Abstract
Background
Renal cell carcinoma (RCC) is a prevalent malignancy of the urinary system that presents significant health and economic burdens. Despite existing treatments such as surgery and targeted therapies, challenges remain due to suboptimal efficacy and high recurrence rates. Previous studies have indicated that metformin and everolimus individually exhibit inhibitory effects on RCC. However, their synergistic potential when combined has not been fully elucidated. Therefore, this paper identified the antiproliferative effect and the hub genes that undergo significant changes in 786-O cells when treated with the combination drugs and their underlying mechanisms to inform the search for kidney renal clear cell carcinoma (KIRC) therapeutic targets.
Methods
The effects of the combination of metformin and everolimus on 786-O cells viability, migration and invasion were investigated. Differentially expressed genes (DEGs) among different drug treatment groups were identified through ribonucleic acid (RNA) sequencing, raw data processing and differential expression analysis. The target genes were obtained by taking the intersection of different DEGs, and hub genes were identified by Maximal Clique Centrality (MCC) and Molecular Complex Detection (MCODE) algorithms, expression validation, and Kaplan-Meier (K-M) survival curve plotting. Subsequently, transcription factors (TFs) regulating the hub genes were identified and drug-hub gene interactions were explored through molecular docking. In addition, gene set enrichment analysis (GSEA) demonstrated hub gene-related biological functions and pathways, and gene set variation analysis (GSVA) explored differential pathways between different drug treatment groups. Finally, quantitative real-time polymerase chain reaction (qRT-PCR) was performed to verify the expression difference of hub genes among four groups.
Results
The combination of metformin and everolimus is more effective than monotherapy at inhibiting cell viability, migration, and invasion in 786-O cells. In total, 3,030 DEG1, 2,953 DEG2, 3,591 DEG3, 1,571 DEG4 and 4,064 DEG5 were identified, yielding five target genes. After MCC and MCODE algorithms, expression validation, and K-M survival curve plotting, target genes were all noted as hub genes (SPC25, NCAPH, MCM10, UHRF1, SMC4). Eleven TFs regulated more than two hub genes, and the binding energy of metformin with SPC25 and everolimus with SMC4 was the lowest. Hub genes were negatively correlated with lysosome and positively associated with cell cycle, and the P13K/Akt/mTOR signaling pathway was significantly positively correlated with hub genes.
Conclusions
Metformin and everolimus are synergistic in anticancer effects on RCC. Based on transcriptomic data, this study obtained five hub genes associated with everolimus and metformin combination therapy in KIRC to inform KIRC-related research.
Keywords: Kidney renal clear cell carcinoma (KIRC), metformin, everolimus, molecular docking, bioinformatics
Highlight box.
Key findings
• Metformin-everolimus combination therapy significantly inhibits proliferation, migration, and invasion of 786-O KIRC cells.
• Five hub genes (SPC25, NCAPH, MCM10, UHRF1, SMC4) were identified and validated, with high expression linked to poor survival.
• Hub genes are positively correlated with the PI3K/Akt/mTOR pathway, suggesting their role in mediating drug synergy.
• Molecular docking revealed strong binding affinity between metformin-SPC25 (−5.8 kcal/mol) and everolimus-SMC4 (−8.7 kcal/mol).
What is known and what is new?
• Metformin and everolimus have anticancer potential, but their synergistic mechanism in KIRC is unclear. PI3K/Akt/mTOR pathway is a key oncogenic driver in KIRC.
• This is the first systematic identification of hub genes for metformin-everolimus combination therapy in KIRC. KLF16 was identified as a key transcription factor regulating multiple hub genes.
What is the implication, and what should change now?
• Scientific implications: hub genes (e.g., SPC25) are potential therapeutic targets for KIRC.
• Clinical actions: future clinical trials should validate the efficacy of this combination therapy in patient populations. Future studies need to explore these hub genes as potential biomarkers for patient stratification.
• Future research: subsequent research should focus on functional validation of the identified hub genes and development of targeted inhibitors.
Introduction
Renal cell carcinoma (RCC) ranks among the most prevalent cancers affecting the urinary system, with an annual incidence increase of 1.5% (1). The most common pathologic type is clear cell renal cell carcinoma [ccRCC, also referred to as kidney renal clear cell carcinoma (KIRC)], accounting for approximately 75–80% of all RCC cases (2). The mortality rate of RCC has been decreasing rapidly due to progress in early diagnosis and treatment. Around 20–30% of individuals are diagnosed with metastatic disease despite early detection (3). Additionally, nearly 20–40% of patients with localized RCC still experience distant metastases after surgical treatment (4). RCC is insensitive to traditional chemotherapy and radiotherapy. While novel therapies like immunotherapy and targeted therapy have improved outcomes for RCC patients, their severe side effects and high costs still restrict their use (5). The prognosis is still extremely poor for those with metastatic disease. Therefore, searching for effective therapeutic agents and functioning hub genes is urgent and necessary.
Everolimus, an mTOR inhibitor, has been approved as a second-line treatment for patients with advanced RCC, showing improved safety assessment of progression-free survival (PFS, 4.0 vs. 1.9 months) compared to placebo (6). However, since the efficacy of everolimus monotherapy is limited, there is a necessity to explore strategies to enhance the response to mTOR inhibitors through the combination of other agents. Metformin is the most widely prescribed medication for diabetes management globally and has established itself as the primary treatment option for individuals newly diagnosed with type 2 diabetes mellitus (T2DM). This is due to its long-term safety and effectiveness, minimal risk of hypoglycemia, associated cardiovascular advantages, mortality reduction benefits, synergistic properties when used in combination therapies, and its cost-effectiveness (7-10). Increasing evidence demonstrates that metformin has potential efficacy in anticancer therapy, including breast cancer, cervical cancer, pancreatic cancer, and colorectal cancer (11-14), among others. In addition, metformin has been used as a synergistic cancer treatment because it enhances sensitivity to radiotherapy (15), chemotherapy (16), and immunotherapy (17) and reduces the side effects of anticancer treatments at lower therapeutic dosages. Several studies have shown the synergic effect of metformin and everolimus in RCC, neuroendocrine cancers, cervical cancer, and endometrial cancer (18-21). This suggests that metformin may enhance the sensitivity of everolimus in treating RCC. However, the underlying mechanism of metformin and everolimus synergistic anticancer effect has not been fully understood. Thus, further exploration of the hub genes and molecular mechanisms in the combination of metformin and everolimus for treating KIRC is essential.
In this study, the relationship between the two drugs (metformin, everolimus) and KIRC was further investigated by molecular biology methods. The hub genes related to the drugs after treatment in KIRC were identified by screening through ribonucleic acid (RNA) sequencing and bioinformatics analysis. A regulatory network was constructed and analyzed, which is important for improving patient prognosis and therapeutic efficacy. We present this article in accordance with the MDAR reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-277/rc).
Methods
Cell line and culture condition
The human RCC cell line 786-O was obtained from the Key Laboratory of Urological Diseases in Gansu Province. 786-O was maintained in RPMI-1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS) (PAN, Adenbach, Germany) and 1% penicillin-streptomycin mixed solution (Beyotime, Shanghai, China). The cells were cultured at 37 ℃ in a humidified environment containing 5% CO2 and were passaged when they grew to 80% confluence.
Cell viability assay
Cell viability was evaluated using the Cell Counting Kit-8 (CCK-8) assay (Biosharp, China). Briefly, 5×103 cells were seeded into 96-well plates and incubated overnight in a complete medium. Then, cells were treated with either 10 mM metformin (S1950, Selleck, China), 25 µM everolimus (S1120, Selleck, China), or a combination of 10 mM metformin and 25 µM everolimus for 24 hours. Subsequently, 10 µL of CCK-8 reagent was introduced to each well containing the culture medium, and cells were incubated for an additional hour. The absorbance of each well was measured at 460 nm using a microplate reader (BioTeck, Vermont, USA).
Wound healing assay
To investigate the impact of drugs on cell migration, a wound healing assay was conducted. 786-O cells were seeded into a six-well plate until reaching approximately 90% density. Straight scratch wounds were made with a plastic pipette tip in each well. The detached cells were then washed twice with PBS, and the cells were treated with metformin or everolimus for 24 hours, alone or in combination. The width of the scratch at 0 hour and 24 hours were monitored in order to evaluate the migration ability of cells.
Transwell assay
786-O cells (30×104 per well) were seeded into a six-well plate overnight. The cells were treated with metformin or everolimus, alone or in combination. Twenty hours later, the cells were transferred to the upper Transwell chamber pre-coated with Matrigel (Corning, New York, USA). Next, culture medium (with 20% FBS) was added in lower chambers. After 24 hours of incubation, cells invading the gel were fixed with 4% paraformaldehyde and subsequently stained with crystal violet. Images of the invading cells were captured using a microscope.
Data source
In this study, a total of twelve 786-O cell samples were included, which were evenly divided into four groups: group A, 786-O cell samples; group B, 786-O cell samples + 10 mM metformin; group C, 786-O cell samples + 25 µM everolimus; group D, 786-O cell samples + 10 mM metformin + 25 µM everolimus. The Cancer Genome Atlas Kidney Renal Clear Cell Carcinoma (TCGA-KIRC) served as a validation set, comprising 535 KIRC samples and 72 control samples. Of these, 531 KIRC samples contained survival information. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.
RNA sequencing
Following the manufacturer’s procedure, total RNA was isolated and purified using TRIzol reagent (Accurate Biology, Hunan, China). The poly(A) RNA was then fragmented into small pieces and reverse transcribed to produce complementary deoxyribonucleic acid (cDNA), and the processed cDNA products were amplified by polymerase chain reaction (PCR). The average insert size for the final cDNA library was 300±50 bp. At last, we performed the 2×150 bp paired-end sequencing (PE150) on an Illumina NovaseqTM 6000 (LC-Bio Technology Co., Hangzhou, China) following the vendor’s recommended protocol. The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in National Genomics Data Center (Nucleic Acids Res 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA-Human: HRA009715, HRA009723, HRA009724, HRA009722) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human.
Raw data procession
We first evaluated the sequencing data quality using FastQC (version 0.11.9) (22). The raw data was pre-processed to filter low-quality data and remove contamination and joint sequences, and the clean data were finally obtained utilizing Trimmomatic (version 0.39) (23) (adapter all.fa:2:30:10 LEADING:3TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36). To compare the reads of sequencing data, the reads of clean data and reference genome were contrasted by hisat2 (version 2.2.1) (24). Through FeatureCount (version 2.0.0) (25), the expression levels of each gene in the samples were further computed.
Differential expression and enrichment analyses
The differentially expressed genes (DEGs) between groups A and B were obtained via differential expression analysis by “DESeq2” (version 1.26.0) (26) (|log2fold change (FC)| > 0.5, adj.P<0.05), using as DEGs1. The “clusterProfiler” (version 3.14.3) (27) was employed to execute Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses to search biological functions and pathways related to DEGs1 (padjCutoff =0.05, OrgDb = org.Hs.eg.db, ont = “ALL”). The same analyses were performed between groups C and A, D and B, D and C, D and A to identify DEG2, DEG3, DEG4 and DEG5, separately, and their function were explored by GO and KEGG enrichment analyses as well.
Identification of hub genes
To analyze changes in transcriptional profiles among the disease group, drug alone groups, and combination group, the up/down regulated genes were separately intersected among DEG1–5 to screen target genes. For the aim of investigating the genes interaction at protein level, target genes were then used to construct a protein-protein interaction (PPI) network through Search Tool for the Retrieval of Interaction Gene (String, https://string-db.org/, confidence =0.4). Moreover, the top five genes were sifted by Maximal Clique Centrality (MCC) algorithms in CytoHubba plug-in, and the pathways of five genes were enriched using BioGO algorithm. The module with highest score was contained via Molecular Complex Detection (MCODE) in Cytoscape, and the pathways of genes in the module were enriched using BioGO. The genes selected by MCC and MCODE were intersected to contain key genes. The expressions of key genes in both sequencing data and TCGA-KIRC were further explored. Patients in TCGA-KIRC were generated into high and low expression groups based on the optimal cutoff, and Kaplan-Meier (K-M) survival curves were plotted to analyze the correlation of patient’s survival with genes expressions. The key genes with differential expression between KIRC and control groups and same expression trends in sequencing data and TCGA-KIRC were recorded as hub genes (P<0.05), as well as holding P<0.05 in K-M survival curves.
Regulatory network construction and molecular docking
The Encyclopedia of DNA Elements (ENCODE, https://www.encodeproject.org/) was used to predict transcription factors (TFs) targeting hub genes to construct a TF-hub gene regulatory network. To evidencing the interactions between drugs and hub genes, the 2-dimensional (2D) structures of metformin and everolimus and the 3-dimensional (3D) structure of metformin were demonstrated utilizing PubChem (https://pubchem.ncbi.nlm.nih.gov/). The PDB documents of hub genes were downloaded from ProteinData Bank (PDB, https://www.rcsb.org/), the 2D structure of everolimus was transformed into 3D structure using pybel mode. The preparation and pre-processing of drugs and gene protein structures were proceeded by CB dock (https://cadd.labshare.cn/cb-dock2/php/index.php), and Pymol was utilized to analyze and visualize the docking configuration.
Gene set enrichment analysis (GSEA) and gene set variation analysis (GSVA)
In the dataset, with the hallmark gene sets in the Molecular Signatures Database (MSigDB) (https://www.gsea-msigdb.org/gsea/msigdb) serving as the reference gene sets, a single gene enrichment analysis was conducted on the Hub genes using the “clusterProfiler” (version 3.14.3) (27). Subsequently, the correlation coefficients between the expression of all genes and the five genes were calculated and utilized as the sorting criteria for the GSEA enrichment analysis. This was done to explore the potential functions and pathways implicated in the disease. Specifically, the correlation coefficients between the expression of all genes and each of the five genes were separately computed as the sorting criterion, followed by the performance of GSEA enrichment analysis to investigate the potentially relevant functions and pathways associated with the disease. Subsequently, in order to examine the enrichment status of different pathways among different samples in each group, a GSVA enrichment analysis was carried out using hallmark gene sets. Hallmark gene sets are a super gene set consisting of multiple known gene sets, where each hallmark gene set corresponds to multiple basic gene sets of other categories. The “limma” (version 3.54.0) (28) was used to calculate the differential pathways among different groups (P<0.05).
Quantitative real-time polymerase chain reaction (qRT-PCR) validation
Twelve cell samples were equally divided into four groups (control, metformin, everolimus, and metformin + everolimus) to extract total RNA using the TRIzol method (Accurate Biology, China). The concentration of total RNA was measured by NanoPhotometer N50 (Thermo Fisher, Massachusetts, USA). Afterward, SureScript-First-strand-cDNA-synthesis-kit (Servicebio, Hubei, China) was used for messenger RNA (mRNA) reverse transcription to gain cDNA. The cDNA was then diluted for qRT-PCR detection with GAPDH as the internal reference gene. qRT-PCR was used to detect the expression level of genes using Aptamer PCR SYBR Green Master Mix (Novogene, Beijing, China). The primer sequences for genes are displayed in Table S1. The data were analyzed using the 2−ΔΔCt calculation method.
Statistical analysis
Bioinformatics analyses were conducted in the R program (version 4.2.2). All cell experiments were conducted in triplicate and the data were expressed as mean ± standard deviation. One-way analysis of variance (ANOVA) was used to detect the significance among multiple groups using GraphPad Prism 9.0 software. The results were regarded as significant at P<0.05.
Results
Effects of metformin and everolimus on the biological behaviors of 786-O cells
In order to evaluate the impact of these therapeutic agents on cell proliferation, migration and invasion, we conducted CCK-8, wounding healing and Transwell assays. Cells were subjected to treatment with 10 mM metformin and 25 µM everolimus for 24 hours, either separately or in combination. The findings from the CCK-8 assay indicated that the antiproliferative effect of the combination of everolimus with metformin was more significant compared with the groups treated with everolimus or metformin alone (P<0.0001, Figure 1A). Moreover, both the wound healing and the Transwell assays revealed that treatment with metformin or everolimus alone effectively hindered the wounding healing and invasive capabilities of the 786-O cells compared to the control group. Still, the combination treatment showed the best inhibitory effect (Figure 1B-1E).
Figure 1.
Effects of metformin and everolimus on the biological behaviors of 786-O cells. (A) CCK-8 assay results. (B,C) Wound-healing assays graph and corresponding bar graph. (D,E) The Transwell’s transfection graph and the corresponding bar graph (crystal violet staining, scale bar =100 µm). Data were presented as mean ± SD. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. CCK-8, Cell Counting Kit-8; Eve, everolimus; Met, metformin; SD, standard deviation.
Clean data were filtered by quality control
The quality of the sequencing data was assessed as being of good quality. After filtering and quality control of raw data, the nucleobase quality of each sequenced fragment was higher than QC30 and the nucleobase error rate was less than 0.001, suggesting the data was satisfied for the subsequent analysis (Table 1). The comparison rates of sequencing data were all higher than 90%, which indicated the sequencing data was not disturbed by exogenous RNA molecules (Table 2). In addition, gene expressions in each sample were demonstrated in the online table (available at https://cdn.amegroups.cn/static/public/10.21037tcr-2025-277-1.xlsx) showing MTATP6P1, AGRN, ERRFI1 and so many genes highly expressed in each sample.
Table 1. The result after filtering and quality control of the raw data.
| Sample | Length | ReadS | QC30 (%) | GC% |
|---|---|---|---|---|
| A1 | 36–150 | 17420887*2 | 96.74 | 47.07 |
| A2 | 36–150 | 16417084*2 | 96.88 | 48.28 |
| A3 | 36–150 | 15092806*2 | 96.96 | 47.88 |
| B1 | 36–150 | 19037232*2 | 97.03 | 47.92 |
| B2 | 36–150 | 22232642*2 | 96.91 | 47.86 |
| B3 | 36–150 | 21681754*2 | 96.98 | 47.26 |
| C1 | 36–150 | 15635038*2 | 96.9 | 48.04 |
| C2 | 36–150 | 14632443*2 | 96.9 | 48.17 |
| C3 | 36–150 | 15825457*2 | 96.61 | 48.02 |
| D1 | 36–150 | 16995976*2 | 96.22 | 48.01 |
| D2 | 36–150 | 19224036*2 | 97 | 48.09 |
| D3 | 36–150 | 18336847*2 | 96.91 | 48.64 |
A: 786-O cell samples; B: 786-O cell samples + 10 mM metformin; C: 786-O cell samples + 25 μM everolimus; D: 786-O cell samples + 10 mM metformin + 25 μM everolimus. GC, guanine-cytosine; QC, quality control.
Table 2. The comparison rates of sequencing data.
| Sample | Total reads (QC-passed reads + QC-failed reads) | Total_Mapped | Paired in sequencing | Properly paired |
|---|---|---|---|---|
| A1 | 38,189,373 | 37,250,017 (97.54%) | 34,841,774 | 32,779,226 (94.08%) |
| A2 | 36,249,979 | 35,312,925 (97.42%) | 32,834,168 | 30,992,060 (94.39%) |
| A3 | 33,247,203 | 32,343,333 (97.28%) | 30,185,612 | 28,453,944 (94.26%) |
| B1 | 42,121,797 | 41,038,959 (97.43%) | 38,074,464 | 35,963,012 (94.45%) |
| B2 | 49,218,402 | 48,006,319 (97.54%) | 44,465,284 | 42,209,110 (94.93%) |
| B3 | 48,178,058 | 46,985,237 (97.52%) | 43,363,508 | 41,037,570 (94.64%) |
| C1 | 34,830,304 | 33,954,814 (97.49%) | 31,270,076 | 29,542,138 (94.47%) |
| C2 | 32,388,318 | 31,552,368 (97.42%) | 29,264,886 | 27,571,064 (94.21%) |
| C3 | 35,877,087 | 34,930,317 (97.36%) | 31,650,914 | 29,804,490 (94.17%) |
| D1 | 37,838,969 | 36,926,276 (97.59%) | 33,991,952 | 32,113,516 (94.47%) |
| D2 | 43,346,165 | 42,289,997 (97.56%) | 38,448,072 | 36,391,354 (94.65%) |
| D3 | 40,616,850 | 39,661,674 (97.65%) | 36,673,694 | 34,809,260 (94.92%) |
A: 786-O cell samples; B: 786-O cell samples + 10 mM metformin; C: 786-O cell samples + 25 μM everolimus; D: 786-O cell samples + 10 mM metformin + 25 μM everolimus. QC, quality control.
PI3K-Akt signaling pathway related DEGs were obtained
In total, 3,030 DEG1 were found, of which 1,557 were up-regulated genes and 1,473 were down-regulated genes (Figure 2A). Besides, 2,953 DEG2 (1,511/1,442 up-/down-regulated genes), 3,591 DEG3 (1,781/1,810 up-/down-regulated genes), 1,571 DEG4 (741/830 up-/down-regulated genes), 4,064 DEG5 (2,103/1,961 up-/down-regulated genes) were identified between groups C-A, D-B, D-C and D-A (Figures S1-S4). Functional and pathway enrichment analysis of DEG4 identified a total of 1,160 significant GO terms and 337 KEGG pathways, with top 5 most significant GO terms being DNA-templated DNA replication, DNA replication, cell cycle DNA replication, nuclear DNA replication, and double-strand break repair, while notable KEGG pathways included the PI3K-Akt signaling pathway, human papillomavirus infection and MAPK signaling pathway. In addition, GO analysis revealed that the common biological function of DEG3, DEG4 and DEG5 was DNA replication, DNA-templated DNA replication and cell cycle DNA replication (Figure 2B). Besides the PI3K-Akt signaling pathway, other pathways like the MAPK signaling pathway were commonly enriched in DEG1–5 for KEGG (Figure 2C).
Figure 2.
Screening of differentially expressed genes and their functional enrichment. (A) Volcano map and heatmap of differentially expressed genes. Each point in the graph represents a gene. Red represents up-regulated differentially expressed genes, and blue represents down-regulated differentially expressed genes. (B) GO enrichment analysis of differentially expressed gene. (C) Bubble map of the DEG KEGG (A: 786-O cell samples; B: 786-O cell samples + 10 mM metformin). DEG, differentially expressed gene; FC, fold change; mRNA, messenger RNA; ncRNA, non-coding RNA; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; rRNA, ribosomal RNA.
SPC25, NCAPH, MCM10, UHRF1 and SMC4 were recorded as hub genes
The intersection of DEG1–5 yielded 76 target genes, including 48 up-regulated genes and 28 down-regulated genes in KIRC (Figure 3A). A PPI network consisting of 36 genes and 59 edges were constructed, revealing ACTB, SPC25, KIF20B and so on target genes interacted with each other frequently (Figure 3B). By MCC, SPC25, NCAPH, MCM10, UHRF1 and SMC4 were screened, which were related to mitotic cell cycle, cell cycle phase and cell cycle process, etc. (Figure 3C). Moreover, SPC25, NCAPH, GTSE1, UHRF1, MCM10, KIF20B, SMC4 were selected after MCODE, and they were correlated with chromosome segregation, mitosis, and M phase of mitotic cell cycle, etc. (Figure 3D). By overlapping genes obtained by MCC and MCODE, five genes (SPC25, NCAPH, MCM10, UHRF1, SMC4) were selected as hub genes (Figure 3E). In both sequencing data and TCGA-KIRC, five hub genes were all differentially expressed among groups A–D and shared the same expression tendency between KIRC and normal patients, where target genes were all least expressed in group D (Figure 4A). Further, patients in high expression groups of different target genes had a lower survival probability (P<0.05) (Figure 4B). Thus, five genes were noted as hub genes in this study.
Figure 3.
Identification of candidate genes. (A) Get intersecting genes. The upper panel shows the intersection up gene; the lower panel shows the intersection down gene. (B) Candidate gene PPI network. (C) Top five gene network and enrichment results in the MCC algorithm. (D) PPI network and enrichment results for MCODE screened genes. (E) PPI intersection genes for MCODE and MCC methods. DEG, differentially expressed gene; MCC, Maximal Clique Centrality; MCODE, Molecular Complex Detection; PPI, protein-protein interaction.
Figure 4.
Identification of candidate genes. (A) Expression analysis of candidate genes in TCGA. Group A: 786-O cell samples; group B: 786-O cell samples + 10 mM metformin; group C: 786-O cell samples + 25 µM everolimus; group D: 786-O cell samples + 10 mM metformin + 25 µM everolimus. (B) Survival analysis of the hub gene. ****, P<0.0001. TCGA, The Cancer Genome Atlas.
TF-hub gene network and molecular docking analysis
In the TF-hub gene regulatory network, ten TFs regulated two hub genes, for instance, IRF1 regulated UHRF1 and NCAPH, ELK1 regulated SMC4 and MCM10. Only KLF16 regulated three hub genes (NCAPH, MCM10, SMC4) (Figure 5A). The binding energy of metformin with SPC25 and everolimus with SMC4 were lower which were −5.8 and −8.7 kcal/mol, showing the strong compound-hub gene binding affinity (Figure 5B, Table 3).
Figure 5.
Network construction and molecular docking of hub genes. (A) Construction of the TF-hub gene relationship network. (B) 2D structure of the drug and the hub gene docked to the active ingredient molecule. On the left is metformin and on the right is everolimus. The left side is the global view, and the right side is the local view. 2D, two-dimensional; TF, transcription factors.
Table 3. The binding energy of drugs and hub genes.
| Drug | Gene | Molecular binding energy (kcal/mol) |
|---|---|---|
| Metformin | SPC25 | −5.8 |
| Metformin | NCAPH | −4.8 |
| Metformin | MCM10 | −4.6 |
| Metformin | UHRF1 | −3.8 |
| Metformin | SMC4 | −4.7 |
| Everolimus | SPC25 | −8.3 |
| Everolimus | NCAPH | −7.4 |
| Everolimus | MCM10 | −8.1 |
| Everolimus | UHRF1 | −6.2 |
| Everolimus | SMC4 | −8.7 |
Hub genes were positively correlated with cell cycle and P13K/Akt/mTOR signaling pathway
In the result of GSEA, hub genes were all negatively correlated with lysosome and positively associated with cell cycle and spliceosome (Figure 6). There were 6, 15, 13, 15, and 10 differential pathways found between groups B-A, C-A, D-A, D-B and D-C, respectively (Figure 7A). In the GSVA analysis of differential pathways between all the different drug treatment groups, the HALLMARK_PI3K_AKT_mTOR _SIGNALING pathway scores higher in group A between B and A, C and A, and D and A. It scored higher in groups B and D between D and B, and D and C, separately. P13K/Akt/mTOR signaling pathway was significantly positively correlated with hub genes (Figure 7B).
Figure 6.
Hub gene enrichment analysis. Horizontal coordinates were genes sorted by correlation coefficients, the top part of the vertical coordinates are enrichment scores, the second part has no practical significance, and different colors represent different pathways.
Figure 7.
Hub gene enrichment analysis. (A) GSVA enrichment analysis. Enrichment scores of different pathways in different samples between different groups. Darker colors represent higher enrichment (group A: 786-O cell samples; group B: 786-O cell samples + 10 mM metformin; group C: 786-O cell samples + 25 µM everolimus; group D: 786-O cell samples + 10 mM metformin + 25 µM everolimus). (B) Correlation of hub genes with enrichment scores of different pathways. Blue represented a negative correlation, and red represented a positive correlation. The deeper the color, the stronger the correlation. DEG, differentially expressed gene; GSVA, gene set variation analysis.
Expression validation of hub genes
The melting curves demonstrated decent primers specificity (Figure S5). The expressions of five hub genes (SPC25, NCAPH, MCM10, UHRF1 and SMC4) were significantly higher in the control group than in the other three groups (Figure 8). The expression trend of only MCM10 was consistent with bioinformatics results. Besides, the expressions of SPC25, NCAPH, MCM10 and UHRF1 in the metformin group and everolimus group were significantly higher than those in the combination group. SMC4’s expression in the combination group was not significantly different from single drug groups.
Figure 8.
Expression results of five hub genes. (A) MCM10; (B) NCAPH; (C) SMC4; (D) SPC25; (E) UHRF1. Data are shown as mean ± SD. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. Eve, everolimus; Met, metformin; ns, no significant difference; SD, standard deviation.
Discussion
RCC is a prevalent malignancy accounting for approximately 90% of all kidney cancers, posing significant health challenges globally. The incidence of RCC has been steadily increasing, with an estimated 430,000 new cases diagnosed, leading to approximately 180000 deaths in 2020 (29). Current therapeutic strategies for patients with advanced KIRC or cancer recurrence, including targeted therapies and immunotherapy, have shown promise; however, resistance to these therapies remains a key issue. Exploring novel therapies is crucial for improving treatment efficacy and addressing resistance mechanisms. Therefore, in this study, SPC25, NCAPH, MCM10, UHRF1, and SMC4 were found to be the hub genes for the combination of metformin and everolimus for the treatment of KIRC by RNA-seq, molecular biological methods and bioinformatics analysis. The results from our experiments lay the groundwork for further exploration of the molecular mechanisms underlying this combination therapy and its implications for personalized treatment strategies in KIRC.
SPC25 serves as a critical protein that facilitates chromosomal alignment and spindle formation during meiosis (30), thereby maintaining genetic stability. Consequently, abnormal levels of SPC25 expression have been linked to various oncogenic processes, including tumorigenesis, tumor progression, metastasis, and stemness (31,32). Recent studies have reported the high expression of SPC25 in KIRC and predicted poor survival (33,34). As a result, SPC25 is anticipated to emerge as a promising therapeutic target for RCC.
NCAPH, situated on chromosome 2q11.2, serves as a crucial regulatory component of condensing I (35). It plays a significant role in maintaining chromosome stability and morphology, managing DNA damage, and regulating mitotic kinetics in normal cells (36). A previous study showed a significant correlation between the expression level of NCAPH and overall survival and recurrence-free survival time in KIRC (37). In addition, Chen et al. reported that NCAPH can increase the expression of PD-L1 by inhibiting the degradation of β-catenin in KIRC cells, thus further promoting the aerobic glycolysis and immune tolerance of KIRC (38). MCM10 is crucial for preserving the stability and elongation of DNA replication forks throughout the process of DNA replication. A previous study suggested that there is an increase in MCM10 expression in KIRC, and upregulation is closely linked to clinical staging and immune cell infiltration of RCC (39).
UHRF1 is categorized within the nucleoprotein UHRF family and is crucial for the preservation of DNA methylation, playing significant roles in various biological processes, including cell growth, regulation of cell cycle, and apoptosis (40,41). Research indicates that UHRF1 exhibits oncogenic properties and has a strong correlation with the progression and prognosis of KIRC patients. It is suggested that UHRF1 may facilitate the progression of RCC through the epigenetic modulation of TXNIP and may act as a p53 regulator, inhibiting p53 activation and enabling RCC cells to evade p53-dependent apoptosis (42-44). SMC4, a member of SMC family, which is located in 3q25.33. SMC4 regulates chromosome assembly and segregation (45). A previous study reported that SMC4 is overexpressed in various tumors and is involved in tumorigenesis and progression (46). These results found that SMC4 was highly expressed in KIRC and positively correlated with prognosis. The use of metformin and everolimus both decreased the expression of SMC4, but there was no significant difference between the combination and the use of the single agent. The above study demonstrated that the expression level of SMC4 influences the development of KIRC, but the specific gene function still needs further experimental verification.
By constructing the TF-hub gene regulatory network, we predicted that 10 TFs could regulate two key genes while only one TF (KLF16) could regulate three. Studies have shown that the expression of KLF16 is upregulated in KIRC and is associated with poor prognosis. The KLF16 regulates circular RNA forkhead box O3 (cirFOXO3). The expression of circFOXO3 in cells regulates the growth of ccRCC and the cytotoxic activity of natural killer cells by sponges miR-29a-3p and miR-122-5p (47,48). In addition, through molecular docking, we found that the strongest binding ability to the two drugs was SPC25 and SMC4, respectively. However, whether these two drugs can affect KIRC by regulating the transcriptional and translational processes of SPC25 or SMC4 remains to be further investigated.
In the GSVA analysis, the combination treatment exhibited higher enrichment of the PI3K/Akt/mTOR pathway compared to the everolimus group. This discrepancy may arise because everolimus, as an mTOR inhibitor, directly inhibits mTOR signaling, whereas metformin exerts anticancer effects through multiple pathways that interact with each other. mTOR is mediated by mTORC1 and mTORC2. A negative feedback mechanism exists between mTORC1 and mTORC2: increased mTORC1 activity upregulates mTORC2, which further activates Akt (49). A study demonstrated that metformin activates mTORC2/Akt via the AMPK/mTORC1/mTORC2 feedback loop (50). Further GSEA analysis of these key genes revealed that they mainly enriched in the cell cycle, and the PI3K-Akt-mTOR pathway was significantly different between all the different drug treatment groups. This is pivotal in regulating cell proliferation, survival, and metabolism. The activation of the PI3K-Akt-mTOR pathway has been implicated in various malignancies, including RCC, where it contributes to tumor growth and resistance to therapy (51). The interaction between metformin and everolimus could interfere with these signaling pathways, potentially enhancing the antitumor effect.
In this study, five hub genes were obtained through bioinformatics analysis and expression validation. It provides an important reference for enriching the therapeutic choices of KIRC and improving patient outcomes. However, there are some limitations to this study. First, the small sample size may affect the generalizability of our findings. Additionally, the absence of long-term clinical follow-up data restricts our understanding of the prognostic significance of these findings in clinical practice. Therefore, we plan to expand the sample size and conduct additional in vitro and in vivo experiments to validate the synergistic effects of the drug combination. Furthermore, clinical studies will be incorporated to further explore its clinical applicability.
Conclusions
In conclusion, metformin combined with everolimus exhibited a synergic anticancer effect on RCC. Five hub genes (SPC25, NCAPH, MCM10, UHRF1, SMC4) for metformin combined with everolimus for the treatment of KIRC were obtained based on transcriptomic data and expression validation identification and were co-enriched in the PI3K-Akt-mTOR pathway. These findings are important for developing novel therapeutic approaches for KIRC and improving patient survival.
Supplementary
The article’s supplementary files as
Acknowledgments
We acknowledge TCGA databases for providing their platforms and data, and we thank the contributors who upload their datasets for public access and use.
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with Declaration of Helsinki and its subsequent amendments.
Footnotes
Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-277/rc
Funding: None.
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-277/coif). The authors have no conflicts of interest to declare.
Data Sharing Statement
Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-277/dss
References
- 1.Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin 2024;74:12-49. 10.3322/caac.21820 [DOI] [PubMed] [Google Scholar]
- 2.Rose TL, Kim WY. Renal Cell Carcinoma: A Review. JAMA 2024;332:1001-10. 10.1001/jama.2024.12848 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bukavina L, Bensalah K, Bray F, et al. Epidemiology of Renal Cell Carcinoma: 2022 Update. Eur Urol 2022;82:529-42. 10.1016/j.eururo.2022.08.019 [DOI] [PubMed] [Google Scholar]
- 4.Ivanyi P, Fröhlich T, Grünwald V, et al. The Treatment of Metastatic Renal Cell Carcinoma. Dtsch Arztebl Int 2024;121:576-86. 10.3238/arztebl.m2024.0147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Barata PC, Rini BI. Treatment of renal cell carcinoma: Current status and future directions. CA Cancer J Clin 2017;67:507-24. 10.3322/caac.21411 [DOI] [PubMed] [Google Scholar]
- 6.Motzer RJ, Escudier B, Oudard S, et al. Efficacy of everolimus in advanced renal cell carcinoma: a double-blind, randomised, placebo-controlled phase III trial. Lancet 2008;372:449-56. 10.1016/S0140-6736(08)61039-9 [DOI] [PubMed] [Google Scholar]
- 7.Ahmad E, Sargeant JA, Zaccardi F, et al. Where Does Metformin Stand in Modern Day Management of Type 2 Diabetes? Pharmaceuticals (Basel) 2020;13:427 . 10.3390/ph13120427 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Flory J, Lipska K. Metformin in 2019. JAMA 2019;321:1926-7. 10.1001/jama.2019.3805 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Matthews DR, Paldánius PM, Proot P, et al. Glycaemic durability of an early combination therapy with vildagliptin and metformin versus sequential metformin monotherapy in newly diagnosed type 2 diabetes (VERIFY): a 5-year, multicentre, randomised, double-blind trial. Lancet 2019;394:1519-29. 10.1016/S0140-6736(19)32131-2 [DOI] [PubMed] [Google Scholar]
- 10.Schernthaner G, Schernthaner GH. The right place for metformin today. Diabetes Res Clin Pract 2020;159:107946 . 10.1016/j.diabres.2019.107946 [DOI] [PubMed] [Google Scholar]
- 11.Bashraheel SS, Kheraldine H, Khalaf S, et al. Metformin and HER2-positive breast cancer: Mechanisms and therapeutic implications. Biomed Pharmacother 2023;162:114676 . 10.1016/j.biopha.2023.114676 [DOI] [PubMed] [Google Scholar]
- 12.Chen YH, Wang PH, Chen PN, et al. Molecular and Cellular Mechanisms of Metformin in Cervical Cancer. Cancers (Basel) 2021;13:2545 . 10.3390/cancers13112545 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Eibl G, Rozengurt E. Metformin: review of epidemiology and mechanisms of action in pancreatic cancer. Cancer Metastasis Rev 2021;40:865-78. 10.1007/s10555-021-09977-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kamarudin MNA, Sarker MMR, Zhou JR, et al. Metformin in colorectal cancer: molecular mechanism, preclinical and clinical aspects. J Exp Clin Cancer Res 2019;38:491 . 10.1186/s13046-019-1495-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gulati S, Desai J, Palackdharry SM, et al. Phase 1 dose-finding study of metformin in combination with concurrent cisplatin and radiotherapy in patients with locally advanced head and neck squamous cell cancer. Cancer 2020;126:354-62. 10.1002/cncr.32539 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Saif MW, Rajagopal S, Caplain J, et al. A phase I delayed-start, randomized and pharmacodynamic study of metformin and chemotherapy in patients with solid tumors. Cancer Chemother Pharmacol 2019;84:1323-31. 10.1007/s00280-019-03967-3 [DOI] [PubMed] [Google Scholar]
- 17.Crist M, Yaniv B, Palackdharry S, et al. Metformin increases natural killer cell functions in head and neck squamous cell carcinoma through CXCL1 inhibition. J Immunother Cancer 2022;10:e005632 . 10.1136/jitc-2022-005632 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hong SH, Lee KS, Hwang HJ, et al. Synergic Effect of Metformin and Everolimus on Mitochondrial Dynamics of Renal Cell Carcinoma. Genes (Basel) 2022;13:1211 . 10.3390/genes13071211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Vitali E, Boemi I, Tarantola G, et al. Metformin and Everolimus: A Promising Combination for Neuroendocrine Tumors Treatment. Cancers (Basel) 2020;12:2143 . 10.3390/cancers12082143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chen YH, Wu JX, Yang SF, et al. Metformin Potentiates the Anticancer Effect of Everolimus on Cervical Cancer In Vitro and In Vivo. Cancers (Basel) 2021;13:4612 . 10.3390/cancers13184612 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Soliman PT, Westin SN, Iglesias DA, et al. Everolimus, Letrozole, and Metformin in Women with Advanced or Recurrent Endometrioid Endometrial Cancer: A Multi-Center, Single Arm, Phase II Study. Clin Cancer Res 2020;26:581-7. 10.1158/1078-0432.CCR-19-0471 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Trivedi UH, Cézard T, Bridgett S, et al. Quality control of next-generation sequencing data without a reference. Front Genet 2014;5:111 . 10.3389/fgene.2014.00111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 2014;30:2114-20. 10.1093/bioinformatics/btu170 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Perelo LW, Gabernet G, Straub D, et al. How tool combinations in different pipeline versions affect the outcome in RNA-seq analysis. NAR Genom Bioinform 2024;6:lqae020 . 10.1093/nargab/lqae020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 2014;30:923-30. 10.1093/bioinformatics/btt656 [DOI] [PubMed] [Google Scholar]
- 26.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014;15:550 . 10.1186/s13059-014-0550-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yu G, Wang LG, Han Y, et al. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284-7. 10.1089/omi.2011.0118 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ritchie ME, Phipson B, Wu D, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015;43:e47 . 10.1093/nar/gkv007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. 10.3322/caac.21660 [DOI] [PubMed] [Google Scholar]
- 30.Liu XS, Zhang Y, Ming X, et al. SPC25 as a novel therapeutic and prognostic biomarker and its association with glycolysis, ferroptosis and ceRNA in lung adenocarcinoma. Aging (Albany NY) 2024;16:779-98. 10.18632/aging.205418 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cui F, Tang H, Tan J, et al. Spindle pole body component 25 regulates stemness of prostate cancer cells. Aging (Albany NY) 2018;10:3273-82. 10.18632/aging.101631 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Yang J, Huang Y, Song M, et al. SPC25 promotes proliferation and stemness of hepatocellular carcinoma cells via the DNA-PK/AKT/Notch1 signaling pathway. Int J Biol Sci 2022;18:5241-59. 10.7150/ijbs.71694 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Xia F, Yang H, Wu H, et al. Spindle component 25 predicts the prognosis and the immunotherapy response of cancers: a pan-cancer analysis. Sci Rep 2024;14:8452 . 10.1038/s41598-024-59038-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wang Y, Wang H, Xu Y, et al. Mefloquine Suppresses Metastasis in Renal Cell Carcinoma Through Targeting SPC25. Cancer Sci 2025;116:1239-54. 10.1111/cas.70001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Neuwald AF, Hirano T. HEAT repeats associated with condensins, cohesins, and other complexes involved in chromosome-related functions. Genome Res 2000;10:1445-52. 10.1101/gr.147400 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Liu C, Han X, Zhang S, et al. The role of NCAPH in cancer treatment. Cell Signal 2024;121:111262 . 10.1016/j.cellsig.2024.111262 [DOI] [PubMed] [Google Scholar]
- 37.Wei W, Lv Y, Gan Z, et al. Identification of key genes involved in the metastasis of clear cell renal cell carcinoma. Oncol Lett 2019;17:4321-8. 10.3892/ol.2019.10130 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Chen Z, Ruan W, Guo C, et al. Non-SMC condensin I complex subunit H participates in anti-programmed cell death-1 resistance of clear cell renal cell carcinomas. Cell Prolif 2023;56:e13400 . 10.1111/cpr.13400 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Chen D, Zhong N, Guo Z, et al. MCM10, a potential diagnostic, immunological, and prognostic biomarker in pan-cancer. Sci Rep 2023;13:17701 . 10.1038/s41598-023-44946-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Tien AL, Senbanerjee S, Kulkarni A, et al. UHRF1 depletion causes a G2/M arrest, activation of DNA damage response and apoptosis. Biochem J 2011;435:175-85. 10.1042/BJ20100840 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Bostick M, Kim JK, Estève PO, et al. UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science 2007;317:1760-4. 10.1126/science.1147939 [DOI] [PubMed] [Google Scholar]
- 42.Jiao D, Huan Y, Zheng J, et al. UHRF1 promotes renal cell carcinoma progression through epigenetic regulation of TXNIP. Oncogene 2019;38:5686-99. 10.1038/s41388-019-0822-6 [DOI] [PubMed] [Google Scholar]
- 43.Luo Y, Shen D, Chen L, et al. Identification of 9 key genes and small molecule drugs in clear cell renal cell carcinoma. Aging (Albany NY) 2019;11:6029-52. 10.18632/aging.102161 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Ma J, Peng J, Mo R, et al. Ubiquitin E3 ligase UHRF1 regulates p53 ubiquitination and p53-dependent cell apoptosis in clear cell Renal Cell Carcinoma. Biochem Biophys Res Commun 2015;464:147-53. 10.1016/j.bbrc.2015.06.104 [DOI] [PubMed] [Google Scholar]
- 45.Losada A, Hirano T. Dynamic molecular linkers of the genome: the first decade of SMC proteins. Genes Dev 2005;19:1269-87. 10.1101/gad.1320505 [DOI] [PubMed] [Google Scholar]
- 46.Zhao Z, Wang X, Ding Y, et al. SMC4, a novel tumor prognostic marker and potential tumor therapeutic target. Front Oncol 2023;13:1117642 . 10.3389/fonc.2023.1117642 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Fu M, Du Y, Liu F, et al. Prognostic value of KLFs family genes in renal clear cell carcinoma. Sci Rep 2024;14:20204 . 10.1038/s41598-024-69892-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Yang F, Chen Y, Luo L, et al. circFOXO3 Induced by KLF16 Modulates Clear Cell Renal Cell Carcinoma Growth and Natural Killer Cell Cytotoxic Activity through Sponging miR-29a-3p and miR-122-5p. Dis Markers 2022;2022:6062236 . 10.1155/2022/6062236 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Lu G, Wu Z, Shang J, et al. The effects of metformin on autophagy. Biomed Pharmacother 2021;137:111286 . 10.1016/j.biopha.2021.111286 [DOI] [PubMed] [Google Scholar]
- 50.Ling S, Song L, Fan N, et al. Combination of metformin and sorafenib suppresses proliferation and induces autophagy of hepatocellular carcinoma via targeting the mTOR pathway. Int J Oncol 2017;50:297-309. 10.3892/ijo.2016.3799 [DOI] [PubMed] [Google Scholar]
- 51.Glaviano A, Foo ASC, Lam HY, et al. PI3K/AKT/mTOR signaling transduction pathway and targeted therapies in cancer. Mol Cancer 2023;22:138 . 10.1186/s12943-023-01827-6 [DOI] [PMC free article] [PubMed] [Google Scholar]