The chemical structure and SMILES format of GIT were obtained from PubChem (https://pubchem.ncbi.nlm.nih.gov/). Possible GIT target predictions were made with PharmMapper (http://www.lilab-ecust.cn/pharmmapper/), with conversion to protein symbols with UniProt (https://www.uniprot.org/). LC-associated genes were identified in GeneCards (https://www.genecards.org/) by searching for “laryngeal cancer,” with a correlation score threshold of ≥10. Overlapping targets between GIT and LC were determined with Draw Venn Diagram tool (https://bioinformatics.psb.ugent.be/webtools/Venn/), and the expression of the shared targets in LC tissues was assessed using UALCAN (http://ualcan.path.uab.edu/) (Accessed date: 2024.11.29).

The top 20 common targets, ranked by relevance scores, were selected for further analysis (Supplementary Table S1; Table 1). These core genes were uploaded to the STRING database (https://string-db.org/) to construct the PPI network. Additionally, the GeneMANIA database (http://www.genemania.org/) was used to develop the GGI network (Accessed date: 2024.12.01).

To evaluate the functional relevance of key genes, the top 20 genes were analyzed using the R package GOSemSim [2.22.0]. The 96 overlapping genes and these top 20 genes were then subjected to GO and KEGG analyses were performed using DAVID (https://david.ncifcrf.gov/). The R package ggplot2 [3.4.4] was employed to generate visualization charts for the top three enriched terms in each category (Accessed date: 2024.12.01).

GIT (CAS No.: 511-96-6) was purchased from MedChemExpress (MCE, NJ, United States). Other reagents, including DMSO, CCK-8 kit, staining kits, paraformaldehyde, propidium iodide (PI), TRI reagent, and ECL chemiluminescence reagent, were obtained from Biosharp (Hefei, China). Matrigel, cell lysis buffer, protease and phosphatase inhibitor cocktails, and the BCA protein assay kit were sourced from Beyotime (Shanghai, China). Antibodies were supplied by Affinity Biosciences (Liyang, China).

The Hep-2 and TU212 human LC cell line was obtained from iCell Bioscience (Shanghai, China). Cells were grown in RPMI 1640 (Gibco, MA, United States) containing 10% fetal bovine serum (CLARK, VA, United States) at 37 °C in a humidified 5% CO₂ atmosphere.

Cells were inoculated in 96-well plates (1 × 10⁴/well) for 24 h, and then treated with various GIT doses for 24 h. After subsequently adding CCK-8 solution, a 2-h incubation was performed, and absorbance was quantified at 450 nm using a microplate reader (Tecan, Männedorf, Switzerland).

Cells (1,000/well) were inoculated in 6-well plates and grown for 48 h. Cells were treated using different concentrations of GIT for 24 h, followed by further incubation with media replenishment every 3 days until colony appearance. Cells were fixed with 4% paraformaldehyde, stained using a commercial staining kit, and photographed via camera.

Cells were added to glass-bottom dishes and followed by GIT treatment for 24 h, after which the cells were rinsed with PBS and stained with PI solution as directed. Brightfield and fluorescence images were captured using a fluorescence microscope (Olympus, Tokyo, Japan).

Cells were seeded into 6-well plates and cultured until confluent. A linear scratch was created with a sterile pipette tip, and the cells were washed using PBS to eliminate debris. Subsequently, the cells were treated using GIT solution and incubated for 24 h. The wounds were imaged at 0 and 24 h using a microscope, and migration was assessed by measuring the relative size of the wounds at 24 h compared with at 0 h post scratch with Fiji (ImageJ distribution, WI, United States).

Cells were suspended using serum-free medium containing varying concentrations of GIT and added to the upper chamber of Matrigel-coated (1:8 in PBS) Transwell inserts. After filling the lower compartment with 10% FBS-containing medium, plates were incubated for 24 h, and PBS was used to wash cells followed by 4% paraformaldehyde fixation, after which a staining kit was used for staining. Migratory cells (cells attached to the underside of the Transwell chambers) were visualized using a microscope (Olympus, Tokyo, Japan) and counted.

The prognostic value of genes associated with the PI3K-Akt axis was evaluated with Kaplan-Meier Plotter (http://kmplot.com/analysis/). A cohort of 499 patients was stratified into high- and low-expression groups for PIK3CA and AKT1 based on median gene expression levels. Overall survival (OS) was compared using Kaplan-Meier survival curves.

Cells were grown in 12-well plates followed by treatment with GIT for 24 h. Total RNA was extracted with the TRI reagent and processed into cDNA with the HiScript III RT SuperMix kit (Vazyme, Nanjing, China). qRT-PCR was conducted using SYBR Green Master Mix (ThermoFisher, MA, United States) on an ABI StepOne Plus system (ABI, CA, United States). Primer sequences: PIK3CA (F: 5′-TCT​GTC​TCC​TCT​AAA​CCC​TG-3′, R: 5′-TTC​TCC​CAA​TTC​AAC​CAC-3′), AKT1 (F: 5′-GCG​AGC​TGT​TCT​TCC​ACC-3′, R: 5′-CCT​TGT​CCA​GCA​TGA​GGT​T-3′), and β-Actin (F: 5′-TTC​AAC​ACC​CCA​GCC​ATG-3′, R: 5′-CTC​GTA​GAT​GGG​CAC​AGT-3′). β-Actin represented the internal reference, with relative gene expression determined using the 2^−ΔΔCT method.

Cells in 6-well plates were treated for 24 h using various GIT doses, followed by lysis with a lysis buffer containing protease and phosphatase inhibitor cocktails. After quantitification by BCA assays, equal protein amounts were separated via SDS-PAGE and transferred onto PVDF membranes. Following a 1-h block, the blots were treated overnight at 4 °C with primary antibodies against PI3K, p-PI3K, AKT1, p-AKT1, and GAPDH. Secondary antibodies were then used to probe blots for 1 h at room temperature. Protein band detection was performed with an ECL chemiluminescence reagent and visualized with a UVP ChemSolo Auto system (Analytic Jena, Jena, Germany).

GraphPad Prism 8 (GraphPad Software, CA, United States) was employed for all analyses. Data are given as means ± SD and compared with one-way ANOVAs. P < 0.05 was deemed significant.

The chemical structure of GIT is presented in Figure 2A. In total, 253 GIT-associated targets were retrieved from the PharmMapper database, while 1,226 LC-related genes were identified from the GeneCards database. By analyzing their intersection using the Draw Venn Diagram tool, 96 potential GIT targets for LC were identified (Figure 2B). The expression levels of these 96 overlapping genes in head and neck cancer tissues relative to normal tissues were assessed using the UALCAN database (Figures 3A–D).

The top 20 overlapping genes, ranked by relevance score, were used for constructing a PPI network in STRING, with the network comprising 20 nodes and 109 edges (Figure 4A). The same genes were additionally analyzed in the GeneMANIA database to generate the GGI network (Figure 4B).

A Friends analysis of the top 20 overlapping genes revealed that FGFR1 exhibited the highest connectivity, suggesting its potential role in LC treatment (Figure 5A). GO annotation analysis identified 725 biological processes (BPs), 10 cellular components (CCs), and 69 molecular functions (MFs) (P < 0.05). The top three enriched terms based on P-values are shown in Figure 5B. The most significant BPs included peptidyl-tyrosine modification, peptidyl-tyrosine phosphorylation, and gland development. The top CCs included membrane microdomains, membrane rafts, and late endosomes, while the most significant MFs involved transmembrane receptor protein tyrosine kinase activity, protein tyrosine kinase activity, and protein serine/threonine/tyrosine kinase activity. KEGG analyses revealed pathways associated with proteoglycans in cancer, the PI3K-Akt axis, and central carbon metabolism in cancer as the three most significantly enriched pathways. Similar KEGG pathway results were obtained when analyzing all 96 overlapping genes, confirming the importance of the PI3K-Akt axis in LC (Supplementary Figure S1). As this pathway is known to be linked to tumor progression and potential relevance in LC treatment, the effects of GIT on the axis were further investigated.

To evaluate the impact of GIT on LC progression, Hep-2 and TU212 cells were treated using varying doses of GIT. As presented in Figure 6A, CCK-8 assays demonstrated that GIT significantly decreased LC cell viability in a dose-dependent fashion. Accordingly, 20, 30, and 40 μM GIT concentrations were selected for subsequent Hep-2 cell experiments, while 5, 10 and 15 μM GIT concentrations were selected for TU212 cell experiments. Consistently, colony formation assays demonstrated a significant reduction in the proliferative capacity of both cells treated with GIT relative to controls (Figure 6B). Additionally, PI staining indicated a dose-dependent increase in apoptosis following GIT treatment, confirming its cytotoxic effects on LC cells (Figure 6C).

Cell migration and invasion are central to tumor metastasis. The results of a wound healing assay demonstrated that GIT treatment markedly and dose-dependently reduced the migratory capacity of Hep-2 and TU212 cells (Figures 7A,B). Similarly, Transwell invasion assays showed a substantial reduction in invasive cell numbers following GIT treatment compared to the control group (Figures 7C,D). These findings show that GIT can inhibit LC cell migration and invasion effectively, further supporting its therapeutic potential.

The preceding results revealed that GIT effectively inhibits LC cell proliferation, viability, migration, and invasion. Network pharmacology showed that the close involvement of the PI3K-Akt axis in GIT-mediated LC treatment. To determine whether GIT exerts its anti-cancer effects through this pathway, we conducted both database analyses and experimental verification. Analysis using the UALCAN database revealed that the key genes PIK3CA and Akt1 were significantly overexpressed in head and neck squamous cell carcinoma tumors relative to normal tissues (Figures 8A,B). Furthermore, the expression of PIK3CA and Akt1 in head and neck cancer based on nodal metastasis status was also investigated (Supplementary Figure S2). Additionally, Kaplan-Meier analyses demonstrated that elevated AKT1 expression was related to poorer prognosis (Figures 8C,D). These results suggest the close involvement of the PI3K-Akt axis in LC progression and its potential for therapeutic intervention.

To further evaluate this hypothesis, qRT-PCR and Western blotting were conducted. qRT-PCR analysis showed that treatment with medium and high concentration GIT led to a slight upregulation of PIK3CA and AKT1 expression in Hep-2 cells and TU212 cells (Figure 9A). Western blotting further demonstrated that GIT markedly reduced PI3K and Akt protein phosphorylation (Figure 9B), indicating that GIT effectively inhibits the activation of the PI3K-Akt axis in LC cells. These combined bioinformatics and experimental findings confirm that the antitumor effects of GIT in LC are driven through the suppression of this pathway.