MINCR is a MYC-induced lncRNA able to modulate MYC’s transcriptional network in Burkitt lymphoma cells

Significance

Gains of the MYC gene are the most common imbalances in cancer and are associated with poor prognosis, particularly in B-cell lymphoma. Recent advances in DNA sequencing have revealed the existence of thousands of long noncoding RNAs (lncRNAs) with unknown functional relevance. We have here identified a MYC-regulated lncRNA that we named MYC-induced long noncoding RNA (MINCR) that has a strong correlation with MYC expression in cancer. We show that MINCR is functional and controls cell cycle progression by influencing the expression of MYC-regulated cell cycle genes. MINCR is, therefore, a novel player in MYC’s transcriptional network, with the potential to open new therapeutic windows in the fight against malignant lymphoma and, possibly, all cancers that rely on MYC expression.

Abstract

Despite the established role of the transcription factor MYC in cancer, little is known about the impact of a new class of transcriptional regulators, the long noncoding RNAs (lncRNAs), on MYC ability to influence the cellular transcriptome. Here, we have intersected RNA-sequencing data from two MYC-inducible cell lines and a cohort of 91 B-cell lymphomas with or without genetic variants resulting in MYC overexpression. We identified 13 lncRNAs differentially expressed in IG-MYC-positive Burkitt lymphoma and regulated in the same direction by MYC in the model cell lines. Among them, we focused on a lncRNA that we named MYC-induced long noncoding RNA (MINCR), showing a strong correlation with MYC expression in MYC-positive lymphomas. To understand its cellular role, we performed RNAi and found that MINCR knockdown is associated with an impairment in cell cycle progression. Differential gene expression analysis after RNAi showed a significant enrichment of cell cycle genes among the genes down-regulated after MINCR knockdown. Interestingly, these genes are enriched in MYC binding sites in their promoters, suggesting that MINCR acts as a modulator of the MYC transcriptional program. Accordingly, MINCR knockdown was associated with a reduction in MYC binding to the promoters of selected cell cycle genes. Finally, we show that down-regulation of Aurora kinases A and B and chromatin licensing and DNA replication factor 1 may explain the reduction in cellular proliferation observed on MINCR knockdown. We, therefore, suggest that MINCR is a newly identified player in the MYC transcriptional network able to control the expression of cell cycle genes.

MYC is a transcription factor belonging to the basic helix–loop–helix zipper family that was originally identified in Burkitt lymphoma (BL) because of a chromosomal translocation that juxtaposes the MYC oncogene with one of three immunoglobulin (Ig) loci (1–3). In BL, the deregulation of the oncogenic transcription factor MYC is considered to be the major driving force in lymphoma development (4, 5). MYC overexpression is not restricted to BL and has been found to be a common feature of many tumors of different origins (6). Amplification of the genomic region containing MYC was actually found to be the most frequent somatic copy number alteration in a panel of 3,131 cancer specimens belonging to 26 histological types (7).

The correlation between the changes in gene expression and the phenotypic changes induced by MYC has been the focus of several reports. Although activation of genes involved in protein biosynthesis, energy metabolism, and cell cycle regulation by MYC is consistent with its ability to drive cell growth and proliferation, the search for “the oncogenic” MYC target genes has been relatively frustrating. This inconsistency is partly due to the fact that MYC is able to modulate the transcription of up to 15% of all protein coding genes (8, 9). Another layer of complexity is added by the finding that MYC can interact with several proteins able to induce chromatin modifications associated with active transcription, like TRRAP (10), GCN5 (11), and TIP60 (12). MYC was also found bound to components of the P-TEFb complex and to influence general transcriptional pause release (13). Recent findings suggest that MYC may behave as an amplifier of a cell/tissue-specific transcriptional program (14, 15). Nevertheless, this effect cannot be observed universally and has been suggested to be an indirect effect of the regulation by MYC of specific sets of genes (16, 17).

Next generation sequencing experiments have revealed that up to 70% of the human genome is actively transcribed from one or both strands (18), with only 2% of the genome being dedicated to coding genes (19, 20). The functional relevance of this large amount of noncoding transcripts remains mostly unknown, but the number of reports showing effects on cellular physiology and aberrant expression in several diseases is quickly increasing (21–25).

The largest portion of the mammalian noncoding transcriptome is made of a class of heterogeneous RNAs: the long noncoding RNAs (lncRNAs) (26). The GENCODE v7 catalog of human lncRNAs estimates a total number of 14,880 lncRNAs (27), but with the rapid increase in RNA sequencing (RNA-seq) data, the number of lncRNAs may soon overcome that of coding genes. lncRNAs have been associated with diverse functions, including X-chromosome inactivation (28, 29), differentiation (30), modulation of p53 activity (21), and formation of subnuclear domains (31, 32). Several lncRNAs were found to recruit RNA binding proteins and affect histone modifications, with immediate ramifications on gene regulation (33, 34). lncRNAs have also been shown to be differentially expressed in specific cancer types, where they can have a key role in the control of cellular proliferation (22, 23). Evidence for a strong requirement of lncRNA expression for normal mammalian development and physiology has also come from functional studies in knockout mouse models (24).

Much work has been done on the analysis of specific patterns of transcription associated to different subtypes of B-cell lymphomas (35–37). Theses analyses typically addressed the expression level of annotated coding genes by means of microarray technology. In this study, we used RNA-seq to focus on gene expression changes associated to lncRNAs, with the aim of identifying MYC-regulated lncRNAs potentially involved in lymphoma development. To do this, we analyzed RNA-seq data of samples from the major subtypes of mature B-cell lymphomas, namely BL, diffuse large B-cell lymphoma (DLBCL), and follicular lymphoma (FL), compared with data from normal germinal center (GC) B cells. These data were intersected with data coming from cell lines expressing MYC in an inducible manner to identify MYC-regulated lncRNAs. This approach led to the identification of a group of lncRNAs differentially expressed in BL relative to normal GC B cells and regulated by MYC in the same direction in two cell model systems. We focused our analysis on an up-regulated transcript showing the highest positive correlation with MYC expression, which we named MYC-induced long noncoding RNA (MINCR). By using RNAi, we show that MINCR knockdown is associated with reduced cellular proliferation in three different cell types. In line with this finding, RNA-seq of cells knocked down for MINCR expression showed a significant reduction in the expression of genes functionally important for cell cycle progression.

Results

Identification of BL-Specific MYC-Regulated lncRNAs.

With the aim of identifying MYC-regulated lncRNAs that could play a driving role in lymphoma development, we set up a bioinformatic pipeline to screen for lncRNAs differentially regulated in MYC-inducible model cell lines and a set of BL samples. In particular, we used RNA-seq data obtained from (i) hT-RPE-MycER cells, an immortalized retinal pigment epithelial cell line expressing the MycER fusion protein (a cell line shown to be a relevant tool for the study of several aspects of MYC activity in oncogenic transformation) (38, 39); (ii) P493-6 cells, an immortalized B-lymphoblastic cell line carrying a tetracyclin-inducible MYC construct that was shown to be a good model to study MYC-induced lymphomas (40, 41); and (iii) 91 GC B cell-derived lymphoma samples subjected to RNA-seq in the framework of the International Cancer Genome Consortium project on malignant lymphoma [ICGC MMML-Seq (International Cancer Genome Consortium Molecular Mechanisms in Malignant Lymphoma by Sequencing)], including 16 BLs, 35 DLBCLs, 40 FLs, and 4 controls (normal GC B-cell samples) (Table S1).

Table S1.

List of the lymphoma samples of the ICGC MMML-Seq Cohort used in this study

Subgrouping in this study*	PID	Age at first diagnosis (y)	Sex	Tumor cell content (%)	Classification based on gene expression	Histopathological panel diagnosis	MYC_STATUS analyzed with FISH
BL	4119027	Below 18	Male	95	mBL	BL	IG-MYC–positive, SH
BL	4125240	Below 18	Male	80	mBL	BL	IG-MYC–positive, SH
BL	4182393	Below 18	Male	90	mBL	BL	IG-MYC–positive, SH
BL	4190495	Below 18	Male	70	mBL	DLBCL	IG-MYC–positive, SH
BL	4112512	Below 18	Female	80	mBL	BL	IG-MYC–positive, SH
BL	4177434	Below 18	Female	85	mBL	Intermediate	IG-MYC–positive, SH
BL	4133511	Below 18	Male	80	mBL	BL	IG-MYC–positive, SH
BL	4177856	Below 18	Male	90	mBL	BL	IG-MYC–positive, SH
BL	4189998	Below 18	Male	75	mBL	BL	IG-MYC–positive, SH
BL	4193278	Below 18	Male	85	mBL	BL	IG-MYC–positive, SH
BL	4194218	Below 18	Male	70	mBL	BL	IG-MYC–positive, SH
BL	4194891	Below 18	Male	85	mBL	BL	IG-MYC–positive, SH
BL	4127766	Below 18	Male	85	mBL	BL	IG-MYC–positive, SH
BL	4130003	Below 18	Male	85	mBL	BL	IG-MYC–positive, SH
BL	4146289	Below 18	Male	60	mBL	BL	IG-MYC–positive, SH
BL	4110996	Below 18	Male	80	mBL	BL	IG-MYC–positive, SH
Non-BL	4116738	Below 18	Male	65	Non-mBL	DLBCL	MYC-negative
Non-BL	4105105	40	Female	70	Non-mBL	FL	MYC-negative
Non-BL	4108101	66	Male	70	Non-mBL	DLBCL	MYC-negative
Non-BL	4121361	74	Male	80	Non-mBL	FL	MYC-negative
Non-BL	4134005	67	Male	80	Non-mBL	FL	MYC-negative
Non-BL	4135099	49	Male	80	Non-mBL	DLBCL	MYC-negative
Non-BL	4163639	75	Female	80	Non-mBL	DLBCL	MYC-negative
Non-BL	4101316	74	Female	60	Non-mBL	DLBCL	MYC-negative
Non-BL	4104893	Below 18	Male	85	Non-mBL	DLBCL	MYC-negative
Non-BL	4158726	48	Male	70	Non-mBL	FL	MYC-negative
Non-BL	4159170	43	Male	70	Non-mBL	FL	MYC-negative
Non-BL	4175837	74	Female	70	Non-mBL	FL	MYC-negative
Non-BL	4188900	76	Male	75	Non-mBL	FL	MYC-negative
Non-BL	4189200	51	Female	60	Non-mBL	FL	MYC-negative
Non-BL	4131257	72	Male	80	Non-mBL	DLBCL	MYC-negative
Non-BL	4160468	62	Male	75	Non-mBL	FL	MYC-negative
Non-BL	4171706	70	Female	70	Non-mBL	FL	MYC-negative
Non-BL	4174905	72	Male	75	Non-mBL	FL	MYC-negative
Non-BL	4177987	71	Male	70	Non-mBL	FL	MYC-negative
Non-BL	4102009	64	Male	75	Non-mBL	DLBCL	MYC-negative
Non-BL	4107137	59	Male	70	Non-mBL	DLBCL	MYC-negative
Non-BL	4120193	41	Female	60	Non-mBL	DLBCL	MYC-negative
Non-BL	4166706	62	Male	70	Non-mBL	DLBCL	MYC-negative
Non-BL	4170686	56	Male	80	Non-mBL	FL	MYC-negative
Non-BL	4177601	52	Female	60	Non-mBL	FL	MYC-negative
Non-BL	4184094	57	Female	75	Non-mBL	DLBCL	MYC-negative
Non-BL	4112447	46	Male	70	Non-mBL	FL	MYC-negative
Non-BL	4113825	74	Female	70	Non-mBL	FL	MYC-negative
Non-BL	4115001	70	Female	75	Non-mBL	DLBCL	MYC-negative
Non-BL	4119279	62	Female	80	Non-mBL	DLBCL	MYC-negative
Non-BL	4120157	46	Male	75	Non-mBL	DLBCL	MYC-negative
Non-BL	4134434	84	Male	85	Non-mBL	DLBCL	MYC-negative
Non-BL	4145056	67	Female	75	Non-mBL	FL	MYC-negative
Non-BL	4149246	41	Male	65	Non-mBL	FL	MYC-negative
Non-BL	4157186	74	Male	75	Non-mBL	DLBCL	MYC-negative
Non-BL	4177810	47	Female	75	Non-mBL	FL	MYC-negative
Non-BL	4178655	50	Male	80	Non-mBL	FL	MYC-negative
Non-BL	4189035	46	Male	60	Non-mBL	DLBCL	MYC-negative
Non-BL	4198542	68	Female	60	Non-mBL	FL	MYC-negative
Non-BL	4166503	43	Male	80	Non-mBL	FL	MYC-negative
Non-BL	4110378	69	Female	70	Non-mBL	FL	MYC-negative
Non-BL	4109956	63	Male	75	Non-mBL	FL	MYC-negative
Non-BL	4142761	54	Female	80	Non-mBL	FL	MYC-negative
Non-BL	4128477	54	Male	60	Non-mBL	FL	MYC-negative
Non-BL	4184011	61	Male	60	Non-mBL	FL	MYC-negative
Non-BL	4177844	68	Male	70	Non-mBL	FL	MYC-negative
Non-BL	4138885	55	Female	65	Non-mBL	FL	MYC-negative
Non-BL	4181460	70	Female	75	Non-mBL	DLBCL	MYC-negative
Non-BL	4177406	72	Female	65	Non-mBL	FL	MYC-negative
Non-BL	4124188	73	Male	80	Non-mBL	DLBCL	MYC-negative
Non-BL	4138527	85	Female	75	Non-mBL	DLBCL	MYC-negative
Non-BL	4147968	65	Female	85	Non-mBL	DLBCL	MYC-negative
Non-BL	4108992	68	Male	65	Non-mBL	FL	MYC-negative
Non-BL	4113191	62	Female	75	Non-mBL	FL	MYC-negative
Non-BL	4124542	72	Female	70	Non-mBL	FL	MYC-negative
Non-BL	4131095	56	Female	60	Non-mBL	FL	MYC-negative
Non-BL	4145528	62	Male	80	Non-mBL	DLBCL	MYC-negative
Non-BL	4147081	42	Female	80	Non-mBL	FL	MYC-negative
Non-BL	4161781	36	Female	70	Non-mBL	DLBCL	MYC-negative
Non-BL	4165379	64	Female	70	Non-mBL	FL	MYC-negative
Non-BL	4193638	55	Female	70	Non-mBL	DLBCL	MYC-negative
Non-BL	4199714	53	Female	70	Non-mBL	DLBCL	MYC-negative
Non-BL	4128852	57	Female	80	Non-mBL	DLBCL	MYC-negative
Non-BL	4140531	66	Male	70	Non-mBL	DLBCL	MYC-negative
Non-BL	4158268	54	Female	30	Non-mBL	FL	MYC-negative
Non–BL-MYC	4133263	40	Male	85	m-Intermediate	DLBCL	IG-MYC–positive, SH
Non–BL-MYC	4188879	52	Male	70	Non-mBL	DLBCL	IG-MYC–positive, SH
Non–BL-MYC	4187640	58	Male	80	Non-mBL	FL	Non–IG-MYC–positive, DH
Non–BL-MYC	4130194	55	Female	80	Non-mBL	DLBCL	Non–IG-MYC–positive, DH
Non–BL-MYC	4103141	60	Female	60	Non-mBL	FL	Non–IG-MYC–positive, DH
Non–BL-MYC	4111326	74	Female	80	nonmBL	DLBCL	Non–IG-MYC–positive, DH
Non–BL-MYC	4166151	74	Female	90	mBL	DLBCL	IG-MYC–positive, DH
Non–BL-MYC	4199996	73	Male	65	Non-mBL	FL	Non–IG-MYC–positive, DH
Non–BL-MYC	4107559	72	Male	80	m-Intermediate	DLBCL	Non–IG-MYC–positive, DH
Non–BL-MYC	4122063	47	Male	70	m-Intermediate	DLBCL	Non–IG-MYC–positive, DH

^*A sample was considered BL if one of two of the following conditions existed: (i) the sample is mBL according to gene expression profiling and BL according to diagnosis and/or (ii) FISH analysis shows that the sample is IG-MYC–positive and a single hit (SH). DH, double hit; IG-MYC, translocation of MYC with one of the three immunoglobulins; m-Intermediate, intermediate between BL and DLBCL; mBL, molecular Burkitt lymphoma; PID, patient identifier.

We identified lncRNAs that are differentially regulated on MYC activation in either of two model cell systems. This analysis revealed 960 and 143 lncRNAs significantly differentially expressed by activation of MYC in the P493-6 cells and the hT-RPE-MycER cells, respectively. Parallel analysis of 16 BL samples of the ICGC MMML-Seq Cohort compared with the normal GC B-cell samples revealed 514 and 367 lncRNAs significantly up- and down-regulated in BL, respectively. Finally, to identify lncRNAs that are both MYC-regulated and -deregulated in BL, we intersected the data coming from the two model cell lines with those obtained from the lymphomas. This approach led to the identification of 20 lncRNAs that were regulated in the same direction by MYC in the cell model systems and primary BL (Fig. 1A). A manual inspection of the identified lncRNAs revealed that 7 of these 20 could be considered false positives, corresponding to misannotated 3′UTRs of coding genes or pseudogenes. A list of the remaining 13 bona fide BL-specific and MYC-regulated lncRNAs is presented in Fig. 1C, and their sequence-based expression values in BL and normal GC B cells are shown in Fig. 1B.

Fig. 1.

Identification of BL-specific, MYC-regulated lncRNAs. (A) Venn diagram showing the intersection of BL-specific lncRNAs (those differentially expressed in BL compared with normal GC B cells) and lncRNAs regulated by MYC in the same direction in hT-RPE-MycER cells and P493-6 cells. (B) Heat map showing an unsupervised hierarchical clustering of the expression intensities in the BL samples of 13 of 20 lncRNAs found at the intersection between three datasets compared as in A. The heat map clearly separates the control (Ctrl) and BL samples. Yellow indicates high expression, and red indicates low expression. (C) List of 13 lncRNAs identified as differentially expressed in BL samples relative to controls (normal GC B cells) and regulated by MYC in the same direction in the two MYC-inducible cell lines P493-6 and hT-RPE-MycER. For each lncRNA, the figure shows the Ensembl gene identification, the chromosomal location, the Spearman’s rank correlation with MYC expression level (MYC correlation) and the significance of this correlation (P value; n.s., not significant), the mean expression (defined as reads per kilobase of transcript per million mapped reads, RPKM) in the four Ctrl samples (mean RPKM Ctrl), the mean expression in 16 BL samples (mean RPKM BL), the presence of MYC binding regions around the lncRNA’s TSS (MYC ChIP) in P493-6 cells expressing low (L) or high (H) MYC according to ref. 16 and at least one BL cell line according to ref. 42. For the MYC-bound region in P493-6 cells, the peak intensity values are also shown. For each lncRNA the table also shows information on the chromatin status and the level of promoter accessibility in terms of DNA methylation and presence of DNase I hypersensitive sites (DHS) at the TSS according to data from ENCODE (43) and the BLUEPRINT Consortium. Analysis of TSS methylation was performed using whole-genome bisulfite sequencing data of normal GC B cells (Ctrl), BL samples (BL), and DLBCL and FL samples (non-BL). In the chromatin states analysis, the following abbreviations were used: GM, B lymphoblastoid cell line GM17828; BL, BL cell lines BL-2 and DG-75; NBL, non-BL cell lines SU-DHL-5 and KARPAS-422.

The finding that the identified lncRNAs are regulated by MYC in two different MYC-inducible model cell systems suggested that MYC could be either directly or indirectly involved in transcriptional regulation of those RNAs. To address this point, we analyzed 13 lncRNAs at the genomic and transcriptomic levels to find out if (i) the promoters of these transcripts are bound by MYC in ChIP experiments, (ii) the promoters of these transcripts are characterized by other chromatin marks associated with active transcription, and (iii) there is a positive correlation between MYC expression and the expression of 13 lncRNAs. To answer the first question, we analyzed available data from MYC ChIP-sequencing (ChIP-seq) experiments performed with P493-6 cells under MYC low or MYC high expression (16) as well as MYC ChIP-seq experiments performed in BL cell lines (42). We found that 7 of 13 lncRNAs were bound by MYC under MYC high conditions in the region around the transcriptional start site (TSS) (Fig. 1C). Furthermore, five of them were also bound by MYC in at least one of the BL cell lines analyzed in ref. 42. These data suggest that some of the identified lncRNAs may be direct MYC target transcripts. Next, we asked if the 13 lncRNAs could be identified as actively transcribed regions by looking at chromatin marks and the DNA methylation status of their promoters. We took advantage of data generated by the BLUEPRINT Project on B-cell malignancies using the cell lines BL-2, DG-75, SU-DHL-5, and KARPAS-422 as well as data on a B-lymphoblastoid cell line (GM12878) from the Encyclopedia of DNA Elements (ENCODE) Project (43) and whole-genome bisulfite sequencing data from the IGCG MMML-Seq Project. The data presented in Fig. 1C show that at least 10 of 13 lncRNAs were defined as either transcribed regions or active promoters based on the analysis of several chromatin marks. Furthermore, most of them were characterized by hypomethylation of the region around the TSS and the presence of a DNase I hypersensitive site, suggesting that they can be considered active transcriptional units, at least in the B-cell lineage. Finally, as additional evidence of a direct regulation by MYC, we compared the expression level of the identified lncRNAs with MYC expression in 16 BL samples that are part of the ICGC MMML-Seq Cohort. In Fig. 1C, we ranked these lncRNAs according to their Spearman’s rank correlation with MYC expression level and the significance of this correlation (P value). The correlation analysis shows that, although 6 of 10 lncRNAs identified had a positive correlation with MYC expression, only for 1 of them (ENSG00000253716) did the correlation have a significant P value (Fig. 1C).

Characterization of the lncRNA MINCR.

Among 13 lncRNAs shown in Fig. 1C, ENSG00000253716 seemed to be the best candidate as the MYC-regulated transcript that could play a role in the development of MYC-positive B-cell lymphomas. ENSG00000253716 had the highest and most significant positive correlation with MYC expression levels and showed the presence of MYC binding regions around its TSS in BL cells and P493-6 cells, with increasing peak intensity in MYC high conditions in P493-6 cells. Finally, ENSG00000253716 had all of the marks associated with open chromatin and active transcription in normal and malignant B cells, and its MYC inducibility could be validated by quantitative RT-PCR (qRT-PCR) in both hT-RPE-MycER and P493-6 cell lines (Fig. 2B). We also found that the expression of ENSG00000253716 was reduced after MYC knockdown in BL cell lines (Fig. S1A). Based on these observations, we decided to concentrate additional experiments on this transcript, which we named MINCR.

Fig. 2.

Characterization of MINCR, a MYC-induced, BL-specific lncRNA. (A) Schematic representation of MINCR genomic locus on chromosome 8 (Chr8) with the mean expression profiles in BL (gray; n = 16) and the control samples (normal GC B cells; black; n = 4). The predicted isoforms based on the ENCODE annotations and the mean methylation of BL and control (Ctrl) are also shown as well as the intersection of chromatin states and MYC-bound regions (MYC BS) between the following cell lines: GM12878, K562, HeLa-S3, H1-hESC, HepG2, HUVEC and K562, HeLa-S3, HepG2, MCF-7 and NB4. (B) qRT-PCR analysis of MINCR expression in hT-RPE-MycER and P493-6 cells after MYC activation for 24 and 48 h. The expression of MINCR was calculated using GAPDH and RPL31 to normalize for cDNA content. T = 0 refers to cells treated with vehicle for hT-RPE-MycER cells and cells grown for 72 h in medium containing tetracyclin for P493-6 cells. (C) qRT-PCR analysis of MINCR isoforms expression in RNA derived from a BL sample of the ICGC MMML-Seq Cohort and the BL cell line BL-2.

Fig. S1.

(A) MYC dependency of MINCR expression in BL cell lines. qRT-PCR analysis of MINCR and MYC expression in BL-2 and Raji cells after MYC knockdown using two different shRNA constructs (shMYC#1 and shMYC#2). Gene expression was calculated using RPL31 to normalize for cDNA content and the sample treated with a nontargeting shRNA (shCtrl) as the reference (P values < 0.01). (B) Both MINCR isoforms are induced after MYC activation in hT-RPE-MycER cells. qRT-PCR analysis of MINCR_S and MINCR_L expression in hT-RPE-MycER cells treated for 48 h with 4-OHT (T = 48) or vehicle (T = 0) to induce MYC activation. Gene expression was calculated using RPL31 to normalize for cDNA content. Expression is relative to T = 0.

The MINCR gene is located on chromosome 8q24.3 and intergenic to the two coding genes GLI4 and ZNF696, with distances of 3 and 9.5 kb, respectively. Using BLAST, we found MINCR to be conserved throughout primates, with the exception of Galago (bushbaby), but not in other vertebrates (alignment shown in Fig. S2). MINCR also has a RefSeq entry defined as uncharacterized LOC100507316. ENCODE annotates at least six different isoforms transcribed from the MINCR gene locus, with a long isoform (MINCR_L) composed of three exons and all others containing two exons. A schematic depiction of MINCR intron–exon structure is shown in Fig. 2A. Analysis of sequence reads mapping of RNA-seq data from both MYC-inducible cell lines and lymphoma samples fits well with the junctions of the annotated transcripts and additionally, suggests a certain degree of intron retention (Fig. 2A). Our analysis also shows that the number of reads spanning exons 1 and 2 is much higher than those spanning exons 2 and 3, arguing that the expression of the short isoform (MINCR_S) may be two times higher than the expression of MINCR_L. A qRT-PCR analysis performed using isoform-specific sets of primers confirmed that, indeed, the most abundant isoforms are MINCR_S and MINCR_L, with MINCR_S being more expressed than MINCR_L (Fig. 2C). Finally, both MINCR_S and MINCR_L are induced after MYC activation in hT-RPE-MycER cells (Fig. S1B).

Fig. S2.

ClustalW multiple sequence alignment showing the conservation of MINCR between primates.

MINCR_L and MINCR_S are 696 and 379 nt long, respectively. To define MINCR coding potential, we analyzed the coding potential calculator (CPC) score for both isoforms. Fig. S3A shows that the CPC scores for MINCR_L and MINCR_S are −1.146 and −1.148, respectively, and they are in a range similar to that observed for the experimentally validated lncRNA XIST (−0.95). On the contrary, the CPC scores for the coding genes MYC and GAPDH are 6.77 and 12.4, respectively. MINCR can be, therefore, considered a lncRNA. Because lncRNAs were shown to be preferentially enriched in nuclear fractions (27) and because their subcellular localization can be indicative of their mechanism of action, we sought to define the intracellular localization of the MINCR_L and MINCR_S transcripts by using RNA fractionation followed by quantitative PCR. The analysis was performed in the BL cell line BL-2. As shown in Fig. S3B, similar to the nuclear lncRNA MALAT1, both MINCR transcripts L and S were preferentially enriched in the nuclear RNA fraction. We conclude that transcription of the MINCR gene produces two main transcripts, MINCR_L and MINCR_S, with very poor coding potential and preferential nuclear localization.

Fig. S3.

(A) Graph representing the CPC scores of the two MINCR isoforms MINCR_L and MINCR_S, the X-inactive specific transcript lncRNA XIST, and the two coding genes MYC and GAPDH. The CPC scores were calculated using the web-based tool available at www.cpc.cbi.pku.edu.cn/. (B) Intracellular enrichment of the two MINCR isoforms MINCR_L and MINCR_S relative to the cytosolic RNA HPRT and the nuclear RNA MALAT1. The graph shows the ratio between the relative nuclear and cytosolic expressions of the RNAs tested. RNAs with ratio values above one are supposed to be relatively enriched in the nucleus. RNAs with ratio values below one are supposed to be relatively enriched in the cytosol. (C) MINCR expression levels across 22 human tissues and cell lines. Fragments per kilobase of transcript per million mapped reads (FPKM) values relative to MINCR (XLOC_007249) were extracted from the Human lincRNA Catalog (www.broadinstitute.org/genome_bio/human_lincrnas/) as published in ref. 44. (D) Boxplot showing the expression of MYC in all samples of the ICGC MMML-Seq Cohort divided as control (Ctrl; normal GC B cells; n = 4), BL (n = 16), non-BL (DLBCLs and FLs MYC break-negative; n = 65), and non–BL-MYC (DLBCLs and FLs MYC break-positive; n = 10). Expression is defined in reads per kilobase of transcript per million mapped reads (RPKM). The significance of MYC differential expression is characterized by the following P values: Ctrl vs. BL, P value = 0.0004128; Ctrl vs. non-BL, P value = 0.0142; Ctrl vs. non–BL-MYC, P value = 0.001998; non-BL vs. non–BL-MYC, P value = 4.011e-05.

Although MINCR has been identified based on its differential expression in BL relative to normal GC B cells and found to be induced by MYC in an immortalized B cell line, it is clearly expressed and also, MYC-regulated in the retinal pigment epithelial cell line hT-RPE-MycER, suggesting that MINCR does not have a B cell-restricted expression pattern. Accordingly, analysis of MINCR expression in the Human Body Map lincRNAs Database shows that MINCR is ubiquitously expressed, with the highest expression values in brain, prostate, and testis (Fig. S3C) (44). A similar conclusion can be drawn by analysis of the genome segmentations based on the ENCODE data. This analysis is presented in Fig. 2C and shows that the intersection of the chromatin states around the TSS region of MINCR is typical of an active promoter region in the cell lines GM12878, K562, HeLa-S3, H1-hESC, HepG2, and HUVEC. Interestingly, in most of these cell lines (K562, HeLa-S3, HepG2, MCF-7, and NB4), MYC ChIP-seq data showed the presence of a uniform peak in the TSS region of MINCR, suggesting that MYC may regulate MINCR expression not only in B cells.

Analysis of MINCR Expression in Cancer.

Translocations involving the MYC oncogene are a typical feature of BL, but they are also known to occur in subsets of DLBCL and FL (45). A total of seven DLBCL and three FL samples in the ICGC MMML-Seq Cohort contained a MYC break as detected by FISH (Table S1). For subsequent analyses, we, therefore, divided the non-BL lymphomas (DLBCL and FL) in MYC break-negative non-BLs (non-BLs) and MYC break-positive non-BLs (non–BL-MYCs). As shown in Fig. S3D, the non–BL-MYC lymphomas show a clearly higher MYC expression level relative to the non-BL samples negative for the MYC break (P value = 4.011e-05). We, therefore, analyzed the expression of MINCR in the entire ICGC MMML-Seq Cohort divided as described above. As shown in Fig. 3A, MINCR expression was found to be higher in the non–BL-MYC lymphomas relative to the non-BL without the MYC break. We, therefore, extended the correlation analysis between MYC and MINCR expression levels to all of the MYC break-positive lymphomas of the cohort and found that there is a highly significant correlation (P value < 10⁻³) in the comparison of the expression values of MINCR and MYC in all MYC-positive lymphomas (16 BL cases + 10 non–BL-MYC cases) (Fig. 3B). The finding that MINCR expression correlates with MYC expression in all lymphomas examined (not only in BL) led us speculate that, because of the ubiquitous expression of MINCR, this correlation may also exist in other cancer types. To test this hypothesis, we searched in the total ICGC data for datasets where MINCR expression was detected. We found 356 datasets representing four groups of neoplasms (endocrine pancreatic neoplasm, pancreatic ductal adenocarcinoma, chronic lymphocytic leukemia, and ovarian cancer) where both MYC and MINCR were expressed. Interestingly, when we compared the MYC and MINCR expression intensities, we found a significant correlation in the pancreatic ductal adenocarcinomas group (Fig. 3C). These data suggest that the expression of the lncRNA MINCR strongly correlates with MYC expression in not only BL but more in general, GC-derived MYC-positive lymphomas and other cancer types, like pancreatic ductal adenocarcinomas, that are known to rely on MYC expression (46, 47).

Fig. 3.

Correlation between MINCR and MYC expression in cancer. (A) Boxplot showing the expression of MINCR in all samples of the ICGC MMML-Seq Cohort divided as control (Ctrl), normal GC B cells (n = 4), BL (n = 16), non-BL (DLBCLs and FLs negative for a MYC break; n = 66), and non–BL-MYC (DLBCLs and FLs positive for a MYC break; n = 10). Expression is defined in reads per kilobase of transcript per million mapped reads (RPKM). The significance of MINCR differential expression is characterized by the following P values: Ctrl vs. BL, P value = 1.618e-06; Ctrl vs. non-BL, P value = 0.0001043; Ctrl vs. non–BL-MYC, P value = 0.001919. (B) Correlation between the RPKM expression levels of MYC and MINCR in all lymphomas samples of the ICGC MMML-Seq Cohort carrying a MYC break; BL samples (n = 16) are indicated with red circles, and non–BL-MYC samples (n = 10) are indicated with blue triangles. The statistical dependence between the expressions of MYC and MINCR is characterized by a Spearman’s correlation coefficient of 0.625 and a P value = 0.00083. (C) Correlation between the RPKM expression levels of MYC and MINCR in expression data of different cancer types from the ICGC data repository. CLL, chronic lymphocytic leukemia (correlation = −0.1247623; P value = 0.3221); OV, ovarian cancer (correlation = −0.0048745; P value = 0.9595); PACA, pancreatic ductal adenocarcinoma (correlation = 0.52926; P value = 4.88e-11); PAEN, pancreatic cancer endocrine neoplasm (correlation = −0.25568; P value = 0.1506).

MINCR Regulates Cell Cycle Progression by Controlling the Expression of Cell Cycle Genes.

Changes in gene expression driven by oncogenes could be either essential for the development of the cancer phenotype and/or continuously required for cancer maintenance. In the second case, cancer cells become addicted to the expression of the gene of interest, and they will eventually cease growing or even die after the expression of the gene is reduced to basal level. We, therefore, decided to use RNAi to knockdown the expression of MINCR and measure a possible change in cell proliferation/viability. MINCR knockdown was performed in the MYC-inducible cell line hT-RPE-MycER using with two independent siRNAs (knockdown efficiency of the siRNAs used is shown in Fig. S4A). We also took advantage of the easy inducibility of the MYC protein in this cell line to analyze the effect of MINCR knockdown in cells with or without MYC activation. As shown in Fig. 4A, transfection of siRNAs directed against MINCR resulted in a time-dependent decrease in cellular proliferation independent of the MYC activation status. A more modest but reproducible effect of MINCR knockdown was also observed in the P493-6 cell line and the BL cell line BL-2 using a cell viability assay (Fig. 4 B and C). The weaker effect of the knockdown on cell viability in BL-2 cells is consistent with the lower MINCR silencing achieved in these cells relative to hT-RPE-MycER cells (Fig. S4B). To understand if the reduction in cell proliferation observed was because of a defect in cell cycle progression caused by MINCR knockdown, we analyzed the DNA content of hT-RPE-MycER cells with or without MYC activation 48 h after the transfection of control or MINCR-targeting siRNAs. As shown in Fig. 4D, MINCR knockdown resulted in a significant reduction in the percentage of cells having a DNA content corresponding to cells in S phase and G2/M phase. Concomitantly, an increase in the percentage of cells in G0/G1 phase was also observed. Finally, to investigate if apoptosis could also contribute to the reduction of cell growth observed in Fig. 4A, we measured the intracellular levels of caspase 3/7 activity after MYC activation in knocked down cells. Interestingly, although very few dead cells could be detected by visual inspection, we found that, in cells transfected with either of the siRNAs directed against MINCR, there was a significant increase in caspase 3/7 activity after MYC activation (Fig. S4C), suggesting that, at later time points, apoptosis could also be affecting cellular viability on MINCR knockdown, specifically when MYC is switched on.

Fig. S4.

(A and B) MINCR knockdown efficiency using two independent siRNAs (siMINCR_1 and siMINCR_2). The figure shows the analysis by qRT-PCR of MINCR expression level in cells transfected with siMINCR_1 and siMINCR_2 compared with cells transfected with a control siRNA (siCtrl) in (A) hT-RPE-MycER cells and (B) BL-2 and P493-6 cells. In hT-RPE-MycER cells, the analysis was performed in cells with MycER active (MYC on) or inactive (MYC off). The significance of MINCR knockdown is characterized by the following P values: hT-RPE-MycER cells, P value < 5e-05; BL-2 cells, P value < 0.002; P493-6 cells, P value < 0.01. (C) Reduction of MINCR expression in hT-RPE-MycER cells using two independent siRNAs (siMINCR_1 and siMINCR_2) results in increased intracellular caspase 3/7 activity specifically when MYC is switched on by the addition of 4-OHT. Caspase activity was measured using the Apo-ONE Assay (Promega).

Fig. 4.

MINCR expression is required for normal cellular viability. (A) Reduction of MINCR expression in hT-RPE-MycER cells using two independent siRNAs (siMINCR1 and siMINCR2) results in a significant decrease in cellular proliferation when MYC is either switched on by the addition of 4-hydroxytamoxifen (P value at 72 h < 0.01) or kept in an off status (P value at 72 h < 0.02). (B and C) Reduction of MINCR expression using two independent siRNAs (siMINCR1 and siMINCR2) results in decreased cellular viability of both (B) P493-6 cells grown without tetracycline (MYC high; P value < 0.02) and (C) the BL cell line BL-2 (P value < 0.01 with siMINCR2). Viability was measured using the Cell Titer Blue reagent (Promega). (D) Reduction of MINCR expression in hT-RPE-MycER cells using two independent siRNAs (siMINCR1 and siMINCR2) results in an increase of the percentage of cells in the G1 phase of the cell cycle and a decrease of the percentage of cells in the S and G2/M phases of the cell cycle (P values < 0.02). siCtrl, control siRNA.

Given that MINCR knockdown had a reproducible effect on cellular viability and cell cycle progression, we decided to investigate gene expression changes induced by MINCR knockdown in hT-RPE-MycER cells, the cell line where the effect was most pronounced. We performed RNA-seq of hT-RPE-MycER cells 48 h after transfection with the two siRNAs directed against MINCR. The analysis was performed using two biological replicas. Furthermore, to exclude possible off-target effects on gene expression associated with single siRNAs, we focused only on genes regulated in the same direction with both siRNAs in the two replicates. As shown in Fig. 5A, MINCR knockdown in hT-RPE-MycER cells resulted in the up-regulation of 568 genes and the down-regulation of 784 genes.

Fig. 5.

Cells with reduced MINCR expression show a significant reduction in the expression of cell cycle genes. (A) Venn diagram showing the intersection between genes differentially regulated in hT-RPE-MycER cells after transfection of the two independent siRNAs against MINCR (siMINCR1 and siMINCR2). The diagram also shows the number of genes differentially up-regulated and down-regulated common to the two siRNA transfections and the number of cell cycle genes enriched among the down-regulated genes. (B) The cell cycle-related biological processes are significantly enriched among the genes down-regulated after MINCR RNAi. The graph shows the P values of the three most highly significant biological processes. (C) Heat map showing hierarchical clustering of the expression intensities of 125 cell cycle genes identified in hT-RPE-MycER cells transfected with two independent siRNAs against MINCR (siMINCR1 and siMINCR2) in two biological replicas (1 and 2). Yellow indicates high expression, and red indicates low expression. (D) qRT-PCR analysis of the expression of a selection of 125 identified cell cycle genes in hT-RPE-MycER cells after MINCR RNAi. The graph shows the mean of the change in expression relative to a control siRNA (siCtrl) observed after transfection of the two independent siRNAs against MINCR.

lncRNAs have been described to be able to regulate gene expression of neighboring genes (48, 49). We, therefore, asked if the expression of the most proximal genes upstream and downstream from MINCR were changed after MINCR knockdown. Given that no significant change was observed, we conclude that MINCR is not a cis-regulator of neighboring genes. Next, we looked at the groups of genes differentially up-regulated or down-regulated after MINCR knockdown to see if there was any enrichment in particular biological processes. Interestingly, we found a strongly significant enrichment of the biological process category “cell cycle” in the analysis of the genes down-regulated after MINCR knockdown (Fig. 5B). A heat map with the expression values of the down-regulated cell cycle genes in all conditions examined is shown in Fig. 5C, and a list of the same genes is shown in Table S2. To this list, we also added a few genes belonging to the same functional category that failed to pass the stringent significance criteria applied to the analysis but were still clearly down-regulated on MINCR knockdown. A selection of these genes chosen for their degree of down-regulation in MINCR knockdown cells and their relevance within lymphoma cell physiology was successfully validated by qRT-PCR (Fig. 5D).

Table S2.

List of the genes belonging to the cell cycle biological process down-regulated in hT-RPE-MycER cells after MINCR knockdown

Gene name	siMINCR vs. siRNA control MYC off	siMINCR vs. siRNA control MYC on	MYC on vs. MYC off siRNA control
PHGDH	−1.69*	−0.71	−0.51*
CDC25B	−1.58*	−0.88*	−0.59*
CDC20	−1.58*	−0.84	−0.4
E2F2	−1.54*	−0.68	−0.38
CDT1	−1.50*	−0.71	−0.3
UBE2C	−1.46*	−0.66	−0.78*
CCNF	−1.45*	−0.72	−0.70*
PLK1	−1.38*	−0.7	−0.23
SEPT14	−1.37*	−0.39	−0.85*
GSG2	−1.37*	−0.29	−0.70*
AURKB	−1.34	−0.75	−0.53*
TUBB2B	−1.32*	−0.78	−0.68*
INCENP	−1.31*	−0.42	−0.80*
AURKA	−1.29*	−0.62	−0.4
PSRC1	−1.29*	−0.80*	−0.74*
FOXM1	−1.25*	−0.45	−0.62*
CDCA8	−1.25*	−0.62	−0.85*
CENPF	−1.24*	−0.46	−1.03*
MCM2	−1.24*	−0.49	−0.21
GTSE1	−1.23*	−0.55	−0.96*
MKI67	−1.20*	−0.17	−0.83
CIT	−1.20*	−0.07	−0.95*
DBF4B	−1.20*	−0.53	−0.53
CLSPN	−1.18*	−0.3	−1.17*
E2F1	−1.18*	−0.36	−0.91*
TACC3	−1.17*	−0.61	−0.70*
PTTG1	−1.16*	−0.57	−0.51
KIF2C	−1.16*	−0.38	−0.87*
ESPL1	−1.15*	−0.21	−0.62*
HJURP	−1.12*	−0.41	−1.38*
TPX2	−1.12*	−0.53	−0.68*
CDCA5	−1.12*	−0.31	−0.60*
CHAF1B	−1.12*	−0.45	−0.41
H2AFX	−1.11*	−0.3	−0.60*
PKMYT1	−1.10*	−0.27	−0.37
RACGAP1	−1.10*	−0.46	−0.94*
CCNB2	−1.10*	−0.6	−0.38
KIF22	−1.10*	−0.49	−0.44
PARD6A	−1.08*	−0.26	−0.95*
CGREF1	−1.08*	−0.26	0.28
CHAF1A	−1.08*	−0.27	−0.57*
LIG1	−1.08*	−0.42	−0.44*
NCAPD2	−1.07*	−0.56	−0.36
TUBB2A	−1.06*	−0.75	−0.57*
KPNA2	−1.06*	−0.51	−0.50*
CHTF18	−1.06*	−0.41	−0.52
BIRC5	−1.05*	−0.42	−0.15
CCNE1	−1.05*	−0.05	0.31
CABLES2	−1.04*	−0.44	−1.47*
CDCA3	−1.03*	−0.55	−1.05*
RAD51	−1.02*	−0.3	−0.74*
NUDC	−1.02*	−0.5	0.18
CENPJ	−1.01*	−0.17	−1.05*
TUBB	−1.01*	−0.57	−0.74*
BUB1B	−1.00*	−0.16	−1.03*
FAM83D	−0.99*	−0.07	−0.59*
SUV39H1	−0.99*	−0.22	−0.42
SKA1	−0.98*	−0.25	−1.00*
TUBB3	−0.98*	−0.52	−1.26*
TUBG1	−0.98*	−0.34	−0.35
BRCA1	−0.98*	−0.11	−0.63
NCAPG	−0.96*	−0.44	−0.72*
BLM	−0.96*	−0.05	−0.09
POLD1	−0.96*	−0.37	0.09
STMN1	−0.96*	−0.63	−1.25*
MCM7	−0.95*	−0.37	−0.44*
MCM6	−0.93	−0.3	0.3
HAUS8	−0.93*	−0.2	−0.5
MCM10	−0.92	−0.2	−0.43
CDCA2	−0.91*	−0.13	−0.70*
CNTRL	−0.91*	−0.21	−0.95*
TIMELESS	−0.90*	−0.24	−0.66*
ERCC6L	−0.88*	−0.22	−0.62
CDC45	−0.88*	−0.13	−0.44
ADCY3	−0.85*	−0.1	0.12
XRCC2	−0.85*	−0.08	−0.36
NASP	−0.85*	−0.41	−0.35
CDKN2D	−0.84*	−0.2	−1.46*
NEK2	−0.84*	−0.35	−0.73*
MCM8	−0.83*	0.03	−0.56
E2F4	−0.83*	−0.53	0.44*
SAC3D1	−0.82*	−0.27	0.12
MCM4	−0.82	−0.23	−0.24
PSMC3	−0.81*	−0.41	−0.21
PPM1G	−0.81*	−0.29	0.17
CALM3	−0.81*	−0.45	−0.59
CEP55	−0.80*	−0.21	−1.21*
NDE1	−0.78*	−0.22	−0.59*
BRCA2	−0.77*	0.03	−0.55
RANBP1	−0.76*	−0.24	0.19
SUN2	−0.75*	−0.5	−0.68*
MAD1L1	−0.75*	−0.2	−0.35
DTYMK	−0.74*	−0.26	−0.03
MYH10	−0.74*	−0.09	−0.24
FZR1	−0.72*	−0.45	−0.62*
DSN1	−0.71*	−0.09	−0.98*
MDC1	−0.70*	−0.35	−0.2
SPC24	−0.70*	−0.59	−0.49*
CKS2	−0.70*	−0.21	0.32
KIF23	−0.69	−0.15	−0.92*
SPECC1L	−0.68*	−0.15	−0.56
SGOL2	−0.67*	−0.17	−0.63
SETD8	−0.66*	−0.09	−0.26
HAUS4	−0.63*	−0.42	−0.57*
NOLC1	−0.62*	0	1.21*
HCFC1	−0.60*	−0.1	−0.38
CENPA	−0.58*	−0.38	0.07
GPS2	−0.57*	−0.32	0.5
PAPD7	−0.54*	−0.11	−0.18
TFDP1	−0.53*	−0.18	0.19
TCF3	−0.52*	−0.11	−0.01
PSMB2	−0.51*	−0.51	0.2
TP53	−0.50*	−0.43	−0.23
MSH2	−0.5	−0.01	−0.25
KATNB1	−0.48*	−0.04	0.33
RCC2	−0.47*	−0.49	0.59*
PSMF1	−0.47*	−0.3	0.13
MSH6	−0.47*	−0.11	−0.28
PML	−0.46*	−0.6	−0.08
RAD21	−0.45*	−0.04	−0.34
ARHGEF2	−0.45*	0.04	−0.31
PES1	−0.43*	−0.23	0.96*
MRE11A	−0.43*	0.14	−0.45
MFN2	−0.41*	−0.1	0.23
AKT1	−0.41*	−0.03	−0.23
PCNT	−0.40*	0.08	−0.44
KHDRBS1	−0.36*	−0.29	−0.03
CCNK	−0.35*	−0.1	0.71*
KPNA6	−0.35	−0.33	0.05
TARDBP	−0.35*	−0.05	−0.06
GSPT1	−0.34*	−0.12	0.31
PSMC4	−0.30*	−0.18	0.26
AKT2	−0.24	0	−0.01
PLK2	0.29	−0.37	0.32

^*Significant.

We also analyzed the changes in gene expression observed in hT-RPE-MycER cells on knockdown of MINCR expression and treatment with 4-hydroxytamoxifen (4-OHT) to activate MYC. On MYC activation, the number of genes significantly differentially expressed was much lower than in the absence of MYC activation. This finding could be due to the fact that many of the cell cycle genes identified are, indeed, weakly down-regulated on MYC activation in this cell system, a behavior not shared with all cell cycle genes (Fig. S5 A and B and Table S2). Nevertheless, our data clearly show that most of the cell cycle genes identified as down-regulated on MINCR knockdown in cells with MYC off are still down-regulated when MYC is activated but that they fail to pass the stringent significance criteria that we used for the definition of differentially expressed genes.

Fig. S5.

Heat map showing hierarchical clustering of the expression intensities of (A) the identified 125 cell cycle genes or (B) 125 randomly selected cell cycle genes in hT-RPE-MycER cells transfected with two independent siRNAs against MINCR (siMINCR_1 and siMINCR_2) or a control siRNA (Ctrl) and treated with 4-OHT to activate MYC (M) or vehicle (C). Yellow indicates high expression, and red indicates low expression. (C) DNA motifs enriched in the promoters of 125 cell cycle genes down-regulated in hT-RPE-MycER cells after MINCR RNAi.

Overall, the data presented in Figs. 4 and 5 suggest that the reduction of proliferation observed in hT-RPE-MycER cells after MINCR knockdown depends on the down-regulation of a set of genes important for cell cycle progression. Next, we asked if these findings could be also extended to BL cell line BL-2. Initially, we found that only a few of the identified genes were also down-regulated in BL-2 cells (Fig. 6A). This observation might be explained by the lower degree of MINCR knockdown achieved in BL-2 cells relative to hT-RPE-MycER cells (Fig. S4B). Nevertheless, we found that the genes coding for Aurora kinase A (AURKA), AURKB, and chromatin licensing and DNA replication factor 1 (CDT1) were also reproducibly down-regulated in BL-2 cells after MINCR knockdown (Fig. 6A). Furthermore, these genes are among the top down-regulated on MINCR knockdown when MYC is activated (Table S2).

Fig. 6.

Cell cycle genes down-regulated in BL-2 cells after MINCR knockdown. (A) qRT-PCR analysis of the expression of a selection of 125 identified cell cycle genes in BL-2 cells after MINCR RNAi. The graph shows the mean of the change in expression relative to a control siRNA (siCtrl) observed after transfection of two independent siRNAs against MINCR (P values > 0.05 only for BIRC5 and PLK1). (B) Expression of some of the identified cell cycle genes in B-cell lymphomas. Boxplots showing the expressions of eight cell cycle genes in all samples of the ICGC MMML-Seq Cohort divided as BL (n = 16), non-BL (DLBCLs and FLs MYC break-negative; n = 65), and non–BL-MYC (DLBCLs and FLs MYC break-positive; n = 10). Expression is defined in reads per lilobase of transcript per million mapped reads (RPKM). (C) ChIP–quantitative PCR analysis of hT-RPE-MycER cells after transfection of two independent siRNAs against MINCR (siMINCR1 and siMINCR2) using either control IgG or MYC-specific antibodies.

MINCR-Regulated Genes Are also MYC Targets.

By closer inspection of the list of genes down-regulated on MINCR knockdown, we also noticed the presence of many known MYC target genes. To establish if there was a significant enrichment in MYC binding sites in the promoters of 125 cell cycle genes down-regulated on MINCR knockdown, we identified 242 promoter regions overlapping with our set of genes (each gene extended by 1,500 nt upstream). The promoter regions were defined based on the chromatin states of the GM12878 cell line (ENCODE 2012). Using this approach, we found 63 MYC binding sites (defined as in ref. 42) in these promoter regions. To check for significance, we analyzed the number of MYC-bound regions in a background set of cell cycle genes created by randomly selecting 125 of 1,324 genes belonging to the cell cycle functional category (Gene Ontology ID GO:0007049). Given that we found only 40 MYC binding sites in 259 promoter regions, we conclude that the 125 genes down-regulated on MINCR knockdown are significantly more enriched in MYC binding sites (P value = 0.004).

We, therefore, asked if the identified genes could differentiate between the lymphoma subtypes carrying and not carrying the MYC translocation. The analysis was done on eight genes found to be down-regulated on MINCR knockdown in BL-2 cells. Interestingly, the expression pattern of all of them clearly mirrored the MYC status of the lymphomas, with higher expression in BLs and non–BL-MYC lymphomas relative to the non-BLs without the break (Fig. 6B).

Given the observed enrichment of MYC binding sites in the promoters of the identified cell cycle genes, we asked if their reduced expression could be because of a reduced binding of MYC to their promoters. To answer this question, we performed ChIP using MYC and control IgG antibodies on hT-RPE-MycER cells transfected with two siRNAs directed against MINCR 24 h after MYC activation. Using quantitative PCR, we analyzed regions of the promoters of four cell cycle genes (AURKA, AURKB, CDK2, and NCAPD2) found to be bound by MYC in at least one of the ENCODE cell lines. The data presented in Fig. 6C show that at least three of four genes (AURKA, AURKB, and NCAPD2) are also significantly bound by MYC in hT-RPE-MycER cells. The result also shows that MYC binding is reduced in all promoter regions after knocking down of MINCR using siMINCR2, the siRNA that has the strongest effect on cellular proliferation (Fig. 4A), suggesting that MINCR silencing might reduce the ability of MYC to engage a transcriptionally active complex, at least for the genes tested.

The data presented suggest that MINCR is able to influence the transcription of a subset of MYC target genes at least in part by affecting the ability of MYC to bind their promoters. Given that not all MYC-regulated genes are affected by MINCR expression, we speculated that an additional specificity determinant might be required for transcriptional regulation by MINCR. We, therefore, analyzed the promoter regions of 125 cell cycle genes to search for motifs discriminating these genes from background cell cycle genes. This analysis showed a statistically significant enrichment of the two motifs shown in Fig. S5C. Interestingly, one of these motifs (motif 1) is almost identical to the DNA binding site of the homeobox protein family Sine Oculis Homeobox Homolog 1–6 (Six1–6).

AURKA, AURKB, and CDT1 Down-Regulation Recapitulates the Effect of MINCR Knockdown.

Finally, we asked if at least some of the cell cycle genes found down-regulated on MINCR knockdown could be responsible for the phenotypic changes observed after MINCR silencing. We selected a set of cell cycle genes based on their degree of down-regulation in hT-RPE-MycER and BL-2 cells and their functional relevance. As shown in Fig. 7, among the chosen genes, knockdown of AURKA, AURKB, and CDT1 was responsible for a reproducible reduction in cellular viability in hT-RPE-MycER cells. Furthermore, the effect was even more pronounced after MYC activation. These data, therefore, suggest that down-regulation of cell cycle genes, like AURKA, AURKB, and CDT1, may be, at least in part, the cause of the reduced basal proliferation observed after MINCR knockdown.

Fig. 7.

Knockdown of AURKA, AURKB, and CDT1 recapitulates the effect of MINCR RNAi on hT-RPE-MycER cellular viability. hT-RPE-MycER cells were transfected with the indicated siRNAs directed against a set of cell cycle genes and then grown in the presence of 150 nM 4-hydroxytamoxifen (MYC on) or vehicle (MYC off). Viability was measured 48 h later using the Cell Titer Blue reagent (Promega). Cells transfected with siRNAs against AURKA, AURKB, and CDT1 had a significant reduction in cellular viability (P values with MYC on < 0.005). siCtrl, control siRNA.

Discussion

Since the discovery that eukaryotic cells express thousands of long RNAs with no clear coding potential, one of the main challenges is to understand the biological functions associated to these novel transcripts. Several lncRNA have already been shown to be important players in the execution of physiological cellular and organismal homeostasis (24). Accordingly, altered expression of lncRNAs has been linked to several disease conditions, particularly cancer (22, 23). There are now several examples of lncRNAs that are able to influence the cellular transcriptional program, acting either at the transcriptional level by influencing the establishment of chromatin marks (34, 50) or the posttranscriptional level by controlling mRNA stability (51), compartimentalization (31), or translation (52). Given that the oncogenic transcription factor MYC is known to play a crucial role in the development of many neoplasms by controlling the expression of thousands of genes, we asked in this study if lncRNAs could also play a role in MYC-induced oncogenic transformation. We approached this aim by investigating the extent of changes that MYC is able to induce in lncRNA expression. We used two different cell line model systems carrying a MYC-inducible construct and through the analysis of BL, a cancer in which MYC is known to play a fundamental role. BL can be considered the best cancer model to study the role of MYC in oncogenic transformation for at least three reasons: it is invariably characterized by an IG-MYC translocation (resulting in the deregulated overexpression of MYC), MYC translocation is considered to be the primary oncogenic event in BL development, and it is usually characterized by a relatively simple karyotype, with secondary aberrations mostly representing singular events (53, 54). We, therefore, hypothesized that deregulated lncRNAs in BL are the result of MYC overexpression and could, therefore, play a major role in the development of the lymphoma. Using this approach, we identified more than 50 lncRNAs regulated by MYC in two different cell lines carrying two different MYC-inducible constructs and almost 1,000 lncRNAs differentially regulated in BL patient samples relative to normal GC B cells. Although several reports investigating the effect of MYC activation on lncRNA transcription were recently published (55–58), our study is the first, to our knowledge, to report an RNA-seq–based systematic analysis of MYC-regulated lncRNAs and at the same time, lncRNAs deregulated in GC-derived B-cell lymphomas.

The intersection of the cell line data and the data obtained from the sequencing of BL patient samples led us to the identification of the MINCR lncRNA. MINCR turned out to be the best candidate as a MYC-regulated BL-specific lncRNA, but it will certainly not remain the only lncRNA playing a role in BL development. To increase the chance to identify true MYC target transcripts, we decided to filter our RNA-seq data using two cell model systems carrying inducible MYC expression. Although this choice has probably shifted the focus toward lncRNAs with broader tissue-specific expression, the findings that MINCR shows a high correlation with MYC expression levels in MYC-positive lymphoma samples and that it is bound by MYC at the TSS region suggest that it was a successful strategy to identify MYC-regulated lncRNAs. Furthermore, the observation that MINCR has broad tissue specificity (Fig. S3C) suggests that MINCR could play an important role in not only MYC-positive lymphomas but also, several other cancers that rely on MYC overexpression. Accordingly, we identified a significant correlation between MINCR and MYC expression in RNA-seq data from pancreatic ductal adenocarcinomas (Fig. 3C). Given that amplification of the genomic region containing MYC is a very frequent event in cancer (7), we can forecast a general role of MINCR in mediating the effect of MYC overexpression in cancer development.

Analysis of the transcriptional changes observed after MINCR knockdown suggested that the lncRNA MINCR is able to control the expression level of a set of genes involved in cell cycle initiation and progression. In particular, genes playing a role in initiation of genome replication and the assembly of the mitotic spindle are found strongly reduced after MINCR RNAi. This observation is in complete agreement with the data that we present in Fig. 4D, showing a strong perturbation of the cell cycle phases after MINCR knockdown. The other interesting observation was that those genes showed a significant enrichment of MYC-binding sites in their promoters. MINCR, therefore, seems to be involved in controlling a subset of MYC target genes. Interestingly, also, the MYC-regulated lncRNAs identified in colorectal cancer by Kim et al. (56) seem to share with MINCR the ability to modulate the expression of MYC target genes involved in cell cycle regulation. Although more experiments are needed to shed light on MINCR mechanism of action, we can speculate that at least two possible mechanisms may take place. One possibility is that MINCR directly influences MYC transcriptional activity. In this model, we can imagine MINCR binding either directly to MYC or to a MYC binding partner and influencing the ability of the MYC transactivation complex to bind promoters and affect transcription. The lncRNA PCGEM1, for instance, has recently been shown to bind MYC directly and enhance its ability to regulate the transcription of metabolic genes (59). In support of this model, we have observed a decrease in MYC binding to the promoters of a set of the identified cell cycle genes on MINCR knockdown (Fig. 6C). Alternatively, MYC has been shown to bind at least two of the core subunits of KMT2 methyltransferase complexes that target histone H3 lysine 4, a mark associated with open chromatin (60). Interestingly, one of these subunits, WDR5, has recently been shown to bind several lncRNAs through an RNA binding pocket essential for its function as a transcriptional activator (61).

In a similar scenario, MINCR could bind to a transcription factor that cooccupies the MYC binding regions in the promoters of the target genes, influencing in this way the ability of MYC to transactivate those genes. We have, indeed, identified two DNA motifs significantly enriched in the promoters of the cell cycle genes (Fig. S5C). One of these motifs is almost identical to the DNA binding site for the family of homeobox proteins Six1–6. Based on this finding, we can speculate that the transcriptional regulation of the identified cell cycle genes could depend on the interplay between MYC, MINCR, and one of the members of the Six family of transcription factors. Interestingly, the Six1 protein has recently been found to promote DNA replication and cell proliferation in cervical cancer by regulating the expression of multiple genes related to the initiation of DNA replication (62).

It is interesting to speculate that MINCR and lncRNAs in general, by conferring increased selectivity to a transcriptional factor like MYC that recognizes an extremely common DNA binding motif, may, in the end, constitute the missing link that could explain the choice between MYC as a general transcriptional amplifier and MYC as a selective gene regulator (16).

Whatever the mechanism involved, our data show that MINCR has a clear effect on the establishment of the expression level of the genes coding for AURKA and AURKB as well as CDT1. Additionally, we also show that down-regulation of these three genes may in the end explain the reduced proliferation observed after MINCR RNAi, because their knockdown results in reduced cellular viability (Fig. 7). Interestingly, AURKA and AURKB have already been shown to play a crucial role in MYC-induced oncogenic transformation, with AURKB showing a synthetic lethal interaction with MYC (63). Of note, the synthetic lethal effect of AURKB inhibition was identified in the same cellular system used in this study, the hT-RPE cells, albeit comparing vector-transfected cells with cells stably transfected with MYC. An oncogenic role in lymphoma development was also previously reported for CDT1 (64).

In conclusion, our findings suggest that the expression level of a group of genes involved in cell cycle initiation and progression may be regulated through the interplay between the transcription factor MYC and a newly identified lncRNA, which we named MINCR. The identification of the mechanism of MINCR action could in the future suggest novel therapeutic opportunities for the fight against not only malignant lymphoma but possibly, all cancers that rely on MYC expression.

Materials and Methods

Extended materials and methods are provided in SI Materials and Methods. The sequences of all primers and siRNAs/shRNAs used in this study are listed in Tables S3 and S4, respectively. The ICGC Malignant Lymphoma study was approved by the Institutional Review Board of the Medical Faculty of the University of Kiel (A150/10).

Table S3.

List of the qRT-PCR primers used in this study

Target gene	Forward primer (5′ to 3′)	Reverse primer (5′ to 3′)
MINCR_L	TGTGGCAAAACTTGAATGGA	GGGGGAGGACAAGAGAAAGA
MINCR_S	GGCCAGCTCACCATGAGTC	CTTTAATGGCAAACAAGGCTCT
MINCR_S1	CCCAGTCTGAACTCACCATG	TTTGCCACATGGCACAGTAT
MINCR_S2	CGGGGCCAGCCCTAGCGTAG	TTGGGCATGTACTGGAGTGA
MINCR_S3	CGGGGCCAGCTCCTTCGGTAG	TTGGGCATGTACTGGAGTGA
GAPDH	CAGCCTCAAGATCATCAGCA	TGTGGTCATGAGTCCTTCCA
G6PD	CGTCACCAAGAACATTCACG	AGGAGATGTGGTTGGACAGC
RPL31	TGCCATCAACGAAGTGGTAA	CATGGCAAATTTCCGAATCT
HPRT	TGTTGTAGGATATGCCCTTGACT	GGCTTTGTATTTTGCTTTTCCA
MALAT1	TGGGGGAGTTTCGTACTGAG	TTTCTCCAGGACTTGGCAGT
CDC20	GCTTTGCACAGAACCAGCTA	CAACTCAAAACAGCGCCATA
CDC25B	CCCTGTAGCCTGGACAAGAG	GCCGCCTTTCAGGATATACA
CDCA8	AGACAGGCAGAACCTCCTCA	GCCTGTTTGTTTCCTCCAAG
CDK2	GTGGTACCGAGCTCCTGAAA	AAAGATCCGGAAGAGCTGGT
PLK1	GCCCCTCACAGTCCTCAATA	CTGCAGCATGTCACTGAGGT
TCF3	GGGACCTCCCAGTACTACCC	CCTGCCGTAGTCCTCACCT
E2F2	ACCAGGATATCCGTGCTGTT	TCCTCTGGGCACAGGTAGAC
NCAPD2	GTTGAAAAGTGGAGGGGTGA	TCGAAAGTGATGCAAAATGC
BIRC5	GTTGCGCTTTCCTTTCTGTC	TCTCCGCAGTTTCCTCAAAT
AURKA_ChIP	TGGAGATAAAGTCCAAGGAGGT	TATCGATCGGTCCCACTCTC
AURKB_ChIP	ACCTGATCATCTGCCCACTC	GGGAGAGTAGCAGTGCCTTG
CDK2_ChIP	CAAATTGACAAGAGCGAGAGG	CGAGGAGGGGGAAACAGG
NCAPD2_ChIP	CTAATTGGTGCGTGTTGTGC	CCACATCTTACCTGGGGATCT
CKS2	QantiTect Primer Assay QT00014693 (Qiagen)
UBE2C	QantiTect Primer Assay QT01015882 (Qiagen)
NOLC1	QantiTect Primer Assay QT00092806 (Qiagen)
NPM1	QantiTect Primer Assay QT00091875 (Qiagen)
MCM2	QantiTect Primer Assay QT00070812 (Qiagen)
KPNA2	QantiTect Primer Assay QT01003527 (Qiagen)
PHGDH	QantiTect Primer Assay QT00083279 (Qiagen)
PA2G4	QantiTect Primer Assay QT00083090 (Qiagen)
RRM2	QantiTect Primer Assay QT00039480 (Qiagen)
TACC3	QantiTect Primer Assay QT00017619 (Qiagen)
AURKA	QantiTect Primer Assay QT00038213 (Qiagen)
AURKB	QantiTect Primer Assay QT00067403 (Qiagen)
MYBL2	QantiTect Primer Assay QT00030058 (Qiagen)
TPX2	QantiTect Primer Assay QT00029169 (Qiagen)
CDT1	QantiTect Primer Assay QT00020601 (Qiagen)

Table S4.

List of the siRNAs and shRNAs used in this study

Target gene	Sense RNA (5′ to 3′)	Antisense RNA (5′ to 3′)
siMINCR_1	GGGAAGAGUGCGUCUGUGAUU	UCACAGACGCACUCUUCCCUU
siMINCR_2	GAGCCUUGUUUGCCAUUAAUU	UUAAUGGCAAACAAGGCUCUU
shMYC#1	GATCCCTATGAACTTGTTTCAAATGTCAAGAGCATTTGAAACAAGTTCATAGGTTTTT
shMYC#2	GATCCGATTCCTTCTAACAGAAATGTCAAGAGCATTTCTGTTAGAAGGAATCGTTTTT
shNT	GATCGATCTCGATATCTAAGCTATTTCAAGAGAATAGCTTAGATATCGAGATCTTTTT
siCtrl	AllStars Negative Control siRNA (catalog no. 1027281; Qiagen)
siKIF22	FlexiTube siRNA (catalog no. SI03019856; Qiagen)
siMCM2	FlexiTube siRNA (catalog no. SI02653525; Qiagen)
siUBE2C	FlexiTube siRNA (catalog no. SI00302071; Qiagen)
siAURKA	FlexiTube siRNA (catalog no. SI02223305; Qiagen)
siAURKB	FlexiTube siRNA (catalog no. SI02622032; Qiagen)
siCDT1	FlexiTube siRNA (catalog no. SI00342384; Qiagen)
siPHGDH	FlexiTube siRNA (catalog no. SI00090384; Qiagen)

siCtrl, control siRNA.

Experimental procedures for whole-genome, whole-genome bisulfite, and transcriptome sequencing of the ICGC MMML-Seq have been published previously (54) and are briefly repeated in SI Materials and Methods. Samples were sequenced using HiSeq 2500 (Illumina), and mapping was performed with the segemehl suite, version 0.1.7 (65, 66).

RNAi in BL-2 and P493-6 cells was performed using an Amaxa Nucleofector using a two-step transfection protocol, and viability was scored after growing cells for 24 h in 3% (vol/vol) FBS. hT-RPE-MycER cells were knocked down with HiPerFect (Qiagen) 5 or 24 h before activation of the MycER protein. Total RNAs were extracted with the RNeasy Mini Kit (Qiagen) and reverse-transcribed using the QuantiTect RT Kit (Qiagen). qRT-PCR was performed using the Luminaris Color HiGreen qPCR Master Mix (Thermo Fisher Scientific). Cell cycle analysis was performed using propidium iodide 48 h after RNAi and 24 h after MYC activation. ChIP was performed according to protocols published by Lee et al. (67) 48 h after RNAi and 24 h after MYC activation.

SI Materials and Methods

Basic Characterization of the Lymphoma Samples in the ICGC MMML-Seq.

Experimental procedures for whole-genome, whole-genome bisulfite, and transcriptome sequencing of the ICGC MMML-Seq have been published previously (54). They are briefly repeated below.

Histopathology.

Lymphoma samples were reviewed by expert hematopathologists on individual microscopes independent of each other. The pathologists assigned each lymphoma to the categories according to the WHO 2008 classification. A consensus diagnosis was reached by a single independent microscope analysis if at least five of seven pathologists agreed by discussion. For discrepant cases, a consensus was reached after discussing the respective cases at a multiheaded microscope.

Immunohistochemistry.

The morphologic and immunophenotypic features were studied on formalin-fixed and paraffin-embedded tissue sections of the diagnostic biopsies. The immunohistochemical panel included antibodies directed against CD20, CD10, BCL2, BCL6, and MUM1/IRF4. Lymphoma samples were scored according to their percentage content of positive tumor cells from 0 to 4 (0 = 0%; 1 = 1–25%; 2 = 26–50%; 3 = 51–75%; 4 = 76–100%). Immunohistochemical staining for K_i-67 was assessed as the percentage of positive tumor cells on formalin-fixed and paraffin-embedded material. In addition, in situ hybridization staining for Epstein–Barr encoding region was done in tissue sections.

FISH.

FISH on interphase nuclei was performed on frozen tissue sections with the specific probes LSI BCL6, LSI MYC, LSI IGH/MYC, CEP8 Tricolor, LSI IGH, and LSI BCL2 probes. All of the probes were provided by Abbott Molecular Diagnostics. Hybridizations, digital image acquisition, processing, and evaluation were carried out as previously described (68). The signal distribution was evaluated by two independent observers.

Pretreatment tumor tissues were studied with informed consent of the patients and/or for minors, the caretakers. The ICGC Malignant Lymphoma Study was approved by the Institutional Review Boards of the Medical Faculty of the University of Kiel (A150/10) and the recruiting centers.

RNA Extraction.

Extraction of total RNA, including the small RNA fraction from frozen tissue and body fluids (blood and buffy coats), was performed using the Ambion mirVana Kit (Life Technologies) and the mirVana PARIS Kit (Life Technologies), respectively. Bioanalyzer (Agilent Technologies) and the RNA 6000 Nano Kit (Agilent Technologies) were used for size determination and quality control of the RNA. RNA quantification was done using the Qubit instrument at appropriate settings (5–100 ng; Quant-iT RNA HA Assay Kit).

Extraction of total RNA from cell lines (hT-RPE-MycER, P493-6, and BL-2 cells) was performed using the RNeasy Mini Kit (Qiagen). RNA quality was determined using an Experion Automated Electrophoresis System (Bio-Rad) according to the manufacturer’s instructions on RNA Analysis Kits.

RNA-Seq Library Preparation, Sequencing, and Downstream Analyses.

TruSeq RNA Sample Preparation v2 Kit (low-throughput protocol; Illumina) was used to prepare the barcoded libraries from 1 µg total RNA. Libraries were validated and quantified using DNA 1,000 and high-sensitivity chips on Bioanalyzer (Agilent); 6 pM denatured libraries were used as input into cBot followed by deep sequencing using HiSeq 2500 (Illumina) for 51 cycles, with an additional 7 cycles for index reading.

After sequencing, the RNA-seq samples were quality trimmed with FASTX-Toolkit, version 0.0.13 (hannonlab.cshl.edu/fastx_toolkit/) and adapter-clipped with Cutadapt, version 1.2.1 (69). The splice aware mapping process onto the human reference genome (hg19) was carried out with the segemehl suite, version 0.1.7 (65, 66). Based on the GENECODE annotation (version 19), individual sample supports were calculated with NGSUtils (70), and differentially expressed genes were identified using the program DESeq2, version 2.13 (71) using a significance threshold of q < 10⁻². Because of the lack of strand specificity in the sequencing protocol, we focused mainly on intergenic lncRNAs by filtering out all transcripts showing any overlap with protein coding genes. This approach was also used to reanalyze published data analyzing the transcriptional changes induced by MYC in P493-6 cells (Gene Expression Omnibus DataSet accession no. GSE58168). Specifically, we used the sample GSM1402581 for MYC low conditions (cells treated with 0.1 µg/mL doxycycline) and the sample GSM1402579 for MYC high conditions (nontreated cells). Enrichment analyses of biological processes were performed with DAVID (https://david.ncifcrf.gov/). Discriminative motif discovery was done with DREME, and motif comparison was done with TOMTOM (72).

Cell Culture.

hT-RPE-MycER cells were grown in DMEM supplemented with 10% (vol/vol) FBS. Activation of the MycER protein was performed as described previously (38) using 150 nM 4-hydroxytamoxifen (4-OHT). Control cells were treated with equal volumes of vehicle (ethanol). BL-2 and P493-6 cells were grown in RPMI supplemented with 10% (vol/vol) FBS and 1% glutamine. Transcriptional shutdown of MYC in P493-6 cells was achieved by treating the cells with 0.1 µg/mL tetracyclin for 72 h. Cells were then washed two times with growth medium to remove the tetracyclin, and the transcriptional activation of MYC was followed by qRT-PCR at different time points.

RNAi and Viability Assays.

siRNAs against MINCR were custom designed based on the MINCR_S sequence using the online siDESIGN Center on the Dharmacon web page and purchased from Dharmacon. Sequences are provided in Table S4. BL-2 and P493-6 cells were transfected with 50 pmol indicated siRNAs using an Amaxa Nucleofector. For both cell lines, 800,000 cells were transfected using Nucleofector Kit V and Program C-09. After transfection, cells were plated in 2 mL normal growth medium. After 24 h, cells were collected by centrifugation at 90 × g for 5 min, and transfection was repeated using the same conditions of the first transfection. For qRT-PCR analysis, an aliquot of cells was collected 24 h after the second transfection. The rest of the cells were shifted to 3% (vol/vol) FBS and grown for an additional 24 h to measure cell viability. Viability was measured on 100 µL cell suspension in triplicate incubated with 20 µL CellTiter-Blue Assay (Promega). After incubation for 2 h, fluorescence (560_Ex/590_Em) was measured using a Synergy Multiplate Reader (BioTek).

Transfection of hT-RPE-MycER cells was achieved with the HiPerFect Transfection Reagent (Qiagen) using siRNAs at the final concentration of 50 nM. Activation of MYC was achieved 5 h after transfection by adding 150 nM 4-OHT (or an equivalent volume of ethanol) to the cells without changing the medium. For RNA extraction, cells were plated the day before transfection at 8 × 10⁴ per well in a six-well plate and collected 48 h later directly in the lysis buffer of the RNeasy Mini Kit (Qiagen). For viability and apoptosis assays, cells were plated the day before transfection in triplicate in a 96-well plate at 2,000 cells per well in 75 µL medium. The day after, siRNA/HiPerFect complexes were prepared in 25 µL serum-free DMEM and added to the cells; 48 h after transfections, 20 µL CellTiter-Blue Assay (Promega) was added to the cells, and fluorescence (560_Ex/590_Em) was measured 2 h later using a Synergy Multiplate Reader (BioTek). Apoptosis was measured by adding to the same samples 120 µL premixed Apo-ONE Homogeneous Caspase-3/7 Assay (Promega) according to the manufacturer. Fluorescence (485_Ex/530_Em) was measured 2 h later using a Synergy Multiplate Reader (BioTek).

Lentiviral Transduction.

shRNA constructs were generated using lentiviral Gene Ontology–GFP (LeGO-G) vector systems. Briefly, after insertion of shRNAs into pSUPER, a BglII/HindIII cassette, including the H1 promoter and the shRNA sequence, was subcloned into LeGO-G plasmids. Lentiviral particles were produced in HEK 293T after cotransfection of corresponding LeGO-G shRNA together with psPAX2 packaging and pMD2.G envelope plasmids. Supernatants containing viral particles were collected 72 h after transfection, and HEK 293Ts were infected with serial dilutions to determine virus concentrations. A multiplicity of infection of five was used for spin infection of cell lines. Three days after infection of cell lines, shRNA-transduced cells were sorted by GFP on an FACSAria III (BD Biosciences). Sequences of the shRNA used are provided in Table S4.

qRT-PCR.

For qRT-PCR analysis of gene expression, 800 ng total RNA was reverse-transcribed using the QuantiTect Reverse Transcription Kit from Qiagen. qRT-PCR was performed using the Luminaris Color HiGreen qPCR Master Mix (Thermo Fisher Scientific) and run on a Light Cycler 480 (Roche). A list of primers used is presented in Table S3.

Cell Cycle Analysis.

For cell cycle analysis, 1 × 10⁶ hT-RPE-MycER cells were transfected in 6 × 10-cm² dishes the same day of plating using siRNAs at the final concentration of 50 nM and 50 µL HiPerFect (Qiagen). After 24 h, cells were treated with either 150 nM 4-OHT or an equal volume of ethanol. After 24 h, cells were trypsinized, washed two times in cold PBS, and fixed with an equal volume of 100% ethanol for 30 min. Fixed cells were washed again in PBS and treated with 1 mg/mL RNase A. Cells were finally resuspended in propidium iodide (50 µg/mL) and analyzed by flow cytometry using a BDAccuri C6 Analyzer and the BD CSampler software.

ChIP.

ChIP was performed according to the protocols published by Lee et al. (67) and the Blueprint Project (www.blueprint-epigenome.eu) with some modifications. Briefly, every siRNA was transfected at the final concentration of 50 nM in 6 × 10-cm² dishes each plated the same day with 1 × 10⁶ hT-RPE-MycER cells using 50 µL HiPerFect (Qiagen). After 24 h, cells were treated with 150 nM 4-OHT to activate MYC. After 24 h, cells were fixed on the plate for 10 min at 4 °C in medium containing 1% formaldehyde. After blocking with 0.1 M glycine and washing with PBS, the cells were incubated for 10 min at 4 °C with cold LB1, harvested by scraping, and frozen for later use. After thawing on ice, the cells were resuspended in LB2, incubated for 10 min at room temperature, and dissolved in 3,000 µL LB3 after centrifugation. Sonication was performed for 45 min [three cycles of 15 min each at high power in pulsed mode (30 s on and 30 s off)] with a Bioruptor Sonicator (Diagenode). After the addition of 300 µL 10% (vol/vol) Triton X-100 and centrifugation, the supernatant was removed; 1,500 µL supernatant was incubated with 10 µg MYC antibody (N-262; Sc-764; Santa Cruz Biotechnology) or 10 µg IgG antibody (2729; Cell Signaling Technology) at 4 °C overnight. After the addition of 100 µL Dynabeads Protein G (Life Technologies/Thermo Fisher Scientific) and incubation for 3 h at 4 °C, the beads were washed and eluted. The eluate was reverse cross-linked and digested with RNase A and Proteinase K; then, the DNA was phenol/chloroform-extracted and precipitated. The dried DNA pellet was dissolved in 30 µL Tris⋅HCl; 50 µL supernatant was saved as an input DNA sample and processed as the eluate.