C-MYC inducible onco-lncRNA LINC00036 acting as EGFR mRNA stabilizer via RNA-protein and RNA-RNA interactions decreases the sensitivity of getinib in human cancer

Background: The oncogenic lncRNA based strategies for combating cancer may usher in a new and promising paradigm in cancer therapy. However, few studies have been performed to solve such a critical issue. The complex traits and molecular mechanism of such lncRNAs in tumorigenesis and their relationship with sensitivity of getinib in human cancer have not been investigated. Methods: We aimed to identify and validate such a novel oncogenic LINC00036 using transcriptome sequencing approach and a large number of tissue samples of different types of cancer from the our cancer center cohort and public data cohorts from the Cancer Genome Atlas (cid:0) Gene Expression Omnibus and Cancer Cell Line Encyclopedia. Moreover, series of in vitro and in vivo experiments were performed to examine its roles in tumorigenesis and the sensitivity of getinib in different types of cancer cells. Special nanoparticle via a more potent delivery system was developed to investigate the feasibility of targeting LINC00036 in vivo. Furthermore, chromatin immunoprecipitation (ChIP)-sequencing, ChIP, actinomycin D assay, dual-luciferase reporter assay, RNA pull-down and RNA immunoprecipitation were performed were developed to uncover the molecular mechanism. Results: LINC00036 that associated with poor prognosis is signicantly upregulated in human cancer tissues. Series of in vitro and in vivo experiments reveal that LINC00036 promotes tumorigenesis and decreases the sensitivity of getinib in different types of cancer cells. LINC00036 targeting nanoparticle markedly reduced the growth of human cancer xenografts. Mechanistically, LINC00036 is a direct transcriptional target of c-MYC and a positive feedback loop of the c-MYC-LINC00036-EGFR axis exists in human cancer. LINC00036 acts as an EGFR mRNA stabilizer via RNA-protein and RNA-RNA interactions, inducing the hyper-activation of the downstream AKT and MAPK signaling pathways, which in turn decreases the sensitivity of getinib in human cancer. Conclusions: LINC00036, a c-MYC inducible onco-lncRNA, acts an in human cancer decreases the sensitivity of getinib through positive feedback loop of the c-MYC-LINC00036-EGFR axis. Overall, this study broadens knowledge regarding novel onco-lncRNAs and will assist in developing feasible onco-lncRNAs based-targeted therapeutic strategies to improve the sensitivity of getinib in human cancer.

mechanisms with diverse modes of action, including 1) inactivation by small interfering (si)RNAs and/or antisense oligonucleotides (ASOs) [9,10], aptamers, and ribozymes; 2) modulation by altering the lncRNA-encoding promoter activity [11]; and 3) blocking the interactions between lncRNAs and regulatory interaction sites using factors (such as small synthetic molecules/peptides) or aptamers [12]. In addition, the large number of regulatory interaction sites in lncRNAs provides a platform for actively pursuing novel oncogenic lncRNA (onco-lncRNA)-based therapeutic strategies against cancer [13,14].
Although the Food and Drug Administration of USA has recently approved the use of several RNA-based therapeutics for infectious diseases (such as hepatitis C virus and human immunode ciency virus infections) and rare diseases (polyneuropathy), the utility of onco-lncRNAs is still being evaluated in large pre-clinical or early-stage clinical trials [15]. Antisense oligonucleotides-mediated inhibition of MALAT1 reduces cancer progression [16][17][18] and metastasis in breast tumors [19]. Apart from direct anti-cancer therapy, onco-lncRNAs can also be considered for combination therapy to improve the speci city, sensitivity, and therapeutic efficacy of existing anti-cancer therapeutic drugs [20][21][22]. Knockdown of the lncRNA HOTAIR increased sensitivity of patients with head and neck squamous cell cancer to platinum therapy [23]. The eld of lncRNA therapeutics is currently undergoing a renaissance with advancements in chemotherapy, radiotherapy, and immunotherapy for various malignancies [24][25][26][27]. Hence, it is evident that targeting onco-lncRNA might be used for tumor therapy and can improve the speci city, sensitivity and therapeutic effect of existing anti-cancer therapeutic drugs. However, considering the large number of aberrantly expressed lncRNAs in various cancers, identi cation of the most fundamental onco-lncRNAs with broad anticancer activity remains highly challenging.
Epidermal growth factor receptor (EGFR) is a transmembrane receptor tyrosine kinase protein, and constitutive activation of EGFR has been shown connected with tumorigenesis, progression and unfavorable prognosis in human cancers [28,29]. Activation of EGFR results in ignition of downstream signaling pathways, such as phosphatidylinositol 3-kinase/serine threonine protein kinase (PI3K/Akt) and phosphorylating mitogen-activated protein kinase (MAPK) pathways [30][31][32], which play major roles in the proliferation, survival, drug sensitivity and resistance in many different cell types [33,34]. Clinical trials have veri ed that novel therapeutic regimens targeting EGFR including monoclonal antibodies, tyrosine kinase inhibitors, phosphatidylinositol 3-kinase (PI3K) inhibitors and antisense gene therapy, have become one of the most widely studied therapeutic targets in human cancers [35][36][37]. Ge tinib is a rstgeneration tyrosine kinase inhibitor that binds to the kinase domain of EGFR, preventing its phosphorylation and blocking downstream signaling cascades and is approved for the rst-line treatment of patients with EGFR mutation-positive metastatic non-small cell lung cancer [38]. However, the therapeutic e ciency of single-agent ge tinib treatment is variable [39,40]. The reduction of ge tinib sensitivity is one of the main obstacles and one of the important reasons leading to unfavorable prognosis. Many view co-targeting therapies as an important next step to improve the clinical outcome of cancer patient [38,41,42]. Currently, onco-lncRNAs have attracted a lot of attention as potential targets of drug-sensitivity in cancers due to their functions in cell proliferation, metastasis, and epithelialmesenchymal transition. In this study, we focused on the role of onco-lncRNAs in drug-sensitivity.
Hence, the aim of this study was to identify novel onco-lncRNAs associated with sensitivity of ge tinib and develop feasible onco-lncRNA-targeted therapeutic strategies against cancer. In this current study, LINC00036 was identi ed as a novel broad spectrum and potent cancer propellant via highthroughput transcriptome sequencing and further validation in human cancer samples from the Harbin Medical University Cancer Center (HMUCC) and public data cohorts. We further demonstrate an uncharacterized positive feedbck loop of the c-MYC-LINC00036-EGFR axis and conclude that LINC00036 acts as an EGFR mRNA stabilizer via both RNA-protein and RNA-RNA interactions, inducing the hyperactivation of the downstream AKT and MAPK signaling pathways, which in turn reciprocally enhances c-MYC transcription, and potentiates the receptor tyrosine kinase cascade during tumorigenesis and decreases sensitivity of ge tinib. It is noteworthy that targeting of LINC00036 represents a promising lncRNA-based therapeutic strategy for combating cancer and may usher in a new paradigm for next generation anti-cancer therapy.

Clinical sample collection
The 185 breast cancer tissues, 36 lung cancer, 32 glioma, and 21 gastric cancer tissues and their corresponding normal tissues were obtained from HMUCC. Eligible patients with a histological diagnosis of cancer who had received neither chemotherapy nor radiotherapy before surgical resection were recruited to the present study. All samples were frozen in liquid nitrogen immediately after surgical resection, and only tumors with > 80% tumor cells were selected for RNA extraction. Two independent senior pathologists con rmed the pathological diagnosis and molecular subtype of each cancer tissue.
This study conformed to the clinical research guidelines and was approved by the research ethics committee of the Harbin Medical University Cancer Center. We obtained written informed consent from all patients. Transcriptome sequencing for breast specimens Transcriptome sequencing of 15 breast cancer tissues, 15 adjacent normal tissues and three breast tissues were obtained from HMUCC. A total of 3 µg of RNA per sample was applied for RNA sample preparation. Ribosomal RNA was rst removed using the Ribo-Zero TM Gold kit (Epicentre, Wisconsin, USA). Next, sequencing libraries were prepared according to the manufacturer's protocol, and the libraries were then sequenced on an Illumina HiSeq 2500 platform to generate 100-bp paired-end reads. Raw sequencing and processed RNASeq data have been deposited into the NCBI GEO database under accession number GSE71651 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE71651).
BGISEQ-500 RNA-sequencing, ChIP-sequencing and small RNA-sequencing Total RNA was isolated from the cancer cells using the Promega isolation kit (Promega) according to the manufacturer's instructions. Raw sequencing and processed RNA-seq data were deposited in the NCBI GEO database under the accession number GSE124961 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE124961). For BGISEQ-500 ChIP-sequencing, six steps used for BGISEQ-500 ChIP-sequencing included sample detection, end repair and adaptor ligation, PCR, circularization, library quality control, and sequencing. Raw sequencing and processed CHIP-seq data were deposited in the NCBI GEO database under the accession number GSE124836 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE124836). For BGISEQ-500 small RNAsequencing, total RNA was isolated from the cancer cells using the Promega isolation kit (Promega) according to the manufacturer's instructions. Raw sequencing and processed small RNA-seq data were deposited in the NCBI GEO database under the accession number GSE124961 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc =GSE124961).
All transcripts were normalized using log 2 transformation. The expression of LINC00036 or EGFR was dichotomized by median expression as the cut-off to de ne high values (at or above the median) versus low values (below the median). An enrichment peak of c-MYC in the LINC00036 promoter was obtained from the ChIP-seq data of MCF-7 cells in the ENCODE portal (http://www.encodeproject.org). The putative c-MYC-binding site in the LINC00036 promoter was inspected using the JASPAR database (jaspar.genereg.net/). Putative TF-binding site in the transcriptional start of LINC00036 locus are listed in Table S1. The predicted microRNAs binding to the 3'UTR of EGFR TargetScan and RNA22 are listed in Table S2. Venn diagrams were drawn using the VENNY 2.1 online software (https://bioinfogp.cnb.csic.es/tools/venny/).  Table S1.

Cell proliferation assay
The cell counting kit-8 (Dojindo, Kumamoto, Japan) was used to assess cell growth. Approximately 2 × 10 3 cells per well were seeded in 96-well plates in a nal volume of 100 μL and then transfected with si-LINC00036-1 or si-negative control (NC). The absorbance was detected 0, 12, 24, 48, and 72 h after gene transfection. CCK-8 solution (10 μl) was added to each well and incubated for 2 h at 37 °C. The absorbance levels were measured at a wavelength of 450 nm. All experiments were performed in triplicate. To calculate the half inhibitory concentration (IC50), data were tted in GraphPad Prism 7.0, and the dose-response curve was plotted using the equation log (inhibitor) vs. response-variable slope. The IC50 was obtained using the following formula: Y=Bottom + (TopBottom)/(1 + 10^((Log IC50-X) *HillSlope)).

Cell apoptosis assay using ow cytometry
The transfected cells were seeded in 6-well plates (5 × 10 5 cells/well) and treated with 1.0 mg/l adriamycin. The cells were then digested with trypsin (Gibco® trypsin-EDTA, Thermo Fisher Scienti c), washed thrice with PBS, suspended in 500 μl binding buffer, and then incubated withFITC Annexin V Apoptosis Detection Kit I (BD Biosciences, 556547) according to the manufacturer's protocol. The stained cells were detected using the BD FACS Aria II ow cytometer (BD Biosciences, Hercules, CA, USA).
Transwell migration/invasion assays Transwell migration assays (Corning ® BioCoat TM control insert-No ECM, 8 micron pore size) and invasion assays (Corning BioCoat TM Matrigel invasion chamber) were performed according to the manufacturer's instructions. Cells which had migrated/invaded to the bottom of the insert were xed with 70% ethanol and stained with 0.5% crystal violet. The migrated/invaded cells were photographed under an inverted microscope, quanti ed using ImageJ, and represented as percentage of total area.

Western blot analysis
Brie y, the protein concentrations were determined using a bicinchoninic acid protein assay kit and bovine serum albumin as the standard. Equal amounts of protein were fractionated via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene uoride (PVDF) membrane (Millipore, Bedford, MA). The membranes were blocked for 2 h using 5% non-fat milk in Trisbuffered saline with Tween (TBST), and then probed overnight at 4 °C. Following incubation with the primary antibodies, the membranes were incubated with secondary antibody (1: 8000 dilution, Alexa Fluor® 700 goat anti-mouse IgG (H+L) or Alexa Fluor® 800 goat anti-rabbit IgG (H + L); Invitrogen) in PBS at room temperature for 1 h. The protein bands were imaged using the Odyssey infrared imaging system (LI-COR Biosciences, Lincoln, NE, USA) and quanti ed using the Odyssey v1.2 software (LI-COR Biosciences, Lincoln, NE, USA) by measuring the band intensity (area × OD) in each group. The band intensities were normalized to that of GAPDH or tubulin as the internal control. The western blot experiments were repeated thrice unless otherwise stated. The antibodies used are listed in Additional le 1: Table S2.

Immunohistochemistry
Para n-embedded tissue sections from patients with breast cancer were depara nized in xylene, rehydrated in a graded series of ethanol solutions, and then incubated for 20 min in 3% H 2 O 2 to block the endogenous peroxidase activity. Next, the sections were heated in target retrieval solution (Dako) for 15 min in a microwave oven (Oriental Rotor) to retrieve the antigen. Non-speci c binding was blocked via incubation with 10% goat serum for 2 h. The slides were then incubated overnight at 4 °C with the anti-EGFR primary antibody. An appropriate secondary antibody was added and incubated for 20 min at 37°C , and binding was visualized with 3, 39-diaminobenzidine tetrahydrochloride (DAB). After each treatment, the slides were washed thrice with TBST for 5 min. Tissues were xed in 4% neutral paraformaldehyde for at least 48h and then embedded in para n. For safety evaluation of the nanocomplex, consecutive para n wax-embedded tissue sections of vital organs (4-5 µm) were dewaxed, rehydrated, and stained with hematoxylin & eosin (H&E). The antibodies used are listed in Additional le 1: Table S1. Quantitative real-time PCR Total RNA samples from cells and tissue samples were isolated using the Trizol reagent (Invitrogen, Carlsbad CA, USA) according to manufacturer's protocols. Total RNA (0.5 μg) was then reverse transcribed using the High-Capacity cDNA reverse transcription kit (Applied Biosystems, MA, USA). The SYBR Green PCR master mix (Applied Biosystems, USA) was used to quantify the RNA levels, and GAPDH or U6 was used as an internal control. The qRT-PCR was performed on a 7500 FAST real-time PCR system (Applied Biosystems, USA). The primers used are listed in Additional le 1: Table S3, 4. LINC00036 and mRNA levels were normalized to those of GAPDH, while microRNAs levels were normalized to those of small nuclear U6. The 2 −ΔΔC t method was used to calculate relative expression levels.
Terminal deoxynucleotidyl transferase dUTP nick end labeling assay Apoptosis-induced DNA fragmentation was determined using the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay. The cells or tissues under different experimental conditions were xed with 4% (w/v) paraformaldehyde and processed using a commercial kit (Roche) in accordance with the manufacturer's instructions. Tumor samples were xed for 30 min, rinsed with PBS, blocked for 10 min using a solution of 96% methanol mixed with 4% H 2 O 2 at room temperature, and permeabilized with 0.2% Triton X-100 in PBS for 5 min at 4 °C. The TUNEL staining was performed using the in situ cell death detection kit (Roche) and the nuclei were stained with DAPI (Beyotime, China) for 10 min. The number of TUNEL-positive cells and the total number of cells were captured using a uorescence microscope (IX71 + DP72, Olympus, Japan) and apoptosis was determined using the ImagePro Plus software.
Isolation of nuclear/cytoplasmic fractions Nuclear/cytoplasmic fractions were isolated using the NE-PERTM nuclear and cytoplasmic extraction reagents (Catalog number 78835; Thermo Fisher) according to the manufacturer's protocol. Cytoplasmic and nuclear fractions were divided for RNA extraction. The GAPDH and U1 enzymes were used as qRT-PCR markers for the cytoplasmic and nuclear RNAs, respectively.

RNA uorescent in situ hybridization (RNA-FISH)
ViewRNA®Probe (Catalogue Number VA1-3016120, Santa Clara) was purchased to perform Fluorescence in situ hybridization (FISH) assay according to the manufacturer's protocol. LINC00036 hybridization was carried out in a moist chamber. After digestion with a working protease solution, slides were incubated with RNase III (AM2290, Life Technologies, USA) or RNase A (AM2272, Life Technologies) for 2 h. Standard immuno uorescence and imaging were performed by confocal microscopy.
Actinomycin D assay Cells were treated with 5 μM ( nal concentration) actinomycin D (Catalog number A1410; Sigma; dissolved in 100% ethanol). Actinomycin D was added to cells 0, 1, 2, 4, 6, 12, or 24 h prior to RNA extraction with the TRIzol reagent. Subsequently, qRT-PCR was used to analyze the changes in the RNA levels.
RNA immunoprecipitation (RIP) and RNA pull down assays RIP was performed using a Magna RNA-binding protein immunoprecipitation kit (Millipore, Bedford, MA) according to the manufacturer's instructions. Brie y, cell lysates were incubated with RIP buffer containing magnetic beads conjugated with negative control normal mouse IgG or human anti-PPP1R150 antibody. The samples were then incubated with proteinase K to isolate the immunoprecipitated RNA. Finally, puri ed RNAs were extracted and analyzed using real-time PCR to con rm the presence of the binding targets. RNA pull-down was performed using a magnetic RNA-protein pull-down kit (Pierce Biotechnology, USA) in accordance with the manufacturer's instructions.

Chromatin immunoprecipitation
Chromatin immunoprecipitation (ChIP) was performed using the EZ ChIP kit (Millipore, Bedford, MA, USA) according to the manufacturer's protocol. In total, 1 × 10 6 cells were xed in 1% formaldehyde at room temperature for 10 min, and the nuclei were isolated with nuclear lysis buffer supplemented with a protease inhibitor. The chromatin DNA was sonicated and sheared to 100-200 bp fragments. The sheared chromatin was immunoprecipitated at 4 °C overnight using an anti-c-MYC antibody (Cell Signaling Technology). Normal mouse IgG was used as the negative control and an anti-RNA pol II antibody (Millipore) was used as the positive control.

Luciferase reporter assay
Luciferase reporter vector-harboring lentiviral particles were constructed with the wild type or mutant 3¢ UTR of EGFR or LINC00036 and packaged by Shanghai GeneChem, Co., Ltd. The H293T cells were transfected with wild type or mutant reporter plasmid vector using Lipofectamine 2000 and were then cotransfected with microRNA mimics or the negative control. After 48 h, luciferase assays were performed using the dual luciferase assay system (Promega, catalogue E1910). Fire y luciferase activity was normalized to the Renilla luciferase activity for each sample.

LINC00036 targeting nanoparticles preparation and biodistribution
The LINC00036 targeting ASO was designed and synthesized by Integrated DNA Technologies (Skokie, Illinois, USA). The ASO and DOTAP liposomal transfection reagent were dissolved in methylene dichloride and allowed to stand for 20 min at room temperature according previous studies [43,44]. Next, the mixture was rehydrated in normal saline for future use. The mice of the control group were administrated with 0.9% normal saline via tail vein injection. The treatment groups were administrated with 20μl ASO or 20μl LINC00036 targeting nanoparticles (TTN) via tail vein injection (t.v.i). Three groups (four mice per group) of female BALB/c nude mice were used to evaluate the biodistribution of the TTN in vivo. Cyanine 5.5 was used to label TTN and ASO. The mice of the control group were administrated 200μl 0.9% normal saline and the treatment groups were administrated 150μl cyanine 5.5/ASO or 150μl cyanine 5.5/TTN via tail vein injection at time intervals of 1, 3, and 24 h. Imaging of cyanine 5.5-labeled ASO DNA oligo was monitored over time by detecting near infrared (NIR) uorescence in the whole animal according to the instructions of the IVIS Lumina Imaging System 100 series (Xenogen, Alameda, CA, USA).

Statistical analysis
All statistical analyses were performed using the R 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria) software program or GraphPad Prism 7 software (San Diego, CA). All immunohistochemistry, immunoprecipitation, western blot, and in vitro assays were conducted for more than three times to ensure reproducibility. The number of mice in each group has been indicated for speci c in vivo experiments. Correlations between the LINC00036 and EGFR expression were assessed using Pearson correlation coe cients. Unpaired Student's t-tests were used to detect signi cant differences among tumors or between tumor and normal samples. The OS was calculated as the time from surgery until the occurrence of death. The DSS indicated the duration of disease-speci c survival. Survival curves were constructed using the Kaplan-Meier method and were compared between subgroups with the log-rank test. All P values were two-tailed, and P < 0.05 was considered statistically signi cant. All data are presented as the mean ± standard deviation (SD) from at least three independent replicates.

LINC00036 is overexpressed and is a c-MYC inducible lncRNA in human cancer
To identify novel onco-lncRNAs associated in breast cancer we determined the changes in lncRNAs expression by comparing the expression pro le in 15 breast cancer tissues and paired adjacent normal tissues , and 3 normal breast tissues using high-throughput transcriptome sequencing. In the present study, we mainly focused on the upregulated lncRNAs given that these lncRNAs might serve as therapeutic targets or biomarkers. Based on the high-throughput transcriptome sequencing, LINC00036, which was one of the prominently upregulated lncRNAs in breast cancer tissues, was chosen for further evaluation (Fig. 1a). Next, 185 breast cancer samples, 36 lung cancer samples, 32 gastric cancer samples and 21 glioma samples from HMUCC were applied to validate the result. According to the qRT-PCR con rmation results, LINC00036 was found to be upregulated in these human cancers (Fig.1b-e).
Next, the possible transcription factor-binding sites in the LINC00036 promoter were detected and putative c-MYC-binding site was identi ed (Fig. 1g and Additional le 1: Table S5). Recent studies have shown that aberrant c-MYC activation drives key gene networks contributing to the pathogenesis of many human cancers, and an increasing number of lncRNAs have been found to be involved in c-MYCmediated signaling [45][46][47]. Gene Expression Omnibus (GEO) and TCGA were used to investigate relationship between c-MYC and LINC00036 in the human cancer and the knockdown of c-MYC decreased the expression of LINC00036 in GSE5823 (Additional le 2: Fig. S1A). Moreover, c-MYC expression was positively correlated with LINC00036 expression in 9,670 patients in 33 types of cancer from TCGA data ( Fig. 1h and Additional le 2: Fig.S1B-D). Notably, to validate the result, c-MYC knockdown signi cantly reduced LINC00036 expression in ve cancer cell types (Fig. 1i), whereas c-MYC overexpression observably upregulated LINC00036 in these cancer cells (Fig. 1j). ChIP-seq was performed to avoid false positive results and correctly interpret the mechanism of LINC00036 regulation. The results revealed signi cant enrichment of c-MYC peaks in the LINC00036 promoter (chr19-5571145-5571442) in UACC-812 cells (Fig. 1k, upper panel). Furthermore, obvious enrichment of c-MYC peaks was observed in the promoter region of LINC00036 in MCF-7 cells in the ENCODE database (Fig. 1k, bottom panel). This observation was veri ed using ChIP-PCR, which is in agreement with the results of ChIP-seq (Fig.1l).

LINC00036 acts as an oncogenic lncRNA in human cancer
To address biological function of LINC00036 in human cancer, the UACC-812 and MDA-MB-453 cells were transduced with lentiviru encoding shortin terfering RNAs (shRNA) for targeting LINC00036 expression (Additional le 2: Fig. S2A). The CCK-8 assay revealed that knockdown of LINC00036 signi cantly impaired breast cancer cell proliferation in UACC-812 and MDA-MB-453 cells (Additional le 2: Fig.S2B). Moreover, the measurement of apoptosis using ow cytometry analysis and western blot analysis revealed that LINC00036 knockdown increased the apoptosis, in contrast to the Sh-NC group (Additional le 2: Fig.S2C-E).  Fig. S2J). Of note, the CCK-8 assay revealed that LINC00036 knockdown also signi cantly inhibited these cells proliferation (Fig. 2a).
To further verify the result of LINC00036 on tumor growth in vivo, we developed mouse subcutaneous xenografts model using UACC-812, NCI-H1975, and SGC-7901 cell transfected with lentivirus. The LINC00036 knockdown signi cantly retarded the growth of breast cancer, lung cancer, and gastric cancer xenografts ( Fig. 2b-g). Next, the TUNEL assay indicated an increase in the number of apoptotic cells in the LINC00036 knockdown group than that of the Sh-NC group (Fig. 2h, i). Furthermore, compared to that in the Sh-NC groups, tumors in the LINC00036 knockdown group showed increased Bax expression and reduced Bcl-2 expression (Fig. 2j). Collectively, these in vivo experiment results were consistent with in vitro data that not only indicative of the oncogenic activity of LINC00036, but also suggested that it can be used as a potential therapeutic target for cancer treatment.

LINC00036 is a promising novel therapeutic target
A potent effective therapeutic model, ASO-targeting LINC00036 coupled with cationic lipid 1,2-dioleoyl-3 trimethylammonium-propane (DOTAP) [44,48], which is hereafter named as LINC00036 targeting nanoparticle (TTN) , was developed and applied in the following experiments. On an average, TTN werẽ 90 nm in diameter (Fig. 3a) and showed higher tumor accumulation than ASO alone at different time points. It was retained for at least 24 h in xenograft models of breast cancer, gastric cancer, and lung cancer (Fig. 3b). Compared to that in the ASO group and blank control group, treatment with TTN markedly reduced tumor growth (Fig. 3c-e). The TUNEL assay indicated the presence of more apoptotic cells in the TTN treatment group than in the other groups ( Fig. 3f and Additional le 2: Fig. S3A).
Furthermore, in accordance with the above results, tumor samples in the TTN treatment group showed increased Bax expression and decreased Bcl-2 expression (Fig. 3g).
In addition, different organs were harvested to investigate the toxicity effects of TTN treatment. No observable toxicity was observed in these organs in TTN treatment, although mild apoptosis and necrosis were observed in the TUNEL assay and H&E staining (Additional le 2: Fig.S3B-D). Venous blood samples were also collected to investigate the biochemical indexes after TTN treatment and the result showed that the TTN treatment negligibly affected the function of the liver, kidney, and heart (Additional le 2: Fig.  S3E).Thus, LINC00036 is a promising novel therapeutic target and targeting of LINC00036, especially coupled with a potent DOTAP, which may represent a novel next generation anti-cancer therapeutic strategy.
LINC00036 ignites the PI3K/AKT signaling pathway and MAPK signaling pathway through up-regulation

EGFR expression
To determine the function of LINC00036, RNA-seq was performed using UACC-812 cells transfected with three siRNAs individually targeting LINC00036. Results revealed that EGFR transcription decreased signi cantly in the LINC00036 knockdown group compared to that in the scramble group (Fig. 4a). EGFR has been reported to promote progression via the AKT and MAPK signaling pathway in human cancer [49][50][51]. EGFR expression was signi cantly inhibited by LINC00036 knockdown both at mRNA and protein levels in UACC-812, NCI-H1975 and SGC-7901 (Fig. 4b). Furthermore, LINC00036 knockdown decreased p-EGFR, p-AKT, p-MEK and p-ERK expression in the above three cancer cell lines and xenografts (Fig. 4c, d).
Positive correlation expression between LINC00036 and EGFR was observed in 50 samples from breast cancer patients in HMUCC (Fig. 4e). Furthermore, the relationship between LINC00036 and EGFR protein expression was next analyzed in 154 breast cancer tissue samples from HMUCC using immunohistochemical staining. Results demonstrated that EGFR was abundant in the tissues with LINC00036 -high expression group and vice versa (Fig. 4f, g). Next, we integrated multi-dimensional data from TCGA and CCLE projects to characterize the relationship between the EGFR and LINC00036 transcription landscapes. As expected, LINC00036 expression correlated positively with EGFR mRNA levels in 9 common types of cancer tissues including breast cancer in TCGA cohort ( Fig. 4h and Additional le 2: Fig.S4A-I). In addition, a similar result was obtained in 602 cancer cell lines across 33 cancer types from the CCLE project, which was in agreement with TCGA data (Fig. 4i).
LINC00036 stabilizes EGFR expression via RNA-protein interaction and acts as a competing endogenous RNA (ceRNA) LncRNAs can regulate gene expression via transcriptional interference in cis or in trans manner. The increased EGFR mRNA levels may be result of increased transcription or enhanced mRNA stability. To further investigate how LINC00036 contributes to EGFR expression, we rst detected LINC00036 subcellular localization in eight diffrent cancer cells by nucleocytoplasmic fractionation PCR analyses. The results showed that LINC00036 was located both in nuclear and cytoplasm, while predominant within the cytoplasm of cancer cells, which was further veri ed by single-molecule RNA uorescence in situ hybridization (FISH) (Additional le 2: Fig.S5A, B), and suggesting that LINC00036 might be involved in post-transcriptional regulation and RNA interactions [52,53]. The half-life of the EGFR mRNA was shorter in LINC00036 knockdown cells than in the control cells, as observed after treatment of the breast cancer cells with actinomycin D (Fig. 5a, b). A report indicates that PPP1R150 increases the stability of differentiation-related mRNAs in somatic tissues [54,55].Therefore, siRNA targeting PPP1R150 was designed to investigate the role of PPP1R150 in EGFR mRNA stability (Additional le 2: Fig.S5C). Results showed that the half-life of the EGFR mRNA was shorter in the PPP1R150 knockdown group than that in the control (Fig. 5c). Furthermore, RNA binding protein immunoprecipitation assays showed signi cant LINC00036 and EGFR enrichment in the PPP1R150 group compared to that in the IgG group (Fig. 5d, e). The RNA pull-down assays showed signi cant enrichment of PPP1R150 in the LINC00036 group than in the control group (Fig. 5f). Therefore, LINC00036 enhanced EGFR mRNA stability via the LINC00036-PPP1R150 interaction.
In addition to the regulation of mRNA stability, lncRNAs located within the cytoplasm affect miRNAmediated gene regulation by acting as ceRNA or natural miRNA sponges [54]. To investigate whether LINC00036 functions as a ceRNA for regulating EGFR mRNA expression, small RNA sequencing was performed and 138 potential microRNAs candidates that potentially bind to the LINC00036 transcript were identi ed (Fig. 5g). Furthermore, public databases, including TargetScan and RNA22, were used to predict the microRNAs binding to the 3¢ UTR of EGFR (Additional le 1: Table S6). Next, overlapping microRNAs were analyzed among the upregulated microRNAs in the LINC00036 knockdown sequencing data and TargetScan and RNA22 databases. Finally, 74 candidate microRNAs were obtained though this ltration process (Additional le 2: Fig.S5D). Quantitative reverse transcription (qRT)-PCR was performed to validate these 74 candidate microRNAs, and 12 microRNAs were nally identi ed to be upregulated signi cantly after LINC00036 knockdown (Fig. 5h). Furthermore, inhibitors for these 12 microRNAs were designed to examine LINC00036 expression. As expected, increase in LINC00036 expression was observed after treatment with these microRNA inhibitors (Additional le 2: Fig.S5E, F). Subsequently, EGFR mRNA and protein levels signi cantly decreased or increased when microRNA mimics or inhibitors, respectively, were applied (Additional le 2: Fig.S5G, H and Fig. 5i, j). The luciferase reporter assays showed that the wild-type 3¢ UTR of EGFR or LINC00036 was regulated by miR-125b-2-3p and miR-424-5p and that this effect was abolished by mutating the binding sites (Fig. 6a-d). Our results indicated that EGFR and LINC00036 were bona de direct targets of miR-125b-2-3p and miR-424-5p. In light of the above ndings, we hypothesized that c-MYC/LINC00036/miR-125b-2-3p/miR-424-5p/EGFR axis might play vital role in the tumorigenesis. The UACC-812 and MDA-MB-453 cells were transfected with different vectors/inhibitor/sh-LINC00036 and LINC00036 knockdown decreased cell proliferation signi cantly, whereas miR-125b-2-3p and miR-424-5p inhibitors markedly rescued theses phenotypes (Fig. 6e, f).
LINC00036 knockdown enhances human cancer cell sensitivity to ge tinib in vitro and in vivo To determine the role of LINC00036 on ge tinib sensitivity, multiple types of cancer cells were exposed with different concentrations of ge tinib for 48 h. CCK8 assays showed that knockdown of LINC00036 decreased the ge tinib IC50 value compared with that of the control cells, respectively (Additional le 2: Fig.S7A-H). To further con rm the role of LINC00036 in ge tinib sensitivity, we examined the effects of LINC00036 on cell proliferation with sh-IC50 concentration ge tinib treatment. As shown in Fig.7a, CCK8 proliferation assays showed that LINC00036 knockdown signi cantly inhibited cell proliferation compared with the control cells with sh-IC50 concentration ge tinib treatment. Moreover, LINC00036 knockdown decreased phosphorylation levels of EGFR signi cantly in human cancer cells (Fig. 7b).
In addition, we attempted to develop co-targeting therapies by combining ge tinib with TTN and investigate the synergistic anti-tumor effects. As expected, the combination of ge tinib with TTN improved tumor suppression (Fig. 7c, d) and increased the cancer cell apoptosis compared to that with ge tinib or TTN alone ( Fig. 7e and Additional le 2: Fig.S7I). Furthermore, western blot analysis showed that combination of ge tinib with TTN inhibited the cascade of EGFR and its downstream signaling pathway (Fig. 7f). Moreover, lung adenocarcinoma patients with high expression of LINC00036 showed less ge tinib sensitivity than patients with low expression of LINC00036 in GSE11969 dataset (Additional le 2: Fig. 7J).
High LINC00036 expression is associated with poor prognosis The association of LINC00036 expression with patient survival was analyzed in 185 patients with breast cancer in the HMUCC cohort. Patients harboring tumors with upregulated LINC00036 expression were associated with potential poor overall survival (Fig. 8a). Owning to the limit of patient follow-up span and the sample number in the HMUCC cohort, 33 cancer types in 14,586 cancer patients in TCGA were used to compensate for the shortage of HMUCC data and to evaluate whether the upregulated c-MYC-LINC00036-EGFR axis was robustly associated with the poor survival of patients. As expected, patients in LINC00036high /EGFR -high group was found to be consistently associated with poor overall survival (Fig. 8b).
Furthermore, patients with LINC00036 -high /MYC -high expression also signi cantly correlated with poor disease special survival (Fig. 8c). Notably, patients with LINC00036 -high /EGFR -high /c-MYC -high expression showed worse overall survival and disease special survival (Fig. 8d, e). Overall, these data indicated that the c-MYC-LINC00036-EGFR axis correlated with poor prognosis of cancer patients (Fig. 8f, left).

Discussion
Recently, aberrant expression of onco-lncRNAs has been reported in various cancers [2,56,57], and onco-lncRNAs have been shown to play important roles in tumor progression [58,59]. Identi cation of novel onco-lncRNAs, characterization of their function and regulatory mechanisms, and development of novel lncRNA-based targeted strategies for combating cancer appear promising and may introduce a new paradigm in cancer research. In this study, we identi ed that LINC00036, a c-MYC-activated onco-lncRNA, promotes cancer progression and decreases the sensitivity of ge tinib in human cancer via RNA-protein and RNA-RNA interactions. Here, we attempted to identify a novel onco-lncRNA and develop a feasible onco-lncRNA-targeted therapeutic strategy against cancer.
Base on transcriptome RNA sequencing and the human cancer samples analysis from HMUCC, we demonstrated for the rst time that LINC00036 is overexpressed in human cancer (Fig.1) and the oncoprotein c-MYC was the transcriptional regulator of LINC00036, as it can bind to its promoter region (chr 19-5571145-5571442) (Fig.1). As a master transcriptional factor, c-MYC regulates the expression of approximately 10-15% genes in the genome [60], including various protein-coding genes [61], as well as various lncRNAs [62,63]. Here, LINC00036 was shown to connect the c-MYC and EGFR axis, which correlated with poor survival. Our study uncovered a distinctive facet of tumorigenesis, which involves the novel positive feedback loop of the c-MYC-LINC00036-EGFR signaling pathway. It also provides a broad framework for understanding the regulation of EGFR tyrosine kinase signaling by lncRNAs, suggesting that LINC00036 may be a new therapeutic target in pharmacological strategies for human cancer. Thus, this is the rst study reporting the existence and function of the lncRNA-EGFR axis.
Recent clinical trials have demonstrated that novel therapeutic regimens targeting EGFR, including monoclonal antibodies, tyrosine kinase inhibitors, phosphatidylinositol 3-kinase inhibitors, and antisense gene therapy, can effectively treat diverse tumors, such as head and neck [64][65][66], bladder [67], lung [32], ovarian [68,69], and cervical cancers [70]. However, the therapeutic e ciency of single-agent treatment is variable [71]. Co-targeting therapies as an important step toward improving the clinical outcome of patients with advanced cancer should be encouraged [72][73][74]. Effective to co-targeting requires the identi cation of other critical modulators of the oncogenic cascade that drives disease progression and the development of strategies that can target EGFR pathways. In this study, co-targeting of LINC00036 and EGFR via LINC00036 knockdown and ge tinib synergistic therapy signi cantly enhanced antitumor e cacy and ge tinib-sensitivity (Fig.8), indicating that LINC00036 alone or in combination with EGFR inhibitors can overcome resistance to EGFR-targeted therapies.
Generally, lncRNAs interact with chromatin DNA, mRNA, or protein, thereby regulating chromatin accessibility, mRNA stability, and protein activity or stability in cis or in trans manner [74][75][76][77]. Unlike the classical theory, the identi cation of LINC00036's RNA-RNA and RNA-protein-mediated dual regulation in this study sheds light on the lncRNA-mediated complex regulation of its target genes. We showed that LINC00036 acts as an EGFR mRNA stabilizer, thereby inducing the cascade of the downstream AKT and MAPK signaling pathways, which in turn directly promote tumorigenesis (Fig.4). Mechanistically, LINC00036 upregulated EGFR expression by recruiting PPP1R150, promoting EGFR mRNA stability (Fig.5). In addition to the RNA-protein model, we have revealed another parallel ceRNA model. A systematic screening of microRNAs has been performed in this study to identify potential ceRNA candidates regulating LINC00036 and EGFR expression. Furthermore, our study suggests that LINC00036 functions as a ceRNA to upregulate EGFR expression by acting as a sponge for miR-125b-2-3p and miR-424-5p ( Fig.5 and Fig.6). Thus, this study lls in the theoretical gap in the lncRNA regulation model and provides comprehensive understanding regarding the above-mentioned dual regulatory network in human cancer.
LncRNAs can be targeted by multiple mechanisms with diverse modes of action. Of them, ASOs are attractive candidates for targeted therapeutic delivery. On one hand, post-transcriptional RNA degradation pathways can knock down pathogenic RNAs. This can be achieved using siRNAs that trigger a Dicer and Argonaute-dependent cleavage pathway. Alternatively, ASOs with chemical modi cations can be used to target the RNA of interest for degradation via an RNAse H-dependent mechanism. On the other hand, one can also achieve loss of function via steric inhibition of RNA-protein interactions or prevention of secondary structure formation. RNA binding small molecules or ASOs can be used in this case. However, a drug delivery system with limited enhanced permeability and retention effect may result in poor therapeutic e cacy due to ine cient drug delivery to the tumor. To avoid this disadvantage, we have developed a more potent delivery system, namely, the ASO-targeting LINC00036 coupled with a potent DOTAP, to inhibit LINC00036 expression in mice xenograft models (Fig.3). However, LINC00036 was targeted using this delivery system in mice xenograft models, which is a limitation of this study, further investigations on large animal or primate models are desirable.

Conclusions
In summary, we have identi ed LINC00036, a c-MYC inducible onco-lncRNA, acts an oncogene in human cancer and decreases the sensitivity of ge tinib through positive feedback loop of the c-MYC-LINC00036-EGFR axis (Fig.8, left pannel). Furthermore, the identi cation of RNA-RNA and RNA-protein dual regulation model sheds light on the lncRNA-mediated complex regulation of target genes. Finally, a new paradigm may be discovered, in which targeting of LINC00036, especially coupled with a potent DOTAP, represents a novel next generation anti-EGFR therapeutic strategy via improving the sensitivity of ge tinib (Fig.8