HCC cell lines (Huh7, Hep3B, SNU-182, and SK-HEP-1) and normal epithelial cell line (THLE-3) were involved in this study. Huh7 and Hep3B were provided by Yaji Biotechnology Co., Ltd. (Shanghai, China) and incubated in RPMI-1640 (A4192301, Gibco, Rockville, MD, United States) and minimum essential medium (MEM; A4192101, Gibco), separately. THLE-3, SNU-182, and SK-HEP-1 were procured from American Type Culture Collection (ATCC; Manassas, VA, United States) and cultured in BEGM (CC-3170, Lonza, Basel, Switzerland), RPMI-1640 and Eagle’s Minimum Essential Medium (EMEM; 11095080, Gibco) separately. All media contained 10% fetal bovine serum (FBS; 16140071, Thermo Fisher Scientific, Rockford, IL, United States) and were maintained under a moist environment with 5% CO₂ at 37°C.

Total RNA extraction was realized using TRIzol Reagent (15596018, Invitrogen, Carlsbad CA, United States). Moreover, PrimeScript™ II Reverse Transcriptase (2690A; TaKaRa, Shiga, Japan) was utilized to obtain cDNA of LINC01134 and MAPK1, following the standard method. The One Step miR cDNA Synthesis Kit (D1801, HaiGene, Harbin, China) was applied to synthesize cDNA of miRNAs, based on the user manual. Quantitative analysis of LINC01134 and MAPK1 was done with SYBR Green PCR Kit (4309155, Applied Biosystems, Foster city, CA, United States) on ABI Prism 7900HT sequence detector (PRISM^® 7900HT, Applied Biosystems). For miRNA quantification, HG miRNA SYBR Green PCR Kit (ZY-61500, HaiGene, Harbin, China) was used. Results were analyzed using the 2^−ΔΔCt method and standardized to GAPDH or U6. The sequences of primers involved in these assays were listed as follows: LINC01134 (F: TTG​GAC​CAT​GTC​AGT​GAC​GG; R: CAG​AGC​CAG​GTA​GGG​TGT​TG), MAPK1 (F: TCC​TTT​GAG​CCG​TTT​GGA​GG; R: GGT​CAG​CAG​GGC​ATC​ATG​TA), IGF2BP2 (F: GGA​ACA​AGT​CAA​CAC​AGA​CAC​A; R: CGC​AGC​GGG​AAA​TCA​ATC​TG), GAPDH (F: GAC​AGT​CAG​CCG​CAT​CTT​CT; R: GCG​CCC​AAT​ACG​ACC​AAA​TC), U6 (F: TCC​CTT​CGG​GGA​CAT​CCG, R: AAT​TTT​GGA​CCA​TTT​CTC​GAT​TTG​T), miR-140-3p (tac​cac​agg​gta​gaa​cca​cgg), miR-676-3p (ctg​tcc​taa​ggt​tgt​tga​gtt), miR-708-5p (aag​gag​ctt​aca​atc​tag​ctg​gg), miR-28-5p (aag​gag​ctc​aca​gtc​tat​tga​g), miR-1271-5p (ctt​ggc​acc​tag​caa​gca​ctc​a), miR-342-3p (tct​cac​aca​gaa​atc​gca​ccc​gt), miR-2355-3p (att​gtc​ctt​gct​gtt​tgg​aga​t) and miR-6512-3p (ttc​cag​ccc​ttc​taa​tgg​tag​g). The reverse sequence for each miRNA used was “CTC​AAC​TGG​TGT​CGT​GGA”.

The specific short hairpin RNAs (sh-RNAs) to LINC01134 or IGF2BP2 were synthesized by GenePharma (Shanghai, China), and non-specific shRNAs worked as negative control (NC). The pcDNA3.1 vectors (GenePharma) were inserted with MAPK1, and the empty pcDNA3.1 vectors were procured for gene overexpression. MiR-342-3p mimics/inhibitor along with NC mimics/inhibitor was designed by Ribobio (Guangzhou, China). Cell transfection was undertaken for 48 h with Lipofectamine 3000 (L3000075, Invitrogen), as instructed by the manufacturer. Sequences of sh-RNAs and sh-NC used herein were listed as follows: sh-NC (for LINC01134) (5’-CCG​GAA​AGA​TCA​GCA​AAC​ACT​CCG​ACT​CGA​GTC​GGA​GTG​TTT​GCT​GAT​CTT​TTT​TTT​G-3’), sh-LINC01134#1 (5’-CAC​CAT​ACA​ATT​TTA​CTT​TCA​GGC​CCT​CGA​GGG​CCT​GAA​AGT​AAA​ATT​GTA-3’), sh-LINC01134#2 (5’-CAC​CAC​TTC​AAG​TGG​TTT​CTA​GCT​CCT​CGA​GGA​GCT​AGA​AAC​CAC​TTG​AAG-3’), sh-LINC01134#3 (5’-CAC​CGT​GCA​TTT​GCT​GTT​CAT​GTC​CAC​TCG​AGT​GGA​CAT​GAA​CAG​CAA​ATG​CA-3’), sh-NC (for IGF2BP2) (5’-CCG​GTT​CTC​TAA​TTA​TCT​CAG​CAC​ACC​TCG​AGG​TGT​GCT​GAG​ATA​ATT​AGA​GAA​CCG​GTT​TTT​G-3’), sh-IGF2BP2#1 (5’-CAC​CAT​GCA​ATT​CCA​CTT​TAC​CCG​ACT​CGA​GTC​GGG​TAA​AGT​GGA​ATT​GCA-3’) and sh-IGF2BP2#2 (5’-CAC​CAT​GAT​TTC​AAG​AAT​CAT​GCG​GCT​CGA​GCC​GCA​TGA​TTC​TTG​AAA​TCA-3’).

CCK-8 assay was conducted with the use of the CCK-8 Kit. At first, HCC cells were plated in 96-well plates. After cells were cultured for 24, 48, 72, or 96 h, 10 μL of CCK-8 reagent was added into the small wells. Cell viability was measured based on the optical density (OD) value at 450 nm. In the end, a microplate reader was used for OD value analysis.

After transfection, SNU-182 and SK-HEP-1 cells were cultured in six-well plates with about 600 cells in each well. After 14 days, colonies were rinsed by PBS and then fixed in methanol for staining with 0.5% crystal violet (V5265, Sigma-Aldrich, St. Louis, MO, United States) solution. Visible clones were manually counted. The experiment was independently performed in triplicate.

Cells were inoculated into six-well plates (600 cells/well) and treated with 0, 2, 4, 6, and 8 Gy of irradiation. Next, the irradiated cells were cultured in a complete medium. Two weeks later, surviving colonies were counted manually after being stained with 0.1% crystal violet. The experiment was independently carried out in triplicate.

HCC cells fixed by 4% paraformaldehyde (PFA) were treated with TUNEL reagent (12156792910, Roche, Basel, Switzerland) that contained TdT and TMR-dUTP. After cells were stained with DAPI (D9542, Sigma-Aldrich), optical microscopy (DMi1, Leica, Wetzlar, Germany) was used to analyze TUNEL positive cells. The experiment was independently implemented in triplicate.

Immunofluorescence staining was conducted to examine the formation of gamma histone 2AX (γ-H2AX) and 53BP1 foci. In short, cells were seeded in 24-well plates to receive 0 and 4 Gy of irradiation. After 12, 24, or 36 h, the cells were subjected to 4% PFA fixing and 0.1% Triton X-100 (Sigma) permeabilization. Afterwards, the cells were blocked with 1% goat serum (16210072, Gibco), followed by incubation with primary anti-γH2AX (1:100; Abcam) and anti-53BP1 (1:100; Abcam). Then, the cells were co-cultured with secondary antibodies combining with fluorescein isothiocyanate after detached primary antibodies were washed off. Finally, nuclei were counterstained with DAPI. The percentage of γ-H2AX or 53BP1-positive cells was analyzed by a fluorescent microscope. The quantification of foci was achieved by Image J software. The experiment was independently conducted in triplicate.

DNA damage was assessed by comet assay. The neutral comet assays were conducted with the help of a Reagent Kit for Single Cell Gel Electrophoresis Assay (4250-050-K, Trevigen, United States). The user manual was strictly followed. DNA was stained by 100 μL propidium iodide (PI; 2 μg/ml, HY-D0815, MedChemExpress, NJ, United States) per slide for 30 min in darkness. Photos were captured by the Olympus BX61 capture system (magnification, 20×, Olympus, Tokyo, Japan). Finally, 50–100 cells were analyzed in each group using CaspLab software (CASP 1.2.3 beta 1). This experiment was independently conducted in triplicate.

Subcellular fractionation assay was operated based on the guidance of PARIS™ Kit (AM1921, Invitrogen, Carlsbad, CA, United States). Cell fractionation buffer was used to separate cell cytoplasm. Then, a cell disruption buffer was utilized to acquire cell nuclei. LINC01134 level in two parts was examined by RT-qPCR. U6 and GAPDH were considered as nuclear control or cytoplasmic control, respectively. The experiment was independently performed in triplicate.

The two assays were utilized to localize LINC01134 and IGF2BP2 in SNU-182 and SK-HEP-1 cells. At first, 4% PFA was utilized for 15 min fixation at 37°C. After being permeabilized with 0.5% Triton X-100 (R00285, Leagene, Beijing, China), HCC cells were co-cultured with LINC01134 FISH probe in hybridization buffer and then stained with DAPI.

For the IF assay, IGF2BP2 primary antibody was added to incubate with cells at 4°C for a whole night. Then, FITC-conjugated secondary antibody (7076, Cell Signaling Technology, Boston, MA, United States) was added. With the help of a confocal laser microscope (Axio-Imager_LSM-800, Zeiss, Oberkochen, Germany), images were gained. The experiment was independently conducted in triplicate.

RNA pull-down assay was completed with the application of Pierce Magnetic RNA-Protein Pull-Down Kit (20164, Thermo Fisher Scientific, Rockford, IL, United States) in light of the provided instruction. SNU-182 and SK-HEP-1 cells were lysed and then cultivated with a biotinylated (Bio)-LINC01134 probe. Bio-NC acted as a control. After magnetic beads (HY-K0208, MedChemExpress, NJ, United States) were added, the precipitated complexes were collected and purified. Finally, relative RNA or protein enrichment was assessed by RT-qPCR or western blot. The experiment was independently performed in triplicate.

The fragments of LINC01134 or MAPK1 3’-UTR covering wild-type and mutant-type miR-342-3p binding sites were inserted into the pmirGLO plasmid (Promega, Madison, WI, United States). HCC cells were co-transfected with miR-342-3p mimics or NC mimics along with pmirGLO plasmid for 48 h. Later, the luciferase activity was analyzed by a dual-luciferase reporter assay system (E1910, Promega, Madison, WI, United States). Renilla luciferase activity acted as an internal reference. As for the luciferase activity detection of different signaling pathways, cells were first plated into 96-well plates. Next, the Cignal Finder Reporter Array Kit (336841, QIAGEN, Dusseldorf, Germany) was utilized to measure the luciferase activity of NOTCH pathway, Wnt pathway, Hedgehog pathway, PI3K/AKT pathway, MAPK pathway, and NF-κB pathway, severally. Each experiment was performed in triplicate.

RIP assay was done by means of Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (638970, Merck, Darmstadt, Germany). After cells were lysed in RIP lysis buffer, collected cell lysate was cultivated with the magnetic beads linked to Ago2 antibody (MABE-253, Sigma-Aldrich) or IGF2BP2 antibody. IgG antibody (ab172730, Abcam, Cambridge, MA, United States) served as NC. The enrichment of RNAs was assessed using RT-qPCR. The experiment was independently performed in triplicate.

Total protein from HCC cells was extracted by RIPA (R0278, Sigma-Aldrich) first. Next, proteins were shifted onto PVDF (IPVH00010, Millipore) membranes after separation on 12% SDS-PAGE (1610174, Bio-Rad Laboratories, Shanghai, China) and then blocked in 5% nonfat milk. The primary antibodies against γH2AX (80312, CST), Cleaved PARP (5625, CST), p-ATM (13050, CST), Rad50 (3427, CST), p-Chk2 (2197, CST), Ku80 (2753, CST), MRE11 (4847, CST), NBS1 (3001, CST), DNA-PKcs (38168, CST), p53 (48818, CST), MAPK1 (4370, CST), ERK (ab32537, Abcam), p-ERK (ab229912, Abcam), JNK (ab76125, Abcam), p-JNK (ab176662, Abcam), p38 (ab170099, Abcam), p-p38 (ab31828, Abcam), HNRNPA1 (ab177152, Abcam), IGF2BP2 (ab124930, Abcam), GAPDH (ab8245, Abcam), and β-actin (ab181092, Abcam) were used for incubating cellular protein overnight at 4°C. GAPDH and β-actin worked as an internal reference. Secondary antibodies (ab7090, Abcam) were labeled with horseradish peroxidase (HRP) and added for 2 h incubation at room temperature. The western blots were developed by ECL luminous liquid (Pierce, Rockford, IL, United States). The experiment was independently performed in triplicate.

A total of 24 male BALB/c nude mice (4 weeks old) were bought from the Model Animal Research Center of Nanjing University. All animal experiment procedures were approved by Shandong Provincial Hospital Affiliated to Shandong University. The mice were divided into four groups based on random selection. HCC cells transfected with 0 Gy/sh-NC, 0 Gy/sh-LINC01134#1, 4 Gy/sh-NC, 4 Gy/sh-LINC01134#1, were incubated with Matrigel (1:1), and then 1.0 × 10⁵ HCC cells were subcutaneously inoculated into the mice. When tumors grew to about 8.0 mm in diameter, mice in each group were exposed to radiation. Mice mechanically fixed by a clamp were exposed to 4 Gy radiation at a dose rate of 0.955 Gy/min. Tumor volume was monitored every 5 days. After 30 days, tumors were excised from all sacrificed mice, and tumor weight was measured.

Experimental results in triplicate were presented as the mean ± standard deviation (SD) and analyzed with the help of GraphPad PRISM 6 (GraphPad, San Diego, CA, United States). Group difference was analyzed by Student’s t-test or one-way/two-way analysis of variance (ANOVA), which was regarded to be significant when p < 0.05.

According to starBase (http://starbase.sysu.edu.cn/index.php) prediction, LINC01134 was discovered to be obviously upregulated in liver hepatocellular carcinoma (LIHC) tissues compared with the control group (Figure 1A). RT-qPCR was implemented to examine LINC01134 expression in HCC cell lines (Huh7, Hep3B, SNU-182, and SK-HEP-1) and normal cell line (THLE-3). The result indicated that LINC01134 was significantly upregulated in HCC cells, particularly in SNU-182 and SK-HEP-1 cells (Figure 1B). To cut down the expression of LINC01134, sh-LINC01134#1/2/3 plasmids were transfected into SNU-182 and SK-HEP-1 cells. It turned out sh-LINC01134#1/2/3 efficiently knocked down LINC01134, and sh-LINC01134#1/2 had higher efficiency than sh-LINC01134#3 (Figure 1C). Afterwards, we conducted a CCK-8 assay to assess the changes in cell viability after LINC01134 knockdown. It was found that LINC01134 downregulation weakened HCC cell viability (Figure 1D). Subsequently, data collected from colony formation assay revealed that, in HCC cells with different radiation doses (0, 2, 4, 6, and 8 Gy), knockdown of LINC01134 lowered the survival fraction of HCC cells (Figure 1E). As the survival fraction markedly declined at 4 Gy radiation, this radiation dose was adopted for the following experiments. Overall, LINC01134 expression is higher in LIHC tissues and HCC cells compared to their corresponding control groups, and LINC01134 deficiency led to a decline in HCC cell viability.

To assess the function of LINC01134 in the radioresistance of HCC cells, functional experiments were implemented. Firstly, LINC01134 expression was examined in HCC cells exposed to 0 Gy or 4 Gy radiation via RT-qPCR assay. It turned out that 4 Gy radiation caused no significant change in LINC01134 expression (Supplementary Figure S1A). Then, colony formation assay manifested that 4 Gy radiation dramatically restricted HCC cell proliferation, and LINC01134 downregulation further suppressed cell proliferation (Figure 2A). Conversely, the results of the TUNEL assay disclosed that HCC cell apoptosis facilitated by 4 Gy radiation was further stimulated by LINC01134 depletion (Figure 2B). Moreover, the effects of 4 Gy radiation and LINC01134 knockdown rose with radiation time (Figure 2C). γ-H2AX foci are reckoned as a biomarker for DNA damage (Lobachevsky et al., 2020). 53BP1 is a DNA damage response factor (Mirman and de Lange, 2020). Hence, we conducted immunofluorescence staining and found that the formation of γ-H2AX and 53BP1 foci increased in cells treated with 4 Gy radiation, and LINC01134 knockdown promoted this trend (Figure 2D). Moreover, the effect of 4 Gy radiation and LINC01134 was the strongest after 12 h radiation and then gradually weakened after 24 and 36 h (Supplementary Figure S1B). Further, the comet assay results also uncovered that after exposure to 4 Gy radiation, the DNA repair capability of SNU-182 and SK-HEP-1 cells was weakened, and LINC01134 inhibition further repressed DNA repair capacity (Figure 2E). Furthermore, xenograft assay was implemented to explore the effects of LINC01134 depletion and 4 Gy radiation in vivo. The results showed that tumor growth was highly restricted after exposure to 4 Gy radiation. Under 4 Gy radiation, downregulated LINC01134 further hindered tumor growth (Figures 3A–C; Supplementary File S1). Taken together, silencing of LINC01134 restricts the radioresistance in HCC and impedes HCC tumor growth.

Through subcellular fractionation assay, we observed that LINC01134 was prominently localized in HCC cell cytoplasm (Figure 4A), which indicated that LINC01134 might be involved in post-transcriptional events. RIP assay demonstrated that LINC01134 could be detected in the Ago2-precipitated complex, suggesting that LINC01134 might function as a ceRNA by interacting with miRNAs (Figure 4B). With the application of starBase, eight possible miRNAs were predicted on the condition of Pan-Cancer ≥ 4. Data from the RNA pull-down assay proved that only miR-342-3p had a strong affinity with LINC01134 (Figure 4C). The alignment between LINC01134 and miR-342-3p was exhibited in Figure 4D. After transfection of miR-342-3p mimics, miR-342-3p was overexpressed in SNU-182 and SK-HEP-1 cells (Figure 4E). Subsequently, the luciferase reporter assay attested that the luciferase activity in the LINC01134-Wt group rather than the LINC01134-Mut group was decreased after transfection of miR-342-3p mimics (Figure 4F). To conclude, miR-342-3p is sponged by LINC01134 in HCC cells.

Given the experimental results of the luciferase reporter assay, we uncovered that depletion of LINC01134 lessened the luciferase activity of the MAPK signaling pathway (Figure 5A). In addition, we found that MAPK1 (also known as ERK2), a crucial regulator in the MAPK signaling pathway (Reyes-Gibby et al., 2016), was also a potential target gene of miR-342-3p. Accordingly, we conducted a western blot to analyze whether LINC01134 deficiency influenced MAPK1 and downstream factors of the MAPK signaling pathway. The results manifested that the protein levels of MAPK1, p-ERK (also referring to p-ERK1), p-JNK, and p-p38 were all cut down after LINC01134 was knocked down (Figure 5B; Supplementary File S2). Outcomes of RT-qPCR also demonstrated that miR-342-3p upregulation resulted in the decline of MAPK1 expression (Figure 5C). The projected binding region between miR-342-3p and MAPK1 was demonstrated in Figure 5D. Moreover, the RIP assay showed that the LINC01134, miR-342-3p, and MAPK1 were all enriched in the Ago2 groups (Figure 5E). The binding relation between miR-342-3p and MAPK1 was then validated by luciferase reporter assay as the luciferase activity of MAPK1 3’-UTR-Wt declined due to miR-342-3p overexpression (Figure 5F). Moreover, the decrease in MAPK1 expression on account of LINC01134 interference was partially restored by a miR-342-3p inhibitor (Figures 5G,H; Supplementary File S2). To summarize, LINC01134 modulates MAPK1 expression via sponging miR-342-3p to activate MAPK signaling pathway.

Based on the previous findings that miR-342-3p inhibitor could not entirely counteract the effect of LINC01134 knockdown on MAPK1 expression, we speculated there might exist other potential regulatory mechanisms for LINC01134 to regulate MAPK1. Venn diagram exhibited five candidate RBPs (NPM1, SND1, IGF2BP2, FBL, and HNRNPA1) screened out by starBase and GEPIA database (http://gepia.cancer-pku.cn/index.html) (Figure 6A). Given that HNRNPA1 and IGF2BP2 have been widely discovered to be able to regulate the progression of various cancers, including HCC (Simon et al., 2014; Wen et al., 2020), these two candidate RBPs were selected to engage in the following experiments. RNA pull-down assay uncovered that IGF2BP2 rather than HNRNPA1 was pulled down by biotin-labeled LINC01134 probe (Figure 6B; Supplementary File S2). The binding between LINC01134/MAPK1 and IGF2BP2 was further testified by the RIP assay (Figure 6C). FISH and IF analysis further disclosed the co-localization of LINC01134 and IGF2BP2 in HCC cell cytoplasm (Figure 6D). Next, the knockdown efficacy of sh-IGF2BP2#1/2 was confirmed to be high by RT-qPCR and western blot (Figures 6E,F; Supplementary File S2). From the data of RT-qPCR and western blot detection, MAPK1 expression at mRNA and protein levels was decreased when IGF2BP2 was downregulated (Figures 6G,H; Supplementary File S2). Moreover, the stability of MAPK1 mRNA in the HCC cells treated with actinomycin D (Act D) was lowered by IGF2BP2 inhibition (Figure 6I). In conclusion, LINC01134 interacts with IGF2BP2 protein to stabilize MAPK1 mRNA.

Before implementing rescue assays, MAPK1 was upregulated by transfection of pcDNA3.1/MAPK1 into HCC cells (Figure 7A). Consequences of colony formation assay illustrated that 4 Gy radiation hampered cell proliferation, and LINC01134 downregulation further restrained cell proliferation. At the same time, this effect was partially restored by miR-342-3p inhibitor and completely restored by MAPK1 overexpression (Figure 7B). The increased apoptotic cells due to 4 Gy radiation were further elevated by LINC01134 depletion, which was partially recovered by miR-342-3p downregulation while being completely recovered by MAPK1 upregulation (Figure 7C). Similarly, the suppressive effect of 4 Gy radiation on the formation of γ-H2AX and 53BP1 foci was enhanced by the LINC01134 knockdown, which was partially offset by miR-342-3p inhibition and completely counteracted by upregulation of MAPK1 (Figure 7D). The comet assay results also suggested that, under the conditions of 4 Gy radiation, the limited DNA repair capacity induced by LINC01134 deficiency was partially rescued by miR-342-3p inhibitor and completely rescued by overexpression of MAPK1 (Figure 7E). To sum up, the inhibiting effect of LINC01134 interference on the radioresistance of HCC cells is partially abrogated by miR-342-3p downregulation and fully abolished by MAPK1 upregulation.