The c-Myc-regulated lncRNA NEAT1 and paraspeckles modulate imatinib-induced apoptosis in CML cells

Chronic myelogenous leukemia (CML) is a malignant, clonal hematopoietic stem cell disorder characterized by a reciprocal translocation between chromosomes 9 and 22 [1]. This translocation results in the formation of the BCR-ABL oncogene, which encodes a chimeric BCR-ABL protein with constitutively active tyrosine kinase activity [2]. Imatinib (IM) was the first BCR–ABL tyrosine kinase inhibitor (TKI) and a frontline drug used in CML therapy. IM binds the ATP-binding site of BCR-ABL, thus preventing a conformational switch to the active form of the oncoprotein, blocking its downstream signal transduction.

Compared with the research progress of microRNAs, there are thousands of longer transcripts, long non-coding RNAs (lncRNAs), whose functions are unknown [3]. We previously demonstrated that the lncRNA NEAT1 (nuclear paraspeckle assembly transcript 1) plays a significant role in the regulation of acute promyelocytic leukemia (APL) cell differentiation [4]. NEAT1, which was demonstrated to be an essential structural component of the subnuclear structure paraspeckle, has two isoforms: 3.7 kb NEAT1_1 and 23 kb NEAT1_2. Recent studies have demonstrated that NEAT1 plays a crucial role in cellular transformation [5]. Despite the importance of NEAT1 in various cancers, little is known about the precise function and mechanism of regulation of nuclear paraspeckles. In addition, the involvement of NEAT1 in CML has not been defined, and the mechanism by which NEAT1 functions to impact BCR-ABL-induced tumorigenesis is unknown.

To identify lncRNAs associated with CML, we began by comparing the transcriptomes of peripheral blood obtained from patients with CML and healthy donors. RNA sequencing data processing and assessment of the transcriptome were conducted as detailed in our previous study [6]. The RNA-seq data analysis revealed 200 upregulated lncRNAs and 177 downregulated lncRNAs in CML patients (Fig. 1a and Additional file 1: Table S1). We selected several differentially expressed, well-studied lncRNAs for validation by qRT-PCR (Additional file 2: Figure S1A). From these analyses, we found that the lncRNA NEAT1 was downregulated in CML. Next, we examined the NEAT1 expression level in peripheral blood mononuclear cells (PBMCs) from 26 cases with de novo CML. qRT-PCR revealed that both NEAT1 and NEAT1_2 were significantly decreased in CML patient samples compared with granulocytes from healthy individuals (Fig. 1b). A similar pattern of expression was also observed in T-ALL (T cell acute lymphoblastic leukemia) patients (Fig. 1c) and several leukemia cell lines (Fig. 1d). This result suggested that NEAT1 may be involved in CML pathogenesis.

The BCR-ABL fusion protein is known to be the initiating factor for CML development via multiple pathways; thus, we explored whether NEAT1 downregulation was a consequence of BCR-ABL expression. We found that silencing BCR-ABL expression markedly increased the expression of NEAT1 (total NEAT1 and NEAT1_2) in CML cell line K562 cells (Fig. 1e and Additional file 2: Figure S1B). Additionally, we found that IM treatment could enhance NEAT1 expression in K562 cells (Fig. 1f). Importantly, this effect was also observed in primary CML cells that were freshly isolated from CML patients (Additional file 2: Figure S1C). The expression of NEAT1 could be restored after treatment with inhibitors of BCR-ABL-mediated pathways or repressed by an activator of a BCR-ABL-mediated pathway (Additional file 2: Figure S1D), suggesting that enhanced NEAT1 expression may depend on BCR-ABL-mediated pathways.

Since NEAT1 is regulated by BCR-ABL-mediated pathways, the luciferase activity of the NEAT1 promoter (methods see Additional file 3) was significantly activated by si-BCR-ABL treatment (Fig. 2a), and it was also activated by inhibitors of BCR-ABL-mediated pathway (Fig. 2b). These results suggest that NEAT1 might be transcriptionally regulated by a downstream effector of BCR-ABL-mediated pathways. NEAT1 has already been identified as a p53 target [7]. However, unlike most cancer types, the frequency of p53 mutations in CML is low [8]. c-Myc was previously identified as a common regulator of pathways downstream of BCR-ABL [9]; thus, we hypothesized that c-Myc might regulate NEAT1 expression. As expected, we found that NEAT1 expression was increased by c-Myc knockdown, and treatment of the c-Myc knockdown cells with IM further increased NEAT1 expression (Fig. 2c). Consistent with previous reports, the c-Myc target gene miR-17HG and p27 were regulated by c-Myc knockdown (Additional file 2: Figure S2A). Importantly, NEAT1 promoter activity was induced by transfection with si-Myc (Fig. 2d). Deletion analysis demonstrated that promoter activity is retained in the ~ 500 bp region upstream of the transcription start site (Fig. 2e). To test whether c-Myc could interact with the endogenous NEAT1 promoter, a ChIP assay was performed using K562 cells. Primers flanking the 500 bp region upstream and the transcriptional start site were then used in qRT-PCR assays in which we could detect c-Myc binding at the NEAT1 promoter (Fig. 2f). Collectively, our results demonstrated that c-Myc can act directly with the NEAT1 promoter in CML cells.

We next explored the functional role of NEAT1 in the IM-induced apoptosis in CML cells. NEAT1 was shown to be effectively knocked down by siRNA (Additional file 2: Figure S3A and S3B). The results demonstrated that NEAT1 knockdown combined with IM treatment could promote the apoptosis of K562 cells (Fig. 3a and Additional file 2: Figure S3C). Similarly, NEAT1-knockdown CML cells were more sensitive to proteasome inhibition (Additional file 2: Figure S3D). NEAT1 has been reported to be critical for nuclear body paraspeckle formation. It is known that SFPQ (splicing factor proline/glutamine-rich), NONO (p54nrb), and PSPC1 (paraspeckle component 1) are protein components of paraspeckles [10]. We found that siRNA-mediated depletion of SFPQ, but not p54nrb or PSPC1, prevented the apoptotic effects induced by IM in K562 cells (Fig. 3b). Accordingly, specific depletion of SFPQ displayed a less severe effect compared with control cells in proteasome inhibitor MG132-induced apoptosis (Additional file 2: Figure S3E). These results suggested that the SFPQ is required for NEAT1 mediated-apoptosis. Strikingly, we observed that the distribution of SFPQ protein in the nucleus was altered. SFPQ adopted a more punctate distribution in normal granulocytes (Fig. 3c), indicating decreased formation of nuclear paraspeckles in CML cells. These data suggested that NEAT1 may sequester SFPQ within paraspeckles, thereby limiting the availability of SFPQ to promote apoptosis [10].

To reveal the downstream targets of SFPQ, we analyzed differential gene expression using RNA-seq data from SFPQ-knockdown cells treated with or without IM (Additional file 2: Figure S3F and Additional file 4: Table S2). Of note, pathway analysis using DAVID Bioinformatics revealed that cell growth and death pathway genes were modulated by IM but resisted in the presence of si-SFPQ. (Fig. 3d and Additional file 5: Table S3), which is consistent with our observation that SFPQ participates in IM-induced apoptosis. Validation by quantitative real-time PCR demonstrated that SFPQ knockdown significantly reduces the expression of Bcl-2 family pro-apoptosis genes such as BBC3, BAX and BMF in IM-induced apoptosis (Fig. 3e). Our findings suggested that SFPQ regulates IM-induced apoptosis via cell death pathway proteins. Taken together, our findings that inactivation of c-Myc regulates NEAT1 expression, and NEAT1 paraspeckles, in turn, attenuated IM-induced apoptosis, suggesting an autoregulatory negative feedback loop that exerts c-Myc activity in apoptosis (Fig. 3f).

In summary, we showed that NEAT1 is an essential mediator of the apoptosis induced by IM in BCR-ABL-expressing cells. Moreover, we demonstrated that NEAT1 is a direct transcriptional target of c-Myc, c-Myc represses NEAT1 expression by binding to the NEAT1 promoter, and SFPQ is required for NEAT1-mediated apoptosis in K562 cells. Together, these findings suggest that targeting NEAT1 might represent a novel and important therapeutic strategy for treating leukemia.