific T315I inhibitor as a bridge therapy that would be followed by an allo-HSCT.

Upcoming therapeutic options targeting tyrosine kinase inhibitor resistance

Despite extraordinary progress, a true cure for CML is not generally achieved by ABL kinase inhibitors. TKIs are potent inhibitors of BCR-ABL kinases (among others), resulting in rapid reduction of the majority of cells carrying the Ph chromosomal marker. However, suppression of ABL-driven hematopoiesis may be insufficient to eradicate quiescent stem cells. Studies assessing the combination of TKIs with promising agents are ongoing. These combinations include TKI and hedgehog inhibitors, omacetaxine, vaccines and hypomethylatings agents. If successful, this strategy could lead to a safe and permanent discontinuation of therapy in patients with a good response. The future of CML therapy may include early use of these potent agents, perhaps in combination with new molecules, to help more patients achieve CMR, which could lead to therapy discontinuation and cure.

Ponatinib

Ponatinib (Ariad Pharma, Cambridge, MA, USA) is a potent, oral pan-BCR-ABL TKI which is under active investigation and is promising for patients with CML and Ph+ acute lymphoblastic leukemia (ALL) whose condition fails to respond to imatinib, dasatinib and nilotinib. Importantly, it is active against T315I and other imatinib-resistant mutations. In addition to inhibiting BCR-ABL, ponatinib also inhibits FLT3, FGFR, VEGFR, PDGFR and c-Kit, some of which have been implicated in the pathogenesis of AML [O’Hare et al. 2009]. In March 2010, ponatinib was granted Orphan status in the European Union for CML and Ph+ ALL.

In vitro, ponatinib potently inhibited native ABL and also several clinically relevant mutations, including T315I. In a mouse model of human CML, treatment with ponatinib resulted in complete tumor regression. Ponatinib is currently being investigated in a phase II clinical trial in patients with relapsed or refractory CML. Early results from these trials have been promising. A phase I study of oral ponatinib in patients with refractory CML/ALL or other hematologic malignancies recently reported that 66% and 53% of patients with CML-CP achieved MCyR and CCyR respectively [Cortes et al. 2010a]. Dose-limiting toxicities (DLTs) reported were elevated pancreatic enzymes, pancreatitis and rash. The study was an open-label dose-escalation study that evaluated safety and clinical responses in patients with refractory CML (CML-CP, CML-AP and CML-BP), Ph+ ALL, AML and other hematologic malignancies. Of 67 patients enrolled, 48 Ph+ patients were evaluable for response. These included 32 patients in CML-CP and 16 other patients in CML-AP and CML-BP, or Ph+ ALL. Ninety-four percent of the patients in CML-CP had CHR and 63% had a MCyR (12 complete and 8 partial). Among 11 patients in CML-CP with T315I mutations, all had a CHR and 82% had a MCyR. Among the evaluable patients with CML-AP, CML-BP and Ph+ ALL, 31% had a major hematologic response, 19% had a MCyR and 6% had a minor CyR. Of nine patients in CML-AP, CML-BP, or Ph+ ALL with T315I mutations, 33% had a major hematologic response and 20% had a MCyR. Responses were also observed in patients with highly refractory disease with either no mutations or other mutations resistant to approved TKIs (e.g. M351T, F359C, F317L, M244V and G250E). Early MMR occurred in 12 patients who were on treatment for up to 4 months, and 4 patients achieved MMR within 2 months or less. MMR was also achieved in patients with M351T, F359C, F317L, M244V, G250E mutations, and 1 patient with no mutation.

Ponatinib was generally well tolerated. In studies in which patients were treated at doses as high as 60 mg, frequently occurring drug-related AEs (≥10% any grade) included thrombocytopenia (24%), headache (14%), nausea (14%), arthralgia (13%) and fatigue (13%). Other AEs were anemia (11%), elevated lipase (11%), pancreatitis (10%), muscle spasms (11%), rash (11%) and myalgias (10%). DLT in 4/14 patients treated at 60 mg were pancreatic enzyme elevations and pancreatitis. One of 22 patients treated at the 45 mg dose had a DLT of grade 3 rash. All DLTs were reversible [Cortes et al. 2010a].

These results led to the phase II PACE (Ponatinib Ph+ALL and CML Evaluation) trial reported by Cortes and colleagues. The PACE trial enrolled 449 patients with all stages of CML or Ph+ ALL who were resistant or intolerant to dasatinib or nilotinib, or developed the T315I mutation after therapy with any TKI. Among these patients, 315 were either resistant or intolerant to imatinib therapy and 64 patients had the T315I mutation. A large majority of the patients with each stage of disease, approximately 80–90%, were resistant to dasatinib or nilotinib. The primary endpoint was a MCyR for CML-CP and a major hematologic response for patients with advanced stage disease. At a median follow up of 6.6 months, among patients in CML-CP, 49%, 41% and 26% achieved MCyR, CCyR and MMR respectively. Among patients in CML-AP, 67%, 38% and 17% attained MHR, MCyR and CCyR respectively and among those in CML-BP, 37%, 34% and 27% attained MHR, MCyR and CCyR respectively [Cortes, 2011].

Omacetaxine

Omacetaxine, a subcutaneously administered first-in-class cetaxine agent that has a mechanism of action independent of tyrosine kinase inhibition has also been evaluated in patients with Ph+ CML who have the T315I mutation and resistance to imatinib therapy. It acts through reversible, transient inhibition of protein elongation that does not depend on BCR-ABL binding. By blocking ribosomal function, the drug decreases intracellular levels of several antiapoptotic regulatory proteins, inducing antitumor activity via apoptosis [Perez-Galan et al. 2007].

More phase II/III studies investigating the benefit of omacetaxine in patients with CML post multiple TKI failure, with a significant proportion of these patients with baseline mutations, are ongoing. Preliminary data in patients with CML-CP with T315I mutations who are resistant to imatinib show achievement of an 85% CHR rate, 28% CyR rate, 15% MCyR rate, 15% MMR rate and 57% reduction in the T315I clone. OS was not met in this group of patients. In CML-AP, 37.5% showed hematologic response with OS of 18.8 months. In CML-BP, hematologic response was demonstrated in 30% with OS of 1.8 months [Cortes-Franco et al. 2009]. An update combining both of the previous study populations with and without T315I mutations treated with omacetaxine was presented at the ASH 2010 annual meeting [Cortes et al. 2010b].

An update combining the two studies that included omacetaxine in patients with the T315I mutation whose condition had failed to respond to prior imatinib and in patients with resistance or intolerance to at least two TKIs was recently presented. In the CML-CP group, 16 patients (20%) achieved a MCyR and 4 (5%) had a minor CyR; the median duration of MCyR was 18 months. A total of 56 (69%) achieved a CHR with a median response duration of 12.2 months (range 8–26); the median OS was 34 months.

In the CML-AP group, 11 patients (27%) had a MHR, including 10 (24%) with CHR and 1 (2%) with no evidence of leukemia; the median duration of MHR was 9 months (range 4–14). In addition, 2 (5%) achieved a return to chronic phase and 3 (7%) had hematologic improvement. Three (7%) patients had a minor and 3 (7%) had a minimal CyR. The median OS was 16 months. Omacetaxine was well tolerated in this population. This study suggests that patients with CML whose condition has failed to respond to previous treatment with at least two TKIs may attain a clinically meaningful response to omacetaxine therapy.

Conclusion

The introduction of TKIs and their implementation in the treatment of CML have changed the management and outcome of this disease dramatically. They have transformed this disease from an immediately life-threatening leukemia, with a 10–20% mortality rate per year, to a chronic disease, managed with oral medications, and with 1–2% mortality per year. Although good results were obtained with the first-generation TKI, imatinib, the results were not consistent among some patients who developed imatinib resistance. As the resistance mechanisms become more evident, newer drugs are being developed to overcome this problem. Mutational analysis should be performed in those with imatinib failure, escalating BCR-ABL transcript levels and those with suboptimal response. Mutational data also help in choosing the right second-generation TKI based on the point mutations that led to imatinib failure. Of particular importance is the development of ponatinib, a pan-BCR-ABL TKI that targets multiple point mutations, including T315I, which is notoriously associated with imatinib and multiple second-generation TKI failure. The introduction of novel therapies may further improve outcomes and address the common mechanisms of resistance in the treatment of CML.