Introduction
Acute myeloid leukemia (AML) is a malignant disease characterized by proliferation with maturation arrest of myeloid blasts in bone marrow and blood.1 Mounting evidence supports the notion that this disease is constituted by a group of distinct entities that are being recognized and categorized based on clinical, cytogenetic, and molecular features.2 Despite recent progress in our understanding of the leukemogenic mechanisms of AML and the use of intensive therapeutic approaches, the prognosis for these patients remains suboptimal, underscoring the critical need for novel diagnostic, risk-stratification, and therapeutic approaches. To achieve this goal it is imperative to dissect the biologic differences that determine the outcome of distinct clinical, cytogenetic, and molecular subsets of AML. The expectation is that the understanding of such differences will allow implementation of "personalized" molecularly targeted therapeutic programs according to the genetic make-up of the disease for each individual patient.
Recurrent structural and numerical chromosomal aberrations have been identified as one of the most important prognostic factors in AML.3 Approximately half of AML patients, however, present with a normal karyotype. Cytogenetically normal (CN) patients are typically classified in an intermediate-risk prognostic category.
Recently, it has been evident that the genomic heterogeneity observed within distinct cytogenetic groups can determine differences in outcomes.4 For example, several distinct molecular subsets of CN-AML have been identified to harbor specific genetic aberrations that are predictive of either adverse (e.g., the internal tandem duplication of FLT3 [FLT3 ITD], partial tandem duplication of MLL [MLL PTD], and overexpression of the BAALC or ERG genes) or favorable (i.e., mutations in the CEBPA or NPM1 genes) outcome. Several groups have also proposed the use of genome-wide analysis to identify specific gene expression signatures that could substitute or complement the prognostic and therapeutic significance of the aforementioned single gene markers, as has been done with success for childhood (ALL).5
MicroRNA Expression to Identify AML Subtypes
MicroRNAs are naturally occurring 19- to 25-nucleotide RNAs cleaved from 70 to 100 nucleotide hairpin precursors that hybridize to complementary mRNA targets and either lead to their degradation or inhibit their translation of the corresponding proteins.6,7 Initially discovered as regulators of normal cell homeostasis, microRNAs have recently been shown to be a new class of genes altered in several human malignancies and play an active role in malignant transformation.8 The genomic position of a large number of microRNAs is correlated with the location of cancer-associated genomic regions.9 Recently, functional and prognostic studies confirmed a role for microRNAs in hematologic malignancies and have been proposed as prognostic markers and therapeutic targets in AML.10
Following up on a pioneer study that showed AML could be distinguished from acute lymphoblastic leukemia based on gene expression profiling,11 Mi, et al. recently showed that the same could be accomplished by microRNA expression profiling.12 Subsequent studies have shown that microRNA signatures are capable of distinguishing not only between different leukemias but also between cytogenetic subtypes of AML.
Two microRNA expression profiling studies indicate that t(8;21), inv(16), and t(15;17) have unique microRNA expression signatures capable of distinguishing them from other subtypes of AML.13,14 As is common with expression profiling studies, there is not a perfect concordance between the signatures derived from the two studies, but there are some commonalities, including the up-regulation of microRNAs in genes transcribed from genes at the 14q32 region in t(15;17) and the down-regulation of miR-133a in t(8;21). Patterns of microRNA expression associated with t(11q23), trisomy 8, and CN-AML have also been reported.15
MicroRNA Expression Associated with Morphological and Molecular Characteristics
Aberrant microRNA expression patterns associated with morphological and molecular characteristics have been identified in AML, some within particular cytogenetic subtypes.
Up-regulation of miR-155 in patients with an internal tandem duplication (ITD) of the FLT3 gene has been independently reported by two groups.13,15,16 This observation fits well with the reported high-blast proliferation and poor survival duration in FLT3-ITD+ AML, since it has also recently been shown using a mouse model that miR-155 can drive granulocyte/monocyte expansion and result in pathological features characteristic of myeloid neoplasia.17
Mutations of the NPM1 gene also have a characteristic microRNA expression signature, including the up-regulation of miR-10a, -10b, and -196a.14,16 Interestingly, these microRNAs reside in the genomic cluster of homeobox (HOX) genes, and up-regulation of HOX genes is a prominent feature of NPM1 mutated gene expression signatures. Furthermore, Garzon, et al. reported down-regulation of miR-204 and -128a as additional features of the NPM1-associated microRNA signature and showed in cell line experiments that miR-204 inhibits expression of HOXA10 and MEIS1, two members of the HOX gene cluster.16
Up-regulation of mir-181a and -335 in AML patients carrying mutations of the CEBPA genes was observed in two studies,13,18 the latter solely focused on CN-AML. Another study showed that miR-124a is a target of CEBPA in vitro, and that miR-124a is epigenetically silenced in leukemic cell lines.19
In CN-AML, expression of miR-181a20 and miR-181b21 has been reported to be associated with French American British (FAB) morphological phenotype, with lower expression in M4 and M5 compared to M1 and M2. Also in CN-AML, miR-10a, -10b, and -196a-1 correlated with expression of HOX genes,20 consistent with the reported NPM1-associated gene signatures, that in turn was shown to be associated with HOX gene overexpression.22
MicroRNA Expression Associated with Clinical Outcome
Evidence is beginning to accumulate that microRNA expression is associated with clinical outcome in AML. Dixon-McIver, et al. reported that miR-9 and -let7b correlated with risk groups determined by cytogenetics, with higher expression in cytogenetic subgroups assigned intermediate and adverse risk.14 Garzon, et al. reported that, across all cytogenetic subgroups, miR-199a, -199b, -191, -25, and -20a, when over-expressed, adversely affected overall survival.15
In CN-AML with high risk molecular characteristics (i.e., FLT3-ITD+ or NPM1 wild-type, or both), we have recently identified a microRNA expression signature associated with event-free survival and validated its prognostic significance.23 The prominent feature of this signature was increased expression of miR-181a and -181b associated with decreased risk of an event. One of the particular features of this report was the integration of the genome-wide microRNA expression profile with gene expression in the attempt to identify genes that are micro-RNA-regulated and contribute to leukemogenesis in high molecular risk CN-AML. Expression levels of 452 genes significantly correlated with the prognostic microRNA signature. We showed that genes involved in mechanisms of innate immunity,24,25 including genes encoding Toll-like receptors (TLR2, TLR4, TLR8), interleukin-1ß, and its regulators (IL1B, CARD8, CARD12 [NLRC4], CARD15 [NOD2], ASC [PYCARD], CASP1), were overly represented in the microRNA-dependent gene expression signature. Some of these genes have already been reported to sustain growth and proliferation of AML blasts and may represent therapeutic targets.26-28
Conclusion
Altogether, the current studies suggest that microRNA expression is associated with cytogenetics, molecular and morphological alterations, and clinical outcomes in AML. We are only beginning to unravel the functional relationships that exist between microRNAs and these other factors, but there is great opportunity for gaining biological insights by studying altered microRNA expression in AML. Furthermore, it will be interesting to see if altered microRNA expression can serve as a valuable prognostic marker that adds information beyond the current collection of cytogenetic and molecular abnormalities.