A total of 10 cancer and adjacent tissue samples from LUAD patients (admitted to the hospital during March 2021 and 2022) were used in the analysis. These tissues were obtained during tumor resection surgery and obtained the samples from 7 male and 3 female patients (median age, 59 years; range, 35‒83 years). Of these, 2 had stage I + II cancers while 8 had stage III + IV. Before the study, patients were provided written consent. The study was approved by Ethics Committee of Shandong Provincial Third Hospital (DWKYLL-2021039).

The human bronchial epithelial cell line (16 HBE) and lung cancer cells (NCI-H1975, H1299, H23, A549) were purchased from Procell Life Science & Technology Co., Ltd., (Wuhan, China) and grown in RPMI-1640 media (Gibco, Carlsbad, CA, USA) with 10% fetal bovine serum (FBS, Gibco, USA). Cells were maintained at 37°C with 5% CO₂. RiboBio (Guangzhou, China) synthesized siRNA molecules targeting STIL (si-STIL-1 and si-STIL-2), E2 promoter binding factor 1-1 (E2F1-1), E2F1-2, and a negative control (si-NC). The amplified forms of STIL were inserted into the pc-DNA3.1 Vector from Tsingke Biotechnology Co., Ltd., in Beijing, China. The sequences were as follow: si-STIL-1: Sense, 5′ GAAUAUUUCUCAAGUUCAAGG-3′; Anti-sense, 5′ UUGAACUUGAGAAAUAUUCAG-3′; si-STIL-2: Sense, 5′-GAACUUUGUUCAAGAGAAAUG-3′; Anti-sense, 5′-UUUCUCUUGAACAAAGUUCUU-3′; si-E2F1-1: Sense, 5′-GAUGGUUAUGGUGAUCAAAGC-3′; Anti-sense, 5′-UUUGAUCACCAUAACCAUCUG-3′; si-E2F1-2: Sense, 5′-CACUGAAUCUGACCACCAAGC-3′; Anti-sense, 5′-UUGGUGGUCAGAUUCAGUGAG-3′; si-NC: Sense, 5′-UUCUCCGAACGUGUCACGUTT-3′; Anti-sense, 5′-ACGUGACACGUUCGGAGAATT-3′.

Logarithmic growth phase H1299 and H23 cells were plated in 6-well plates (Gibco) and grown for 24 h before transfection. The si-NC, si-STIL-1, si-STIL-2, and STIL overexpression vector were tansfected using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA). The transfection effectiveness was determined by Western blotting. Cells in si-STIL-1 + Z-VAD-FMK group were treated with 10 µM Z-VAD-FMK (Sigma-Aldrich, St. louis, MO, USA) for 24 h after transfection.

Cells were lysed with RIPA buffer (E-BC-R327, Elabscience, Wuhan, China) for protein extraction. Proteins (20 μg) were electrophoresed on SDS-PAGE and transferred to PVDF membranes which were treated with antibodies against STIL (ab89314, Abcam, Cambridge, UK), E-Cadherin (ab231303, Abcam), N-Cadherin (ab76011, Abcam), Vimentin (A2584, Abclonal, Wuhan, China), E2F1 (ab288369, Abcam), Hexokinase II (HK-2, ab209847, Abcam), glucose transporter 1 (Glut-1, A6982, Abclonal), Lactate dehydrogenase A (LDHA, #3582, Cell Signalling, Technology, Danvers, MA, USA), cyclin-dependent kinase 4 (CDK4, ab108357, Abcam), cyclin E1 (CCNE1, ab33911, Abcam), cyclin D1 (CCND1, ab16663, Abcam), and GAPDH (ab181603, Abcam) at 4°C overnight. It was then incubated at room temperature for 2 h with the secondary antibody (ab7090 and ab96879, Abcam). The chemiluminescent reagent Vazyme, Nanjing, China was added onto the PVDF membrane. The membrane was scanned and analyzed using the Tanon 4800 image system (Beijing Yuanpinghao Biotechnology Co., Ltd, Beijing, China), and the grayscale valure of the bands used to assess relative expression. Assessment of protein levels was performed 48 h following transfection.

All respective cell lines were allowed to grow in 96-well plates (Beyotime Biotechnology, Shanghai, China), with 3 replicate wells in each group. A growth medium (100 μL/well) was placed in each well and incubated in 5% CO₂ for 24 h at 37°C. A volume of 10 μL/well of CCK-8 detection reagent (Dojindo, Kumamoto, Japan) was then added. Theabsorbance of different time (0, 24, 48, 72, and 96 h) were detected at 450 nm via a microplate reader (BIO-TEK, Biotek Winooski, Vermont, USA).

Cells were allowed to grow in 24-well plates (Thermo Fisher Scientific, Waltham, MA, USA) for 24 h. Each well was enriched with EdU medium (100 μL, 50 μmol/L, Beyotime Biotechnology) for 2 h after which cells were placed in paraformaldehyde (PFA, 4%) and 0.5% Triton X-100 solution for fixation followed by EdU staining. Nuclei were labeled with DAPI (Beyotime Biotechnology). All fluorescence microscopic images were quantified for the detection of the proportion of EdU-positive cells.

Cells were trypsinized, rinsed with PBS, and resuspended in Binding buffer to 1 × 10⁶ cells/mL. After adding Annexin V-FITC (BD, San Jose, USA), the sample was kept for 15 min at 25°C (dark). Next, PI was added for 5 min. Results were quantified and analyzed via flow cytometry (BD) and Cell Quest software (BD).

For cell cycle detection, H1299 and H23 cells were fixed in ice-cold 70% ethanol at 4°C for 24 h, washed with PBS, incubated with a staining solution containing PI (Sigma-Aldrich, USA) and RNase A (Sigma-Aldrich) at 4°C for 30 min. Samples were loaded and assessed using a FACS Calibur flow cytometer (BD Biosciences, USA).

H1299 and H23 cells were plated in 6-well plates and grown to monolayers after which the cell surface was scratched vertically with a sterilized tip. The detached or free-floating cells were then rinsed with PBS before the serum-free medium (SFM) was added. They were kept at 37°C with 5% CO₂. The images were collected at 0 and 24 h.

The dissolved Matrigel gel (BD) was combined with an serum free medium and distributed uniformly across the chamber of the transwell (upper chamber surface). To turn the gel into gelatin, it was allowed to solidify at 37°C for approximately 2 h. A specific volume of the serum-free culture medium was diluted before digesting and counting the H1299 and H23 cell lines. Cells were inoculated uniformly to the upper portion of the transwell chamber, the chamber was transferred to a plate containing 10% FBS medium. The transwell cells were removed after 24 h of cell culture. Any remaining matrigel adhesive was removed using cotton swabs, and the wells were rinsed thrice with PBS. After PFA fixation and crystal violet staining, the cells migrating the the bottom of the chamber were examined by counting analysis.

Initial steps involved cloning the sequence of the wild-type or mutant STIL promoter into the pGL3-luciferase reporter construct. After this, HEK-293T cells were co-transfected with pcDNA3.1-Vector (Vector, Tsingke Biotechnology Co., Ltd.) or E2F1 (Tsingke Biotechnology Co., Ltd.) via Lipofectamine3000 (Invitrogen) for 48 h before assessement of luciferase activity assay via a Promega luciferase activity kit.

Total RNA was extracted with TRIzol (Invitrogen) and reverse-transcribed to cDNA with a PrimeScript RT Reagent Kit (TaKaRa, Tokyo, Japan). qRT-PCR was conducted using a SYBR Green PCR Kit (Qiagen, Duesseldorf, Germany), on a LightCycler 480 II (Roche, Switzerland).

Beijing Tsingke Biotech supplied primers [STIL, F primer: 5′-CTACCCTGCAAACAGACCTCA-3′, R primer: 5′-AAAGGATATATAGGCTCCATG-3′; E2F1, F primer: 5′-GGGACTTTGCAGGCAGCGGCG-3′, R primer: 5′-GACGATCTGCGAGGAGTCGAG-3′; GAPDH (internal control), F primer: 5′-CTGCCGGTGACTAACCCTGCG-3′, R primer: 5′-GCCCAATACGACCAAATCAGA-3′]. Expression was determined via the 2^−ΔΔCt method.

This experiment was conducted using ChIP commercial kit (cat:53008; Active Motif, Carlsbad, CA, USA). Cells were cultured into a 10 cm² plate and allowed to proliferate until reaching a state of 70%–80% confluency. They were fixed and collected for further experiments. All nuclear precipitates were re-suspended in ChIP buffer for ultrasonic treatment and incubated for 24 h with a 5 μg antibody. The DNA protein complex is cross-linked and eluted through a series of washes. The DNA was prepared in TE buffer for analysis. We used RT-qPCR to evaluate STIL’s relative enrichment after the ChIP assay. The primer sequences were: F:5′-CGCCCCGGAACTTAACCGTTG-3′; R:Primer: 5′-TCCGCGCTCGACCAATCCCAA-3′.

Glucose uptake, lactate levels, and ATP/ADP ratios were detected by Glucose Uptake Assay Kit (J1343, Promega, Beijing, China), Lactate Assay Kit (L256, Dojindo), and ADP/ATP Ratio Assay Kit (A552, Dojindo), respectively.

The oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were observed via the Seahorse XF96 Extracellular Flux Analyzer. (Agilent, USA), in a assay medium (Agilent, Santa Clara, CA, USA) within a non-CO₂ incubator for 1 h under basal conditions.

The GEPIA website (http://gepia.cancer-pku.cn/index.html) was used to evaluate differences in STIL and E2F1 levels between normal lung tissue and LUAD tissue, and the effect of STIL and E2F1 on the prognosis of LUAD. Clinical data were analyzed by Cox risk regression using the survival package in R (Version 3.4.1). The relationship between STIL and E2F1 in The Cancer Genome Atlas Lung Adenocarcinoma (TCGA-LUAD) was tested by Spearman correlation. Kaplan-Meier curves with log-rank tests were used to assess survival, and then the expression of STIL and clinical indicators in TCGA-LUAD were analyzed using univariate Cox regression. LASSO regression was performed using the glmnet package method. For potential predictive factors, a Cox model was employed for estimating hazard ratios (HRs) with confidence intervals (CIs). The pROC package in R was used for drawing ROC curves. The metadata for TCGA-LUAD patients was provided in Table S1.

Gene set enrichment analysis (GSEA) was used to analyze the transcription factor for STIL based on the TCGA-LUAD database (ENCODE_TF_ChIP-seq_2015.txt). The JASPAR software (https://jaspar.genereg.net/) was used for predicting potential transcription factor interaction sites in the STIL promoter. The association between STIL and E2F1 was examined using R software. GSEA was also used to study the relationship between LUAD-related biological pathways and STIL expression based on the TCGA-LUAD database (h.all.v2023.1.Hs.entrez.gmt).

All results were analyzed via GraphPad Prism 5.0 (GraphPad Software, La Jolla, CA, USA). The data are expressed as the mean ± standard deviation (SD). Data were compared using a two independent sample t-test, and the rates were evaluated via χ² test or Fisher’s exact probability method. Substantial differences were indicated by p < 0.05. Triplicate experiments were conducted for statistical analysis.

We analyzed the expression pattern of the STIL gene in LUAD tissue as compared with the normal lung tissue using the GEPIA database. The results showed that STIL expression was markedly upregulated in LUAD tissue relative to normal lung tissue (Fig. 1A). Moreover, STIL levels were significantly related to the overall survival (OS). The OS rate of LUAD patients with the STIL low expression being substantially higher compared with those with the high STIL expression (Fig. 1B). Univariate and multivariate analyses were performed using the TCGA-LUAD dataset (Table 1). Through Lasso regression analysis, we selected variables combining the expression level of STIL with clinical indicators based on data of TCGA-LUAD. Aiming to establish a predictive model, we integrated clinical data from TCGA-LUAD patients, examining a comprehensive set of seven clinical parameters including STIL expression, sex, age, stage, (T_stage, N_stage, M_stage) through Lasso regression analysis (Figs. S1A, S1B). Ultimately, three variables with positive coefficients were deemed significant: STIL expression, age, and cancer stage (Fig. 1C). Using these, a Cox proportional hazards model was constructed with the formula: Hazard Ratio = exp (0.920 × Stage II + 1.113 × Stage III + 1.361 × Stage IV + 0.172 × STIL + 0.016 × age). As shown in Fig. 1C, higher stages (II, III, IV) significantly increased risk ratios (HRs) to 2.51, 3.04, and 3.90, respectively, indicating disease progression’s impact. STIL’s elevated levels were associated with an HR of 1.19 (p = 0.034), signifying its prognostic relevance. Age influenced risk marginally yet significantly (HR of 1.02, p = 0.049). Risk scores calculated from this model stratified patients into high and low-risk groups, revealing marked differences in survival, and underscoring the model’s predictive strength (Fig. S1C). Moreover, our model’s risk score demonstrated excellent prognostic prediction for LUAD patients (AUC = 0.683, Fig. S1D). It was also found that STIL was upregulated in LUAD tissue relative to normal samples (Fig. 1D). Moreover, the level of STIL expression in cancer cells was markedly elevated compared to normal cells (16HBE), particularly in H23 and H1299 cells (Fig. 1E). Collectively, the level of STIL was significantly elevated in LUAD, and increased STIL levels were related with unfavorable prognosis.

As seen in Fig. 2A and Fig. S2A, STIL expression was notably suppressed or elevated by transfection with siRNAs (si-STIL-1 and si-STIL-2) or STIL-overexpressing vectors. H23 and H1299 cell viability in si-STIL-1 and si-STIL-2 cells was substantially lower than si-NC group (Fig. 2B). Furthermore, the knockdown of STIL markedly reduced cell proliferation in both H23 and H1299 cell lines, as depicted in Fig. 2C. In contrast, an increase in the number of EdU upon STIL overexpression was observed in H23 and H1299 cells (Fig. S2B). Moreover, Knockdown of STIL increased cancer cell apoptosis, as shown in Fig. 2D. In contrast, STIL overexpression inhibited apoptosis (Fig. S2C). Collectively, STIL promoted LUAD cell proliferation and restrained cell apoptosis.

A significant reduction in the rate of wound healing was found in both the si-STIL-1 and si-STIL-2 groups, as assessed by wound healing assays (Fig. 3A), and invading cells, as measured by Transwell assays (Fig. 3B). In contrast, STIL overexpression increased the wound healing rate (Fig. S2D). Moreover, the results obtained from the Western blot analysis (Fig. 3C) showed the molecular mechanisms underlying these cellular changes. Markedly increased E-Cadherin levels were observed after STIL knockdown, suggesting an improvement in epithelial characteristics. Simultaneously, knockdown of STIL induced a significant reduction in the levels of N-Cadherin and Vimentin, implying a concurrent decrease in mesenchymal features. In addition, we also found that knockdown of STIL caused cell cycle arrest at the G1 phase (Fig. 3D). Further measurement of the cell cycle-related proteins, CDK4, CCNE1, and CCND1, confirmed the suppression of cell cycle progression under STIL inhibition (Fig. 3E). Collectively, these observed alterations in protein expression substantiate the conclusion that STIL effectively affected the migratory, invasive, and cycle progression potential of LUAD cells.

Based on a single-gene GSEA analysis of TCGA-LUAD using data from “ENCODE_TF_ChIP-seq_2015.txt”, the upstream mechanism of STIL prediction was examined. The results indicate that E2F1 potentially functions as a transcription factor that positively regulates the level of STIL (Figs. 4A, 4B). It was noted that the STIL promoter contains a possible E2F1 binding site upstream of the translation initiation site using the JASPAR software to predict potential transcription factor binding sites (Fig. 4C). Further, GEPIA analysis found that E2F1 was higher expressed in LUAD patients’ samples, and higher E2F1 was correlated with poor prognosis (Figs. 4D, 4E). A positive correlation was observed between E2F1 and STIL expression in LUAD tissues (Fig. 4F). The STIL-E2F1 interaction was verified using dual-luciferase reporter experiment and ChIP assays. The results of the experiment indicated that the luciferase activity of STIL promoter-WT-transfected cells was markedly enhanced by E2F1 upregulation. In contrast, the stimulatory impacts of E2F1 upregulation on luciferase activity were considerably reduced when the binding sequence was mutated. This suggests that E2F1 may interact with the promoter of STIL to increase its expression (Fig. 4G). Relative enrichment of the STIL promoter in the E2F1 group was markedly greater than in the IgG group (Fig. 4H). Further, the down-regulation of E2F1 remarkably inhibited the level of STIL (Fig. 4I). Altogether, E2F1 promoted the STIL expression level. Moreover, the findings of Fig. 4I also confirmed the successful transfection of E2F1 siRNA into H1299 and H23 cells.

Associations between E2F1 and STIL in cell viability and invasion were further studied. Firstly, we verified the expression levels of E2F1 in 16HBE, H1299, and H23. This showed that E2F1 levels were markedly higher in the LUAD cells relative to the normal cells (Fig. 5A). Secondly, we transfected STIL into cells. We found that after transfection, the level of intracellular STIL was markedly up-regulated, suggesting that the transfection was successful (Fig. 5B). As shown in Fig. 5B, down-regulation of E2F1 could significantly reduce the expression level of STIL. In contrast, overexpression of STIL could reverse this reduction. Moreover, relative to the si-NC + Vector group, the knockdown of E2F1 significantly reduced cell viability, invasion, and the levels of cycle-related proteins, while overexpression of STIL could significantly increase cell viability, invasion, and the levels of cycle-related proteins. Meanwhile, compared with the si-E2F1-1 + Vector group, overexpression of STIL counteracted the effects of E2F1 knockdown on cell viability, invasion, and the levels of cycle-related proteins (Figs. 5C–5E). Collectively, the influence of E2F1 on cell viability, invasion ability, and cell cycle was mediated by STIL.

GSEA enrichment analysis, leveraging the TCGA-LUAD database (h.all.v2023.1.Hs.entrez.gmt), identified glycolysis as a significantly associated pathway with STIL expression (Figs. 6A, 6B). Further investigations revealed that downregulation of STIL notably inhibited intracellular glucose uptake, lactate production, and the ATP/ADP ratio (Figs. 6C–6E). Conversely, overexpression of STIL enhanced glycolytic processes, including glucose uptake, lactate levels, and theATP/ADP ratios in H1222 and H23 cells (Figs. S2E–S2G). Additionally, STIL downregulation reduced the protein levels of key glycolytic enzymes HK-2, GLUT-1, and LDHA (Fig. 6F), while their levels increased following STIL overexpression (Fig. S2H). Notably, STIL knockdown significantly diminished the ECAR (Fig. 6G) and elevated the OCR (Fig. 6H), suggesting a shift from aerobic glycolysis towards enhanced aerobic oxidation. To further explore STIL’s regulatory role in glycolysis, we employed the apoptosis inhibitor Z-VAD-FMK. The findings indicated that STIL downregulation continued to decrease the protein levels of HK-2, GLUT-1, and LDHA even after treatment with Z-VAD-FMK (Fig. 6I). The results revealed that the effect of STIL on the expression of key glycolysis enzymes was not caused by apoptosis. Collectively, the results demonstrate the involvement of STIL in regulating glycolysis in LUAD.