Childhood Leukemia — New Advances and Challenges
http://www.100md.com
《新英格兰医药杂志》
Since 1970, the rate of cure of acute lymphoblastic leukemia (ALL) in children has increased dramatically, from less than 30 percent to approximately 80 percent. This remarkable improvement has resulted from the marriage of laboratory and clinical science. The identification of effective agents in randomized cooperative-group studies, the application of treatment to the central nervous system before the onset of symptoms, the intensification of treatment, and the use of "risk-adapted therapy" (therapy tailored to the predicted risk of relapse) have led to today's impressive cure rates.
Despite these successes, however, much work remains. Many of the children with ALL who are cured by current therapies are overtreated and thus unnecessarily exposed to the risk of short- and long-term adverse effects. Even among the subgroups with the most favorable prognostic factors, 10 to 20 percent will have a relapse and most of these children will ultimately die of their disease. Subgroups of children with ALL, including infants, those with unfavorable genotypes such as Philadelphia chromosome–positive ALL, and those who do not have a complete remission initially or who relapse after a remission, still have an extremely poor prognosis. Further intensification of existing therapies is unlikely to improve the cure rate substantially in these populations. Therefore, the identification of additional prognostic variables that can be used to tailor therapy more precisely and the discovery of drugs that can modify pathways involved in transformation and resistance to therapy are top priorities.
A risk-adapted stratification scheme emerged from the identification of clinical and laboratory variables unique to the host or leukemic blast that are predictive of the outcome. Age and white-cell count at diagnosis, sex, and immunophenotypic and cytogenetic differences among lymphoblast populations can be used to tailor therapy so that the chances of cure are maximized while exposure to unnecessary risk is minimized.1 The prognostic value of an early response, as determined by a morphologic assessment of blast clearance or by means of sensitive methods to detect minimal residual disease, including flow cytometry and polymerase-chain-reaction techniques, has been used as the basis for risk stratification.2,3,4 Clinical trials have shown that patients with a "slow early response" may benefit from early application of intensified therapy, thus helping a subgroup of patients who would otherwise be destined to have a poor outcome.5
In only a few instances has there been a mechanistic understanding of why certain variables correlate with the outcome. For example, a T-cell immunophenotype, present in about 15 percent of children with ALL, can be used to identify those who fare poorly with early conventional therapies. The observation, in a review by Gorlick et al.,5 that T lymphoblasts have a lower level of expression of folylpolyglutamate synthetase and, therefore, less polyglutamate methotrexate than B-lineage blasts prompted the randomized assessment of high-dose methotrexate in children with T-cell ALL, which significantly improved the outcome.6 The finding that B-lineage blasts with more than 51 chromosomes or the TEL-AML1 gene translocation are more sensitive to the cytotoxic effects of asparagine depletion in vitro than are other forms of leukemia7,8,9 prompted an upcoming randomized trial of the effect of additional doses of asparaginase on the outcome among children with these cytogenetic features.
These well-tested prognostic factors must now be considered in the context of the emerging field of high-throughput genomics and proteomics, which allows investigators to determine simultaneously the levels of expression of tens of thousands of genes and proteins.1 Microarray, or DNA-chip, technology has already been used successfully to differentiate known subgroups of leukemias and to identify previously unrecognized and prognostically significant subgroups among leukemias with indistinguishable morphologic characteristics or karyotypes.10,11 Holleman et al.12 have used this powerful technique to determine gene-expression patterns that predict the outcome of ALL, and they report their results in this issue of the Journal. Members of this group had previously demonstrated that the in vitro sensitivity of ALL blasts to the chemotherapeutic agents used to treat childhood ALL during the induction period predicted the outcome.8 Using supervised clustering analysis, based on in vitro drug sensitivity, they were able to identify several sets of genes whose expression profiles were closely linked to the outcome. Though the ability of the expression signatures to classify samples as sensitive or resistant to a particular agent was moderate (approximately 80 percent), Holleman and colleagues were able to use the signatures in aggregate to predict the outcome in two cohorts of children treated with different regimens. This finding indicates that the observed gene-expression profiles represent fundamental biochemical features, not epiphenomena, and suggests that gene-expression profiles could be used to alter therapy instead of the more cumbersome method of in vitro sensitivity testing.
A critical challenge comes in integrating these new data with current methods for assigning risk-directed therapy. It is uncertain whether the gene-expression–signature approach used by the authors will prove to be superior to current risk-based systems predicated on clinical data, genotype, and early response. As Holleman et al. state, the genes predictive of a response to individual agents did not overlap. This finding suggests that critical defects in common or more terminal drug-induced apoptotic pathways are not crucial mediators of resistance; the problem may lie in more proximal pathways unique to each drug. Thus, as with other data from microarrays, there is no single gene whose expression accurately predicts clinical characteristics or the outcome of ALL, emphasizing what clinicians have always known — cancer is a complex disease and needs to be attacked on many fronts.
If patients are found to have a gene-expression profile that is predictive of a poor response to one or several chemotherapeutic agents, should those agents be omitted from their treatment? Not yet. The subgroup of patients with ALL classified as "highly resistant" by Holleman and colleagues still had a disease-free survival rate of over 60 percent. Furthermore, there are few proven alternative therapies. The gene-expression patterns these authors describe can be used to begin to define mechanisms of resistance and will stimulate the development of alternative treatment strategies, targeted to those with resistant disease identified at diagnosis. The identification of a gene-expression profile that predicts the outcome of treatment on the basis of sensitivity to drugs currently used universally to treat children with ALL creates a unique opportunity to modulate therapy early on. It would potentially avoid the administration of drugs to which a given leukemia is resistant and maximize the use of drugs to which a clone is sensitive. Most important, this discovery provides the basis for the creation of critically important molecules that will modulate the expression of previously unknown proteins whose actions render malignant lymphoblasts immune to currently available best therapies.
The treatment of childhood ALL has served as a paradigm for the treatment of other cancers since Farber first demonstrated, more than 50 years ago, that a pharmaceutical agent could induce a clinical remission. The past five decades of progress in the treatment of childhood ALL have shown that clinical trials, largely empirically based, can dramatically improve the outcome of treatment. Although the number of lives saved through advances in the treatment of childhood ALL may seem small in comparison to the number lost as a result of the common cancers of adulthood, the savings in terms of the cumulative years of life gained are enormous. Work such as that of Holleman and colleagues will lead to the creation of treatment protocols that are based on an understanding of the underlying biology of childhood ALL and will thus continue the tradition started by Farber and other pioneers. The obvious and certain goal is to increase the current cure rate of approximately 80 percent to the previously unthinkable rate of 100 percent.
Source Information
From the University of Texas Southwestern Medical Center, Dallas (N.J.W.); New York University School of Medicine and Mount Sinai School of Medicine, New York (W.L.C.); and the University of Florida College of Medicine, Gainesville (S.P.H.).
References
Carroll WL, Bhojwani D, Min D-J, et al. Pediatric acute lymphoblastic leukemia. Hematology (Am Soc Hematol Educ Program) 2003;1:102-131.
van Dongen JJ, Seriu T, Panzer-Grumayer ER, et al. Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood. Lancet 1998;352:1731-1738.
Donadieu J, Hill C. Early response to chemotherapy as a prognostic factor in childhood acute lymphoblastic leukaemia: a methodological review. Br J Haematol 2001;115:34-45.
Coustan-Smith E, Sancho J, Behm FG, et al. Prognostic importance of measuring early clearance of leukemic cells by flow cytometry in childhood acute lymphoblastic leukemia. Blood 2002;100:52-58.
Gorlick R, Goker E, Trippet T, Waltham M, Banerjee D, Bertino JR. Intrinsic and acquired resistance to methotrexate in acute leukemias. N Engl J Med 1996;335:1041-1048.
Asselin BL, Shuster JJ, Amylon M, et al. Improved event-free survival (EFS) with high dose methotrexate (HDM) in T-cell lymphoblastic leukemia (T-ALL) and advanced lymphoblastic lymphoma (T-NHL): a Pediatric Oncology Group (POG) study. Prog Proc Am Soc Clin Oncol 2001;20:367a. abstract.
Kaspers GJ, Smets LA, Pieters R, Van Zantwijk CH, Van Wering ER, Veerman AJ. Favorable prognosis of hyperdiploid common acute lymphoblastic leukemia may be explained by sensitivity to antimetabolites and other drugs: results of an in vitro study. Blood 1995;85:751-756.
Kaspers GJ, Veerman AJ, Pieters R, et al. In vitro cellular drug resistance and prognosis in newly diagnosed childhood acute lymphoblastic leukemia. Blood 1997;90:2723-2729.
Ramakers-van Woerden NL, Pieters R, Loonen AH, et al. TEL/AML1 gene fusion is related to in vitro drug sensitivity for L-asparaginase in childhood acute lymphoblastic leukemia. Blood 2000;96:1094-1099.
Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999;286:531-537.
Yeoh EJ, Ross ME, Shurtleff SA, et al. Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling. Cancer Cell 2002;1:133-143.
Holleman A, Cheok MH, den Boer ML, et al. Gene-expression patterns in drug-resistant acute lymphoblastic leukemia cells and response to treatment. N Engl J Med 2004;351:533-542.(Naomi J. Winick, M.D., Wi)
Despite these successes, however, much work remains. Many of the children with ALL who are cured by current therapies are overtreated and thus unnecessarily exposed to the risk of short- and long-term adverse effects. Even among the subgroups with the most favorable prognostic factors, 10 to 20 percent will have a relapse and most of these children will ultimately die of their disease. Subgroups of children with ALL, including infants, those with unfavorable genotypes such as Philadelphia chromosome–positive ALL, and those who do not have a complete remission initially or who relapse after a remission, still have an extremely poor prognosis. Further intensification of existing therapies is unlikely to improve the cure rate substantially in these populations. Therefore, the identification of additional prognostic variables that can be used to tailor therapy more precisely and the discovery of drugs that can modify pathways involved in transformation and resistance to therapy are top priorities.
A risk-adapted stratification scheme emerged from the identification of clinical and laboratory variables unique to the host or leukemic blast that are predictive of the outcome. Age and white-cell count at diagnosis, sex, and immunophenotypic and cytogenetic differences among lymphoblast populations can be used to tailor therapy so that the chances of cure are maximized while exposure to unnecessary risk is minimized.1 The prognostic value of an early response, as determined by a morphologic assessment of blast clearance or by means of sensitive methods to detect minimal residual disease, including flow cytometry and polymerase-chain-reaction techniques, has been used as the basis for risk stratification.2,3,4 Clinical trials have shown that patients with a "slow early response" may benefit from early application of intensified therapy, thus helping a subgroup of patients who would otherwise be destined to have a poor outcome.5
In only a few instances has there been a mechanistic understanding of why certain variables correlate with the outcome. For example, a T-cell immunophenotype, present in about 15 percent of children with ALL, can be used to identify those who fare poorly with early conventional therapies. The observation, in a review by Gorlick et al.,5 that T lymphoblasts have a lower level of expression of folylpolyglutamate synthetase and, therefore, less polyglutamate methotrexate than B-lineage blasts prompted the randomized assessment of high-dose methotrexate in children with T-cell ALL, which significantly improved the outcome.6 The finding that B-lineage blasts with more than 51 chromosomes or the TEL-AML1 gene translocation are more sensitive to the cytotoxic effects of asparagine depletion in vitro than are other forms of leukemia7,8,9 prompted an upcoming randomized trial of the effect of additional doses of asparaginase on the outcome among children with these cytogenetic features.
These well-tested prognostic factors must now be considered in the context of the emerging field of high-throughput genomics and proteomics, which allows investigators to determine simultaneously the levels of expression of tens of thousands of genes and proteins.1 Microarray, or DNA-chip, technology has already been used successfully to differentiate known subgroups of leukemias and to identify previously unrecognized and prognostically significant subgroups among leukemias with indistinguishable morphologic characteristics or karyotypes.10,11 Holleman et al.12 have used this powerful technique to determine gene-expression patterns that predict the outcome of ALL, and they report their results in this issue of the Journal. Members of this group had previously demonstrated that the in vitro sensitivity of ALL blasts to the chemotherapeutic agents used to treat childhood ALL during the induction period predicted the outcome.8 Using supervised clustering analysis, based on in vitro drug sensitivity, they were able to identify several sets of genes whose expression profiles were closely linked to the outcome. Though the ability of the expression signatures to classify samples as sensitive or resistant to a particular agent was moderate (approximately 80 percent), Holleman and colleagues were able to use the signatures in aggregate to predict the outcome in two cohorts of children treated with different regimens. This finding indicates that the observed gene-expression profiles represent fundamental biochemical features, not epiphenomena, and suggests that gene-expression profiles could be used to alter therapy instead of the more cumbersome method of in vitro sensitivity testing.
A critical challenge comes in integrating these new data with current methods for assigning risk-directed therapy. It is uncertain whether the gene-expression–signature approach used by the authors will prove to be superior to current risk-based systems predicated on clinical data, genotype, and early response. As Holleman et al. state, the genes predictive of a response to individual agents did not overlap. This finding suggests that critical defects in common or more terminal drug-induced apoptotic pathways are not crucial mediators of resistance; the problem may lie in more proximal pathways unique to each drug. Thus, as with other data from microarrays, there is no single gene whose expression accurately predicts clinical characteristics or the outcome of ALL, emphasizing what clinicians have always known — cancer is a complex disease and needs to be attacked on many fronts.
If patients are found to have a gene-expression profile that is predictive of a poor response to one or several chemotherapeutic agents, should those agents be omitted from their treatment? Not yet. The subgroup of patients with ALL classified as "highly resistant" by Holleman and colleagues still had a disease-free survival rate of over 60 percent. Furthermore, there are few proven alternative therapies. The gene-expression patterns these authors describe can be used to begin to define mechanisms of resistance and will stimulate the development of alternative treatment strategies, targeted to those with resistant disease identified at diagnosis. The identification of a gene-expression profile that predicts the outcome of treatment on the basis of sensitivity to drugs currently used universally to treat children with ALL creates a unique opportunity to modulate therapy early on. It would potentially avoid the administration of drugs to which a given leukemia is resistant and maximize the use of drugs to which a clone is sensitive. Most important, this discovery provides the basis for the creation of critically important molecules that will modulate the expression of previously unknown proteins whose actions render malignant lymphoblasts immune to currently available best therapies.
The treatment of childhood ALL has served as a paradigm for the treatment of other cancers since Farber first demonstrated, more than 50 years ago, that a pharmaceutical agent could induce a clinical remission. The past five decades of progress in the treatment of childhood ALL have shown that clinical trials, largely empirically based, can dramatically improve the outcome of treatment. Although the number of lives saved through advances in the treatment of childhood ALL may seem small in comparison to the number lost as a result of the common cancers of adulthood, the savings in terms of the cumulative years of life gained are enormous. Work such as that of Holleman and colleagues will lead to the creation of treatment protocols that are based on an understanding of the underlying biology of childhood ALL and will thus continue the tradition started by Farber and other pioneers. The obvious and certain goal is to increase the current cure rate of approximately 80 percent to the previously unthinkable rate of 100 percent.
Source Information
From the University of Texas Southwestern Medical Center, Dallas (N.J.W.); New York University School of Medicine and Mount Sinai School of Medicine, New York (W.L.C.); and the University of Florida College of Medicine, Gainesville (S.P.H.).
References
Carroll WL, Bhojwani D, Min D-J, et al. Pediatric acute lymphoblastic leukemia. Hematology (Am Soc Hematol Educ Program) 2003;1:102-131.
van Dongen JJ, Seriu T, Panzer-Grumayer ER, et al. Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood. Lancet 1998;352:1731-1738.
Donadieu J, Hill C. Early response to chemotherapy as a prognostic factor in childhood acute lymphoblastic leukaemia: a methodological review. Br J Haematol 2001;115:34-45.
Coustan-Smith E, Sancho J, Behm FG, et al. Prognostic importance of measuring early clearance of leukemic cells by flow cytometry in childhood acute lymphoblastic leukemia. Blood 2002;100:52-58.
Gorlick R, Goker E, Trippet T, Waltham M, Banerjee D, Bertino JR. Intrinsic and acquired resistance to methotrexate in acute leukemias. N Engl J Med 1996;335:1041-1048.
Asselin BL, Shuster JJ, Amylon M, et al. Improved event-free survival (EFS) with high dose methotrexate (HDM) in T-cell lymphoblastic leukemia (T-ALL) and advanced lymphoblastic lymphoma (T-NHL): a Pediatric Oncology Group (POG) study. Prog Proc Am Soc Clin Oncol 2001;20:367a. abstract.
Kaspers GJ, Smets LA, Pieters R, Van Zantwijk CH, Van Wering ER, Veerman AJ. Favorable prognosis of hyperdiploid common acute lymphoblastic leukemia may be explained by sensitivity to antimetabolites and other drugs: results of an in vitro study. Blood 1995;85:751-756.
Kaspers GJ, Veerman AJ, Pieters R, et al. In vitro cellular drug resistance and prognosis in newly diagnosed childhood acute lymphoblastic leukemia. Blood 1997;90:2723-2729.
Ramakers-van Woerden NL, Pieters R, Loonen AH, et al. TEL/AML1 gene fusion is related to in vitro drug sensitivity for L-asparaginase in childhood acute lymphoblastic leukemia. Blood 2000;96:1094-1099.
Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999;286:531-537.
Yeoh EJ, Ross ME, Shurtleff SA, et al. Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling. Cancer Cell 2002;1:133-143.
Holleman A, Cheok MH, den Boer ML, et al. Gene-expression patterns in drug-resistant acute lymphoblastic leukemia cells and response to treatment. N Engl J Med 2004;351:533-542.(Naomi J. Winick, M.D., Wi)