Background
Myocyte enhancer factor 2 (MEF2) proteins, composed of four family members in vertebrates, are transcription factors that were initially studied in the control of muscle development [
1]. In particular, gene deletion studies in mice identified essential functions of
MEF2C in cardiac myogenesis and right ventricular development [
2]. However, subsequent studies have indicated that
MEF2C plays a much broader biological role and is involved in the function and generation of tissues other than cardiac and skeletal muscle, including bone development and osteoclast-mediated bone resorption, neuronal development, and craniofacial and melanocyte development [
3].
Increasing evidence also suggests an important role of
MEF2C in the normal hematopoietic system, particularly for the production of immature and mature lymphoid cells and as a modulator of the cell fate decision between monocyte and granulocyte differentiation [
3‐
6]. This is indicated by genetic studies in mice showing that
Mef2c deficiency is associated with reduced levels of monocytes in response to cytokines [
4] as well as profound defects in the production of B cells, T cells, natural killer cells, and common lymphoid progenitor cells, as well as enhanced myeloid output [
5]. In turn, constitutive expression of
Mef2c in the bone marrow results in increased monopoiesis at the expense of granulopoiesis [
4]. In human acute myeloid leukemia (AML) cell line models, 1,25-dihydroxyvitamin D3 induces monocytic differentiation and CD14 expression, an effect that is mediated through activation of
MEF2C signaling via regulation of CCAAT-/enhancer-binding protein alpha (CEBPA) [
6]. Consistent with these central functions,
MEF2C has been found to be aberrantly expressed in subsets of T cell acute lymphoblastic leukemia (T-ALL) and in early thymocyte precursor (ETP) T-ALL in particular, an aggressive leukemia that tends to be refractory to chemotherapy and shares genetic features with AML [
7‐
10]. In AML,
MEF2C has been found to be overexpressed in distinct molecular subsets of adult onset AML, including mixed lineage leukemia (
MLL) gene-rearranged and ectropic virus integration site 1 (
EVI1)-overexpressing leukemias [
11]. Models of both MLL and EVI1 leukemias have been, and continue to be, instrumental in our understanding of fundamental principles of leukemogenesis and the identification of pathways that confer tumor aggressiveness and resistance to chemotherapy [
12‐
21]. Functional studies using mouse leukemia models demonstrate that
Mef2c is a potent oncogene, causing fully penetrant AML in cooperation with
SOX4 [
11,
22,
23]. In addition,
Mef2c is required for the growth of mouse leukemias induced by
MLL-AF9 [
11].
Together, these studies suggest that
MEF2C participates in key molecular mechanisms of AML pathogenesis and could serve as a marker of poor-risk AML and, therefore, have prognostic significance. Here, we tested this hypothesis by retrospectively quantifying
MEF2C expression in pretreatment bone marrow specimens and by associating
MEF2C expression level with disease characteristics and outcome in participants of the Children’s Oncology Group (COG) AML protocol, AAML0531 (NCT00372593). AAML0531 was a multicenter, randomized phase 3 study, which found that the addition of gemtuzumab ozogamicin to intensive chemotherapy improved the event-free survival (EFS) through reduction of the relapse risk (RR) relative to intensive chemotherapy alone in patients aged <30 years with newly diagnosed de novo non-APL AML, excluding those with bone marrow failure syndromes, juvenile myelomonocytic leukemia, or Down syndrome (if ≤3 years of age) between 2006 and 2010 [
24].
Discussion
Recent studies have highlighted a possible role of
MEF2C in the molecular pathogenesis and therapy response of AML [
3]. Using over 750 pretreatment bone marrow specimens from pediatric patients enrolled in a recent cooperative group phase 3 trial, ours is the first study to quantify
MEF2C mRNA abundance by RT-PCR and comprehensively examine the relationship between
MEF2C expression and disease characteristics as well as treatment outcome in pediatric AML. The findings from these investigations support three main conclusions. First,
MEF2C is widely expressed in pediatric AML, with relative levels that vary considerably (>3000-fold) across bone marrows of patients with active disease. Second, high
MEF2C expression is associated with adverse treatment outcome in pediatric AML. Specifically, in our cohort, patients with the highest relative
MEF2C expression (4th quartile) less likely achieved a CR after one course of chemotherapy than the other patients; they also had an inferior OS and EFS and higher RR than patients within the lower 3 quartiles of
MEF2C expression. And third, high
MEF2C expression is associated with several adverse-risk features. Specifically, in participants of AAML0531, high relative expression of
MEF2C was associated with a lower prevalence of cytogenetically/molecularly defined
low-risk disease and higher prevalence of
standard-risk disease, largely because of a lower prevalence of CBF leukemias or mutations in
NPM1 or
CEBPA and a higher prevalence of leukemias with monosomy 7 or abnormalities involving 11q23. Conversely, high relative expression of
MEF2C was associated with some better risk features, particularly a lower prevalence of
FLT3/ITD (10 vs. 18 %; Table
2). Still, the associations between adverse cytogenetic or molecular disease risk features with high
MEF2C expression dominated and largely accounted for the association between
MEF2C expression and outcome. In fact, after multivariable adjustment,
MEF2C expression was not apparently associated with outcome. As
MEF2C expression does not provide prognostic information that is independent of established risk factors,
MEF2C may not be particularly useful as a response biomarker. Nonetheless, high
MEF2C expression was found to be associated with inferior efficacy of curative-intent, intensive AML chemotherapy. These data may, ultimately, provide a strong rationale for therapeutic targeting of
MEF2C transcriptional activation in this disease.
Because of the genetic, molecular, and immunophenotypic heterogeneity of human AML, identification of pharmacologic drugs suitable for reasonably large subsets of patients has remained challenging. Therefore, unraveling signaling aberrancies shared by many of the leukemias could be useful for the development of risk-directed, mechanism-based therapies. Our data suggest the possibility that targeting
MEF2C-induced signaling could serve as one such strategy. Very recent studies have identified
MEF2C as a key factor in regulating
suppressor of cytokine signaling-2 (
SOCS2) in normal and malignant hematopoiesis and indicated that the
MEF2C/
SOCS2 regulatory network might confer leukemic stemness features to a neoplastic hematopoietic clone [
25]. Consistent with a close relationship between
MEF2C and
SOCS2, we [
26] and subsequently others [
25] have provided evidence that high
SOCS2 expression is associated with poor survival in AML. Studies in T-ALL and colon cancer cells have indicated that
MEF2C may inhibit BCL2-regulated apoptosis and can function as a regulator of cell proliferation [
7,
27]. A similar mechanism of apoptosis resistance induced by
MEF2C in AML cells may explain the apparent association between
MEF2C overexpression and failure of AML chemotherapy. Further experimental studies will be required to elucidate the mechanisms of
MEF2C-induced leukemogenesis and effective therapeutic strategies to block them.
It is a strength of our analysis that we included a large number of diagnostic specimens from patients treated homogeneously on a recent cooperative group trial, thereby increasing the precision of the outcome estimates. On the other hand, our studies have some limitations that need to be acknowledged. First, despite the use of over 750 specimens, our study was not large enough to allow for extensive multivariate adjustments. Because of the sample size of the individual risk groups, our ability to perform subset analyses was similarly limited. Second, since unsorted bone marrow specimens were used for our studies, differences in
MEF2C abundance between specimens may not necessarily reflect differences in AML blasts but, rather, other (i.e., non-leukemic) cells or varying compositions of less mature and more mature AML cells. Gene expression studies in human material indicate that higher
MEF2C mRNA levels are found in less mature hematopoietic cells, including LSC populations [
28,
29]. Additional studies will be required for the identification of the exact cellular origins of the greatly variable amounts of
MEF2C and more detailed analyses of relative expression levels along the cellular differentiation path of AML cells. Third, we only had cryopreserved specimens available for our analyses. Future studies will be necessary to determine to what degree, if any,
MEF2C expression changes in the cryopreservation process. And fourth, we were unable to formally study whether high
MEF2C mRNA expression leads to high MEF2C protein expression, a relationship that would provide a strong rationale for therapeutic targeting of MEF2C transcriptional activation in AML. However, preliminary data from ongoing laboratory studies indeed suggest that dysregulated
MEF2C transcription results in MEF2C protein overexpression and confers enhanced AML cell survival (A. Kentsis, personal communication). If clinically exploitable strategies to counteract MEF2C signaling were developed, it is conceivable that
MEF2C expression could become a biomarker of interest for successful drug development [
30], e.g., to identify the subsets of patients most suitable for MEF2C-directed therapy.
Competing interests
The authors have declared no conflicts of interests.
Authors’ contributions
GSL and RBW designed and performed research, analyzed and interpreted data, and wrote the manuscript. TAA, RBG, and Y-CW performed statistical analyses and analyzed interpreted data. CJG, KHH, and RER performed research and wrote the manuscript. AK analyzed and interpreted data and wrote the manuscript. SCR, BAH, ASG, and SM collected, analyzed, and interpreted data. All authors revised the manuscript critically, and gave final approval to submit for publication.