Long Non-coding RNA BTG3-7:1 and JUND Co-regulate C21ORF91 to Promote Triple-Negative Breast Cancer Progress

Background Triple-negative breast cancer (TNBC) is a type of highly invasive breast cancer with poor prognosis. Recently, massive data reveal that long non-coding RNAs (lncRNAs) play important roles in cancer progress. Recently, although the role of lncRNAs in breast cancer has been well documented, few focused on TNBC. In this study, we aimed to systematically identify functional lncRNAs and to explore its molecular mechanism on TNBC progress. Methods The recurrence of lncRNAs and their target genes were validated with TNBC biopsies and cell lines. Total one hundred and thirteen TNBC biopsies, including nineteen patient-matched samples, were collected. The profile of TNBC-related lncRNAs and their target genes were characterized by RNA sequencing (RNA-seq) and bioinformatic analysis. Tumor specific lncRNAs, which also showed biological function correlated with TNBC, were identified as potential candidates; and the target genes, which regulated by the identified lncRNAs, were predicted by the analysis of expression correlation and chromosome colocalization. Cross bioinformatic validation was performed with TNBC independent datasets from the cancer genome atlas (TCGA). The biological functions and molecular mechanism were investigated in TNBC model cell lines by cell colony forming assay, flow cytometry assay, western-blot, RNA Fluorescence in situ Hybridization assay (RNA FISH) and chromatin immunoprecipitation-qPCR (ChIP-qPCR). Results Abundant Lnc-BTG3-7:1, which targets gene C21ORF91, was specifically observed in TNBC biopsies and cell lines. Knockdown of Lnc-BTG3-7:1 or C21ORF91 strongly inhibited cell proliferation, promoted cell apoptosis and cell cycle G1-arrested. Meanwhile, investigation of molecular mechanism indicated that Lnc-BTG3-7:1, cooperated with transcription factor JUND, cis-regulated the transcription of C21ORF91 gene, and down-regulation of Lnc-BTG3-7:1/C21ORF91 suppressed GRB2-RAS-RAF-MEK-ERK and GRB2-PI3K-AKT-GSK3β-β-catenin pathways. Conclusions In this study, we identified a TNBC specific lncRNA Lnc-BTG3-7:1, which sustained tumor progress. Up-regulation of Lnc-BTG3-7:1 promoted the transcription of oncogene C21ORF91 and activated PI3K-AKT-GSK3β-β-catenin and MAPK pathways. Taken together, our results not only identified a biomarker for diagnosis but also provided a potential therapeutic target against TNBC.


INTRODUCTION
Triple-negative breast cancer (TNBC) is the most aggressive breast cancer subtype and is characterized by poor survival (Shah et al., 2012;Johnson et al., 2020), occurring from younger age and being prone to distant metastasis. Moreover, it is kind of a heterogeneous disease with complex genetic background (Network, 2012), which represents different prognosis and responds to chemotherapies (Masuda et al., 2013;Burstein et al., 2015). However, current treatment outcomes are unsatisfied even under target therapy and immunotherapy, and unfortunately chemotherapy has to remain as the preferred treatment (Pandy et al., 2019). Thus, finding new target and understanding its mechanisms are essential. Recently, many studies report that long non-coding RNAs (LncRNAs) are likely pervasive in cancers and can drive cancers development, drug resistance and progress (Gupta et al., 2010;Martin and Chang, 2012;Bhan et al., 2017;Chi et al., 2019), which can be considered as important cancer biomarkers and play critical roles in cell proliferation, cell cycle, cell apoptosis, cell invasion, and metastasis (Yang et al., 2014;Collina et al., 2019). Recently, although several lncRNAs in breast cancer have been well documented (Mourtada-Maarabouni et al., 2009), only a few of them focus on TNBC, such as MALAT1, LINK-A, HOTAIR, LINP1, LncRNA-RoR, Lnc-RNA BORG, LOC554202, and HULC (Loewer et al., 2010;Augoff et al., 2012;Tripathi et al., 2013;Shi et al., 2014;Tao et al., 2015;Jin et al., 2016;Lin et al., 2016;Zhang et al., 2016;Gooding et al., 2019), and the mechanisms are still unclear. Thus, potential TNBC specific lncRNAs are expected to be explored, and the molecular features of lncRNAs that underlie the development, metastasis, and relapse of TNBC remain to be elucidated. Therefore, in this study, we aimed to systematically identify specific lncRNA by using RNA sequencing (RNAseq), and to explore its functions and molecular mechanism on TNBC progress.

Patients' Tissue Samples
Total one hundred and thirteen TNBC biopsies, including nineteen patient-matched samples, were collected from Breast cancer biobank of Western China Hospital. This experiment was approved by Ethics Committee of Western China Hospital.

RNA Extraction and High-Throughput Sequencing
TRIzol reagent (Invitrogen, Carlsbad, CA, United States) was used to isolated total RNAs from tissues and cells following the manufacture's protocol. Library preparation and sequencing were performed as described in detail in our previous study (Deng et al., 2019). Briefly, RNA integrity and quality were assessed by 2100 Bioanalyzer RNA Nano Chip (Agilent Technologies, United States) and Nanodrop ND-2000 Spectrophotometer (NanoDrop Technologies, United States). Subsequently, library preparation was performed as described in NEBNext R Ultra TM Directional RNA Library Prep Kit for Illumina R (NEB, United States) following manufacturer's recommendations. The libraries were then sequenced on the Illumina Hiseq platform (Novogene Bioinformatic Technology Co., Ltd., China), following a 2 × 150-bp paired-end protocol. Sequencing data were available at the NCBI Sequence Read Archive under project number PRJNA553096.

Sequencing Data Processing and LncRNAs Identification
After pre-processing filtering of low-quality reads and adaptor trimming, high-quality sequencing reads were aligned to the human reference genome assembly (GRCh37) using TopHat2 (Kim et al., 2013). Transcript assembly and abundance estimation were performed using Cufflinks (Trapnell et al., 2012) and HTSeq (Anders et al., 2015). The Ensembl, LNCipedia, and non-code databases were chosen as references to retrieve known transcripts. Novel transcripts were subjected to coding potential analysis by Coding-Non-Coding-Index (CNCI) (Sun et al., 2013), Coding Potential Calculator (CPC) (Kong et al., 2007), and Pfam-Scan (Chojnacki et al., 2017). The transcripts with lengths greater than 200 nt but without coding function were defined as candidate lncRNAs. To perform differential expression analysis in matching tumor and normal tissues, LPEseq (Gim et al., 2016) was utilized. Features with q value < 0.05 and | log2(Fold Change) | > 1 was considered differentially expressed.

Target Gene Prediction and Function Analysis
The biological functions of differentially expressed lncRNAs were predicted via the positional relationship (cis-target) and expression correlation (trans-target) between lncRNAs and protein-coding genes. To classify lncRNAs cis-target genes, we searched protein-coding genes located in 100 kb upstream and downstream of differentially expressed lncRNAs. Proteincoding genes with absolute correlation value greater than 0.95 were considered to be trans-target genes. The interaction probability between lncRNAs and proteins was predicted by RPISeq (Muppirala et al., 2011). The functional analysis was conducted by KOBAS (Xie et al., 2011), and visualized by the clusterProfiler (Yu et al., 2012) and ggraph (Pedersen, 2017). To scrutinize how Lnc-BTG3-7:1 promotes tumor progress, coexpression network using the CEMiTool package (Russo et al., 2018) to identify functional protein interacted with its target gene C21ORF91 were constructed. Protein-protein network (PPI) from STRING database (Szklarczyk et al., 2015) was further integrated into co-expression modules.

Quantitative Real-Time PCR (qPCR)
The first strand complementary DNAs of the RNAs were synthesized using the ReverseAid First Strand cDNA Synthesis Kit (Thermo Fisher, United States). qPCR was performed using SYBR Green Supermix (Bio-Rad, United States). Lnc-BTG3-7:1 and C21ORF91 expression in cell lines was normalized to β-actin using the comparative Ct method. Random primer was used for reverse transcription. Then, forward primer of Lnc-BTG3-7:1 and C21ORF91: TGC AACAACCCCATTTTTCCTA/GAACGTGTGCATGTGCTAAG and reverse primer: AAGAGTTCGGGCTCATCTCAC/TGAG TACCAGCACCACAAAG, respectively, were used to conduct qPCR.

Short Hairpin RNA (shRNA) Synthesis and Transfection
Lnc-BTG3-7:1 and specific shRNAs were synthesized by Sangon Biotech (Shanghai, China). The shRNAs target Lnc-BTG3-7:1 and C21ORF91 gene were shRNA241/shRNA351 and shRNA73/shRNA764, respectively. Sequences of shRNA241 and shRNA351 were: GCTGCTTTGTACTGATTGTAACTCGAGT TACAATCAGTACAAAGCAGCTTTTTG and GGTGCAGTT AACAGAGTTACGCTCGAGCGTAACTCTGTTAACTGCACC TTTTTG, respectively. The sequences of shRNA73 and shRNA764 were GGAGCAGTTTGTAAACATTGACTCGA GTCAATGTTTACAAACTGCTCCTTTTTG and GCAAAGCT CCTACAGCAAATCCTCGAGGATTTGCTGTAGGAGCTTTG CTTTTTG, respectively. These shRNAs were annealing and then inserted into the pLKO.1 vector between the restriction sites AgeI and EcoRI. The propagated synthetic construct vectors from Escherichia coli were extracted using Plasmid DNA purification kit (Macherey-Nagel, Germany). Lentiviral particles were produced by co-transfecting expression vector sh241_pLKO.1, sh351_pLKO.1, sh73_pLKO.1, and sh764_pLKO.1 with viral particle packaging helper vector into 293T cells. Tires of the four kinds of viral particles were determined by limited serial dilution. The three TNBC cell lines were infected with the four kinds of packaged lentivirus. The efficiency of Lnc-BTG3-7:1 and C21ORF91 single or double knockdown were determined by qPCR.

Colony Forming Assay
Mock and Lnc-BTG3-7:1/C21ORF91 gene single or double knockdown TNBC cells were seeded in 12-well plate at 2 × 10 3 cells/well for MDA-MB-231 and MDA-MB-468 cells, and 1 × 10 3 cells/well for BT-549 cells, respectively. Then, cells were cultured for continuous 8 days. After fixation with 4% paraformaldehyde for 30 min, cells were stained with crystal violet solution for 30 min. Each 10 µg of protein was resolved on SDS-PAGE and transferred to PVDF membranes (Millipore, United States). Immunoblotting was performed overnight at 4 • C. The membranes were then washed with Tris Buffered saline Tween (TBST) three times and incubated with the corresponding secondary antibodies (1:5000) at room temperature for 1 h. Then, these membranes were washed with TBST for three times. Images were captured by Bio-Rad ChemiDoc MP (Bio-Rad, United States).

RNA Fluorescence in situ Hybridization Assay (RNA FISH)
An oligonucleotide probe that was complementary to Lnc-BTG3-7:1 (designed by Stellaris Designer) and labeled with Cy3 dye at 5' was purchased from Sangon Biotech (Shanghai, China). A Cy3-labeled sense oligonucleotide was used as negative control. The sequences of the lncRNA probe and negative control (NC) probe were as follows: CCCAGTCAACACTCATACTT, CCATCCTATACCAATCTCGA, respectively. MDA-MB-231 and BT-549 cells were collected at exponential growth phase and resuspended to obtain single-cell suspensions (5 × 10 4 /ml). Then, cells were seeded into multi-chambered coverglass slides and cultured at normal growth conditions overnight. The cells were fixed in 4% formaldehyde for 15 min after washing 5 min for 3 times by phosphate buffer saline (PBS). Then, cells were treated with RNA FISH Kit (Fluorescent in situ Hybridization Kit, RiboBio, China) as recommended by the manufacturer, and Lnc-BTG3:7:1 probe and NC probe were cultured at 0.5 µM in 37 • C overnight. Then, slides were incubated with JUND antibody (1:400, #5000S, CST, United States) for 2 h and incubated with the corresponding secondary fluorescence antibody (FITC, 1:200, Invitrogen, United States) for 40 min at room temperature after washing by PBS for three times. Last, the slides were sealed with DAPI Fluoromount-G (Yeasen Biotech, China). Images were captured by laser scanning confocal microscope (Nikon, Japan) and analysis with software NIS-Elements BR (Version 5.11.01).

Chromatin Immunoprecipitation (ChIP)-qPCR
ChIP assays (Jing et al., 2020) were performed for the MDA-MB-231, MDA-MB-468, and BT-549 cells. Cells (approximately 1 × 10 7 cells) were digested into cell suspension using trypsin and washed with PBS. Then crosslink was performed using 3% methanol and quenched by 0.125 M glycine. The cells were rinsed and scraped off into conical tubes. Every 0.1 g pellets were resuspended in 1 mL cell lysis buffer and then sonicate (SCIENTZ-II D, Ningbo Scientz Biotechnology, Zhejiang, China) until the cell lysis turned clear to fragmentate chromatin to desired size (100-500 bp). The sonicated lysates were diluted using RNA dilution buffer with RNase inhibitor. After prehybridization, JUND antibody (#5000S, CST, United States) and 40 µL of beads were added together for immunoprecipitation. IgG was used as negative control. The RNA-protein complex was eluted, reverse crosslinked, and purified for qPCR and westernblot, respectively.

Statistical Analysis
The level of significance was defined as P < 0.05. The difference between mean values was assessed by t-test using Prism GraphPad 8.0 (GraphPad Software, United States).

Characterization of LncRNA and Transcriptome of TNBC
According to the specific structural and non-coding features of lncRNA, 58,163 lncRNAs were identified in at least one sample. Of these, 48,551 (83.47%) were identified as known lncRNAs, and 9,612 novel lncRNAs were detected in TNBC tissues. To investigate the key lncRNAs involved in TNBC progress, the expression profiles of lncRNAs between tumor and normal tissues were compared in order to detect the differentially expressed lncRNAs (DE-lncRNAs). We identified 864 lncRNAs with high occurrence (occurrence >10) displayed differential expression in TNBC biopsies. Of the dysregulated lncRNAs, 193 lncRNAs had consistent expression patterns in TNBC biopsies. Among them, 62 lncRNAs were highly expressed in tumor tissues, whereas, 131 lncRNAs were down-regulated, compared with normal tissues. Several DE-lncRNAs with high occurrence were further verified in TCGA TNBC cohorts ( Figure 1A).
To better understand the role of lncRNAs in TNBC progress, functional analysis of target genes was performed to predict the biological functions of DE-lncRNAs. A total of 671 protein-coding genes were significantly correlated with neighbor lncRNAs (Figure 1B). We then conducted pathway analysis to gain insight into the functions of DE-lncRNA target genes, and found that the target genes of DE-lncRNAs were highly enriched in pathways in cancer ( Figure 1C). Moreover, numerous target genes were involved in oncogenic signaling pathways, such as PI3K-AKT signaling pathway (12 target genes), RAS signaling pathway (9 target genes), and MAPK signaling pathway (7 target genes). To find out candidate lncRNAs, we removed DE-lncRNAs whose occurrence less than 10 prior to subsequent analyses ( Figure 1C).
To address the question of how Lnc-BTG3-7:1 function in concert with target gene C21ORF91 to regulate TNBC progress, the possible interaction network was constructed. We then concentrated on module M3, which contained C21ORF91. Gene set enrichment analyses showed module M3 had higher activity in tumor tissues ( Figure 2G). Remarkably, not only genes in module M3 participated in signal pathway (Figure 2H), and directly related to Wnt/β-catenin, RAS/MAPK, and PI3K/AKT pathways (Figure 2I), but also functional enrichment analysis revealed that co-expressed genes of C21ORF91 were substantially enriched in PI3K/AKT pathway ( Figure 2J). The results indicated that Lnc-BTG3-7:1 and C21ORF91 gene might involve in Wnt/βcatenin, RAS/MAPK and PI3K/AKT pathways.
Lnc-BTG3-7:1 and JUND Co-regulate the Transcription of C21ORF91 in TNBC No previous study has reported the mechanism of how Lnc-BTG3-7:1 regulates C21ORF91. To investigate the underlying mechanism, analysis of lncRNA-protein interaction was used to predict the potential way, and the prediction showed that Lnc-BTG3-7:1 might be synergistic with JUND to co-regulate target genes through cis-acting SE-lncRNA mechanism (Figures 5A,B). Furthermore, in order to confirm the prediction, RNA-FISH and immunofluorescence co-localization analysis were performed to determine the sub-location of Lnc-BTG3-7:1 and co-localization with JUND protein in MDA-MB-231 and BT-549 cells. The location of Lnc-BTG3-7:1, which was mainly in nucleus of TNBC cells and co-located with JUND protein, was observed ( Figure 5C). More than that, we further performed ChIP-qPCR (Immunoprecipitation-qPCR), and significant enrichment of Lnc-BTG3-7:1 in JUND specific antibody immunoprecipitation group in MDA-MB-468 (P < 0.05****) and BT-549 (P < 0.05****) cells were observed compared with the negative control group (Figure 5D), which confirm the combination of Lnc-BTG3-7:1 and JUND. Thus, these results further showed that Lnc-BTG3-7:1 co-regulated with JUND in TNBC.
Altogether, based on the results, we found that Lnc-BTG3-7:1 was synergistic with JUND in the TNBC cells and potentially coregulated the transcription of C21ORF91.
Lnc-BTG3-7:1/C21ORF91 Activated AKT-GSK3β and MAPK Pathways in TNBC Previous report indicated that GRB2 is one of C21ORF91 potential target proteins (Wang et al., 2008). In line with this conclusion, we also found that C21ORF91 was involved in regulation of GRB2-RAS-RAF-MEK-ERK and GRB2-PI3K-AKT-GSK3β-β-catenin pathways (The data were shown in Figure 2H). To validate the activation of GRB2 signal, we checked the downstream kinase of MAPK and PI3K-AKT-GSK3β pathways. On the one hand, the result indicated that the phosphorylation of AKT at serine 473 a.a. residue was suppressed upon C21ORF91 and Lnc-BTG3-7:1 knockdown in MDA-MB-231, MDA-MB-468 and BT-549 cell lines ( Figure 5E). The suppression of p-AKTs473 resulted in hyper phosphorylation of GSK3beta, and subsequentially, down regulation of beta-catenin and c-myc. On the other hand, removal of C21ORF91 or Lnc-BTG3-7:1 inhibited the downstream of canonical MAPK kinases MEK1/2 and ERK1/2 ( Figure 5E).

DISCUSSION
Triple-negative breast cancer is a high invasive with poor prognosis subtype breast cancer, occurring from younger age and being prone to distant metastasis (Foulkes et al., 2010;Shah et al., 2012). Based on the National Comprehensive Cancer Network guideline (NCCN), there is few clinical success of current therapy, making most TNBC patients have to still rely on chemotherapy (Bianchini et al., 2016;Park et al., 2018), and the median overall survival (mOS) for advanced TNBC patients are only about 9-12 months (Belli et al., 2019). Meanwhile, identifying effective biomarkers for cancer prognosis and drug responsiveness is of great importance in improving the clinical management of cancer, but unfortunately, precise biomarkers of TNBC treatment response and prognosis have not yet been identified (Singh et al., 2010).
Recently, the whole genome and transcriptome sequencing technology have been widely used in discovering latest genome difference (Marotti et al., 2017), which indicate that lncRNAs contribute to a significant portion of the "dark matter" of the human transcriptome (Kapranov et al., 2010) and show a new insight into the regulator heterogeneity of TNBC. Meanwhile, the differential expression of lncRNAs between normal and tumor tissues suggests that it is similar to protein-coding oncogenes (Huarte and Rinn, 2010), and dysfunction of lncRNA strongly associates with cancer progress (Fang et al., 2020). For example, UCA1, LncRNA human ovarian cancer-specific transcript 2 (HOST2) and LncRNA nuclear enriched abundant transcript1 (NEAT1) are associated with lung cancer, ovarian cancer and prostate cancer, respectively (Chakravarty et al., 2014;Gao et al., 2015;Nie et al., 2016). Nowadays, several lncRNAs have been shown to modify critical breast cancer associated molecular  However, so far, there is rare outstanding results concerning TNBC in the field of lncRNAs (Gooding et al., 2019;Zhang et al., 2019), such as HOTAIR, which relates with luminal androgen receptor (LAR) subtype of TNBC (Gupta et al., 2010;Collina et al., 2019).
Also, previous studies have indicated that the C21ORF91 gene encoding cytosolic protein plays a role in biological processes, such as cerebral cortex neuron differentiation, cell differentiation and regulation of dendritic spine development (Danileviciene et al., 2019). Recently, few researches show relationship between the C21ORF91 gene and herpes labialis, in which C21ORF91 gene is identified as direct targets of miR-194 in hepatocellular carcinoma cells (HCC) (Kriesel et al., 2011;Bao et al., 2015), but its function is not clear. The way in which the C21ORF91 gene influences TNBC progress has not been established yet, and how Lnc-BTG3-7:1 regulates C21ORF91 gene is still unknown.
In this study, we first screened out TNBC specific lncRNA Lnc-BTG3-7:1, which sustained tumor progress, and this TNBC FIGURE 6 | In nucleus, Lnc-BTG3-7:1 co-regulates the transcription of C21ORF91 gene with nuclear synergistic transcription factor JUND. In cytoplasm, C21ORF91 protein regulated AKT phosphorylation, β-catenin, c-MYC and GSK-3β in GRB2-PI3K-AKT-GSK3β-β-catenin pathway and activated ERK phosphorylation in GRB2-RAS-RAF-MEK-ERK pathway. specific lncRNA and its target C21ORF91 gene were involved in Wnt/β-catenin and MAPK pathways, which were associated with cancer cell progress (Liu et al., 2019). According to the bioinformatic analysis, the 1-195th bases of the Lnc-BTG3-7:1 sequence overlapped with the 25-219th bases of the C21ORF91 and the coding DNA sequence (CDS) of C21ORF91 started from the 62nd base. Meanwhile, based on analysis of the gene sequence of C21ORF91, we not only found that Lnc-BTG3-7:1 located in the region of super enhance SE_13873\SE_18618, which belonged to cis-acting SE-lncRNA, but also lncRNA-protein interaction prediction showed that JUND and Lnc-BTG3-7:1 had high possibility of interaction. Thus, we predicted that Lnc-BTG3-7:1 might be synergistic with JUND to co-regulate target genes through SE-LncRNA mechanism, which was associated with Wnt/β-catenin and MAPK pathways.
Furthermore, in this study, we observed the functions of the Lnc-BTG3-7:1 and found its regulation mechanism in TNBC cell lines. (1) In functional analysis, single or double knockdown of Lnc-BTG3-7:1 and C21ORF91 gene, strongly induced diminishing of cell proliferation, increasing cell apoptosis and arresting cell cycle in G1. (2) In mechanism analysis, by using IF/FISH and ChIP-qPCR assay, not only Lnc-BTG3-7:1 was co-located with JUND in nucleus, but also Lnc-BTG3-7:1 was enriched in JUND specific antibody immunoprecipitation groups in TNBC cell lines. Taken together, our result demonstrated that Lnc-BTG3-7:1 co-regulate the transcription of C21ORF91 gene with nuclear synergistic transcription factor JUND. Meanwhile, Western-blot test showed that Lnc-BTG3-7:1 and the target C21ORF91 gene were involved in GRB2-PI3K-AKT-GSK3β-βcatenin and GRB2-RAS-RAF-MEK-ERK pathways in TNBC.
Thus, from the above, our findings suggested that Lnc-BTG3-7:1 and the target C21ORF91 gene were TNBC specific factors and participated in TNBC progress.

CONCLUSION
In our study, a TNBC specific lncRNA Lnc-BTG3-7:1 was screened out and verified, which was involved in TNBC progress and could activate PI3K-AKT-GSK3β and MAPK pathways by regulating transcription of the target C21ORF91 gene. In conclusion, our results not only identified Lnc-BTG3-7:1 as a biomarker for diagnosis, but also provided a potential therapeutic target against TNBC.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm. nih.gov/, PRJNA553096.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of West China Hospital of Sichuan University. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
LY and JJ: conceptualization. ZD and HX: formal analysis. ZD, HX, LT, ZX, and YJ: methodology. ZD, LY, JJ, and ZH: project administration. LY: writing -original draft. ZD, HX, and LY: writing -review and editing. All authors have read and agreed to the published version of the manuscript.