The active ingredients and associated targets of Aloe vera were explored through TCM Systems Pharmacology Database and Analysis Platform (TCMSP; http://old.tcmsp-e.com/index.php) and SwissTargetPrediction (http://www.swisstargetprediction.ch/) platforms. The interaction between ingredients and targets was established using Cytoscape v3.10.2 software (https://cytoscape.org/). A threshold of 30% for the oral bioavailability and 0.18 for the drug likeness was set for TCMSP, with a probability score of 1 for SwissTargetPrediction (19,20).

GSE53355 and GSE117973 were retrieved from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) using the key words ‘HPV positive HNSCC’, with GSE53355 containing 44 samples (26 HPV⁺ HNSCC, 18 HPV⁻ HNSCC) and GSE117973 comprising 77 samples (54 HPV⁺ HNSCC, 23 HPV⁻ HNSCC) (21,22). Differential gene expression analysis was performed using GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/) (P<0.05; logFC >1).

A Venn diagram was employed to analyze the overlap between the predicted genes of Aloe vera and those associated with HPV⁺ HNSCC, thereby identifying potential therapeutic targets for this cancer subtype.

Functional enrichment analysis was performed employing WebGestalt v2024 (http://www.webgestalt.org). Top-ranked Gene Ontology (GO) terms including biological processes (BPs), molecular functions (MFs) and cellular components (CCs), as well as Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were visualized as bar charts and bubble plots (23). The reference gene set was set to all genome protein-coding genes, to serve as the background for enrichment calculations.

PPI networks based on common targets between Aloe vera and HPV⁺ HNSCC were created using the STRING v12.0 platform (https://cn.string-db.org/) and were visualized in Cytoscape. MCODE v2.0.3 (https://apps.cytoscape.org/apps/mcode) and Network Analyzer plugins were employed to calculate the MCODE score, betweenness centrality (BC), closeness centrality (CC) and degree for each node, with the top 10 genes identified as hub genes (24).

The prognostic relevance of hub genes in a cohort of patients with HNSCC (n=500) was analyzed using the Kaplan-Meier Plotter (https://kmplot.com/analysis/) to evaluate overall survival outcomes, employing the automatic selection of the optimal cutoff based on percentile. To ensure the robustness of the findings, the results of the survival analysis were further validated using the Gene Expression Profiling Interactive Analysis (GEPIA; http://gepia.cancer-pku.cn/) and UALCAN (https://ualcan.path.uab.edu/) platforms. These platforms offer valuable insights into the association between gene expression and survival based on TCGA datasets. Significant associations were determined using a threshold of P<0.05.

The cBioPortal (www.cbioportal.org/) and TIMER2.0 (http://timer.comp-genomics.org/) were employed to examine the somatic copy number alterations (SCNAs) in key Aloe vera targets against HPV⁺ HNSCC, and the correlation between SCNAs and immune cell infiltrations. The infiltration levels for each SCNA category were compared with normal controls using a two-sided Wilcoxon rank-sum test.

The LinkedOmics platform (www.linkedomics.org/) was employed to investigate the co-expression of the hub gene, alongside GSEA to investigate their functional implications (25).

The core pharmacological compounds of Aloe vera, along with the protein structures of the key genes, were sourced from the PubChem (https://pubchem.ncbi.nlm.nih.gov/) and Protein Data Bank (PDB; http://www.rcsb.org/) databases. Molecular docking was performed using PyRx software v0.8 (https://pyrx.sourceforge.io/) that was integrated with AutoDock Vina v1.2.x (https://vina.scripps.edu/). This process identified critical molecular interactions, with a binding energy threshold of 7 kcal/mol.

Quercetin (Shanghai Yuanye Biotechnology Co., Ltd.) was dissolved in dimethyl sulfoxide (MilliporeSigma) to create solutions with concentrations of 5 and 10 µM. Post-sonication (30 min, 40 kHz, 25°C) and centrifugation (5 min, 13,500 × g, 4°C), the supernatant was collected. Equivalent concentrations of dimethyl sulfoxide without quercetin were used as a control.

UPCI:SCC154 cells (CRL-3241; American Type Culture Collection) were cultured in DMEM supplemented with 10% FBS and 100 U/ml penicillin/streptomycin (all Gibco; Thermo Fisher Scientific, Inc.) at 37°C in 5% CO₂.

After being treated for 24 h, total protein was extracted using radioimmunoprecipitation assay lysis buffer (Beyotime Institute of Biotechnology) containing protease and phosphatase inhibitors. Protein concentration was determined using a bicinchoninic acid protein assay kit (Beyotime Institute of Biotechnology). Equal amounts of protein (30 µg per lane) were then separated by SDS-PAGE on 10% gels and transferred to polyvinylidene fluoride membranes (Thermo Fisher Scientific, Inc.). The membranes were blocked with 5% non-fat milk in TBS-0.1% Tween-20 (TBST) for 1 h at room temperature, and then incubated overnight at 4°C with the following primary antibodies: Anti-SERPINE1 (1:1,000 dilution; cat. no. 13801-1-AP; Proteintech Group, Inc.) and anti-β-actin (1:5,000 dilution; cat. no. 66009-1-Ig; Proteintech Group, Inc.). After washing three times with TBST, the membranes were incubated with HRP-conjugated secondary antibodies (goat anti-rabbit IgG, 1:5,000, cat. no. SA00001-2; goat anti-mouse IgG, 1:5,000, cat. no. SA00001-1; both from Proteintech Group, Inc.) for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence detection kit (MilliporeSigma), and imaged using a ChemiDoc™ MP Imaging System (Bio-Rad Laboratories, Inc.) Semi-quantification of protein bands was performed using ImageJ software v1.54 (National Institutes of Health).

Cells were treated for 24 h before RNA extraction. Total RNA was extracted using RNAiso Plus reagent (Takara Bio Inc.), following the manufacturer's instructions. RT was performed using the PrimeScript™ RT reagent kit with gDNA Eraser (Takara Bio Inc.) according to the manufacturer's protocol. qPCR was conducted using TB Green^® Premix Ex Taq™ II (Takara Bio Inc.,) on a StepOnePlus™ Real-Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The primer sequences used were as follows: SERPINE1: forward 5′-ACCGCAACGTGGTTTTCTCA-3′, reverse 5′-TTGAATCCCATAGCTGCTTGAAT-3′; and β-actin: forward 5′-AGAGCTACGAGCTGCCTGAC-3′, reverse 5′-GTCACCTTCACCGTTCCAGT-3′.

The qPCR cycling protocol included an initial denaturation step at 95°C for 3 min, followed by 35 cycles of denaturation at 95°C for 20 sec, annealing at 56°C for 20 sec and extension at 72°C for 20 sec. Melting curve analysis was performed with the following steps: 95°C for 15 sec, 60°C for 15 sec, 95°C for 20 min and 95°C for 15 sec. All reactions were performed in triplicate. Relative gene expression was quantified using the 2^−ΔΔCq method, with β-actin used as the internal control (26).

The experimental data were statistically analyzed using GraphPad Prism v6.0 software (Dotmatics). Differences between two groups were assessed using an unpaired Student's t-test, while multi-group comparisons were performed using a one-way ANOVA followed by Dunnett's test for post hoc analysis. P<0.05 was considered to indicate a statistically significant difference.

Using network pharmacology analysis, eight key bioactive components of Aloe vera were identified: Aloe-emodin, aloeresin C, arachidonic acid, β-carotene, campest-5-en-3β-ol, cholesterol, quercetin and sitosterol (Fig. 1A). Concurrently, 244 potential therapeutic targets for Aloe vera were identified (Fig. 1B). Differential gene expression analysis of the GSE53355 dataset revealed 402 HPV⁺ HNSCC-specific genes, including 257 upregulated and 145 downregulated genes (Fig. 2A). From the GSE117973 dataset, 461 HPV⁺ HNSCC-specific genes were detected (194 upregulated and 267 downregulated; Fig. 2B). The integrated analysis of two datasets identified a total of 768 genes specifically associated with HPV⁺ HNSCC. Through the intersection of Aloe vera targets and HPV⁺ HNSCC-specific genes, 34 potential therapeutic targets were identified as targets for Aloe vera against HNSCC(Fig. 2C) and a drug-target-disease network was constructed (Fig. 2D).

GO analysis indicated that the target genes primarily participated in BPs such as ‘response to lipid’ and ‘response to oxygen-containing compound’ (Fig. 3A and Table SI). At the MF level, they were associated with ‘bile acid binding’ and ‘estradiol 17-beta-dehydrogenase [NAD(P)] activity’ and were localized in the ‘extracellular matrix’ and ‘external encapsulating structure’. KEGG pathway analysis revealed their participation in pathways such as ‘AGE-RAGE signaling pathway in diabetic complications’ and ‘IL-17 signaling pathway’ (Fig. 3B).

A PPI network comprising 29 nodes was created using STRING and Cytoscape (Fig. 4A). A total of two significant clusters within the network were identified by employing the MCODE plugin. By integrating top-ranking nodes based on centrality measures such as degree, BC, CC and MCODE scores, PTGS2, IL6, IL1B, CXCL8 and SERPINE1 were recognized as the hub genes (Fig. 4B and C).

SERPINE1 expression was markedly upregulated in HPV⁺ HNSCC tissues compared with that in normal tissues from healthy individuals (Fig. 5A) and was increased across all stages of HNSCC (Fig. 5B), as well as in eight other types of cancer (Fig. 5C). Survival analysis revealed that SERPINE1 expression was significantly associated with overall survival in patients with HNSCC (Fig. 5D).

The overall frequency of genetic alterations in SERPINE1 in HNSCC was ~5%, suggesting a potential genetic predisposition associated with this gene (Fig. 6A). Amplifications were particularly common and constituted a substantial proportion of these alterations. By contrast, missense and truncating mutations occurred at a lower frequency. In HPV⁺ HNSCC, SERPINE1 frequently underwent arm-level gains and deletions, suggesting that gene dosage alterations may contribute to tumorigenesis in this subgroup (Fig. 6B). These gains were associated with the infiltration of B cells, CD8⁺ T cells, CD4⁺ T cells, macrophages, neutrophils and dendritic cells (Fig. 7A). Moreover, the increased expression of SERPINE1 was significantly correlated with reduced B-cell infiltration, which in turn was associated with worse survival outcomes in patients with HPV⁺ HNSCC (Fig. 7B and C). High levels of SERPINE1 expression were also associated with diminished overall survival in patients with HPV⁺ HNSCC.

Co-expression analysis of SERPINE1 in HNSCC revealed a significant association with several genes, including TGA5, INHBA, ACTN1, E2F2 and PDIK1L (Fig. 8A). Heatmaps showed the top 100 genes associated with SERPINE1, suggesting potential regulatory networks in HNSCC (Fig. 8B and C). Furthermore, GSEA revealed that key BPs were significantly associated with SERPINE1, including ‘focal adhesion’, ‘ECM-receptor interaction’ and ‘PI3K-Akt signaling’, all crucial for cell-matrix interactions and survival signaling (Fig. 8D-F). Additionally, pathways associated with ‘human papillomavirus infection’, ‘base excision repair’ and ‘DNA replication’ were significantly enriched, highlighting the potential role of SERPINE1 in the oncogenic mechanisms of HPV⁺ HNSCC (Fig. 8G-I).

Based on the results of network pharmacological analysis and the drug-component-target diagram (Fig. 1B), quercetin was identified as the primary bioactive component in Aloe vera that specifically targets SERPINE1. Subsequent molecular docking analysis revealed a strong binding interaction between quercetin and the SERPINE1 protein (PDB identification, 1LJ5), with a binding energy calculated at −7.1 kcal/mol (Fig. 9). The analysis revealed key interactions between quercetin and critical residues within the SERPINE1 active site, including hydrogen bonds with GLU378 and hydrophobic interactions involving VAL274, ALA239 and ILE237. These interactions suggested that quercetin may effectively inhibit the function of SERPINE1 by stabilizing its active conformation.

The impact of quercetin on SERPINE1 expression was further evaluated in HPV⁺ HNSCC cells. A dose-dependent reduction in SERPINE1 expression was observed following quercetin treatment. At the mRNA level, a significant decrease was detected in response to 5 and 10 µM quercetin compared with 0 µM, with the effect being more pronounced at 10 µM (Fig. 10A). This decrease in mRNA levels was further validated by western blot analysis. This analysis confirmed that the protein levels of SERPINE1 were downregulated following treatment with quercetin, as depicted in Fig. 10B.