LncRNA-IUR1 is a novel lncRNA whose expression is low in Bcr-Abl-positive leukemic cells but highly induced by imatinib
To identify key long noncoding RNAs (lncRNAs) involved in Bcr-Abl-mediated tumorigenesis, an lncRNA microarray was employed to analyze the expression of lncRNAs in Bcr-Abl–positive K562 cells treated with or without the Abl kinase inhibitor imatinib. Numerous lncRNAs were found to display differential expression in response to imatinib treatment (Fig. 1A, B). Notably, the expression of a novel lncRNA, designated as imatinib-upregulated lncRNA 1 (IUR1), was highly induced by imatinib treatment (Fig. 1A-D). Intriguingly, lncRNA-IUR1 was expressed in a very low level in Bcr-Abl-positive cells from chronic myeloid leukemia patients as compared with that in normal subjects (Fig. 1E, F), which prompted us to explore the role of lncRNA-IUR1 in Bcr-Abl-induced tumorigenesis.
LncRNA-IUR1 is located between the purinergic receptor P2Y2 (P2ry2) gene and the FCH and double SH3 domains 2 (Fchsd2) gene on human chromosome 11 (Supplementary Fig. S1A-B). Next, we analyzed the protein-coding potential of lncRNA-IUR1 through software prediction and experiments. As shown in Fig. 1G, the coding potential calculator (CPC) analysis showed that the CPC score of each lncRNA-IUR1 transcript was minus, indicating “noncoding”. Using in vitro translation experiments, we failed to detect any specific protein band for lncRNA-IUR1 while a specific protein band (about 72 kD) was observed for the control gene gag (Supplementary Fig. S1C), supporting that lncRNA-IUR1 had no protein-coding capability. In addition, we examined the localization of lncRNA-IUR1 in Bcr-Abl-positive leukemic cells by subcellular fractionation analysis, and found that lncRNA-IUR1 was localized in both cytoplasm and nucleus (Fig. 1H).
Altering lncRNA-IUR1 expression affects Bcr-Abl-transformed cell survival in vitro and tumor growth in vivo
To dissect the role of lncRNA-IUR1 in Bcr-Abl mediated cellular transformation, we first evaluated the effect of altered lncRNA-IUR1 expression on Bcr-Abl-transformed cell survival. LncRNA-IUR1 knockdown K562 cells stably expressing lncRNA-IUR1 shRNA and control cells were generated, treated with imatinib, and subjected to cell survival analysis (Fig. 2A). We observed that disruption of lncRNA-IUR1 expression in K562 cells resulted in a significant increase in viable cells compared with the control cells after treatment with imatinib (Fig. 2B, Supplementary Fig. S2A-D), while no significant difference in cell cycle progression was found between control and lncRNA-IUR1 knockdown cells (Supplementary Fig. S2E). Consistently, lower levels of cleaved caspase-3, caspase-9 and PARP were observed in lncRNA-IUR1 depleted K562 cells than those in control cells after imatinib treatment (Supplementary Fig. S2F-G). The data suggest that knockdown of lncRNA-IUR1 promotes Bcr-Abl-positive cell survival in response to imatinib treatment. Then, we further examined the effect of lncRNA-IUR1 deficiency on Bcr-Abl-induced tumorigenesis in vivo. Control or lncRNA-IUR1 knockdown K562 cells were injected into nude mice subcutaneously, and tumor growth was examined. It was shown that tumors formed by lncRNA-IUR1 knockdown cells grew much faster than that formed by control cells (Fig. 2C, D). Analysis of Ki-67 expression in the tumors revealed that the level of Ki-67 was elevated in tumors formed by lncRNA-IUR1 depleted K562 cells than those formed by control cells, while a lower degree of apoptosis was observed in tumors formed by lncRNA-IUR1 knockdown K562 cells than those by the control cells (Supplementary Fig. S2H-K). These results demonstrate that silencing of lncRNA-IUR1 promotes Bcr-Abl-transformed leukemia cell survival in vitro and tumor growth in vivo.
On the other hand, we examined the effect of lncRNA-IUR1 overexpression on Bcr-Abl-induced tumorigenesis. LncRNA-IUR1 overexpressing K562 cells and control cells were generated, and analyzed for cell survival in response to imatinib treatment (Fig. 2E). In contrast to the promotion of Abl-positive cell survival evoked by lncRNA-IUR1 deficiency, enforced expression of lncRNA-IUR1 sensitized K562 cells to imatinib-induced apoptosis while had no significant effect on cell cycle progression (Fig. 2F, Supplementary Fig. S2L). Accordingly, overexpression of lncRNA-IUR1 remarkably impeded the growth of K562 cell xenografts in nude mice (Fig. 2G, H, Supplementary Fig. S2M). Collectively, these results demonstrate that lncRNA-IUR1 plays an inhibitory role in Bcr-Abl-induced tumorigenesis.
Identification of murine lncRNA-IUR1
Since human lncRNA-IUR1 critically regulates Bcr-Abl-mediated cellular transformation, this prompted us to identify murine homologous lncRNA-IUR1, which may help to further address the functional relevance of lncRNA-IUR1 in Abl-induced tumorigenesis under a more sophisticated and physiological circumstance. Sequence alignment analysis between human lncRNA-IUR1 transcript and the mouse genome, revealed a 442 bp mouse genome sequence with up to 72% homology to human lncRNA-IUR1, which we call lncRNA-mIUR1 (murine lncRNA-IUR1) (Supplementary Fig. S3A). Of particular interest, lncRNA-mIUR1 is located between the P2ry2 gene and the Fchsd2 gene on mouse chromosome 7 (Fig. 3A), which displays highly concordant genomic location with human lncRNA-IUR1 (Supplementary Fig. S1A), indicating that lncRNA-mIUR1 is probably the murine homologue of lncRNA-IUR1.
Then, we performed reverse transcriptase PCR and quantitative real-time PCR to examine the expression of lncRNA-mIUR1 transcript in response to imatinib treatment. Indeed, the expression of lncRNA-mIUR1 transcript significantly increased after imatinib treatment in v-Abl-transformed mouse cells (Fig. 3B, C). This is consistent with the upregulation of human lncRNA-IUR1 induced by imatinib in human Bcr-Abl-positive cells. These experiments support that lncRNA-mIUR1 is the murine homologous lncRNA-IUR1 and Abl-mediated regulation of lncRNA-IUR1 is evolutionally conserved across species. 5′ RACE and 3′ RACE experiments were performed to determine the full length of lncRNA-mIUR1. The full length of lncRNA-mIUR1 is 2257 nt and the sequence has been submitted to GenBank (MZ643464) (Fig. 3D).
Next, we analyzed the protein-coding potential of murine lncRNA-IUR1 transcript through software prediction and experimental study. As shown in Supplementary Fig. S3B-C, the coding potential calculator (CPC) and open reading frame (ORF) Finder analysis demonstrated that lncRNA-mIUR1 had no protein-coding capacity. Besides, using in vitro translation experiments, we failed to detect any specific protein band for lncRNA-mIUR1 while a specific protein band (about 70 kD) was observed for the control gene KU70 (Fig. 3E), supporting that lncRNA-mIUR1 is a non-coding RNA.
Murine lncRNA-IUR1 regulates v-Abl-transformed cell survival and tumor growth
We next probed whether murine lncRNA-IUR1 also plays an important role in Abl-mediated cellular transformation. We generated control and lncRNA-mIUR1 knockdown v-Abl-transformed NS2 cells, and examined the effect of lncRNA-mIUR1 deficiency on Abl-induced tumorigenesis. We observed that depletion of murine lncRNA-IUR1 in NS2 cells promoted cell survival in vitro and xenografted tumor growth in vivo (Fig. 4A-D). By contrast, overexpression of lncRNA-mIUR1 in NS2 cells dampened cell survival and tumor growth in mice (Fig. 4E-H). Overall, in line with the inhibitory role of human lncRNA-IUR1 in Bcr-Abl-mediated cellular transformation, these results demonstrate that murine homologous lncRNA-IUR1 also suppresses tumorigenesis induced by v-Abl oncogene. Together, these data suggest that lncRNA-IUR1 might be an evolutionarily conserved lncRNA that has critical functions in tumorigenesis induced by Bcr-Abl and v-Abl oncogenes.
Depletion of murine lncRNA-IUR1 in Abl-transformed cells promotes the development of leukemia in mice
To further address the functional involvement of lncRNA-IUR1 in Abl-induced leukemia, we generated a leukemia mouse model by injecting sub-lethally irradiated mice with control or lncRNA-mIUR1 knockdown GFP-positive NS2 cells stably expressing control or lncRNA-mIUR1 shRNA respectively (Fig. 5A). The effect of lncRNA-mIUR1 deficiency on Abl-mediated leukemia was then examined. As shown in Fig. 5B, the body weight of mice injected with lncRNA-mIUR1 knockdown NS2 cells lost much faster than that of mice challenged with control cells. Accordingly, the number of white blood cell (WBC) in the peripheral blood from mice injected with lncRNA-mIUR1 knockdown cells, significantly increased compared with that from mice treated with control cells, and meanwhile the number of red blood cell (RBC) decreased (Fig. 5C-D). No significant difference of platelet (PLT) was observed between mice treated with control or lncRNA-mIUR1 knockdown cells (Supplementary Fig. S4).
In addition, we examined the intensity of GFP signal in the whole body, and found that the GFP signal was much stronger in mice injected with lncRNA-mIUR1 knockdown NS2 cells than that in mice treated with control cells (Fig. 5E). Accordingly, spleens of mice injected with lncRNA-mIUR1 knockdown cells displayed obvious splenomegaly compared with that of mice treated with control cells (Fig. 5F, G). These observations demonstrate that loss of lncRNA-IUR1 promotes Abl-transformed cell survival and the development of Abl-mediated leukemia in mice.
Knockout of murine lncRNA-IUR1 in mice facilitates Abl-mediated transformation of primary bone marrow cells and leukemia formation in mice
To determine the role of lncRNA-IUR1 in malignant transformation by Abl oncogenes under a more physiological circumstance, we generated lncRNA-mIUR1 knockout (KO) mice (Fig. 6A, Supplementary Fig. S5A-B). The deficiency of lncRNA-mIUR1 was confirmed in multiple organs of the lncRNA-mIUR1 KO mice, including the spleen, bone marrow cells, thymus, and white blood cells from peripheral blood (Fig. 6B). Then, primary bone marrow cell (BMC) derived from wild type (WT) or lncRNA-mIUR1 KO mice, were infected with the retrovirus expressing Bcr-Abl oncogene, and the efficiency of transformation was measured by counting the number of Bcr-Abl-transformed cell clones. As shown in Fig. 6C, the clone number of Bcr-Abl-transformed BMCs from lncRNA-mIUR1 KO mice was significantly increased as compared with that of Bcr-Abl-transformed BMCs from WT mice, suggesting that knockout of murine lncRNA-IUR1 facilitated Bcr-Abl-mediated primary bone marrow transformation.
Next, we established the leukemia model using lncRNA-mIUR1 KO mice, and evaluated the effect of lncRNA-mIUR1 knockout on Abl-induced leukemia in mice. Sub-lethally irradiated WT or lncRNA-mIUR1 KO mice were injected with GFP-positive NS2 cells, and the development of Abl-mediated leukemia was examined (Supplementary Fig. S5C). As shown in Fig. 6D, after injection with NS2 cells, the body weight of lncRNA-mIUR1 KO mice lost much faster than that of WT mice. The number of WBC in the peripheral blood from lncRNA-mIUR1 KO mice was increased remarkably compared to that from WT mice, and meanwhile the number of RBCs decreased (Fig. 6E, F). No significant difference of PLT was observed between WT and lncRNA-mIUR1 KO mice challenged with NS2 cells (Supplementary Fig. S5D). Moreover, we examined the intensity of GFP signal in the whole body, and observed that the GFP signal in lncRNA-mIUR1 KO mice injected with NS2 cells, was much stronger than that in challenged WT mice (Fig. 6G). Additionally, spleens of lncRNA-mIUR1 KO mice displayed obvious splenomegaly compared with that of WT mice (Fig. 6H, Supplementary Fig. S5E). Collectively, these results reveal that knockout of lncRNA-IUR1 in mice facilitates Abl-mediated primary bone marrow transformation and leukemia formation in mice.
LncRNA-IUR1 negatively regulates STAT5-mediated GATA3 expression
To decipher the molecular mechanism by which lncRNA-IUR1 regulates Abl-mediated cellular transformation, we performed RNA sequencing (RNA-Seq) to analyze the differential expression of genes in control and lncRNA-IUR1 knockdown K562 cells (Fig. 7A). Of particular interest, the expression of GATA3, a critical transcription factor involved in multiple cell processes including T-cell differentiation, tumor progression and metastasis [38, 39], significantly increased in lncRNA-IUR1 knockdown K562 cells compared with the controls (Fig. 7A). Notably, it has been shown that GATA3 could promote leukemic transformation by driving MYC enhancer activity, and inherited GATA3 variants are associated with Ph-like childhood acute lymphoblastic leukemia and risk of relapse [38, 39], suggesting a possibility that lncRNA-IUR1 is involved in Abl-induced tumorigenesis through regulating GATA3 expression.
To test this possibility, we first determined whether the expression of GATA3 was regulated by lncRNA-IUR1 in Abl-transformed leukemic cells. Indeed, the expression of GATA3 vastly increased in lncRNA-IUR1 knockdown K562 cells, whereas lncRNA-IUR1 overexpression led to a significant decrease of GATA3 levels in the cells (Fig. 7B and Supplementary Fig. S6A). Disruption of lncRNA-mIUR1 expression in v-Abl-transformed NS2 cells, also led to a significant increase of GATA3 expression (Fig. 7C). Besides, the level of GATA3 in WBCs and BMCs derived from lncRNA-IUR1 knockout mice was obviously higher than that in these cells from WT mice (Fig. 7D, Supplementary Fig. S6B). In addition, expression of GATA3 was markedly decreased in imatinib-treated K562 cells in which the expression of lncRNA-IUR1 was highly induced (Fig. 7E). These results indicate that lncRNA-IUR1 negatively regulates GATA3 expression in Abl-transformed leukemic cells.
STAT5 has been reported as a transcription factor of GATA3 , which prompted us to probe whether STAT5 contributes to the expression of GATA3 in Abl-positive leukemic cells. As shown in Fig. 7F, GATA3 RNA levels were significantly decreased in STAT5 knockdown K562 cells, suggesting that STAT5 is required for the transcription of GATA3 in Abl-transformed cells. Therefore, we next asked whether STAT5 is involved in lncRNA-IUR1-mediated regulation of GATA3 in Abl-positive leukemic cells. To this end, we examined the effect of altering lncRNA-IUR1 expression on STAT5 activation in Abl-transformed cells. Indeed, elevated phosphorylation levels of STAT5 were observed in lncRNA-IUR1 depleted Abl-transformed cells, whereas overexpression of lncRNA-IUR1 caused an obvious reduction of STAT5 phosphorylation in cells (Fig. 7G and H, Supplementary Fig. S6C-E). Importantly, the increased GATA3 expression caused by depletion of lncRNA-IUR1 in the cells, could be reversed by the treatment with STAT5 inhibitor (Fig. 7I). In addition, we also evaluated the effect of lncRNA-IUR1 on the JAK2/STAT3 signaling. Neither depletion nor forced expression of lncRNA-IUR1 had significant effect on the phosphorylation of JAK2 and STAT3 in K562 cells (Supplementary Fig. S6F-G). These experiments demonstrate that lncRNA-IUR1 negatively regulates GATA3 expression through suppression of STAT5 activation.
LncRNA-IUR1 inhibits Abl-induced tumorigenesis by suppressing GATA3 expression
Next, we asked whether lncRNA-IUR1 suppresses Abl-induced tumorigenesis through regulating GATA3 expression. To test this, we first evaluated the role of GATA3 in Abl-mediated transformation. Control and GATA3 knockdown NS2 cells stably expressing control or GATA3 shRNA were generated, treated with imatinib, and subjected to cell survival analysis (Fig. 8A). We observed that proportion of viable cells in GATA3 knockdown cells was significantly reduced as compared with that in control cells after imatinib treatment, suggesting that GATA3 is associated with Abl-transformed cell survival in response to imatinib treatment (Fig. 8B). Then, we further investigated the effect of GATA3 depletion on Abl-induced tumorigenesis in vivo. Control or GATA3 knockdown NS2 cells were injected into nude mice subcutaneously, and tumor growth was examined. As expected, tumors formed by GATA3 knockdown cells grew much slower than that formed by control cells (Fig. 8C). Moreover, we observed that overexpression of GATA3 promoted cell survival of Abl transformants upon imatinib treatment (Supplementary Fig. S7A-B). These results demonstrate that GATA3 is required for efficient tumorigenesis induced by Abl oncogenes.
Then, we further investigated the involvement of GATA3 in regulation of Abl-induced tumorigenesis by lncRNA-IUR1. We generated K562 cells stably overexpressing empty vector (EV), lncRNA-IUR1, or combination of lncRNA-IUR1 and GATA3. These cells were treated with imatinib, and subjected to cell survival analysis (Fig. 8D, Supplementary Fig. S7C). As shown in Fig. 8E, less viable cells were detected in lncRNA-IUR1 overexpressing K562 cells compared with that in control cells. Intriguingly, forced expression of GATA3 in lncRNA-IUR1 overexpressing K562 cells can rescue the decreased cell survival caused by enhanced expression of lncRNA-IUR1. In line with these results, tumors formed by lncRNA-IUR1 overexpressing K562 cells grew much slower than that formed by control cells in the xenograft mouse model, whereas enhanced expression of GATA3 in lncRNA-IUR1 overexpressing cells can reverse the inhibitory effect of lncRNA-IUR1 overexpression on xenografted tumor growth in mice (Fig. 8F). Collectively, these results reveal that lncRNA-IUR1 inhibits Abl-induced tumorigenesis through suppression of GATA3 expression.