Unraveling the “Known Unknowns”: Lessons and Reflections from the New Directions in Leukemia Research 2012 Conference

Statement of Purpose

New Directions in Leukemia Research (NDLR) is the only biennial international conference focused on leukemia. The goals of these meetings are to bring together scientists and clinicians to discuss and debate current concepts in our understanding of the molecular basis of leukemia, emerging paradigms and breakthroughs at the forefronts of leukemia research, and new therapies emerging in the clinic. There were 8 international speakers, 10 national speakers, 17 selected oral presentations, and 53 posters. In addition, the meeting hosted an early career forum, sponsored by the Leukemia Foundation of Australia, where some of the international and national speakers shared the trials and tribulations of their journey in science with 60 early-career researchers.

Success in CML and Why Current Targeted Therapies Don't Suit Everybody

Tim Hughes (Centre for Cancer Biology, Adelaide, Australia) gave the NDLR 2012 Special Oration reflecting on what imatinib has taught us about chronic myeloid leukemia (CML) and targeted therapy. He discussed the advancements that have enabled CML to change from an invariably fatal disease to a disease in which the 5-year survival rate exceeds 85%. These include the development of a highly sensitive and quantitative test that monitors disease burden, a predictive test that examines the intrinsic sensitivity of patients to tyrosine kinase inhibitor therapy, and an assay to determine the functionality of drug transporters, in particular OCT-1 (major active influx pump for imatinib). Importantly, Tim Hughes and Deb White (Adelaide, Australia) emphasized that, although tyrosine kinase inhibitor (TKI) therapy has improved the outlook for CML patients, not all respond well, and that failure to respond is largely related to variability in the intrinsic sensitivity of patients to kinase inhibitors; this can now be assayed for at diagnosis. White showed that understanding the underlying biology of individual patients could, in turn, provide previously unexplored insight into resistance mechanisms, paving the way for the development of combinational strategies to overcome primary resistance, which may further reduce development of secondary resistance. Hughes discussed sustained disease control in some patients who had stopped taking imatinib. His comments may link into an observation made by Jerry Radich (Fred Hutchinson Cancer Research Center, Seattle, WA) that 20% of all individuals have constitutively activated IFN-α genes, which might explain why 10% to 25% of CML patients respond to IFN-α treatment and the fact that the response rate to IFN-α is determined by the underlying activation status of the immune system. Collectively, these observations indicate a hitherto unexplored role for the immune system in drug-induced disease control and eradication.

Acute Myeloid Leukemia–Present and Future

Allan Burnett (School of Medicine, Cardiff University, Cardiff, UK) gave an overview of treatment in acute myeloid leukemia (AML), highlighting the mammoth task of evaluating both currently used treatments and strategies for evaluating emerging treatments. The use of large randomized trials in combination with cytogenetic and mutation status has brought improvement of outcomes for several groups of young AML patients. However, significant challenges posed are the heterogeneity of AML, the lack of treatments to improve the outcome of elderly patients, and the inability of the current trial designs to test new compounds rapidly. Burnett outlined the “pick a winner trial design” that enables the rapid and progressive incorporation of new compounds into the current low-dose Ara-C standard-of-care treatment for elderly patients and allows for agents without promise to be quickly dropped (1). For patient subgroups with poor survival, this may transform clinical trial approaches in an era in which new treatments are being developed at an increasing rate.

The genomics era of cancer has been led by leukemia, with the increasing number of whole-genome sequences identifying a plethora of new mutations and initiating new research directions. Thus, it was salient for Ross Levine (Memorial Sloan-Kettering Cancer Center, New York, NY) to discuss the profiling of many of these mutations in a large cohort of AML patients (E1900 phase III clinical trial) and how outcome classifications have changed. Internal tandem duplications of the FLT3 tyrosine kinase receptor were confirmed as the most frequent alteration (30%) and associated with reduced overall survival, emphasizing its continuing position as an important target for therapy. Importantly, integrating mutational analysis with current cytogenetics identified that mutations in TET2, ASXL1, MLL-PTD, PHF6, and DNMT3A could reclassify patients, identified as intermediate risk by cytogenetics alone, as those who will have an adverse outcome (2). Epigenetic modifiers were of particular focus as the mutual exclusivity of TET2 and IDH1/2 mutations indicated a shared mechanism of action. This overlapping function was explored further by both Levine and Ari Melnick (Weill Cornell Medical College, New York, NY), who used patient data and animal models to delineate a pathway by which neomorphic mutations in IDH1/2 proteins lead to increased levels of 2-hydroxy-glutarate in the cell and that subsequently inhibit TET2 function, leading to widespread hypermethyation. In addition, Melnick used new techniques of methylation profiling on AML patient cohorts to show that both aberrant hypermethylation (IDH mutations) and hypomethyation (MLL rearrangements) can be associated with the pathologic state, showing that the MLL-AF9 fusion protein directly leads to a hypomethylated state. Moreover, he hinted at studies in progress looking at modules of hypermethylated genes that may lead to chemoresistance in models of leukemia as well as the changing epigenetic landscape in relapsed leukemia. Undoubtedly, these findings will lead the way in identifying new treatment strategies for AML.

Acute Lymphoblastic Leukemia in the Genomics Era–Not Lost in Translation

Charles Mullighan (St Jude Children's Research Hospital, Memphis, TN) dissected the genetics of early T-cell precursor acute lymphoblastic leukemia (ETP-ALL) and discussed methods that have identified mutations that fall into functional clusters, such as activating mutations affecting cytokine receptor and RAS signaling (NRAS, KRAS, JAK1/3, and IL7R), mutations disrupting normal hematologic development (GATA3, RUNX1, ETV6, and IKZF1), and mutations in histone-modifying genes (EZH2, SUZ12, EED, and SETD2) with novel recurrent mutations found in DNM2, ECT2L, and RELN (3). Importantly, many of these genes are also mutation targets in myeloid malignancy, and transcriptional profiling highlighted a similarity between ETP-ALL and high-risk AML as well as normal hematopoietic stem cells (HSC). Such detailed profiling has defined ETP as a “stem cell leukemia” and emphasizes the importance of this information in diagnosis, classification, and downstream treatment approaches. Kathryn Roberts (St Jude Children's Research Hospital) extended this approach in Ph-like high-risk pre-B-ALL in children to identify activating mutations in the interleukin 7 (IL-7) receptor, as well as fusion proteins involving ABL1, platelet-derived growth factor receptor B, and JAK. Roberts further showed the therapeutic effectiveness of several small-molecule inhibitors using in vitro and in vivo murine models of these novel fusions.

This theme of better drug selection and better patient outcome based on understanding the underlying genetics of the disease was extended by Ross Dickins (Walter & Eliza Hall Institute, Melbourne, Australia). Building on NDLR 2008 where Charles Mullighan described the essential role of Ikaros in pre–B-cell development and the very poor outcome of patients with these deletions, Dickins and his team sought to understand why Ikaros deletions predict for treatment failure. Glucocorticoids are first-line treatments for ALL patients, with glucocorticoid resistance being linked to poor outcome. Ikaros restoration is associated with restored glucocorticoid sensitivity resulting in apoptosis-driven cell death. This elegant study shows how an understanding of drug and cellular metabolism pathways that are affected by genetic lesions leads to rational drug choices for patients and improved outcomes.

Novel Small-Molecule Inhibitors–More New Kids on the Block!

Understanding the genetic defects that drive leukemogenesis has facilitated the development of novel small-molecule inhibitors. Several speakers used this rationale to develop novel small-molecule inhibitors that are currently in preclinical and early-phase clinical trial stages of development. Some of these new approaches include (i) BEZ235, a dual DDR kinase/mTOR inhibitor, which has been shown to induce apoptosis in a p53-independent manner leading to enhanced BMF expression and clearance of the leukemic burden in animal models (Jake Shortt, Peter MacCallum Cancer Centre, Melbourne, Australia); (ii) blocking the interaction between MYB and p300 to shut down the transformation process while leaving normal cells intact (Thomas Gonda, Diamantina Institute, University of Queensland, Brisbane, Australia); (iii) CX-5461 inhibition of Pol 1 transcription, as sensitivity to this agent is dependent on wild-type p53 expression, which is maintained in the absence of Arf. Due to changes in nucleolar morphology as a consequence of the inhibition of ribosomal RNA gene transcription, cells die via p53-induced apoptosis and caspase-3 is activated within 2 hours after treatment subsequent to increased p53 and p21 protein levels. Treatment with CX-5461 leads to increased survival in the Eu-Myc lymphoma and the MLL-ENL leukemic model (Megan Bywater, Peter MacCallum Cancer Centre, Melbourne, Australia); (iv) a novel dual inhibitor of FLT3 and Aurora kinase (CCT241736) has been validated in an in vitro model of FLT3 resistance. Oral administration of this drug in preclinical animal models shows that CCT241736 overcomes the resistance to the selective FLT3 inhibitor, AC220, caused by secondary FLT3 mutations (Andrew Moore, Institute of Cancer Research, London, UK); and (v) inhibition of FLT3 signaling, using the FLT3 inhibitor AC220. Wally Langdon (University of Western Australia, Perth, Australia) has previously shown that a mouse MPN/AML model expressing a RING finger mutant of c-CBL (C379A), which abolishes E3 ubiquitin ligase activity, is dependent on intact FLT3 signaling for disease progression (4). He showed that disease symptoms were reversed with reduced numbers of FLT3⁺LSK cells, and decreased splenomegaly and extramedullary invasion of myeloid cells. In addition, inhibitor treatment was associated with reduced proliferation of cells without causing apoptosis, and was itself reversible upon withdrawal of drug with rapid recovery of FLT3⁺ proliferative cells. These observations are similar to those seen with TKI treatment in CML patients, for whom long-term imatinib treatment provides disease control but not cure. Cure is likely to require combination strategies to destroy the leukemic stem cells (LSC).

Cure: A Balance between Compliance and Quality of Life? Is There a Role for the Immune System?

Catriona Jamieson (Moores UC San Diego Cancer Center, San Diego, CA) spoke about novel therapies to treat myleoproliferative neoplasms (MPN) that were developed as a result of increased understanding of normal hematopoiesis. MPN granulo-myelo-progenitors acquire β-catenin and undergo reprogramming to acquire stem cell characteristics. Polycythemia vera HSCs have the JAK2V617F mutation, and the selective JAK2/FLT3 inhibitor, TG101348, decreases mutant burden and leukocytosis, reducing the level of myelofibrosis after 12 months of treatment, but it also causes anemia. The oral Sonic Hedgehog inhibitor, PF04449913, has been successfully used to treat AML in the elderly, reducing marrow blasts and decreasing fibrosis. Although these new inhibitors are effective, they require prolonged exposure to mitigate symptoms associated with MPN. If these therapies are to be effective, particularly in diseases of the elderly, they need to be faster acting and not compromise quality of life.

Only one presentation at NDLR 2012 addressed the role of the immune system in the treatment of leukemia. David Ritchie (Peter MacCallum Cancer Centre, Melbourne, Australia) discussed the potency of natural killer (NK) cells, which kill leukemic cells via activating receptors and the release of cytotoxic granules that induce FasL/TRAIL and IFN-γ–mediated apoptosis. Ritchie discussed immune editing of tumors showing that normal myeloid cells have NK ligands and, through analysis of paired diagnosis and relapse samples, can show further that NK-sensitive AML cells at diagnosis are absent in relapse.

Myelodysplasia–Common but Misunderstood?

Myelodysplasia (MDS), although predisposing to AML and sharing many common genetic lesions, is a very different entity to AML at the cellular level, with apoptosis being a prominent feature in early-stage disease. Again, the theme of identification and understanding the genetic drivers of disease to develop models and new therapies was discussed. Hamish Scott (Centre for Cancer Biology, Adelaide, Australia) showed the power of familial genetics describing GATA2 as a target of mutation in familial MDS/AML (5). Louise Purton (St Vincent's Institute, Melbourne, Australia), generated a novel mouse model of MDS on the basis of deregulated alternative splicing of the Hoxa1 gene. She is using this model to further understand how MDS occurs, and to identify better treatments for this disease. David Curtis (Australian Centre for Blood Diseases, Melbourne, Australia) used the NHD13 transgenic mouse, which is characterized by the expression of the NUP98-HOXD13 to investigate mechanisms and implications of apoptosis in MDS. NHD13 mice develop MDS, AML, and ALL. The disease is characterized by prominent apoptosis in myeloid progenitors, and premalignant progenitors have increased cycling and double-strand breaks. Apoptosis in the NHD13 mouse model does not require TNF or FASL, and premalignant progenitors do not have increased caspase 8. Early MDS progenitors have reduced BCL2 expression. Expression of BCL2 blocks apoptosis of premalignant progenitors and rescues normal progenitor growth. Blocking apoptosis rescues blood counts and dysplasia but the mice retain a low platelet count. Apoptosis promotes proliferation of premalignant cells in MDS, potentially through secretion of factors such as Wnts or prostaglandins. Blocking apoptosis prevents AML transformation by restoring quiescence and reducing DNA damage, indicating that blocking apoptosis in MDS may lead to disease control.

HSC and LSC in the Niche–Time to Leave Home

LSCs, as the presumptive source of relapse in leukemia, are masters of “laying low” by maintaining quiescence and localizing to sanctuary sites in the marrow, thereby evading therapy. The stromal cells of the marrow play an important role in facilitating this therapy resistance and, therefore, dissecting out the stem cell–stromal interaction in normal and neoplastic hematopoiesis is essential to target LSCs, as recently illustrated by the observation that disruption of the CXCR4/CXCL12 axis with CXCR4 antagonists sensitizes LSCs to conventional chemotherapy. Following and extending this logic, the application of a CXCR4 antagonist, AMD3100, to mobilize HSCs was examined by Linda Bendall (University of Sydney, Sydney, Australia), who described how the bioactive lipid S1P forms a gradient between tissue and blood to direct lymphocyte trafficking. Using both pharmacologic inhibition of S1P signaling and conditional mouse knockouts of the S1P receptor S1P1 and sphingosine kinase, Bendall and colleagues have shown that S1P is required for HSC egress from the marrow after mobilization with CXCR4 antagonists (6). As predicted by this model, the use of S1P agonist SEW2871 enhanced HSC mobilization. The dual use of CXCR4 antagonists with S1P agonists may, therefore, be an efficient alternative to the use of granulocyte colony-stimulating factor for stem cell mobilization.

Jean-Pierre Levesque (Mater Medical Research Institute, Brisbane, Australia) discussed the microenvironment of the normal and leukemic marrow. He showed an association between the degree of perfusion of and hypoxia in the bone marrow, with HSCs capable of serial repopulation located at the poorly perfused endosteum. Pharmacologic stabilization of the oxygen-labile transcription factor hypoxia-inducible factor 1α (HIF-1α) increases HSC quiescence, further showing that hypoxia is a quiescence signal for HSCs via HIF-1–dependent mechanisms. Leukemic hyperproliferation alters the marrow environment by increasing hypoxia and stabilizing HIF-1α protein. Leukemic cells have altered hypoxia signaling, which enables them to overcome the hypoxia constraint of the leukemic marrow and proliferate in a very hypoxic environment. This results in a stem cell niche hijack by leukemic cells. These observations are in support of the use of hypoxia-induced prodrugs to target LSCs.

Disclosure of Potential Conflicts of Interest

D.L. White has a commercial research grant from Novartis Oncology and Bristol-Myers Squibb.

Authors' Contributions

Conception and design: D.L. White, R. D'Andrea, A.M. Rice

Writing, review, and/or revision of the manuscript: D.L. White, A.L. Brown, R. D'Andrea, A.M. Rice