Liver cancer cell lines HepG2 (cat. no. HB-8065) and SNU-449 (cat. no. CRL-2234), and the non-tumorigenic human hepatocyte cell line THLE-2 (cat. no. CRL-2706) were obtained from the American Type Culture Collection. The JHH-4 cell line (cat. no. JCRB0435) was sourced from the Japanese Collection of Research Bioresources Cell Bank. The human 293FT cell line (cat. no. R70007) was obtained from Thermo Fisher Scientific, Inc. Cell lines were authenticated by short tandem repeat analysis (Macrogen, Inc.). Cells were cultured in complete growth media containing 10% fetal bovine serum (cat. no. A5256701; Gibco; Thermo Fisher Scientific, Inc.) following the manufacturers' protocols and maintained at 37°C in a 5% CO₂ incubator.

To generate liver cancer cells stably expressing miR-885-5p, a second lentiviral vector system was utilized comprising the packaging plasmids pMD2.G (plasmid #12259; Addgene, Inc.) and psPAX2 (plasmid #12260; Addgene, Inc.) in 293FT cells (cat. no. R70007; Invitrogen; Thermo Fisher Scientific, Inc.). The pre-miRNA sequence of miR-885-5p was amplified from human genomic DNA and cloned into an EcoRI/BamHI-cut pLVX-EF1α-IRES-Puro vector (cat. no. 631988; Takara Bio USA, Inc.). For each 6-well plate, 3.08 μg pLVX-EF1α-pre-miR -885-5p-IRES-Puro, 3.08 μg psPAX2 and 2 μg pMD2.G were co-transfected into 293FT cells with a calcium phosphate transfection kit (cat. no. CAPHOS; Sigma-Aldrich; Merck KGaA) for 8 h at 37°C according to the established protocol (20). Lentiviral particles were collected at 48 h post-transfection and were concentrated before being used to transduce target cells at a multiplicity of infection of 3 using a spinoculation method (800 × g, 40 min, 37°C) (21). Following transduction, cells exhibiting stable integration were selected using puromycin (cat. no. P8833; Sigma-Aldrich; Merck KGaA) at 1-2 μg/ml until all non-transduced control cells were eliminated (typically 7-10 days). Subsequently, the cells were maintained in 0.5-1 μg/ml puromycin for further experimentation. Successful overexpression of miR-885-5p was confirmed via reverse-transcription quantitative PCR (RT-qPCR). Primer sequences used for cloning and RT-qPCR analysis are detailed in Table SI.

Cell proliferation was measured using a standard MTT assay (cat. no. M6494; Invitrogen; Thermo Fisher Scientific, Inc.). Cells were incubated with 0.5 μg/ml MTT for 1 h. The formazan crystals were dissolved in dimethyl sulfoxide (DMSO), and the absorbance was read at a wavelength of 570 nm (BioTek Synergy HTX; Agilent Technologies, Inc.). Colony formation was assessed by plating 500-2,000 cells/well of a 6-well plate and culturing for 2-3 weeks. Colonies were fixed in 100% methanol for 30 min at room temperature, stained with phosphate-buffered saline containing 0.1% crystal violet (cat. no. B21932.14; Thermo Fisher Scientific, Inc.) at room temperature for 30 min and counted for colonies comprising ≥50 cells.

mRNA and miRNA fractions were both isolated from liver cancer cells using the GeneAll Hybrid-R miRNA kit (cat. no. HB3520; GeneAll Biotechnology, Co., Ltd.). Complementary DNA was synthesized from the mRNA fraction using iScript Reverse Transcription Supermix (cat. no. 1708840; Bio-Rad Laboratories, Inc.) for mRNA analysis according to the manufacturer's instructions. For miRNA analysis, complementary DNA was synthesized from the miRNA fraction using RevertAid First Strand cDNA Synthesis kit (cat. no. K1621; Thermo Fisher Scientific, Inc.) according to the previous study (22). qPCR was performed on a QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific, Inc.) with CAPITAL qPCR Green Mix HRox (cat. no. BR0501901; Biotechrabbit GmbH) under the following conditions: 95°C for 10 min, followed by 35 cycles at 95°C for 15 sec and 60°C for 20 sec. Relative gene expression was calculated using the 2^−ΔΔCq method (23). RPL19 served as the internal reference gene for mRNA analysis and U6 served as the internal reference gene for miRNA analysis. The sequences of the primers used are listed in Table SI.

To identify key miRNAs in HCC, the Database of Differentially Expressed MiRNAs in Human Cancers (https://www.biosino.org/dbDEMC/index) was utilized (24). The top three datasets derived from HCC tissue samples with the highest number of differentially expressed miRNAs [absolute fold change (|FC|)>1.5 and adjusted P-value <0.05] were selected. Applying this criterion, datasets EXP00409, EXP00410 and EXP00576 were chosen for further analysis. Datasets EXP00409 (comparing miRNA expression between cancerous and normal liver tissues) and EXP00410 (comparing miRNA expression between stage 1 and stage 3 HCC tumors) were obtained from The Cancer Genome Atlas (TCGA) Liver Hepatocellular Carcinoma Collection project (https://portal.gdc.cancer.gov/projects/TCGA-LIHC). Dataset EXP00576, retrieved from the GSE147889 repository, also compared miRNA expression between normal and cancerous tissues from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE147889). For reanalysis, miRNA expression profiles from EXP00576 were processed using the GEO2R web tool (version 3.26.8; https://www.ncbi.nlm.nih.gov/geo/geo2r/). Expression data from the TCGA Liver Hepatocellular Carcinoma Collection datasets were retrieved using the TCGAbiolinks R package (R Core Team; Posit Software, PBC). All expression values were transformed using log₂(expression + 1).

RNA-sequencing (RNA-seq) analysis was performed by Beijing Biomarker Technologies Co., Ltd. RNA quality and integrity were assessed using the Agilent Bioanalyzer 2100 system (Agilent Technologies, Inc.) to ensure suitability for sequencing. Libraries were prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina^® (cat. no. E7770; New England Biolabs, Inc.). The final library concentration was examined using Qubit 2.0 (Thermo Fisher Scientific, Inc.), diluted to a loading concentration of 2 nM, and sequenced in paired-end 150 bp mode on the Illumina NovaSeq 6000 platform (Illumina, Inc.).

Raw sequencing reads were filtered for quality and adapter contamination. The clean reads were then mapped to the reference genome using HISAT2 (version 2.2.1; https://daehwankimlab.github.io/hisat2/). Gene expression levels were quantified using StringTie (version 2.2.1; https://ccb.jhu.edu/software/stringtie/index.shtml). Differentially expressed genes (DEGs) between control and miR-885-5p-overexpressing JHH-4 cells were identified using DESeq (version 1.30.1; http://www.bioconductor.org/packages/release/bioc/html/DESeq.html) with a threshold of |FC|>1.5 and P<0.05.

Functional enrichment analysis of DEGs was performed using g:Profiler web tool (version e109_eg56_p17_63f6c3d4; https://biit.cs.ut.ee/gprofiler_archive3/e109_eg56_p17/gost) to identify enriched Gene Ontology (GO) terms (https://geneontology.org), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (https://www.genome.jp/kegg/pathway.html), Reactome pathways (https://reactome.org), and WikiPathways (https://www.wikipathways.org/organisms/human.html). Statistical significance for GO and KEGG pathway enrichment was determined using a false discovery rate (FDR) <0.05 and P<0.05. Gene Set Enrichment Analysis (GSEA) was performed to identify significantly enriched biological pathways associated with DEGs. The analysis was conducted for KEGG and Reactome pathways using the clusterProfiler and ReactomePA R packages (Posit Software, PBC). Differential expression analysis results, obtained from the RNA-seq data, were used to calculate a ranking metric for each gene, computed as Ranking Metric=sign(logFC)*-log₁₀ (P-value). The most significant negatively enriched pathways in KEGG and Reactome were then visualized using GSEA enrichment plots.

Cell proliferation was assessed by measuring bromodeoxyuridine (BrdU) incorporation. Cells were incubated with 10 μM BrdU (cat. no. ab142567; Abcam) for 30-60 min at 37°C, fixed with 4% paraformaldehyde for 15 min at room temperature and permeabilized with 2N hydrochloric acid for 30 min at room temperature. BrdU incorporation was detected by immunofluorescence using an anti-BrdU antibody (1:200; cat. no. sc-32323; Santa Cruz Biotechnology, Inc.) overnight at 4°C, followed by Alexa Fluor 488-conjugated secondary antibody (1:500; cat. no. A-21202; Invitrogen) for 1 h at room temperature and DAPI staining (1 μg/ml; cat. no. D9542; Sigma-Aldrich; Merck KGaA) for 10 min at room temperature. Images were captured using an EVOS M7000 cell imaging system (Invitrogen; Thermo Fisher Scientific, Inc.) and BrdU-positive cells were quantified from 10 random fields/group with >2,000 nuclei counted, using ImageJ software (version 1.54g; National Institutes of Health).

To determine the cell cycle distribution, liver cancer cells were harvested and fixed in ice-cold 70% ethanol overnight. After washing with 1% fetal bovine serum in phosphate-buffered saline, cells were stained by incubation in phosphate-buffered saline containing 50 μg/ml propidium iodide (PI; cat. no. P1304MP; Thermo Fisher Scientific, Inc.) and 100 μg/ml RNase A (cat. no. EN0531; Thermo Fisher Scientific, Inc.) for 30 min at 37°C in the dark. DNA content was then analyzed using a MACSQuant X Flow Cytometer (Miltenyi Biotec, Inc.) by detecting PI fluorescence in the PE channel. Debris was excluded using FSC-A/SSC-A gating, and doublets were removed by pulse-processing (FSC-H vs. FSC-A). The proportion of cells in each phase (G₁, S and G₂/M) was quantified using the FlowJo software (version 10.10; BD Biosciences).

Apoptosis was assessed using the Apoptosis/Necrosis Assay Kit (blue, green, red; cat. no. ab176749; Abcam). Liver cancer cells were washed twice with Assay Buffer and resuspended in Assay Buffer containing 1X Apopxin Green Indicator at room temperature. Cells were incubated at room temperature for 30 min while protected from light. After incubation, cells were washed twice with Assay Buffer and replaced with fresh Assay Buffer. Apoptotic cells were visualized using the fluorescein isothiocyanate channel (Ex/Em=490/525 nm) on an EVOS M7000 cell imaging system (Invitrogen; Thermo Fisher Scientific, Inc.). Images were captured using an EVOS M7000 cell imaging system (Invitrogen; Thermo Fisher Scientific, Inc.), and quantification of Apopxin Green-positive cells was performed from 10 random fields/group with >2,000 nuclei counted, using the ImageJ software (version 1.54g; National Institutes of Health).

To identify potential miRNA-mRNA interactions, TargetScan web tool (version 8.0; https://www.targetscan.org/vert_80/) was used to predict miRNA targets from a pool of downregulated DEGs involved in the G₁-to-S transition, and binding sites were further predicted using RNAhybrid web tool (version 2.1.1; https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid/). Based on these analyses four genes (CDK6, ORC1, E2F1 and E2F2) were selected for experimental validation using a dual luciferase reporter assay. The 3′ UTRs of these genes containing either wild-type (WT) or mutated (MT) binding sites were cloned into the pmirGLO vector (cat. no. E133A; Promega Corporation). The pmirGLO vector with no insert was used as the negative control. The pre-miRNA sequence of miR-885-5p was cloned into the pSilencer 3.0 H1 vector (cat. no. AM7210; Thermo Fisher Scientific, Inc.). Human 293FT cells (cat. no. R70007; Invitrogen; Thermo Fisher Scientific, Inc.) were co-transfected with these constructs by a standard calcium phosphate transfection (20). Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (cat. no. E1910; Promega Corporation) 48 h post-transfection using a BioTek Synergy HTX Multi-Mode Microplate Reader (BioTek; Agilent Technologies, Inc.). Firefly luciferase activity was normalized to Renilla luciferase activity for data analysis. Primer sequences used for cloning are detailed in Table SI.

Protein extracts from liver cancer cells were prepared using radioimmunoprecipitation assay lysis buffer supplemented with a protease inhibitor cocktail (cat. no. 04693132001; Roche Diagnostics). Protein concentrations were then determined using a Bradford protein assay (cat. no. 5000001; Bio-Rad Laboratories, Inc.). Equal amounts of protein (40 μg) were denatured and separated by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis. Proteins were then transferred to a nitrocellulose membrane using a semidry method (Bio-Rad Laboratories). Following transfer, membranes were blocked for 30 min in a blocking buffer (cat. no. BM01-500; Visual Protein; Energenesis Biomedical, Co., Ltd.). Membranes were subsequently incubated overnight at 4°C with the following primary antibodies: CDK6 (1:1,000; cat. no. sc-7961; Santa Cruz Biotechnology); Origin Recognition Complex Subunit 1 (ORC1; 1:1,000; cat. no. A14756; ABclonal Biotech Co., Ltd.) and β-actin (1:2,000; cat. no. sc-47778; Santa Cruz Biotechnology). Membranes were then incubated with HRP-conjugated goat anti-rabbit or anti-mouse IgG secondary antibody (1:5,000; cat. no. 7074 and 7076, respectively; Cell Signaling Technology) for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence substrate (cat. no. RPN3004; Cytiva) and imaged using a ChemiDoc Touch Imaging System (Bio-Rad Laboratories, Inc.). Densitometric quantification of protein band intensity was performed using the ImageJ software (version 1.54g; National Institutes of Health) and levels were normalized against those of β-actin as a loading control.

Abemaciclib (cat. no. HY-16297A), palbociclib (cat. no. HY-50767) and ribociclib (cat. no. HY-15777B) were purchased from MedChemExpress and dissolved in DMSO to prepare stock solutions at a concentration of 20 mM. Serial dilutions were performed to achieve final drug concentrations ranging from 0 to 50 μM. The final concentration of DMSO in all treatment conditions did not exceed 0.5% (v/v), and an equivalent volume of DMSO was added to control cells to ensure consistency across experimental conditions. Cells were seeded at a density of 1×10⁴ cells/well, and drug treatments were carried out for 72 h at 37°C in a 5% CO₂ incubator. The half-maximal inhibitory concentration (IC₅₀) of each drug was determined in both control and miR-885-5p-overexpressing cells using the MTT assay as previously mentioned. Dose-response curves were generated and IC₅₀ values were calculated using non-linear regression analysis using the GraphPad Prism software (version 9; Dotmatics).

Data are presented as mean ± standard deviation. Statistical significance was determined using GraphPad Prism (version 10; Dotmatics). Unpaired, two-tailed Student's t-tests were used for comparisons between two groups. One-way analysis of variance was used for comparisons among ≥3 groups. Spearman correlation analysis was used to study the relationship between miR-885-5p and cell cycle genes in HCC tissues. Non-linear regression analysis was employed to analyze the effects of CDK4/6 inhibitors on liver cancer cells. P<0.05 was considered to indicate a statistically significant difference. All experiments were performed independently at least twice, with each condition tested in triplicate.

To identify miRNAs with potential roles in liver cancer pathogenesis, miRNA expression profiles were reanalyzed from normal and HCC tissue pairs (dataset IDs. EXP00409 and EXP00576) and low-grade vs. high-grade HCC tissues (dataset ID. EXP00410) using the Database of Differentially Expressed MiRNAs in Human Cancers (24). miRNAs that were significantly downregulated in HCC or high-grade HCC were prioritized, as it was hypothesized that these miRNAs may potentially function as tumor suppressors. The analysis demonstrated five commonly downregulated miRNAs across all three datasets: miR-99a-5p, miR-122-3p, miR-122-5p, miR-139-5p, and miR-885-5p (Fig. 1A and Table SII). While the tumor-suppressive roles of several of these miRNAs are well-documented (25-29), presently, the specific function of miR-885-5p in HCC is the least known. Therefore, investigation was conducted to uncover novel mechanistic insights into its tumor-suppressive role in HCC. Supporting the present hypothesis, miR-885-5p was consistently downregulated in HCC tissues compared with normal tissues across both the TCGA (dataset EXP00409, P=0.005) and GEO (dataset EXP00576, P<0.001) datasets (Fig. 1B). Furthermore, it showed significant downregulation in high-grade tumors (dataset EXP00410, P=0.001). Furthermore, 5-year Kaplan-Meier survival analysis was performed on TCGA-LIHC cohort. Patients were stratified into high- and low-expression groups for miR-885-5p; the high-expression group was defined as patients with expression at or above the 80th percentile, while the low-expression group consisted of those with expression below this threshold. This analysis demonstrated that high miR-885-5p expression was associated with significantly prolonged overall survival in patients with HCC (P=0.0204; hazard ratio, 0.61; 95% confidence interval, 0.40-0.93; Fig. 1C).

To investigate the functional roles of miR-885-5p in HCC, the pre-miRNA sequence of miR-885-5p was cloned into a lentiviral vector and transduced into three liver cancer cell lines: HepG2, JHH-4 and SNU-449. These cell lines exhibited reduced expression levels of endogenous miR-885-5p compared with that of a normal hepatocyte cell line, THLE-2 (P<0.001; Fig. 1D). Cells transduced with a lentivirus containing a scrambled miRNA sequence served as controls. Following puromycin selection, successful miR-885-5p overexpression (885-OE) was confirmed by RT-qPCR. Significant miR-885-5p induction was achieved in each cell line compared with its respective control (HepG2, P<0.001; JHH-4, P=0.001; SNU-449, P=0.002; Fig. 1E). While the absolute levels of overexpression varied among cell lines, consistent and robust relative induction was observed. MTT assays demonstrated that 885-OE cells significantly impaired cell proliferation in all three liver cancer cell lines compared with that of control cells (HepG2, P=0.01; JHH-4, P=0.002; SNU-449, P<0.001; Fig. 1F). Consistent with this defect, 885-OE cells also demonstrated reduced colony formation ability compared with that of control cells (HepG2, P=0.013; JHH-4, P=0.001; SNU-449, P=0.045; Fig. 1G).

To elucidate the molecular mechanisms underlying the tumor-suppressive function of miR-885-5p, RNA-seq on control and 885-OE JHH-4 cells (GEO ID. GSE290378) was performed. Transcriptome analysis demonstrated 1,190 upregulated and 990 downregulated DEGs in 885-OE cells compared with controls (|FC|>1.5 and P<0.05; Fig. 2A; Tables SIII and SIV). Consistent with the typical inhibitory effect of miRNAs on mRNA targets, further analysis on downregulated DEGs was conducted. KEGG analysis indicated that miR-885-5p overexpression significantly affected the cell cycle, DNA replication and homologous recombination (P<0.001; Fig. 2B). GO enrichment analysis demonstrated the G₁/S transition as a key affected process (P<0.001; Fig. 2B). Reactome and WikiPathways analyses further supported this finding, demonstrating the potential role of miR-885-5p in the G₁/S checkpoint (P<0.001; Fig. 2B). GSEA corroborated these findings, demonstrating a significant association between miR-885-5p overexpression and downregulation of key G₁/S transition genes, including CDK6, E2F Transcription Factor (E2F1), E2F2 and ORC1 (GSEA analysis, q<0.001; Fig. 2C and D). To validate these findings, RT-qPCR analysis was performed on a representative list of top downregulated genes involved in the G₁/S transition, including CDK6, CCNA2, E2F1, E2F2, PPP2CA and ORC1 in two independent liver cancer cell lines, which confirmed their significant downregulation in 885-OE cells (Fig. 2E).

The G₁/S transition, a critical checkpoint for DNA replication, is frequently dysregulated in cancer, leading to uncontrolled proliferation (30,31). To investigate if miR-885-5p regulates this transition, BrdU incorporation assays were performed. This method identifies cells in S phase and assesses G₁/S progression, where a decrease in BrdU incorporation indicates impaired transition from G₁ (gap phase 1) into S phase (DNA synthesis) (32). A significant decrease in BrdU incorporation in 885-OE cells was demonstrated, indicating a reduced proportion of cells in S phase (HepG2, P=0.004; JHH-4, P<0.001; SNU-449, P=0.019; Fig. 3A). Cell cycle analysis by flow cytometry was subsequently performed to precisely quantify the proportion of cells in each cell cycle phase (G₁, S and G₂/M). This analysis confirmed G₁ arrest in 885-OE cells, demonstrating an increased proportion of cells in the G₁ phase (HepG2, P=0.002; JHH-4, P=0.007; SNU-449, P<0.001; Fig. 3B). These findings were consistent across three liver cancer cell lines and aligned with the reduced proliferation and colony formation observed in aforementioned MTT and colony formation assays. Furthermore, the G₁ accumulation in 885-OE cells was consistent with the enrichment of pathways governing G₁/S transition identified in the present RNA-seq analysis. To further assess the impact of miR-885-5p on cell fate, its effect on apoptosis was investigated. Apoptosis, which is characterized by the translocation of phosphatidylserine to the outer plasma membrane leaflet, was measured using Apopxin Green Indicator staining. Overexpression of miR-885-5p significantly increased the proportion of apoptotic cells (HepG2, P<0.001; JHH-4, P<0.001; SNU-449, P=0.003; Fig. S1).

To investigate whether miR-885-5p directly modulates genes involved in the G₁/S transition, potential miR-885-5p targets within the 3′ UTRs of key genes identified from the G₁/S transition term highlighted by the present RNA-seq analysis were analyzed. The TargetScan tool was used to predict four genes as potential direct targets. CDK6, ORC1, E2F1 and E2F2 were prioritized for further validation and determined their lowest predicted minimum free energy of binding sites using the RNAhybrid web tool (Fig. 4A). WT and MT 3′ UTR constructs were generated. To confirm the specificity of miR-885-5p binding, the 5′ seed regions within the predicted binding sites of the MT constructs were engineered to abolish interaction. Cotransfection of miR-885-5p expression plasmids with the WT constructs significantly decreased luciferase activity for the firefly luciferase reporter fused with the 3′ UTRs of CDK6 (P<0.001 vs. no-insert control) and ORC1 (P<0.001 vs. no-insert control) but not those of E2F1 and E2F2 (Fig. 4B). Mutations within the miR-885-5p binding sites in the MT constructs significantly abolished this repressive effect compared with their respective WT constructs (CDK6, P<0.001 vs. WT; ORC1, P<0.001 vs. WT), thereby demonstrating the specificity of the interaction (Fig. 4B). Clinical data further supported the regulation of miR-885-5p as evidenced by a stronger negative correlation between miR-885-5p and CDK6/ORC1 (r=−0.448 and −0.221) compared with E2F1/E2F2 (r=-0.095 and −0.173; Fig. 4C). To validate the physiological relevance of this direct targeting, the endogenous protein expression levels of CDK6 and ORC1 upon miR-885-5p overexpression were assessed. Western blot analysis confirmed a significant downregulation of both CDK6 and ORC1 protein levels in 885-OE cells compared with control cells (P<0.001; Fig. S2).

Given that miR-885-5p downregulated CDK6 expression, its potential to sensitize liver cancer cells to CDK4/6 inhibitors was investigated. Drug sensitivity was quantified by IC₅₀, where lower IC₅₀ values indicated increased drug sensitivity, and higher IC₅₀ values indicated decreased sensitivity or resistance. Overexpression of miR-885-5p significantly increased the sensitivity of liver cancer cells to palbociclib, ribociclib and abemaciclib as demonstrated by significantly reduced IC₅₀ values in the liver cancer cell lines examined (Fig. 5A-C). In SNU-449 cells, the IC₅₀ for palbociclib decreased from 7.49±0.15 μM in control cells to 1.50±0.17 μM in 885-OE cells (P<0.001). In JHH-4 cells, the IC₅₀ for palbociclib decreased from 8.25±1.12 μM in control cells to 3.35±0.41 μM in 885-OE cells (P=0.028). Similar significant reductions in IC₅₀ values were observed for ribociclib (SNU-449, P=0.034; JHH-4, P= 0.024) and abemaciclib (SNU-449, P= 0.011; JHH-4, P= 0.035).