Leukemia (2003) 17, 2358–2382 & 2003 Nature Publishing Group All rights reserved 0887-6924/03 $25.00 www.nature.com/leu

REVIEW

Treatment by design in leukemia, a meeting report, Philadelphia, Pennsylvania, December 2002 RA Larson1, GQ Daley2, CA Schiffer3, P Porcu4, C-H Pui5, J-P Marie6, LS Steelman7, FE Bertrand7 and JA McCubrey7,8 1Section of Hematology/Oncology, University of Chicago Pritzker School of Medicine, Chicago, IL, USA; 2Whitehead Institute for Biomedical Research and Harvard Medical School, Cambridge, MA, USA; 3Wayne State University School of Medicine, Karmanos Cancer Institute, Detroit, MI, USA; 4Division of Hematology Oncology, Ohio State University College of Medicine and Public Health, Columbus, OH, USA; 5St Jude Children’s Research Hospital and the University of Tennessee Health Science Center, Memphis, TN, USA; 6Onco-Hematology Department, Hoˆtel-Dieu de Paris, Paris, AP-HP, France; 7Department of Microbiology and Immunology, The Brody School of Medicine at East Carolina University, Greenville, NC, USA; and 8Leo Jenkins Cancer Center, The Brody School of Medicine at East Carolina University, Greenville, NC, USA

Novel approaches have been designed to treat leukemia based disseminated cancer was ushered in approximately 50 years on our understanding of the genetic and biochemical lesions ago, when investigators recognized that inhibiting folate present in different malignancies. This meeting report sum- marizes some of the recent advances in leukemia treatment. metabolism led to the death of childhood leukemia cells. Based on the discoveries of cellular oncogenes, chromosomal Methotrexate, which inhibits dihydrofolate reductase and other translocations, monoclonal antibodies, multidrug resistance enzymes, is still a cornerstone of treatment for acute lympho- pumps, signal transduction pathways, genomics/proteonomic blastic leukemia (ALL), although it has a minimal role in acute approaches to clinical diagnosis and mutations in biochemical myeloid leukemia (AML). However, methotrexate is not pathways, clinicians and basic scientists have been able to selective. It inhibits the target enzymes in both normal and identify the particular genetic mutations and signal transduc- tion pathways involved as well as design more appropriate malignant cells. treatments for the leukemia patient. This meeting report This symposium reviewed a number of pathways that are discusses these exciting new therapies and the results known to be important in the regulation of cell proliferation, obtained from ongoing clinical trials. Furthermore, rational apoptosis, drug resistance and metabolism in acute and chronic approaches to treat complications of tumor lysis syndrome by leukemia. Promising new anticancer drugs and the current status administration of the recombinant urate oxidase protein, also of clinical trials of investigational agents that have been known as rasburicase, which corrects the biochemical defect present in humans, were discussed. Clearly, over the past 25 designed to specifically inhibit or alter these pathways and years, molecular biology and biotechnology has provided the interact with leukemia-specific targets were discussed at this hematologist/oncologist novel bullets in their arsenal that will meeting. allow treatment by design in leukemia. This meeting report covers the role and requirement of the Leukemia (2003) 17, 2358–2382. doi:10.1038/sj.leu.2403156 BCR-ABL oncoprotein in chronic myelogenous leukemia (CML), Published online 9 October 2003 the transition from the chronic to the blast crisis stage and the Keywords: leukemia therapy; signal transduction inhibitors; chromosomal translocations; multidrug resistance; monoclonal permissive role for additional chromosomal translocations antibodies; rasburicase; elitekt; tumor lysis syndrome involving transcription factor genes and the roles of FLT-3 and RAS mutations in leukemia. The treatment of CML and other leukemias with Gleevect, and FLT-3 and RAS inhibitors is also Introduction and overview of meeting: Richard A Larson, MD discussed. Certain chromosomal translocations involve the expression of novel transcription factors that can exert dominant Several decades of prospective clinical trials have allowed the negative (DN) effects on gene expression by silencing transcrip- identification of important determinants of response for patients tion through chromatin remodeling by tethering histone with leukemia that correlate with both treatment and survival. decacetylases (HDAC). Treatment of certain leukemias with These have included age, the presence of comorbid illnesses, retinoic acids and HDAC inhibitors may prove efficacious. and the number of white blood cells (WBC) at time of diagnosis, A frequent consequence observed in leukemia treatment with and precise morphological diagnoses. Recently, more sophisti- common anticancer drugs is drug resistance. Hematopoietic cated techniques have been used to examine the biological precursor cells may be intrinsically resistant to many drugs since characteristics of the leukemia itself including immunopheno- they were ‘designed’ to survive repeated exposure to hetero- typing, karyotyping and gene expression. This knowledge has geneous toxins and natural products. Drug resistance can arise led to risk-adapted treatment plans in which the use of particular by many different mechanisms; however, a common mechan- cytotoxic drugs and supportive care measures have been ism is increase in expression of a family of proteins called drug designed for individual patients. Empiricism is gradually being transporters (P-glycoprotein (P-gp), MRP-1, breast cancer replaced by individualized therapy. resistance protein (BCRP) and others). These transmembrane Targeted therapy did not start with STI571, now known as proteins contain an ATP-binding cassette (ABC) and are imatinib mesylate or Gleevect. The possibility of curing associated with energy-dependent efflux of natural products. Many antileukemia drugs are substrates for multidrug resistance Correspondence: JA McCubrey, Department of Microbiology and (MDR)-mediated efflux. P-gp, the protein product of the MDR-1 Immunology, The Brody School of Medicine at East Carolina gene, is frequently overexpressed in stem cells. P-gp over- University, Brody Building 5N98C, Greenville, NC 27858, USA; Fax: þ 1 252 744 3104 expression in AML is associated with a poorer prognosis. A Received 1 August 2003; accepted 26 August 2003; Published online means to circumvent MDR is the use of MDR inhibitors. 9 October 2003 Numerous MDR inhibitors have been developed and their Novel approaches to treat leukemia RA Larson et al 2359 effects are being evaluated in clinical trials (cyclosporine, PSC- Promising targets in leukemia: translating molecular 833, LY335979, VX710, XR9575, R101933, PK11195 and mechanisms into rational therapies: George Q Daley, MD GG120918). The commonly used immunosuppressive drug cyclosporine is known to reverse the effects of P-gp. Dr Daley (Whitehead Institute and Harvard Medical School, Chronic lymphocytic leukemia (CLL) is the most frequent Cambridge, MA, USA) discussed some of the scientific basis for leukemia observed in the Western Hemisphere. Historically, rational drug design in leukemia. The dramatic success of CLL has been treated with alkylating drugs. Now, however, Gleevec has galvanized the drug development community fludarabine is the best single agent for initial therapy. Chimeric toward the identification of further such agents. This develop- and humanized monoclonal antibodies (mAbs) represent novel ment has been facilitated by insights into the molecular therapeutic options for CLL. The humanized mAb targeting mechanisms of cancer. s s CD52 (CAMPATH ), a chimeric aCD20 Ab (Rituxan ), an mAb There is an ongoing debate as to whether Gleevec will prove targeting CD23 (IDEC 152) and a humanized mAb targeting an to be a rare example of great success or whether the principles epitope of the HLA-DR b chain (Hu1D10, Remitogent) are in responsible for Gleevec’s usefulness can be generalized to other various stages of clinical trials in CLL patients. These modified tumor types. It is likely that our understanding of the rather antibodies may exert their cytotoxic effects by inducing limited set of pathways in cancer cells that lead to transforma- apoptosis. tion can in many instances be targeted in the next decade or ALL can be cured in approximately 80% of children and 30– two. 40% of adults. While up to 75% of ALL patients show How many mutations are required to transform a normal cell identifiable genetic abnormalities associated with prognostic into a malignant cell? This fundamental issue has been studied and therapeutic relevance, there remains considerable genetic extensively in solid tumors by Vogelstein and colleagues at heterogeneity, which can complicate effective treatment. Some Johns Hopkins University (Baltimore, MD, USA). By studying high-risk ALL patients with certain chromosomal translocations the occurrence of various mutations in colon cancers, Vogel- (eg, BCR-ABL and MLL-AF4) can be effectively treated with stein has derived a cancer model referred to by some as a allogeneic hematopoietic stem cell transplantation. The ability ‘Vogelgram’, which assigns a sequence of stepwise mutations to classify ALL by genomic/proteomic approaches represents a across the spectrum of colon cancer. It is estimated that at least significant advance in our understanding of the disease and will four independent mutations must cooperate in the creation of a eventually provide the clinician with more precise therapeutic full-blown colonic malignancy.1,2 A diagram representing this options for the patient. concept is presented in Figure 1. Tumor lysis syndrome (TLS) is a serious and potentially life- Weinberg and colleagues at MIT modeled that work in threatening complication of cancer chemotherapy, especially in primary cells,3 and recently it has become clear that at least rapidly proliferating malignancies such as Burkitt’s lymphoma. three or four genetic mutations are required for the malignant TLS involves an increase in the levels of uric acid, potassium, transformation of primary cells. In initial rodent studies in the phosphorous and other intracellular components due to the lysis 1980s, it was shown that mutations at only two oncogenes myc of tumor cells. This increase in uric acid and other metabolites plus ras were needed to generate a tumor cell.4 However, when can lead to renal impairment primarily due to the insolubility of those studies were attempted in human primary tissue, no uric acid in the urine. Humans lack the ability to metabolize uric tumors arose. It was not until the isolation of telomerase and the acid to allantoin, which is five times more soluble than uric identification of its role in cellular immortalization did acid, due to a mutation in the urate oxidase gene. For the past 30 malignant transformation experiments in human cells start to years, urate oxidase has been purified from Aspergillus flavus work.5 and used to treat Burkitt’s lymphoma, ALL and AML patients in Now it is clear that in primary human tissues, several France and Italy. Rasburicase is a recombinant urate oxidase. collaborating oncogenes are required for malignant transforma- Rasburicase is effective in preventing and treating hyperurice- tion. These genes fall into general categories, biological path- mia. Thus, this meeting covered various aspects of how basic ways or functional pathways that need to be deranged in all scientific discoveries over the past 30 years have been applied tumor cells in order for full tumorigenic phenotype to be clinically to improve significantly treatment of leukemia realized. Immortalization and prevention of apoptosis need to patients. be overcome. Genetic lesions that drive cell proliferation are

Figure 1 Multiple mutations and time are required for the transformation of a malignant colon cancer cell.

Leukemia Novel approaches to treat leukemia RA Larson et al 2360 necessary, and ras pathway mutations are common. Also, the It took a decade, however, to make this process of retroviral problem of cellular senescence must be overcome and that is gene transfer and CML development more efficient, which the function of the telomerase gene product. confirmed that BCR-ABL, acting as a single agent, was in fact Standing in contrast to solid tumors, CML is an unusual responsible for CML.9,10 It is known that every hematopoietic disease that is classified as a myeloproliferative disorder. It is a stem cell that gets infected by the BCR-ABL virus will contribute weak malignancy, which some physicians consider a hyperpla- to the disease, and under experimental conditions in mice, a sia. The myeloproliferative phase of CML is associated with a polyclonal disorder is observed. This suggests that at least in consistent chromosomal abnormality recognized as a balanced mice, the BCR/ABL oncoprotein alone can drive CML (Figure 2). translocation between distal elements of chromosomes 9 and This concept has now been generalized. There are many 22.6 The human homologue of the Abelson murine leukemia different chromosomal translocations that have been identified virus oncogene, c-abl, is located on chromosome 9 and gets and characterized in the context of these hyperplasia-like translocated and juxtaposed to a breakpoint cluster region on myeloproliferative conditions. The translocation of 9 and 22, chromosome 22. This results in the Philadelphia chromosome which juxtaposes BCR-ABL, is the best characterized and the (Ph), which encodes among other things the BCR-ABL onco- most common, but it is now known that there are chromosomal proteins.7 translocations such as fusion of PDGF receptor and other genes Over 10 years ago, Daley and colleagues at MIT inserted the such as TEL.11 TEL also fuses with ABL, and to other kinases like BCR-ABL oncogene into a mouse retrovirus.8 Bone marrow cells JAK.11 In general, the phenomenon seems to be that an activated of a mouse were infected with the recombinant retrovirus and a kinase is responsible for driving the myeloproliferative pheno- bone marrow transplant was performed. A CML-like myelopro- type. The global sites where these and other chromosomal liferative condition in the recipient mouse was observed. This translocations may act in the cell are presented in Figure 3. led to the conclusion that BCR-ABL was indeed the causative Genetic mutations that destroy the kinase activity of BCR-ABL genetic lesion driving CML. abrogate all transforming functions of this oncoprotein. This

Figure 2 Steps in transformation of human hematopoietic cells.

Figure 3 Conversion from chronic to acute myelogenous leukemia requires at least two mutations.

Leukemia Novel approaches to treat leukemia RA Larson et al 2361 structure/function information provided the rationale for the repressor complex includes nuclear co-repressor proteins and development of the small molecule inhibitor Gleevec, which is principally the histone deacetylase complex (HDAC) that is a highly potent and specific inhibitor of the kinase activity responsible for modulating chromatin, leading to gene silen- developed by Novartis. If CML patients are treated with Gleevec cing.22–24 immediately after diagnosis in the chronic phase, greater than This repression can be overcome by the addition of 98% will achieve a complete normalization of the WBC retinoic acid (RA), which releases the repressor complex, counts.12–14 The vast majority of patients achieve a major and recruits an activator complex containing P300.25,26 RA cytogenetic response, that is, virtually complete suppression of acts as a potent differentiation agent in various hematopoietic evidence of the Ph in their cells. Clearly this is a dramatic result. cell types.27 The chronic phase of CML is unstable, and over time, these The PML-RARa fusion protein encoded by the (15;17) patients, left untreated, will progress to a malignancy typical of translocation in APL binds the promoter and acts as DN acute leukemia. This is called blast crisis. Blast crisis is the inhibitors of transcription through stable transcriptional repres- consequence of additional genetic abnormalities. This has been sion.28 The repressor complex binds to the promoter, blocking known at the cytogenetic level for many years, where patients gene transcription and blocking expression of key genes who evolve blast crisis have an extra Ph, abnormalities of involved in hematopoietic cell differentiation. This repressor chromosome 17, trisomy 8, and a whole host of other complex is refractory to normal physiologic doses of RA. superimposed genetic events. Pharmacologic doses of RA release the repressor complex, It thus appears that AML is at least a two-hit disease. This can recruit the co-activator complex and restore the differentiation be modeled in experimental murine systems, and supporting of blood cells.29 A diagram of the PML-RARa complex and the evidence has been gathered by cytogenetic analysis of clinical effects on gene transcription by RA and histone deacetylases is specimens. Patients have been described who present initially presented in Figure 4. with the chronic phase of CML [t(9;22) bearing], and later This admittedly oversimplified model has nonetheless stimu- evolve into blast crisis with leukemia cells bearing secondary lated the idea that a target for this entire class of translocations is genetic lesions such as the activation of homeobox genes or the transcriptional repressor complex. The most attractive target core-binding factor (CBF). This has also been seen in patients of this complex is the HDAC enzyme. Pharmacologic inhibition with myeloproliferative syndrome (MDS) characterized by the of HDAC has been shown to relieve the repressor complex, (5;12) translocation that evolved into AML with an additional recruit the co-activator and restore hematopoietic differentia- (8;21) translocation. tion.30,31 A concept promulgated by Dr Gary Gilliland (Harvard Recently, many HDAC inhibitors have been developed with University, Boston, MA, USA) is that myeloproliferative condi- micromolar potencies and introduced into clinical trials.32,33 In tions seem to evolve from activated kinases, whereas the the laboratory, these have been shown to induce differentiation progression to acute leukemia requires a block in differentiation of a variety of hematopoietic cells, and there is encouraging that is typically seen in the context of translocations, which alter clinical data in a growing variety of solid tumors as well as transcription factors.15 This concept has also been modeled in leukemias. experimental systems. Gilliland et al have shown that BCR-ABL Another emerging leukemia target is FLT3.34 FLT3 is a type 3/ alone can induce CML. But if murine bone marrow cells are class 3-receptor tyrosine kinase receptor that binds the growth infected with a combination of viruses, one encoding BCR-ABL factor FLT3 ligand, an important cytokine for the maintenance of and the other carrying an altered transcription factor, AML is hematopoietic stem cells. Naoe and colleagues have deter- immediately observed in the transplanted mice.16 mined that about a quarter of AML patients has an internal Gleevec is highly effective in chronic phase, perhaps because tandem duplication of a juxtamembrane region of the receptor, the chronic phase involves only a single hit. However, in blast which leads to its constitutive activation.35–38 Moreover, crisis patients, Gleevec results in initial responses in virtually all another cohort of AML patients harbor point mutations that patients, but a rapid relapse and resistance follows. Much of this involve the activation loop, a critical regulatory region around resistance is due to mutations in the target BCR-ABL kinase12,17 the enzymatic cleft.39 Therefore, almost a third of all patients indicating that the BCR-ABL remains an important target in this with AML carry mutations in the FLT3 receptor in their leukemia disease. Perhaps other drugs can be identified to target the cells. mutant Gleevec-resistant forms of the BCR-ABL oncoprotein. It A number of FLT3 inhibitors are being evaluated in clinical is also likely that there are other pathways activated in blast trails that have previously been shown to be active against crisis. Therefore, the ultimate treatment will need target-directed transformed cell lines and in mouse models.40,41 The clinical agents against both BCR-ABL and the other activated pathways. activity of these agents is among the most exciting new Identification of these addition targets is crucial for complete observations in target-directed treatment of AML. CML therapy. Another excellent pathway for targeting in leukemia is the Ras Studies in AML have indicated the recurrent activation of or pathway. Ras is one of the canonical genetic lesions that is alteration of various transcription factors. The first and best- responsible for driving leukemia and solid tumor cell prolifera- characterized translocation was the t(15;17) translocation in tion.42 The Ras pathway can be activated at many levels. acute promyelocytic leukemia (APL), which juxtaposes the PML Indeed, the FLT3 activating mutations themselves as well as gene onto the retinoic acid receptor alpha (RARa) subunit essentially all mutations in tyrosine kinases are upstream of the gene.18–20 The most commonly targeted transcription factors in Ras pathway, driving its activation and resulting in cell growth. AML translocations include the RAR and the CBF, which has as A diagram of activation of the Ras pathway in leukemia is its central DNA-binding component AML1 or RUNX-1, as well presented in Figure 5. as various homeobox genes.21 For example, BCR-ABL can associate with adaptor proteins The activity of RAR is a paradigm for the function of altered like Shc and Grb2 and activate Ras. Direct modulators of Ras transcription factors in AML. The RAR acts as a dimer in concert activity, such as the Ras GAP proteins, can be mutated or with a partner, RXR. It sits on the transcriptional control region activated in juvenile CML.43 There are many cases of of a gene and acts to recruit a repressor complex.22–24 That myelodysplasia in AML, which have direct mutational activation

Leukemia Novel approaches to treat leukemia RA Larson et al 2362

Figure 4 PML-RARa translocation in AML. (a) In a normal cell in the absence of RA, a repressor complex is present on certain genes that may modulate differentiation. (b) RA can induce a disassociation of transcription complexes on certain genes that may be involved in differentiation. (c) In cells containing PML-RARa chromosomal translocation, differentiation is inhibited as it acts as a DN transcription factor and displaces RXRa and secures HDAC.

Figure 5 Activation of the RAS pathway in leukemia.

of Ras. The Ras pathway can be activated in many ways and is ABL transgenic mouse model, the mice treated with FTIs had thus an excellent therapeutic target. prolonged survival compared to vehicle-alone-treated mice.49 The enzymatic activity of Ras itself is not a particularly good FTIs also have activity in hematopoietic colony-forming target, but Ras is activated in an obligate way by a post- assays performed on cells taken from patient biopsies. BCR- translational isoprenylation.44 A fatty acid moiety is added to the ABL-positive colonies have a greater sensitivity to the drug than C terminus of Ras by farnesyl transferase (FT). Ras is thereby colonies from normal individuals.48 These results suggest there targeted to cell membranes, and a number of companies have is therapeutic selectivity with FTI compounds, which could be developed inhibitors of the FT that leave Ras in the cytosol as an useful in the treatment of CML. inactive enzyme.45–47 Together with Drs Mahon and Melo’s groups (Department of BCR-ABL is known to activate the Ras pathway. Dr Daley’s Haematology, Imperial College of Science, Technology and laboratory has investigated whether farnesyl transferase inhibi- Medicine, Hammersmith Hospital, London, UK), Dr Daley has tors (FTIs) might serve as blockers of CML. In two different shown that FTIs are also effective against cells from patients who mouse models, his laboratory has shown that treatment of mice are resistant to Gleevec.50 Hematopoietic colony-forming assays injected with BCR-ABL-transformed hematopoietic cells with indicate that treatment with one micromolar Gleevec leads to Ras inhibitors led to enhanced survival.48 Moreover, in a BCR- very low toxicity of BCR-ABL-positive cells from resistant

Leukemia Novel approaches to treat leukemia RA Larson et al 2363 patients. However, addition of an FTI at clinically achievable of BCR-ABL, TEL-PDGFR, and various activating mutations of levels generates complete suppression, suggesting that resis- FLT3 and the Ras pathway. tance to Gleevec does not imply cross-resistance to FTI. It is Class I mutations exist in AML in concert with class II therefore possible that FTI could be used successfully in mutations, which typically lead to loss of function of hemato- combination with Gleevec in CML patients. poietic transcription factors, perhaps through stable recruitment Whether Ras is the true target of FTIs remains controversial. of HDAC or transcriptionally repressive complexes. Class II Ras is one of several hundred proteins in the cell that require mutations act to block differentiation, and when seen alone in post-translational modification by isoprenylation. There is now a experimental models, induce an MDS-like condition. large body of literature implicating various other targets within How many drugs will it take to treat malignancies that harbor the cell that might be clinically relevant targets of FTIs.51–54 multiple genetic mutations? If patients with the CML-like FTIs have been tested in early-stage clinical trials. Karp and syndrome are treated early post diagnosis with targeted kinase colleagues at the University of Maryland Greenebaum Cancer inhibitors or perhaps the FTIs, significant single-agent activity Center (Baltimore, MD) have shown that the compound can be observed. However, in the more genetically complex Zarnestras, which has an oral route of delivery, was effective acute leukemias, combinations of target-directed agents will be in approximately 33% of the patients with highly refractory, needed for effective therapy. drug-resistant AML.55 There were also some CML blast crisis There is an enormous amount of enthusiasm because of the patients in this study who had partial responses (PRs). Another tremendous success of Gleevec. However, this enthusiasm FTI, lonafarnib, has also been tested in early phase clinical trials. should be tempered by the enormous challenges for targeted Approximately 20–30% of patients had some clinical response therapy that remain. It is daunting to consider the task of with this compound. There has been some modest compassio- inhibiting multiple targets with highly specific agents. Each of nate use of this compound together with Gleevec in some these drugs will have to be put through the very complicated resistant patients. Combination clinical trials are being planned. series of trials that are required before FDA approval. We are at A summary of where the several oncoproteins that have been the beginning of a very exciting era of target-directed leukemia discussed function in signal transduction pathway and where therapy and once again, hematology seems to be paving the way various inhibitors act is presented in Figure 6. for all of oncology in understanding how to translate molecular In summary, understanding the total number of mutations in a mechanisms into better therapies. cancer cell will provide insight into which target-directed agents are optimal for treating various malignancies. The CML-like syndromes appear to be the result of single genetic abnormal- Questions from audience ities and therefore target-directed agents are likely to have significant activity when used as monotherapy for these Q: My question may be futile in this therapeutic new area, but disorders. how do you reconcile the one-hit basis of CML with Dr In contrast, the more acute malignancies, like acute leukemia, Fialkow’s data? typified by CML in blast crisis, arise from at least two genetic George Daley, MD: The question is that there are classical lesions and perhaps more. Against acute leukemia, target- data from Dr Fialkow and colleagues using clonality analysis directed agents are likely to be needed in combination. Dr that suggested in some patients that there might be a pre-existing Gilliland has developed a very nice heuristic model to promote clonal phase of hematopoiesis onto which the BCR-ABL this hypothesis. He classifies mutational lesions in AML as translocation is superimposed.56 falling into two classes. Class I mutations, which typically drive These data suggest that there may be pre-existing genetic cell proliferation, create a CML-like or myeloproliferative lesions that contribute to the disease. I interpret the Fialkow data condition. These have been mimicked in mice by expression to suggest that a pre-existing tendency toward clonal expansion

Figure 6 Site of action of oncogenes and signal transduction pathway inhibitors (black circles).

Leukemia Novel approaches to treat leukemia RA Larson et al 2364 may indeed predispose individuals to the BCR-ABL transloca- of eight or nine transmembrane protein pumps that contain tion. CML is age-dependent and it tends to hit older rather than ABCs.60–62 This is an energy-dependent process by which drugs the younger individuals. So perhaps clonality represents a that passively cross the cell membrane are very quickly pumped general precondition of the marrow, perhaps through decay of out. A number of these proteins have been characterized and telomeres, which generates increased frequency of chromoso- may be clinically relevant in patients with AML. This meeting mal translocations. It is by no means clear, however, that the report will focus predominantly on MDR-1 or P-gp, because Fialkow data imply a necessary prior genetic event. most of the in vitro and clinical work has been carried out with We know in model systems, at least in the mouse, that the inhibitors of this particular pump. Many of the drugs used in the single hit is functional. Emerging data from primary human cells treatment of AML are substrates for MDR-1 and are rapidly generated by transferring BCR-ABL into primary human tissue extruded from drug-resistant cells.62,63 A diagram of the drug indicate that a myeloproliferative phenotype can be seen in pumps and the effects of pump modulators is presented in primary CD34 þ cells. Dr Fialkow’s data and the experimental Figure 7. data remain to be fully reconciled. Q: You covered most of the genes, which are involved in translocations. However, the ‘bad’ prognosis AML, for example, partial deletions of chromosomes 5 and 7, cannot be cured with chemotherapy. Even after bone marrow transplantation, the outcome is poor. What is known about critical genes that are involved in 5 and 7 chromosomal deletions? George Daley, MD: This is indeed a good point. Chromoso- mal deletion syndromes are frequently seen in myelodysplasias and secondary AMLs,57 and imply loss of tumor suppressor gene function. A general principle I also did not touch on is gene silencing – not just through transcriptional repression, but also through methylation.58 This leads to the therapeutic concept of demethylation agents, which may act to derepress or reactivate some of these classes of tumor suppressor genes. Q: Using very sensitive PCR, it is possible to demonstrate many of these translocations in totally normal individuals who never get leukemia. If one hit is enough, what is happening? George Daley, MD: It is actually quite alarming when one takes a sample of their own blood, and performs a sensitive PCR and finds BCR-ABL, which some in my lab have done. The reality is that the genetic lesion needs to occur in the right cell. The presumption is that as the hematopoietic stem cell under- goes dramatic proliferative expansion through all the various transit-amplifying populations, many translocations indeed occur. However, most of the target cells are unable to sustain the leukemia, because they are fated to ultimately differentiate, apoptose or die. So at least two bits of bad luck are needed for malignancy: the translocation has to occur in the first place and it has to occur in a cell that can sustain a long-term leukemia, like the hematopoietic stem cell.

Multidrug resistance in AML (Charles A Schiffer, MD)

Drug resistance represents the major barrier to the successful treatment of leukemia, and AML in particular. Why is AML so resistant? First, AML is a disorder of hematopoietic precursors, which are designed to ‘live forever’ until failure occurs in other organ systems, such as the heart, kidneys or lungs. Excretory organs are in constant contact with multiple types of toxins, particularly from the organisms that we live with, and have developed mechanisms to survive such exposure. For example, most resistance mechanisms are amplified in gastrointestinal (GI) tract precursors. Similarly, bone marrow purging experi- ments indicate that hematopoietic stem cells can survive exposure in vitro to large doses of a variety of chemotherapeutic agents.59 Therefore, it is not particularly surprising that leukemias that derive from cells that resemble hematopoietic stem cells (eg, Ph-positive diseases, myelodysplasia and all the myeloproliferative diseases) are very resistant to chemotherapy. A major cause of resistance is a consequence of active efflux Figure 7 Effects of drug pumps and modulators on chemother- of drugs that have entered the cells, usually mediated by a group apeutic drug resistance.

Leukemia Novel approaches to treat leukemia RA Larson et al 2365 Overexpression of P-gp can be detected in AML blasts at clinically relevant definition of ‘positive’ in any of these assays is diagnosis, particularly in older patients, and may be further not well defined and the magnitude of the in vitro fluorescence increased at relapse.63–67 In many studies, P-gp overexpression necessary to confer drug resistance in the patient is not known. is associated with a poorer outcome, as is overexpression of Moreover, it is not clear which cells are killed – the clonogenic some of the other efflux pumps, although the data on P-gp seem or the bystander cells. Lastly, it is not clear that the behavior of to be the strongest in this regard.64,66 Importantly, P-gp- rhodamine, one of the substrates commonly used in drug efflux mediated resistance can be reversed in vitro by a variety of experiments, is similar to the effects of clinically important drugs drugs, such as cyclosporine, PSC-833 and quinine.63,66 These in such as the anthracyclines. It will take cumbersome clinical vitro observations led to a series of clinical trials evaluating trials to answer these questions, which can take a long time and modulators of drug resistance in patients with AML. often can be difficult to analyze. A complicating issue is that the administration of some of the MDR1 inhibitors affects the metabolism and disposition of many chemotherapeutic drugs because of their effect on normal Clinical trials with MDR modulators tissues such as biliary lining cells. In general, when modulators of these efflux pumps are given to patients, there is impaired A number of clinical trials evaluating MDR modulators in AML excretion of the chemotherapeutic drug, increases in AUC and have recently been completed. The CALGB completed a Phase I drug exposure and the potential for increased toxicity. study, which was a prelude to a planned Phase III study, in older patients with previously untreated de novo AML.70 Patients received combination therapy with daunorubicin, etoposide and In vitro assays of P-gp cytarabine with or without the MDR-1 modulator, PSC-833. The maximally tolerated dose without the modulator was When in vitro assays on hematopoietic cells from AML patients 60 mg/m2 Â 3 doses of daunorubicin, 100 mg/m2 Â 3 doses of are performed, they are usually performed on whole populations etoposide and 100 mg/m2 of cytarabine for 7 days. With the of leukemia cells. A problem, however, is that these cells are addition of PSC-833, however, the MTD were approximately 1/ quite heterogeneous. For example, within a population of 3 lower (40 mg/m2 of daunorubicin and 60 mg/m2 of etoposide) leukemia cells with similar morphology, there are dramatic with mucositis as the dose-limiting toxicity. This was one of the differences in expression of different surface antigens. Further- largest Phase I trials that has ever been performed in acute more, only a small fraction of the leukemia cells have the leukemia, with results consistent with other studies demonstrat- capacity to grow in long-term culture or in SCID mouse models. ing the necessity of dose reduction because of the effect of the This subpopulation has been characterized by cell sorting as MDR inhibitor on the pharmacokinetics (PK) of the chemother- CD34 þ , HLA and DR negative. Animal experiments have apeutic agents. Trials using such modulators with reduced doses shown that injection of this subpopulation of cells is capable of of chemotherapy are based on the hope that intracellular drug producing sustained tumors.68 Experiments with blasts from levels and resultant cytoxicity of the clonogenic cells will be patients with AML have indicated that o1% of cells had the increased to a much greater extent than can be achieved with potential to produce long-term persistence either in culture or in higher doses alone. Indeed, earlier clinical trials suggest that animals, and might therefore represent leukemic ‘stem cells’ higher doses of etoposide and/or daunorubicin are not sufficient P-gp expression can be measured with antibodies directed to improve the outcome of treatment with AML. against membrane P-gp epitopes by flow cytometric analysis or CALGB 9720 was an extension of the Phase I study in which by measuring the function of the pump.69 Rhodamine 123 dye older people with de novo AML were randomized to receive the can be used to measure drug pump function as a surrogate for two regimens described above.71 Unfortunately, this study was chemotherapeutic drugs. In ‘normals’, incubation with rhoda- closed early because of excessive toxicity and mortality in the mine results in increased fluorescence. In resistant cells group receiving PSC-833. After about 120 patients were however, there is decreased fluorescence because of rapid analyzed, there was a somewhat lower complete response rate excretion of the dye. The effect of different modulators or pump in the experimental group, but a much higher early death rate inhibitors can also be evaluated in such assays. associated with mucositis and infection. These results could not It can be difficult to detect drug resistance in a heterogeneous be explained by differences in the clinical characteristics in the hematopoietic cell population. If only 0.5–1% of the cell two groups of patients. Overall, MDR-1 expression seemed to be population are actually responsible for persistence of leukemia a prognostic factor, in that the CR rate was higher in patients after treatment, even very sensitive flow cytometric assays might who were efflux-negative. be incapable of assessing the characteristics of this very small The ECOG did a study evaluating PSC-833, which was also subpopulation of cells that may be most important in terms of stopped early because of poor results. This study enrolled very drug resistance. To place this point in clinical perspective, even high-risk patients including those in early relapse, with primary marginally effective drugs produce a multilog reduction in the refractory disease, high-risk MDS or secondary leukemias and number of AML cells that are circulating or in the bone marrow. others who were post marrow transplantation. The patients were Thus, it is the very small subpopulation of residual cells that are treated with a regimen of mitoxantrone, etoposide and ara-C responsible for ultimate drug resistance and the patient’s death. with or without PSC-833.72,73 The doses of mitoxantrone and Interpretation of in vitro assays of MDR can also be difficult. It etoposide were reduced in the PSC-833 cohort. There was no has been known for years that there is variable and sometimes difference in overall or disease-free survival, but because of a poor concordance among the immune, molecular and func- significant decrease in the CR rate in the PSC-833 arm, this study tional assays used to assess the effects of P-gp expression. Most was stopped prematurely. laboratories now stress the importance of functional assays In addition to the CALGB and ECOG studies, another large measuring the efflux/retention of the particular drug of interest. It randomized trial conducted in Europe has been completed. is inferred in flow cytometry experiments that increased While this study has not yet been published, it also failed to immunofluorescence representing increased chemotherapeutic show an advantage of adding PSC-833 to induction therapy for drug retention results in increased cytotoxicity. However, a older patients with AML.

Leukemia Novel approaches to treat leukemia RA Larson et al 2366 Somewhat in contrast are the results of a study performed by when clinical trials are designed and affect sample size and SWOG using cyclosporine as an MDR modulator.73 PSC 833 is study feasibility. a nonimmunosuppressive, non-nephrotoxic analog of cyclos- porine, which was much more active in vitro and hence was Other MDR modulators preferred by some groups in clinical trials. The SWOG study included high-risk patients with AML who received cytarabine There are a number of other drugs that affect MDR, which are in and daunorubicin with or without cyclosporine. The same dose early clinical trials. These include Zosuquidar.HCl trihy- of daunorubicin was used in each arm, although a potentially drochloride (LY335979), which is reported not to have a PK important difference was that the daunorubicin was adminis- effect on anthracycline disposition.74,75 Phase I trials in tered by continuous infusion rather than by the standard bolus combination with standard dose chemotherapy have been used in the CALGB and ECOG studies. Responders received a completed and ECOG is contemplating a Phase III trial in second course of the same regimen. More than 200 patients AML patients utilizing this drug. were randomized, of whom more than half were in first relapse Other compounds include VX710, which has the potential and almost 15–20% were primarily refractory. Among the advantage of inhibiting MRP1 but which also had a significant patients who achieved remission, the relapse-free survival was effect on paclitaxel disposition and was associated with significantly higher in the group of patients who received hyperbilirubinemia suggesting an effect on hepatic function.76,77 cyclosporine compared to the control group and what would KXR9576 is an anthranilic acid derivative, which can be have been predicted from historical experience. This translated administered orally and intravenously.78,79 It has a prolonged into an overall survival advantage from the time of randomiza- effect after in vitro exposure to cells and it is presently in Phase I tion for the cyclosporine recipients. and II trials. In these studies, natural killer (NK) cells were What can be learned from these trials that can be applied in isolated and characterized ex vivo after XR9576 infusion. NK future studies and why have attempts to modulate P-gp in AML cells normally overexpress MDR and in some patients, there was been disappointing? An obvious issue is that there are additional 73 marked inhibition of rhodamine efflux after infusion of this MDR mechanisms of resistance not affected by PSC-833. An inhibitor, demonstrating a significant in vivo effect. intriguing possibility is that cyclosporine might have an effect R101933 is an MDR inhibitor that can be administered orally on these other resistance factors accounting for the positive and intravenously. Side effects have included somnolence; no results in the SWOG trial. There are also competing causes of PK effect was demonstrated when used in combination with treatment failure in AML, which present a design and statistical docetaxel.80 Lastly, GG120918 may have more pleiotropic conundrum. At least a third of older AML patients will die of effects in that it also modulates the BCRP.66 complications of treatment, which have nothing to do with whether MDR is successfully modulated. Although all patients should be included in the analysis of randomized trials, a high Future considerations fraction of patients with early death dilute the ability to detect a biologic effect of a modulator of MDR. What is next in drug resistance studies in AML? As noted, there Furthermore, it is not known whether the modulator, even in are modulators with less PK effects, which are of considerable vitro, enhances the killing efficiency of the critical clonogenic interest. It might be relevant to focus only on newly diagnosed leukemia cells. Although it is possible to achieve levels in vivo younger patients, in whom an additional goal could be the that are effective in vitro, it is not known whether this occurs in prevention of the emergence of resistance. The CALGB has every patient since there is considerable PK variability among completed such a Phase I trial in younger patients with PSC-833 individuals and higher levels may be needed in vivo to reverse and has a randomized study in younger patients in progress. P- drug resistance effectively. Another possibility is inadequate gp is also overexpressed in patients with acute lymphocytic exposure to chemotherapy in the patients who are receiving leukemia, and there is enormous room for improvement in the reduced doses of chemotherapy with the expectation that the treatment of adult ALL. Given the multiagent regimens used in modulator will actually increase the drug exposure. If this PK ALL, however, ‘ideal’ trials with MDR modulators will be effect does not occur reliably in every patient, it is possible that difficult to design. some individuals are actually being underdosed with che- Finally, although a lot is known about drug resistance, little is motherapy with the potential for inferior outcome. known as to what promotes apparent drug sensitivity. For AML may also not be the ideal tumor to evaluate the concept example, even in the poorest prognosis leukemias, a fraction of of MDR modulation. Only complete response is clinically patients are cured with chemotherapy alone. The blasts from relevant in AML, presenting a significant challenge for any new patients with acute progranulocytic leukemia, which are therapy. In contrast, in other diseases such as multiple myeloma exquisitely sensitive to the effects of anthracyclines, have low or in certain solid tumors, partial remission is clinically very expression of MDR1. Certain transcription factor mutations, TEL, AML1, AML1-ETO, can actually repress MDR expression by meaningful and may be more readily achievable. In addition, 81,82 early death is not a major issue in such tumors. interacting with the promoter region of the MDR. These turn Lastly, clinical trials have focused on high-risk patients out to be highly curable leukemias. In other patients, differences because there are lots of them and if a true benefit is observed, in the cytokinetic state of the leukemic stem cell or immune- it could be very obvious and detected more rapidly than in mediated suppression of minimal residual disease may con- patients with a better prognosis. However, these are also the tribute to disease cure. It is hoped that greater understanding of hardest AML patients to treat, and small but clinically important these mechanisms of drug sensitivity and resistance will differences could be missed particularly since such trials include translate into improved outcomes in the future. a biologically heterogeneous population. There is also some disagreement about what the relevant end point should be. Questions from audience Obviously overall survival is critical, but some researchers have focused on the CR rate, or the fraction of patients who are cured Q: I just wanted to make a comment about biricodar. I should rather than median survival. These issues have to be considered point out in the interest of fairness that I am working with the

Leukemia Novel approaches to treat leukemia RA Larson et al 2367 company that has tried to develop that drug. Your point about The evolving role of monoclonal antibodies in the treatment elevations in bilirubin is well taken; certainly if this occurs, of CLL (Pierluigi Porcu, MD) it could affect the PK of anthracyclines. However, those elevations were very modest, and when one actually looks at CLL is the most common type of adult leukemia in the Western very carefully done studies with PK of anthracycline alone and world.83 Most of the patients who develop CLL will need then in combination with biricodar, there are no substantial treatment at some point during the course of the disease, and differences in the PK. These studies have also been extended to many eventually will die from complications related to the other types of drugs, such as mitoxantrone, and the results seem disease or its treatment. CLL is not curable with conventional fairly consistent. So while I share your concerns, particularly in treatments. For a long time, CLL was considered a relatively this very delicate population of patients, the results are simple and even unexciting form of leukemia, compared to the encouraging. acute leukemias. This was more a reflection of the simplicity I have another point and I ask your thoughts about this. and limitations of the treatments available rather than lack of There have been a lot of questions raised about the biological or clinical complexity. The complexity of CLL was cyclosporine results. I certainly view them as extremely initially recognized with the staging classifications of Rai and encouraging. Is it conceivable that any of these results could Binet. Greater therapeutic sophistication in CLL was first occur by other effects of cyclosporine? I am curious about introduced with the independent discovery by Dr Grever (Ohio your views on this or do you think the data are done carefully State University, Columbus, OH, USA) and Dr Keating (MD enough to suggest that this occurs primarily by an MDR Anderson Cancer Center, Houston, TX, USA) that the purine mechanism? analog fludarabine was able to induce significant responses in Charles Schiffer, MD: With regard to the first comments, I patients with refractory disease.84,85 Eventually, fludarabine was thank you for those. I could not find literature or publications taken to front-line therapy with impressive results86 and is now that showed that, but if indeed there is no effect on PK, being used alone or in combination with alkylating agents and particularly of the leukemia-relevant drugs, that is very monoclonal antibodies, producing a high number of complete important. I think it is easier to do a randomized study where responses. you do not have to drop the dose of your chemotherapy in the Drs Rai (Division of Hematology/Oncology, Long Island experimental arm. Jewish Medical Center, New Hyde Park, NY, USA) and Binet Cyclosporine is a little more pleiotropic than we thought (Laboratory of Hematology, CHRU Bretonneau, Tours, France) in terms of inhibition of other ABC pumps and that could recognized the existence of different prognostic groups of CLL be one explanation for the SWOG results, because it clearly patients.87–89 Even though the classifications are slightly had some benefit in that trial with patients who were not different, both recognized that three major prognostic groups MDR-positive. were with different survivals. More recently, advances in our Q: I have a question regarding secondary neoplasias, maybe knowledge of the molecular biology of CLL cells have led to the as a consequence of MDR1 inhibition. You mentioned in the recognition of new biological prognostic markers, which can be beginning of your talk the essential meaning of MDR1 in integrated into the classifications of Drs Rai and Binet. Intriguing physiologic tissues. What about late consequences of MDR1 data about mutations in the variable region of the heavy inhibition when anthracyclines or other toxic drugs are being immunoglobulin (Ig) chain (IgVH) in CLL cells have recently administered? been reported.90,91 It is now recognized that there are two forms Charles Schiffer, MD: I must say, I have never thought about of CLL, one where the IgVH region contains somatic hypermu- it. I have always assumed that these patients would be happy to tations (SHM), indicating that these cells are derived from have had that opportunity. In addition, the period of exposure is germinal center (GC) B cells. These patients have a much better very brief. I guess, however, one could hypothesize that if you prognosis compared to patients with the second form of CLL, had an early intestinal lesion where MDR1 is an important who do not have SHM in the IgVH region. A second type of component in the GI tract stem cell, that you might protect it in biological risk stratification relies on the detection of specific some way I am not aware of human or any animal data, but it is chromosomal translocations in CLL cells.92–94 Three different an interesting thought. subgroups of patients have been recognized who have Q: My question concerns the issue of stem cells and how remarkably different lengths of survival and treatment-free important MDR is to the basic stem cells. When one examines intervals (TFI), depending on which chromosomal aberration the negative data – not the stuff by List, but the ECOG things you they carry. The presence of 17p deletion or 11q deletion are showed – is it lack of effect of the drug in PSC or is it the associated with clinically more aggressive disease and worse increased toxicity of the combination that results in the negative prognosis; patients with normal karyotype or trisomy 12 have an trial? I mean, is MDR so important in the basic stem cell that it is intermediate-risk prognosis; finally, patients with 13q deletion where the problem lies? have a better outcome. Charles Schiffer, MD: I think it is both. With regard to the first A class of compounds that must be considered among question, there is increased toxicity and some increased targeted therapies is that of mAbs. They can be designed to mortality, and that becomes particularly evident if more fragile target specific molecules important for the biology CLL cells, patients are examined. For example, in the CALGB study, in such as survival, proliferation and trafficking. A list of chimeric younger patients the toxicity was less of an issue. So I think it is a and humanized mAbs that are approved or in clinical function of both. development for the treatment of CLL is presented in Figure 8. Recall that a positive trial in this disease is often obtained if Monoclonal antibodies have become a new paradigm in the the outcomes are improved by 10%. Oncologists would be treatment of hematologic malignancies.95,96 First, the toxicity is ecstatic if that happened. However, with all the variables I have non-overlapping with that of chemotherapy and generally listed, one could see how even in large trials that have a couple myelosuppression is not observed. They also have excellent of hundred patients, small differences could be obscured. This is biodistribution in key anatomical areas, such as the bone why this has really been a very difficult field to study and to marrow, peripheral blood and spleen, that are critically involved perform clean clinical trials. in CLL and lymphoma. They have a novel, albeit incompletely

Leukemia Novel approaches to treat leukemia RA Larson et al 2368

Figure 8 Use of engineered antibodies to treat CLL.

understood, mechanism of action. Finally they have very broad plasma levels of rituximab and also was associated with spectrum of targets, such as surface antigens, cytokines, significant response rates in previously treated patients. These adhesion molecules, etc. two studies illustrate how part of the resistance of CLL cells to One of the best-known monoclonal antibodies is rituximab rituximab might be overcome with higher plasma levels of (Rituxans), a chimeric antibody, which binds to the CD20 antibody. molecule expressed on B cell. It has significant single-agent Several questions remain, however, as to the specific activity in low-grade B-cell lymphoproliferative disorders and mechanism of rituximab’s action in CLL as opposed to low- modest single-agent activity in diffuse large-cell lymphomas, grade lymphoma. One of the currently favored mechanisms of with a very favorable toxicity profile. In CLL, early studies in action of rituximab in B-cell malignancies is that it works previously treated patients using the standard weekly  4 through antibody-dependent cell cytotoxicity (ADCC). A sig- schedule did not show a significant activity, but when nificant amount of evidence suggests that ADCC is very previously untreated patients were examined it was recognized important for the activity of rituximab. However, many of the that standard-dose rituximab had reasonable activity.97 How- antibodies that are currently used in the treatment of cancer are ever, it is clear that, comparatively, standard-dose rituximab in also known to induce intracellular signaling when tumor cells CLL and small lymphocytic lymphoma (SLL) is substantially less are examined in vitro, and the possibility that rituximab-induced active than in low-grade lymphomas. signaling plays a role in its antitumor activity in vivo should be There may be different reasons for this phenomenon. One is carefully examined. the different PK of rituximab in CLL and SLL as opposed to low- ADCC is mediated by a set of cell surface receptors that bind grade lymphomas. Other possibilities include a deficient or the Fc portion of the immunoglobulin G (FcgR). While many different mechanism of mAb-mediated killing compared to low- isoforms of FcgR exist, they can generally be distinguished as grade lymphomas, resistance to complement-mediated killing activating or inhibitory,100 depending on the effect that their and decreased CD20 density in CLL cells. Additionally, there engagement has on the ability of the effector cell to perform may be a deficiency in immune-effector cells in CLL. These ADCC. Many innate immune effector cells carrying these patients have significant underlying immunodeficiency related receptors can efficiently mediate ADCC, remarkably among to their disease and they then are treated with alkylating agents, them NK cells, which only express the activating receptors.101 In steroids and fludarabine, which add to the immunosuppression. contrast, other cells such as B cells and monocytes express some Finally, defective signaling events leading to reduced sensitivity inhibitory receptor that can turn off ADCC. to apoptosis-inducing signals may also play a role. Jeffrey Ravetch’s group (Laboratory of Molecular Genetics and The problem of unfavorable PK has been addressed clinically Immunology, Rockefeller University, New York, NY, USA), has in two different ways. O’Brien et al (Leukemia Department, The very elegantly demonstrated the importance of FcgR in anti- University of Texas MD Anderson Cancer Center, Houston, TX, body-mediated tumor killing in vivo.102 They performed a study USA) performed a dose-escalating trial with rituximab, which with mice infused with B16 melanoma cells. In wild-type mice was taken up to 2.25 g/m2 as a weekly dose.98 This study given no treatment, massive pulmonary metastases were showed that dose escalation could be accomplished safely, and observed. If antibodies against the B16 melanoma cells were that the response rates were increasing with higher doses of administered to the wild-type mice, a reduction of the incidence rituximab. A second approach was adopted by Byrd et al of pulmonary metastases was observed, although a substantial (Division of Hematology-Oncology, Ohio State University, amount of disease was still present. However, when knockout Columbus, OH, USA), which used a three times a week mice for the inhibitory form, but not the activating form, of the administration of rituximab.99 This led to substantially higher FcgR were administered the antibody, there was a near-

Leukemia Novel approaches to treat leukemia RA Larson et al 2369 complete clearance of pulmonary metastases, indicating To determine whether rituximab can be taken to the upfront that in vivo the presence of inhibitory FcgR was decreasing the stage in combination with fludarabine, CALGB conducted a efficacy of the antimelanoma monoclonal antibodies. clinical trial (CALGB 9712) where patients with newly These experiments strongly suggest that in vivo, ADCC depends diagnosed CLL were randomized to receive either fludarabine on the balance between inhibitory and activating FcgR followed by rituximab or a concomitant regimen of fludarabine and that alterations of this balance have profound effects on and rituximab followed by maintenance rituximab.110 A total of the outcome of therapy with mAbs. A way to increase the 104 patients were involved. A good number of them had balance in favor of ADCC, with the goal of enhancing intermediate risk and about half of them had high risk. The the efficacy of mAb-based therapy, is to increase the response analysis was intriguing, particularly in terms of the number of cells that are carrying the activating FcgR by giving complete responses, indicating that there was a significant a cytokine such as interleukin-2 (IL-2) that expends NK cells. difference between the concurrent and the sequential arms, with Some of these approaches are being explored in lymphoma by 47% and 28% response rates respectively. These results suggest many different research groups.103 On the other hand, Th2 that rituximab was increasing the sensitivity of CLL cells to the cytokines such as IL-4 and IL-10, which increase the expression cytotoxic effect of fludarabine. The other important observation of inhibitory FcgR on these cells, are associated with a reduced in this trial was that this result was achieved at no cost of ADCC effect. increased toxicity. In terms of the incidence of opportunistic CLL, however, is a difficult disease target from the standpoint infections, there were more opportunistic infections in the of ADCC for a number of reasons. First, ADCC usually requires a sequential as opposed to the concurrent arms. In terms of favorable effector-to-target ratio, which is very difficult to hematological and nonhematological toxicity, there was slightly achieve in CLL because of the high number of circulating tumor more, but not significant, hematological toxicity in the cells. Second, patients with CLL tend to have a Th2 cytokine concurrent arm. profile104–106 that downmodulates ADCC. Third, CLL cells Some of the correlative studies performed by Dr Byrd et al in express the inhibitory FcgR.107 Finally, experimental studies of their initial trial of single-agent rituximab in CLL patients aimed ADCC in CLL are difficult to perform because there is no good at understanding the mechanism of the severe infusional preclinical model for human CLL at this time and it is hard to reactions to rituximab, observed in CLL patients. Byrd et al isolate sufficient effector cells from the peripheral blood to observed that patients who were having severe infusional perform in vitro ADCC assays. Based on these considerations, reactions had significantly higher levels of TNF-a as opposed one would predict that ADCC may not play a major role in the to those who did not. TNF-a appears to be associated not only antitumor activity of rituximab in CLL. with toxicity, but it is also likely to be responsible for protecting Induction of apoptosis may be an alternative possibility for the CLL cells from apoptosis. TNF-a signaling through its receptor effect of rituximab in CLL.108 If CLL cells are incubated with induces activation of NF-kB, which increases Bcl-2 transcrip- rituximab in vitro, particularly when crosslinking antibodies are tion. This signal therefore not only increases survival, but also added, a substantial amount of apoptosis is observed. Evidence blocks the proapoptotic signaling induced by rituximab binding for a role of induction of apoptosis also comes from studies to CD20 in CLL cells. Therefore, TNF-a may become an examining the signaling activity that follows rituximab binding attractive target from the therapeutic standpoint in addition to to CLL cells in vivo. The proform of caspase-3 is converted into being a target in terms of reducing the infusional reactions. A the active p16 form in cells taken from patients after a 1- to 3- clinical trial examining the combination of rituximab and day exposure to rituximab.109 The level of antiapoptotic etanercept, a dimeric fusion protein that binds TNF-a, blocking molecules such as XIAP and MCL-1 following exposure to its interaction with cell surface receptors,111,112 is in progress at rituximab is also reduced. Of note, caspase activation correlates the Ohio State University. The hypothesis behind this trial is that with reduction in lymphocyte counts following rituximab by reducing the availability of TNF-a to CLL cells, there will be therapy. A model for the mechanism of action of rituximab in an increase in the efficacy of rituximab and potentially a CLL has therefore been proposed.109 Exposure to rituximab reduction in the incidence of severe infusion reactions. results in activation of the caspase pathway of apoptosis. Both Campath-1H is a humanized anti-CD52 antibody and has caspase-3 and caspase-9 are activated. This results in activation significant activity in T-cell prolymphocytic leukemia (T-PLL), of the downstream effectors of apoptosis. Additionally, two CLL and cutaneous T-cell lymphomas (CTCL).113 The target antiapoptotic molecules, MCL-1 and XIAP, are down-modulated antigen, CD52, is a GPI-linked glycoprotein whose function is in CLL cells following binding of the antibody, potentially not well known. CD52 is broadly expressed in B cells, T cells, making the cells more sensitive to the action of other drugs monocytes, NK cells and dendritic cells.115 The mechanism of as well. action of Campath-1H is not known. A clinical trial with What has been learned from the single-agent rituximab Campath-1H was conducted in fludarabine-refractory pa- studies in CLL? First, rituximab is active as a single agent, tients.114–116 Campath-1H was administered intravenously by particularly when the dose and the schedule of administration a stepped-up dosing with 3, 10 and 30 mg over a week and then are targeted to reach a significant blood level of the antibody. followed by about 12 weeks of three times a week administra- Second, rituximab is safe, and excessive myelosuppression or tion, with PCP and varicella zoster prophylaxis. In this pivotal infectious complications are not observed. Third, the apoptotic trial, 92 patients were enrolled. The overall response rate was pathways activated in response to the binding of rituximab to 33%. This was remarkable for fludarabine-refractory patients, CD20 are likely to have a major role in the antitumor activity of who before the availability of Campath-1H had no viable option this antibody in CLL. for therapy. The median survival was 16 months. Importantly, How to improve on rituximab therapy? One obvious way is to among the people who actually responded, there was a combine the antibody with other active agents, such as significant increase in median survival. fludarabine or alkylating agents. Another is to combine it with Toxicity was substantial, both at the hematological and agents that down-modulate known survival signals for CLL cells nonhematological level. A 50% incidence of grade 3/4 coming from the microenvironment, among them tumor neutropenia and thrombocytopenia was observed. Infections, necrosis factor-a (TNF-a). fever, rigors, rashes, nausea and vomiting, which were difficult

Leukemia Novel approaches to treat leukemia RA Larson et al 2370 to manage, were also observed. Patients required antiemetics. and for treatment. This is a significant advance in terms of the Some deaths occurred from infection. An additional complica- planning of further clinical trials. The mAbs active in CLL may tion was frequent reactivation of cytomegalovirus (CMV). work differently in CLL as compared to non-Hodgkin’s CALGB is now conducting two studies with Campath-1H. lymphomas. Intriguing is the possibility that different mAbs One consists of fludarabine induction, followed by randomiza- may eliminate specific subsets of CLL clones, depending on tion to intravenous Campath-1H vs a new subcutaneous (SQ) which molecular abnormalities they carry. Therefore, combin- schedule of administration, which seems to be associated with a ing different antibodies and biologic therapies should be the significantly lower incidence of infusional reactions. A second focus of the next generation of clinical trials. CLL trial in CALGB is CALGB10101, which is looking at the fludarabine–cyclophosphamide–rituximab (FCR) regimen to be followed by subcutaneous Campath-1H. In both of these trials, Questions from the audience the principal goal is to achieve the highest possible level of complete remission in front-line patients with CLL. Q: One of the hazards that seems to have emerged is the A third mAb in clinical trials for CLL is Hu1D10.117 Hu1D10 proximity of giving fludarabine followed by Campath-1H. A targets an epitope on the HLA-DR beta chain, which is recent German trial has shown prolonged cytopenias accom- expressed on many hematopoietic cells, including B cells, panied by very profound fungal infections. Would you like to monocytes, T cells and dendritic cells. Altogether, about 50– comment on that and how long a gap you really need to leave 60% of B-cell lymphomas and approximately 75% of CLL between the two? patients express this epitope. Hu1D10 can mediate ADCC and Pierluigi Porcu, MD: I would say that the approaches that are complement-dependent cytotoxicity (CDC), and induce apop- being taken right now are to wait several weeks, 4–6 weeks. I tosis in vitro.118,119 Similar to rituximab, when Hu1D10 is think that it also depends on the severity of myelosuppression combined with crosslinking antibodies, the amount of cells that following fludarabine, which may be significantly different from undergo apoptosis is significantly higher. Unlike rituximab, patient to patient. So, it is going to be difficult to have a though, Hu1D10 does not activate caspase-3 or PARP cleavage. homogenous approach that will be effective for all the patients. Hu1D10 reduces the mitochondrial membrane potential in CLL Q: I think the German trials are really quite worrying, because cells (JC Byrd, unpublished results), suggesting that the they did not leave even as much as 2 months, but there are mechanism of induction of apoptosis in CLL cells with several trials that are leaving 8 weeks. I am hesitant to say that is Hu1D10 might be different from that of rituximab. going to be enough, because they were getting 6 months of A Phase I study in lymphomas was recently completed. There neutropenia after that combination. were four drug levels with a standard weekly  4 schedule, Pierluigi Porcu, MD: Yes, I agree that the issue of hemato- stepping up from 0.15 to 5 mg/kg. In one cohort, there was an logical toxicity is extremely important. attempt to administer a more intensified treatment with daily Richard Larson, MD: I can comment on that, perhaps. In the incremental doses of Hu1D10 over a week. Some intriguing CALGB study, there were only four cycles of fludarabine responses were observed in this Phase I trial in lymphomas, administered followed by a 2-month interval and then followed which led to a Phase II trial in low-grade lymphoma and an by 4 weeks of Campath-1H therapy. Fungal infections were not additional pilot Phase I trial in CLL and ALL. In the Phase II study observed but reactivation of CMV occurred. Therefore, there is in low-grade lymphoma, preliminary data do not appear to some concern about the combined immunotoxicity of that confirm the activity of this antibody. In CLL and in ALL, a Phase I sequential regimen. study was initiated in collaboration between the Ohio State Q: I think that the major challenge that we are facing now is University and the University of Chicago. The schedule of when to treat early-stage A Binet CLL. I think that approximately administration was thrice weekly. This schedule was adopted 50% will never be treated and will live 20–30 years without any because the antibody has a very short half-life. A stepped-up complications. So do you have any idea of the place of these dosing was chosen, similar to the one used for Campath-1H, to antibodies in the treatment of CLL? decrease infusional toxicity. The study design was to gradually Pierluigi Porcu, MD: I think that most of the combination increase the dose of Hu1D10 to reach the target dose level for approaches that are being explored now will have to be tested in each particular cohort, which were 1.5, 3 and 5 mg/kg. In seven Stage A Binet CLL patients who have high-risk disease defined of 11 evaluable patients, there was a significant decrease in by FISH or IgVH mutational status, as the toxicity of these tumor volume at each dose level. Several patients had combinations remains a big concern. This is only a minority of stabilization of their disease. There was a PR in a patient with Binet A patients but one with no good expectations for a a 17p deletion. This was very interesting, as these patients are prolonged treatment-free interval or survival. I think that for known to be insensitive to rituximab. Moderate toxicity was patients who have good-risk early-stage CLL, deciding whether observed with relatively unusual adverse events, such as or not they should go on one of the clinical trials with delayed infusion toxicity with hypotension even several hours combination of antibodies and fludarabine or antibodies, after the completion of the infusion, grade 3 hypophosphatemia, fludarabine, and alkylating agents remains a great challenge. rash and urticaria. There were some cases of thrombocytopenia. A patient who might need therapy and is not willing to be Nausea was reported in these studies as well as headaches, treated on one of those trials, or has concerns about toxicity, which occurred in a significant number of patients. Based on the could be treated with single-agent antibodies, since they are observation that Hu1D10-induced apoptosis in CLL cells substantially more active in newly diagnosed patients. follows different signaling pathways compared to rituximab (caspase-independent vs caspase-dependent, respectively) and the preliminary evidence of clinical activity in CLL, further The remaining challenges in ALL (Ching-Hon Pui, MD) development of this antibody in the treatment of CLL is being pursued. As cure rates in pediatric ALL reach 80%, emphases are placed In conclusion, cytogenetics and IgVH mutational status are on precise risk-directed therapy to avoid over- or under- new powerful tools for risk stratifying CLL patients at diagnosis treatment, to elucidate the mechanisms involved in drug

Leukemia Novel approaches to treat leukemia RA Larson et al 2371 resistance and to develop new therapeutic strategies.120 Genetic insights into the biology of these leukemia subgroups. For analyses have contributed greatly to our understanding of the example, E2A-PBX1 leukemias were characterized by high pathogenesis and prognosis of ALL, and have begun to identify expression of the c-MER receptor tyrosine kinase, a known molecular targets for specific treatment.121 Heretofore, specific transforming gene, suggesting that c-MER may be involved in genetic abnormalities with prognostic relevance were limited to the abnormal growth of these cells. Overexpression of FLT-3 cases of B-cell precursor ALL. By use of DNA microarrays, Tom receptor tyrosine kinase was detected in MLL-rearranged Look and colleagues (Harvard University, Boston, MA, USA) leukemias, a finding confirmatory of that recently reported by recently classified T-cell ALL into four major subtypes with Stan Korsmeyer and colleagues.128 On the basis of this finding, a prognostic significance: MLL-ENL and HOX11 clusters are Phase I pediatric trial of inhibitor of FLT-3 tyrosine kinase will associated with a favorable prognosis and TAL1 and LYL1 soon be initiated. clusters with inferior outcome.122 Altogether, approximately To elucidate the genomics of cellular responses to cancer 75% of childhood ALL cases now have identifiable specific treatment, Dr William Evans and colleagues (St Jude Children’s genetic abnormalities with prognostic and therapeutic rele- Research Hospital, Memphis, TN, USA) analyzed the expression vance.121 of over 9600 human genes in ALL cells before and after in vivo Although risk classification based on the primary genetic treatment with methotrexate and mercaptopurine given alone or abnormalities of leukemic cells has great intuitive appeal, its in combination.130 Changes in expression of 124 genes predictive value is not especially high,123 probably because of accurately discriminated among the four treatments. Discrimi- secondary cooperative mutations, host pharmacogenetics and nating genes included those involved in apoptosis, mismatch the over-riding impact of treatment on the prognostic informa- repair, cell cycle control and stress response. Only 14% of genes tion conveyed by these genetic abnormalities. For example, in that changed when these medications were given as single an international collaborative study, Maurizio Arico` and agents also changed when they were given together. Several associates showed that among Ph-positive ALL, an age of 1–9 gene clusters responded in a treatment-specific manner. For years and a low presenting leukocyte count confer a relatively example, expression of the ataxia telangiectasia mutated (ATM) favorable prognosis.124 Ph-positive cases with a leukocyte count gene was higher after treatment with methotrexate alone but not above 100 Â 109/l have an extremely dismal outcome. More after the other treatments. Consistent with G1 arrest by ATM, the recently, the same group of investigators demonstrated the percentage of ALL cells in S phase was lower after treatment clinical heterogeneity among cases with 11q23 abnormalities with methotrexate alone but higher after the other three (Table 1).125,126 Age was found to be the most important treatments. Ongoing studies by Dr Mary Relling and colleagues prognostic factor; infants younger than 1 year fared significantly (St Jude Children’s Research Hospital, Memphis, TN, USA) are worse than patients 1 year of age or older. Among patients with testing the hypothesis that gene expression profile can enhance t(11;19) ALL and MLL-ENL fusion, those with a T-lineage our ability to identify patients at risk of developing specific immunophenotype, who were all over 1 year of age, had a complications. Preliminary data suggested that supervised superior outcome as compared to patients over 1 year of age hierarchical clustering and principal component analysis can with B-lineage ALL.126 Notably, there was heterogeneity even identify selected genes expressed by ALL blasts that were among infants with t(4;11) ALL; those with a poor early response associated with the development of secondary AML or brain to prednisone had an especially dismal prognosis. tumors. However, these findings need to be confirmed by Gene expression profiling using DNA microarrays can not studying independent test sets. only accurately identify the major, important subtypes of ALL, Inherited differences in the metabolism and disposition of but also provide insights into their underlying biology and the drugs and polymorphisms in the genes that encode drug- nature of cellular responses to antileukemic therapy.122,127–129 metabolizing enzymes, transporters or targets can profoundly Dr James Downing and colleagues (St Jude Children’s Research influence the efficacy and toxicity of drug therapy, thereby Hospital, Memphis, TN, USA) studied 360 pediatric ALL affecting the antileukemic response.131 Essentially all genes patients. Distinct expression profiles identified six important encoding drug-metabolizing enzymes responsible for key leukemic subtypes, including T-cell ALL, E2A-PBX1, BCR-ABL, modifications of functional groups (phase I reactions) or for TEL-AML1, MLL rearrangement, and hyperdiploid 450 chro- conjugation with endogenous substituents (phase II reactions) mosomes.127 In addition, another ALL subgroup was identified exhibit polymorphism. In general, phase I enzymes activate based on its unique expression profile. Examination of the endogenous or exogenous substrates, and phase II enzymes genes comprising the expression signatures provided important deactivate or detoxify them. Thiopurine methyltransferase (TPMT) is a phase II enzyme that converts thiopurine prodrugs into methylated metabolites which are mostly inactive. This Table 1 The 5-year event-free survival estimates in ALL with enzyme competes with other enzymes involved in the activation 11q23 rearrangements, according to age and phenotype of thiopurine to their thioguanine nucleotides, which are incorporated into DNA and induce an antileukemic effect. Type Category % EFS (s.e.) P-value TPMT activity is inherited as an autosomal codominant trait. Approximately 90% of the population are homozygous for the t(4;11) o1 year 19 (3) o0.001 wild-type allele and have full enzyme activity; about 10% are 41 year 42 (5) heterozygous for the polymorphism and have intermediate t(11;19) B-lineage, o1 year 27 (9) 0.01 levels of enzyme activity; and one of 300 persons carries two B-lineage, 41 year 46 (14) T-lineage 88 (13) mutant TPMT alleles and does not express functional TPMT. t(9;11) o1 year 38 (15) 0.27 Patients with TPMT deficiency accumulate higher levels of 41 year 46 (14) thioguanine nucleotides, and experience more hematopoietic del(11)(q23) o1 year 40 (22) 0.05 toxicity when treated with standard doses of mercaptopurine, as 41 year 73 (5) compared to those with normal enzyme activity.132 Not Other 11q23 o1 year 22 (8) o0.001 surprisingly, patients who have deficient TPMT activity also 41 year 65 (7) tend to have better leukemic control than do those with normal

Leukemia Novel approaches to treat leukemia RA Larson et al 2372 activity.133 It should be noted that patients who have deficient Table 2 Risk factors for traumatic lumbar puncture at diagnosis of TPMT activity are at an increased risk of epipodophyllotoxin- ALL related AML or irradiation-induced brain tumor, in the context of antimetabolite therapy.134,135 In fact, therapy-related AML Risk factor Odds ratio (95% CI) has also been reported in these patients even when their Black 1.5 (1.2–1.8) treatment consisted primarily of antimetabolites.136 Age o1 year 2.3 (1.7–3.0) Thymidylate synthase, an essential enzyme in proliferating Early eraa 1.4 (1.2–1.7) cells, is an important target of methotrexate. A homozygous Less experience 1.4 (1.1–1.8) triple-tandem-repeat polymorphism of the thymidylate synthase Platelet count o100 Â 109/l 1.5 (1.2–1.8) enhancer has been associated with increased enzyme expres- aLack of dedicated procedure area and general anesthesia. sion and an inferior treatment outcome in childhood ALL.137 The homozygous polymorphism variant (substitution of T for C at position 677) of methylenetetrahydrofolate reductase, on the 6%, respectively.150,151 We believe that with more precise other hand, correlates with an increased risk of oral, GI or assessment of initial risk factors, and use of optimal intrathecal hepatic toxicity following low-dose methotrexate,138,139 and and systemic chemotherapy, CNS relapse hazard can be with greater in vitro sensitivity of leukemic blasts to methotrex- substantially reduced.152 In this regard, besides high-risk genetic ate.140 features and large leukemic cell burden, the presence of Early response to therapy is one of the most important leukemic cells in cerebrospinal fluid (even from iatrogenic prognostic factors because it reflects both the leukemic-cell- and introduction via traumatic lumbar puncture) is also associated host-related characteristics. Measurement of MRD, by flow with an increased risk of CNS relapse.153–155 We recently cytometric detection of aberrant immunophenotype or by PCR identified several factors associated with the risk of traumatic of clonal antigen-receptor gene rearrangements, affords a level lumbar puncture, including experience of the clinician, platelet of sensitivity and specificity not attainable by traditional count and the use of sedation or anesthesia (Table 2).156 In our morphologic assessment of treatment response.141,142 Dario center, the lumbar puncture is now performed by the most Campana and colleagues (St Jude Children’s Research Hospital, experienced clinician and under deep sedation or general Memphis, TN, USA) have shown that patients who attain a anesthesia. We transfuse all thrombocytopenic patients with molecular or immunologic remission (defined as leukemic platelets before lumbar puncture at diagnosis, and administer involvement of less than 0.01% of nucleated bone marrow intrathecal chemotherapy immediately after collection of cells) after 4–6 weeks of remission induction have a significantly cerebrospinal fluid. better treatment outcome than do those who do not achieve this Autologous transplantation has little if any therapeutic role in status.142,143 In fact, almost half of all patients have an MRD childhood ALL, and the indications for allogeneic stem cell level p0.01% after only 2 weeks of remission induction and transplantation from an HLA-matched related or unrelated these patients have an exceptionally good prognosis, with less donor during first remission of ALL are generally based on than 5% cumulative risk of relapse.144,145 By contrast, patients anecdotal evidence. A recent international collaborative study with an MRD level of X1% at the end of remission induction, a by Arico` and colleagues124 clearly demonstrated that allogeneic level that is difficult to detect by morphologic examination, stem cell transplantation from an HLA-matched related donor is fared as poorly as those who were considered induction failures superior to intensive chemotherapy alone in prolonging initial by conventional criteria (ie, X5% marrow blast cells). Tandem complete remission in children with Ph-positive ALL, regardless application of flow cytometry and PCR testing allows us to of other risk features. However, transplantation did not improve successfully detect MRD in virtually 100% of cases of childhood outcome in patients with t(4;11).125,126 Studies are needed to ALL.146 We have therefore incorporated MRD detection into our determine definitively if transplantation benefits patients with current risk assessment system.120 Comparison of paired bone severe hypodiploidy, near-diploidy or leukemia that has marrow and blood samples obtained concomitantly disclosed responded very poorly to initial remission-induction therapy. virtually identical levels from both sources in T-cell ALL but Several new agents are being tested. Arabinosylguanine is generally higher levels from bone marrow in B-lineage quite effective for T-cell ALL but it has narrow therapeutic index ALL.147,148 This finding suggested that T-cell ALL originates and is associated with neurotoxicity (eg, seizure and somno- from extramedullary sites (eg, thymus) and subsequently invades lence).157 Imatinib mesylate, which inhibits the BCR-ABL fusion bone marrow, and that T-cell cases can be monitored by protein and other tyrosine kinases, has induced a response rate examination of peripheral blood. of 70% with 20% complete (albeit transient) in patients with While many advances have been made in the treatment of relapsed Ph-positive ALL,158 and will be tested in newly ALL, only CNS-directed therapy and hematopoietic stem cell diagnosed cases. Other promising drugs include clofarabine159 transplantation will be briefly addressed here. Most contempor- and FLT-3 inhibitors.129 Further refinements in molecular ary clinical trials limit the use and dose of cranial irradiation classification, together with the identification of leukemic- and because of concern over late sequelae. We recently reviewed host-related genetic features that affect the efficacy and toxicity the survival experience of our 10-year event-free survivors and of antileukemic therapy, will afford unique opportunities to observed an over 20% cumulative risk of secondary neoplasms devise treatment plans for individual patients, and thus to further at 30 years from diagnosis among those who had received CNS advance the cure rate. irradiation, resulting in a slight excess in mortality rate as compared to the general population.149 The irradiated patients also had a lower employment rate and among women, a lower Questions from audience marital rate. This finding should provide additional impetus to omit the use of cranial irradiation. Q: I am curious about the data on the MRD. Do you think that Two studies have omitted cranial irradiation in all patients will lead eventually to a change in how these patients are regardless of their risk classification, resulting in rates of isolated treated? There is the maintenance phase, which has the potential CNS relapse of 4.2 and 3% and of any CNS relapse of 8.3 and for toxicities. It is a heritage of the past where patients were

Leukemia Novel approaches to treat leukemia RA Larson et al 2373 treated for a very long period of time. Now we have some high lysozyme levels and dehydration. Dehydration is a frequent mechanisms to see whether or not their disease is there or not event in leukemia at diagnosis. All these abnormalities could be and do you think that this will be changed in the future. life threatening and obviously TLS has to be prevented or Ching-Hon Pui, MD: While MRD determination can identify treated. A diagram of TLS and kidney dysfunction is presented in patients at increased risk of relapse who may benefit from Figure 9. intensified therapy, I do not think it can be used to determine Different types of TLS are observed, for example, spontaneous when we can stop therapy. ALL uniquely requires long-term TLS can occur before any chemotherapy in rapidly growing continuation treatment. Although half of the patients have no tumors and the best example is Burkitt’s lymphoma. Induced measurable disease even by flow cytometry or PCR at the end of TLS is also observed in tumors highly sensitive to chemotherapy, remission induction, all of them will relapse if continuation like Burkitt’s, but also in acute leukemia. Tumor burden in acute treatment is not given. leukemia can represent approximately 1 kg of tumor. After 1 The Tokyo Children’s Cancer Study Group had carried out an week, 99% of the tumor is destroyed, which can create interesting study that featured only 1 year of therapy, and physiological disorders. yielded an approximately 60% rate of long-term event-free Acute renal insufficiency is due to the renal tubular damage. survival. This finding indicates that 60% of patients may have Normally, phosphorus, uric acid and calcium are filtered by the therapy stopped at 1 year. However, this group of patients glomerulus, and are partly reabsorbed in the proximal tubules. cannot be reliably distinguished from those who require In the distal tubules the urine concentrates, and in the collecting prolonged therapy. Unfortunately, MRD determination cannot tubules precipitation of uric acid could occur in case of be used to guide cessation of therapy. For example, virtually all hyperuricemia, especially due to the acidic pH of the urine. 1 patients have negative MRD at the end of 22 years of Also, precipitation of calcium phosphate can occur if the pH is continuation therapy but 10–20% of them will relapse. There- basic. The situation is worsened when lysozymes also pre- fore, MRD cannot be used to determine the length of cipitate in tubules. This can occur in a patient with either a continuation therapy. However, it may prove to be useful to monocytic component or an M4 or M5 leukemia (according to determine the intensity of therapy. FAB classification). This precipitation leads to acute tubulo- pathy, leading to anuria and acute renal failure.

TLS in leukemias: frequency, diagnosis and management (Jean-Pierre Marie, MD) Frequency of TLS in adult acute leukemia

TLS is a group of metabolic abnormalities associated with a In the scientific literature it is very difficult to find an article on rapid cell death in patients with cancer.160–187 All the contents the frequency of TLS in adult acute leukemia. Dr Marie of the cell will be released in the circulation when the cell dies. presented a retrospective analysis looking at TLS and its This means that released potassium could lead to hyperkalemia. prevention or treatment in adults treated for acute leukemia Phosphorus release could lead to hyperphosphatemia and during the last 3 years at the Hoˆtel Dieu Hospital in Paris. hypocalcemia due to phosphorus–calcium complexes, which The files of 73 patients hospitalized at Hoˆtel Dieu hospital are responsible for nephrocalcinosis. Hyperkalemia and hypo- were reviewed. The mean age was 56. There were 63 AML and calcemia could be responsible for cardiac arrhythmia. 10 ALL patients and the mean WBC count was 35 500. This is The DNA present in the cell’s nucleus is degraded in the liver, quite representative of the acute leukemia patients hospitalized which can give rise to hyperuricemia. Hyperuricemia and at the Hoˆtel Dieu Hospital. The files of these patients were nephrocalcinosis are the major disorders associated with TLS; reviewed for TLS parameters. The levels of potassium, phos- they could lead to acute renal insufficiency. This renal phorus, uric acid and creatinine were examined at the time of insufficiency could be worsened by other factors, including diagnosis before any chemotherapy to ascertain the possibility

Figure 9 TLS and kidney dysfunction. This diagram provides an overview of the effects of TLS on kidney function and the effects of no treatment vs rasburicase treatment.

Leukemia Novel approaches to treat leukemia RA Larson et al 2374 of spontaneous TLS. The patient files were reviewed also during oxidation of hypoxanthine and xanthine oxidase. Uric acid is the first week of intensive chemotherapy, to determine the not very soluble, and can crystallize in the urine. In other frequency of chemotherapy-induced TLS. species, the terminal product is not uric acid but allantoin (Figure 10). Allantoin is the degradation product of uric acid by urate oxidase. Allantoin is five times more soluble than uric TLS at diagnosis acid, avoiding a precipitation in the tubule. In humans, the urate oxidase protein is not functional because of a nonsense Elevated life-threatening potassium levels were never observed. mutation. Significantly elevated phosphorus levels were observed in 15% of the cases. With regard to uric acid levels, 18% of the patients had a significant hyperuricemia. The mean values of creatinine Prevention of uric acid overproduction relies on for women and men according to EDGE were borderline. allopurinol However, 26% of the patients presented with elevated creatinine levels at diagnosis. This could have resulted from Allopurinol is a very potent inhibitor of xanthine oxidase and it dehydration. blocks the formation of uric acid156 (Figure 11). But this drug is The study at Hoˆtel Dieu Hospital was unable to find any unable to degrade the uric acid already formed. So it would be correlation between the WBC count and the potassium, very useful to have a recombinant urate oxidase able to destroy phosphorus, creatinine or uric acid levels. But interestingly, uric acid and to transform it into allantoin avoiding renal uric there was a very good correlation between the creatinine and acid crystallization. uric acid levels. A nonrecombinant urate oxidase, Uricozymes, was A correlation between FAB subtype and creatinine levels was extracted from Aspergillus flavus and has been available for observed. Patients with a monocytic, monoblastic component over 30 years in France and Italy.188–194 It is widely used to (M4 and M5 of the FAB classification) had a higher creatinine avoid TLS in patients with Burkett’s lymphoma, ALL and level at diagnosis than the other myeloid leukemia patients. AML. Severe allergic reactions and anaphylactic shock are Lymphoblastic leukemia patients also had higher creatinine rarely observed. But the limited production of Uricozyme and levels than the myeloid non-M4-M5 leukemia patients. its subsequent purification for use in humans precluded its The frequency of chemotherapy-induced TLS was examined. worldwide distribution. The variation of this parameter between day 0, before Recently, rasburicase, a recombinant urate oxidase, has chemotherapy and during the first week of chemotherapy was become available in the USA and in Europe.160,189,195–201 In examined. The percentage of variation of these parameters was the USA, its use is restricted to pediatric patients with tumors. examined. A mean variation of þ 11% was observed with Compared to the nonrecombinant Uricozyme, this recombinant potassium levels, but only four patients with more than a 50% product has a greater purity and also a greater enzyme activity. increase in potassium levels were observed during the first The increased production of rasburicase has permitted world- week. A mean variation of þ 27% was observed with wide distribution of its product. phosphorus, and 13% of patients had an increase in phosphorus Rasburicase is much more active than allopurinol: 4 h after levels of more than þ 50%. Phosphorus levels probably the first injection, there was a dramatic decrease of the serum represent the best surrogate marker for TLS, because preventa- uric acid levels after rasburicase compared to slow decrease tive treatments for decreasing phosphorus levels were not after allopurinol. At 48 h, there was no difference in serum uric administered. On the other hand, a sharp decrease of uric acid acid level between allopurinol and rasburicase. levels with more than a 50% decrease was observed in the Concerning safety, in a European study of 107 patients, only majority of the patients, and only 5% of patients experienced a 3% of adverse events related to the drug were observed. Two 450% increase of serum uric acid level. patients had rash and one patient had grade 3 hemolysis, but this With regard to the creatinine levels, the mean variation was patient was a G6PD-deficient patient. not significantly different before and during chemotherapy, and It was also observed that 7% of the patients generated two different patient populations were observed. A total of 42% antibodies (positive ELISA test), after the drug was administered. of patients had a sharp (450%) decrease in creatinine levels. It is important to remember that urate oxidase is a foreign On the other end, 11% of the patients had a significant increase protein, not synthesized by humans. Three deaths were (450%) of their creatinine levels. Patients with uric acid and observed during the study, but none of them were related to creatinine increases are the patients developing TLS most rasburicase. frequently. The prevention of TLS in leukemic patients involves In summary, TLS in acute leukemia, according to this heavy hydration and antiuric acid treatment. A total of 77% of retrospective analysis, is present in about 20% of the patients the patients in our series received three or more liters of at the time of diagnosis. A renal impairment was present in 26% hydration plus bicarbonate. Allopurinol was administered in of the patients, and uric acid and creatinine levels before the 91% of the cases, often alone but sometimes with urate beginning of chemotherapy were highly correlated with TLS. oxidase. In all, 29% of the patients received urate oxidase, sometimes alone but usually in combination with allopurinol. These measures to prevent hyperuricemia have to be initiated Treatment of TLS before chemotherapy in order to prevent increases in levels of creatinine during chemotherapy. When hyperuricemia Measures taken to contain TLS helped avoid major metabolic was prevented before chemotherapy, only 10.5% of the disorders during chemotherapy treatment. These measures patients showed elevated levels of creatinine during therapy. consisted of vigorous hydration and alkalinization, control of On the other hand, close to half of the patients displayed the uric acid production as well as its degradation. elevated levels of creatinine when the initiation of prevention Uric acid is a terminal product in human and primate and treatment of hyperuricemia were performed after the catabolism of urine (Figure 10). Uric acid is a product of beginning of chemotherapy.

Leukemia Novel approaches to treat leukemia RA Larson et al 2375

Figure 10 Purine catabolism and prevention of TLS. This diagram depicts biochemical pathways responsible for degradation of purines following rapid cell lysis induced by chemotherapy and also illustrates the ability of either purified or recombinant urate oxidase to convert uric acid to allantoin.

the patients with a high risk of TLS, such as FAB M4–M5 patients with lymphoblastic leukemia. The best time to begin urate oxidase treatment is before day 0 of chemotherapy, usually day À2 or day À3. Sometimes urate oxidase was administered later, but only as a salvage therapy when unexpected TLS developed. The levels of uric acid were examined 3 days after initiation of chemotherapy. A 90% decrease in uric acid levels was observed in patients treated with urate oxidase (7allopurinol), compared with a 36% decrease in patients treated with allopurinol alone. In summary, an early prevention of TLS before day 1 of chemotherapy reduces the risk of creatinine increases during chemotherapy. Allopurinol was administered to most patients, and urate oxidase was given preventively in an emergency setting to patients with monocytic components, in ALL patients Figure 11 Allopurinol is an inhibitor of xanthine oxidase. This with a large tumor burden, in patients with increased creatinine diagram depicts inhibition of xanthine oxidase by allopurinol to at presentation, and also in patients with high hyperleukocytosis prevent conversion of hypoxanthine to xanthine to uric acid. needing an immediate chemotherapy within 1–2 days. According to those results, guidelines were proposed for the prevention and treatment of TLS in adult acute leukemia. In Use of urate oxidase patients with a low risk of TLS, it is important to start hydration (1.5 l/m2 including bicarbonates) and allopurinol administration Urate oxidase was administered to two-thirds of the patients (600 mg/day) as soon as possible, usually 2 days before with increased creatinine levels at initial visit and to one-half of chemotherapy and until WBC decrease to normal. This

Leukemia Novel approaches to treat leukemia RA Larson et al 2376 treatment is sufficient to control hyperuricemia and to prevent apeutic drugs, as they are ‘designed’ to survive repeated increase of creatinine levels in 75% of the cases. exposure to heterogeneous toxins. The overexpression of one For patients who develop hyperuricemia despite this treat- of these pumps, P-gp, is detected in AML blasts, particularly in ment, it is important to increase hydration and administer urate older patients, and may be increased at relapse. P-gp over- oxidase at 0.2 mg/kg/day until uric acid is definitively con- expression is associated with a poorer outcome. Modulators of trolled. Usually, 2–3 days are enough to reduce uric acid levels the drug pumps have been discovered. These modulators (eg, to normal. cyclosporine, PSC-833) can reverse the effects of the drug- In patients with high risk of TLS or even if TLS is already resistant pumps. Thus, it is may be possible to treat patients with present or highly anticipated, it is important to begin as soon as lower concentrations of the chemotherapeutic drug and the possible before chemotherapy with vigorous alkaline hydration, modulators. The goal of these studies with drug pump 2–3 l/m2/day. Urate oxidase (0.2 mg/kg/day) is administered modulators is to raise the intracellular chemotherapeutic drug during the first few days of chemotherapy. levels in the leukemic cells to a greater extent than can be In conclusion, TLS was found in roughly 20% of adult patients achieved by administration of the chemotherapeutic drug by with acute leukemia. The main risk is acute tubulopathy, caused itself. A problem remains that these ‘drug pumps’ also serve mainly by crystallization of uric acid. Urate oxidase catalyzes important physiological roles in their normal tissues. Thus there the conversion of uric acid to allantoin, avoiding tubular is often toxicity associated with interference with the normal crystallization. TLS can be prevented through hydration plus functions of the drug pumps. antiuric acid treatment, mainly allopurinol or in selected CLL is the most common type of leukemia observed in the patients, urate oxidase. Severe TLS has to be treated/prevented Western Hemisphere with over 10,000 cases/year and is not with vigorous hydration, and urate oxidase. curable with currently available therapy. Most patients are initially asymptomatic but eventually require therapy. Until the advent of monoclonal antibodies, the therapeutic options Summary included treatment with chlorambucil, prednisone and purine analogs. Fludarabine is the best single drug for initial treatment In this meeting report, exciting new therapies for leukemia and results in improved response rates and progression-free treatment were discussed. Through molecular biology and survival. However, there exists significant opportunities to cytogenetic approaches, many of the genes responsible for improve fludarabine therapy. Humanized antibodies have been transformation of hematopoietic cells have been identified. Due generated to treat CLL. The humanized antibodies have led to to the elucidation that aberrant expression of certain oncopro- improved efficacy and diminished toxicity and represent new teins causes cancer, fundamental increases in our understanding hope in the treatment of CLL. These humanized antibodies are of the mechanisms of leukemogenesis have been obtained, directed at different proteins including CD22, CD23, CD52, which has yielded many novel therapeutic approaches. Many of HLA-DR and they exert their effects by different mechanisms these oncogenes activated in human cancer are created by the including ADCC, CDC and apoptosis. Some of these antibodies fusion of genes by chromosomal translocation. These transloca- will induce caspase 3-dependent (Rituxan) and caspase 3- tions result in two broad classes of oncoproteins, kinases (eg, independent (Hu1D10) apoptosis. Hu1D10 induces apoptosis in BCR-ABL, TEL-JAK) or transcription factors (TEL-AML-1, PML- CLL cells and it does not induce ADCC or CDC, and the RARa) with modified activities. Some oncogenes are activated apoptosis induced by Hu1D10 is caspase independent but leads by point mutations (RAS) or small amplifications of certain to the production of AIF release, which may involve changes in regions of the gene (FLT-3). Some chromosomal translocations the mitochondrial membrane potential. The apoptosis induced (BCR-ABL, TEL/PDGFbR) may occur singly in myeloproliferative by Hu1D10 may occur by inhibition of Bcl-2 function leaving disorders; however, complimenting translocations must occur many of the other apoptotic machinery intact, as overexpression for the transition to acute leukemia (NUP98/HoxA9, AML1/ of Bcl-2 is protective and neutralizes the apoptotic effects of EVI1, AML1/ETO). Hu1D10. Many other monoclonal antibodies have been More hope has been brought into the treatment of leukemia developed to treat CLL and are currently being evaluated. Some by the characterization of small molecular weight cell mem- of the antibodies target the CLL microenvironment and include brane permeable signal transduction pathway inhibitors. Glee- aTNF, aVEGF, aIL4, aSDF and abFGF. Other antibodies are vec is an inhibitor of BCR-ABL that can have astounding results conjugated to toxins (aCD22–diptheria toxin, aCD19–genestein) on the chronic phase of CML with 495% normalization of or some are radiolabeled (aHLA-DR-I131, aCDC20Y90, aCD22, blood counts. Other inhibitors have been developed that target aCD52I131). the FLT-3 and Ras pathway (FLT-3, Raf and MEK inhibitors). There remain many challenges to improve the survival rate in These inhibitors may be especially useful in the treatment of ALL. The type of chromosomal translocation the patient has and CML patients whose leukemia cells have become resistant to the type of hematopoietic cell that the lesion is present in often Gleevec. Another class of inhibitors is represented by RA, which determines how long a patient will have event-free survival. acts as a differentiating agent and can be effective in treating Leukemogenomics is now a real concept as gene expression APL patients with PML-RAR chromosomal translocation. profiling studies performed at St Jude Children’s Research Furthermore, as more is known about how normal and chimeric Hospital have led to exciting methods that can classify patients transcription factors work, it has been clear that compounds that into particular leukemia subtypes and can also predict whether inhibit histone deacetylation (HDAC inhibitors) may prove those patients will have relapses over a given time period. This is effective in certain leukemia therapies. These compounds may also being coupled with determining the patient’s genotypes for prevent the block in differentiation induced by the chimeric genes involved in intermediary metabolism, drug metabolism, transcription factor oncoproteins. drug transport, drug receptor, host susceptibility and disease Drug resistance remains a nagging problem in cancer pathogenesis. Since many of these are enzymes involved in drug chemotherapy. Eukaryotes have evolved pumps to remove metabolism, these differences can affect the patient’s outcome toxins they encounter in their environments. Hematopoietic after treatment with a particular drug. Thus leukemogenomics precursor cells may be particularly resistant to many chemother- and pharmacogenetics will eventually provide the physician

Leukemia Novel approaches to treat leukemia RA Larson et al 2377 with the proper drug and concentration to treat individual 3 Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000; leukemia patients. 100: 57–70. Complications of TLS can be prevented by either allopurinol 4 Land H, Parada LF, Weinberg RA. Tumorigenic conversion of or rasburicase treatment. This meeting also discussed the results primary embryo fibroblasts requires at least two cooperating oncogenes. Nature 1983; 304: 596–602. of recent clinical trials designed to prevent TLS by either 5 Hahn WC, Counter CM, Lundberg AS, Beijersbergen RL, Brooks allopurinol or rasburicase treatment. Clearly, rasburicase is MW, Weinberg RA. Creation of human tumour cells with defined more effective and rapid in reducing uric acid levels resulting genetic elements. Nature 1999; 400: 464–468. from TLS than allopurinol. As a result of the cloning of urate 6 Deininger MWN, Goldman JM, Melo JV. The molecular biology oxidase and the subsequent manufacturing of this recombinant of chronic myeloid leukemia. Blood 2000; 96: 3343–3356. protein, clinicians will be able to treat potential TLS patients 7 Ben-Neriah Y, Daley GQ, Mes-Masson AM, Witte ON, Baltimore D. The chronic myelogenous leukemia-specific P210 protein with rasburicase. Proper identification of patients of high risk for is the product of the bcr/abl hybrid gene. Science 1986; 233: TLS and their subsequent treatment with rasburicase will reduce 212–214. renal problems associated with TLS. Adverse immune reactions 8 Daley GQ, Van Etten RA, Baltimore D. Induction of chronic resulting from administration of a foreign protein may develop if myelogenous leukemia in mice by the P210bcr/abl gene of the cancer patients are re-treated with rasburicase. However, this Philadelphia chromosome. Science 1990; 247: 824–830. may not be a major problem if patients are cured of their cancer 9 Pear WS, Miller JP, Xu L, Pui JC, Soffer B, Quackenbush RC et al. Efficient and rapid induction of a chronic myelogenous leukemia- after a single treatment. Rasburicase is an intriguing example of like myeloproliferative disease in mice receiving P210 bcr/abl- how basic science can be applied to prevent a disease. A transduced bone marrow. Blood 1998; 92: 3780–3792. recombinant enzyme absent in humans is infused to catabolize 10 Li S, Ilaria Jr RL, Million RP, Daley GQ, Van Etten RA. The P190, uric acid that could otherwise accumulate to toxic levels in P210, and P230 forms of the BCR/ABL oncogene induce a similar certain disease states. chronic myeloid leukemia-like syndrome in mice but have Cancer research has surpassed many important milestones in different lymphoid leukemogenic activity. J Exp Med 1999; the past 30 years. These advances have stemmed from 189: 1399–1412. 11 Look AT. Oncogenic transcription factors in the human acute technological discoveries that have allowed the cloning of leukemias. Science 1997; 278: 1059–1064. DNA sequences, the discovery of oncogenes and their roles in 12 Azam M, Latek RR, Daley GQ. Mechanisms of autoinhibition and human cancer, the knowledge of how transcription factors, STI-571/imatinib resistance revealed by mutagenesis of BCR-ABL. promoters and chromatin structure affect gene expression, the Cell 2003; 112: 831–843. sequencing of the human genome, and the discovery and 13 Nagar B, Hantschel O, Young MA, Scheffzek K, Veach D, clinical applications of monoclonal antibodies. Many scientific Bornmann W et al. Structural basis for the autoinhibition of c-Abl tyrosine kinase. Cell 2003; 112: 859–871. advances have been achieved by brute-force screening of 14 O’Brien SG, Guilhot F, Larson RA, Gathmann I, Baccarani M, pharmaceutical libraries. This has led to numerous clinical Cervantes F et al. Imatinib compared with interferon and low- trials, which have tested the effects of particular compounds or dose cytarabine for newly diagnosed chronic-phase chronic antibodies to suppress cell growth, which might under certain myeloid leukemia. N Engl J Med 2003; 348: 999–1004. circumstance cure the leukemia patient. Undoubtedly and most 15 Deguchi K, Gilliland DG. Cooperativity between mutations in often, these drug screening and clinical trials yielded negative tyrosine kinases and in hematopoietic transcription factors in results. But we must always be optimistic with the advances that AML. Leukemia 2002; 16: 740–744. 16 Dash AB, Williams IR, Kutok JL, Tomasson MH, Anastasiadou E, the cancer field has made, and continue to test possible Lindahl K et al. A murine model of CML blast crisis induced by mechanism and therapies. Unquestionably, advances have been cooperation between BCR/ABL and NUP98/HOXA9. Proc Natl made that have improved the quality of care and the extended Acad Sci USA 2002; 99: 7622–7627. survival of leukemia patients. 17 Shah NP, Nicoll JM, Nagar B, Gorre ME, Paquette RL, Kuriyan J et al. Multiple BCR-ABL kinase domain mutations confer polyclonal resistance to the tyrosine kinase inhibitor imatinib Acknowledgements (STI571) in chronic phase and blast crisis chronic myeloid leukemia. Cancer Cell 2002; 2: 117–125. GQ Daley was supported in part by the Grants NIH CA76418, 18 Kakizuka A, Miller Jr WH, Umesono K, Warrell Jr RP, Frankel SR, Murty VV et al. Chromosomal translocation t(15;17) in human CA86991, DK59279 and HL71265, and a sponsored research acute promyelocytic leukemia fuses RAR alpha with a novel agreement from the Schering-Plough Research Institute. GQ putative transcription factor, PML. Cell 1991; 66: 663–674. Daley is a Birnbaum Scholar of the Leukemia and Lymphoma 19 de The H, Lavau C, Marchio A, Chomienne C, Degos L, Dejean Society of America. P Porcu was supported in part by a grant A. The PML-RAR alpha fusion mRNA generated by the t(15;17) (K23CA102155) from the NCI. C-H Pui was supported in part by translocation in acute promyelocytic leukemia encodes a the Grants CA 21765, CA 31566, CA 51001, CA 78824, CA functionally altered RAR. Cell 1991; 66: 675–684. 20 Goddard AD, Borrow J, Freemont PS, Solomon E. Characteriza- 29139, CA 37379 and GM 61393 from the US National Institutes tion of a zinc finger gene disrupted by the t(15;17) in acute of Health; a Center of Excellence grant from the State of promyelocytic leukemia. Science 1991; 254: 1371–1374. Tennessee, USA; and the American Lebanese Syrian Associated 21 Pandolfi PP. Oncogenes and tumor suppressors in the molecular Charities. C-H Pui is the American Cancer Society FM Kirby pathogenesis of acute promyelocytic leukemia. Hum Mol Genet Clinical Research Professor. JA McCubrey was supported in part 2001; 10: 769–775. by the Grants CA 51025 and CA098195. 22 Fenrick R, Hiebert SW. Role of histone deacetylases in acute leukemia. J Cell Biochem Suppl 1998; 30-31: 194–202. 23 Lin RJ, Nagy L, Inoue S, Shao W, Miller Jr WH, Evans RM. Role of the histone deacetylase complex in acute promyelocytic leukaemia. Nature 1998; 391: 811–814. References 24 Grignani F, De Matteis S, Nervi C, Tomassoni L, Gelmetti V, Cioce M et al. Fusion proteins of the retinoic acid receptor-a 1 Cho KR, Vogelstein B. Genetic alterations in the adenoma – recruit histone deacetylase in promyelocytic leukaemia. Nature carcinoma sequence. Cancer 1992; 70 (6 Suppl): 1727–1731. 1998; 391: 815–818. 2 Fearon ER, Vogelstein B. A genetic model for colorectal 25 Chen A, Licht JD, Wu Y, Hellinger N, Scher W, Waxman S. tumorigenesis. Cell 1990; 61: 759–767. Retinoic acid is required for and potentiates differentiation of

Leukemia Novel approaches to treat leukemia RA Larson et al 2378 acute promyelocytic leukemia cells by nonretinoid agents. Blood 49 Reichert A, Heisterkamp N, Daley GQ, Groffen J. Treatment of 1994; 84: 2122–2129. Bcr/Abl-positive acute lymphoblastic leukemia in P190 trans- 26 Lin RJ, Sternsdorf T, Tini M, Evans RM. Transcriptional regulation genic mice with the farnesyl transferase inhibitor SCH66336. in acute promyelocytic leukemia. Oncogene 2001; 20: 7204– Blood 2001; 97: 1399–1403. 7215. 50 Hoover RR, Mahon FX, Melo JV, Daley GQ. Overcoming STI571 27 Collins SJ. The role of retinoids and retinoic acid receptors in resistance with the farnesyl transferase inhibitor SCH66336. normal hematopoiesis. Leukemia 2002; 16: 1896–1905. Blood 2002; 100: 1068–1071. 28 Lin RJ, Sternsdorf T, Tini M, Evans RM. Transcriptional regulation 51 Lebowitz PF, Prendergast GC. Non-Ras targets of farnesyltrans- in acute promyelocytic leukemia. Oncogene 2001; 20: 7204– ferase inhibitors: focus on Rho. Oncogene 1998; 17: 1439–1445. 7215. 52 Du W, Lebowitz PF, Prendergast GC. Cell growth inhibition by 29 Doucas V, Brockes JP, Yaniv M, de The H, Dejean A. The PML- farnesyltransferase inhibitors is mediated by gain of geranylger- retinoic acid receptor a translocation converts the receptor from anylated RhoB. Mol Cell Biol 1999; 19: 1831–1840. an inhibitor to a retinoic acid-dependent activator of transcription 53 Law BK, Nørgaard P, Gnudi L, Kahn BB, Poulson HS, Moses HL. factor AP-1. Proc Natl Acad Sci USA 1993; 90: 9345–9349. Inhibition of DNA synthesis by a farnesyltransferase inhibitor 30 Minucci S, Nervi C, Lo Coco F, Pelicci PG. Histone deacetylases: involves inhibition of the p70S6k pathway. J Biol Chem 1999; a common molecular target for differentiation treatment of acute 274: 4743–4748. myeloid leukemias? Oncogene 2001; 20: 3110–3115. 54 Sepp-Lorenzino L, Rosen N. A farnesyl-protein transferase 31 Melnick A, Licht JD. Histone deacetylases as therapeutic targets inhibitor induces p21 expression and G1 block in p53 wild type in hematologic malignancies. Curr Opin Hematol 2002; 9: 322– tumor cells. J Biol Chem 1998; 273: 20243–20251. 332. 55 Karp JE, Lancet JE, Kaufmann SH, End DW, Wright JJ, Bol K et al. 32 Johnstone RW. Histone-deacetylase inhibitors: novel drugs for the Clinical and biologic activity of the farnesyltransferase inhibitor treatment of cancer. Nat Rev Drug Discov 2002; 1: 287–299. R115777 in adults with refractory and relapsed acute leukemias: 33 Kelly WK, O’Connor OA, Marks PA. Histone deacetylase a phase 1 clinical–laboratory correlative trial. Blood 2001; 97: inhibitors: from target to clinical trials. Expert Opin Invest Drugs 3361–3369. 2002; 11: 1695–1713. 56 Fialkow PJ, Martin PJ, Najfeld V, Penfold GK, Jacobson RJ, 34 Gilliland DG, Griffin JD. The roles of FLT3 in hematopoiesis and Hansen JA. Evidence for a multistep pathogenesis of chronic leukemia. Blood 2002; 100: 1532–1542. myelogenous leukemia. Blood 1981; 58: 158–163. 35 Kiyoi H, Ohno R, Ueda R, Saito H, Naoe T. Mechanism of 57 Padua RA, McGlynn A, McGlynn H. Molecular, cytogenetic and constitutive activation of FLT3 with internal tandem duplication genetic abnormalities in MDS and secondary AML. Cancer Treat in the juxtamembrane domain. Oncogene 2002; 21: 2555–2563. Res 2001; 108: 111–157. 36 Hayakawa F, Towatari M, Kiyoi H, Tanimoto M, Kitamura T, 58 Gojo I, Karp JE. The impact of biology on the treatment of Saito H et al. Tandem-duplicated Flt3 constitutively activates secondary AML. Cancer Treat Res 2001; 108: 231–255. STAT5 and MAP kinase and introduces autonomous cell growth 59 Kaizer H, Stuart RK, Brookmeyer R, Beschorner WE, Braine HG, in IL-3-dependent cell lines. Oncogene 2000; 19: 624–631. Burns WH et al. Autologous bone marrow transplantation in 37 Kiyoi H, Towatari M, Yokota S, Hamaguchi M, Ohno R, Saito H acute leukemia: a phase I study of in vitro treatment of marrow et al. Internal tandem duplication of the FLT3 gene is a novel with 4-hydroperoxycyclophosphamide to purge tumor cells. modality of elongation mutation which causes constitutive Blood 1985; 65: 1504–1510. activation of the product. Leukemia 1998; 12: 1333–1337. 60 Borges-Walmsley MI, Walmsley AR. The structure and function of 38 Yokota S, Kiyoi H, Nakao M, Iwai T, Misawa S, Okuda T et al. drug pumps. Trends Microbiol 2001; 9: 71–79. Internal tandem duplication of the FLT3 gene is preferentially 61 Ban T. Pleiotropic, multidrug-resistant phenotype and P-glyco- seen in acute myeloid leukemia and myelodysplastic syndrome protein: a review. Chemotherapy 1992; 38: 191–196. among various hematological malignancies. A study on a large 62 Karp JE. MDR modulation in acute myelogenous leukemia: is it series of patients and cell lines. Leukemia 1997; 11: 1605–1609. dead? Leukemia 2001; 15: 666–667. 39 Yamamoto Y, Kiyoi H, Nakano Y, Suzuki R, Kodera Y, Miyawaki 63 Leith CP, Chen IM, Kopecky KJ, Appelbaum FR, Head DR, S et al. Activating mutation of D835 within the activation loop Godwin JE et al. Correlation of multidrug resistance (MDR1) of FLT3 in human hematologic malignancies. Blood 2001; 97: protein expression with functional dye/drug efflux in acute 2434–2439. myeloid leukemia by multiparameter flow cytometry: identifica- 40 Tse KF, Novelli E, Civin CI, Bohmer FD, Small D. Inhibition of tion of discordant MDRÀ/efflux+ and MDR1+/effluxÀ cases. Blood FLT3-mediated transformation by use of a tyrosine kinase 1995; 86: 2329–2342. inhibitor. Leukemia 2001; 15: 1001–1010. 64 Borg AG, Burgess R, Green LM, Scheper RJ, Liu Yin JA. 41 Levis M, Tse K-F, Smith BD, Garrett E, Small D. A FLT3 tyrosine Overexpression of lung-resistance protein and increased P- kinase inhibitor is selectively cytotoxic to acute myeloid glycoprotein function in acute myeloid leukaemia cells predict leukemia blasts harboring FLT3 internal tandem duplication a poor response to chemotherapy and reduced patient survival. Br mutations. Blood 2001; 98: 885–887. J Haematol 1998; 103: 1083–1091. 42 Shannon K. The Ras signaling pathway and the molecular basis of 65 van den Heuvel-Eibrink MM, Sonneveld P, Pieters R. The myeloid leukemogenesis. Curr Opin Hematol 1995; 2: 305–308. prognostic significance of membrane transport-associated multi- 43 Largaespada DA. Genetic heterogeneity in acute myeloid drug resistance (MDR) proteins in leukemia. Int J Clin Pharmacol leukemia: maximizing information flow from MuLV mutagenesis Ther 2000; 38: 94–110. studies. Leukemia 2000; 14: 1174–1184. 66 Marie J-P. Drug resistance in hematologic malignancies. Curr 44 Pe´rez-Sala D, Rebollo A. Novel aspects of Ras proteins biology: Opin Oncol 2001; 13: 463–469. regulation and implications. Cell Death Differ 1999; 6: 722–728. 67 Covelli A. Modulation of multidrug resistance (MDR) in 45 Adjei AA. Blocking oncogenic Ras signaling for cancer therapy. hematological malignancies. Ann Oncol 1999; 10 (Suppl 6): J Natl Cancer Inst 2001; 93: 1062–1074. 53–59. 46 Le DT, Shannon KM. Ras processing as a therapeutic target in 68 Blair A, Hogge DE, Sutherland HJ. Most acute myeloid leukemia hematologic malignancies. Curr Opin Hematol 2002; 9: 308–315. progenitor cells with long-term proliferative ability in vitro and in 47 Chang F, Steelman LS, Lee JT, Shelton JG, Navolanic PM, Blalock vivo have the phenotype CD34+/CD71À/HLA-DRÀ. Blood 1998; WL et al. Signal transduction mediated by the Ras/Raf/MEK/ERK 92: 4325–4335. pathway from cytokine receptors to transcription factors: poten- 69 Beck WT, Grogan TM, Willman CL, Cordon-Cardo C, Parham tial targeting for therapeutic intervention. Leukemia 2003; 17: DM, Kuttesch JF et al. Methods to detect P-glycoprotein- 1263–1293. associated multidrug resistance in patients’ tumors: consensus 48 Peters DG, Hoover RR, Gerlach MJ, Koh EY, Zhang H, Choe K recommendations. Cancer Res 1996; 56: 3010–3020. et al. Activity of the farnesyl protein transferase inhibitor 70 Lee EJ, George SL, Caligiuri M, Szatrowski TP, Powell BL, Lemke SCH66336 against BCR/ABL-induced murine leukemia and S et al. Parallel phase I studies of daunorubicin given with primary cells from patients with chronic myeloid leukemia. cytarabine and etoposide with or without the multidrug resistance Blood 2001; 97: 1404–1412. modulator PSC-833 in previously untreated patients 60 years of

Leukemia Novel approaches to treat leukemia RA Larson et al 2379 age or older with acute myeloid leukemia: results of cancer and 88 Binet JL, Auquier A, Dighiero G, Chastang C, Piguet H, Goasguen leukemia group B study 9420. J Clin Oncol 1999; 17: 2831– J et al. A new prognostic classification of chronic lymphocytic 2839. leukemia derived from a multivariate survival analysis. Cancer 71 Baer MR, George SL, Dodge RK, O’Loughlin KL, Minderman H, 1981; 48: 198–206. Caligiuri MA et al. Phase 3 study of the multidrug resistance 89 Rai KR. Chronic lymphocytic leukaemia. Current strategy and modulator PSC-833 in previously untreated patients 60 years of new perspectives of treatment. Haematologica 1999; 84 (Suppl age and older with acute myeloid leukemia: Cancer and EHA-4): 94–95. Leukemia Group B Study 9720. Blood 2002; 100: 1224–1232. 90 Stilgenbauer S, Bullinger L, Lichter P, Do¨hner H, for the German 72 Advani R, Saba HI, Tallman MS, Rowe JM, Wiernik PH, Ramek J CLL Study Group (GCLLSG). Genetics of chronic lymphocytic et al. Treatment of refractory and relapsed acute myelogenous leukemia: genomic aberrations and VHgene mutation status in leukemia with combination chemotherapy plus the multidrug pathogenesis and clinical course. Leukemia 2002; 16: 993–1007. resistance modulator PSC 833 (Valspodar). Blood 1999; 93: 91 Dreger P, Brand R, Hansz J, Milligan D, Corradini P, Finke J et al. 787–795. Treatment-related mortality and graft-versus-leukemia activity 73 List AF, Kopecky KJ, Willman CL, Head DR, Persons DL, Slovak after allogeneic stem cell transplantation for chronic lymphocytic ML et al. Benefit of cyclosporine modulation of drug resistance in leukemia using intensity-reduced conditioning. Leukemia 2003; patients with poor-risk acute myeloid leukemia: a Southwest 17: 841–848. Oncology Group study. Blood 2001; 98: 3212–3220. 92 Stilgenbauer S, Lichter P, Do¨hner H. Genetic features of B-cell 74 Shepard RL, Cao J, Starling JJ, Dantzig AH. Modulation of P- chronic lymphocytic leukemia. Rev Clin Exp Hematol 2000; 4: glycoprotein but not MRP1- or BCRP-mediated drug resistance by 48–72. LY335979. Int J Cancer 2003; 103: 121–125. 93 Karnolsky IN. Cytogenetic abnormalities in chronic lymphocytic 75 Rubin EH, de Alwis DP, Pouliquen I, Green L, Marder P, Lin Y leukemia. Folia Med (Plovdiv) 2000; 42: 5–10. et al. A phase I trial of a potent P-glycoprotein inhibitor, 94 Do¨hner H, Stilgenbauer S, Do¨hner K, Bentz M, Lichter P. Zosuquidar.3HCl trihydrochloride (LY335979), administered Chromosome aberrations in B-cell chronic lymphocytic leuke- orally in combination with doxorubicin in patients with advanced mia: reassessment based on molecular cytogenetic analysis. J Mol malignancies. Clin Cancer Res 2002; 8: 3710–3717. Med 1999; 77: 266–281. 76 Rowinsky EK, Smith L, Wang YM, Chaturvedi P, Villalona M, 95 Dearden C. Monoclonal antibody therapy of haematological Campbell E et al. Phase I and pharmacokinetic study of paclitaxel malignancies. BioDrugs 2002; 16: 283–301. in combination with biricodar, a novel agent that reverses 96 Cheson BD. Hematologic malignancies: new developments and multidrug resistance conferred by overexpression of both MDR1 future treatments. Semin Oncol 2002; 29 (4 Suppl 13): 33–45. and MRP. J Clin Oncol 1998; 16: 2964–2976. 97 Montserrat E. Rituximab in chronic lymphocytic leukemia. Semin 77 Peck RA, Hewett J, Harding MW, Wang YM, Chaturvedi PR, Oncol 2003; 30 (1 Suppl 2): 34–39. Bhatnagar A et al. Phase I and pharmacokinetic study of the novel 98 O’Brien SM, Kantarjian H, Thomas DA, Giles FJ, Freireich EJ, MDR1 and MRP1 inhibitor biricodar administered alone and Cortes J et al. Rituximab dose-escalation trial in chronic in combination with doxorubicin. J Clin Oncol 2001; 19: lymphocytic leukemia. J Clin Oncol 2001; 19: 2165–2170. 3130–3141. 99 Byrd JC, Murphy T, Howard RS, Lucas MS, Goodrich A, Park K 78 Stewart A, Steiner J, Mellows G, Laguda B, Norris D, Bevan P. et al. Rituximab using a thrice weekly dosing schedule in B-cell Phase I trial of XR9576 in healthy volunteers demonstrates chronic lymphocytic leukemia and small lymphocytic lymphoma modulation of P-glycoprotein in CD56+ lymphocytes after oral demonstrates clinical activity and acceptable toxicity. J Clin and intravenous administration. Clin Cancer Res 2000; 6: Oncol 2001; 19: 2153–2164. 4186–4191. 100 Ravetch JV, Bolland S. IgG Fc receptors. Annu Rev Immunol 79 Mistry P, Stewart AJ, Dangerfield W, Okiji S, Liddle C, Bootle D 2001; 19: 275–290. et al. In vitro and in vivo reversal of P-glycoprotein-mediated 101 Cooper MA, Fehniger TA, Caligiuri MA. The biology of human multidrug resistance by a novel potent modulator, XR9576. natural killer-cell subsets. Trends Immunol 2001; 22: 633–640. Cancer Res 2001; 61: 749–758. 102 Clynes R, Takechi Y, Moroi Y, Houghton A, Ravetch JV. Fc 80 van Zuylen L, Sparreboom A, van der Gaast A, Nooter K, Eskens receptors are required in passive and active immunity to FALM, Brouwer E et al. Disposition of docetaxel in the presence melanoma. Proc Natl Acad Sci USA 1998; 95: 652–656. of P-glycoprotein inhibition by intravenous administration of 103 Porcu P, Caligiuri MA. Cytokine–antibody combinations in the R101933. Eur J Cancer 2002; 38: 1090–1099. therapy of lymphoma. Biol Ther Lymphoma 2002; 5: 8–11. 81 Javed A, Guo B, Hiebert S, Choi JY, Green J, Zhao SC et al. 104 Dancescu M, Rubio-Trujillo M, Biron G, Bron D, Delespesse G, Groucho/TLE/R-esp proteins associate with the nuclear matrix Sarfati M. Interleukin 4 protects chronic lymphocytic leukemic B and repress RUNX (CBF(alpha)/AML/PEBP2(alpha)) dependent cells from death by apoptosis and upregulates Bcl-2 expression. J activation of tissue-specific gene transcription. J Cell Sci 2000; Exp Med 1992; 176: 1319–1326. 113: 2221–2231. 105 Foa R, Massaia M, Cardona S, Tos AG, Bianchi A, Attisano C et al. 82 Lutterbach B, Sun D, Schuetz J, Hiebert SW. The MYND motif is Production of tumor necrosis factor-alpha by B-cell chronic required for repression of basal transcription from the multidrug lymphocytic leukemia cells: a possible regulatory role of TNF in resistance 1 promoter by the t(8;21) fusion protein. Mol Cell Biol the progression of the disease. Blood 1990; 76: 393–400. 1998; 18: 3604–3611. 106 Podhorecka M, Dmoszynska A, Rolinski J, Wasik E. T type 1/type 83 Dyer MJS, Oscier DG. The configuration of the immunoglobulin 2 subsets balance in B-cell chronic lymphocytic leukemia – the genes in B cell chronic lymphocytic leukemia. Leukemia 2002; three-color flow cytometry analysis. Leuk Res 2002; 26: 16: 973–984. 657–660. 84 Grever MR, Kopecky KJ, Coltman CA, Files JC, Greenberg BR, 107 Vervoordeldonk SF, Merle PA, van Leeuwen EF, van der Schoot Hutton JJ et al. Fludarabine monophosphate: a potentially useful CE, von dem Borne AEGK, Slaper-Cortenbach ICM. Fcg receptor agent in chronic lymphocytic leukemia. Nouv Rev Fr Hematol II (CD32) on malignant B cells influences modulation induced by 1988; 30: 457–459. anti-CD19 monoclonal antibody. Blood 1994; 83: 1632–1639. 85 Keating MJ, Kantarjian H, O’Brien S, Koller C, Talpaz M, 108 Bannerji R, Kitada S, Flinn IW, Pearson M, Young D, Reed JC et al. Schachner J et al. Fludarabine: a new agent with marked Apoptotic-regulatory and complement-protecting protein expres- cytoreductive activity in untreated chronic lymphocytic leuke- sion in chronic lymphocytic leukemia: relationship to in vivo mia. J Clin Oncol 1991; 9: 44–49. rituximab resistance. J Clin Oncol 2003; 21: 1466–1471. 86 Rai KR, Peterson BL, Appelbaum FR, Kolitz J, Elias L, Shepherd L 109 Byrd JC, Kitada S, Flinn IW, Aron JL, Pearson M, Lucas D et al. et al. Fludarabine compared with chlorambucil as primary The mechanism of tumor cell clearance by rituximab in vivo in therapy for chronic lymphocytic leukemia. N Engl J Med 2000; patients with B-cell chronic lymphocytic leukemia: evidence of 343: 1750–1757. caspase activation and apoptosis induction. Blood 2002; 99: 87 Rai KR, Sawitsky A, Cronkite EP, Chanana AD, Levy RN, 1038–1043. Pasternack BS. Clinical staging of chronic lymphocytic leukemia. 110 Byrd JC, Peterson BL, Morrison VA, Park K, Jacobson R, Hoke E Blood 1975; 46: 219–234. et al. Randomized phase 2 study of fludarabine with concurrent

Leukemia Novel approaches to treat leukemia RA Larson et al 2380 versus sequential treatment with rituximab in symptomatic, 131 Pui C-H, Relling MV, Evans WE. Role of pharmacogenomics and untreated patients with B-cell chronic lymphocytic leukemia: pharmacodynamics in the treatment of acute lymphoblastic results from Cancer and Leukemia Group B 9712 (CALGB 9712). leukaemia. Best Pract Res Clin Haematol 2002; 15: 741–756. Blood 2003; 101: 6–14. 132 Relling MV, Hancock ML, Rivera GK, Sandlund JT, Ribeiro RC, 111 Reed JC, Kitada S, Kim Y, Byrd J. Modulating apoptosis pathways Krynetski EY et al. Mercaptopurine therapy intolerance and in low-grade B-cell malignancies using biological response heterozygosity at the thiopurine S-methyltransferase gene locus. J modifiers. Semin Oncol 2002; 29: 10–24. Natl Cancer Inst 1999; 91: 2001–2008. 112 Keating M, Hallek M. Alemtuzumab, the first monoclonal 133 Relling MV, Hancock ML, Boyett JM, Pui C-H, Evans WE. antibody (MAb) directed against CD52. Med Oncol 2002; 19 Prognostic importance of 6-mercaptopurine dose intensity in (Suppl): S1–S2. acute lymphoblastic leukemia. Blood 1999; 93: 2817–2823. 113 Domaga"a A, Kurpisz M. CD52 antigen: a review. Med Sci Monit 134 Pui C-H, Relling MV. Topoisomerase II inhibitor-related acute 2001; 7: 325–331. myeloid leukaemia. Br J Haematol 2000; 109: 13–23. 114 Rai KR, Freter CE, Mercier RJ, Cooper MR, Mitchell BS, 135 Relling MV, Rubnitz JE, Rivera GK, Boyett JM, Hancock ML, Felix Stadtmauer EA et al. Alemtuzumab in previously treated chronic CA et al. High incidence of secondary brain tumors after lymphocytic leukemia patients who also had received fludar- radiotherapy and antimetabolites. Lancet 1999; 354: 34–39. abine. J Clin Oncol 2002; 20: 3891–3897. 136 Thomsen JB, Schrøder H, Kristinsson J, Madsen B, Szumlanski C, 115 Keating MJ, Flinn I, Jain V, Binet J-L, Hillmen P, Byrd J et al. Weinshilboum R et al. Possible carcinogenic effect of 6- Therapeutic role of alemtuzumab (Campath-1H) in patients who mercaptopurine on bone marrow stem cells: relation to have failed fludarabine: results of a large international study. thiopurine metabolism. Cancer 1999; 86: 1080–1086. Blood 2002; 99: 3554–3561. 137 Krajinovic M, Costea I, Chiasson S. Polymorphism of the 116 Kennedy B, Rawstron A, Carter C, Ryan M, Speed K, Lucas G et al. thymidylate synthase gene and outcome of acute lymphoblastic Campath-1H and fludarabine in combination are highly active leukaemia. Lancet 2002; 359: 1033–1034. in refractory chronic lymphocytic leukemia. Blood 2002; 99: 138 Urano W, Taniguchi A, Yamanaka H, Tanaka E, Nakajima H, 2245–2247. Matsuda Y et al. Polymorphisms in the methylenetetrahydrofolate 117 Mavromatis B, Cheson BD. Monoclonal antibody therapy of reductase gene were associated with both the efficacy and the chronic lymphocytic leukemia. JClinOncol2003; 21: 1874–1881. toxicity of methotrexate used for the treatment of rheumatoid 118 Shi JD, Bullock C, Hall WC, Wescott V, Wang H, Levitt DJ et al. arthritis, as evidenced by single locus and haplotype analyses. In vivo pharmacodynamic effects of Hu1D10 (remitogen), a Pharmacogenetics 2002; 12: 183–190. humanized antibody reactive against a polymorphic determinant 139 Ulrich CM, Yasui Y, Storb R, Schubert MM, Wagner JL, of HLA-DR expressed on B cells. Leukemia Lymphoma 2002; 43: Bigler J et al. Pharmacogenetics of methotrexate: toxicity among 1303–1312. marrow transplantation patients varies with the methylenetetra- 119 Stockmeyer B, Schiller M, Repp R, Lorenz H-M, Kalden JR, hydrofolate reductase C677T polymorphism. Blood 2001; 98: Gramatzki M et al. Enhanced killing of B lymphoma cells by 231–234. granulocyte colony-stimulating factor-primed effector cells and 140 Taub JW, Matherly LH, Ravindranath Y, Kaspers G-JL, Rots MG, Hu1D10 – a humanized human leucocyte antigen DR antibody. Zantwijk CH. Polymorphisms in methylenetetrahydrofolate re- Br J Haematol 2002; 118: 959–967. ductase and methotrexate sensitivity in childhood acute lympho- 120 Pui C-H, Campana D, Evans WE. Childhood acute lymphoblastic blastic leukemia. Leukemia 2002; 16: 764–765. leukaemia – current status and future perspectives. Lancet Oncol 141 Pui C-H, Campana D. New definition of remission of childhood 2001; 2: 597–607. acute lymphoblastic leukemia. Leukemia 2000; 14: 783–785. 121 Pui C-H, Relling MV, Downing JR. Acute lymphoblastic 142 Szczepanski T, Orfa˜o A, van der Velden VJH, San Miguel JF, van leukemia. N Engl J Med (in press). Dongen JJM. Minimal residual disease in leukaemia patients. 122 Ferrando AA, Neuberg DS, Staunton J, Loh ML, Huard C, Lancet Oncol 2001; 2: 409–417. Raimondi SC et al. Gene expression signatures define novel 143 Coustan-Smith E, Sancho J, Hancock ML, Boyett JM, Behm FG, oncogenic pathways in T cell acute lymphoblastic leukemia. Raimondi SC et al. Clinical importance of minimal residual Cancer Cell 2002; 1: 75–87. disease in childhood acute lymphoblastic leukemia. Blood 2000; 123 Pui C-H, Evans WE. Acute lymphoblastic leukemia. N Engl J Med 96: 2691–2696. 1998; 339: 605–615. 144 Panzer-Gru¨mayer ER, Schneider M, Panzer S, Fasching K, Gadner 124 Arico` M, Valsecchi MG, Camitta B, Schrappe M, Chessells J, H. Rapid molecular response during early induction chemother- Baruchel A et al. Outcome of treatment in children with apy predicts a good outcome in childhood acute lymphoblastic Philadelphia chromosome-positive acute lymphoblastic leuke- leukemia. Blood 2002; 95: 790–794. mia. N Engl J Med 2000; 342: 998–1006. 145 Coustan-Smith E, Sancho J, Behm FG, Hancock ML, Razzouk RI, 125 Pui C-H, Gaynon PS, Boyett JM, Chessells JM, Baruchel A, Kamps Ribeiro RC et al. Prognostic significance of measuring early W et al. Outcome of treatment in childhood acute lymphoblastic clearance of leukemic cells by flow cytometry in childhood acute leukaemia with rearrangements of the 11q23 chromosomal lymphoblastic leukemia. Blood 2002; 100: 52–58. region. Lancet 2002; 359: 1909–1915. 146 Neale GAM, Coustan-Smith E, Pan Q, Chen X, Gruhn B, Stow P 126 Pui C-H, Chessells JM, Camitta B, Baruchel A, Biondi A, Boyett et al. Tandem application of flow cytometry and polymerase JM et al. Clinical heterogeneity in childhood acute lymphoblastic chain reaction for comprehensive detection of minimal residual leukemia with 11q23 rearrangements. Leukemia 2003; 17: disease in childhood acute lymphoblastic leukemia. Leukemia 700–706. 1999; 13: 1221–1226. 127 Yeoh EJ, Ross ME, Shurtleff SA, Williams WK, Patel D, Mahfouz R 147 Coustan-Smith E, Sancho J, Hancock ML, Razzouk BI, Ribeiro et al. Classification, subtype discovery, and prediction of RC, Rivera GK et al. Use of peripheral blood instead of bone outcome in pediatric acute lymphoblastic leukemia by gene marrow to monitor residual disease in children with acute expression profiling. Cancer Cell 2002; 1: 133–143. lymphoblastic leukemia. Blood 2002; 100: 2399–2402. 128 Armstrong SA, Staunton JE, Silverman LB, Pieters R, den Boer ML, 148 van der Velden VHJ, Jacobs DCH, Wijkhuijs AJM, Comans-Bitter Minden MD et al. MLL translocations specify a distinct gene WM, Willemse MJ, Ha¨hlen K et al. Minimal residual disease expression profile that distinguishes a unique leukemia. Nat levels in bone marrow and peripheral blood are comparable in Genet 2002; 30: 41–47. children with T cell acute lymphoblastic leukemia (ALL), but not 129 Armstrong SA, Kung AL, Mabon ME, Silverman LB, Stam RW, in precursor-B-ALL. Leukemia 2002; 16: 1432–1436. Den Boer ML et al. Inhibition of FLT3 in MLL. Validation of 149 Pui C-H, Cheng C, Leung W, Rai SN, Rivera GK, Sandlund JT a therapeutic target identified by gene expression based et al. Extended follow-up of long-term survivors of childhood classification. Cancer Cell 2003; 3: 173–183. acute lymphoblastic leukemia. N Engl J Med 2003; 349: 640– 130 Cheok MH, Yang W, Pui C-H, Downing JR, Cheng C, Naeve CW 649. et al. Treatment-specific changes in gene expression discriminate 150 Vilmer E, Suciu S, Ferster A, Bertrand Y, Cave´ H, Thyss A et al. in vivo drug response in human leukemia cells. Nat Genet 2003; Long-term results of three randomized trials (58831, 58832, 34: 85–90. 58881) in childhood acute lymphoblastic leukemia: a CLCG-

Leukemia Novel approaches to treat leukemia RA Larson et al 2381 EORTC report. Children Leukemia Cooperative Group. Leukemia 171 Kedar A, Grow W, Neiberger RE. Clinical versus laboratory tumor 2000; 14: 2257–2266. lysis syndrome in children with acute leukemia. Pediatr Hematol 151 Manera R, Ramirez I, Mullins J, Pinkel D. Pilot studies of species- Oncol 1995; 12: 129–134. specific chemotherapy of childhood acute lymphoblastic leuke- 172 Lawrence J. Critical care issues in the patient with hematologic mia using genotype and immunophenotype. Leukemia 2000; 14: malignancy. Semin Oncol Nurs 1994; 10: 198–207. 1354–1361. 173 Veenstra J, Krediet RT, Somers R, Arisz L. Tumour lysis syndrome 152 Pui C-H. Toward optimal central nervous system-directed and acute renal failure in Burkitt’s lymphoma. Description of treatment in childhood acute lymphoblastic leukemia. J Clin 2 cases and a review of the literature on prevention and Oncol 2003; 21: 179–181. management. Neth J Med 1994; 45: 211–216. 153 Mahmoud HH, Rivera GK, Hancock ML, Krance RA, Kun LE, 174 Lorigan PC, Woodings PL, Morgenstern GR, Scarffe JH. Tumour Behm FG et al. Low leukocyte counts with blast cells in lysis syndrome, case report and review of the literature. Ann cerebrospinal fluid of children with newly diagnosed acute Oncol 1996; 7: 631–636. lymphoblastic leukemia. N Engl J Med 1993; 329: 314–319. 175 Jones DP, Mahmoud H, Chesney RW. Tumor lysis syndrome: 154 Bu¨rger B, Zimmermann M, Mann G, Ku¨hl J, Lo¨ning L, Riehm H pathogenesis and management. Pediatr Nephrol 1995; 9: et al. Diagnostic cerebrospinal fluid examination in children with 206–212. acute lymphoblastic leukemia: significance of low leukocyte 176 Kjellstrand CM, Campbell II DC, von Hartitzsch B, Buselmeier TJ. counts with blasts or traumatic lumbar puncture. J Clin Oncol Hyperuricemic acute renal failure. Arch Intern Med 1974; 133: 2003; 21: 184–188. 349–359. 155 Gajjar A, Harrison PL, Sandlund JT, Rivera GK, Ribeiro RC, 177 Mahmoud HH, Leverger G, Patte C, Harvey E, Lascombes F. Rubnitz JE et al. Traumatic lumbar puncture at diagnosis Advances in the management of malignancy-associated hyper- adversely affects outcome in childhood acute lymphoblastic uricaemia. Br J Cancer 1998; 77 (Suppl 4): 18–20. leukemia. Blood 2000; 96: 3381–3384. 178 Ten Harkel ADJ, Kist-Van Holthe JE, Van Weel M, Van der Vorst 156 Howard SC, Gajjar AJ, Cheng C, Kritchevsky SB, Somes GW, MMJ. Alkalinization and the tumor lysis syndrome. Med Pediatr Harrison PL et al. Risk factors for traumatic and bloody lumbar Oncol 1998; 31: 27–28. puncture in children with acute lymphoblastic leukemia. JAMA 179 Wolf G, Hegewisch-Becker S, Hossfeld DK, Stahl RAK. Hyper- 2002; 288: 2001–2007. uricemia and renal insufficiency associated with malignant 157 Gandhi V, Plunkett W, Rodriguez Jr CO, Nowak BJ, Du M, disease: urate oxidase as an efficient therapy? Am J Kidney Dis Ayres M et al. Compound GW506U78 in refractory hema- 1999; 34: E20. tologic malignancies: relationship between cellular pharmaco- 180 Dillman RO. Infusion reactions associated with the therapeutic kinetics and clinical response. J Clin Oncol 1998; 16: use of monoclonal antibodies in the treatment of malignancy. 3607–3615. Cancer Metastasis Rev 1999; 18: 465–471. 158 Druker BJ, Sawyers CL, Kantarjian H, Resta DJ, Reese SF, Ford JM 181 Bishop MR, Coccia PF. Tumor lysis syndrome. In: Abeloff MD, et al. Activity of a specific inhibitor of the BCR-ABL tyrosine Armitage JO, Lichter AS, Niederhuber JE (eds). Clinical Oncology kinase in the blast crisis of chronic myeloid leukemia and acute 2nd edn. New York: Churchill Livingstone, 2000, pp 750–754. lymphoblastic leukemia with the Philadelphia chromosome. N 182 Flombaum CD. Metabolic emergencies in the cancer patient. Engl J Med 2001; 344: 1038–1042. Semin Oncol 2000; 27: 322–334. 159 Kantarjian HM, Gandhi V, Kozuch P, Faderl S, Giles F, Cortes J 183 Arrambide K, Toto RD. Tumor lysis syndrome. Semin Nephrol et al. Phase I clinical and pharmacology study of clofarabine in 1993; 13: 273–280. patients with solid and hematologic cancers. J Clin Oncol 2003; 184 Ezzone SA. Tumor lysis syndrome. Semin Oncol Nurs 1999; 15: 21: 1167–1173. 202–208. 160 Navolanic PM, Pui C-H, Larson RA, Bishop MR, Pearce TE, Cairo 185 Jeha S. Tumor lysis syndrome. Semin Hematol 2001; 38 (4 Suppl MS et al. Elitekt-rasburicase: an effective means to prevent and 10): 4–8. treat hyperuricemia associated with tumor lysis syndrome: a 186 Hogan DK, Rosenthal LD. Oncologic emergencies in the patient Meeting Report, Dallas, Texas, January 2002. Leukemia 2003; with lymphoma. Semin Oncol Nurs 1998; 14: 312–320. 17: 499–514. 187 Kelly KM, Lange B. Oncologic emergencies. Pediatr Clin N Am 161 Hande KR, Hixson CV, Chabner BA. Postchemotherapy purine 1997; 44: 809–830. excretion in lymphoma patients receiving allopurinol. Cancer Res 188 Easton J, Noble S, Jarvis B. Rasburicase. Paediatr Drugs 2001; 3: 1981; 41: 2273–2279. 433–437. 162 Cunningham SG. Fluid and electrolyte disturbances associated 189 Laboureur P, Langlois C. Urate-oxydase d’Aspergillus flavus:I. with cancer and its treatment. Nurs Clin N Am 1982; 17: Obtention, purification, proprie´te´s. Bull Soc Chim Biol (Paris) 579–593. 1968; 50: 811–825. 163 Boles J-M, Dutel J-L, Briere J, Mialon P, Robasckiewicz M, Garre 190 Chanteclair G, Cartault F, Humbert J-C, Olive D, Neimann N. M et al. Acute renal failure caused by extreme hyperpho- Place de l’urate-oxydase dans la pre´vention de l’hyperurice´mie sphatemia after chemotherapy of an acute lymphoblastic en he´matologie infantile. Rev Int Pediatr 1974-1975; 50: 47–56. leukemia. Cancer 1984; 53: 2425–2429. 191 Leplatois P, Le Douarin B, Loison G. High-level production of a 164 Stapleton FB, Strother DR, Roy III S, Wyatt RJ, McKay CP, Murphy peroxisomal enzyme: Aspergillus flavus uricase accumulates SB. Acute renal failure at onset of therapy for advanced stage intracellularly and is active in Saccharomyces cerevisiae. Gene Burkitt lymphoma and B cell acute lymphoblastic lymphoma. 1992; 122: 139–145. Pediatrics 1988; 82: 863–869. 192 Pui C-H, Relling MV, Lascombes F, Harrison PL, Struxiano A, 165 Fleming DR, Doukas MA. Acute tumor lysis syndrome in Mondesir J-M et al. Urate oxidase in prevention and treatment of hematologic malignancies. Leukemia Lymphoma 1992; 8: hyperuricemia associated with lymphoid malignancies. Leuke- 315–318. mia 1997; 11: 1813–1816. 166 Stokes DN. The tumour lysis syndrome: intensive care aspects of 193 Leach M, Parsons RM, Reilly JT, Winfield DA. Efficacy of urate paediatric oncology. Anaesthesia 1989; 44: 133–136. oxidase (uricozyme) in tumour lysis induced urate nephropathy. 167 Hande KR, Garrow GC. Acute tumor lysis syndrome in patients Clin Lab Haematol 1998; 20: 169–172. with high-grade non-Hodgkin’s lymphoma. Am J Med 1993; 94: 194 Fam AG. Difficult gout and new approaches for control of 133–139. hyperuricemia in the allopurinol-allergic patient. Curr Rheumatol 168 Stucky LA. Acute tumor lysis syndrome: assessment and nursing Rep 2001; 3: 29–35. implications. Oncol Nurs Forum 1993; 20: 49–59. 195 Goldman SC, Holcenberg JS, Finklestein JZ, Hutchinson R, 169 Chasty RC, Liu-Yin JA. Acute tumour lysis syndrome. Br J Hosp Kreissman S, Johnson FL et al. A randomized comparison between Med 1993; 49: 488–492. rasburicase and allopurinol in children with lymphoma or 170 Dietz KA, Flaherty AM. Oncologic emergencies. In: Groenwald leukemia at high risk for tumor lysis. Blood 2001; 97: 2998–3003. SL, Frogge MH, Goodman M, Yarbro CH (eds). Cancer Nursing: 196 Pui C-H. Introduction – optimal treatment of malignancies Principles and Practice 3rd edn. Boston: Jones & Bartlett, 1993, associated with hyperuricemia. Semin Hematol 2001; 38 pp 821–824. (4 Suppl 10): 1–3.

Leukemia Novel approaches to treat leukemia RA Larson et al 2382 197 Patte C, Sakiroglu O, Sommelet D. European experience in the associated hyperuricemia in pediatric and adult patients: results treatment of hyperuricemia. Semin Hematol 2001; 38 (4 Suppl of a compassionate-use trial. Leukemia 2001; 15: 1505–1509. 10): 9–12. 200 Pui C-H, Mahmoud HH, Wiley JM, Woods GM, Leverger G, 198 Pui C-H. Urate oxidase in the prophylaxis or treatment of Camitta B et al. Recombinant urate oxidase for the prophylaxis or hyperuricemia: the United States experience. Semin Hematol treatment of hyperuricemia in patients with leukemia or 2001; 38 (4 Suppl 10): 13–21. lymphoma. J Clin Oncol 2001; 19: 697–704. 199 Pui C-H, Jeha S, Irwin D, Camitta B. Recombinant urate oxidase 201 Pui C-H. Rasburicase: a potent uricolytic agent. Expert Opin (rasburicase) in the prevention and treatment of malignancy- Pharmacother 2002; 3: 433–452.

Leukemia