Tyrosine Kinase Inhibitor Use in Pediatric Philadelphia Chromosome–Positive Acute Lymphoblastic Anemia

Nowell and Hungerford described a small chromosome present in 2 patients with chronic myelogenous leukemia (CML) in 1960, which was eponymously named the Philadelphia chromosome (Ph) after their city.¹  In 1973, Janet Rowley discovered that this chromosome was the der(22) produced by the reciprocal translocation t(9;22)(q34;q11.2).²  After this, molecular studies showed that this translocation joined the chromosome 9 ABL1 gene, the human homolog of the Abelson murine leukemia virus, to a 5.8-kb region of chromosome 22 in a gene that came to be called the Breakpoint Cluster Region gene (BCR).^3,4  Later studies revealed that t(9;22) produced BCR-ABL1 fusion transcripts that encoded for a 210-kD chimeric protein with tyrosine kinase (TK) activity.⁵  BCR-ABL1–transgenic mice developed a myeloproliferative disorder reminiscent of human CML.^6,7  A major observation made in parallel was that an intact TK domain was absolutely required for BCR-ABL1 to transform cells in vitro.⁸ 

The observations summarized above opened the door to a new generation of targeted therapy for human cancer, because they suggested that one might be able to treat CML by developing a drug that could interfere with the BCR-ABL1 TK function. Screens of chemical libraries were under way to identify tyrosine kinase inhibitors (TKIs), which led to the identification of STI571 (imatinib mesylate), a potent and relatively specific BCR-ABL1 inhibitor that could kill CML cells in vitro.⁹  In 1998, a phase 1 dose-escalation study of imatinib in patients with CML refractory to other therapies was begun and yielded stunning results. Almost all (53 of 54; 98%) chronic-phase CML (CML-CP) patients who were resistant/intolerant to IFN attained a complete hematological response when treated with at least 300 mg/d of imatinib, and 60% had a decrease of Ph⁺ metaphases to < 35%.¹⁰  Toxicity was mild compared with standard cytotoxic drugs, and no maximally tolerated dose of imatinib was defined. The US Food and Drug Administration (FDA) approved imatinib for the treatment of CML in 2000, and subsequent randomized trials established it to be the best-available first-line therapy for patients with CML-CP.¹¹ 

The Ph chromosome also occurs in patients with acute lymphoblastic leukemia (ALL), but the genomic break points are typically different from those that occur in CML, producing a 190-kD fusion protein in most cases that has more potent transforming activity than p210 BCR-ABL1. Approximately 3% of children with ALL have Ph⁺ ALL, with approximately 90% having the “ALL-type” break points that produce p190 BCR-ABL1.^12,13  Ph⁺ ALL accounts for approximately 15%-25% of adult ALL cases, with rates starting to increase in adolescence.¹⁴ 

Until recently, Ph⁺ ALL was one of the most challenging subtypes of pediatric ALL because it was associated with a dismal prognosis. In 2000, the Ponte di Legno childhood ALL consortium reported a retrospective, multicenter review of outcomes of 326 Ph⁺ ALL patients younger than 20 years of age who were diagnosed between 1986 and 1996 and treated with a variety of different protocols by cooperative groups and major pediatric cancer centers. The outcome was quite poor, with a 7-year event-free survival (EFS) rate of 25% and an overall survival (OS) rate of 36%.¹²  Whereas risk factors that are favorable in other subtypes of pediatric ALL (eg, younger age, lower initial WBC, and favorable early morphologic response to therapy) also predicted for better outcome in Ph⁺ ALL, the outcome of these “good-risk” Ph⁺ ALL patients was still poor. For example, Ph⁺ patients with an initial WBC < 25 000/mL still had an EFS of only ∼ 40%. Matched related, but not unrelated donor hematopoietic stem cell transplantation (HSCT) produced better outcomes than chemotherapy alone. Recently, the Ponte di Legno group reported retrospective analyses on the outcome of 610 Ph⁺ ALL patients younger than 18 years of age who were diagnosed in 1995-2005 and treated by a variety of cooperative groups and pediatric cancer centers with different treatment regimens, none of which included TKI therapy. The results were modestly better than those from the prior era, with a 7-year EFS of 31% and an OS of 44%.¹⁵  Age, initial WBC, and early response remained prognostic. In this study, HSCT with either matched related or unrelated donors produced better results than chemotherapy alone, but outcomes were still poor even with HSCT. When analysis was limited to the 89% of patients who attained a complete remission after induction chemotherapy and disease-free survival was measured from the median time from complete remission to HSCT (5.1 months), the 5-year disease-free survival and OS rates were 43.5% and 54.0% after HSCT compared with 34.2% and 48.3% after chemotherapy treatment, respectively.

Single-agent imatinib is effective in Ph⁺ ALL, but responses are generally transient in patients with advanced-stage disease. Promising results were reported with imatinib plus chemotherapy in adults with Ph⁺ ALL, but the imatinib exposure was generally intermittent and studies for younger adults have typically used HSCT for consolidation therapy.^16–18  The COG AALL0031 trial (2002-2006) incorporated imatinib, starting after completion of induction therapy, into a very intensive chemotherapy regimen in a stepwise fashion.¹⁹  Patients in the final dose cohort of AALL0031 (cohort 5) received continuous treatment with imatinib 340 mg/m²/d from the start of consolidation, with the drug administered on a 2-week on/2-week off schedule for the last year of maintenance therapy. Patients treated with chemotherapy plus imatinib had no signs of increased toxicity compared with Ph⁻ ALL patients treated in the same study with identical chemotherapy without imatinib. The 3-year EFS was 80% for patients in cohort 5, which was more than double the EFS rate (35% ± 4%; P < .0001) of historical controls treated in the pre-imatinib era. There was no advantage of HSCT, with the 3-year EFS similar for patients in cohort 5 treated with chemotherapy plus imatinib, related-donor HSCT, or off-protocol therapy unrelated-donor HSCT.^19,20  Whereas these results are based on relatively small patient numbers, they have been stable with longer follow-up, and suggest that the addition of imatinib to intensive chemotherapy can dramatically improve the outcome of children with Ph⁺ ALL.

Starting in 2004, the major European pediatric cooperative groups have conducted the European Intergroup Study on Post Induction Treatment of Philadelphia Positive Acute Lymphoblastic Leukemia with Imatinib (EsPhALL) study for children with Ph⁺ ALL. EsPhALL took a different approach from the COG trial, and originally the poor-risk (prednisone-poor response with ≥ 1000 leukemia cells/microliter remaining after 7 days of prednisone and a single intrathecal methotrexate administration, and/or > 25% marrow blasts at induction day 15, and/or ≥ 5% marrow blasts at induction day 21, and/or failure to enter remission at the end of induction therapy) Ph⁺ ALL patients were nonrandomly assigned to chemotherapy plus imatinib, and the remaining good-risk patients were randomized to receive chemotherapy with or without imatinib. However, this trial had much more intermittent exposure to imatinib than that given in COG AALL0031 cohort 5, and by study design, > 70% of patients underwent HSCT. In 2010, the EsPhALL trial was amended to halt the ±imatinib randomization and to start continuous imatinib treatment for all patients at day 15 of induction therapy. In addition, the indications for HSCT have evolved such that many more patients will be treated with chemotherapy plus imatinib rather than HSCT. This trial will complete its planned accrual in the next 1-2 years.

Some patients with CML-CP and many with advanced-stage CML or Ph⁺ ALL develop imatinib resistance, most commonly due to outgrowth of cells with point mutations in the Abl kinase domain that interfere with imatinib binding.²¹  Two second-generation Abl class TKIs, dasatinib and nilotinib, have been FDA approved based on trials demonstrating their superiority in meeting early-response end points compared with imatinib in adults with CML-CP.^22,23  Nilotinib (formerly termed AMN107) is 10- to 30-fold more potent than imatinib against BCR-ABL1 mutants resistant to imatinib, with the prominent exception of the T315I mutation; however, like imatinib, it inhibits a relatively limited number of TKs, including the Abl-class, KIT, and PDGFR TKs.²⁴  The SRC kinases are closely related to Abl kinases, but imatinib and nilotinib do not inhibit SRC kinase activity. Dasatinib was originally developed as an SRC kinase inhibitor, and was later found to be a potent BCR-ABL1 kinase inhibitor that was active against most imatinib-resistant BCR-ABL1 mutants, again, with the exception of T315I.²⁵ 

Relatively little data are available on the use of nilotinib in Ph⁺ ALL. There has been more interest to date in dasatinib because it was shown in murine models that signaling through the SRC family kinases HCK, LYN, and FGR is required for the development of Ph⁺ ALL but not CML.²⁶  In addition, dasatinib has much better CNS penetration than imatinib (and likely nilotinib), so it may have advantages in ALL, a disease in which the CNS is known to be an important sanctuary site.²⁷ 

In 2010, Ravandi et al published the first report of an M.D. Anderson Cancer Center trial of 35 adults (median age, 53 years; range, 21-79 years) treated using hyper-CVAD (cyclophosphamide, vincristine, Adriamycin, and dexamethasone) with dasatinib given the first 14 days of each 28-day course.²⁸  These results were quite impressive, with 94% of patients entering complete remission (the other 2 patients had good responses but died before completing induction therapy) and a 2-year OS rate of 64% and EFS of 57%. Only 4 of these patients underwent HSCT in the first complete remission. Longer follow-up is needed to fully assess these data, but they suggest that the routine use of HSCT might also be reconsidered in adults with Ph⁺ ALL.

The COG developed a successor study to AALL0031 (AALL0622) that uses essentially the same chemotherapy regimen as AALL0031 with dasatinib 60 mg/m²/d replacing imatinib. In this study, dasatinib is started at day 15 of multiagent induction therapy, rather than waiting until after induction therapy is completed, as was done in AALL0031. This has been an important change. Approximately 10% of Ph⁺ ALL patients failed to enter complete remission after 4 weeks of chemotherapy alone in AALL0031,¹⁹  a rate similar to historical control data.^12,15  With > 50 patients enrolled to date, there have been no induction failures in COG AALL0622 (unpublished observations, April 2011). Similar to AALL0031, AALL0622 was designed as a dose-escalation trial to ensure that adding dasatinib to intensive chemotherapy did not lead to excess toxicity. The first cohort of patients received dasatinib for 2 weeks of every 3- to 4-week chemotherapy block. That phase of the trial has now been completed successfully, and newly enrolled patients are receiving continuous dasatinib treatment.

There are several important unanswered questions in pediatric Ph⁺ ALL. What is the optimal TKI to combine with chemotherapy? How intensive a chemotherapy backbone is needed? What is the role of HSCT in Ph⁺ ALL? As summarized above, there are theoretical reasons to favor dasatinib over imatinib or nilotinib, but no direct comparisons are available at this time. Given the small number of pediatric Ph⁺ ALL patients available to study and the rapid evolution of therapeutic options in this field, such a comparison may never be performed in children. Therefore, one may have to extrapolate data on the optimal TKI from adult CML to pediatric ALL, which is intrinsically unsatisfying and perhaps risky. It might be possible to design a randomized TKI study in children and young adults with Ph⁺ ALL up to the age of 30-40 years. Because there is no significant experience to date with the intensive COG AALL0031/AALL0622 or EsPhALL chemotherapy regimens in patients older than 21-22 years, it is likely that such a trial would need to use a less-intensive chemotherapy backbone. It will also be important to determine whether the various TKIs have different side effect profiles when combined with chemotherapy, because the optimal TKI might be the one that can be used most intensively with combination chemotherapy.

There are several reasons to consider the intensity of the chemotherapy backbone in Ph⁺ ALL. At this time, it is not known how intensive a backbone is needed. Were the highly promising results from COG AALL0031^19,20  due to the chemotherapy, the TKI, or both? It might be better to use less-intensive chemotherapy, maximize TKI therapy, and reserve “room” for additional new drugs that might potentiate TKI therapy and/or overcome resistance. Very encouraging results have been obtained in elderly adults with Ph⁺ ALL who were treated with imatinib plus relatively little chemotherapy.²⁹  Most importantly perhaps, as cure rates for Ph⁺ ALL improve and approach those of other subtypes of childhood ALL, one must begin to place increased emphasis on issues of long-term toxicity. Efforts to diminish the use of HSCT are generally based on the avoidance of early mortality and long-term toxicity. The AALL0031 and AALL0622 chemotherapy backbones are quite intensive, and many males treated with these regimens will likely be infertile. Such outcomes are not acceptable for a disease in which cure rates might exceed 75%.

What is the role of HSCT in pediatric Ph⁺ ALL today? There is no reason to advocate that all children with Ph⁺ ALL be treated with HSCT early after obtaining remission. A majority will almost certainly be cured without HSCT. Can the ones who won't be cured with chemotherapy plus TKI be identified reliably? We know that age, initial WBC, and early response are prognostic in Ph⁺ ALL; current COG and EsPhALL trials focus on using early minimal residual disease response to allocate patients to treatment with chemotherapy plus TKI or HSCT. Allocation to HSCT based on minimal residual disease response is a logical extrapolation of data from other subsets of childhood ALL, but has not yet been validated prospectively in those with Ph⁺ ALL.^30,31  It would not be surprising if treatment evolves such that very few children with Ph⁺ ALL undergo HSCT in first remission. For those who do receive HSCT, should a TKI be used after HSCT? If so, which one and for how long?

With improved outcomes, the rarity of Ph⁺ ALL makes it very hard for even the large North American (COG) or European (EsPhALL) cooperative groups to conduct informative clinical trials in this uncommon subset of pediatric ALL. Because of this, COG and EsPhALL are working together to develop a joint trial for children with Ph⁺ ALL, which they hope to open to patient accrual in late 2011 or early 2012. It is hoped that this trial will serve as a paradigm for international collaboration in other small, high-risk subsets of pediatric ALL.

High-throughput genomic studies and next-generation sequencing projects are leading to major new discoveries about the genomic landscape of childhood cancer.³²  One important discovery has been the recognition of a subtype of ALL cases with a gene-expression profile similar to that of Ph⁺ ALL and with a poor outcome.^33–35  This profile suggested that activated TK signaling could be occurring in these cases, prompting searches for underlying mutations. Approximately half of these Ph-like ALL cases have cryptic translocations and interstitial deletions that activate expression of CRLF2, a cytokine receptor not previously known to be involved in the pathogenesis of ALL.^36–38  Among cases with CRFL2 lesions, approximately half have mutations that target the kinase or pseudokinase domains of one of the 3 JAK genes, especially JAK2.³⁹  Several small-molecule inhibitors of JAK kinase activity are in clinical trials for adults with myeloproliferative neoplasms characterized by JAK2 V617F mutations,⁴⁰  and the COG has developed a phase 1 trial of one of these inhibitors (INCB018424) in refractory pediatric solid tumors and leukemias with the hope of developing a trial testing chemotherapy plus JAK2 inhibitor therapy, which is analogous to the approach taken in Ph⁺ ALL. Recent transcriptome-sequencing studies have shown that a significant majority of the Ph-like ALL cases that lack CRLF2 lesions contain chromosome rearrangements and mutations that dysregulate cytokine receptor and kinase signaling.⁴¹  The targets of these rearrangements include ABL1, JAK2, and PBGFRB, all of which could be targeted by available TKIs. Each of these rearrangements is uncommon individually, but taken together, the Ph-like ALL subset may be twice as common as Ph⁺ ALL. International collaboration will be needed to test new therapies in these rare subsets of pediatric ALL.

Conflict-of-interest disclosure: The author has consulted for, has equity ownership in, and is on the board of directors or an advisory committee for Bristol Myers Squibb. Off-label drug use: TKI therapy for pediatric ALL is not approved at this time.

Stephen P. Hunger, MD, Center for Cancer and Blood Disorders, The Children's Hospital, 13123 East 16th Ave B115, Aurora, CO 80045; Phone: (720) 777-8855; Fax: (720) 777-7279; e-mail: hunger.stephen@tchden.org.