Acute Lymphoblastic Leukaemia
Genetic changes in key progenitor cells in acute lymphoblastic leukaemia (ALL) are typically either point mutations or (more frequently) translocation of proto-oncogenes to active promoter sites, with such genetic rearrangements leading to aberrant protein production. In most childhood disease the first genetic events arise in utero, with one or more further events usually being required to convert a preleukaemic clone or clones into overt ALL.
Clinical features and diagnosis
Clinical features—typical presentation is with manifestations including pallor, petechiae/bruising, fever with or without lymphadenopathy, and abdominal and/or limb pain (arising from organ or bone infiltration). Life-threatening risks include Gram-negative or -positive septicaemia and fungaemia, bleeding (especially into the brain), and tumour lysis.
Diagnosis—the key investigations are examination of peripheral blood and bone marrow for blast cell infiltration. Immunophenotyping, cytogenetics, and gene expression arrays are important in defining subsets with variable prognostic and drug resistance profiles.
Treatment and prognosis
Treatment—aside from supportive care, the key requirements are (1) induction and consolidation of remission—usually with steroids (dexamethasone), vincristine, and L-asparaginase, sometimes with the addition of an anthracycline in high risk patients; (2) or treatment directed to the central nervous system—repeated administration of intrathecal methotrexate. Only very high risk patients including some older adults now require cranial irradiation; (3) delayed intensification and maintenance of remission—long-term survival is improved by giving pulses of second-line drugs (e.g. cytarabine, cyclophosphamide), along with more steroids, vincristine and L-asparaginase; and sometimes (4) intensive therapy followed by allogeneic bone marrow transplantation—considered for patients whose disease is refractory or relapse quickly.
Prognosis—the single most important factor is the highest white cell count recorded at presentation before introduction of any intervention—those in whom it is >50 × 109/litre fare worse than those with lower counts. Most children with ALL can be cured, but 20 to 25% relapse, many unpredictably. Improvement in survival has been less marked in adults, but appears to be getting better with use of more sustained, continuous therapy rather than the previously adopted ‘pulsed’ regimen of cytotoxic treatment. Identification of minimal residual disease by new molecular technology and stratification of therapy consequently has enabled reduction of intensivity of therapy for low risk and escalation for high risk patients.
After Sidney Farber and his team (1947/8) showed that the folic acid antagonist aminopterin could induce temporary remission in acute lymphoblastic leukaemia of childhood, the search for other active agents started. The therapeutic value of adrenal corticoid steroids (1949, especially prednisolone), thiopurines (1953), methotrexate (1956, replacing aminopterin) and vincristine (1962) heralded in the era of potential cure.
The concept of total therapy was pioneered by Pinkel et al. at St Jude Children’s Research Hospital in Memphis (United States of America) and encompassed remission induction, consolidation of remission, central nervous system disease control, and maintenance or continuing therapy. Emphasis was on sustained, almost continuous therapy rigorously applied without significant gaps in therapy. The result has been an improvement in survival from little expectation of cure in the 1960s to over 80% survival for children in resource-rich countries in the 21st century. For teenagers and adults the success has been more limited but recent strict adherence to ‘paediatric’ protocols has improved their survival also by 15 to 20%. Previous approaches in adults involved more intermittent pulsed therapy which may have allowed malignant cell recovery as well as recovery from myelosuppression.
No ‘curative’ new agents have been identified since the early 1970s although a few ‘possible’ drugs are currently under investigation, some of which may offer the possibility of targeted therapy and toxicity reduction within protocols which are inevitably intensive and aggressive. Most emphasis has been on better delivery of standard drugs. Rapidity of blood and bone marrow clearance of leukaemic cells, more recently supplemented by molecular identification of minimal residual disease (not visible by conventional microscopy) has facilitated stratification of patients into standard, high, and very high risk groupings. Nevertheless, numerically most relapses still occur in standard-risk patients. Trying to understand why is the focus of much current research. Most adults with acute lymphoblastic anaemia (ALL) fit into the high or very high risk groupings. Current ‘curative’ therapy is also only available to about 20% of the children worldwide who acquire ALL. Most children and adults affected by acute leukaemia in the world do not even receive any supportive or palliative care.
Reduction in induction and remission deaths has contributed to the improved survival, and involves vigilance, early, and intensive support for febrile patients with antimicrobial prophylaxis, e.g. use of cotrimoxazole for prevention of pneumocystis infection.
Presentation and diagnosis
Historically, ALL has been diagnosed by clinical suspicion followed by peripheral blood and bone marrow examination for blast cell infiltration.
Growth and proliferation of lymphoid precursors arrested at an immature stage of differentiation suppress production of normal bone marrow cells. This causes thrombocytopenia and bleeding/petechiae, myelosuppression and immunosuppression increasing risk of infection, fever and anaemia. The classic presentation is with a constellation of signs and symptoms including pallor, petechiae/bruising, fever with or without lymphadenopathy (including mediastinal swelling in T cell ALL) and abdominal and/or limb pain. The latter arise from organ or bone infiltration with leukaemia. About 3 to 5% of patients present with overt central nervous system infiltration. Hepatosplenomegaly is variable but more frequently seen in childhood ALL (especially in T-ALL and mature B-ALL disease). Organ infiltration of the liver and kidneys can complicate diagnosis and induction of remission. Especially if the white cell count is high, tumour lysis syndrome (hyperkalaemia, hyperuricaemia, uraemia, hyperphosphataemia and hypocalcaemia) can occur before the start of treatment or concurrent with it. Presentation with joint swelling and pain can cause confusion with rheumatoid arthritis. Isolated lytic and/or sclerotic bone lesions are not unusually seen in childhood and may lead to vertebral body collapse.
At presentation and during the induction phase of treatment, there are several life-threatening risks: Gram-negative or -positive septicaemia and fungaemia; bleeding (especially into the brain) and tumour lysis. Childhood early death rates have fallen from 9 to 10% in the 1980s to about 1% currently.
Diagnosis should be prompt and appropriate measures put in place to hydrate and support the patient whilst blast cell typing is taking place.
The French–American–British (FAB) staging classification into three subtypes—L1 (small monomorphic cells), L2 (large heterogeneous cells), and L3 (Burkitt-like cells with deeply basophilic cytoplasm and vacuoles)—is useful in separating out mature B-cell leukaemia, which requires quite different treatment, but is no longer used for stratification of other subtypes. Immunophenotyping, (increasingly using flow cytometry), cytogenetics, and gene expression arrays have become more important in defining subsets with variable prognostic and resistance profiles. In childhood about 75% of cases are of precursor B cell type (CD10, CD 19, HLA DR, and Tdt positivity); in adults about 50%. Coexpression of one or more myeloid markers (e.g. CD 13, CD 33, CDw 65) is not of prognostic significance and does not denote a true mixed lineage leukaemia. About 10 to 15% of all ALL cases express cytoplasmic immunoglobulin and are designated pre-B-ALL. T-cell phenotype (CD 7, CD 2 and/or CD 1 positivity) accounts for 11 to 12% of childhood cases and roughly double that in adults. Mature B-cell leukaemia requires a completely different lymphoma-type therapeutic approach from that used in precursor B, pre-B-, and T-ALL, all of which respond to sustained continuous therapy.
The greatest advance in subtype definition and prognostication has come from the definition of the genetic subtypes which vary in frequency between children and adults. These represent aberrant expression of proto-oncogenes, fusion genes resulting from chromosomal translocations (usually encoding for transcription factors or kinases) and chromosomal replication (e.g. high hyperdiploidy with >50 chromosomes). In childhood B-ALL, high hyperploidy is seen in about 25% of cases; t(12;21) (TEL-AML1 fusion gene) in 22%; both with favourable outcome on current therapy; and 3% with t(9;22) (BCR-ABL1); 8% with MLL rearrangements (especially t(4;11), t(11;19), or t(9;11) and 1% with hypodiploidy (<45 chromosomes); all with more adverse outcome. In adults the percentages are 7%, 2%, 25%, 10%, and 2%. The higher incidence of adverse prognostic groups goes some way to explain the poorer outcome in adult ALL. In T-ALL not only is the overall incidence higher in adults, but, in most, T-cell receptor genes are rearranged and there is an increased expression of certain HOX genes (50% NOTCHI). Given the relatively poor survival with standard treatment, especially in adults, the involvement of these genes is exciting interest as targets for novel therapeutic approaches. In the last few years a new subgroup with intrachromosomal amplification of chromosome 21 (iAMP21) which represents about 2% of ALL cases in childhood has been identified with an adverse prognosis, at least on standard therapy, but treatment intensification appears to have reduced relapse rates. How frequently this is seen in adults is as yet unknown.
The single most important prognostic factor is the highest white cell count recorded at presentation before introduction of any intervention. Those with a white blood cell count greater than 50 × 109/litre fare worse than those with lower counts. Patients under 1 year and over 10years of age fare worse. As already mentioned, part of the reason for older patients having a worse prognosis is the increasing frequency of adverse molecular rearrangements with associated resistance patterns, but reports also suggest poorer tolerance of therapy, decreased adherence to therapy especially following greater experienced toxicity, and less effective treatment regimens, traditionally used in adults partially because of worries about such toxicity. The impact of genetic factors has been alluded to for leukaemias involving BCR-ABL1, MLL gene rearrangements, and hypodiploidy, and all three present a special therapeutic challenge. For all these subtypes treatment has improved and individual prognostic factors have altered in their significance.
The key elements of successful management of ALL are good supportive care— induction of remission and consolidation of remission; treatment directed to the central nervous system; delayed intensification; and maintenance of remission with a total duration of therapy of 2 to 3 years for most patients. Treatment of any presenting infection, alkalinization of urine especially for high-risk patients, with the use of allopurinol or urate oxidase (to reduce the risks of hyperphosphataemia and hyperuricaemia respectively), are essential before treatment is started. Once diagnosis is established, transfusion of red cells and platelets for any mucosal haemorrhages or overt bleeding are important supportive measures. An initial platelet count less than 50 × 109/litre requires transfusion before the diagnostic lumbar puncture is carried out. The presence of blast cells in cerebrospinal fluid requires more aggressive intrathecal therapy with methotrexate and is one of the few current indications for cranial radiotherapy, especially in adult ALL.
The three most useful induction agents are steroids, vincristine, and L-asparaginase. Recent evidence suggests that dexamethasone is more effective than prednisolone (the traditional steroid used) especially because of its greater penetrance into brain tissue. Older patients appear to have greater steroid toxicity (especially avascular necrosis of weight-bearing joints, e.g. hips and knees) and either capping of dosage or intermittent high-dose courses are recommended for them. Traditionally steroids have been given for 4 weeks orally along with five to six doses of intravenous vincristine. L-Asparaginase is a crucial element of treatment but carries with it risks of overt anaphylaxis, silent neutralizing antibody production, coagulation disorders, and pancreatitis. The most effective formulation providing some protection against antibody production in induction is the peglyated form of Escherichia coli asparaginase, with a half life of 5.7 days. This means that fewer injections are required (two in induction rather than six to nine with native asparaginase) and more effective depletion achieved of the essential amino acid asparagine which appears essential for efficacy. The exact mechanism of action of >L-asparaginase is still in doubt. A percentage of patients (20–25%), mostly with high-risk ALL, have blasts secreting proteases which cleave asparaginase, producing both greater risk of inactivation and anaphylaxis. Dexamethasone and asparaginase have some synergy. For high or very high risk patients the addition of an anthracycline, doxorubicin or daunorubicin, is recommended. In childhood ALL remission rates of 93 to 96% are now achieved, but rates are slightly lower in adults. The speed of early response is used for stratification of patients. Traditionally this is measured by peripheral blood and marrow clearance, but in the modern era minimal residual disease monitoring, by molecular or flow cytometric measures, has been added. Those with slow early responses require greater intensification of therapy, and very slow responders or refractory patients appear to benefit from subsequent bone marrow transplantation. However in childhood the proportion of patients who require such therapy is ever decreasing. In those with BCR-ABL1, hypodiploid, and MLL gene rearrangement ALL it is a more likely requirement for cure. For those under 1 year of age with MLL gene rearrangements (most commonly t 4:11) a combination of ALL and AML-type therapy appears to be yielding better results.
Intrathecal methotrexate given two or three times during induction at 2-weekly intervals (weekly if overt blasts are present) followed by ongoing intrathecal therapy during the first year of treatment has replaced traditional cranial irradiation (18–24 Gy) in childhood because of the latter’s neurotoxicity and effects on growth. More effective delivery of steroids and asparaginase contributes to better control of disease in the central nervous system (failure rate now <2% in children and 5% in adults). There is more toxicity with triple intrathecal therapy (methotrexate, hydrocortisone, and cytosine arabinoside) without obvious benefit.
Post-remission consolidation and use of intensification pulses of sustained treatment principally with second-line drugs such as cytarabine and cyclophosphamide along with more steroids, vincristine, and Lasparaginase has improved long-term survival to 75 to 80% in children. For the reasons outlined before, results are still less favourable in teenagers and adults than in childhood. The reduction in use of pulsed therapy and adoption of ‘childhood’-type protocols has been shown to increase survival in younger adults and is now being tested in those up to 50 years. Control of lymphoblastic leukaemia appears to be improved with sustained, almost continuous immunosuppression, without marked myelosuppression. Prolonged neutropenia enhances the risk of sepsis and death while at the same time affording leukaemic blasts time to recover. The number of intensification modules required is currently being tested in randomized clinical trials.
In childhood ALL any attempt to shorten the length of continuing therapy (maintenance) to less than 2 years has decreased survival. Its effectiveness using oral methotrexate weekly and daily thiopurine, (usually 6-mercaptopurine), probably reflects influences on immune function and/or on bone marrow stroma cells rather than any ability to kill blast cells directly.
Truly refractory patients and those relapsing early on treatment or within 6 months of stopping therapy are considered for intensive therapy followed by allogeneic bone marrow transplantation (BMT). No clear benefit from autologous BMT has been demonstrated in patients with ALL. Effective cytoreductive therapy, prevention of anti-graft-vs-host and infection control have improved disease and event-free survival after marrow transplantation, but still the risks outweigh the benefits in most childhood and young adult ALL patients, unless they have true refractory disease. A combination of relapses after BMT and death from infection/or graft-vs-host diesease still result in 25 to 30% mortality; with matched related and unrelated donor BMT having similar outcomes in the modern era.
The remarkable success of what has essentially been empirical treatment has characterized much fruitful development, but we are left with the challenges of the unpredictable relapses among patients with ‘low risk’ features (numerically the largest group) and of very high risk ALL. Some of the ‘unpredictable’ relapses have been identified as being due to novel genetic rearrangements such as those occurring in the iAMP21 ALL group. For such patients intensive therapy and particularly the use of higher dose methotrexate and/or more effective asparaginase appears to have improved survival. Delivering each agent most effectively has improved disease control to its current high point, at least in childhood ALL. Can understanding of the genetic alterations leading to ALL also provide clues to novel ways of treatment?
Leukaemogenesis and novel therapies
It is now recognized that in most childhood ALL (and possibly AML) the first genetic events in key progenitor cells arise in utero but that one or more further events are usually required to convert a preleukaemic clone or clones into overt ALL. Response to infection may play a part in promoting development of ALL. Whether this sequence leads to ALL in older age is not yet known, but all events may occur postnatally. Those initial genetic changes alter cellular functioning, e.g. by facilitating unlimited capacity for self-renewal; altering the controls on proliferation; blocking partially or completely differentiation; and/or providing resistance to apoptosis. The BCR-ABL1 fusion protein seen in Philadelphia-positive ALL is a kinase that affects the signalling pathways which do indeed control cell proliferation, survival, and self-renewal. Identification of the kinase not only facilitates diagnosis of this form of ALL but an inhibitor of BCR-ABL1 tyrosine kinase has been produced (imatinib mesylate) which can inhibit blast growth and contribute to apoptosis. Alternative BCR-ABL1 kinase inhibitors are in trial. Good response rates (70%) in BCR-ABL1-positive ALL and in chronic myeloid leukaemia are observed with imatinib, but with time resistance can develop. Hence the search for alternative ways to inhibit the fusion product. Current trials are testing long-term efficacy of inhibition given in addition to standard ALL therapy.
Most of the genetic changes identified in ALL involve oncogenes rather than tumour suppressor malfunction. Such genetic rearrangement, due either to point mutations or, more frequently, translocation of proto-oncogenes to active promoter sites, leads to aberrant protein production.
The TEL-AML1 fusion gene seen in 20 to 25% of childhood ALL cases but only in a small minority of adult ALL combines the TEL gene involved in the homing of progenitors to the marrow and AML1, which is a DNA-binding component of the heterodimeric transcription factor core binding factor (CBFα plus CBFβ), with its crucial role in haematopoiesis.
The HOX genes, so frequently involved in T-ALL, appear to function downstream of the transcriptional cascade initiated by CBF. This pathway is also targeted by MLL fusion proteins, seen especially in infancy and in pre-B-ALL carrying the 1;19 chromosomal translocation (E2A-PBX1 rearrangement). The TEL-AML1 fusion protein inhibits the transcription normally activated when AML1 binds to a DNA region known as the core enhanced sequence. It recruits histone deacetylases which lead to closure of the chromatin structure and alter self-renewal and differentiation. Exploration of the use of small-molecule inhibitors of histone deacetylases has occurred usually in combination with cytotoxics. Similarly, investigation of other molecules along the HOX regulatory pathway as potential drug targets may yield a more targeted approach to therapy especially in T-ALL. Such work has focused on γ-secretase inhibition causing interference with NOTCH signalling. Another drug on trial in T-cell malignancies is nelarabine, a deoxyguanosine analogue.
As described earlier, the initiating genetic events in leukaemogenesis may require secondary changes to induce overt leukaemia. Primary changes which impair differentiation, e.g. TEL-AML1, require further mutations altering proliferation and survival of cells. For example, the commonest secondary change in TEL-AML1 cells is deletion of the other TEL gene.
Another example is the overexpression of FLT3, a receptor tyrosine kinase (role in development stem cells) seen both in ALL with MLL rearrangements and in high hyperdiploidy. Inhibitors of the kinase are now being tested in clinical trials (e.g. PKC412, MLN518, CEP701).
Very frequently in ALL the complex pathways involving the tumour suppressor retinoblastoma protein (RB), p130, p107, and Tp53 are altered by deletion or epigenetic silencing of P16ink 4a or P15ink 4b (especially in childhood T-ALL). As in many cancers, components of the Tp53 pathway are frequently mutated. Similarly, some lymphoid malignancies overexpress BCL2 with consequent blocking of apoptosis. All of these pathways are being investigated as potential drug targets, especially using small molecule inhibitors.
Sensitivity/resistance to chemotherapy
It has long been recognized that some subtypes of ALL have differential sensitivity/resistance to certain classes of cytotoxic drug. For example, hyperdiploid ALL appears to be very sensitive to antimetabolites (methotrexate especially). Such leukaemic cells accumulate high intracellular levels of methotrexate and its active polyglutamates following therapy. Most patients with high hyperdiploidy have three or more copies of chromosome 21 including the folate transporter gene for cellular influx of methotrexate. Effective therapy for such patients should include significant doses of systemic methotrexate. Only the blasts cells take up the methotrexate selectively. Conversely, T-ALL blasts appear to accumulate fewer active polyglutamates than B-lineage cells.
TEL-AML1 leukaemic blasts appear to be selectively sensitive to L-asparaginase; conversely, BCR-ABL1 and iAMP21 cells are frequently not. In these latter two it has been recently identified that they more commonly over express proteases especially asparaginyl endopeptidase which can cleave E. coli L-asparaginase, and inactivate the drug.
It has been reported that t(4;11)-ALL in infants and adults is preferentially sensitive to high-dose cytarbine. This may relate to increased secretion of hENT1, another cell membrane transporter. Genetic polymorphism of key metabolizing genes have been implicated in causation and response to therapy/resistance, e.g. NAD(P)H quinine oxidoreductase for causation and thiopurine methyl transferase for response.
Gene expression arrays have facilitated identification of unique leukaemia-associated markers which are then available as response identifiers; the arrays have also enabled identification of differentially expressed genes in drug-sensitive and drug-resistant ALL. Patterns of resistance to one or more of the key induction drugs (steroids, vincristine, daunorubicin and L-asparaginase) have been identified and strongly correlated with survival. This methodology has uncovered new potential therapeutic targets, e.g. in steroid-resistant ALL, overexpression of the anti-apoptotic MCL1 gene with subsequent attempts made to inhibit it using rapamycin. In addition, under-expression of several transcription genes was identified. The potential to modulate glycolytic pathways and increase steroid responsiveness has also emerged from such array studies. This research has emphasized the complexity of resistance mechanisms which formerly concentrated on studies of the multidrug-resistance gene product and its inhibition by calcium channel blocking agents.
Other potential targets
An imbalance of DNA methylation, involving widespread hypomethylation, regional hypermethylation, and increased cellular capacity for methylation, is characteristic of human neoplasia. Beginning in preneoplastic cells, methylation becomes extensive in subsequent stages of tumour progression. This aberrant methylation, particularly of cytosine, may mark abnormalities of chromatin reorganization and mediate progressive loss of gene expression associated with tumour development.
Abnormal methylation sites have been detected in lymphoid malignancies but to date we have little evidence of affective use of agents aimed at interfering with such events in ALL. The drug decitabine is being tested for potential efficacy.
Increasingly there is scientific interest in what supports survival of leukaemic cells within the marrow and in extramedullary sites and what facilitates migration of cells.
The growth of leukaemic cells requires the aggressive infiltration of capillaries into nests of tumour in excess of that seen in normal bone marrow. Relative selectivity of this neovascularization results in selective tumour regression on administration of antiangiogenic agents that block the effects of vascular endothelial cell growth factor (VEGF). Other drugs act to block oncoprotein function by inhibiting signal transduction, e.g. Ras farnesyltransferase inhibitors. Such factors might inhibit the neovascularization and supply of blood to new tumour growth if given at the time of maximal leukaemia-cell reduction. To date, the application of such agents in ALL has been limited. The stromal cells which support such nests of cells are of increasing research interest. It has been speculated that the effectiveness of ‘maintenance’ therapy may depend on its impact on the stromal cell component rather than directly on the leukaemia blasts.
Cell-surface-directed approaches are now in large clinical trials with some encouraging responses. Constructs aimed directly at receptors including monoclonal antibodies specific for the cell-surface receptors CD20, CD33, CD52, and CD22 linked with antibiotics or endotoxins are being tested. Monoclonal antibodies have achieved utility in the treatment of lymphoma particularly with the use of rituximab, an anti-CD20 chimeric murine–human monoclonal antibody.