Targeting LncRNA LLNLR-299G3.1 with antisense oligonucleotide inhibits malignancy of esophageal squamous cell carcinoma cells in vitro and in vivo

Accumulating evidence has indicated that long non-coding RNAs (lncRNAs) play critical roles in the development and progression of cancers, including esophageal squamous cell carcinoma (ESCC). However, the mechanisms of lncRNAs in ESCC are still incompletely understood and therapeutic attempts for in vivo targeting cancer-associated lncRNA remain a challenge. By RNA-sequencing analysis, we identified that LLNLR-299G3.1 was a novel ESCC-associated lncRNA. LLNLR-299G3.1 was up-regulated in ESCC tissues and cells and promoted ESCC cell proliferation and invasion. Silencing of LLNLR-299G3.1 with ASO (antisense oligonucleotide) resulted in opposite effects. Mechanistically, LLNLR-299G3.1 bound to cancer-associated RNA binding proteins and regulated the expression of cancer-related genes, including OSM, TNFRSF4, HRH3, and SSTR3. ChIRP-seq (chromatin isolation by RNA purification and sequencing) revealed that these genes contained enriched chromatin binding sites for LLNLR-299G3.1. Rescue experiments confirmed that the effects of LLNLR-299G3.1 on ESCC cell proliferation were dependent on interaction with HRH3 and TNFRSF4. Therapeutically, intravenous delivery of placental chondroitin sulfate A binding peptide-coated nanoparticles containing antisense oligonucleotide (pICSA-BP-ANPs) strongly inhibited ESCC tumor growth and significantly improved animal survival in vivo. Overall, our results suggest that LLNLR-299G3.1 promotes ESCC malignancy through regulating gene-chromatin interactions and targeting ESCC by pICSA-BP-ANPs may be an effective strategy for the treatment of lncRNA-associated ESCC.


Introduction
Esophageal cancer (EC) is one of the most fatal malignancies with high mortality and morbidity worldwide [1,2]. EC consists of two major pathological subtypes: esophageal adenocarcinoma (EA) and esophageal squamous cell carcinoma (ESCC), with the latter accounting for 90% of all EC cases in the world [3,4]. Although significant progress has been achieved in the diagnosis and treatment for ESCC in the past decades, the five-year survival rate for patients with ESCC remains less than 30% [5,6]. Therefore, it is still necessary to explore the molecular mechanisms underlying ESCC pathogenesis in order to develop more effective therapy for ESCC.
Long noncoding RNAs (lncRNAs) are RNA transcripts with lengths exceeding 200 nucleotides (nt) but without protein coding potential [7]. It has been found that the expression of lncRNAs is strikingly cell-and tissue-type specific [8], indicating that lncRNAs may serve as sensitive biomarkers for different diseases. Importantly, lncRNAs play essential regulatory roles in proliferation, apoptosis, invasion, and differentiation in many cancers including breast, colon, liver, pancreatic, and lung cancers [9]. ESCC-related lncRNAs, such as TTN-AS1, LINC01503, and PCAT1 have also been identified in vitro and in vivo [10][11][12]. However, the regulatory mechanisms of lncRNAs in ESCC remains largely unclear. Moreover, while it has been proposed that targeting lncRNAs with specific antisense oligonucleotides (ASOs) may be a promising cancer therapeutics, several limitations, including tissue-specific delivery and the ability to penetrate into cancer cells, remain to be unsolved challenges for the clinical application of ASO-lncRNAs [13,14].
Previously, we have identified a novel lncRNA LLNLR-299G3.1 (ENSG00000272448.1) that was up-regulated in exosomes from patients with early-stage ESCC (GSE104926) [15]. However, whether and how LLNLR-299G3.1 may play a role in ESCC is unclear. The primary objective of the current study was to investigate the biological functions and the regulatory mechanisms of LLNLR-299G3.1 in ESCC. The second aim of this study was to investigate whether a novel type of nanoparticles conjugated with glycosaminoglycan placental chondroitin sulfate A binding peptide (plCSA-BP) was able to deliver ASO-LLNLR-299G3.1 for targeting inhibition of ESCC cell in vivo.

Materials and Methods
Patients and specimen Tumor tissues and adjacent normal tissues were collected from patients with ESCC undergoing surgical resections in the Department of Thoracic Surgery at Shenzhen People's Hospital, China [15]. No local or systemic treatments were administered to these patients before surgery. All collected tissue samples were immediately snap-frozen in liquid nitrogen and stored at −80°C. This study was approved by the Medical Ethics Committee of Shenzhen University Health Science Center (protocol no. 2016001). A written informed consent was obtained from each subject before participating in this study.

Construction of LLNLR-299G3.1-overexpression lentivirus vector
To generate lentiviruses for stably overexpression of LLNLR-299G3.1, full length cDNA of LLNLR-299G3.1 was cloned into the pHBLV-CMV-MCS-3flag-EF1-ZsGreen-T2A-PURO vector system (Hanbio Biotech., Shanghai China). The vectors were verified by sequencing and electrophoresis after digesting by EcoRI and BamHI (Thermo Fisher, Shanghai, China). The lentiviral vectors were co-transfected into 293T cells with packaging plasmid pSPAX2 and envelope plasmid pMD2G using the Lipofiter TM transfection reagent (Hanbio Biotech., Shanghai, China). After 48 h of infection, lentiviral particles in the supernatant of cell culture were collected and concentrated by ultracentrifugation at 4°C. Concentrated viruses were used to infect 5 × 10 5 cells in a 6-well plate, and then subjected to selection with 1.5 µg/ml puromycin in 10% FBS for 1 week. Stable overexpression cell lines or NC cell lines were identified by qRT-PCR assay.
Antisense oligonucleotide (ASO) and cell transfection Antisense oligonucleotide (ASO) against LLNLR-299G3.1 was synthesized by RiboBio (Guangzhou, China). To determine the in vitro interference efficiency, ASO was transfected into TE1 cells using the Lipofectamine 3000 kit (Life Technologies, Shanghai, China) and the expression level of LLNLR-299G3.1 was analyzed by qRT-PCR assay. For the in vitro ASO functional assays, cells were seeded in 96-well plates and transfected with 100 nM of the ASO using the Lipofectamine 3000 kit following the manufacturer's instruction. At Table S1.
RNA extraction and qRT-PCR Total RNA was extracted from either cells or tissues using the Trizol reagents (Invitrogen, CA, USA) according to the user's guide. 1 µg of total RNA was reversely transcribed into cDNA in a final volume of 20 µl using the PrimeScript RT reagent kit (Takara, Tokyo, Japan). Real-time PCR was performed using the SYBR Premix RT-PCR kit (Takara, Tokyo, Japan) with quantitative fluorescence gene amplification instrument (qTOWER 3 analytik jena, German) under the following conditions: 95°C for 5 min followed by 40 cycles at 95°C for 10 s, 60°C for 30 s, 95°C for 15 s, and then 60°C for 1 min. The results were normalized to the expression level of GAPDH (internal control) and the fold change of gene expression was calculated using the 2 −DDCt method [16]. All primers used in this study were synthesized by RiboBio (Guangzhou, China) and are listed in Suppl. Table S1.
Cell proliferation assay Cell viability was evaluated using the cell counting kit-8 (CCK-8) (DojinDo, Tokyo, Japan) following the manufacturer's guidelines. A total of 5000 cells per well were seeded in 96-well culture plates and incubated overnight at 37°C with 5% CO 2 . Subsequently, 10 µl CCK8 reagent was added into each well and incubated for 2h. Lastly, cell viability was measured every 24 h (0, 24, 48, 72, and 96 h) at the absorbance of 450 nm using the Synergy HTX multi-mode reader (BioTek, Germany).

Transwell assay
For transwell assay, 1 × 10 5 cells in fetal bovine serum (FBS)free medium were seeded to the upper chamber of the Transwell insert coated with Matrigel (BD Biosciences, CA, USA). RPMI-1640 medium containing 10% FBS was added into the lower chamber. After incubation for 24 h, the cells remaining on the upper chamber were removed with cotton wool. Cells that had migrated or invaded through the membrane were fixed by cold methanol and stained by 0.1% crystal violet. The stained cells were counted under a microscope in six randomly selected areas and the number of migrated cells was presented relative to untreated controls. All experiments were performed three times in triplicate.

Wound healing assay
The motility of cells was measured by wound healing assays. Cells were seeded in 6-well plates and cultured to 90-100% confluent for 48 h. Wounds were made in each well using pipette tips. The cell migration distance was observed at 0 and 48 h under a microscope. Then the relative percentage of the wound closure was calculated. The areas between scratches were measured using Image Pro Plus 6.0 software under the microscope (Nikon, ECLIPSE NI-U, Tokyo, Japan) and the relative percentage of the wound closure was calculated according to the formula (area of 0 h-area of 24 h)/area of 0 h.

Apoptosis assay
Apoptosis of cells was analyzed using the Annexin V-FITC cell apoptosis kits (BD Biosciences, CA, USA) according to the manufacturer's instruction. Briefly, ESCC cells were seeded into 6-well plates and cultured for 24 h. Cells were then stained with Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI). The apoptotic rates of cells were determined by flow cytometer (Beckman Coulter, Inc., IN, USA) equipped with CytExpert software (version 2.3).
RNA pull-down and mass spectrometry assays RNA pull-down analysis was carried out using the Pierce TM Magnetic RNA-Protein Pull-Down Kit (Thermo Fisher, Shanghai, China). Briefly, cells were firstly lysed in dilution buffer and RNA was extracted by Trizol reagent (Invitrogen, Shanghai, China). Then large Scale RNA Production System-T7 Kit (Promega, Beijing, China) was used to synthesize large quantities of RNA and RNA 3′ End Desthiobiotinylation Kit (Thermo Fisher, Shanghai, China) was applied to label the RNA with biotin. Thereafter, biotin-labeled RNAs were mixed with cell extracts and incubated with Pierce Nucleic-Acid Compatible Streptavidin Magnetic Beads (Thermo Fisher, Shanghai, China). The RNA-protein complex was isolated from magnetic beads using Biotin Elution buffer and boiled in SDS buffer. After RNA pull-down, equal amounts of samples pulled down by sense and anti-sense LLNLR-299G3.1 were loaded on SDS-PAGE (sodium dodecyl sulphate polyacrylamide gel electrophoresis). Then the gel was stained with Fast silver stain kit (Beyotime, Shanghai, China) according to the manufacturer's instructions. Specific bands were cut and analyzed by Q-Exactive mass spectrometer (Thermo Fisher Scientific, NH, USA). Protein identification was retrieved in the human RefSeq protein database (National Center for Biotechnology Information), using Mascot version 2.4.01 (Matrix Science, London, UK; http://www.matrixscience.com/).

Subcellular fractionation
The nuclear and cytoplasmic RNA fractions were extracted from TE1 cells using the NE-PER TM Nuclear and Cytoplasmic Extraction Reagents (Cat: 78835, Thermo Fisher) according to the manufacturer's protocols. LLNLR-299G3.1 expression levels were quantified by qRT-PCR assay, using U1 gene as nuclear control and GAPDH as cytoplasmic control.
Construction of ASO-loaded plCSA-BP nanoparticles (plCSA-ANPs) The plCSA-conjugated biodegradable lipidnanoparticles were synthesized using our previously published EDC/NHS bioconjugate techniques [17,18] ASO-LLNLR-299G3.1 was encapsulated in plCSA-BP-nanoparticles using our prior reported one-step sonication method [19] with minor modifications. Briefly, Dlin-MC3-DMA, DSPC, cholesterol and DSPE-PEG-COOH were solubilized in ethanol at a molar ratio of 50/10/38.5/1.5. The lipid mixture was added to an aqueous buffer (50 mM citrate, pH 4) and mixed to a final ethanol and lipid concentration of 30% (vol/vol) and 6.1 mg/ml, respectively. After mixing, the hydrated lipids were passed through an extrusion device (T&T scientific corporation, Knoxville, TN, USA) with an 80 nm diameter filter, following the vendor's instructions. The ASO (solubilized in a 50 mM citrate, pH 4 aqueous solution containing 30% ethanol) was added to the nanoparticles (pre-equilibrated to 35°C) using the extrusion device at a rate of 1 ml/min and extruded with an 80 nm diameter filter for 21 times. The final ASO to total nanoparticles ratio were 0.06 (wt/wt). The ethanol was then removed, and the external buffer was replaced with PBS by dialysis.

Animal experiments
Four-week-old female BALB/c nude mice were purchased from the Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China) and housed under SPF (Specific Pathogen Free) conditions at the animal care facility of the Experimental Animal Center of Shenzhen Institute of Advanced Technology, Chinese Academy of Science (China).
For ASO in vivo treatment experiment, fluconazole (fluc) labeled-TE1 cells stably overexpressed LLNLR-299G3.1 or the negative control cells (1 × 10 6 cells/mouse) were subcutaneously inoculated into the right flank area of each nude mice. One week after injection, mice were randomly divided into 4 groups (n = 5/group), treated with PBS (control), the plCSA-NNPs (plCSA-coated lipid nanoparticles loading with ASO-NC, 80 mg/kg), the ANPs (lipid nanoparticles loading with ASO, 80 mg/kg), and the plCSA-ANPs (plCSA-coated lipid nanoparticles loading with ASO, 80 mg/kg), respectively. Treatments were administered via the tail vein injection once a week. The state of the nude mice was observed every day and tumors were measured twice a week, and tumor sizes were calculated using the equation V = length × width 2/2 (V, volume). Tumor growth was monitored by the bioluminescence intensity using the IVIS spectroscopic imaging system (Perkin-Elmer, Waltham, MA) [19]. To investigate tumor metastasis, 21 days after injection, 200 mg/kg D-luciferin potassium salt was injected into the peritoneal cavity of nude mice. The mice were sacrificed by using 30% urethane, and the primary tumor, lung, liver and lymph nodes were dissected for examination of metastasis by fluorescence imaging. All animal procedures were approved and performed in accordance with the Guidance Suggestions for the Care and Use of Laboratory Animals by the Animal Care and Use Committee from the Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences (China).

RNA-sequencing (RNA-seq)
Total RNA was extracted from ASO-treated TE1 cells using the TRIzol (Invitrogen, Shanghai, China) reagents. RNA quantity and quality were measured using the NanoDrop ND-2000 Bioanalyzer (Agilent, CA, USA). RNA libraries were prepared using the TruSeq Stranded mRNA LT Sample Prep Kit (Illumina CA, USA) with 500 ng of purified total RNA according to the manufacturer's protocol. The libraries were sequenced on an Illumina HiSeq X Ten Platform at OeBiotech (Shanghai, China) and 150 bp paired-end reads were generated. Raw reads were processed using Trimmomatic software and the clean reads were aligned to the human genome database (GRCh38) using the HISAT2 software (2.0.5). Gene expression level were calculated by the FPKM (fragments per kilobase of transcript per million mapped reads) by HTseq-count softwares. Differential expression was analyzed using the DESeq R package (1.16.1) [15].

ChIRP-seq
ChIRP-seq (Chromatin isolation by RNA purification sequencing) experiments were performed as previously described with minor modifications [20]. In brief, seveal DNA probes against the LLNLR-299G3.1 full-length sequence were ordered from Dianxi Biotech (Shanghai, China) and were biotinylated at the 3′ end. TE1 cells were cross-linked, lysed sonicated, incubated with probe sets, and LLNLR-299G3.1-bound chromatin was retrieved. The ChIRPed RNA was purified with proteinase K buffer and then subjected for library construction using the NEBNext Ultra II RNA Library Prep kit (New England Biolabs, MA, USA). The purified DNA fragments were sequenced on an Illumina Novaseq PE150 system at Dianxi Biotech (Shanghai, China). Raw reads were mapped to human reference genome assembly (GRCh38) using the STAR (Spliced Transcripts Alignment to a Reference) version 2 software. Peaks were called using the MACS package. Motif enrichment analysis was performed using the online tool Homer (http://homer.ucsd.edu/homer/ngs/peakMotifs.html). The sequences of probes for LLNLR-299G3.1 are listed in Suppl. Table S1.

Statistical methods
Statistical analyses were performed using SAS v. 9.4 program (SAS Corp., NC, USA) or Prism GraphPad 8.0. Normality of data distribution was examined using the Shapiro-Wilk test. Normal distribution continuous variables between subgroups were analyzed using paired or unpaired student's t-test. Wilcoxon rank sum test, and Wilcoxon signed-rank test were used for comparing continuous data that were not in normal distribution. The Mann-Whitney U test was performed for the comparison of the xenograft tumor weights and volumes. All statistical analyses were performed using two-tailed p-values. Statistical significance was defined as a p-value of less than 0.05.
To verify the expression levels of LLNLR-299G3.1 in ESCC, we used qRT-PCR to analyze its expression in exosomes. The results showed that LLNLR-299G3.1 expression levels in exosomes from ESCC patients were significantly higher than that in healthy controls ( Fig. 2A). LLNLR-299G3.1 expression levels in several ESCC cell lines (KYSE180, KYSE30, KYSE70, and TE1) were also significantly higher than that in normal esophageal epithelial cells (NE3) (Fig. 2B). Moreover, the expression level of LLNLR-299G3.1 in ESCC tumor tissues was significantly increased, compared to that in adjacent tissues (Fig. 2C). Collectively, these results consistently indicated that LLNLR-299G3.1 was up-regulated in ESCC.

LLNLR-299G3.1 bound to cancer-associated proteins
To explore the mechanisms by which LLNLR-299G3.1 promotes ESCC cell proliferation and migration, RNA pulldown assays were performed to identify the protein partners binding to LLNLR-299G3.1. After silver staining (Fig. 3A), differential bands were subjected to mass spectrometry assays, revealing that LLNLR-299G3.1 bound to a number of RNA binding proteins (Suppl . Table S2). Among the RNA binding proteins identified, 15 proteins that had the highest  Mascot scores have been reported to be involved in the regulation of proliferation and migration in cancer cells, supporting the findings of in vitro analyses (Fig. 3B). KEGG Functional pathway analysis revealed that proteins derived from RNA pull-down assay were associated with cancerrelated pathways such as ribosome composing (hsa03010), carbon metabolism (hsa01200), and citrate cycle (TCA cycle) (hsa00020) (Fig. 3C). Gene Ontology (GO) analysis showed that these proteins were enriched in cellular components of intracellular ribonucleoprotein complex (GO: 0030529), nucleus (GO: 0005634), and extracellular matrix (GO: 0031012) (Fig. 3D). Molecular functions were mainly related to poly (A) RNA binding (GO: 0044822), double-stranded RNA binding (GO: 0003725), and RNA binding (GO: 0003723) (Fig. 3E), supporting the validity of RNA pull-down experiment. Nevertheless, biological processes were more enriched in mRNA catabolic process (GO: 0000184), translational initiation (GO: 0006413), and gene expression (GO: 0010467) (Fig. 3F).

LLNLR-299G3.1 regulated the expression of cancer-associated pathway genes
To get a better understanding of the regulatory mechanisms of LLNLR-299G3.1 in gene expression, we first determined the cellular localization of LLNLR-299G3.1 in the cytoplasm and nucleus. We found that LLNLR-299G3.1 was mainly expressed in the nucleus (Fig. 4A). Since nuclear lncRNAs are known to be mainly involved in regulating gene expression [21], we conducted RNA-seq to identify differentially expressed genes regulated by LLNLR-299G3.1 in ESCC cells. After silencing LLNLR-299G3.1 by ASO, both LLNLR-299G3.1-ASO and NC treated cells were subjected to RNA-seq analysis. Using a fold change S | 1.5 | and p-value < 0.05 as cut-off, we found a total of 1039 differentially expressed (DE) genes on LLNLR-299G3.1 silencing (Suppl .  Table S3; Figs. 4B and 4C). Following the identification of DE genes, KEGG pathway analysis was carried out to identify pathways that were perturbed upon LLNLR-299G3.1 knockdown. There were 38 signaling pathways significantly enriched in differentially expressed genes affected by ASO silencing (Suppl. Table S4; Fig. 4D). Among them, cytokinecytokine receptor interaction and neuroactive ligand-receptor interaction pathways were the top two most significantly ones, suggesting that these two pathways may play important roles in the regulatory mechanisms of LLNLR-299G3.1 in ESCC cells. In agreement with this, gene set enrichment analysis (GSEA) showed that there were significant correlations between knockdown of LLNLR-299G3.1 and expression of these two pathway genes (Figs. 4E and 4F). QRT-PCR analyses on selected genes in these two pathways confirmed the results of RNA-seq (Figs. 4G-4I).

LLNLR-299G3.1 regulated ESCC-associated genes through directly interacting with chromatins
Given that nuclear lncRNAs may regulate gene expression through chromatin interactions [22], we further examined      Table S5). The enrichment of LLNLR-299G3.1 chromatin binding regions is highly suggestive of its functional consequences on gene expression.
Next, we performed motif enrichment analysis of these regions and found that the motifs, such as PB0156.1, INSM1. PB0118.1, PB0133.1, MEIS1, and Rfx5, were significantly enriched (Fig. 5C), suggesting that LLNLR-299G3.1 might affect gene expression by regulating the transcriptional function of these motifs. By overlapping the DE genes (FC S 1.5, p < 0.05) detected by RNA-seq with the directly bound genes (peak score S 300) identified by ChIRP-seq, we found 186 overlapped genes between two sets of data (Suppl . Table S6). Among these overlapped genes, we found that several cytokine-cytokine receptor interaction pathway genes (OSM, TNFRSF4) and neuroactive ligand-receptor interaction pathway genes (HRH3, SSTR3) were physically occupied with LLNLR-299G3.1 and also significantly dysregulated upon LLNLR-299G3.1 silencing (Fig. 5D). As shown in Fig. 5E, LLNLR-299G3.1 directly bound to the promoter and gene body regions of these genes, indicating an important function of LLNLR-299G3.1-chromatin interaction in gene regulation.
Since DE-genes regulated by LLNLR-299G3.1 were enriched in cytokine-cytokine receptor interaction and neuroactive ligand-receptor interaction pathways, and ChIRP-seq analysis revealed that genes in these pathways had enriched chromatin binding sites for LLNLR-299G3.1, we speculated that LLNLR-299G3.1 could regulate the malignant phenotypes of ESCC cells by inducing genechromatin interactions. To test this hypothesis, we conducted rescue experiments by treating LLNLR-299G3.1overexpression ESCC cells with si-RNAs for cytokinecytokine receptor interaction pathway gene (TNFRSF4) and neuroactive ligand-receptor interaction pathway gene (HRH3), respectively. As shown in Fig. 6A, si-TNFRSF4 and si-HRH3 significantly suppressed the expression of TNFRSF4 and HRH3, respectively. And over-expression of LLNLR-299G3.1 increased proliferation of ESCC cells. However, the impacts of LLNLR-299G3.1 on cell proliferation were abolished by co-transfection with si-TNFRSF4 and si-HRH3, respectively (Figs. 6B and 6C). These findings suggest that LLNLR-299G3.1 may contribute to ESCC cell proliferation, at least in part, through interacting with cytokine-cytokine receptor interaction and neuroactive ligand-receptor interaction pathway genes. Since these genes contained enriched chromatin binding sites for LLNLR-299G3.1, our results also indicate that the effect of LLNLR-299G3.1 on cell proliferation of ESCC cells were dependent on chromatin-gene interactions.

Characterization and cellular uptake of the pICSA-ANPs
The pICSA-ANPs were assembled from PLGA, lecithin, DSPE-PEG-COOH, and ASO through a single-step sonication method (Fig. 7A). The mean diameters of the ANPs and pICSA-ANPs were 88.5 ± 5.6 and 113.4 ± 4.7, respectively, suggesting that nanoparticles peptide conjugation increased the size of the nanoparticles (Fig. 7B). TEM images of nanoparticles demonstrated that the particles had generally spherical morphologies (Fig. 7C). To evaluate the cellular uptake of nanoparticles, Cy5.5-labeled pICSA-ANPs were incubated with TE1 cells at 37°C for 30 min, and then the unbound nanoparticles were removed. Subcellular localization of pICSA-ANPs in the cells were observed using confocal laser scanning microscopy (TCS, SP5, Leica, Hamburg, Germany). Red fluorescence signals from pICSA-ANPs were detected in the cytoplasm of TE1 cells, suggesting that pICSA-ANPs were able to enter into TE1 cells (Fig. 7D).

Intravenous delivery of pICSA-ANPs inhibited ESCC tumor growth in vivo
To assess the therapeutic effect of pICSA-ANPs on ESCC cell growth in vivo, fluc-TE1 cells stably overexpressed LLNLR-299G3.1 were grafted into nude mice. The tumor burden was monitored using bioluminescence imaging and caliper measurement. As can be seen in Fig. 8A, the bioluminescence intensities of both pICSA-ANPs and ANPs groups were significantly lower than that of pICSA-NNPs and PBS groups, indicating inhibitive effect of ASOs on tumor growth. Interestingly, the bioluminescence signals from pICSA-ANPs treatment group were evidently weaker than that of ANPs treatment groups, indicating that pICSA-ANPs displayed stronger inhibitive effect on tumor growth than that of ANPs (Fig. 8B). Similarly, tumor volume of pICSA-ANPs treated group was significantly lower than that of ANPs group (Fig. 8C). Furthermore, survival analysis confirmed that pICSA-ANPs treatment resulted in longer survival probability than that of ANPs administration (Fig. 8D). No significant differences in body weight were found among the four nude mice groups during the entire period of animal experiments, suggesting the low toxicity of pICSA-ANPs, ANPs, pICSA-NNPs, and PBS (Fig. 8E). Together, these data demonstrated the effectiveness and safety of systemic administration of pICSA-ANPs for treatment of ESCC tumor growth in mice.

Discussion
Through genome-wide profiling of lncRNA landscapes, we have previously identified a novel lncRNA LLNLR-299G3.1 that was significantly enriched in exosomes from early-stage ESCC patients [15]. In the present study, we further verified that LLNLR-299G3.1 was up-regulated in ESCC tumor tissues and in ESCC cells. We showed that higher expression of LLNLR-299G3.1 enhanced the malignancy of ESCC cells. Whereas knockdown of LLNLR-299G3.1 resulted in contrary effects. Our findings suggested that LLNLR-299G3.1 was an oncogenic lncRNA in ESCC. Moreover, we showed that intravenous delivery of ASO-LLNLR-299G3.1 by plCSA-BP-ANPs significantly suppressed ESCC tumor growth with high efficiency and low toxicity, highlighting the potential of plCSA-BP-ANPs for ASO-based therapies in ESCC.
Antisense oligonucleotides (ASOs) are synthetic singlestrand DNAs that can specifically bind to target mRNA through Watson-Crick hybridization [23]. This DNA-RNA hybrid may recruit and activate the endonuclease RNase-H to cleave the RNA strand, degrade the mRNA, and prohibit protein translation [24,25]. Due to its ability to specifically inhibit transcription and translation of oncogenes, ASOs have been used for experimental therapy in some types of cancers [26]. However, several limitations, including low stability, weak affinity with cell membrane, and poor targeting restrict the translation of ASOs for clinical application [14]. Recent progressions in nanotechnology provide new opportunity for safer and more efficient delivery of anticancer agents to target cells or tissues [27].
However, results from in vivo treatment did not result in the desired therapeutic effects [28], partly due to low tissue specificity and varied tumor microenvironments [29]. It has been proposed that conjugate-mediated delivery might improve nanoparticle pharmacokinetics and efficacy in tissues [30]. For example, we previously developed a novel type of nanoparticles coated with glycosaminoglycan placental chondroitin sulfate A binding peptide (plCSA-BP), which could specifically bind to many types of cancer cells  [17]. We demonstrated that plCSA-BP-conjugated nanoparticles (plCSA-NP) had higher specificity and efficiency than conventional nanoparticles for delivering chemotherapeutic drugs to cancer cells and tumor tissues [17,18]. However, whether plCSA-NP could also effectively deliver lncRNA-targeting ASOs to cancer tissues has not been characterized.
It has been suggested that lncRNA-targeting ASOs could regulate the biological functions of lncRNA in cancers. For instance, BCAR4-targeted ASOs were found to be able to effectively inhibit metastasis in breast cancer mouse models [31]. Intravenous delivery of ASO-LINC00673 by liposomes significantly suppressed breast cancer cell proliferation in vivo [32]. However, the actual translation of ASOs into clinical applications has been hampered by several difficulties, such as the need to prevent ASO from degradation by nucleases in biological fluids, the transportation of ASO to target tissues, and the intracellular delivery of ASO to the target cell organs [33]. To overcome these limitations, several RNAi delivery approaches have been developed. Among them, lipid-based nanoparticles showed more promising potential due to their low toxicity and high biocompatibility compared with inorganic nanoparticles or viral systems [34]. Nevertheless, the instability of lipid nanoparticles in the biological environment, and the low specificity to tumor cells, have restrained their application in vivo [35,36]. plCSA is a fragment of protein VAR2CSA that could specifically bind to 90% of known cancer cells, and drug-conjugated VAR2CSA could inhibit tumor growth in mouse models in vivo [37]. PlCSA-BP, similar to VAR2CSA, was able to bind to multiple human cancer cells that expressed plCSA on the cell surface [18]. Previously, we and others have demonstrated that plCSA-BP was a specific guiding peptide for delivery of nanoparticle-loaded chemotherapeutic drugs to target tumor xenografts [17][18][19]. However, no attempts have been made regarding the use of plCSA-BP to guide the ASO-loaded nanoparticles to tumors. In this study, we showed that ASOs-loaded lipid nanoparticles (ANPs) were able to deliver ASO-LLNLR-299G3.1 to tumor xenografts and inhibited the growth of xenograft tumors. Moreover, we demonstrated that conjugation of plCSA-BP to ANPs significantly increased the inhibitive effects of ANPs on the tumor growth. To the best of our knowledge, this is the first report using plCSA-BP-nanoparticles to deliver lncRNAs-ASO to targeting tumors. Our results highlight the usefulness of plCSA-BP-nanoparticles as a novel delivery tool for ASO-targeted therapeutics in cancers.
Increasing evidence have suggested that RNA binding proteins (RBP) may play a role in modulating cancer initiation and progression through binding to RNAs and regulating their processing, stability, localization, modification or translation [38,39]. Our RNA pull-down analysis revealed that LLNLR-299G3.1 could bind to several proteins, including IGF2BP1, DHX9, YBOX1, and ANXA2P2. Consistent with our hypotheses, these lncRNA binding proteins have been associated with cell proliferation and migration in multiple cancer cells. For instance, IGF2BP1 could post-transcriptionally regulate the expression of some essential genes required for the control of tumor cell proliferation, invasion, and chemo-resistance, and was associated with a poor overall survival and metastasis in various types of human cancers [40]. DHX9 has been implicated in many cellular processes, including regulation of DNA replication, transcription, translation, RNA processing and transport, tumor cell maintenance and drug response [41]. YBX1 promoted tumorigenesis and metastasis via interacting with c-Myc and linc02042 in vitro and in vivo in ESCC [42]. ANXA2P2 was up-regulated in hepatocellular carcinoma (HCC) tumor tissues and in HCC cells. Higher expression of ANXA2P2 was associated with worse prognosis of HCC patients [43]. ILF3 was overexpressed in gastric cancer and contributed to cell proliferation and poorer prognosis in patients with gastric cancer [44]. Our findings, together with results from previous studies by others, suggested an important mechanism by which LLNLR-299G3.1 promoted ESCC cell proliferation and migration via its RNA-binding activities.
Our RNA sequencing analysis showed that knockdown of LLNLR-299G3.1 led to the dysregulation of multiple genes in ESCC cells. KEGG pathway analysis revealed that differentially expressed genes regulated by LLNLR-299G3.1 were significantly enriched in the cytokine-cytokine receptor interaction and neuroactive ligand-receptor interaction, pathways. The cytokine-cytokine receptor interaction pathway was involved in the initiation and progression of many types of cancers [45,46], including ESCC [47]. The neuroactive ligand-receptor interaction pathway has been associated with tumorigenesis and poor prognosis in cancers [48][49][50]. It has been found that lncRNAs contributed to cancer prognosis through a competing endogenous RNA (ceRNA) regulatory network with the "neuroactive ligandreceptor interaction" pathways [51].
It has been suggested that lncRNAs may regulate tumor progression through interacting with chromatin [52]. For example, HOTTIP promoted leukemogenesis by altering the chromatin signature at the FAS promoter and increasing its expression [53]. The oncogenic LINC00284 contains enriched chromatin binding sites for genes it regulated in breast cancer [54]. Meg3 was a cancer suppressor lncRNA which bound to chromatin on unique genomic regions in the c-Met gene in pancreatic neuroendocrine tumors (PNETs) [55]. Nevertheless, whether lncRNA and chromatin interaction may play a role in ESCC remains to be elucidated. Here, we showed that LLNLR-299G3.1 was predominantly nuclear and had genomic interactions. Among genic regions, the LLNLR-299G3.1-chromatin interactions were most common in regulatory regions such as intronic and intergenic regions. It was noteworthy that although a large number of genome loci were found to be associated with LLNLR-299G3.1 in ChIRP-seq analysis, only a small number of intersected genes were overlapped with corresponding DE-genes identified by RNA-seq. It could be attributed to the fact that binding of lncRNA to chromatin is a physical phenomenon while regulation of genes by lncRNA is a mechanical phenomenon [56]. These observations highlighted that a lncRNA might not regulate all the physically associated genomic region in every cell, but highly cell specific and context dependent. In RNA-seq analysis, LLNLR-299G3.1-dysregulated genes were significantly enriched in cytokine-cytokine receptor interaction and the neuroactive ligand-receptor interaction pathway genes, including OSM, TNFRSF4, HRH3, and SSTR3. The enrichment of chromatin interactions in these genes identified by further ChIRP-seq suggested that binding of LLNLR-299G3.1 on chromatin had functional consequence. Indeed, in subsequent rescue experiments, it was confirmed that the impacts of LLNLR-299G3.1 on ESCC cell proliferation were dependent on interacting with these genes. Collectively, our data suggest that LLNLR-299G3.1 is a chromatin regulator of ESCC-associated genes. Future experiments will be required to determine if LLNLR-299G3.1-chromatin interaction may alter chromatin structure and the expression of LLNLR-299G3.1 gene in ESCC cells.

Conclusion
In summary, we showed for the first time that LLNLR-299G3.1 was an oncogenic lncRNA in ESCC. Moreover, we found that LLNLR-299G3.1 may enhance ESCC progression through genome-wide interactions with chromatins among the genes it regulated. Finally, we demonstrated that plCSA-BP nanoparticle was an efficient tool for specific delivery of ASO-lncRNA to ESCC tumor tissues and may be a novel approach for targeted cancer therapy (Fig. 9).

Acknowledgement:
The authors thank all study subjects for their participation in this study. Availability of Data and Materials: All relevant data are available from the corresponding author upon request.
Ethics Approval: This study was approved by the Medical Ethics Committee of Shenzhen University Health Science Center (protocol no. 2016001). All animal experiment protocols were approved by the Institutional Animal Care Committee of Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences.

Conflicts of Interest:
The authors declare that they have no conflicts of interest to report regarding the present study.