LncRNA DANCR upregulates PI3K/AKT signaling through activating serine phosphorylation of RXRA

Abstact Conventional therapies and novel molecular targeted therapies against breast cancer have gained great advances over the past two decades. However, poor prognosis and low survival rate are far from expectation for improvement, particularly in patients with triple negative breast cancer (TNBC). Here, we found that lncRNA DANCR was significantly overregulated in TNBC tissues and cell lines compared with normal breast tissues or other type of breast cancer. Knockdown of DANCR suppressed TNBC proliferation both in vitro and in vivo. Further study of underlying mechanisms demonstrated that DANCR bound with RXRA and increased its serine 49/78 phosphorylation via GSK3β, resulting in activating PIK3CA transcription, and subsequently enhanced PI3K/AKT signaling and TNBC tumorigenesis. Taken together, Our findings identified DANCR as an pro-oncogene and uncoverd a new working pattern of lncRNA to mediate TNBC tumorigenesis, which may be a potential therapeutic target for improving treatment of TNBC.


Introduction
Breast cancer is the most prevalent malignant tumor in women, and the efficacy of currently available therapies seems far from satisfactory, which severely threats the health of females from all over the world 1,2 . Generally, breast cancer can be classified into four subtypes, including luminal type A, luminal type B, human epidermal growth factor receptor-2 (HER2) positive type, and triple-negative breast cancer (TNBC) type 3 . TNBC demonstrates a lack of ER, PR, and HER2 expression by immunohistochemical results, and acts as a highly invasive subtype comprising about 20% of all breast cancer patients 4,5 . Furthermore, TNBC can exhibit high invasiveness, metastasis, high recurrence risk, and mortality rates, resulting in poor prognosis 6 . The lack of molecular targeted therapies and the poor survival of TNBC patients have fostered great endeavors to discover precise molecular targets for clinical treatment strategies. A growing body of evidence suggests that long non-coding RNAs (LncRNAs) are involved in TNBC progression through regulating tumor-related gene expression [7][8][9] . However, the precise molecular mechanisms by which lncRNA mediates TNBC progression remain unclear.
Differentiation antagonizing non-protein coding RNA (DANCR), an lncRNA encoded on human chromosome 4q12, has been identified as an oncogene in multiple malignant tumors, including colon cancer 10 , esophageal cancer 11 , hepatocellular carcinoma 12 , osteosarcoma 13 , TNBC 14 . DANCR was previously demonstrated to contribute to suppression of cell differentiation due to acting as a negative regulator 15,16 . Recent studies demonstrated that DANCR directly interacts with miR-758-3p 17 and miR-577 18 in non-small cell lung cancer and colorectal cancer, respectively. DANCR represses the expression of TIMP2/3 through physical binding with EZH2 in prostate cancer 19 . DANCR is also shown as a direct target of MYC in cancer 20 . Furthermore, DANCR expression is correlated with survival and/or prognosis of patients with hepatocellular carcinoma 12 , gastric cancer 21 , and colorectal cancer 10 . However, due to the molecular and phenotypic heterogeneity within and between different tumor types, mechanism explorations are required to elucidate the precise biological behaviour of DANCR in tumors, especially TNBC.
Here, we showed that DANCR is more highly expressed in TNBC compared with that in normal breast tissues, which exhibits poor prognosis. DANCR promoted proliferation and tumorigenesis in TNBC through activating ser49/78 phosphorylation of RXRA, and thus promoting PIK3CA expression. We also showed that signaling axis DANCR-RXRA-PI3K/AKT plays important roles in TNBC proliferation and tumorigenesis in vitro and vivo, respectively. In conclusion, the present study elucidated the function of DANCR in TNBC and might provide a novel of signaling pathway in the treatment of TNBC.

Clinical samples
Between August 2013 and August 2015, clinical specimens containing breast cancer tissues including 60 triplenegative (TNBC) type, 15 HER2 type, 15 Luminal A type, and 15 Luminal B type, and 10 normal breast tissues were obtained from department of Breast Surgery in Meizhou People' s Hospital. All these patients had not received chemotherapy and radiotherapy before the operation and all clinical samples were confirmed by pathology. This study protocol approval from the Research Ethics Committee of the Meizhou People' s Hospital, and written informed consent from each participant were obtained.

RNA extraction and quantitative RT-PCR
Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and extracted according to the manufacturers' protocol (Sigma-Aldrich). Reverse transcription (RT) was performed using the M-MLV Reverse Transcription Kit(Thermo Fisher Scientific). Quantitative RT-PCR was performed by means of Power qPCR Premix (SYBR Green) (Shang hai Generay Biotech CoCo., Ltd). GAPDH was used as a control. Primers are contained in Supplenentary Table S1.

ChIP-qPCR
ChIP assay was performed using EZ-Magna ChIP™ A/G Chromatin Immunoprecipitation Kit (Millipore-17-408) referring to the manufacturer's protocol.
Input genomic DNA and the purified immunoprecipitated DNA were used for qRT-PCR. Primers are contained in Supplementary Table 1.
Luciferase promoter assay pGL3-PIK3CA promoter wild or mutant type of the RXRA binding was performed co-transfection with or without RXRA using Lipofectamine RNAiMAX Reagent (Thero Fisher) referring to the manufacturer's protocol. pRL Renilla luciferase control reporter vector (Promega) was presented as a control. A dual-luciferase assay was analyzed 48 h after co-transfection using the Promega E1960 Dual-Luciferase® Reporter System in accordance with the manufacturer recommendation.

RNA immunoprecipitation (RIP) and RNA pull-down assays
The EZ-Magna RIP Kit (17-701, Millipore) was used for RIP assay with 10 μg anti-RXRA (21218-1-AP, Proteintech) referring to the manufacturer recommendation. The purified immunoprecipitated RNA and input genomic RNA were detected by quantitative RT-PCR. Biotinlabeled RNA was transcribed with Biotin RNA Labeling Mix (Roche 11685597910) and T7 RNA polymerase (Roche 10881775001), mixed with DNase I recombinant (Roche 04716728001), and purified with RNeasy Mini Kit (Qiagen 74904). Cell nuclear proteins were extracted using the pierce 78833 NE-PER (R) Nuclear and Cytoplasmic Extraction Reagents. Cell nuclear extract was mixed with Biotin-labeled RNA. Washed streptavidin agarose beads (Sigma-Aldrich) were added to each reaction. The binding protein was analyzed by Western blot assay.

Cell proliferation and soft agar colony formation
A WST-1 Assay Kit (Roche) was used for cell proliferation assay. Cells were seeded in a 48-well plate and then incubated at 37°C. Cell numbers were assessed with the WST-1 Assay Kit. For soft agar colony formation, cells were were split into the suspension of a singlecell, and then seeded in a media containing 0.4% top layer agar and 0.8% bottom layer agar in a 6-well plate. Cell culture media was changed every 4 days after seeding. Colonies were fixed with 4% paraformaldehyde and stained with 2% crystal violet solution after 2-3 weeks, respectively. The visible colony numbers scored and data were analyzed.

Tumorigenesis studies
BALB/c-nude mice female at an age of 4-5 weeks (SLAC, Shanghai, China) were randomly divided into 4 per group, and then MDA-MB-231 cells (3 × 10 6 ) were implanted subcutaneously into mammary fat pads of each mice. All animal experiments procedures were approved by Zhejiang Provincial People's Hospital the Guidance of Institutional Animal Care and Use Committee (IACUC). The IVIS Lumina imaging station (Caliper Life Sciences) was used for bioluminescence imaging. Different investigators independently performed mice allocation, surgery and the outcome assessing.

Statistical analysis
All statistical analyses were performed using the GraphPad Prism version 5.0. The significance of the data from patient specimens was determined by Pearson' s correlation coefficient. The significance of data from the vitro and vivo between experimental groups was determined by the Student' s test or Mann-Whitney U-test *P < 0.05 was considered statistically significant.

Expression of DANCR in clinical TNBC specimens
We examined DANCR expression using clinical samples consisting of TNBC tissues and normal breast tissues. DANCR was found to be significantly upregulated in TNBC tumor tissues compared with that in normal breast tissues (Fig. 1a). To support our finding, we downloaded the TCGA RNA-seq dataset and microarray dataset of invasive breast cancer specimens, respectively. In these datasets, There was a clear trend that the TNBC tissues exhibits higher DANCR expression compared to the paired peritumoral tissues ( Fig. 1e and Supplementary Figure 1A). Interestingly, higher level of DANCR was demonstrated to be associated with bigger tumor size (Fig. 1b). We then detected DANCR expression in breast cancer samples of various subtypes. DANCR showed significantly higher expression level in TNBC than that in the other subtypes of breast cancer (Fig. 1c). Furthermore, patients with high DANCR expression (n = 30) suffered poorer overall survival (OS) as compared to low expression group (n = 30) (Fig. 1d). Similarly, TCGA microarray dataset also demonstrated the result (Supplementary Figure 1B). Gene amplifications in DANCR were observed in mRNA overexpression (TCGA database from cBio-Portal) (Supplementary Figure 1C). Taken together, this data demonstrated that DANCR is amplified and overexpressed in TNBC tumors.

DANCR inhibition suppressed cell proliferation and tumor growth in TNBC
To further demonstrate the involvement of DANCR in TNBC, we tested DANCR expression in six breast cancer cell lines (BT549, MCF7, T47D, MDA-MB-231, MDA-MB-453, and MDA-MB-468) and a normal mammary epithelial cell line (MCF10A). DANCR expression was markedly upregulated in breast cancer cell lines as compared to that in MCF10A cell lines (Fig. 2a). Importantly, among all the six cancer cell lines, DANCR exhibited highest expression levels in two TNBC cell lines (MDA-MB-231 and MDA-MB-468).
We then tested the effect of DANCR inhibition on TNBC cell growth. Here, we knocked down the DANCR expression through treating MDA-MB-231 and MDA-MB-468 cells with the short hairpin RNA-mediated DANCR silencing (shRNA) or non-silencing control (shC), and found that depletion of endogenous DANCR significantly suppressed cell proliferation in both TNBC cell lines compared with the controls (Fig. 2b, c). Moreover, DANCR knockdown also impaired soft agar colony formation in MDA-MB-231 and MDA-MB-468 cells (Fig. 2d, e).
To further elucidate the TNBC tumorigenesis with DANCR, analysis of the orthotopic breast cancer model was performed. MDA-MB-231 cell lines transduced with shDANCR-1, shDANCR-2 or shControl (shC) were separately injected into the mice mammary gland fat pads. Remarkably, knockdown of DANCR was significantly decreased TNBC tumor growth compared with controls ( Fig. 2f, g). These data suggested that DANCR contributes to TNBC cell proliferation and tumor growth.

DANCR interacts with RXRA in TNBC cells
Recent studies have demonstrated that lncRNAs mainly function as sponges to bind functional proteins and then influence their downstream genes expression [23][24][25] . Thus, We hypothesized that DANCR-regulated tumorigenesis depends on its binding proteins. DANCR was examined for transcription factor binding sites using a JASPAR database of transcription factor binding profiles, which identified two RXRA-binding sites as the most potential candidates. It is predicted that RXRA may bind with DANCR at both 211 to 225 and 269 to 283 sites (Fig. 3a).
To test this, we performed RIP quantitative PCR with an antibody against RXRA from nuclear extracts of both MDA-MB-231 and MDA-MB-468 cells. We demonstrated that DANCR specifically bound to endogenous RXRA protein (Fig. 3b).
We next constructed DANCR plasmids containing a wild type and three mutant types including DANCR- RXRA binding site1 (Mut1), site2 (Mut2), or both sites (Mut1/2). As we predicted, re-expression of shRNA-resistant DANCR cDNA encoding the wild type restored the binding of DANCR with RXRA, whereas re-expression of a DANCR shRNA-resistant plasmid containing the DANCR-RXRA binding site1, site 2, or two sites mutants did not rescue it, suggesting that these two sites are critical for DANCR-RXRA binding (Fig. 3c, d).
To further validate the binding between DANCR and RXRA, we performed RNA pulldown (Fig. 3e), and deletion-mapping methods (Fig. 3f, g) to demonstrate whether RXRA would bind within specific regions of DANCR. These data identified a 300nt region at the 5′ end of DANCR required for the binding with RXRA (Fig. 3g). Taken together, the RIP, RNA pulldown, and deletionmapping data validate a specific binding between RXRA and DANCR. To explore the role of RXRA in DANCR-mediated TNBC tumor growth, we first assessed RXRA protein and expression in DANCR depletion TNBC cells. We found that DANCR knockdown did not affect the protein level of RXRA in MDA-MB-231 and MDA-MB-468 cells (Fig. 4a). Moreover, DANCR depletion had no effect on RXRA mRNA level (Fig. 4b). However, Western blot analysis showed that lower phosphorylation level of RXRA was found in the DANCR-knockdown group as compared to that of the controls (Fig. 4c). Furthermore, re-expression of shRNA resistant DANCR wild type rescued DANCR knockdown-inhibited RXRA phosphorylation, whereas reexpression of shRNA resistant DANCR mutant type of the RXRA binding did not affect it (Fig. 4d). Together, Our data suggest that DANCR enhance RXRA protein phosphorylation in TNBC.

DANCR facilitates PIK3CA transcription in a RXRA-mediated manner
Since PI3K signaling is critical for TNBC proliferation and Akt is a dominant downstream effector of PI3K signaling pathway 26 , we explored the effects of DANCR depletion on expression of phosphatidylinositol-4,5-biophosphate 3-kinase catalytic subunit alpha (PIK3CA) in MDA-MB-231 and MDA-MB-468 cells. As shown in Fig. 5a, b, knockdown of DANCR significantly inhibited PIK3CA protein and mRNA expression levels in both TNBC cells. However, knockdown of RXRA significantly rescued DANCR depletion-inhibited PIK3CA expression (Fig. 5c). These data suggest that DANCR may mediates RXRA to regulate PIK3CA expression.
To validate that RXRA regulates PIK3CA expression in TNBC cells, we predicted RXRA-binding sites to the promoter of PIK3CA using the JASPAR database of , targeted with streptavidin beads, and binding proteins were resolved in a gel. Western blotting assay of the specific binding of RXRA and DANCR. f, g RNAs corresponding to fragments in different regions of DANCR were treated as in (e), and binding RXRA was detected by western blotting assay. Error bars ± SD. *P < 0.05. **P < 0.01. ***P < 0.001. Data are representative from two independent experiments transcription factor binding profiles. One RXRA binding site was detected in the promoter of PIK3CA at the −1148 to −1134 site (Fig. 5d). RXRA Knockdown increased PIK3CA mRNA levels by qRT-PCR assays (Supplementary Figure 3). ChIP-qPCR assays using the antibody against RXRA demonstrated that RXRA bind to PIK3CA promoter (Fig. 5e). Promoter luciferase assays further validated that RXRA overexpression markedly decreased PIK3CA promoter transcriptional activity compared with the control, whereas the transcriptional activity was definitely restored after the mutation of RXRA binding site in PIK3CA promoter (Fig. 5f). Taken together, these data indicate that RXRA acts as a transcription repressors to inhibit PIK3CA expression.

DANCR-mediated RXRA phosphorylation depends on GSK3β
Since the glycogen synthase kinase 3 beta (GSK3β) had been reported to promote RXRA phosphorylation in colorectal cancer cells 27 , we detected whether DANCRregulated RXRA phosphorylation depends on GSK3β. As shown in Fig. 6a, compared with the control group, overexpression of DANCR increased RXRA binding with GSK3β, RXRA phosphorylation, Akt phosphorylation and PIK3CA expression in both TNBC cells (Fig. 6a). Knockdown of DANCR decreased RXRA binding with GSK3β, RXRA phosphorylation, Akt phosphorylation and PIK3CA expression in both TNBC cells (Fig. 6b). Furthermore, overexpression of GSK3β increased RXRA phosphorylation, Akt phosphorylation and PIK3CA expression (Fig. 6c) inhibited by DANCR knockdown. Consistent with previous study 27 , overexpression of RXRA increased its binding with GSK3β and decreased Akt phosphorylation and PIK3CA expression in MDA-MB-231 and MDA-MB-468 cells (Fig. 6d). Overexpression of RXRA significantly restored GSK3β association and Akt phosphorylation and PIK3CA expression inhibited by DANCR knockdown (Fig. 6d). These data support that DANCR depends on GSK3β to mediate RXRA phosphorylation in both TNBC cells.  Since the roles of RXRA phosphorylation mediated by DANCR in TNBC is not well understood, we focused on the role of RXRA phosphorylation in DANCRregulated TNBC cell proliferation. To detect whether RXRA phosphorylation is necessary to activate PI3K/AKT signaling pathway, we generated the GSK3β kinase deficient mutant plasmid, GSK3β-K58A. As shown Fig. 7a, overexpression of GSK3β cDNA encoding the wild type restored DANCR knockdown-inhibited RXRA phosphorylation, Akt phosphorylation and PIK3CA expression, whereas transfection of GSK3β kinase deficient mutant type did not rescue it. Next, we predicted potential serine phosphorylation sites of the RXRA protein using the Phosphor Motif Finder Program. Two serine phosphorylation candidate sites at Ser49 and Ser78 were identified. Transfection of RXRA wild type in MDA-MB-231 cells restored its binding with GSK3β inhibited by DANCR knockdown, and further decreased DANCR depletion-inhibited Akt phosphorylation, PIK3CA expression, cell proliferation, and soft agar colony formation (Fig. 7b-e). However, transfection of RXRA S49A/S78A mutant did not rescued it compared with the wild type group (Fig. 7b-e). Taken together, these data demonstrates that serine phosphorylation of RXRA plays a crucial role in activating the PI3K/AKT signaling pathway.

Discussion
In this study, we demonstrated a new mechanism by which DANCR-upregulated PI3K/AKT signaling pathway through activating serine phosphorylation of RXRA protein is important for TNBC cell proliferation and tumor growth (Fig. 8). By means of facilitating connection between RXRA and GSK3β, DANCR promotes the phosphorylation of RXRA, leading to suppress RXRA induced inhibition of PIK3CA transcription, thereby activating PI3K/AKT downstream signaling and ultimately promoting TNBC tumorigenesis.
Our data demonstrate that DANCR promotes TNBC tumorigenesis. DANCR was previously identified to play an important role in maintaining the undifferentiated cell state 15,16 . Previous work demonstrated that DANCR interacts with EZH2 to form a complex to promote stem cell characteristics 16,28 . It has been well documented that DANCR was overexpressed and amplified in breast cancer, and knockdown of DANCR significantly inhibited cell proliferation and invasion in breast cancer through facilitating binding of EZH2 to the promoters of ABCG2 and CD44 genes 14 . Here, we report that DANCR is upregulated in clinical TNBC samples, and higher DANCR level is positively correlated with poorer prognosis of TNBC patients. Knockdown of DANCR significantly inhibited TNBC cell proliferation, colony formation in vitro, and tumor growth in vivo. These results strongly support that DANCR is critical for TNBC tumorigenesis.
Our results also demonstrate that DANCR mediates TNBC through RXRA. RXRA had been reported to play  29,30 . Phosphorylation of the RXRA at serine 260 decreases its coactivator recruitment ability, leading to inhibition of RXRA transcriptional activity and enhanced cancer cell proliferation 31,32 . DANCR has also been identified as a molecular sponge mediating miR-758-3p 17 and miR-577 18 in non-small cell lung cancer and colorectal cancer, respectively. Furthermore, some studies showed that DANCR regulate breast cancer through binding and phosphorylating EZH2 25 . In this study, our data showed that DANCR directly bound with RXRA and mediated its phosphorylation. Knockdown of DANCR enhanced RXRA phosphorylation. Re-expression of shRNA resistant DANCR wild type impaired DANCR knockdownpromoted RXRA phosphorylation, whereas re-expression of shRNA resistant mutant type of the DANCR-RXRA binding did not affect it. Knockdown of RXRA rescued DANCR depletion-inhibited cell proliferation and soft agar colony formation. Taken together, our data demonstrate that DANCR regulate TNBC cell proliferation through RXRA phosphorylation.
Our results further suggest that DANCR mediates TNBC through RXRA-downregulating PI3K/AKT signaling pathway. Activating PI3K/AKT signaling was reported to be associated with poor OS in patients with breast cancer 33 . PIK3CA activation was also demonstrated to be critical for enhancing PI3K/AKT signaling 22 . RXRA was reported to mediate PI3K/AKT signaling in response to stem cell differentiation and provoke tumor suppression 34 . Inhibition of the N-terminally truncated RXRA association with the PI3K p85α subunit also resulted in suppression of PI3K/AKT signaling activation 35 . Here, we found that RXRA binds with the promoter of PIK3CA and downregulate PIK3CA transcription. We also observed that knockdown of DANCR inhibited RXRA protein phosphorylation and PIK3CA expression level. Interestingly, knockdown of RXRA rescued DANCR depletion-inhibited PIK3CA expression, cell proliferation, soft agar colony formation, and the promoter transcriptional activity of PIK3CA. Furthermore, We found DANCR markedly promotes RXRA association with GSK3β. In consistent with the previous study 31,32 , our data demonstrate that DANCR-induced RXRA phosphorylation suppresses RXRA-inhibited PIK3CA transcription in TNBC cells and ultimately activates the downstream PI3K/AKT signaling, whereas overexpression of GSK3β mutant type with deficient kinase activity or RXRA with mutated binding site with GSK3β dramatically blocked it. These data suggest that DANCR promoted TNBC tumorigenesis depends on its binding and phosphorylating RXRA, which leads to PI3K/AKT signaling pathway activation. In summary, our results identified DANCR as an oncogene promotes TNBC tumorigenesis through a distinctive mechanism by which DANCR facilitated RXRA phosphorylation depending on GSK3β, thereby inhibiting the function of RXRA as a transcription repressors, and ultimately enhanced downstream PI3K/AKT signaling. Our findings have shed light into a novel roles of DANCR in TNBC tumorigenesis which have significant implications on better understanding the function of DANCR in human cancers. We hope that the newly established roles of DANCR and RXRA in tumorigenesis may provide a strong rationale for targeting them to improve the treatment of TNBC patients. DANCR interacts with RXRA, and highly promotes GSK3β phosphorylating RXRA, and inhibits RXRA binding with PIK3CA promoter to upregulate its transcription, resulting in activating PI3K/ AKT signaling pathway and TNBC tumorigenesis