Long noncoding RNA LINC01088 inhibits esophageal squamous cell carcinoma progression by targeting the NPM1-HDM2-p53 axis

Abstract

Esophageal squamous cell carcinoma (ESCC) is characterized by extensive metastasis and poor prognosis. Long noncoding RNAs (lncRNAs) have been shown to play important roles in ESCC. However, the specific roles of lncRNAs in ESCC tumorigenesis and metastasis remain largely unknown. Here, we investigate LINC01088 in ESCC. Differentially expressed LINC01088 levels are screened from the GEO database. We find that LINC01088 is expressed at low level in collected clinical samples and is correlated with vascular tumor emboli and poor overall survival time of patients after surgery. LINC01088 inhibits not only ESCC cell migration and invasion in vitro, but also tumorigenesis and metastasis in vivo. Mechanistically, LINC01088 directly interacts with nucleophosmin (NPM1) and increases the expression of NPM1 in the nucleoplasm compared to that in the nucleolar region. LINC01088 decreases mutant p53 (mut-p53) expression and rescues the transcriptional activity of p53 by targeting the NPM1-HDM2-p53 axis. LINC01088 may also interfere with the DNA repair function of NPM1 by affecting its translocation. Our results highlight the potential of LINC01088 as a prognostic biomarker and therapeutic target of ESCC.

Introduction

Esophageal cancer (EC) is the eighth most commonly diagnosed cancer and the sixth leading cause of cancer death worldwide [1]. Moreover, esophageal squamous cell carcinoma (ESCC) is the predominant histological subtype in China, accounting for more than 90% of all EC cases [ 2, 3] . Despite remarkable progress in both diagnostic and therapeutic techniques over the past decades, ESCC still has a poor prognosis, and the 5-year survival rate is less than 25% [ 4, 5] . The main reasons for this high mortality are extensive metastasis and tumor recurrence. Although various risk factors have been identified, knowledge of the genetic drivers of ESCC is still limited [ 6– 8] . There is an urgent need to identify the molecular mechanisms underlying ESCC and to develop new and effective therapeutic approaches.

Over the past decades, aberrant expression of long noncoding RNAs (lncRNAs) has been reported to correlate with tumorigenesis and metastasis [ 9– 11] . It was reported that the lncRNA CASC9 promotes ESCC by recruiting EZH2 to the promoter of programmed cell death protein 4, and HEIH (a lncRNA) depletion slowed esophageal carcinoma cell progression by sponging microRNA-185 [ 12, 13] . More than 90% of the human genome is functional and encodes large numbers of noncoding RNAs (ncRNAs), and lncRNAs comprise over 90% of ncRNAs [ 14, 15] . Because of their genome-wide and tissue-specific expression patterns, lncRNAs are ideal novel biomarkers and therapeutic targets for cancer [ 11, 15] .

Zhang et al. [16] first reported that the expression of LINC01088 in ovarian epithelial cells (EOCs) is significantly downregulated using microarray analysis and that its expression is associated with the progression of EOC. In human non-small cell lung cancer, LINC01088 is a scaffold that recruits EZH2 to inhibit p21 expression [17]. LINC01088 physically interacts with small nuclear ribonucleoprotein polypeptide A (SNRPA) and promotes the growth and invasion of glioma cells by regulating SNRPA transcription [18]. Recently, it has also been reported that LINC01088 promotes the apoptosis of trophoblast cells by binding with Arginase-1 and activating the JNK and MAPK signaling pathways [19]. However, the role of LINC01088 in ESCC has not been reported, and its underlying molecular mechanisms remain to be elucidated.

In the present study, we screened and characterized LINC01088 as a new target of ESCC, which is associated with tumorigenesis and metastasis. The functions of LINC01088 in ESCC were explored in vitro and in vivo. The downstream pathways were further investigated to determine its role in ESCC progression.

Materials and Methods

Bioinformatics analysis

The datasets (GSE106185, GSE17351, GSE67508, GSE21293, and GSE42363) analyzed during the current study are available in the GEO repository. A computational pipeline was established to reannotate several million probes of the Affymetrix Human Genome U133 plus 2.0 array ( Supplementary Figure S1). To further increase accuracy, we adjusted a P value of less than 0.01 and log ₂ fold change of more than 1.5 as a threshold.

RNA-seq and GO enrichment analysis

According to the sequencing requirements of BGI (Beijing Genomics institution Company, Shenzhen, China), we prepared qualified stable cell line EC9706 that overexpressed LINC01088. The company was responsible for sequencing details, and the corresponding sequencing results were presented in the form of volcano map. GO enrichment were carried out for downregulated and upregulated genes using R package cluster Profiler. We used R package ggplot2 to visualize the important projects of each group function and pathway enrichment analysis.

Clinical specimens

Primary ESCC specimens ( n=134, containing matched adjacent normal tissues) were obtained from patients who underwent surgery at Peking University Cancer Hospital from 2012 to 2015. The following inclusion/exclusion criteria were applied for patients: (i) with pathological diagnosis of ESCC; (ii) with follow-up information; (iii) without preoperative chemotherapy; and (iv) without other major esophageal and tumor diseases. The detailed clinicopathological features of these patients are shown in Table 1. This study was approved by the Research and Ethical Committee of Peking University Cancer Hospital (No. 2017KT79), and it was conducted according to the principles expressed in the Declaration of Helsinki. Written informed consent was obtained from all subjects involved in the study before collection of specimens.

Table 1 Correlation between LINC01088 expression and clinicopa-thological parameters of patients with ESCC

Clinical parameters	Characteristics	Low (No)	High (No)	P value
Gender	Male	57	55	0.939
Female	11	11
Age (year)	≥9	35	27	0.375
<60	33	39
Section (position)	Up	22	21	0.377
Middle	34	33
Low	12	12
Living state	Survive	36	33	0.667
Demise	31	7
Differentiation degree	High	15	7	0.190
Middle	41	44
Low	12	15
TNM stage	I	6	9	0.576
II	22	18
III	37	38
IV	3	1
Lymphatic metastasis	Yes (N1,N2,N3)	41	36	0.501
No (N0)	27	30
Vascular tumor emboli	Yes	26	16	0.040
No	42	50

Cell lines and cell culture

Human ESCC cell lines (KYSE30, KYSE150, KYSE180, KYSE450, EC9706 and EC109) were purchased from the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences (Shanghai, China). In a humidified incubator, the cells were cultured in RPMI-1640 medium (Gibco, Grand Island, USA) at 37°C with 5% CO ₂. The culture medium was supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin (Gibco).

RNA interference and lentivirus shRNA vector for LINC01088 knockdown

Small interfering RNAs (siRNAs) targeting NPM1 were customized by Genepharma (Shanghai, China). The cells were transfected with siRNA for transient knockdown using Lipofectamine 2000 (Invitrogen, Carlsbad, USA). Three short hairpin RNAs (shRNAs) targeting LINC01088 were designed and cloned into the pLenti6 vector (Genepharma) with the U6 promoter for stable knockdown. The siRNA and shRNA sequences are listed in Table 2. The shuttle vectors containing the shRNAs were then transfected into HEK293T cells with packaging vectors using Lipofectamine 2000 (Invitrogen). The infectious lentiviruses were harvested 72 h after transfection and filtered through a 0.45-μm filter. KYSE150 or KYSE450 cells were infected with lentiviruses, followed by two weeks of selection with blasticidin (Thermo Fisher, Waltham, USA). The expression of LINC01088 in the infected cells was validated by real-time polymerase chain reaction (RT-qPCR).

Table 2 The sequences of shRNAs and siRNAs used in this study

Sample	Sequence
INC01088-shNC	Sense: 5′-GATCCACTACCGTTGTTATAGGTGCTCGAGCACCTATAACAACGGTAGTTTTTTG-3′
Antisense: 5′-AATTCAAAAAACTACCGTTGTTATAGGTGCTCGAGCACCTATAACAACGGTAGTG-3′
LINC01088-sh37	Sense: 5′-GATCCGGATCTCTCTCTCTCAGATTTCTCGAGAAATCTGTGAAGAGAGAGAGATCCTTTTTG-3′
Antisense: 5′-AATTCAAAAAGGATCTCTCTCTCTCAGATTTCTCGAGAAATCTGTGAAGAGAGAGAGATCCG-3′
LINC01088-sh101	Sense: 5′-GATCCGGTGAAGACCTGTGGTAAAGTCTCGAGACTTTACCACAGGTCTTCACCTTTTTG-3′
Antisense: 5′-AATTCAAAAAGGTGAAGACCTGTGGTAAAGTCTCGAGACTTTACCACAGGTCTTCACCG-3′
LINC01088-sh122	Sense: 5′-GATCCGCTGGCAGAGAGGAAGCTAAACTCGAGTTTAGCTTCCTCTCTGCCAGCTTTTTG-3′
Antisense: 5′-AATTCAAAAAGCTGGCAGAGAGGAAGCTAAACTCGAGTTTAGCTTCCTCTCTGCCAGCG-3′
LINC01088-sh164	Sense: 5′-GATCCGCTCAGCGTTTCACAGCTAAGCTCGAGCTTAGCTGTGAAACGCTGAGCTTTTTG-3′
Antisense: 5′-AATTCAAAAAGCTCAGCGTTTCACAGCTAAGCTCGAGCTTAGCTGTGAAACGCTGAGCG-3′
siNPM1-NC	Sense: 5′-UUCUCCGAACGUGUCACGUTT-3′
Antisense: 5′-ACGUGACACGUUCGGAGAATT-3′
siNPM1-1	Sense: 5′-GGAUGAGUUGCACAUUGUUTT-3′
Antisense: 5′-AACAAUGUGCAACUCAUCCTT-3′
siNPM1-2	Sense: 5′-GGAAGCCAAAUUCAUCAAUTT-3′
Antisense: 5′-AUUGAUGAAUUUGGCUUCCTT-3′
siNPM1-3	Sense: 5′-GGAAUGUUAUGAUAGGACATT-3′
Antisense: 5′-UGUCCUAUCAUAACAUUCCTT-3′

Overexpression vector construction

EC9706 cells were transfected with the pCDNA3.1-LINC01088 overexpression plasmids or empty vector (as a negative control) using Lipofectamine 2000 (Invitrogen), followed by two weeks of selection with G418 (Geneticin; Invitrogen) to establish the EC9706 cell line stably expressing LINC01088.

Cell viability assay

Tumor cells were seeded in 96-well plates at a density of 3×10 ³ cells/well. The number of viable cells was assessed by optical density examined at 450 nm using a CCK8 assay (Dojindo, Kumamoto, Japan) at 0, 12, 24 and 48 h according to the manufacturer’s instructions.

Colony formation assay

A total of 500 cells/well were plated in 6-well plates and cultured for 2 weeks. The number of cell colonies was counted after fixed with 4% formalin and stained with 1% crystal violet.

Wound healing assay

To image the same position every time, we marked the six-well plates with six straight lines in advance. Cells were seeded into 6-well plates and scratched with a 20 μL pipette tip. After washing three times with PBS, the culture medium was changed to RPMI 1640 medium without FBS to eliminate the effects of proliferation. Detached cells were removed carefully with PBS wash for three times and the wounded area was photographed at 0, 12, 24 and 48 h (Leica, Wetzlar, Germany).

Transwell assay

Transwell insert chamber plates were used for migration and invasion assays. Transfected cells (1–3×10 ⁵) in 200 μL of RPMI 1640 medium supplemented with 1% FBS was added to the upper chamber and incubated at 37°C for 20 h. The lower chamber was filled with 1 mL RPMI 1640 medium supplemented with 10% FBS, and the upper chamber was coated with 60 μL diluted Matrigel (1:10; Corning Inc., Corning, USA) for the invasion assay. Non-migratory and non-invaded cells in the upper chamber were scraped off using cotton swabs. The migratory and invading cells were fixed in 4% paraformaldehyde and stained with 0.1% crystal violet. Cells were then photographed under a microscope (100×, magnification) in three fields randomly, and then analyzed using Image J software (National Institutes of Health, Bethesda, USA).

CRISPR/Cas9-mediated NPM1 knockout

L29161 NPM1 knockout lentivirus was purchased from Beyotime Biotechnology (Shanghai, China). It can mediate the expression of pLenti-NPM1-sgRNA, which is a plasmid that can simultaneously express Cas9, sgRNA of NPM1, and puromycin resistance gene in animal cells. KYSE150 and EC9706 cells were infected with L29161 knockout lentivirus, followed by two weeks of selection with puromycin (MedChemExpress, New Jersey, USA). The knockout expression of NPM1 in the infected cells was validated by western blot analysis.

Reverse transcription and quantitative PCR

Total RNA was extracted from ESCC tissues and cell lines using Trizol reagent (Invitrogen) based on the manufacturer’s instructions and was reverse transcribed to cDNA with M-MuLV Reverse Transcriptase (BioLabs, San Diego, USA). According to the manufacturer’s instructions, RT-qPCR was performed with SYBR Green qPCR mix (Toyobo, Tokyo, Japan) on the 7500 Fast Real-time PCR system (Applied Biosystems, Foster City, USA). The 2 ^–△△Ct method was used for quantification, and the fold change of target genes was normalized to that of the internal control. The primer sequences are shown in Table 2.

Coimmunoprecipitation (Co-IP)

Cells were lysed using NP-40 lysis buffer (Beyotime, Shanghai, China) supplemented with 0.5% Triton X-100 (T8200; Solarbio Life Sciences, Beijing, China), RNase inhibitor (RNasin; Solarbio Life Sciences), a protease inhibitor, and phosphatase inhibitors (complete, EDTA-free; Roche Applied Science, Basel, Switzerland). The total protein concentration was determined with a BCA protein assay kit (Tiangen, Beijing, China). After normalization of the protein concentrations, lysates were immunoprecipitated with specific antibodies (anti-NPM1; ab10530; Abcam, Cambridge, UK; anti-HDM2, a mix of sc-965 and SMP-14; Santa Cruz Biotech, Santa Cruz, USA; anti-p53; ab32389; Abcam) at 4°C overnight. The immunocomplexes were collected using Protein G Sepharose beads (101241; Thermo Fisher) after 2 h of incubation. The cell lysates or immunocomplexes were separated by 10% or 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore, Bedford, USA). Immunoblotting was carried out using the above antibodies, followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies, after which the proteins were detected with enhanced chemiluminescence reagent (ECL; Pierce, Rockford, USA).

Western blot analysis

Protein samples from tissues or cells were acquired using RIPA lysis buffer (Biosharp, Hefei, China) and separated by 10% or 12% SDS-PAGE, followed by electrotransfer onto PVDF membranes. The membranes were blocked with 5% nonfat milk for 1 h, and then incubated with primary antibodies against NPM1, HDM2 (sc-965), ARF (ab185620; Abcam), p53, γH2AX (9718T; Cell Signaling Technology, Danvers, USA), p21 (12D1; Cell Signaling Technology) and GAPDH (ab8245; Abcam) overnight at 4°C. After extensive wash, membranes were incubated with the corresponding HRP-conjugated secondary antibodies. The protein bands were visualized using ECL substrate and signal intensities were measured by ImageJ software. GAPDH was used as the loading reference.

Subcellular fractionation assay

To determine the subcellular localization of LINC01088, the Cytoplasm & Nuclear RNA Purification kit (Norgen Biotek, Thorold, Canada) was used to separate the cytoplasmic and nuclear fractions according to the manufacturer’s protocol. After RNA extraction, RT-qPCR was performed using SYBR Green qPCR mix (Toyobo). GAPDH was used as an internal reference for the cytoplasm, and U6 was used as an internal reference for the nucleus. The sequences of the primers are shown in Table 3.

Table 3 Sequences of primers used in the study

Gene	Sequence
U6	Forward: 5′-CTCGCTTCGGCAGCACA-3′
Reverse: 5′-AACGCTTCACGAATTTGCGT-3′
LINC01088	Forward: 5′-TATTACCACAAGGCCGGCAA-3′
Reverse: 5′-GGGCTTAGCTGTAAGGACGAA-3′
GAPDH	Forward: 5′-GGAGCGAGATCCCTCCAAAAT-3′
Reverse: 5′-GGGCTTAGCTGTAAGGACGAA-3′
p53	Forward: 5′-CAGCACATGACGGAGGTTGT-3′
Reverse: 5′-TCATCCAAATACTCCACACGC-3′
p21	Forward: 5′-TTAGCAGCGGAACAAGGAGT-3′
Reverse: 5′-GCCGAGAGAAAACAGTCCAG-3′

RNA pull-down and liquid chromatography-mass spectrometry (LC-MS)

KYSE150 or HEK293T cells treated with pCDNA3.1- LINC01088-6xMS2bs or pCDNA3.1-6xMS2bs were cotransfected with pCDNA3.1-Flag-2xMS2. The cells were collected and washed twice with PBS 48 h after transfection. Then, the cells were lysed in NP40 lysis buffer supplemented with RNase inhibitor (EO0381; Solarbio Life Sciences), a protease inhibitor, and phosphatase inhibitors (complete, EDTA-free; Roche Applied Science) at 4°C for 30 min. Finally, the cell lysates were incubated overnight with Protein G Sepharose beads bound with anti-flag antibody (Merck, Darmstadt, Germany). After the beads were washed several times, the retrieved proteins were sent to the Core Facility of Molecular Biology of Peking University Health Science Center (Beijing, China) for mass spectrometry analysis. The candidate target protein was further confirmed by western blot analysis.

RNA immunoprecipitation assay

HEK293T and KYSE150 cells were treated with formaldehyde to cross-link in vivo protein-RNA complexes. Then, the cells were washed with cold PBS and lysed with RIP buffer at 4°C for 30 min. The cell lysates were incubated overnight with Protein G Sepharose beads (101241; Thermo Fisher) conjugated to antibodies against NPM1 (Ab10530; Abcam) or normal mouse immunoglobulin G (A7028; Beyotime). After the beads were washed several times, the RNA was purified, reverse transcribed into cDNA and detected by qPCR to quantify the binding targets.

Immunofluorescence microscopy

The cells were fixed with 4% paraformaldehyde in PBS (pH 7.4) for 15 min, followed by permeabilization with 0.5% Triton X-100 for 20 min. Then, the cells were incubated with diluted primary antibody (NPM1, freshly diluted at a ratio of 1:500), placed in a wet box, and incubated at 4°C overnight. After hybridization, m-IgGk BP-CFL 488 (sc-516176; Santa Cruz) were sequentially added to visualize the signal with an inverted microscope (Leica DMi8; Leica, Darmstadt, Germany) or Zeiss LSM780 confocal microscope (Zeiss, Darmstadt, Germany). The nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI).

Fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC)

A FISH assay was carried out using the LINC01088 fluorescence in situ hybridization kit (Boster Biological Technology, Wuhan, China) according to the supplier’s instructions. After being fixed with 4% formaldehyde and washed with PBS, the KYSE150, KYSE450, and EC9706 cells or frozen sections from ESCC tumor tissue were treated with pepsin and fixed again with 1% formaldehyde. Next, the slides were incubated in prehybridization buffer at 40°C for 2–4 h. Then, the prehybridization buffer was carefully washed off with 2× SSC, 0.5× SSC, and 2× SSC for 5 min. The cells were incubated with denatured digoxin (DIG)-labelled FISH probes in hybridization buffer (50% formamide in 2× SSC) overnight at 40°C. After hybridization, anti-DIG primary antibody and fluorescein (FITC) IgG Fraction Monoclonal Mouse Anti-Biotin (AB_2339020; Jackson ImmunoResearch Labs, UK) were sequentially added to visualize the signal with an inverted microscope (Leica DMi8; Leica) or Zeiss LSM780 confocal microscope (Zeiss). The nuclei were counterstained with DAPI. For the frozen tissue sections, anti-DIG antibody labelled with biotin, streptavidin-biotin complex (SABC), and peroxidase labelled with biotin (all from Boster Biological Technology) were sequentially added. The signals were then visualized with DAB substrate. Finally, the results were evaluated by two experienced pathologists from Peking University Cancer Hospital.

Ultraviolet irradiation

In the preexperiment, 1200 J/m ² was determined as the appropriate dose. Therefore DNA damage was induced by ultraviolet (UV) irradiation at a dose of 1200 J/m ² using a Hoefer UVC500 Ultraviolet Crosslinker (Hoefer, San Francisco, USA). The cells were harvested, washed, and lysed with RIPA 3 h after exposure to UV irradiation. DNA damage-related proteins, such as γH2AX, were detected by western blot analysis.

Flow cytometry

An Annexin V-FITC/PI staining experiment was conducted to evaluate the proportion of EC9706 and EC9706-Lnc cells undergoing apoptosis using the flow cytometric methods. Briefly, cells were seeded into a 6-well plate (1×10 ⁵ cells/mL) and then exposed to cisplatin (0, 4, 8, 16, and 32 μM) for 24 h. The harvested cells were centrifuged at 300 g for 3 min. Cells were suspended in binding buffer at 1×10 ⁶ cells/mL. Then, 5 μL Annexin V-FITC and 5 μL PI were added and incubated in the dark for 15 min. The cell apoptotic rate was analysed using a flow cytometer (Becton Dickinson, Franklin Lakes, USA). Each assay was repeated three times.

In vivo model

A chicken chorioallantoic membrane (CAM) model was established to engraft tumor cells and monitor tumor growth and metastasis [20]. The digested KYSE150 and EC9706 cells were suspended in RPMI-1640 medium at 1×10 ⁷ cells/mL and labelled with Cell Tracker CM-Dil (Invitrogen), which is nontoxic to cells and has prolonged stability. A hole of 3 mm in diameter was created on the eggshell with a sterile tweezer. After tumor cell engraftment, the hole was covered with a laboratory wrapping film to prevent dehydration and possible infections. The eggs were then returned to the incubator with a relative air humidity of 65% and a temperature of 37°C in a static position. After 10 days of incubation, the volume and weight of the tumor and the number of metastatic foci in the lungs in each group were evaluated.

Statistical analysis

Each experiment was repeated three times. All data are presented as the mean±standard deviation (SD). SPSS 25.0 (SPSS Inc., Chicago, USA) or GraphPad PRISM 7 (GraphPad, San Diego, USA) was used for statistical analysis. Kaplan-Meier analysis was used to estimate the overall survival (OS) of patients with ESCC. Differences between groups were analyzed using Student’s t test or one-way ANOVA. Differences were considered to be significant when P<0.05.

Results

LINC01088 predicts a poor clinical outcome for ESCC patients

To identify the differentially expressed lncRNAs between ESCC and adjacent tissue samples (GSE106185 and GSE17351), between invasive and noninvasive EC (GSE67508 and GSE21293), and between N-positive and negative tissue samples (GSE42363) [ 21– 24] , a computational pipeline was established to reannotate several million probes of the Affymetrix Human Genome U133 plus 2.0 array ( Supplementary Figure S1). To further increase accuracy, we adjusted a P value of less than 0.01 and log ₂ fold change of more than 1.5 as a threshold. Then, by intersecting the three groups, 9 downregulated lncRNAs (FLG-AS1, EIF3J-AS1, NEAT1 SPRR2C, TTTY14, SH3PXD2A-AS1, LINC01088, LOC440434 and C5orf66-AS1) and 4 upregulated lncRNAs (LOC100507053, LINC00622, MIR18A2HG and LOC642852) were identified ( Figure 1A). The identified lncRNAs were confirmed to be down- or upregulated in ESCC cancer tissues by the GEPIA database. Compared with 286 normal esophageal tissues, the expression of 6 lncRNAs was decreased in 182 esophageal tumor tissues ( Figure 1B). Considering the relationship between lncRNA expression and overall survival (OS) or disease-free survival (DFS) in GEPIA, 5 lncRNAs (LINC01088, C5orf66-AS1, SPRR2C, TTTY14 and SH3PXD2A-AS1) were selected as candidates ( Supplementary Figure S2).

Figure 1 .

Increased LINC01088 expression is related to a favorable prognosis in ESCC patients

(A) A Venn diagram of three independent lncRNA differential analyses. Blue: between tumor and adjacent tissue samples; Green: between invasive and noninvasive EC samples; Red: between N-positive and N-negative EC samples. (B) The expression levels of 6 lncRNAs in 286 normal tissues and 182 ESCC tumor tissues based on data from the GEPIA database. (C) The expression levels of LINC01088 in a cohort of 134 ESCC tumor tissues and paired adjacent tissues collected by Peking University Cancer Hospital. (D) Kaplan-Meier curves of the LINC01088 high-expression and low-expression groups for overall survival time (OS). (E) The patients with vascular tumor emboli in the low-expression group had a significant reduction in OS time. Data are shown as the mean±SD. * P<0.05, ** P<0.01.

Using our cohort of 134 clinical cases collected by Peking University Cancer Hospital, we found that compared with normal adjacent tissues ESCC tissues showed downregulated expression of LINC01088 ( P=0.0023; Figure 1C). The expression of SH3PXD2A-AS1 was upregulated, which was not consistent with the above results. We also detected the expression of LINC01088 by immunohistochemistry, and the results suggested that LINC01088 localizes to the nucleus ( Supplementary Figure S3A). To facilitate analysis, the median expression level was used as a cut-off point to classify 134 patients into the high-expression ( n=67) and low-expression groups ( n=67). Kaplan-Meier analysis was used to evaluate the predicted value of LINC01088 for survival. Compared with the patients in the high-expression group, the patients in the low-expression group showed not only a shorter survival time ( P=0.0374; Figure 1D) but also a higher rate of vascular tumor emboli incidence ( P=0.040; Table 1). The patients with vascular tumor emboli in the low-expression group had a significant reduction in overall survival time ( Figure 1E).

LINC01088 suppresses the malignant phenotype of ESCC cells

The KYSE150 and KYSE450 cell lines have a relatively higher expression of LINC01088, so these cell lines were utilized for knockdown experiments. The EC9706 and KYSE180 cell lines were utilized for the overexpression experiments ( Supplementary Figure S2B). Stable EC9706 and KYSE180 cell lines overexpressing LINC01088 and stable KYSE150 and KYSE450 cell lines in which LINC01088 was knocked down were constructed and validated ( Figure 2A and Supplementary Figure S2C). The CCK-8 assay showed that the proliferation ability was decreased in the EC9706 and KYSE180 cell lines overexpressing LINC01088, while the proliferation ability was increased in the KYSE150 and KYSE450 cell lines with LINC01088 knockdown ( Figure 2B and Supplementary Figure S2D). Furthermore, LINC01088 overexpression markedly decreased colony formation, while knockdown of LINC01088 increased colony formation ( Figure 2C and Supplementary Figure S2E,F). Next, we examined the effect of LINC01088 on cell migration and invasion via wound healing and Transwell assays. Knockdown of LINC01088 led to a significant increase in the migration and invasion ability of the KYSE150 and KYSE450 cell lines, whereas LINC01088 overexpression inhibited the migration and invasion ability of the EC9706 and KYSE180 cell lines ( Figure 2D,E and Supplementary Figure S2G,H). We also established the chicken chorioallantois membrane model (CAM), which displays tumor growth and spontaneous metastatic events. The digested tumor cells were labelled with Cell Tracker CM-Dil. The results showed that the tumor volume and weight and the number of metastatic foci in the lungs in the knockdown group were increased, and the LINC01088 overexpression group showed the opposite results ( Figure 2F,G). Collectively, these results indicate that LINC01088 can negatively affect ESCC cell proliferation, migration and invasion both in vivo and in vitro.

Figure 2 .

LINC01088 expression negatively affects the ESCC cell malignant phenotype

(A) The stable cell lines overexpressing LINC01088 and those with knockdown of LINC01088 were validated by RT-qPCR analysis, and sh37 and sh164 were the effective interference targets. (B) CCK8 assays showed that the proliferation ability was inhibited by LINC01088. (C) Colony formation assays showed that the colony formation ability was negatively correlated with LINC01088 expression. (D) Wound healing assay showed LINC01088 inhibited the migration ability. (E) Transwell assays showed that LINC01088 inhibited cell migration and invasion abilities. (F,G) In the CAM model, tumor volume and weight and metastatic foci were decreased after LINC01088 overexpression. Data are shown as the mean±SD. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

NPM1 is a target of LINC01088

Next, we explored the underlying antitumor mechanisms of LINC01088. According to the cellular fractionation results, LINC01088 was primarily found in the nucleus of ESCC cells ( Figure 3A). This result was consistent with the immunohistochemical results ( Supplementary Figure S2A) and FISH assay results in which LINC01088 was labelled with fluorescence ( Figure 3B). LINC01088 may regulate gene transcription and translation by directly binding to proteins, transcription factors or DNA [25]. To determine the potential interaction targets, RNA pull-down and LC-MS were performed ( Figure 3C,D). HSPD1, nucleophosmin (NPM1) and PCNA were among the top candidates. Next, the existence of NPM1 was validated by western blot analysis ( Figure 3E), and the results for other candidate proteins were negative. We consistently verified this interaction via RIP assays and observed the enrichment of LINC01088 in the NPM1 group ( Figure 3F). Finally, we identified the interaction in vivo. We also applied RNA-seq to the stably overexpressed cell lines ( Supplementary Figure S4A), and found that the expression of NPM1 did not change significantly ( Supplementary Figure S4B). This finding indicates that LINC01088 may not change the expression or protein stability of NPM1. The GO enrichment analysis results confirmed that the transcription factor and transcriptional activator activity were closely related pathways ( Supplementary Figure S3C). The expressions of oncogenes, such as c-Myc, TERT and HNF1A, were reduced, further confirming the antitumor effect of LINC01088 ( Supplementary Figure S3B). HNF1A and c-Myc expression levels were verified by western blot analysis (data not shown).

Figure 3 .

NPM1 is a target of LINC01088

(A) Intracellular localization of LINC01088 was assessed by RT-qPCR. U6 and GAPDH were the preference controls for nuclear RNA and cytoplasmic RNA, respectively. (B) FISH assay showed that LINC01088 was mainly expressed in the nucleus. DAPI staining shows the nucleus. Scale bar: 5 μm. (C) 293T cell lysates were incubated overnight with protein G Sepharose binding of anti-flag antibody for RNA pull-down assay and mass spectrometry (MS). (D) The top candidates obtained by LC-MS analysis are listed. (E) 293T cell lysates were immunoprecipitated with anti‐NPM1 antibody or IgG control. (F) RIP-PCR was applied to determine the amount of LINC01088 associated with NPM1 immunoprecipitation relative to the IgG control. Data are shown as the mean±SD. * P<0.05.

The inhibition of ESCC progression by LINC01088 is dependent on NPM1

To determine whether the inhibition of ESCC progression by LINC01088 is dependent on NPM1, we knocked out NPM1 in EC9706 cells overexpressing LINC01088 (EC9706-Lnc). CRISPR/Cas9-mediated NPM1 knockout led to rapid ESCC death, and we speculated that the normal physiological function of NPM1 was disrupted. Two different siRNAs were confirmed to knockdown NPM1 expression in EC9706 cells by RT-qPCR analysis ( Figure 4A). We found that the viability of EC9706-Lnc cells was increased ( Figure 4B) when NPM1 was downregulated. Similar results were also observed in the cytofunctional experiments. Knockdown of NPM1 promoted migration and invasion of LINC01088-overexpressing EC9706 cells ( Figure 4C,D). All these results indicate that the ability of LINC01088 to inhibit ESCC proliferation, invasion and metastasis is dependent on NPM1.

Figure 4 .

Inhibition of ESCC progression by LINC01088 is dependent on NPM1

(A) Knockdown of NPM1 in EC9706 cells overexpressing LINC01088 (EC-Lnc) was confirmed by RT-qPCR analysis. (B) CCK8 assays showed that the viability of EC-Lnc cells was increased after NPM1 knockdown. (C,D) Wound healing and Transwell assays showed that the ability of LINC01088 to inhibit ESCC migration and invasion was attenuated when NPM1 was knocked down. Scale bar: 500 μm in C and 200 μm in D. Data are shown as the mean±SD. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

LINC01088 regulates mut-p53 by influencing the translocation of NPM1 and its interaction with HDM2

NPM1 is a widely expressed nucleolar phosphoprotein that constantly shuttles between the nucleus and cytoplasm [ 26, 27] . Dysregulation of NPM1 expression and localization contributes to tumorigenesis through different mechanisms [ 26, 28– 30] . The GEPIA database indicated that there is no significant correlation between NPM1 expression and LINC01088 expression (data not shown). Therefore, we suspected that LINC01088 may affect the localization of NPM1. Confocal laser scanning microscopy results showed that the expression ratio of NPM1 in the cytoplasm and nucleoplasm versus the nucleolus in ESCC was changed with LINC01088 overexpression ( Figure 5A). Compared with that in the EC9706-Ctrl group, the number of positive cells with a ratio change was significantly increased in the EC9706-Lnc group ( P=0.008; Figure 5A).Based on data from the STRING database ( https://string-db.org/.), a highly interconnected protein network, NPM1-HDM2-p53, was identified ( Figure 5B). NPM1 is closely associated with tumor protein p53 (TP53 or p53) [ 25, 29, 31] . It directly interacts with p53 and regulates the increase in stability and transcription activity. The ubiquitin-proteasome system plays a role in ESCC, and ubiquitin protein ligase can regulate a number of tumor-associated proteins via ubiquitination. NPM1 could influence p53 degradation by interacting with the human double minute 2 protein (HDM2) [ 29, 32, 33] . NPM1 also has an inhibitory effect on the expression of HDM2 to regulate mut-p53 stability in tumors [ 31, 34] .

Figure 5 .

LINC01088 regulates mut-p53 by influencing the translocation of NPM1 and its interaction with HDM2

(A) Immunofluorescence results showed that the expression ratio of NPM1 in the nucleoplasm versus the nucleolus was changed, and the number of cells with a positive change in the EC9706- LINC01088 group was significantly increased compared with that in the EC9706-Ctrl group. Scale bar: 5 μm. (B) The STRING database showed a highly interconnected protein network of NPM1-HDM2-p53. (C) The Co-IP assay revealed that NPM1 interacted with HDM2, and LINC01088 inhibited NPM1-HDM2 complex formation. (D) LINC01088 overexpression rescued the transcriptional activity of mut-p53, and p21 expression was increased at the mRNA level. (E) LINC01088 overexpression inhibited mut-p53 expression and promoted p21 expression at the protein level. Data are shown as the mean±SD. * P<0.05, ** P<0.01, *** P<0.001.

TP53 is by far the most frequently mutated gene in ESCC [ 35, 36] , and over 80% of mutant p53 proteins (mut-p53) have acquired gain-of-function (GOF) activities [ 37, 38] .These activities enable them to both inactivate wild-type p53 (wt-p53) and promote tumor progression. The ESCC cell lines in this study are all p53-mutant cell lines [ 39– 42] . Under stressful conditions, transient expression of the cyclin-dependent kinase inhibitor (CKI) p21 is promoted in a p53-dependent manner [43]. However, the normal function of the p53 pathway is ubiquitously lost in cancers either through mutation or inactivating interactions [44].

We speculated that LINC01088 may affect mut-p53 expression by affecting the binding of NPM1 to p53, HDM2 or ARF. The Co-IP assay revealed that NPM1 interacted with HDM2, but p53 and ARF were not detected. Furthermore, the amount of HDM2 protein bound to NPM1 was decreased. These findings indicate that LINC01088 disrupted NPM1-HDM2 complex formation in the EC9706 cell line and increased free HDM2 in the cytoplasm and nucleoplasm ( Figure 5C). The amount of NPM1 protein bound to HDM2 was also decreased in the presence of LINC01088. RT-qPCR and western blot analysis results showed that mut-p53 was decreased and p21 was increased at both the protein and mRNA levels in the presence of LINC01088. These findings suggest that the normal transcription activity of p53 was rescued ( Figure 5D,E), and knockdown of LINC01088 increased mut-p53 expression at the protein level ( Figure 5E).

We also tried to predict the potential binding sites of LINC01088 in the promoter region of p53 using the JASPAR 2020 database but were not successful. It is unlikely that LINC01088 directly regulates mut-p53 expression. Based on the above results, it is highly possible that LINC01088 regulates the expression of mut-p53 by affecting the binding of NPM1 to HDM2. Increased level of free HDM2 protein in the nucleoplasm leads to the degradation of mut-p53 protein.

LINC01088 may disrupt the DNA repair function of NPM1

NPM1 acts as a molecular chaperone during DNA damage repair. It not only plays a significant role in the base excision DNA repair (BER) of DNA lesions [33] but also affects translesion DNA synthesis (TLS) [39]. NPM1 SUMOylation is indispensable for the assembly of the BRCA1 complex at the early stage of the DNA-damage response (DDR) [45]. Its expression in the nucleolus is also essential for the caspase-2 activation cascade in response to a variety of forms of DNA damage [ 46, 47] . Considering the distribution change and competitive association, we explored whether LINC01088 disturbs the DSB repair function of NPM1. γH2AX (phospho-H2A. X Variant Histone) is a sensitive marker for DSB after radiation or chemotherapy [48]. As DSBs are repaired, γH2AX foci disappear. Under UV irradiation, γH2AX expression was decreased after LINC01088 knockdown ( Figure 6A), and this effect was rescued by concomitant knockdown of NPM1 ( Figure 6B). These findings indicate its dependence on NPM1-mediated DNA repair. Additionally, LINC01088 overexpression resulted in increased γH2AX expression ( Figure 6C), but there was no difference between the control and overexpression groups in the absence of ultraviolet stimulation. Cisplatin is an antineoplastic chemotherapy agent that works by cross-linking with DNA and causing DNA damage in cancer cells [49]. The half-maximal inhibitory concentration (IC ₅₀) values were measured to assess the sensitivity to cisplatin. The IC ₅₀ of the EC9706-LINC01088 group (log ₂IC ₅₀=7.911±1.128 μM) was significantly lower than that of the EC9706-Ctrl group (log ₂IC ₅₀ =21.87±2.98 μM) ( Figure 6D). To evaluate the effects of LINC01088 on the cisplatin-induced apoptosis of EC9706 or EC9706-Lnc cells, we performed flow cytometric analysis with Annexin V-FITC/PI double staining ( Figure 6E). The proportion of early apoptosis (Annexin V-FITC ⁺/PI ^–) and late apoptosis (Annexin V-FITC ⁺/PI ⁺) in the EC9706 or EC9706-Lnc group was 1.78%±5.60% and 0.72%±5.64%, respectively. After the cells were exposed to 8, 16 or 32 μM cisplatin, the percentages of early and late apoptosis were increased to 3.82%±26.40%, 8.85%±28.30% and 8.85%±50.84%, respectively, which were significantly higher than those of the EC9706 group (1.38%±10.44%, 2.28%±11.33% and 7.40%±14.14%; P<0.01). Our results indicated that LINC01088 increased the sensitivity of EC9706 cells to chemotherapeutic drugs and radiotherapy. The antitumor effects of LINC01088 on ESCC may be related to the impaired DNA repair function of NPM1, at least partially.

Figure 6 .

LINC01088 may disturb the DNA repair function of NPM1

(A) Knockdown of LINC01088 decreased the DNA damage induced by ultraviolet irradiation, and (B) silencing of NPM1 by siRNA rescued the result. (C) Overexpression of LINC01088 enhanced the DNA DSBs induced by ultraviolet irradiation. (D) The IC 50 value of EC9706 to cisplatin was decreased in the presence of LINC01088. (E) Effects of LINC01088 on the cisplatin-induced apoptosis of EC9706 and EC9706-Lnc cells were evaluated by flow cytometry with Annexin V-FITC/PI dual staining. Data are shown as the mean±SD. ** P<0.01, *** P<0.001.

Discussion

A few lncRNAs have been demonstrated to have clinical application value in ESCC [ 50– 54] . The priority for future research in this area should focus on the use of lncRNAs as molecular targets in clinical diagnosis and treatment. The aim of this study was to explore differentially expressed lncRNAs associated with ESCC invasion and metastasis. We identified a novel endogenous lncRNA, LINC01088 that was downregulated in ESCC. This study showed that LINC01088 negatively affected ESCC cell proliferation, migration and invasion abilities in vitro and in vivo. Previous findings provided multiple molecular mechanisms through which LINC01088 executes protumor or antitumor function. LINC01088 could bind to different microRNAs or RNA-binding proteins. These results suggested that LINC01088 regulated cancer progression in a cell type-dependent manner and could play contradictory roles among different cancers.Here, RNA pull-down and RIP analyses suggested that LINC01088 physically interacts with NPM1. The role of NPM1 in hematological tumors has received extensive attention, but there has been little focus on NPM1 in solid tumors. Unlike the abnormal mutations in hematological tumors [ 55, 56] , NPM1 is rarely mutated and is often overexpressed in solid tumors of diverse histological origins [ 57– 59] . Under stress and oncogenic stimuli, increased NPM1 level promotes proliferation and inhibits apoptosis. In this study, we provided experimental evidence that NPM1 plays a role in ESCC. The inhibition of ESCC progression by LINC01088 is dependent on NPM1. In the rescue assay, after NPM1 was knocked down in LINC01088-overexpressing cells, the proliferation, migration and invasion trends were reversed. However, knockdown of NPM1 only slightly affects the proliferation, migration and invasion of ESCC. When LINC01088 is overexpressed, the inhibitory effect on ESCC is more significant. This phenomenon may be related to the dynamic changes in the distribution of NPM1 between the nucleolus and nucleoplasm. In another study, the lncRNA LETN was concentrated in the nucleolus by directly binding to NPM1, which contributed to the formation of NPM1 oligomers [60]. The oligomerization of NPM1 has been linked to its nucleolar localization and role in cell proliferation. These findings indicate that LINC01088 promotes the nucleoplasm location of NPM1 and disturbs its proliferative capacity.

Our work also showed that LINC01088 disturbed the interaction between NPM1 and HDM2. In recent years, a large number of studies have shown that the protein-nucleic acid complex binds to and inhibits the E3 ubiquitin ligase activity of the oncogene MDM2, which ubiquitinates p53 and regulates the level of intracellular p53 protein [ 61– 63] . It has been reported that the lncRNA SAMD12-AS1 interacts with NPM1 by promoting p53 degradation, and its association with HDM2 is impaired [64]. When LINC01088 is overexpressed, the interaction between NPM1 and HDM2 is disrupted, and mut-p53 protein level is decreased. It is highly possible that an increase in the amount of free HDM2 protein in the nucleoplasm leads to the degradation of mut-p53 protein. NPM1 also regulates apoptosis by inhibiting caspases and endonucleases or regulating tumor suppressors, including ARF and p53 [ 26, 65, 66] . It is highly likely that the mechanism by which NPM1 binds with mut-p53 to promote its stability is weakened.

TP53 mutation is significantly correlated with various biological behaviors of ESCC, such as differentiation degree, tumor invasion, and lymph node metastasis [ 67– 70] . Detection of p53 protein mutation and expression in ESCC is helpful to determine the degree of malignancy and prognosis of patients. Increasing evidence indicates that downregulation of mut-p53 expression in cancer cells reduces their aggressive properties [ 71, 72] , restores wt-p53 activity and increases the transcripts of p53 downstream target genes [71]. To date, several lncRNAs, such as lincRNA-p21, PANDA, PINT, NEAT1, GUARDIN, and DINO, have been reported to function as downstream effectors of p53 [ 73– 81] , whereas more lncRNAs with p53 regulatory functions remain to be discovered. Our work revealed that LINC01088 is a novel lncRNA involved in regulating mut-p53 expression via the NPM1-HDM2 pathway. LINC01088 inhibits ESCC proliferation, invasion and metastasis by targeting the NPM1-HDM2-p53 axis. The interaction between HDM2 and mut-p53 needs to be explored and ubiquitination experiments need to be performed, and whether different p53 mutation types affect the antitumor function of LINC00188 remains unclear and needs further exploration.

Aberrantly localized NPM1 may also promote tumor progression by participating in DNA damage repair or rRNA synthesis. NPM1 plays vital roles in centrosome duplication, genome stability and responses to various DNA damaging or cytotoxic agents [ 30, 82, 83] . Previous studies have reported that DNA damage induces a switch in the aberrant cytoplasmic localization of NPM1 mutants back to a predominantly nucleolar localization [83], and the chemotherapeutic oxaliplatin causes cell death via the inhibition of rRNA synthesis accompanied by NPM1 relocalization [30]. Cytoplasmic-localized NPM1 mutants play a dramatic role in the development of AML [ 84– 86] . The DNA repair process involving NPM1 occurs in the nucleoplasm. This study indicated that the DNA repair function of NPM1 was disrupted by nuclear-retained LINC01088. The sensitivity of EC9706 cells to chemotherapeutic drugs or radiotherapy was increased. NPM1 in the nucleoplasm may be a potential target of ESCC. Combined with the effect of mut-p53 on chemotherapy resistance and the DNA damage repair function of NPM1, LINC01088 may be related to the efficacy of chemotherapy drugs on ESCC, which needs further study.

Based on the above results, we proposed a model depicting the critical role of LINC01088 in ESCC ( Figure 7). To the best of our knowledge, our results are the first to prove that LINC01088 suppresses ESCC cell tumorigenesis and metastasis both in vitro and in vivo. LINC01088 impairs the interaction between NPM1 and mut-p53, which is the downstream target. LINC01088 is also involved in disrupting the DNA repair function of NPM1. The exact mechanism of action involved in this process needs further research. Previous studies on NPM1 have predominantly focused on hematological tumors, and the role of NPM1 in solid tumors has only recently been recognized. Our study indicates that NPM1 functions in ESCC and highlights the potential of LINC01088 as a prognostic biomarker and therapeutic target for ESCC.

Figure 7 .

A schematic illustration of the proposed model depicting the role of LINC01088 in inhibiting ESCC progression

Supplementary Data

Supplementary data is available at Acta Biochimica et Biophysica Sinica online.

COMPETING INTERESTS

The authors declare that they have no conflict of interest.

Funding Statement

This work was supported by the grants from the National Natural Science Foundation of China (Nos. 81773144 and 81772632), Capital Health Development Scientific Research Project (No. SF2018-2-2155) and Science Foundation of Peking University Cancer Hospital (No. A002208). The APC was funded by Peking University Cancer Hospital.