Genetic Biomarkers in Acute Myeloid Leukemia

Will the Promise of Improving Treatment Outcomes Be Realized?

Jay Yang; Charles A Schiffer


Expert Rev Hematol. 2012;5(4):395-407. 

In This Article

Clues to the Pathogenesis of AML & Potential Drug Targets

Much of the new data about prognostic markers have not shed much light on the mechanisms of disease. Their mere discovery offers potential targets for novel therapies, but we do not know which, if any, of these are truly 'druggable'. It will be important to delineate whether new mutations are 'driver' or 'passenger' mutations. NPM1 is an example of the former that, at least conceptually, is more likely to have a prominent role in the pathogenesis of AML and, therefore, an important target for therapy. Despite the research focus that has been placed on NPM1 mutations, how they contribute to the AML phenotype is still very unclear. It is also unclear whether (or how) NPM1 itself increases sensitivity to chemotherapy or whether it is associated with the absence of as yet unknown other mechanisms of drug resistance.

Combinations Matter

Not only are there an ever-expanding number of recurring chromosomal and molecular changes in AML, such abnormalities often also overlap. It will be important to identify unique patterns of clustered genetic changes that may suggest biologically distinct types of AML with underlying unique signaling pathways.

In a rather short period of time, NPM1+ AML has been shown to be remarkably heterogeneous. Although NPM1 mutations are generally exclusive of CEBPA-DM and recurrent chromosomal abnormalities, there is an association with mutations in FLT3-ITD, FLT3-TKD, IDH, TET2 and DNMT3A. The favorable impact of NPM1 mutations is largely abrogated by concurrent FLT3-ITD mutations.[34] Coexisting TET2 mutations also appear to have an unfavorable impact[65] whereas DNMT3A, FLT3-TKD and ASXL1 mutations may have the opposite effect.[44,51] Studies analyzing the impact of IDH mutations in NPM1+ AML have had contradictory results.[74,96] Although seen infrequently, other concurrent karyotypic abnormalities do not appear to dramatically affect the prognosis of NPM1+ AML.[101]

Potentially more important is the observation that some mutations appear to be mutually exclusive. As an example, NPM1 mutations and CEBPA-DMs are rarely seen in the same patient. IDH mutations,[63] specifically at R132, are virtually never found together with TET2 mutations.[71] Both the gain-of-function IDH mutation and the loss-of-function TET2 mutation can result in DNA hypermethylation and impaired hematopoietic differentiation, suggestive of a common epigenetic pathway that may contribute to the development of AML.[102]

Drug Targets

There are multiple clinical trials using various FLT3 inhibitors in patients with FLT3 mutations, including lestaurtinib (CEP701), midostaurin (PKC412), sorafenib, sunitinib and quizartinib (AC220). Early studies have demonstrated single-agent activity,[103–105] including reductions in peripheral and bone marrow blast counts in patients with AML harboring the FLT3 mutation. However, most responses were incomplete and transient. FLT3 inhibitors have been safely combined with chemotherapy, prompting Phase III trials comparing standard therapies with or without FLT3 inhibitors. A randomized trial in relapsed AML patients showed no benefit to adding CEP701 to chemotherapy, but the results of other trials are still anticipated.[106] A number of hypotheses have been put forth to explain why this trial showed no improvement, including various leukemic resistance mechanisms, difficulty in achieving consistent serum levels of the drug and elevated levels of free-circulating FLT3, which may decrease the biologic activity of the drug on the receptor.[107] In addition, the different FLT3 inhibitors vary in target specificities and potencies, which may lead to differential clinical activities.[108]

Inhibitors of the KIT oncogene are also being studied, particularly in patients with CBF leukemias in which KIT mutations have been shown to be adverse. Inhibitors of KIT such as imatinib and dasatinib have shown preclinical activity.[109,110] Even CBF leukemias without KIT mutations have upregulated KIT activity, making the use of KIT inhibitors appealing for all patients with this type of leukemia.

Hypomethylating agents are approved in higher-risk myelodysplastic syndrome where azacitidine has shown a survival benefit compared to standard therapies.[111] These agents also have activity in frank AML.[112] Although these drugs have a novel mechanism of action, they do not target a specific genetic abnormality; rather, their efficacy is perhaps due to restoration of global abnormal methylation patterns in AML. There is interest in finding pretreatment markers that would predict for a response to hypomethylating agents.[113] For example, one could surmise that these agents may work preferentially in patients who carry DNMT3A or TET2 mutations, which can result in the hypermethylation of DNA.

A less conventional but promising target for therapy are miRNAs. miRNAs are small noncoding RNA sequences that hybridize to mRNA resulting in the regulation of gene expression.[82] Dysregulation of miRNAs occurs frequently in leukemia, implicating their role in leukemogenesis. Specific miRNAs, such as miR-155 and miR-29, have not only been associated with prognosis; they also work as functional oncogenes or as tumor suppressors. Synthetic oligonucleotides targeting the expression of specific miRNAs have been proposed as a novel class of therapy with potential anti-leukemic activity.

The experience with targeted agents to date has shown some promise of single-agent activity. However, none of these novel drugs have produced deep and durable responses, owing in part to other mutations that confer resistance to therapy. This suggests that combinations of novel agents with or without standard chemotherapy will be necessary to produce clinically significant results.