Chronic Myeloid Leukaemia

Chronic myeloid leukaemia - technical article

Essentials

Chronic myeloid leukaemia has a worldwide incidence of 1 to 2 per 100 000 population. Most cases are caused by translocation of the distal end of chromosome 9 on to chromosome 22 (known as a Philadelphia chromosome), which leads to the creation of a fusion protein from the BCR (break-point cluster region) and ABL genes that is a constitutive tyrosine kinase and appears to operate as an initiator for the development of the leukaemia. Why this translocation occurs is not known.

Clinical features, diagnosis, and (historical) prognosis

Clinical features—many patients are asymptomatic at diagnosis, which is made following a routine blood test. Others present with signs and symptoms including fatigue, sweats, fever, weight loss, haemorrhagic manifestations, and abdominal discomfort (due to splenomegaly).

Diagnosis—this is typically made by the examination of a peripheral blood film (revealing features including increased numbers of myelocytes) and the demonstration of the Philadelphia chromosome by conventional cytogenetics on a bone marrow aspirate sample. PCR analysis of peripheral blood or marrow confirms the presence of a BCR-ABL1 gene and characterizes the BCR-ABL1 junction.

Prognosis—before the introduction of tyrosine kinase inhibitors (see below) the condition, having usually been diagnosed in the chronic phase, then spontaneously progressed after (typically) 3 to 6 years to myeloid (or less commonly lymphoid) blast transformation, which had very poor prognosis.

Treatment

Aside from supportive care, first line management is with the tyrosine kinase inhibitor imatinib, which induces complete haematological remission in 98% of all previously untreated patients and prolongs overall survival very substantially, although it does not totally eradicate disease in most cases. Patients with sub-optimal responses to imatinib can be offered (1) newer tyrosine kinase inhibitors—e.g. dasatinib, nilotinib, and bosutinib; and (2) allogeneic stem cell transplantation—for patients less than 50 years of age and with a suitable donor.

Introduction

Patients with the rare malignancy, chronic myeloid leukaemia (CML), have been well served by translational research over the past half century. Though the disease was first described in 1845 and characterized by the 1920s, it was 60 years before the unravelling of initiating molecular events paved the way to define specific targets for treatment. CML is a clonal disease that results from an acquired molecular change in a pluripotential haematopoietic stem cell. The leukaemia cells have a consistent cytogenetic abnormality, the Philadelphia (Ph) chromosome, which carries a BCR-ABL1 fusion gene. This gene encodes a Bcr-Abl1 oncoprotein with enhanced tyrosine kinase activity, which is generally considered to be the ‘initiating event’ in the chronic phase of CML, though there remains debate as to whether this is the first molecular event in all cases.

Imatinib mesylate can inhibit the enzymatic activity of this dysregulated tyrosine kinase of the Bcr-Abl1 oncoprotein and has now become the preferred treatment for all newly diagnosed patients with CML, except perhaps for some children. Imatinib substantially reduces the number of leukaemia cells in a patient’s body, and comparison with historical data confirms the notion that it prolongs overall survival very substantially. However, complete molecular responses occur only in a minority of patients, and allogeneic stem cell transplantation (allo-SCT) remains the only treatment that can reliably produce complete and durable molecular remission due presumably to eradication of all residual leukaemia stem cells. The second-generation ABL-kinase inhibitors, notably dasatinib and nilotinib, are now firmly established in clinical use for the treatment of imatinib-resistant/refractory CML and Ph-positive acute lymphoblastic leukaemia. Both drugs are now being tested as potential first-line therapies.

Epidemiology

The annual incidence of CML is constant worldwide at about 1.0 to 2.0 per 100 000 of population per annum. In the Western world, it represents approximately 15% of all adult leukaemias and less than 5% of all childhood leukaemias, but it is higher as a percentage of all leukaemias in China and India, where chronic lymphocytic leukaemia is very rare. In the Western world, the median age of onset is about 50 years, and there is a slight male excess. In contrast, the median age of onset may be considerably younger, around 36 years, in some countries such as India.

Aetiology

For most patients with CML, possibly for all, there appear to be no obvious predisposing factors, and the disease arises sporadically. Epidemiology studies have suggested a marginal increment in the number of cases of CML following exposure to high doses of irradiation as occurred in survivors of the Hiroshima and Nagasaki atomic bombs in 1945. A small number of families with a high incidence of the disease have also been reported, though no specific HLA genotypes have been identified. One convincing case has been reported of CML recurring in cells of donor origin following related allo-SCT.

Natural history

CML is a remarkably heterogeneous disease. Before the introduction of tyrosine kinase inhibitors, it typically ran a biphasic or triphasic course. It was usually diagnosed in the chronic phase, which lasted typically 3 to 6 years; the leukaemia then spontaneously progressed to blast transformation. About 70 to 80% of patients had a myeloid blast transformation, and they usually survived 2 and 6 months; the 20 to 30% of patients with a lymphoid blast transformation had a slightly better survival. About half the patients in the chronic phase transformed directly into blast transformation, and the remainder did so following a period of accelerated phase.

Soon after the introduction of imatinib mesylate, it was observed that the natural history for most patients with CML who received this drug as initial therapy, particularly for patients who remain in complete cytogenetic response beyond the fourth year of therapy, was very greatly improved. The recent follow-up of the phase III prospective trial (IRIS) which compared imatinib to the previous best nontransplant therapy, interferon-α (IFN-α) and cytarabine, showed that about 60% of the original cohort randomized to receive the drug were still taking imatinib and were still in complete cytogenetic remission 6 years after starting the drug; none of these patients had entered the more advanced phases of the disease. Patients presenting in the late chronic phase appear to fare less well, and those in the advanced phases, particularly the blast phase, generally do poorly, including those who did initially respond to imatinib mesylate. In patients with lymphoid blast phase CML, there appear to be no durable responses beyond 6 months.

For patients who receive an allo-SCT, most remain in a complete cytogenetic and molecular remission for 10 years or more. Many of these patients are probably completely cured, if one defines cure as the complete eradication of all molecular evidence of the disease. It is of interest, however, that some of these patients do become intermittently positive for Bcr-Abl1 transcripts, albeit at low levels, but the rare patient with a persisting high transcript level is at a high risk of relapse. A very small minority appear to relapse directly into the advanced phases of the disease.

Clinical feature and diagnosis

Current estimates suggest that one-third to one-half of patients with CML are totally asymptomatic at diagnosis, which is made following a routine blood test. The remainder present with signs and symptoms often of about 3 months’ duration and related to altered haematopoiesis, particularly anaemia and platelet dysfunction and increasing disease burden, resulting in splenomegaly. Most patients will have leucocytosis due to increased numbers of myelocytes and segmented neutrophils; basophilia is almost universal, and some patients have a eosinophilia. The anaemia tends to be mild and normochromic normocytic in nature; some patients have a degree of thrombocytosis. Nearly all patients diagnosed in the advanced phases of CML are symptomatic. Occasionally patients may present with an extramedullary event, such as a chloroma.

Classical clinical features include sweats, weight loss, haemorrhagic manifestations such as spontaneous bruising and retinal haemorrhages, abdominal discomfort due to splenomegaly, fatigue often, but not always, related to anaemia and fever (Bullet list 1). The diagnosis is typically made by the examination of a peripheral blood film and the demonstration of the Ph chromosome by conventional cytogenetics on a bone marrow aspirate sample. Most haematologists carry out a bone marrow trephine examination also; this is often hypercellular with complete or near complete loss of fat spaces and a high myeloid to erythroid ratio. There may be up to 10% of blast cells.

Sometimes the diagnosis is made by demonstrating the presence of a BCR-ABL1 gene by fluorescence in situ hybridization (FISH) on a peripheral blood sample. Modern practice dictates the use of a baseline real-time quantitative polymerase chain reaction (RQ-PCR) analysis of peripheral blood or marrow to confirm the presence of a BCR-ABL1 gene and characterize the BCR-ABL1 junction. Such an analysis is particularly useful in the subsequent monitoring of patients.

Bullet list 1 Clinical features of patients with chronic phase CML seen at the Hammersmith Hospital, London

  • Fatigue 33.5%
  • Bleeding 21.3%
  • Weight loss 20.0%
  • Abdominal discomfort (left upper quadrant) 18.6%
  • Sweats 14.6%
  • Bone pain 7.4%
  • Splenomegaly 75.8%
  • Hepatomegaly 2.2%

(Adapted from Savage D, et al. (1997). Clinical features at diagnosis in 430 patients with chronic myeloid leukaemia seen at a referral centre over a 16-year period. Br J Haematol, 96, 111–16.)

Molecular biology

The Ph chromosome is an acquired cytogenetic abnormality present in all leukaemic cells of the myeloid lineage and in some B cells and T cells. It is formed as a result of a reciprocal translocation of DNA from chromosomes 9 and 22, t(9; 22)(q34;q11). The classical Ph chromosome is easily identified in about 90% of CML patients. A further 5% of patients have variant translocations which may be ‘simple’ involving chromosome 22 and a chromosome other than chromosome 9, or ‘complex’, where chromosome 9, 22, and other additional chromosomes are involved. About 5% of patients with clinical and haematological features typical of CML lack the Ph chromosome and are referred to as having ‘Ph-negative’ CML. About half of these patients have a BCR-ABL1 chimeric gene and are referred to as Ph-negative, BCR-ABL1-positive cases; the remainder are BCR-ABL1-negative, and some of these have mutations in the RAS gene. These BCR-ABL1-negative patients have a more aggressive clinical course. Some patients acquire additional clonal cytogenetic abnormalities, in particular +8, +Ph, iso17q–, and +19, as their disease progresses. The emergence of such clones may herald the onset of blastic transformation.

The various genetic events have now been elucidated, and the chimeric BCR-ABL1 gene is believed to play a central role in the pathogenesis of CML, though the precise mechanism(s) are still not fully understood. Three distinct breakpoint locations in the BCR gene in chromosome 22 have been identified. The break in the major breakpoint cluster region (M-bcr) occurs in the intron between exon e13 and e14 or in the intron between exon e14 and e15 (toward the telomere). By contrast, the position of the breakpoint in the ABL1 gene on chromosome 9 is highly variable and may occur at almost any position upstream of exon a2. The Ph translocation results in the juxtaposition of 5′ sequences from the BCR gene with 3′ sequences from the ABL1 gene. This event results in the generation of the chimeric BCR-ABL1 fusion gene transcribed as an 8.5-kbp mRNA. This mRNA encodes a protein of 210 kDa (p210BCR-ABL1) that has a greater tyrosine kinase activity compared with the normal ABL protein. The different breakpoints in the M-bcr result in two slightly different chimeric BCR-ABL1 genes, resulting in either an e13a2 or an e14a2 transcript. The type of BCR-ABL1 transcript has no important prognostic significance.

The second breakpoint location in the BCR gene occurs between exons e1 and e2 in an area designated the minor breakpoint cluster region (m-bcr) and forms a smaller BCR-ABL1 fusion gene. This is transcribed as an e1a2 mRNA which encodes a p190BCR-ABL1 oncoprotein. This protein characterizes about two-thirds of patients with Ph-positive acute lymphoblastic leukaemia (ALL). A third breakpoint location is found in patients with the very rare Ph-positive chronic neutrophilic leukaemia. This has been designated as a micro breakpoint cluster region (µ-bcr) and results in e19a2 mRNA, which encodes a larger protein of 230 kDa (p230BCR-ABL1).

The recognition of several features in the Bcr-Abl1 oncoprotein that are essential for cellular transformation led to the identification of signal transduction pathways activated in BCR-ABL1-positive cells. Much attention has since focused on determining the precise role played by the various Bcr-Abl1 proteins in the pathogenesis of CML. A number of possible mechanisms of BCR-ABL1-mediated malignant transformation have been implicated, which are not necessarily mutually exclusive. These include constitutive activation of mitogenic signalling, reduced apoptosis, impaired adhesion of cells to the stroma and extracellular matrix, and proteasome-mediated degradation of Abl inhibitory proteins. The deregulation of the Abl tyrosine kinase facilitates autophosphorylation, resulting in a marked increase of phosphotyrosine on Bcr-Abl1 itself, which creates binding sites for the SH2 domains of other proteins. A variety of such substrates, which can be tyrosine phosphorylated, have now been identified. Although much is known of the abnormal interactions between the Bcr-Abl1 oncoprotein and other cytoplasmic molecules, the finer details of the pathways through which the ‘rogue’ proliferative signal is mediated, such as the RAS-MAP kinase, JAK-STAT, and the PI3 kinase pathways, are incomplete, and the relative contributions to the leukaemic ‘phenotype’ are still unknown. Moreover, the multiple signals initiated by the Bcr-Abl1 have both proliferative and antiapoptotic qualities, which are often difficult to separate. Much remains to be learned about the significance of tyrosine phosphatases in the transformation process.

It is generally believed that the some CML stem cells, at a cytokinetic level, are in a ‘quiescent’ or ‘dormant’ (G0) phase. These quiescent CML cells appear to be able to exchange between a quiescent and a cycling status, allowing them to proliferate under certain circumstances. This provides some rationale for autografting as treatment for CML. There is also evidence that some Ph-positive cells are quiescent and cannot be eradicated by cycle-dependent cytotoxic drugs, even at high doses, or indeed by imatinib mesylate.

It is likely that the acquisition of a BCR-ABL1 fusion gene by a haematopoietic stem cell and the ensuing expansion of the Ph-positive clone set the scene for acquisition and expansion of one or more Ph-positive subclones that are genetically more aggressive than the original Ph-positive population. The propensity of the Ph-positive clone to acquire such additional genetic changes is an example of ‘genomic instability’, but the molecular mechanisms underlying this instability are poorly defined. Such new events may occur in the BCR-ABL1 fusion gene or indeed in other genes in the Ph-positive population of cells and presumably underlie the progression to advanced phases of the disease. The average length of chromosomal telomeres in the Ph-positive cells is generally less than that in corresponding normal cells, and the enzyme telomerase, which is required to maintain the length of telomere, is up-regulated as the patient’s disease enters the advanced phases. About 25% of patients with CML in myeloid transformation have point mutations or deletions in the p53 gene, and about half of all patients in lymphoid transformation show homozygous deletion in the p16 gene. There is also evidence supporting the role of the RB (retinoblastoma) and the MYC genes in disease progression.

Prognostic factors

Various efforts have been made to establish criteria definable at diagnosis, both prognostic (disease-related) and predictive (treatment-related), that may help to predict survival for individual patients. The most frequently used method is that proposed by Sokal in 1984, whereby patients can be divided into various risk categories based on a mathematical formula that takes into account the patient’s age, blast cell count, spleen size, and platelet count at diagnosis. The Euro or Hasford system is an updated Sokal index, which includes consideration of basophil and eosinophil numbers. Stratifying patients into good-, intermediate-, and poor-risk categories may assist in the decision-making process regarding appropriate treatment options. Recent observations, however, suggest that age per se might not influence the biology of the disease, but rather increases the probability of treatment-related adverse effects by virtue of potential comorbid conditions.

More recent efforts have identified other possible risk-stratification factors. Green and colleagues from Cambridge, United Kingdom, described the presence of small deletions in the region of the reciprocal ABL-BCR1 fusion gene on the derivative 9q+ chromosome arm, which were associated with poor prognosis, at least in the IFN-α era, though not necessarily in the imatinib era. It is of interest that a number of imatinib mesylate–specific parameters, which may carry a prognostic significance, such as the degree of myelosuppression associated with treatment with imatinib and the plasma levels of the agent at a specific time, are now being described. There is also evidence that the depth and quality of the cytogenetic and molecular responses may be influenced by the actual dose of imatinib mesylate, with patients who receive higher doses of imatinib mesylate achieving a complete cytogenetic remission and a major molecular response much earlier than those treated with conventional doses, and there is a trend for an improved event-free survival.

Gene expression changes associated with progression and response in CML have also been introduced and are currently being validated. Other candidate biomarkers include the polycomb group BMI1 gene, which regulates both normal and leukaemic stem cells, and the rate of shortening of telomeres in the leukaemia clone.

A decade ago, it was standard practice to recommend an allo-SCT to all patients younger than 50 years of age with newly diagnosed CML in the chronic phase, provided they had suitable HLA-identical sibling or ‘matched’ unrelated donors. Patients presenting in the advanced phases of CML usually received combination chemotherapy, often followed by an allo-SCT if a ‘second’ chronic phase could be achieved. The treatment algorithm for newly diagnosed patients changed dramatically once the impressive success of imatinib mesylate in inducing durable complete cytogenetic remissions in the majority of newly diagnosed patients with CML in the chronic phase was recognized. Imatinib mesylate is now the preferred treatment for most, if not all, newly diagnosed patients with CML in the chronic phase, and is also useful in the management of patients presenting in the advanced phase.

Imatinib mesylate inhibits the enzymatic action of the activated Bcr-Abl1 tyrosine kinase by occupying the ATP-binding pocket of the kinase component of the Bcr-Abl1 oncoprotein, thereby blocking the capacity of the enzyme to phosphorylate and activate downstream effector molecules that cause the leukaemic phenotype. It also binds to an adjacent part of the kinase domain in a manner than holds the Abl–activation loop of the oncoprotein in an inactive configuration.

Imatinib mesylate induces ‘cumulative best’ complete haematological remissions in 98% of all previously untreated patients with CML in the chronic phase and ‘cumulative best’ complete cytogenetic remission in about 85% of such patients. About 2% of all patients in the chronic phase, progress to advanced-phase disease each year, which contrasts with estimated annual progression rates of >15% for patients treated with hydroxycarbamide and about 10% for patients receiving IFN-α, either with or without cytarabine. Indeed, preliminary evidence suggests that this annual rate of progression, about 2%, might actually be diminishing as the years pass. Complete molecular responses are, however, less common, and imatinib mesylate will probably not eradicate residual CML in most patients.

Nevertheless, the drug prolongs overall survival very substantially compared with historical patients who received IFN-α or hydroxycarbamide. Furthermore, the majority of patients who achieve a ≥3-log reduction in Bcr-Abl1 transcripts remain alive in complete cytogenetic remission at 6 years after initiation of treatment with imatinib. Therefore, a current topical issue is whether total eradication of all residual leukaemia stem cells is actually necessary, since the survival of small numbers of residual leukaemia stem cells might well be compatible with an individual patient’s long-term survival. Conversely, if such a population of cells remained a possible source for progression of the leukaemia, the notion of two contrasting therapeutic algorithms for patients based on prognostic factors—both disease-related, such as the Sokal risk score, and treatment-related, such as the European Group for Blood and Marrow Transplantation (EBMT) risk score—can be considered.

For most patients with CML in the chronic phase, imatinib mesylate at a standard starting dosage of 400 mg/day is the first-line treatment of choice, although several studies suggest that higher doses, up to 800 mg daily, might give better results with a greater proportion of patients achieving a complete molecular response. Such studies also suggest better progression-free and transformation-free survivals, but with potentially more adverse effects, particularly myelosuppression. The higher-dose studies are still ongoing, and until the longer term results are available, it is reasonable to start newly diagnosed patients in the chronic phase on imatinib mesylate at 400 mg/day.

Adverse effects of imatinib mesylate, which appear to be dose-related, include nausea, headache, rashes, bone pains, various skin reactions, and fluid retention. Significant cytopenias and hepatotoxicity occur less commonly and usually in the first 6 to 12 months of therapy. Very rare cases of severe or fatal cerebral oedema have been reported, and there have been some concerns about potential cardiomyopathy, though the recent 6-year IRIS study analysis reassures us that this might not be a major problem, except for older patients, who might have other predisposing cardiac risks and have anaemia. Another potential issue of concern is the potential teratogenicity of imatinib mesylate. A recent study assessing outcomes in 125 of 180 study patients exposed to the agent during pregnancy concluded that about half of the offspring born were normal; 28% of the study cohort elected to undergo termination of pregnancy, including three postidentification of fetal abnormalities. In sum there were 12 infants in whom abnormalities were identified, including 3 who had strikingly similar complex malformations. It would therefore appear sensible to avoid imatinib mesylate exposure during pregnancy.

The alternative treatment option should involve an early allo-SCT for the small minority of patients who would clearly benefit from an immediate transplant compared with continuing imatinib irrespective of its outcome. A retrospective analysis from the Center for International Blood and Marrow Transplant Research (CIBMTR) and the EBMT suggests that for adult patients, including those who are at low risk for transplant-related mortality by EBMT criteria, it is not possible to identify a cohort who would clearly benefit from an immediate stem cell transplant vs continuing imatinib mesylate irrespective of the outcome. The best initial treatment for children, for adults with a potential syngeneic donor, and possibly for adults with high-risk disease by Sokal or Hasford criteria is still uncertain. The current EBMT experience, however, suggests that patients with high-risk disease and a low-transplant risk should probably still be considered for an early transplant. Such a cohort, if treated with imatinib in the first instance should probably not receive a second tyrosine kinase inhibitor on relapse (see below) and rather proceed to stem cell transplantation. For children, many paediatric haematologists recommend initial treatment by allo-SCT for patients under the age of 16 years who have HLA-identical siblings, largely because of a lack of adequate long-term data on the use of imatinib mesylate as first-line therapy in children.

Second line-therapy and issues regarding resistance to imatinib mesylate

Various efforts to define ‘failure’ and ‘suboptimal responses’ to imatinib have resulted in two principal consensus panels. Primary resistance or refractoriness to the drug appears to be very rare and when seen may be related to poor drug compliance, poor gastrointestinal absorption, p450 cytochrome polymorphisms, interactions with other medications, and abnormal drug efflux and influx at the cellular level. In a small cohort of patients, a correlation between the transcription factor OCT-1 expression and response has been observed: the higher the levels of OCT-1, the better the molecular responses.

A somewhat larger proportion of patients, about 20% in the chronic phase, respond initially to imatinib mesylate and then lose their response. This acquired or ‘secondary’ resistance results from a variety of mechanisms, including amplification of the BCR-ABL1 fusion gene, relative overexpression of Bcr-Abl1 protein, and expansion of subclones with point mutations in the BCR-ABL1 kinase domain (KD). Such point mutations code for amino acid substitutions that may impede binding of imatinib but do not impair phosphorylation of downstream substrates that mediate the leukaemia signal. The precise position of the mutation appears to dictate the degree of resistance to imatinib; some mutations are associated with minor degrees of drug resistance, whereas one notorious mutation, the replacement of isoleucine by threonine at position 315 (T315I), is associated with near-total nonresponsiveness to imatinib, as well as with resistance to the newer tyrosine kinase inhibitors, namely dasatinib, nilotinib, and bosutinib. The precise significance and indeed the kinetics of the over 70 currently well-characterized mutations remain largely unknown.

The majority of patients who are resistant/intolerant to imatinib mesylate should receive dasatinib or nilotinib. Dasatinib is a thiazole-carboxamide structurally unrelated to imatinib. Furthermore, it binds to the ABL KD regardless of the conformation of the activation loop—whether open or closed. It also inhibits some of the Src family kinases. Preclinical studies showed that dasatinib was 300-fold more potent than imatinib and is active against 18 of 19 tested imatinib-resistant KD mutant subclones, with the notable exception of the T315I mutant.

Current experience with dasatinib in patients with chronic-phase CML resistant/refractory to imatinib suggest that about 90% of the patients have a complete haematological response and 52% have a complete cytogenetic remission. About 25% of patients with the more advanced phases of CML and Ph-positive ALL also achieve reasonable responses. Responses are seen in patients with most of the currently known ABL-kinase mutations, except the T315I mutation. Haematological toxicity is common, particularly in those with advanced phases of CML and Ph-positive ALL. These include neutropenia (49%), thrombocytopenia (48%), and anaemia (20%). Nonhaematological toxicity includes diarrhoea, headaches, superficial oedema, pleural effusions, and occasional pericardial effusions. Grade 3/4 side effects are rare, and grade 3/4 pleural effusions occurred in 6% of patients.

Dasatinib is also being assessed as a potential first-line treatment, and studies involving patients in the chronic phase, following at least 3 months of dasatinib therapy, show 89% complete haematological response and 79% cytogenetic remission rates. These data compare well with first-line responses to imatinib. The toxicity profile appears to be similar to that seen in the drug-resistance studies, but the number of patients entered so far is quite small. The drug has also been tested in patients with CML in advanced phases whose disease was resistant to both imatinib and nilotinib; remarkably, 57% haematological responses, including 43% complete haematological remission, were observed. Among those patients who had a haematological response, 32% had a cytogenetic response, including 2 patients who achieved cytogenetic remission.

Nilotinib, like imatinib, acts by binding to the closed (inactive) conformation of the Abl-KD, but with a much higher affinity. Like imatinib, it inhibits the dysregulated tyrosine kinase activity of the Abl kinase by occupying the ATP-binding pocket of the oncoprotein and blocking the capacity of the enzyme to phosphorylate downstream effector molecules. In vitro studies suggest that nilotinib is about 30- to 50-fold more potent than imatinib mesylate. Nilotinib is also active in 32 of the currently 33 imatinib-resistant cell lines with mutant Abl kinases, but like imatinib and dasatinib has no activity against the Bcr-Abl1T315I mutation. Phase II studies in patients who are resistant or intolerant to imatinib mesylate are still in progress, and preliminary results suggest a complete haematological response in about 70% or a third of these patients who show a complete cytogenetic remission. Patients in the advanced phases of CML also respond, but to a lesser degree. The most common treatment-related toxicity is myelosuppression, followed by headaches, pruritus, and rashes. Overall, 22% of the patients in the chronic phase experienced thrombocytopenia, with 19% having either grade 3/4 severity; 16% had neutropenia and a further 16% had anemia. Most of the nonhaematological adverse effects were of a grade 1/2 severity. All, including the haematological effects, were fully reversible. About 19% of all patients experience arthralgias, and about 14% experience fluid retention, particularly pleural and pericardial effusions. Importantly, patients with the acquired Bcr-Abl1T315I mutation appear to be refractory to nilotinib.

As discussed earlier, based on current EBMT experience, it is reasonable to consider an early allogeneic transplantation for those patients who are resistant to imatinib and have high-risk disease, by Sokal and Hasford risk stratification, and a low transplant risk, by EBMT criteria, and wish to be transplanted, rather than subjecting them to the next generation tyrosine kinase inhibitors. An alternative would be to prescribe a second tyrosine kinase inhibitor for a defined period and then to proceed to an allo-SCT if the response is suboptimal. In practice, however, many patients will opt to receive a trial of a next generation of the tyrosine inhibitor drugs.

Third-line therapy

For patients who are resistant/refractory to the current generation of inhibitors, and are under the age of 50 years, it is probably best to consider an allo-SCT, provided a suitable donor is identified and the patient remains in the chronic phase of the disease. For those with advanced-phase disease, one could offer combination chemotherapy or an appropriate clinical trial assessing one of the newer drugs and then consider allo-SCT if a second chronic phase is achieved. Clearly, this is an area which is evolving rapidly, so it is difficult to make firm recommendations at present.

Patients who proceed to a transplant after treatment with imatinib appear to have a higher relapse incidence than those who have not previously received the drug. This most probably represents a selection bias for relatively resistant disease. Preliminary data based on small patient series do not, however, suggest that prior treatment with imatinib increases the probability of transplant-related mortality. Moreover, patients with kinase domain mutations appear to fare as well post-transplant as those lacking such mutations. The experience with allo-SCT after initial treatment of advanced-phase disease with imatinib mesylate is still limited.

Investigational approaches

Immunotherapy

Following the realization that a molecular remission and ‘cure’ might not be possible with imatinib alone, many efforts were directed to exploring the potential of developing an active specific immunotherapy strategy for patients with CML by inducing an immune response to a tumour-specific or tumour-associated antigen. The principle involves generating an immune response to the unique amino acid sequence of p210 at the fusion point. Clinical responses to the Bcr-Abl1 peptide vaccination, including complete cytogenetic remissions, have been reported in a small series. In contrast to previous earlier unsuccessful attempts, the current series included administration of granulocyte–macrophage colony-stimulating factor as an immune adjuvant, and patients were only enrolled if they had measurable residual disease and human leucocyte antigene known to which the selected fusion peptides were predicted to bind avidly. If these results can be confirmed, vaccine development against Bcr-Abl1 and other CML-specific antigens could become an attractive treatment for patients who have a minimal residual disease status with imatinib mesylate. Other targets for vaccine therapy now being studied include peptides derived from the Wilms tumor 1 protein, proteinase 3, and elastase, all of which are overexpressed in CML cells.

Investigational drugs

Bosutinib (SKI-606, Wyeth) is an emerging oral dual Abl/Src kinase inhibitor currently in phase I/II study. This drug appears to be about 200 times more potent than imatinib, and unlike imatinib and dasatinib, does not inhibit other targets such as KIT or platelet-derived growth factor receptor. The preliminary results of treating patients with CML in chronic and advanced phase, as well as Ph-positive ALL, appear encouraging, and the toxicity profile appears reasonable, with gastrointestinal cutaneous toxicity being the major grade 3/4 adverse effects.

Other specific inhibitors of signal transduction pathways downstream of Bcr-Abl1 have been tested alone and in combination with imatinib mesylate. Some of these agents, such as 17-allylaminogeldanamycin (17-AAG), are just entering formal clinical trials or might do so in the near future. 17-AAG, a drug which degrades the Bcr-Abl1 oncoprotein by inhibiting the heat shock protein 90, a molecular chaperone required for stabilization of Bcr-Abl1, has just entered phase I studies. 17-AAG appears to have activity in patients with the E255K and T315I mutations. It also down-regulates BCR-ABL1 mRNA, though the precise mechanism remains unclear. Another novel tyrosine kinase inhibitor, PD166 326, also appears to have significant activity in patients with the H396P and M351T mutations. This agent also appears to be superior to imatinib in murine models. Other potential agents include rapamycin, an mTOR inhibitor, and wortmannin, which is a PI3K inhibitor not currently available in a formulation suitable for clinical use. Rapamycin synergizes with imatinib mesylate in inhibiting Bcr-Abl1–transformed cells, including those that are imatinib mesylate resistant.

Recently, there has been considerable interest in combining imatinib mesylate with diverse agents, including hypomethylating agents, farnesyl transferase inhibitors, pegylated IFN-α, arsenic trioxide, bortezomib, and other cytotoxic drugs, such as homoharringtonine. Homoharringtonine is a semisynthetic plant alkaloid that enhances apoptosis of CML cells, is active in combination with imatinib in drug-resistant/refractory patients.

Conclusions

The substantial understanding of the molecular features and pathogenesis of CML has provided important insights into targeting treatment to specific molecular defects. The successful introduction of imatinib mesylate as targeted therapy for CML has made the approach to management of the newly diagnosed patient fairly complex. The second generation of tyrosine kinase inhibitors, dasatinib and nilotinib, have significant activity in selected patients in both chronic and the more advanced phases of the disease, who are resistant to imatinib. Efforts are also in progress to assess the potential first-line role of both these drugs. The notion that the graft-vs-leukaemia effect is the principal reason for success in patients with CML subjected to an allograft transplant has renewed interest in immunotherapy. The use of kinase inhibitors in conjunction with various immunotherapeutic strategies is now being studied.

For the moment, the various treatment options should be assessed carefully in terms of the relative risk–benefit ratios, and a management strategy should be developed accordingly. For a small minority of patients, namely children or adults in whom the risk of transplant-related mortality is relatively low but with high Sokal risk stratification, it appears reasonable to recommend an early allo-SCT. It would also be reasonable to contemplate an early transplant procedure for patients who have an identical twin donor. For all other patients (who constitute the majority), it is best to commence imatinib mesylate at 400 mg/day, increasing to a maximum of 800 mg/day in patients with a suboptimal response.

Patients who are resistant/refractory to imatinib should receive dasatinib or nilotinib. For those who are resistant/refractory to these drugs, it is best to consider an allo-SCT, provided that a suitable donor is identified or after an appropriate clinical trial assessing emerging drugs, for those with a T315I mutation. Clearly, this is an area which is evolving rapidly; it is difficult to make firm recommendations at present.

Finally efforts in improving the technology of allo-SCT, such as the ability to prevent graft-vs-host disease without abrogation of graft-vs-leukaemia, are also in progress. If successful, they might restore allo-SCT as an alternative treatment option for some newly diagnosed patient, who might otherwise have to continue lifelong therapy with a tyrosine kinase inhibitor at considerable expense, both financial and personal.