Cytogenetics and Molecular Genetics

Altered miRNA and gene expression in acute myeloid leukemia with complex karyotype identify networks of prognostic relevance

Article metrics

Abstract

Recently, the p53-miR-34a network has been identified to have an important role in tumorigenesis. As in acute myeloid leukemia with complex karyotype (CK-AML) TP53 alterations are the most common known molecular lesion, we further analyzed the p53-miR-34a axis in a large cohort of CK-AML with known TP53 status (TP53altered, n=57; TP53unaltered, n=31; altered indicates loss and/or mutation of TP53). Profiling microRNA (miRNA) expression delineated TP53 alteration-associated miRNA profiles, and identified miR-34a and miR-100 as the most significantly down- and upregulated miRNA, respectively. Moreover, we found a distinct miR-34a expression-linked gene expression profile enriched for genes belonging to p53-associated pathways, and implicated in cell cycle progression or apoptosis. Clinically, low miR-34a expression and TP53 alterations predicted for chemotherapy resistance and inferior outcome. Notably, in TP53unaltered CK-AML, high miR-34a expression predicted for inferior overall survival (OS), whereas in TP53biallelic altered CK-AML, high miR-34a expression pointed to better OS. Thus, detailed molecular profiling links impaired p53 to decreased miR-34a expression, but also identifies p53-independent miR-34a induction mechanisms as shown in TP53biallelic altered cell lines treated with 15-deoxy-Δ12,14-prostaglandin. An improved understanding of this mechanism might provide novel therapeutic options to restore miR-34a function and thereby induce cell cycle arrest and apoptosis in TP53altered CK-AML.

Introduction

Acute myeloid leukemia (AML) is a genetically heterogeneous clonal stem cell disorder characterized by the accumulation of genetic alterations leading to aberrant proliferation and differentiation.1, 2 In addition to chromosomal abnormalities, several genetic lesions of pathogenic and prognostic relevance could be identified, such as gene mutations and/or gene expression signatures. Moreover, microRNAs (miRNAs) have recently been implicated in the pathogenesis of AML;3 miRNAs consisting of 22-nucleotides regulate post-transcriptional expression of protein-coding mRNAs in proliferating cells.4 Malignant cells, in general, as well as AML blasts show an aberrant miRNA expression pattern,5 and given that miRNAs are located in genomic regions frequently affected by translocations, deletions and/or amplifications in cancer, including leukemia, supports their role in tumorigenesis. Furthermore, functional studies confirmed the important pathogenic role of miRNA deregulation in hematological malignancies, including AML.3

Similar to global gene expression profiling (GEP), recent studies using miRNA expression profiling identified cytogenetic subgroup-associated miRNA signatures in AML. Specific signatures were delineated for core-binding factor leukemias [t(8;21)(q22;q22), inv(16)(p13.1q22)/t(16;16)(p13.1;q22)], acute promyelocytic leukemia [t(15;17)(q22;q12)], 11q23 rearranged AML, as well as for cytogenetically normal AML.3 In addition, miRNA signatures are also correlated with molecular markers of prognostic impact, such as mutations in NPM1, FLT3 and CEBPA. However, recent studies have shown that not only miRNA expression signatures, but also expression values of single miRNAs can be of prognostic relevance, for example, the expression of miR-191 and miR-199a, which were associated with adverse overall survival (OS) in intermediate- and poor-risk AML.6 Similarly, in cytogenetically normal AML, higher expression of miR-181a was correlated with longer OS and a higher complete remission (CR) rate.7

With regard to the biological impact of miRNA deregulation in cancer, recently the inactivation of members of the miR-34 family, which can induce cell cycle arrest, senescence and apoptosis, has been shown to provide a selective advantage for cancer cells.8 Although especially inactivating mutations of TP53 appear to significantly contribute to decreased miR-34 expression in tumors, there have also been reports demonstrating the miR-34-encoding genes themselves to be targets of genomic mutations or epigenetic silencing. Furthermore, located on 1p36, a chromosomal band proposed to harbor a tumor suppressor gene due to homozygous deletions seen in neuroblastoma and in other tumor types, miR-34a was revealed as a candidate tumor suppressor based on its tumor suppressive properties.9 In accordance, in chronic lymphocytic leukemia, 17p deletions and/or TP53 mutations have been associated with decreased expression of the direct p53 target miR-34a, and thus with resistance to chemotherapy. Furthermore, in this leukemia, miR-34a expression itself could be associated with chemotherapy-refractory disease irrespective of TP53 alteration status.10

In AML with complex karyotype (CK-AML),11 little is known about miRNAs despite the fact that frequent TP53 alterations are associated with dismal outcome due to a high rate of chemotherapy resistance.12 Therefore, to gain additional insight into the molecular pathogenesis of CK-AML, including the role of miRNAs, we performed an integrative analysis using miRNA profiling and global GEP in 88 molecularly characterized CK-AML with known TP53 alteration status. Findings were evaluated both with regard to their biological relevance in vitro and a potential clinical impact in primary leukemia.

Patients and methods

Patients

DNA and RNA were extracted from frozen peripheral blood and/or bone marrow samples from 88 adult CK-AML patients. The definition of CK-AML followed recommended criteria.1, 11 The diagnosis of AML was based on French–American–British Cooperative Group criteria13 and on the WHO criteria for cases diagnosed after 2004;14 85/88 (97%) patients were treated on consecutive multicenter treatment trials of the AML Study Group (AMLSG), applying age-adjusted intensive chemotherapy: AMLHD98A (n=21; NCT00146120) and AMLSG07-04 (n=30; NCT00151242) for younger patients (16 to 60 years); AMLHD98B (n=6) and AMLSG06-04 (n=28; NCT00151255) for elderly patients (>60 years).15, 16 All studies were approved by local ethics committees and all patients gave informed consent for treatment, cryopreservation of samples and molecular analyses according to the Declaration of Helsinki.

Cytogenetics and molecular genetics

On the basis of banding analysis, 63 (72%) of the 88 CK-AML patients exhibited a monosomal karyotype (CK+/MK+) as previously defined.17 The most frequent monosomy was monosomy 7 in 23 patients (37% of the CK+/MK+ subgroup). The molecular TP53 status assessed as previously described18 was as follows: TP53 mutations in 51/88 (58%), genomic loss of TP53 in 36/88 (41%) CK-AML, with a total of 57/88 (65%) cases showing TP53 alterations (biallelic, 75%; monoallelic, 25%, see Supplementary Table 1 for detailed clinical and cytogenetic information).

miRNA expression profiling in CK-AML

Expression of miRNAs was analyzed using the Agilent miRNA Microarray System (miRBase V12.0) according to the manufacturer’s protocol (Agilent Technologies Inc., Santa Clara, CA, USA). In brief, total RNA isolated from mononuclear AML cells (TP53unaltered, n=8; TP53altered, n=9) with the mirVana miRNA Isolation Kit (Applied Biosystems/Ambion, Austin, TX, USA) was directly labeled. The microarray data were extracted and processed using the Feature Extraction Software, and data analysis was performed with the GeneSpring GX software using the standard normalization and transformation settings (Agilent Technologies Inc.). A miRNA was considered differentially expressed when the expression difference was at least 1.5-fold, with a P-value of <0.05 between groups (unpaired Mann–Whitney test). The complete filtered miRNA microarray data set is provided as Supplementary Table S2.

Validation and evaluation of miRNA findings

Quantitative reverse-transcription PCR (qRT-PCR) of miR-34a and miR-100 using the TaqMan MicroRNA Reverse Transcription Kit and MicroRNA Assays (Applied Biosystems Inc., Foster City, CA, USA) was performed in 88 CK-AML cases as previously described.19 Expression of RNU6B was measured as endogenous control for normalization of data.

Identification of miR-34a- and TP53 status-associated gene expression profiles

On the basis of our previously published GEP data with available miR-34a expression (n=20) and TP53 mutation status (n=32),20, 21 we performed supervised analyses, Class Comparison and Pathway Class Comparison analyses, using BRB-Array Tools Version 3.6.1 (available at http://linus.nci.nih.gov/BRB-ArrayTools.html) and R Version 2.6.0 (available at www.r-project.org) as previously reported.22 In addition, gene signatures were further evaluated using the Molecular Signatures Database (MSigDB, http://www.broadinstitute.org/gsea/msigdb).23

To further investigate and validate the miR-34a-associated cDNA microarray-based gene expression patterns, we analyzed distinct subgroups defined by TP53 alteration and miR-34a expression status using Human Genome GeneChip U133plus2.0 microarrays according to standard protocols reported elsewhere.24 Here, exemplary samples from TP53unaltered (n=6) and TP53biallelic altered (n=6) CK-AML characterized by either high (CK+/miR-34ahigh expression, above median miR-34a expression of the entire cohort) or low (CK+/miR-34alow expression, below median miR-34a expression of the entire cohort) miR-34a expression (n=3 in each group) were analyzed. Data normalization and analyses were carried out as described above (GEO accession number: GSE39730).

Cell lines and cell culture

Leukemia cell lines (HEL, HL-60 and K-562) were purchased from the German Collection of Microorganisms and Cell Cultures (DSMZ) and cultured according to their guidelines (www.dsmz.de/human_and_animal_cell_lines/) as previously reported.25 Cell lines were treated in triplicate with 15-deoxy-Δ12,14-prostaglandin (PGJ2 dissolved in methyl acetate; Sigma Aldrich, St Louis, MO, USA) at a final concentration of 10 μM for 24 h. For the control experiments, cell lines were treated with an equivalent volume methyl acetate (Sigma Aldrich) for 24 h. From treated cell lines, RNA was isolated and miR-34a expression measured by qRT-PCR as outlined above.

Statistical analyses

Pairwise comparisons between patient characteristics were performed by Mann–Whitney test for continuous variables and by Fisher’s exact test for categorical variables. Clinical endpoints were CR after induction therapy, relapse-free survival (RFS), event-free survival (EFS) and OS, as recommended.1 A conditional logistic regression model incorporating stratification according to treatment intensity was used to analyze associations between baseline characteristics and achievement of CR. A Cox model with stratification to account for treatment intensity was used to identify prognostic variables; in addition to TP53 alteration, age, logarithm of white blood cell count, lactate dehydrogenase serum level, monosomy 7 and miR-34a expression were added as explanatory variables in all regression analyses. To provide quantitative information on the relevance of results, 95% confidence intervals (95% CIs) of hazards ratios (HRs) were computed. Statistical analyses were performed using GraphPad Prism 4 (GraphPad Software, La Jolla, CA, USA) and PASW Statistics 18 (IBM Corporation, Armonk, NY, USA). P-values <0.05 were considered to indicate statistical significance.

Results

TP53 status-associated miRNA profiles

To address the potential impact of the p53-miR-34a network in CK-AML, we performed miRNA expression profiling in a cohort of 17 CK-AML cases (TP53altered, n=9; TP53unaltered, n=8). Supervised analysis identified 22 differentially expressed miRNAs (P<0.05) being correlated with the TP53 alteration status, of which miR-34a and miR-100 were the most significantly down- and upregulated miRNAs, respectively (Figure 1). Although miR-34a was on average 22.8-fold downregulated in TP53altered AML cases, miR-100 showed 12.2-fold higher expression levels in TP53altered compared with TP53unaltered AML cases, an inverse correlation not previously reported.

Figure 1
figure1

TP53 status-associated miRNA profiles. (a) miRNAs differentially expressed between TP53unaltered and TP53altered cases (P<0.05, unpaired Mann–Whitney test). MiRNAs (n=22, rows) and CK-AML patients (TP53unaltered, n=8; TP53altered, n=9; columns) were hierarchically clustered. Mean-centered log2 expression values are depicted by a pseudocolor scale, as indicated. (b) Profile plot view of the respective 22 miRNAs. Ascending lines depict miRNAs with higher expression and descending lines miRNAs with lower expression in TP53altered cases compared with TP53unaltered cases. The slope of the line correlates with the fold change difference of the respective miRNA between the two groups. (c) Table indicating the 22 differentially expressed miRNAs sorted by P-value. Fold change between the two groups and type of regulation (up or down) are indicated (TP53altered vs TP53unaltered).

Evaluation of miR-34a and miR-100 expression in CK-AML

We subsequently investigated miR-34a and miR-100 using qRT-PCR in all 88 CK-AML cases to further evaluate the interconnection between the two miRNAs and with regard to the TP53 status. For miR-34a, we observed a significant correlation of expression levels with TP53 alteration (Figure 2a). However, for miR-100 there was no significant correlation with TP53 alteration (Figure 2b), but there was a significant inverse correlation between miR-34a and miR-100 expression (Figure 2c). Thus, our observations demonstrate that miR-34a and TP53 alteration status are not exclusively linked, as we found TP53unaltered cases with low miR-34a expression and TP53altered cases with high miR-34a expression (Figure 2d), indicating that additional determinants of miR-34a expression in CK-AML remain to be identified.

Figure 2
figure2

Expression of miR-34a and miR-100 in CK-AML. Correlation of miR-34a expression (a) and miR-100 expression (b) with TP53 alteration (two-group comparisons were calculated by Mann–Whitney test). Data are shown for the genotypes TP53unaltered, TP53monoallelic altered and TP53biallelic altered. (c) Inverse correlation of miR-34a and miR-100 expression in CK-AML (Wilcoxon matched-pair test). (d) Subgroup-specific miR-34a expression in TP53unaltered and TP53altered CK-AML delineating cases with low miR-34 expression among TP53unaltered CK-AML, and cases with high miR-34a expression among TP53altered CK-AML (Mann–Whitney test).

Delineation of miR-34a-associated GEP

To further characterize the molecular anatomy of CK-AML, we reanalyzed our previously published GEP data for cases with newly available TP53 alteration status and miR-34a expression information and identified a distinct miR-34a expression-associated gene expression pattern.

Supervised Class Comparison analysis revealed 1219 clones to be significantly differentially expressed among groups (miR-34a expression below median of the entire cohort, CK+/miR-34alow expression (n=10) versus miR-34a expression above median of the entire cohort, CK+/miR-34ahigh expression (n=10)) at a nominal P-value <0.05 (Supplementary Table 3). One of the most differentially expressed genes was TP53, with a significantly lower expression in CK+/miR-34alow expression (P=0.004). Pathway Class Comparison analysis showed the miR-34a signature to be enriched for 66/287 investigated Biocarta gene sets (P<0.05; Supplementary Table 4). These included pathways implicated in cell cycle arrest, apoptosis or tumor suppression (for example, cell cycle G2/M checkpoint, cell cycle G1/S checkpoint, regulation of cell cycle progression by Plk3, or tumor suppressor Arf pathway).

To address the interconnection between miR-34a and p53, and because p53 is having a central role in most of these pathways, we looked for a TP53 alteration-associated gene expression pattern. Thereby, supervised Class Comparison analysis revealed 1631 clones to be significantly differentially expressed among groups (TP53unaltered (n=22) versus TP53altered (n=10)) at a nominal P-value <0.05 (Supplementary Table 5). In addition to TP53, we observed that other genes located in 17p13 also showed decreased expression, including, for example, SENP3, ZBTB4 and DVL2, consistent with a gene dosage effect.26 Pathway Class Comparison analysis showed the TP53 signature to be enriched for 35/287 investigated Biocarta gene sets (P<0.05; Supplementary Table 6). These included several pathways known to be regulated by p53 (for example, the AKT–mTOR or the hTERT (telomerase reverse transcriptase in human) pathway).

Comparing the miR-34a- and the TP53 alteration-associated pathways, an overlap of distinct pathways could be observed, such as the RHOA (Ras homolog gene family, member A) signaling pathway, the vascular endothelial growth factor pathway, the STAT3 (signal transducer and activator of transcription 3) pathway or the AKT signaling pathway; but moreover, we also identified pathways that were mutually exclusive, such as the CXCR4 signaling pathway or the phosphatase and tensin homolog pathway being enriched only in the TP53 signature.

Correlation of miR-34a expression with clinical characteristics, response to therapy and survival

Correlation with clinical characteristics and outcome was restricted to patients enrolled in consecutive multicenter treatment trials of the AMLSG (n=85; median age, 57 years (range 19–77)), and was performed based on: (i) data dichotomized at the median miR-34a expression and (ii) data grouped according to the quartiles of miR-34a expression.

CK-AML with low miR-34a expression (below median expression of the entire cohort, CK+/miR-34alow expression) were older (median age, 59 versus 52 years, P=0.006) and showed, in trend, higher lactate dehydrogenase serum levels (median U/l, 591 versus 415, P=0.05) (Table 1). Subgroup analysis (miR-34a expression 1st quartile, CK+/miR-34a1st quartile, versus 2nd–4th quartiles, CK+/miR-34a2nd-4th quartiles, of the entire cohort) revealed higher white blood cell count (P=0.02) for CK+/miR-34a1st quartile (Table 1).

Table 1 Pretreatment characteristics and outcome according to miR-34a expression

MiR-34a expression was associated with resistance to chemotherapy. Response to induction therapy for CK+/miR-34a1st quartile and CK+/miR-34a2nd-4th quartiles was as follows: CR 14 (3/21) versus 42% (26/62; P=0.03), refractory disease 57 (12/21) versus 34% (21/62; P=0.07) and early/hypoplastic death 29 (6/21) versus 24% (15/62; P=0.77), respectively (Table 1). Other variables predicting for poor response to induction therapy were age (CR 18 (6/34) versus 47% (23/49) for CK-AML >60 years and CK-AML <60 years, respectively; P=0.009) and TP53 alteration (CR 25 (13/52) versus 52% (16/31) for TP53altered and TP53unaltered CK-AML, respectively; P=0.02). In multivariable analysis, only age retained its significance for CR achievement (Table 2).

Table 2 Multivariable analyses of outcome

MiR-34a expression was associated with inferior survival. The 3-year estimated survival rates for CK+/miR-34alow expression and CK+/miR-34ahigh expression patients were as follows: OS, 4 versus 20% (log-rank, P=0.006); EFS, 0 versus 7% (P=0.003); and RFS, 8 versus 33% (P=0.13), respectively, (Table 1 and Figure 3). Other variables predicting for inferior OS in univariable analysis were TP53 alteration (P=0.0007), monosomy 7 (P=0.049) and in trend age (>60 years; P=0.051). Multivariable analysis revealed miR-34a-expression (HR, 1.47; 95% CI, 1.06–2.03; P=0.02), TP53 alteration (HR, 2.93; 95% CI, 1.43–6.02; P=0.003), logarithm of white blood cell count (HR, 1.71; 95% CI, 1.02–2.89; P=0.04) and age (HR, 1.03; 95% CI, 1.01–1.04, P=0.008), but not lactate dehydrogenase serum level or monosomy 7 as significant variables for OS (Table 2).

Figure 3
figure3

Kaplan–Meier survival estimates according to miR-34a expression. Data are shown for the genotypes CK+/miR-34alow expression and CK+/miR-34ahigh expression for overall survival (OS, months; log-rank test).

Owing to the wide range of miR-34a expression in TP53unaltered (n=31) and TP53biallelic altered (n=41) CK-AML, we performed explorative subgroup analyses, revealing that among TP53biallelic altered CK-AML, those with high miR-34a expression (4th quartile (n=11) compared with 1st–3rd quartiles (n=30) of the TP53biallelic altered cohort) had significantly better OS (P=0.0225, Figure 4a); but contrariwise, among TP53unaltered CK-AML, those with high miR-34a expression (4th quartile of the TP53unaltered cohort, n=8) had significantly worse OS (P=0.0104; Figure 4b). There were no significant differences in pre-treatment patient characteristics between the respective groups.

Figure 4
figure4

Kaplan–Meier survival estimates for overall survival (OS) for CK-AML according to molecular genetic findings. OS based on miR-34a expression of the respective cohort for the genotypes TP53biallelic altered (a) and for TP53unaltered (b).

miR-34a-associated gene expression profiles within distinct TP53 genotype subgroups

Further molecular characterization using GEP identified distinct miR-34a-associated gene expression patterns within the TP53unaltered and TP53biallelic altered CK-AML subgroups, respectively.

In the TP53unaltered CK-AML subgroup, supervised Class Comparison analysis revealed 981 probe sets to be significantly differentially expressed at a nominal P-value <0.05 (Supplementary Table 7), not comprising TP53. However, the Pathway Class Comparison analysis showed the miR-34a-associated signature to be enriched for 52/290 investigated Biocarta gene sets (P<0.01; Supplementary Table 8). These included several pathways having a role in onco-/leukemogenesis or tumor suppression, like the RAS, PI3K–AKT–mTOR, or the phosphatase and tensin homolog pathway. Furthermore, we found genes of RHO cell motility and the CXCR4 signaling pathway significantly enriched (Supplementary Table 8).

In the TP53biallic altered CK-AML subgroup, Class Comparison analysis revealed 1197 probe sets to be significantly differentially expressed at a nominal P-value <0.05 (Supplementary Table 9). The Pathway Class Comparison analysis showed the miR-34a-associated signature to be enriched for 59/290 investigated Biocarta gene sets (P<0.01; Supplementary Table 10). These included the pathways mentioned above, but moreover, additional pathways implicated in cell cycle control (cell cycle: G2/M checkpoint, regulation of cell cycle progression by Plk3), DNA damage (ATM signaling) and/or apoptosis (induction of apoptosis through DR3 and DR4/5 death receptors).

TP53-independent miR-34a upregulation via peroxisomal proliferator-activated receptor-γ agonist treatment

To further evaluate the biological relevance of the miR-34a-associated gene expression patterns, we used the Connectivity Map, a repository of gene expression profiles from human cell lines treated with a large number of small bioactive molecules,27 to identify modulators that might reverse the respective transcriptional changes. Specifically, we focused on the gene signature discriminating low and high miR-34a-expressing TP53biallelic altered CK-AML subgroups (1st–3rd quartiles of the TP53biallelic altered cohort (n=30)) to query the Connectivity Map for compounds for which the activity is inversely linked to this signature, thereby possessing the potential ability to induce TP53-independent miR-34a upregulation. Among the perturbagens with negative enrichment scores (indicating compounds that may repress the biological state encoded in the low miR-34a expression signature), we observed PGJ2, a natural peroxisomal proliferator-activated receptor-γ ligand, to be one of the most prominent drugs. To test our hypothesis that this compound might invert the low miR-34a-associated expression signature via the induction of miR-34a expression in a TP53-independent way, we treated three myeloid cell lines with known biallelic TP53 alteration (HEL, HL-60 and K-562) with PGJ2. Although the TP53biallelic altered cell lines showed very low miR-34a expression levels similar to the TP53biallelic altered CK-AML subgroup (data not shown), treatment with PGJ2 for 24 h induced a 3.7 to 18.1-fold upregulation of miR-34a expression in the respective cell lines (Figure 5).

Figure 5
figure5

PGJ2 induced TP53-independent expression of miR-34a. Treatment of TP53biallelic altered cell lines with PGJ2 for 24 h results in a significant upregulation of miR-34a expression in HEL and HL-60, and also in trend in K-562 (see indicated P-values; normalized miR-34a expression depicts the fold-change in PGJ2-treated cells compared with the respective controls that have been set to 1.0).

Identification of potential miR-34a target genes

To identify putative miR-34a-regulated target genes in CK-AML, we performed an integrative analysis comparing predicted miR-34a targets and our TP53 alteration-associated gene expression signatures (see above), as it was recently shown that mammalian miRNAs predominantly act to decrease target mRNA levels.28 Using different prediction algorithms (MicroCosm Targets (formerly miRBase Targets),29 miRWalk,30 DIANA–microT v3.0,31 TargetScan32 and PicTar33), we identified 2019 potential miR-34a target genes predicted by at least one of the algorithms (range: 235–1221 predicted targets). Intersection of this theoretical direct miR-34a target gene list with genes upregulated in CK+/TP53altered cases, which display low miR-34a expression and thus an increased abundance of target mRNA was anticipated, revealed potential miR-34a-regulated candidates (Supplementary Table 11). Although this approach identified known miR-34a target genes such as the transmembrane glycoprotein CD44,34 cyclin-dependent kinase CDK6,35 ACSL1 (acyl-CoA synthetase long-chain family member 1)36 and the transcription factor E2F3,37 our integrative analysis also pointed to novel candidate genes of potential oncogenic relevance, such as BCL6, CCNE2 (Cyclin E2, related to the known miR-34a target CCND1), the colony stimulating factor 1 receptor gene (CSF1R) and ICOSLG (inducible T-cell co-stimulator ligand, previously associated with AML38). Additional genes of potential pathogenic relevance include ALDOA (aldolase A), a leukemia-associated antigen expressed in chronic myeloid leukemia,39 and TPD52 encoding the tumor protein D52.

Discussion

In this study, we delineated the miRNA and mRNA profiles associated with genetically defined CK-AML subtypes. We show here that TP53altered CK-AML harbor a specific miRNA profile, which comprises, among others, miR-34a and miR-100 as the most significantly down- and upregulated miRNAs, respectively. Notably, other known p53-induced miRNAs (for example, miR-145, miR-107, miR-192, miR-215, and miR-16) were not found significantly altered, most likely due to cell type dependent p53-associated gene expression patterns. Moreover, CK+/TP53altered and CK+/miR-34alow expression are characterized by similar gene expression profiles enriched for genes belonging to pathways known to be regulated by p53 and/or implicated in cell cycle progression or apoptosis. These findings fit well with a p53-dependent miR-34a expression reported in many other tumors, which seems to be independent of the tumor type.8, 10 Although p53 protects cellular integrity and p53 activation induces cell cycle arrest, apoptosis, or senescence following DNA damage, these effects are in part mediated through p53-dependent induction of miR-34a, which has been shown to be a direct p53 target.8 Accordingly, ectopic miR-34 expression also induces apoptosis, cell cycle arrest or senescence, and loss of miR-34 expression has been linked to chemotherapy resistance because of impaired apoptosis induction by p53-activating agents.8, 10

Consistent with findings from chronic lymphocytic leukemia, in our study, miR-34a expression was also associated with resistance to chemotherapy. Furthermore, miR-34a expression also predicted for inferior survival. As miR-34a is a direct target of p53, it was not surprising that we observed similar clinical characteristics and endpoints for CR, EFS, OS and RFS in TP53altered CK-AML. In multivariable analysis, only miR-34a expression retained its significance for OS and RFS. Thus, these data further demonstrate the important role of the p53-miR-34a network in CK-AML that warrants further investigation in larger patient cohorts.

The correlation of high miR-100 expression with TP53 alteration seen in our microarray analyses has not been reported previously. Although the analysis of our entire CK-AML patient cohort by qRT-PCR showed miR-100 expression only in trend correlated with TP53 alteration, we could confirm a significant inverse correlation between miR-34a and miR-100 expression. MiR-100 has not been extensively studied, and little is known about miR-100 function in leukemia. Nevertheless, miR-100 expression was recently found to be associated with resistance to chemotherapy in pediatric acute lymphoblastic leukemia,40 and it was shown to have a role in controlling proliferation because of its function as an endogenous repressor of the serine/threonine protein kinase mammalian target of rapamycin (mTOR).41 Furthermore, a recent study in AML showed that miR-100 can regulate G1/S transition and S-phase entry, and block the terminal differentiation of cells.42 These observations, in addition to the deregulation seen in our study, point to an important role of miR-100 in AML pathogenesis.42 Thus, the inverse correlation between miR-100 and miR-34a expression may contribute to the chemoresistance and the enrichment of mTOR pathway members in the TP53unaltered-associated CK-AML signature. Although this speculation of course warrants further investigation, in view of recent reports on the role of miR-100 in cancer,40, 41, 42 our observations strongly suggest a respective role in CK-AML.

Unexpectedly, among TP53biallelic altered CK-AML, not all cases exhibited low miR-34a expression and explorative subgroup analyses revealed that those with high miR-34a expression (4th quartile) had significantly better OS (Figure 4a). High miR-34a expression in TP53biallelic altered CK-AML was not related to the type of biallelic TP53 alteration (hemizygous, homozygous mutations or location of mutation). However, these cases harbored a distinct gene expression profile different from that of low miR-34a-expressing TP53biallelic altered CK-AMLs. Although this signature was enriched for genes belonging to pathways implicated in cell cycle control, DNA damage and/or apoptosis, our data suggest different pathomechanisms leading to a p53-independent miR-34a upregulation in these cases (Figure 6c). Reflecting biological mechanisms, we could use this signature to identify a compound with ability to induce miR-34a expression in a TP53biallelic altered background. Treatment of TP53biallelic altered cell lines with PGJ2 delineated to be negatively linked to the low miR-34a expressing TP53biallelic altered CK-AML signature resulted in a marked upregulation of miR-34a expression, thereby offering the possibility to, in part, restore defective p53 signaling.

Figure 6
figure6

Modeling the p53-miR-34a network and its potential impairment in CK-AML. (a) Intact p53-induced upregulation of miR-34a leads to cell cycle arrest and apoptosis through silencing of miR-34a target genes. (b) Impaired p53 leads to weak or no miR-34 induction resulting in impaired cell cycle arrest and apoptosis, and thereby in degree to genomic instability and proliferation. (c) p53-independent miR-34a induction also leads to cell cycle arrest and apoptosis through silencing of miR-34a target genes.

In accordance, recent findings further underline the hypothesis that miR-34a expression can also be regulated in a p53-independent manner. During oncogene-induced senescence, a mechanism preventing tumorigenesis, miR-34a can be upregulated by ELK1, an E-twenty six family transcription factor, and can act as a tumor suppressor by targeting the proto-oncogene MYC.43 Another transcription factor regulating miR-34a expression is the CCAAT-enhancer binding protein alpha (CEBPA). During granulopoiesis, CEBPA upregulates miR-34a that targets E2F3, thereby blocking myeloid cell proliferation.44 Mutations in CEBPA are found in approximately 10% of AML1 and expression profiling revealed significantly lower miR-34a expression in CEBPA-mutated AML compared with CEBPA wild-type AML.45

Similarly, the inverse prognostic impact of miR-34a expression in TP53unaltered CK-AML (high miR-34a expression (4th quartile) was correlated with significantly worse OS (P=0.0104)) (Figure 4b) might also be related to p53-independent miR-34a induction. In agreement with this idea, the gene expression signature discriminating the high miR-34a expression cohort within the TP53unaltered cases did not contain TP53 or correlate with p53 pathway genes, but we found deregulated CXCR4 and RHOA signaling in these cases that might lead to upregulation of miR-34a via the activation of ERK signaling.46 Alternatively, p53 could be active in all TP53unaltered patients, and in cases with very low expression of miR-34a, other mechanisms could prevent p53 from inducing miR-34a expression. Although only one case exhibited a deletion of the miR-34a locus 1p36, which resulted in the owest expression of miR-34a observed in this group, the other TP53unaltered cases showed no 1p36 alteration (see Rücker et al.12 and unpublished data). Although the miR-34a gene has been reported to be epigenetically inactivated at high frequency in different solid tumors due to aberrant CpG methylation of its promoter,8 no silencing of miR-34a has been reported in AML yet.47 However, only 20 cases of undisclosed cytogenetics were analyzed in this study. Thus, epigenetic inactivation of the miR-34a gene could nevertheless contribute to atypical low expression of miR-34a in TP53unaltered patients, even though a probably rare event in AML.

With regard to the downstream effects of miR-34a in CK-AML, our integrative miRNA and mRNA expression data analysis proved to be useful, as it identified several known miR-34a target genes as well as new candidates with potential oncogenic relevance. For example, TPD52 has recently been reported to be downregulated after induction of miR-34a expression and could be validated as a direct miR-34a target.48 TPD52 is also found overexpressed in many cancer types, and overexpression of its murine ortholog mD52 in cell lines resulted in enhanced proliferation and anchorage-independent cell growth,49 thereby also suggesting a potential role of TPD52 in deregulated p53 signaling in CK-AML. Thus, our study provides a solid basis for further investigation of the complex networks involved in CK-AML.

In summary, we could demonstrate the prognostic relevance of miR-34a expression in CK-AML. As in other cancers, TP53 alterations are associated with repressed miR-34a expression in CK-AML. However, our detailed, multidimensional analysis could further characterize complex molecular networks and delineate subgroups of TP53unaltered CK-AML with poor outcome despite the presence of miR-34a. In addition, we could demonstrate that in TP53altered cases, miR-34a might be expressed independent of p53 by mechanisms also associated with better response to chemotherapy and improved outcome. Although these mechanisms need to be further explored, they ultimately might be targeted therapeutically and could open new strategies to deal with CK-AML.

Accession codes

Accessions

Gene Expression Omnibus

References

  1. 1

    Dohner H, Estey EH, Amadori S, Appelbaum FR, Buchner T, Burnett AK et al. Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood 2010; 115: 453–474.

  2. 2

    Frohling S, Scholl C, Gilliland DG, Levine RL . Genetics of myeloid malignancies: pathogenetic and clinical implications. J Clin Oncol 2005; 23: 6285–6295.

  3. 3

    Marcucci G, Mrozek K, Radmacher MD, Garzon R, Bloomfield CD . The prognostic and functional role of microRNAs in acute myeloid leukemia. Blood 2011; 117: 1121–1129.

  4. 4

    Bartel DP . MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 2004; 116: 281–297.

  5. 5

    Chen J, Odenike O, Rowley JD . Leukaemogenesis: more than mutant genes. Nat Rev Cancer 2010; 10: 23–36.

  6. 6

    Garzon R, Volinia S, Liu CG, Fernandez-Cymering C, Palumbo T, Pichiorri F et al. MicroRNA signatures associated with cytogenetics and prognosis in acute myeloid leukemia. Blood 2008; 111: 3183–3189.

  7. 7

    Schwind S, Maharry K, Radmacher MD, Mrozek K, Holland KB, Margeson D et al. Prognostic significance of expression of a single microRNA, miR-181a, in cytogenetically normal acute myeloid leukemia: a Cancer and Leukemia Group B study. J Clin Oncol 2010; 28: 5257–5264.

  8. 8

    Hermeking H . The miR-34 family in cancer and apoptosis. Cell Death Differ 2010; 17: 193–199.

  9. 9

    Cole KA, Attiyeh EF, Mosse YP, Laquaglia MJ, Diskin SJ, Brodeur GM et al. A functional screen identifies miR-34a as a candidate neuroblastoma tumor suppressor gene. Mol Cancer Res 2008; 6: 735–742.

  10. 10

    Zenz T, Mohr J, Eldering E, Kater AP, Buhler A, Kienle D et al. miR-34a as part of the resistance network in chronic lymphocytic leukemia. Blood 2009; 113: 3801–3808.

  11. 11

    Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H et al WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. IARC Press: Lyon, France, 2008.

  12. 12

    Rücker FG, Schlenk RF, Bullinger L, Kett H, Becker A, Kugler CM et al. In acute myeloid leukemia with complex karyotype TP53 alterations are associated with specific genomic aberrations and predict inferior survival. Blood (ASH Meeting Abstracts) 2009; 114 abstract 2632.

  13. 13

    Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposed revised criteria for the classification of acute myeloid leukemia. A report of the French-American-British Cooperative Group. Ann Intern Med 1985; 103: 620–625.

  14. 14

    Jaffe ES, Harris NL, Stein H, Vardiman JD eds. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. 3rd edn. (IARC Press: Lyon, France, 2001).

  15. 15

    Schlenk RF, Dohner K, Mack S, Stoppel M, Kiraly F, Gotze K et al. Prospective evaluation of allogeneic hematopoietic stem-cell transplantation from matched related and matched unrelated donors in younger adults with high-risk acute myeloid leukemia: German-Austrian trial AMLHD98A. J Clin Oncol 2010; 28: 4642–4648.

  16. 16

    Schlenk RF, Frohling S, Hartmann F, Fischer JT, Glasmacher A, Del Valle F et al. Intensive consolidation versus oral maintenance therapy in patients 61 years or older with acute myeloid leukemia in first remission: results of second randomization of the AML HD98-B treatment Trial. Leukemia 2006; 20: 748–750.

  17. 17

    Breems DA, Van Putten WL, De Greef GE, Van Zelderen-Bhola SL, Gerssen-Schoorl KB, Mellink CH et al. Monosomal karyotype in acute myeloid leukemia: a better indicator of poor prognosis than a complex karyotype. J Clin Oncol 2008; 26: 4791–4797.

  18. 18

    Rucker FG, Schlenk RF, Bullinger L, Kayser S, Teleanu V, Kett H et al. TP53 alterations in acute myeloid leukemia with complex karyotype correlate with specific copy number alterations, monosomal karyotype, and dismal outcome. Blood 2012; 119: 2114–2121.

  19. 19

    Sander S, Bullinger L, Klapproth K, Fiedler K, Kestler HA, Barth TF et al. MYC stimulates EZH2 expression by repression of its negative regulator miR-26a. Blood 2008; 112: 4202–4212.

  20. 20

    Bullinger L, Dohner K, Bair E, Frohling S, Schlenk RF, Tibshirani R et al. Use of gene-expression profiling to identify prognostic subclasses in adult acute myeloid leukemia. N Engl J Med 2004; 350: 1605–1616.

  21. 21

    Kharas MG, Lengner CJ, Al-Shahrour F, Bullinger L, Ball B, Zaidi S et al. Musashi-2 regulates normal hematopoiesis and promotes aggressive myeloid leukemia. Nat Med 2010; 16: 903–908.

  22. 22

    Luck SC, Russ AC, Du J, Gaidzik V, Schlenk RF, Pollack JR et al. KIT mutations confer a distinct gene expression signature in core binding factor leukaemia. Br J Haematol 2010; 148: 925–937.

  23. 23

    Liberzon A, Subramanian A, Pinchback R, Thorvaldsdottir H, Tamayo P, Mesirov JP . Molecular signatures database (MSigDB) 3.0. Bioinformatics 2011; 27: 1739–1740.

  24. 24

    Kohlmann A, Bullinger L, Thiede C, Schaich M, Schnittger S, Dohner K et al. Gene expression profiling in AML with normal karyotype can predict mutations for molecular markers and allows novel insights into perturbed biological pathways. Leukemia 2010; 24: 1216–1220.

  25. 25

    Rucker FG, Sander S, Dohner K, Dohner H, Pollack JR, Bullinger L . Molecular profiling reveals myeloid leukemia cell lines to be faithful model systems characterized by distinct genomic aberrations. Leukemia 2006; 20: 994–1001.

  26. 26

    Rucker FG, Bullinger L, Schwaenen C, Lipka DB, Wessendorf S, Frohling S et al. Disclosure of candidate genes in acute myeloid leukemia with complex karyotypes using microarray-based molecular characterization. J Clin Oncol 2006; 24: 3887–3894.

  27. 27

    Lamb J, Crawford ED, Peck D, Modell JW, Blat IC, Wrobel MJ et al. The Connectivity Map: using gene-expression signatures to connect small molecules, genes, and disease. Science 2006; 313: 1929–1935.

  28. 28

    Guo H, Ingolia NT, Weissman JS, Bartel DP . Mammalian microRNAs predominantly act to decrease target mRNA levels. Nature 2010; 466: 835–840.

  29. 29

    Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, Enright AJ . miRBase: microRNA sequences, targets and gene nomenclature. Nucleic Acids Res 2006: (34 (Database issue)): D140–D144.

  30. 30

    Dweep H, Sticht C, Pandey P, Gretz N . miRWalk—Database: prediction of possible miRNA binding sites by "walking" the genes of three genomes. J Biomed Inform 2011; 44: 839–847.

  31. 31

    Maragkakis M, Reczko M, Simossis VA, Alexiou P, Papadopoulos GL, Dalamagas T et al. DIANA-microT web server: elucidating microRNA functions through target prediction. Nucleic Acids Res 2009: (37 (Web Server issue)): W273–W276.

  32. 32

    Grimson A, Farh KK, Johnston WK, Garrett-Engele P, Lim LP, Bartel DP . MicroRNA targeting specificity in mammals: determinants beyond seed pairing. Mol Cell 2007; 27: 91–105.

  33. 33

    Krek A, Grun D, Poy MN, Wolf R, Rosenberg L, Epstein EJ et al. Combinatorial microRNA target predictions. Nat Genet 2005; 37: 495–500.

  34. 34

    Liu C, Kelnar K, Liu B, Chen X, Calhoun-Davis T, Li H et al. The microRNA miR-34a inhibits prostate cancer stem cells and metastasis by directly repressing CD44. Nat Med 2011; 17: 211–215.

  35. 35

    Sun F, Fu H, Liu Q, Tie Y, Zhu J, Xing R et al. Downregulation of CCND1 and CDK6 by miR-34a induces cell cycle arrest. FEBS Lett 2008; 582: 1564–1568.

  36. 36

    Li WQ, Chen C, Xu MD, Guo J, Li YM, Xia QM et al. The rno-miR-34 family is upregulated and targets ACSL1 in dimethylnitrosamine-induced hepatic fibrosis in rats. Febs J 2011; 278: 1522–1532.

  37. 37

    Welch C, Chen Y, Stallings RL . MicroRNA-34a functions as a potential tumor suppressor by inducing apoptosis in neuroblastoma cells. Oncogene 2007; 26: 5017–5022.

  38. 38

    Tamura H, Dan K, Tamada K, Nakamura K, Shioi Y, Hyodo H et al. Expression of functional B7-H2 and B7.2 costimulatory molecules and their prognostic implications in de novo acute myeloid leukemia. Clin Cancer Res 2005; 11: 5708–5717.

  39. 39

    Zou L, Wu Y, Pei L, Zhong D, Gen M, Zhao T et al. Identification of leukemia-associated antigens in chronic myeloid leukemia by proteomic analysis. Leuk Res 2005; 29: 1387–1391.

  40. 40

    Schotte D, De Menezes RX, Moqadam FA, Khankahdani LM, Lange-Turenhout E, Chen C et al. MicroRNA characterize genetic diversity and drug resistance in pediatric acute lymphoblastic leukemia. Haematologica 2011; 96: 703–711.

  41. 41

    Doghman M, El Wakil A, Cardinaud B, Thomas E, Wang J, Zhao W et al. Regulation of insulin-like growth factor-mammalian target of rapamycin signaling by microRNA in childhood adrenocortical tumors. Cancer Res 2010; 70: 4666–4675.

  42. 42

    Zheng YS, Zhang H, Zhang XJ, Feng DD, Luo XQ, Zeng CW et al. MiR-100 regulates cell differentiation and survival by targeting RBSP3, a phosphatase-like tumor suppressor in acute myeloid leukemia. Oncogene 2011; 31: 80–92.

  43. 43

    Christoffersen NR, Shalgi R, Frankel LB, Leucci E, Lees M, Klausen M et al. p53-independent upregulation of miR-34a during oncogene-induced senescence represses MYC. Cell Death Differ 2010; 17: 236–245.

  44. 44

    Pulikkan JA, Peramangalam PS, Dengler V, Ho PA, Preudhomme C, Meshinchi S et al. C/EBPalpha regulated microRNA-34a targets E2F3 during granulopoiesis and is down-regulated in AML with CEBPA mutations. Blood 2010; 116: 5638–5649.

  45. 45

    Marcucci G, Maharry K, Radmacher MD, Mrozek K, Vukosavljevic T, Paschka P et al. Prognostic significance of, and gene and microRNA expression signatures associated with, CEBPA mutations in cytogenetically normal acute myeloid leukemia with high-risk molecular features: a Cancer and Leukemia Group B Study. J Clin Oncol 2008; 26: 5078–5087.

  46. 46

    Zhao M, Discipio RG, Wimmer AG, Schraufstatter IU . Regulation of CXCR4-mediated nuclear translocation of extracellular signal-related kinases 1 and 2. Mol Pharmacol 2006; 69: 66–75.

  47. 47

    Chim CS, Wong KY, Qi Y, Loong F, Lam WL, Wong LG et al. Epigenetic inactivation of the miR-34a in hematological malignancies. Carcinogenesis 2010; 31: 745–750.

  48. 48

    Kaller M, Liffers ST, Oeljeklaus S, Kuhlmann K, Roh S, Hoffmann R et al. Genome-wide characterization of miR-34a induced changes in protein and mRNA expression by a combined pulsed SILAC and micro-array analysis. Mol Cell Proteomics 2011; 10: M111.010462.

  49. 49

    Lewis JD, Payton LA, Whitford JG, Byrne JA, Smith DI, Yang L et al. Induction of tumorigenesis and metastasis by the murine orthologue of tumor protein D52. Mol Cancer Res 2007; 5: 133–144.

Download references

Acknowledgements

Analyses were performed using BRB-ArrayTools developed by Dr Richard Simon and BRB-ArrayTools Development Team. We thank all members of the German-Austrian AML Study Group (AMLSG) for their participation in this study and for providing patient samples. This study was supported in part by the Bundesministerium für Bildung und Forschung (NGFNplus Grant No. 01GS0871), by the Deutsche José Carreras Stiftung e.V. (DJCLS R 06/41v), and LB was supported in part by the Deutsche Forschungsgemeinschaft (Heisenberg-Stipendium BU 1339/3-1).

Author information

Correspondence to L Bullinger.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Leukemia website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Rücker, F., Russ, A., Cocciardi, S. et al. Altered miRNA and gene expression in acute myeloid leukemia with complex karyotype identify networks of prognostic relevance. Leukemia 27, 353–361 (2013) doi:10.1038/leu.2012.208

Download citation

Keywords

  • acute myeloid leukemia
  • microRNA
  • p53

Further reading