Abstract

Long non-coding RNAs (lncRNAs) have been recently found to be pervasively transcribed in human genome and link to diverse human diseases. However, the expression patterns and regulatory roles of lncRNAs in hematopoietic malignancies have not been reported. Here, we carried out a genome-wide lncRNA expression study in MLL-rearranged acute lymphoblastic leukemia (MLL-r ALL) and established lncRNA/messenger RNA coexpression networks to gain insight into the biological roles of these dysregulated lncRNAs. We detected a number of lncRNAs that were differentially expressed in MLL-r ALL samples compared with MLL-r wild-type and identified unique lncRNA expression patterns between MLL-r subtypes with different translocations as well as between infant MLL-r ALL with other MLL-r ALL patients, suggesting that they might be served as novel biomarkers for the disease. Importantly, several lncRNAs that correspond with membrane protein genes, including a lysosome-associated membrane protein, were identified. No such link between the membrane proteins and MLL-r leukemia has been reported previously. Impressively, the functional analysis showed that several lncRNAs corresponded to the expression of MLL-fusion protein target genes, including HOXA9, MEIS1, etc., while some other associated with histone-related functions or membrane proteins. Further experiments characterize the effect of some lncRNAs on MLL-r leukemia apoptosis and proliferation as the function of the coexpressed HOXA gene cluster. Finally, a set of lncRNAs epigenetically regulated by H3K79 methylation were also discovered. These findings may provide novel insights into the mechanisms of lncRNAs involved in the initiation of MLL-r leukemia. This is the first study linking lncRNAs to leukemogenesis.

INTRODUCTION

MLL-rearranged acute lymphoblastic leukemia (MLL-r ALL) has long been considered distinct from other childhood lymphocytic leukemias. It is characterized by an exceptionally high incidence (i.e. 80% of infant ALL cases) of chromosomal translocations affecting the MLL gene (1,2). The presence of MLL rearrangements is the most important independent predictor of adverse outcome, and the disease remains the most aggressive type of childhood leukemia. Thus, the distinct mechanisms of leukemogenesis must be explored to develop a rational basis for the design of novel therapies for the treatment of this disease.

Several theories have been proposed to explain the leukemogenic potential of the MLL oncoprotein, including transcriptional activation and changes in chromatin structure (3). These studies expand our knowledge of the etiology of MLL-r and provide a potential direction for clinical application. However, the detailed mechanisms of the initiation of MLL-r leukemia remain to be demonstrated. Furthermore, it is not yet clear whether other transcripts interact with the MLL-fusion protein and participate in leukemogenesis, and the pathways involved in the transformation of cells are not fully understood.

The human transcriptome comprises not only large numbers of protein-coding messenger RNAs (mRNAs) but also a large set of non-protein coding transcripts that have structural, regulatory or unknown functions (4). In the past years, studies on small non-coding RNAs, such as microRNAs (miRNAs), in MLL-r leukemia have revealed that a number of non-protein coding transcripts extensively regulate and control proliferation, apoptosis and drug resistance in MLL-r leukemia cells (5,6). More recently, members of a less well-characterized class of non-coding RNAs, designated as long ncRNAs [lncRNAs, >200 nucleotides (nt)], have also been implicated in developmental regulation (7), disease pathogenesis (8) and chromatin state (9). LncRNAs are characterized by the complexity and diversity of their sequences and regulation mechanisms. In hematopoiesis, several lncRNAs have been reported to be associated with blood development (10,11). Notably, a handful of studies have implicated that lncRNAs are emerging as new players in the cancer paradigm demonstrating potential roles in both oncogenic and tumor suppressive pathways, highlighting the emerging functional role of lncRNAs in human cancer (8,12). However, no lncRNA has been reported to be associated with hematopoietic malignancies to date. Importantly, one study showed that the lncRNA Mistral could activate Hoxa6 and Hoxa7 expression by recruiting MLL1 to the chromatin (13). This observation suggested that this class of non-coding RNA might be associated with the MLL gene related function.

To explore whether any lncRNAs are associated with MLL rearrangement in the disease progression and served as novel biomarkers for the disease, in this study, we carried out a genome-wide lncRNA expression study in MLL-r ALL patients. A unique lncRNA expression patterns for the disease and a distinct set of lncRNAs associated with the pathways involved in the biological processes of MLL-r ALL progression were identified. Some of these lncRNAs were epigenetically regulated by histone 3 lysine 79 (H3K79) methylation in MLL-r ALL cells. These results support the existence of MLL-r ALL-specific lncRNAs that are independent of cell lineage and cytogenetic abnormalities and hint at crucial roles for these lncRNAs in the pathogenesis of MLL-r ALL.

RESULTS

The lncRNA profile indicates the differential expression patterns and potentially unique biological characteristics of lncRNA in childhood MLL-r ALL

We hypothesized that a specific lncRNA expression profile might be characteristic of MLL-r ALL. Therefore, we carried out a genome-wide lncRNA expression study using the Human LncRNA Array v2.0 platform. We processed 23 children patients carrying MLL-r ALL (n = 15) or carried untranslocated MLL genes (MLL-wt ALL, n = 8) and 3 healthy controls. The lncRNA microarray platform detected 33 045 lncRNAs, including long intergenic non-coding RNAs (lincRNAs), as well as those associated with the nearby coding genes.

The lncRNA array data were initially processed using an unsupervised analysis in which the samples largely clustered into their biological subtypes. All of the MLL-r samples clustered together and were distinct from the MLL-wt samples or healthy controls. We extended the analysis of the lncRNA array data by creating a supervised list of the most differentially expressed lncRNAs (P < 0.01, fold-change > 2.0) of MLL-r versus MLL-wt and MLL-r versus healthy controls (Fig. 1). These results clearly show that the lncRNAs in the MLL-r samples are globally dysregulated, including 52 upregulated and 59 downregulated, compared with both the MLL-wt samples and the healthy controls, while the MLL-wt and the healthy controls exhibited similarly relative expression. The most upregulated lncRNAs are ENST00000443469, NR_033375 and ENST00000432488, and the most downregulated lncRNAs are BC035666, NR_026844 and ENST00000454882, suggesting that these lncRNAs might be particularly important in the biology of MLL-r ALL.

Figure 1.

Cluster analysis of lncRNA expression in MLL-r and MLL-wt ALL samples and healthy controls. The 111 top-ranked differentially expressed lncRNAs in MLL-r ALL (P < 0.01, fold-change > 2.0). The expression values are represented in yellow and blue, indicating expression above and below the median expression value across all samples, respectively.

Figure 1.

Cluster analysis of lncRNA expression in MLL-r and MLL-wt ALL samples and healthy controls. The 111 top-ranked differentially expressed lncRNAs in MLL-r ALL (P < 0.01, fold-change > 2.0). The expression values are represented in yellow and blue, indicating expression above and below the median expression value across all samples, respectively.

To validate the lncRNA array platform accuracy, we next selected a number of lncRNAs for further analysis by qRT-PCR using a subset of chosen childhood ALL samples (MLL-r ALL, n = 18; MLL-wt ALL, n = 29). The three most upregulated (ENST00000443469, NR_033375 and ENST00000432488) and the two most downregulated (BC035666 and NR_026844) lncRNAs were selected for validation. Two other lncRNAs that are adjacent to the genes MEF2C and CTDSPL (uc003che.2 and uc003kjn.1, respectively) were also selected because both MEF2C and CTDSPL are reported to play important roles in MLL-r leukemic cells proliferation, dissemination and cell cycle (14,15), The qRT-PCR confirmed the lncRNA array data except for NR_026844, which did not show a significant difference (Fig. 2 A–C). The most differentially expressed ones are ENST00000443469 (P < 0.0001), NR_033375 (P = 0.0347) and ENST00000432488 (P = 0.0156), consistent with the lncRNA array data. To further validate these qRT-PCR results, we used three known mRNAs HOXA9, HOXA7 and MEIS1 which have been shown to be deregulated in MLL-r ALL as positive controls (Supplementary Material, Fig. S1). The expression levels of these mRNAs were consistent with previous studies.

Figure 2.

LncRNA expression validation using qRT-PCR. Differential expression of lncRNA cohorts in MLL-r or MLL-wt patients was validated with qRT-PCR, and the relative expression level of lncRNAs was normalized by GAPDH. (A) The most upregulated lncRNAs were ENST00000443469, NR_033375 and ENST00000432488. (B) The most downregulated lncRNA was BC035666. (C) LncRNAs adjacent to genes that play core roles in MLL-r ALL pathogenesis.

Figure 2.

LncRNA expression validation using qRT-PCR. Differential expression of lncRNA cohorts in MLL-r or MLL-wt patients was validated with qRT-PCR, and the relative expression level of lncRNAs was normalized by GAPDH. (A) The most upregulated lncRNAs were ENST00000443469, NR_033375 and ENST00000432488. (B) The most downregulated lncRNA was BC035666. (C) LncRNAs adjacent to genes that play core roles in MLL-r ALL pathogenesis.

The unique lncRNA expression patterns between MLL-r subtypes with different translocations as well as between infant MLL-r ALL with other MLL-r ALL patients

By far the most frequent MLL translocations found among children, especially among infant ALL patients, are t(4;11), t(11;19) and t(9;11) (1,3), giving rise to the fusion proteins MLL-AF4, MLL-ENL and MLL-AF9, respectively. Accumulating evidence suggests that MLL translocations cause deregulated gene expression as a result of translocation-specific modifications (3). Therefore, we asked whether distinct lncRNAs could be identified associated with the type of MLL translocation. For this, we separated our MLL-r ALL samples according to the type of translocation, mainly t(4;11) (n = 6), t(11;19) (n = 4) and t(9;11) (n = 2), and three with unavailable translocations information, and determined the differentially expressed lncRNAs for each subgroup (fold change > 5.0, P < 0.01). MLL-r ALL samples carrying translocations t(4;11), t(11;19) or t(9;11) indeed display translocation-specific lncRNA expression signatures that clearly separate these samples into three distinct patient groups (Fig. 3 A and B).

Figure 3.

Differential expression of lncRNAs in subtypes of MLL-r ALL. (A) Heatmap is based on the most significantly upregulated lncRNAs for each type of translocations. (B) Heatmap is based on the most significantly downregulated lncRNAs for each types of translocations patient group (compared with the other patient groups combined). (C) Differential expressed lncRNAs between MLL-r ALL patients younger than 1 year (infant) and those older children (>1 year). The expression values are represented in yellow and blue.

Figure 3.

Differential expression of lncRNAs in subtypes of MLL-r ALL. (A) Heatmap is based on the most significantly upregulated lncRNAs for each type of translocations. (B) Heatmap is based on the most significantly downregulated lncRNAs for each types of translocations patient group (compared with the other patient groups combined). (C) Differential expressed lncRNAs between MLL-r ALL patients younger than 1 year (infant) and those older children (>1 year). The expression values are represented in yellow and blue.

In total, heatmaps visualized the most significantly upregulated (Fig. 3A) and downregulated (Fig. 3B) lncRNAs in t(4;11), t(11;19) and t(9;11), respectively. The results showed that, although the MLL-r ALL samples all clustered within the same branch of the dendrogram, there was particularly tight clustering of the three most common translocations in children MLL-r ALL, MLL-AF4, MLL-ENL and MLL-AF4. Each MLL-r subtype has a unique lncRNA expression profile, for example, the most differentially upregulated lncRNAs in MLL-AF4 are ENST00000492683, while in MLL-ENL and MLL-AF9 are uc001zfv.1 and uc001wko.1, and the most downregulated lncRNAs in MLL-AF4, MLL-ENL and MLL-AF9 are uc002ybr.1, BC024020 and AJ493605, respectively. The results indeed reveal that apart from a fundamental signature shared by all MLL-r childhood ALL samples, each type of MLL translocation is associated with a translocation-specific lncRNA expression signature.

Infant ALL (<1 year) is characterized by a poor prognosis and a high incidence of MLL translocations and is treated as a distinct subtype of MLL-r ALL (1). Next, we asked whether infant ALL patients bearing MLL translocations have their unique lncRNA signature. Therefore, we compared lncRNA expression profiles of MLL-r infant ALL samples (n = 8) to those of the MLL-r older childhood ALL (>1 years) (n = 7). The expression profile separated the MLL-r infant ALL samples from the MLL-r childhood ALL (Fig. 3C). Remarkably, these lncRNAs clustered MLL-r infant ALL samples tightly as a group regardless of the type of MLL translocation. In general, the 11 upregulated and 35 downregulated lncRNAs in MLL-r infant ALL were found (Fig. 3C; fold change > 3, P < 0.01). The most upregulated ones are BC013423, uc003ysq.2 and NR_033661, whereas the most downregulated lncRNAs are CR601061, AF086267 and NR_024284. The signatures clearly separated MLL-r infant ALL patients with other older children, suggesting that the lncRNAs may function via an alternative mechanisms in the biology of this malignancy.

Functional classification and annotation of the candidate lncRNAs indicated that the dysregulated lncRNAs may participate in a variety of biological processes related to MLL-r ALL pathogenesis

The observations above indeed showed that MLL-r ALL displays a highly characteristic lncRNA expression profile and the unique lncRNA signatures potentially serve as a biomarker for the disease. We next endeavor to explore the universal regulatory mechanisms and the biological importance of the dysregulated candidate lncRNAs in MLL-r ALL; therefore, we performed mRNA microarrays with the same samples used above for lncRNAs array, as shown in Supplementary Material, Figure S2. Comparing with the previous published mRNA microarray data, our results are consistent with those presented in other studies (16,17), and the most differentially expressed genes were the HOXA genes (HOXA4, HOXA5, HOXA7, HOXA9 and HOXA10) and MEIS1. We then used lncRNA/mRNA coexpression networks to cluster thousands of transcripts into phenotypically relevant coexpression modules (Fig. 4A) (18,19). In this gene coexpression network, each gene corresponds to a node, and the edges connecting two nodes indicate either a positive or a negative strong correlation between lncRNAs or mRNAs (r > 0.85).

Figure 4.

LncRNA/mRNA coexpression networks. (A) Graphical view of the human gene coexpression network. Circles correspond to genes, diamonds correspond to lncRNAs and the edges correspond to coexpression links with an r-value >0.85. The most significant regions were marked with background colors, and the labels describe the main functions assigned. (B) The table shows the specific values of the degree of correlation of between the lncRNAs and the functionally related gene group for two coexpression data sets derived from each method at two specific standards: r > 0.85 and r > 0.75. Only lncRNAs with three or more related genes at the r > 0.85 standard were selected.

Figure 4.

LncRNA/mRNA coexpression networks. (A) Graphical view of the human gene coexpression network. Circles correspond to genes, diamonds correspond to lncRNAs and the edges correspond to coexpression links with an r-value >0.85. The most significant regions were marked with background colors, and the labels describe the main functions assigned. (B) The table shows the specific values of the degree of correlation of between the lncRNAs and the functionally related gene group for two coexpression data sets derived from each method at two specific standards: r > 0.85 and r > 0.75. Only lncRNAs with three or more related genes at the r > 0.85 standard were selected.

Because coexpression modules might correspond to biological pathways (20), and an lncRNA that is coexpressed with a set of genes with similar functions might have a related role, we analyzed the functional categories (21,22) of the 545 dysregulated genes (fold change > 2.0, P < 0.05) in the MLL-r samples compared with the MLL-wt samples. These results showed that the dysregulated genes belong to several functional clusters, including proteins involved in apoptosis, chromatin assembly, regulation of transcription, cell signaling, cell secretion and cytoplasmic vesicle or plasma membrane components, as well as zinc finger proteins. The most significant groups of coexpressed lncRNA/mRNAs are highlighted with different background colors in Figure 4A.

To further annotate the function of the dysregulated lncRNAs, we ranked the most similarly expressed lncRNAs in each gene cluster (Fig. 4B). The most noticeable clusters contain a number of genes that function in transcription regulation, several of which were directly affected by the MLL-r complex (3) (Fig. 4A, marked with blue background, Enrichment Score = 0.89). The top-ranked lncRNAs, AK090762, uc001mkl.1 and ENST00000418618, were significantly coexpressed with more than six genes within this cluster, implying crucial roles for these lncRNAs in MLL-r ALL pathogenesis through their interactions with transcription regulators. Another impressive cluster contains lncRNAs associated with a number of the membrane protein genes (Fig. 4A, marked with pink background, Enrichment Score = 1.18). For example, ENST00000443469, the most highly expressed lncRNA in our profile, was found to be correlated with the adjacent gene, LAMP5, a lysosomal-associated membrane protein (23). A third interesting cluster contains the lncRNA ENST00000369385, which is significantly coexpressed with 12 histone-related genes (Fig. 4A, marked with purple background). This lncRNA might be involved in MLL-r formation by regulating the expression of histones and maintaining the stability of the MLL gene and its fusion partners. The functional classification and annotation of the candidate lncRNAs in MLL-r ALL indicated that the dysregulated lncRNAs may interact with mRNA to participate in a variety of biological processes related to MLL-r ALL.

A set of lncRNAs is specifically linked to the HOXA genes and other MLL-r target genes in MLL-r ALL

It has been reported that the MLL-fusion proteins drive leukemogenesis by directly activating several important target genes, including HOXA clusters (24). We then asked whether some lncRNAs are linked to the HOXA genes or other MLL-r target genes in MLL-r ALL. To address this question, we first identified the genes that were significantly differentially expressed between MLL-r, MLL-wt and healthy control samples (Fig. 5A). We then analyzed the lncRNAs that were highly coexpressed with these genes (r > 0.75, P < 0.001). Figure 5B shows these target genes (yellow circles) and their closely related lncRNAs (diamonds). These findings imply that these lncRNAs might regulate different pathways in MLL-r ALL by interacting with the HOXA cluster and MEIS1, respectively. Importantly, some of the lncRNAs were coexpressed with more than one target gene of the MLL-r complex, for example, the lncRNAs ENST00000418618 and NR_027085 were 7.1-fold and 5.6-fold upregulated in the MLL-r samples, respectively, and both were significantly coexpressed with several genes of the HOXA cluster, including HOXA4, HOXA5, HOXA7, HOXA9 and HOXA10 (Fig. 5C). Another example is NR_033375, which was 43.3-fold upregulated in MLL-r ALL and coexpressed with several well-defined targets in MLL-r ALL, including MEIS1, CPEB2, PPP2R5C and RNF220 (Fig. 5C).

Figure 5.

Expression and sub-networks of MLL-r target genes and related lncRNAs. (A) Expression profiles of MLL-r target genes in MLL-r and MLL-wt samples. (B) Sub-networks with the MLL-r complex target genes. The nodes shown as circles correspond to genes, and those in yellow represent targets of the MLL-r complex; the diamond nodes correspond to lncRNAs, and the red nodes represent the lncRNAs that are modified by H3K79 methylation. (C) Correlation between the expression levels of lncRNAs in the microarray (ENST00000418618, NR_027085 and NR_033375) and crucial genes in MLL-r ALL. (D) Correlation between the expression levels of lncRNAs and the crucial genes confirmed by qRT-PCR. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5.

Expression and sub-networks of MLL-r target genes and related lncRNAs. (A) Expression profiles of MLL-r target genes in MLL-r and MLL-wt samples. (B) Sub-networks with the MLL-r complex target genes. The nodes shown as circles correspond to genes, and those in yellow represent targets of the MLL-r complex; the diamond nodes correspond to lncRNAs, and the red nodes represent the lncRNAs that are modified by H3K79 methylation. (C) Correlation between the expression levels of lncRNAs in the microarray (ENST00000418618, NR_027085 and NR_033375) and crucial genes in MLL-r ALL. (D) Correlation between the expression levels of lncRNAs and the crucial genes confirmed by qRT-PCR. *P < 0.05, **P < 0.01, ***P < 0.001.

To further confirm these coexpression relationships, we used qRT-PCR to validate the coexpression of the lncRNAs with the HOXA gene cluster. Three lncRNAs (ENST00000418618, NR_027085 and NR_033375) and five HOXA genes (HOXA4, HOXA5, HOXA7, HOXA9 and HOXA10), as well as MEIS1 transcripts, were validated in the clinical MLL-r and MLL-wt ALL samples. Consistent with the microarray data, we found that the expression of ENST00000418618 was correlated with that of HOXA4, HOXA5, HOXA7, HOXA9 and HOXA10 (r = 0.3010, 0.5639, 0.6353, 0.5853 and 0.5375, respectively, Fig. 5D and Supplementary Material, Fig. S3), NR_027085 was correlated with HOXA9 (r = 0.6085, Fig. 5D), and NR_033375 was strongly associated with MEIS1 (r = 0.6440, Fig. 5D).

To further investigate the correlation between lncRNA and mRNA, we chose two lncRNAs, ENST00000418618 and NR_027085, which were positively correlated with the HOXA genes, and detected whether these lncRNAs could regulate cell fate as HOXA genes. We first performed the knock-down study of the lncRNA ENST00000418618 and NR_027085 to characterize the effect of these two lncRNAs on MLL-r leukemia apoptosis and proliferation in MV4-11 cell line. Both lncRNAs were significantly downregulated after transfection of small-interfering RNAs (siRNAs) at 24 h using the qRT-PCR detection (Fig. 6A). The cell counting kit-8 (CCK-8) was used to detect cell proliferation and fluorescence-activated cell sorting analysis was used to assess the apoptosis. The result revealed a remarkable inhibition of cell proliferation (P < 0.01 at 48 h, and P < 0.001 from 72 to 96 h; Fig. 6B), and a significant increased apoptosis with either si-ENST00000418618 or si-NR_027085 compared with the control transfected with nonsense siRNA (the percentage of apoptotic cells increased from 17.72 to 44.56 and 76.39% in si-ENST00000418618 and si-NR_027085, respectively; P < 0.001; Fig. 6C).Thus, these results showed that the lncRNAs ENST00000418618 and NR_027085 inhibit apoptosis and promote proliferation in the MLL-r cell line MV4-11, which was similar to the HOXA cluster function, highlighting they have the functional synergism.

Figure 6.

ENST00000418618 and NR_027085 promote proliferation and inhibit apoptosis in MV4-11 cell line. (A) The two lncRNAs were downregulated after transfected with siRNA for 24 h. (B) The two lncRNAs significantly promoted cell proliferation in MV4-11 used CCK-8 assay kit (P < 0.01 at 48 h and P < 0.001 from 72 to 96 h). (C) The inhibition of the two lncRNAs induced apoptosis after transfected with siRNAs for 48 h. Data are shown as the mean ± SD from three independent assays. ***P < 0.001 compared with controls.

Figure 6.

ENST00000418618 and NR_027085 promote proliferation and inhibit apoptosis in MV4-11 cell line. (A) The two lncRNAs were downregulated after transfected with siRNA for 24 h. (B) The two lncRNAs significantly promoted cell proliferation in MV4-11 used CCK-8 assay kit (P < 0.01 at 48 h and P < 0.001 from 72 to 96 h). (C) The inhibition of the two lncRNAs induced apoptosis after transfected with siRNAs for 48 h. Data are shown as the mean ± SD from three independent assays. ***P < 0.001 compared with controls.

Taken together, these results indicate a correlation between the lncRNAs and the well-known MLL-r target genes, suggesting that these lncRNAs might contribute in MLL-r ALL by interacting with the MLL-r target genes. However, further functional studies are necessary to validate the essential of these dysregulated lncRNAs for the development of MLL-r ALL.

H3K79 methylation distribution at lncRNA promoters in MLL-r ALL cells is correlated with lncRNA transcription levels

Given that epigenetic mechanisms contribute to the cell-type specific expression of protein coding genes, we investigated whether these dysregulated lncRNAs are epigenetically regulated. In MLL-r ALL, genome-wide H3K79 methylation profiling has revealed that H3K79 methylation is abnormal, which can serve to distinguish MLL-r ALL from other types of leukemias (24,25). The loss of H3K79 methylation affects the expression of MLL-target loci and is detrimental to the leukemogenic activity of MLL-r ALL cells (24,25). Taking advantage of previously published genome-wide location data for H3K79 in MLL-r leukemia (26), we examined this histone methylation mark at the promoters of lncRNAs. We found a small group of unique lncRNAs that are highly differentially expressed in MLL-r ALL and that contain these modifications (Fig. 7A), suggesting a correlation between H3K79 methylation and lncRNAs. A previous report showed that MLL wild-type may play a role in MLL-r protein leukemogenesis through driving H3K4 methylation (3); we then compared the H3K79 with H3K4 ChIP-seq data (26) and found that four H3K79-methylated lncRNAs (ENST00000454595, ENST00000443469, uc004aec.2 and BC047080) were also regulated by H3K4 methylation, implying that their regulation may cooperate with the MLL wild-type protein. Next, we performed the chromatin immunoprecipitation (ChIP) to examine whether H3K79 methylation physically interacts with the lncRNA promoter regions in vitro in the MLL-r leukemic cell line RS4;11. Most of these lncRNAs were confirmed to be associated with H3K79 methylation ranging from −1.0 to +2.0 kb of the lncRNA transcription start sites (Fig. 7B). Collectively, the above results demonstrate that H3K79 methylation is correlated with the activation or repression of lncRNAs in MLL-r ALL.

Figure 7.

Dysregulated lncRNAs modified by H3K79 methylation in MLL-r ALL. (A) Regions of H3K79 methylation in the dysregulated lncRNAs in MLL-r and MLL-wt cells. H3K79 enrichment corresponds to the read depth. The red plot lines represent the transcriptional start site. (B) ChIP-PCR of H3K79-modified lncRNA. One percent of the input DNA was used as a positive control for PCR, and IgG-immunoprecipitated chromatin was used as a control for H3K79.

Figure 7.

Dysregulated lncRNAs modified by H3K79 methylation in MLL-r ALL. (A) Regions of H3K79 methylation in the dysregulated lncRNAs in MLL-r and MLL-wt cells. H3K79 enrichment corresponds to the read depth. The red plot lines represent the transcriptional start site. (B) ChIP-PCR of H3K79-modified lncRNA. One percent of the input DNA was used as a positive control for PCR, and IgG-immunoprecipitated chromatin was used as a control for H3K79.

Interestingly, we noted that some of the lncRNAs that are epigenetically regulated by H3K79 methylation, such as ENST00000443469, uc003kjn.1, BC047080 and NR_024154, were also coexpressed with genes that exhibited H3K79 methylation, such as LAMP5, CPEB2, JMJD1C and CAMK2D, as shown in Figure 5B (lncRNAs are colored in red and mRNAs are colored in yellow). To elucidate the potential pathogenic roles of these genes, we examined the Gene Ontology (GO) “biological process” classifications of the genes that were coexpressed with these H3K79 methylation-regulated lncRNAs (Supplementary Material, Fig. S4). Clustering of the GO classes that were enriched among the genes predicted to be coexpressed with these lncRNAs showed a number of genes involved in signaling pathways and biological regulation. For example, FYN, a member of the Src-family protein-tyrosine kinase which was reported involved in glucocorticoids resistance (27) and several other signaling-related genes including IL1B, STAT5A and PML (28–30), is represented in the signaling pathway networks (Supplementary Material, Table S1). And transcription factors including the HOXA4, HOXA5 and MEIS1, which regulate multiple biological processes, were also over-represented in biological regulation (Supplementary Material, Table S1). This observation implies that lncRNAs might be important to promote or augment the expression of H3K79 methylation-regulated genes, highlighting a previously unknown mechanism of MLL-r ALL pathogenesis.

DISCUSSION

LncRNAs, which are generally defined as non-coding RNAs that are greater than 200 nt in length, have recently been discovered as a special group of RNAs that plays a crucial role in multiple biological processes (31), such as embryogenesis, development (7) and disease (8). And the expression of certain lncRNAs provides tissue-specific or tumor-specific regulatory regions (32,33), which might serve as good biomarkers as other RNAs, such as mRNAs and miRNAs. Several novel lncRNAs have been revealed as not only candidate independent biomarkers but also as key regulators of cancer progression (8,12). However, thus far, no study on lncRNA associated with hematopoietic malignancies has been reported. In this study, we carried out a genome-wide lncRNA expression study on a subtype of leukemia, MLL-r ALL, which revealed several interesting findings. First, in an unsupervised analysis, the samples could be clustered into their biologic subtypes. All of the MLL-r ALL samples clustered together, separately from the MLL-wt samples or healthy controls. An obvious difference and unique lncRNA expression signatures were identified in each MLL-r subtype with different fusion partners. These results suggest that there are remarkable intragroup similarities and intergroup differences in lncRNA expression patterns among the populations and that these patterns are driven by the underlying biology of each group; these patterns are likely primarily determined by recurrent cytogenetic abnormalities in the leukemias. Given that lncRNA are key regulators of chromatin state for important biological processes (9), together with the evidences that many of the MLL fusion partners are part of transcriptional regulation networks that also function through chromatin remodeling (34), it could be perspective that further functional studies on these dysregulated lncRNAs would provide additional insights into the genetic makeup of the MLL-fusion partners. Importantly, we also identified a unique lncRNA expression pattern separating infant MLL-r ALL with other MLL-rearranged childhood ALL patients. Although the clinical usage of lncRNAs is only beginning to be explored and the advantages and disadvantages of lncRNAs serve as diagnostic agents compared with mRNAs and miRNAs need further validation, the findings in the study highlight the potential of these less well-characterized non-protein coding transcripts for the disease diagnosis. It is also worth noting that some dysregulated lncRNAs showed heterogeneity among MLL-r ALL patients, such as NR_033375 and BC035666, which might be caused by the difference between individuals. We speculated that there are some links between the heterogeneous lncRNA expression and clinical manifestations or subtypes. Although we cannot conclude significantly correlations between the expression levels of any dysregulated lncRNA and the classifications including different fusion partners, patient age or sex in this study, we believe that further studies with a larger cohort of patient samples might help explain the heterogeneity of the lncRNAs level in MLL-r ALL patients.

Secondly, to gain further insight into the biological importance of these dysregulated candidate lncRNAs, we established lncRNA/mRNA coexpression networks to cluster thousands of transcripts into phenotypically relevant coexpression modules. This analysis of the global coexpression of lncRNAs and mRNAs revealed that sets of abnormally expressed lncRNAs were associated with apoptosis, chromatin assembly, transcriptional factors, cell secretion, cytoplasmic vesicle or plasma membrane components and zinc finger proteins. Consistent with the previous report, the transcriptional factors we detected in the lncRNA/mRNA coexpression networks including the classic MLL-r ALL pathogenetic factors HOXA4, HOXA7, HOXA9, HOXA10 and MEIS1, implying that the lncRNAs might be important in driving MLL-r ALL. Notably, the largest group of lncRNA-associated mRNAs encodes membrane proteins. For example, ENST00000443469, the most highly expressed lncRNA in our profile, was found to be correlated with its adjacent gene, LAMP5, a lysosomal-associated membrane protein that is upregulated in non-activated human plasmacytoid dendritic cells and may be involved in endocytosis (23,35). And a study about the gene expression profiling of MLL-r infant ALL (36) showed that the LAMP5 (also called C20orf103) is one of the most upregulated genes in the MLL-r ALL compared with other infant ALL and pediatric precursor B-ALL, implying the impressive roles of LAMP5 and ENST00000443469 in the MLL-r ALL. In addition, the correlation between lncRNAs and membrane proteins provides an evidence for a link between membrane proteins and MLL-r ALL, and this linkage has not been investigated yet. However, several of the member proteins which were shown in the network have been reported important for identifying the subtypes of ALL as flow cytometry markers. For example, the expression of the member protein CD27 can discriminate subtypes of acute lymphoblastic leukemia (37), and CD99 could be used to assess minimal residual disease in T-lineage acute lymphoblastic leukemia (T-ALL) (38). Because the MLL-r ALL is a unique subtype compared with other ALLs, these membrane proteins identified in the study might serve as novel flow cytometry markers of MLL-r ALL. In addition, we also found several other lncRNAs that showed histone-related functions, implying that these lncRNAs might be involved in the origin of MLL-r leukemia due to aberrant histone methylation (39). Collectively, our results show the potential role of lncRNA in transcriptional regulation that may affect the MLL-r ALL progression.

Another interesting finding in this study was that a set of lncRNAs were highly correlated with one or more MLL-fusion target genes. It is well known that MLL-fusion genes can cause leukemia through the transcriptional upregulation of several downstream target genes, including the HOXA cluster and MEIS1 (3). The lncRNA, ENST00000418618, was found to be coexpressed with HOXA4, HOXA5, HOXA7, HOXA9 and HOXA10 and the lncRNA NR_027085 was found correlated with HOXA9. Intriguingly, a well-defined miRNA, miR-196b, was also reported to be coexpressed with the HOXA cluster genes (40), and this miRNA has been demonstrated as a key regulator of the early leukemogenesis of MLL-r leukemia (5). Our further experiments showed that both lncRNAs ENST00000418618 and NR_027085 promoted proliferation and inhibited apoptosis in the MLL-r leukemia cell lines, consistent with the function of the HOXA cluster. Based on this finding, we speculated that the coexpression between ENST00000418618 and the HOXA cluster might play a dominant role in the pathogenesis of MLL-r ALL. In addition, several other lncRNAs are correlated with specific MLL-r targets, such NR_033375 with MEIS1, CPEB2, PPP2R5C and RNF220, and ENST00000432488 with MEIS1. Because the lncRNAs could regulate gene expression through both cis and trans mechanisms and participate in diverse biological processes (41), these correlations indicate that the lncRNAs may be associated with the target genes downstream of the MLL-fusion protein and play indispensable roles for MLL-r ALL pathogenesis at the level of gene expression, although the detailed mechanism remains to be described.

Finally, we also investigated the distribution of H3K79 methylation in a number of lncRNA promoters and correlated this epigenetic feature with lncRNA transcription levels in MLL-r ALL leukemic cells. In recent years, an increasing number of studies have shown an important role for H3K79 methylation in the pathogenesis of MLL-r ALL. The major MLL-r-fusion genes, including MLL-AF4 (23) MLL-ENL (42) and MLL-AF9 (25), have been reported to enhance leukemogenic gene expression by modifying H3K79 methylation, and the activated expression of leukemogenic genes play a pivotal role in the MLL-r ALL. The results revealed in this study implied that lncRNA regulation by H3K79 methylation might be an important modulating pathways in MLL-r ALL, which was previously unknown as a mechanism of MLL-r ALL pathogenesis. Notably, the most upregulated lncRNA, ENST00000443469 was measured the analogous H3K79 regulation in the promoter area, and its adjacent gene LAMP5 (C20orf103) was identified as one of the top genes associated with H3K79 in MLL-r ALL in the previous report (24), which support the speculation that LAMP5 and its natural antisense-lncRNA ENST00000443469 might play significant roles in MLL-r leukemia. Besides, two of these lncRNAs, ENST00000454595 and ENST00000435695, were located close to or within the well-defined MLL-r ALL leukemogenic gene MEIS1 (43) and the cell-cycle-regulating gene CDK6 (44), suggesting that lncRNAs modified by H3K79 methylation may obtain their function through regulating these genes, which deserves for further study. GO “biological process” classifications analysis further indicated that the regulation of transcription and signaling pathways was the most enriched function of these lncRNAs, showing that these lncRNAs are potentially capable of influencing the expression of multiple genes. As H3K79 methylation occurs at a very early stage of MLL-r ALL pathogenesis, it can be speculated that the modification and activation of these lncRNAs may act as a primary regulator of other downstream genes that initiate and sustain the hematopoietic malignancies.

In summary, our study is the first to reveal that lncRNAs are dysregulated in childhood MLL-r ALL and that unique lncRNA expression patterns exist between each MLL-r subtype with different fusion partners as well as between infant MLL-r ALL with other MLL-r childhood ALL patients. These lncRNAs could be candidate biomarkers for this disease. Furthermore, by analyzing lncRNA/mRNA coexpression patterns, a set of dysregulated lncRNAs were found to be correlated with diverse biological progresses and to play key roles in the pathogenesis of MLL-r ALL. These lncRNAs can be grouped into two different sets: one set is coexpressed with MLL-fusion gene targets, and the other is epigenetically regulated by H3K79 methylation, which could induce the gene activation and lead the MLL-fusion gene to drive leukemogenesis. Altogether, our findings suggest that lncRNAs are dysregulated in MLL-r ALL and provide a novel insight into the mechanism linking lncRNA function with MLL-r-fusion proteins and leukemogenesis.

MATERIALS AND METHODS

Patient samples

Bone marrow samples were obtained from 47 patients at the time of the initial diagnosis including 18 MLL-r ALL and 29 patients without MLL-r (MLL-wt ALL). Among 47 samples, 15 patients with MLL-r ALL and 8 MLL-wt ALL patients were used for microarray and qRT-PCR analysis, and 18 MLL-r ALL (including 15 used in array) and 29 MLL-wt ALL (including 8 used in array) patients were used for qRT-PCR validation. The majority of the MLL-r ALL patients had B-cell ALL (B-ALL; n = 12), while six patients had T-ALL. The patient characteristics are detailed in Table 1. MLL-r ALL, including MLL-AF4/t(4;11), MLL-AF9/t(9;11) and MLL-ENL/t(11;19), was detected initially by fluorescence in situ hybridization and confirmed by PCR. All bone marrow samples were obtained from the First Affiliated Hospital of Sun Yat-sen University. The study was approved by the ethics committee of the affiliated hospitals of Sun Yat-sen University.

Table 1.

Pediatric ALL patients' characteristics

Type of sample Characteristics Median (range)/no. (%) 
MLL-r Age at diagnosis, years 1 (0.25–13) 
(N = 18) Fusion gene  
  MLL-AF4 7 (38.9) 
  MLL-AF9 3 (16.7) 
  MLL-ENL 4 (22.2) 
  N/A 4 (22.2) 
 Sex  
  Male 10 (55.6) 
  Female 8 (44.4) 
 Immunophenotype  
  B 13 (72.2) 
  T 5 (27.8) 
MLL-wt  Age at diagnosis, years 3 (0.5–13) 
(N = 29) Sex  
  Male 20 (69.0) 
  Female 9 (31.0) 
 Immunophenotype  
  B 21 (72.4) 
  T 8 (27.6) 
Type of sample Characteristics Median (range)/no. (%) 
MLL-r Age at diagnosis, years 1 (0.25–13) 
(N = 18) Fusion gene  
  MLL-AF4 7 (38.9) 
  MLL-AF9 3 (16.7) 
  MLL-ENL 4 (22.2) 
  N/A 4 (22.2) 
 Sex  
  Male 10 (55.6) 
  Female 8 (44.4) 
 Immunophenotype  
  B 13 (72.2) 
  T 5 (27.8) 
MLL-wt  Age at diagnosis, years 3 (0.5–13) 
(N = 29) Sex  
  Male 20 (69.0) 
  Female 9 (31.0) 
 Immunophenotype  
  B 21 (72.4) 
  T 8 (27.6) 

LncRNA microarray profiling and qRT-PCR analysis

Microarray analysis was performed with the Arraystar Human LncRNA Array v2.0 platform (Kangchen, Shanghai, China). The raw signal intensities were normalized with the quantile method by Gene Spring GX v11.5.1, and the results represent differentially expressed lncRNAs and mRNAs with statistical significance that passed Volcano Plot filtering (fold change > 2.0, P < 0.05, t-test). The resultant cDNA quantified in triplicate by SYBR® Premix Ex Taq™ II (Takara, Dalian, China)-based quantitative real-time PCR. The expression level of each lncRNA and mRNA was determined using the 2−△△Ct method. The specificity and reliability of the PCR were confirmed by sequencing the PCR product fragments. The results were normalized to GAPDH and presented as the fold change of each lncRNA/mRNA. All primers were shown in Supplementary Material, Table S2.

Coexpression network and GO analysis

The Mann–Whitney U-test and Spearman's correlation coefficient were used to determine the levels of differences and correlations. DAVID Bioinformatics Resources v6.7 was used to identify enriched functionally related gene groups (20,21). For GO analysis, the biological processes and molecular function categories were as defined in the Gene Ontology Consortium database (http://www.geneontology.org, last accessed on 5 February, 2014) (45), and over-represented functional themes present in the background were mapped on the GO hierarchy using the Cytoscape plugin BINGO (46).

Cell lines and cell cultures

The human MV4-11 and RS4;11 cells (ATCC) were maintained in the RPMI-1640 medium (HyClone, Beijing, China) supplemented with 10% fetal bovine serum (HyClone) at 37°C in a 5% CO2 atmosphere.

Cell transfection

The siRNAs against ENST00000418618 and NR_027085 were purchased from GenePharma (Shanghai, China). And the MV4-11 transfection used the Neon Transfection System (Invitrogen) according to the manufacturer's guidelines.

Cell proliferation CCK-8 assays

Cell proliferation was assessed using the Cell Counting Kit-8 (CCK-8, Dojindo Molecular Technologies, Shanghai, China). For the CCK-8 assay, cells transfected with siRNAs (100 nm) were seeded at a density of 30 000 cells per well in 100 µl of full medium in 96-well plates. And the absorbance was measured at wavelengths of 480 and 630 nm on consecutive 0, 24, 48, 72 and 96 h.

Apoptosis assay (fluorescence-activated cell sorting analysis)

Cells were transfected with the siRNAs for 48 h. The cells were stained using Annexin V/FITC and propidium iodide (PI) (Lianke, China) and then analyzed by flow cytometry (BD, USA).

H3K79 methylation region identification and ChIP

To identify H3K79 methylation regions, ChIP-seq data targeting H3K79 methylation of leukemia cell lines were analyzed. ChIP was performed with the EZ ChIP™ Chromatin Immunoprecipitation Kit (Upstate, Billerica MA, USA). The PCR amplification primers used are listed in Supplementary Material, Table S2.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

AUTHORS’ CONTRIBUTIONS

K.F. and B.W.H. designed and performed the research, analyzed data and wrote the article. Z.H.C., K.Y.L., C.W.Z. and X.J.L. performed the research and analyzed data. J.H.L. and X.Q.L. analyzed data. Y.Q.C. designed the research and wrote the article.

FUNDING

This work was supported by the funds from National Science and Technology Department (973, 2011CB8113015 and 2011CBA0110) and National Science Foundation of China (No. 81270629 and 81300398). This work is also supported in part by the Guangdong Province Key Laboratory of Computational Science and the Guangdong Province Computational Science Innovative Research Team.

Conflict of Interest statement. None declared.

REFERENCES

1
Pieters
R.
Schrappe
M.
De Lorenzo
P.
Hann
I.
De Rossi
G.
Felice
M.
Hovi
L.
LeBlanc
T.
Szczepanski
T.
Ferster
A.
, et al.  . 
A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (Interfant-99): an observational study and a multicentre randomised trial
Lancet
 , 
2007
, vol. 
370
 (pg. 
240
-
250
)
2
Pui
C.H.
Robison
L.L.
Look
A.T.
Acute lymphoblastic leukaemia
Lancet
 , 
2008
, vol. 
371
 (pg. 
1030
-
1043
)
3
Muntean
A.G.
Hess
J.L.
The pathogenesis of mixed-lineage leukemia
Annu. Rev. Pathol.
 , 
2012
, vol. 
7
 (pg. 
283
-
301
)
4
O'Connell
R.M.
Chaudhuri
A.A.
Rao
D.S.
Gibson
W.S.
Balazs
A.B.
Baltimore
D.
MicroRNAs enriched in hematopoietic stem cells differentially regulate long-term hematopoietic output
Proc. Natl Acad. Sci. USA
 , 
2010
, vol. 
107
 (pg. 
14235
-
14240
)
5
Popovic
R.
Riesbeck
L.E.
Velu
C.S.
Chaubey
A.
Zhang
J.
Achille
N.J.
Erfurth
F.E.
Eaton
K.
Lu
J.
Grimes
H.L.
, et al.  . 
Regulation of mir-196b by MLL and its overexpression by MLL fusions contributes to immortalization
Blood
 , 
2009
, vol. 
113
 (pg. 
3314
-
3322
)
6
Mi
S.
Li
Z.
Chen
P.
He
C.
Cao
D.
Elkahloun
A.
Lu
J.
Pelloso
L.A.
Wunderlich
M.
Huang
H.
, et al.  . 
Aberrant overexpression and function of the miR-17–92 cluster in MLL-rearranged acute leukemia
Proc. Natl Acad. Sci. USA
 , 
2010
, vol. 
107
 (pg. 
3710
-
3715
)
7
Ulitsky
I.
Shkumatava
A.
Jan
C.H.
Sive
H.
Bartel
D.P.
Conserved function of lincRNAs in vertebrate embryonic development despite rapid sequence evolution
Cell
 , 
2011
, vol. 
147
 (pg. 
1537
-
1550
)
8
Wapinski
O.
Chang
H.Y.
Long noncoding RNAs and human disease
Trends Cell Biol.
 , 
2011
, vol. 
21
 (pg. 
354
-
361
)
9
Chu
C.
Qu
K.
Zhong
F.L.
Artandi
S.E.
Chang
H.Y.
Genomic maps of long noncoding RNA occupancy reveal principles of RNA-chromatin interactions
Mol. Cell
 , 
2011
, vol. 
44
 (pg. 
667
-
678
)
10
Orom
U.A.
Derrien
T.
Beringer
M.
Gumireddy
K.
Gardini
A.
Bussotti
G.
Lai
F.
Zytnicki
M.
Notredame
C.
Huang
Q.
, et al.  . 
Long noncoding RNAs with enhancer-like function in human cells
Cell
 , 
2010
, vol. 
143
 (pg. 
46
-
58
)
11
Zhang
X.
Lian
Z.
Padden
C.
Gerstein
M.B.
Rozowsky
J.
Snyder
M.
Gingeras
T.R.
Kapranov
P.
Weissman
S.M.
Newburger
P.E.
A myelopoiesis-associated regulatory intergenic noncoding RNA transcript within the human HOXA cluster
Blood
 , 
2009
, vol. 
113
 (pg. 
2526
-
2534
)
12
Spizzo
R.
Almeida
M.I.
Colombatti
A.
Calin
G.A.
Long non-coding RNAs and cancer: a new frontier of translational research?
Oncogene
 , 
2012
, vol. 
31
 (pg. 
4577
-
4587
)
13
Bertani
S.
Sauer
S.
Bolotin
E.
Sauer
F.
The noncoding RNA Mistral activates Hoxa6 and Hoxa7 expression and stem cell differentiation by recruiting MLL1 to chromatin
Mol. Cell
 , 
2011
, vol. 
43
 (pg. 
1040
-
1046
)
14
Schwieger
M.
Schuler
A.
Forster
M.
Engelmann
A.
Arnold
M.A.
Delwel
R.
Valk
P.J.
Lohler
J.
Slany
R.K.
Olson
E.N.
, et al.  . 
Homing and invasiveness of MLL/ENL leukemic cells is regulated by MEF2C
Blood
 , 
2009
, vol. 
114
 (pg. 
2476
-
2488
)
15
Zheng
Y.S.
Zhang
H.
Zhang
X.J.
Feng
D.D.
Luo
X.Q.
Zeng
C.W.
Lin
K.Y.
Zhou
H.
Qu
L.H.
Zhang
P.
, et al.  . 
MiR-100 regulates cell differentiation and survival by targeting RBSP3, a phosphatase-like tumor suppressor in acute myeloid leukemia
Oncogene
 , 
2012
, vol. 
31
 (pg. 
80
-
92
)
16
Armstrong
S.A.
Staunton
J.E.
Silverman
L.B.
Pieters
R.
den Boer
M.L.
Minden
M.D.
Sallan
S.E.
Lander
E.S.
Golub
T.R.
Korsmeyer
S.J.
MLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia
Nat. Genet.
 , 
2002
, vol. 
30
 (pg. 
41
-
47
)
17
Zangrando
A.
Dell'Orto
M.C.
Te
K.G.
Basso
G.
MLL rearrangements in pediatric acute lymphoblastic and myeloblastic leukemias: MLL specific and lineage specific signatures
BMC Med. Genomics
 , 
2009
, vol. 
2
 pg. 
36
 
18
Pujana
M.A.
Han
J.D.
Starita
L.M.
Stevens
K.N.
Tewari
M.
Ahn
J.S.
Rennert
G.
Moreno
V.
Kirchhoff
T.
Gold
B.
, et al.  . 
Network modeling links breast cancer susceptibility and centrosome dysfunction
Nat. Genet.
 , 
2007
, vol. 
39
 (pg. 
1338
-
1349
)
19
Liao
Q.
Liu
C.
Yuan
X.
Kang
S.
Miao
R.
Xiao
H.
Zhao
G.
Luo
H.
Bu
D.
Zhao
H.
, et al.  . 
Large-scale prediction of long non-coding RNA functions in a coding-non-coding gene co-expression network
Nucleic Acids Res.
 , 
2011
, vol. 
39
 (pg. 
3864
-
3878
)
20
Nayak
R.R.
Kearns
M.
Spielman
R.S.
Cheung
V.G.
Coexpression network based on natural variation in human gene expression reveals gene interactions and functions
Genome Res.
 , 
2009
, vol. 
19
 (pg. 
1953
-
1962
)
21
Huang
D.W.
Sherman
B.T.
Lempicki
R.A.
Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources
Nat. Protoc.
 , 
2009
, vol. 
4
 (pg. 
44
-
57
)
22
Huang
D.W.
Sherman
B.T.
Lempicki
R.A.
Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists
Nucleic Acids Res.
 , 
2009
, vol. 
37
 (pg. 
1
-
13
)
23
David
A.
Tiveron
M.C.
Defays
A.
Beclin
C.
Camosseto
V.
Gatti
E.
Cremer
H.
Pierre
P.
BAD-LAMP defines a subset of early endocytic organelles in subpopulations of cortical projection neurons
J. Cell Sci.
 , 
2007
, vol. 
120
 (pg. 
353
-
365
)
24
Krivtsov
A.V.
Feng
Z.
Lemieux
M.E.
Faber
J.
Vempati
S.
Sinha
A.U.
Xia
X.
Jesneck
J.
Bracken
A.P.
Silverman
L.B.
, et al.  . 
H3K79 methylation profiles define murine and human MLL-AF4 leukemias
Cancer Cell
 , 
2008
, vol. 
14
 (pg. 
355
-
368
)
25
Bernt
K.M.
Zhu
N.
Sinha
A.U.
Vempati
S.
Faber
J.
Krivtsov
A.V.
Feng
Z.
Punt
N.
Daigle
A.
Bullinger
L.
, et al.  . 
MLL-rearranged leukemia is dependent on aberrant H3K79 methylation by DOT1L
Cancer Cell
 , 
2011
, vol. 
20
 (pg. 
66
-
78
)
26
Guenther
M.G.
Lawton
L.N.
Rozovskaia
T.
Frampton
G.M.
Levine
S.S.
Volkert
T.L.
Croce
C.M.
Nakamura
T.
Canaani
E.
Young
R.A.
Aberrant chromatin at genes encoding stem cell regulators in human mixed-lineage leukemia
Genes Dev.
 , 
2008
, vol. 
22
 (pg. 
3403
-
3408
)
27
Spijkers-Hagelstein
J.A.
Mimoso
P.S.
Schneider
P.
Pieters
R.
Stam
R.W.
Src kinase-induced phosphorylation of annexin A2 mediates glucocorticoid resistance in MLL-rearranged infant acute lymphoblastic leukemia
Leukemia
 , 
2013
, vol. 
27
 (pg. 
1063
-
1071
)
28
Ennas
M.G.
Moore
P.S.
Zucca
M.
Angelucci
E.
Cabras
M.G.
Melis
M.
Gabbas
A.
Serpe
R.
Madeddu
C.
Scarpa
A.
, et al.  . 
Interleukin-1B (IL1B) and interleukin-6 (IL6) gene polymorphisms are associated with risk of chronic lymphocytic leukaemia
Hematol. Oncol.
 , 
2008
, vol. 
26
 (pg. 
98
-
103
)
29
Joliot
V.
Cormier
F.
Medyouf
H.
Alcalde
H.
Ghysdael
J.
Constitutive STAT5 activation specifically cooperates with the loss of p53 function in B-cell lymphomagenesis
Oncogene
 , 
2006
, vol. 
25
 (pg. 
4573
-
4584
)
30
Puccetti
E.
Beissert
T.
Guller
S.
Li
J.E.
Hoelzer
D.
Ottmann
O.G.
Ruthardt
M.
Leukemia-associated translocation products able to activate RAS modify PML and render cells sensitive to arsenic-induced apoptosis
Oncogene
 , 
2003
, vol. 
22
 (pg. 
6900
-
6908
)
31
Guttman
M.
Amit
I.
Garber
M.
French
C.
Lin
M.F.
Feldser
D.
Huarte
M.
Zuk
O.
Carey
B.W.
Cassady
J.P.
, et al.  . 
Chromatin signature reveals over a thousand highly conserved large non-coding RNAs in mammals
Nature
 , 
2009
, vol. 
458
 (pg. 
223
-
227
)
32
Cabili
M.N.
Trapnell
C.
Goff
L.
Koziol
M.
Tazon-Vega
B.
Regev
A.
Rinn
J.L.
Integrative annotation of human large intergenic noncoding RNAs reveals global properties and specific subclasses
Genes Dev.
 , 
2011
, vol. 
25
 (pg. 
1915
-
1927
)
33
Qi
P.
Du
X.
The long non-coding RNAs, a new cancer diagnostic and therapeutic gold mine
Mod. Pathol.
 , 
2013
, vol. 
26
 (pg. 
155
-
165
)
34
Den Boer
M.L.
van Slegtenhorst
M.
De Menezes
R.X.
Cheok
M.H.
Buijs-Gladdines
J.G.
Peters
S.T.
Van Zutven
L.J.
Beverloo
H.B.
Van der Spek
P.J.
Escherich
G.
, et al.  . 
A subtype of childhood acute lymphoblastic leukaemia with poor treatment outcome: a genome-wide classification study
Lancet Oncol.
 , 
2009
, vol. 
10
 (pg. 
125
-
134
)
35
Defays
A.
David
A.
de Gassart
A.
De Angelis
R.F.
Wenger
T.
Camossetto
V.
Brousset
P.
Petrella
T.
Dalod
M.
Gatti
E.
, et al.  . 
BAD-LAMP is a novel biomarker of nonactivated human plasmacytoid dendritic cells
Blood
 , 
2011
, vol. 
118
 (pg. 
609
-
617
)
36
Stam
R.W.
Schneider
P.
Hagelstein
J.A.
van der Linden
M.H.
Stumpel
D.J.
de Menezes
R.X.
de Lorenzo
P.
Valsecchi
M.G.
Pieters
R.
Gene expression profiling-based dissection of MLL translocated and MLL germline acute lymphoblastic leukemia in infants
Blood
 , 
2010
, vol. 
115
 (pg. 
2835
-
2844
)
37
Vaskova
M.
Mejstrikova
E.
Kalina
T.
Martinkova
P.
Omelka
M.
Trka
J.
Stary
J.
Hrusak
O.
Transfer of genomics information to flow cytometry: expression of CD27 and CD44 discriminates subtypes of acute lymphoblastic leukemia
Leukemia
 , 
2005
, vol. 
19
 (pg. 
876
-
878
)
38
Dworzak
M.N.
Froschl
G.
Printz
D.
Zen
L.D.
Gaipa
G.
Ratei
R.
Basso
G.
Biondi
A.
Ludwig
W.D.
Gadner
H.
CD99 expression in T-lineage ALL: implications for flow cytometric detection of minimal residual disease
Leukemia
 , 
2004
, vol. 
18
 (pg. 
703
-
708
)
39
Krivtsov
A.V.
Armstrong
S.A.
MLL translocations, histone modifications and leukaemia stem-cell development
Nat. Rev. Cancer
 , 
2007
, vol. 
7
 (pg. 
823
-
833
)
40
Schotte
D.
Lange-Turenhout
E.A.
Stumpel
D.J.
Stam
R.W.
Buijs-Gladdines
J.G.
Meijerink
J.P.
Pieters
R.
Den Boer
M.L.
Expression of miR-196b is not exclusively MLL-driven but is especially linked to activation of HOXA genes in pediatric acute lymphoblastic leukemia
Haematologica
 , 
2010
, vol. 
95
 (pg. 
1675
-
1682
)
41
Guttman
M.
Rinn
J.L.
Modular regulatory principles of large non-coding RNAs
Nature
 , 
2012
, vol. 
482
 (pg. 
339
-
346
)
42
Mueller
D.
Bach
C.
Zeisig
D.
Garcia-Cuellar
M.P.
Monroe
S.
Sreekumar
A.
Zhou
R.
Nesvizhskii
A.
Chinnaiyan
A.
Hess
J.L.
, et al.  . 
A role for the MLL fusion partner ENL in transcriptional elongation and chromatin modification
Blood
 , 
2007
, vol. 
110
 (pg. 
4445
-
4454
)
43
Kumar
A.R.
Li
Q.
Hudson
W.A.
Chen
W.
Sam
T.
Yao
Q.
Lund
E.A.
Wu
B.
Kowal
B.J.
Kersey
J.H.
A role for MEIS1 in MLL-fusion gene leukemia
Blood
 , 
2009
, vol. 
113
 (pg. 
1756
-
1758
)
44
Grossel
M.J.
Hinds
P.W.
From cell cycle to differentiation: an expanding role for cdk6
Cell Cycle
 , 
2006
, vol. 
5
 (pg. 
266
-
270
)
45
Ashburner
M.
Ball
C.A.
Blake
J.A.
Botstein
D.
Butler
H.
Cherry
J.M.
Davis
A.P.
Dolinski
K.
Dwight
S.S.
Eppig
J.T.
, et al.  . 
Gene ontology: tool for the unification of biology. The Gene Ontology Consortium
Nat. Genet.
 , 
2000
, vol. 
25
 (pg. 
25
-
29
)
46
Maere
S.
Heymans
K.
Kuiper
M.
BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks
Bioinformatics
 , 
2005
, vol. 
21
 (pg. 
3448
-
3449
)

Author notes

K.F. and B.W.H. contributed equally to this work.

Supplementary data