Cysteine suppresses the progression of oral squamous cell carcinoma via regulating GLDC

Introduction

Oral cancer is one of the most common cancer globally, with higher prevalence in certain regions and among populations with specific lifestyle risk factors. This geographic clustering is closely associated with extensive tobacco and alcohol use, as well as betel nut chewing (1). Oral squamous cell carcinoma (OSCC) is the most prevalent form of oral cancer, accounting for approximately 90% of all cases (2). The prognosis for OSCC is generally poor, with a 5-year survival rate of only 50% (3). Current treatment strategies for OSCC include surgical resection, radiotherapy, chemotherapy, or a combination of these modalities, depending on the stage of the disease. While these treatments are essential for managing OSCC, they come with side effects that can complicate clinical practice.

Recent research highlights that tumors often exhibit abnormal metabolism, and metabolic reprogramming plays a significant role in tumor progression and response to immunotherapy (4). In particular, reprogramming of amino acid metabolism is a critical aspect of tumor metabolism that significantly influences various tumor behaviors (5). Recent studies have identified significant differences in amino acid levels between tumor and normal tissues in patients with OSCC, suggesting that amino acid level regulation may be a new strategy for the treatment of OSCC (6). However, the specific effects and mechanism of amino acid on the progression remain unknown.

In this study, we investigated the function and underlying mechanism of amino acid in the development of OSCC. Transcriptome sequencing was applied to identify the target of cysteine in OSCC. We aimed to explore a novel metabolic cause of OSCC and evaluate the potential value of cysteine as a promising diagnostic and therapeutic target. We present this article in accordance with the ARRIVE and MDAR reporting checklists (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0839/rc).

Methods

Study design

Mendelian randomization (MR) is a genetic epidemiological technique that uses single nucleotide polymorphisms (SNPs), which are strongly associated with exposure factors, as instrumental variables (IVs) to estimate potential causal relationships between exposures and outcomes (7). By leveraging the fact that genotypes are randomly assigned during gamete formation, this approach effectively mitigates confounding issues in observational studies, particularly addressing biases from unmeasured confounding factors (8). This method has been demonstrated to be effective in resolving causal questions in various research contexts.

We performed a dual-sample MR analysis to assess the impact of amino acid levels on OSCC. Exposure and outcome data were obtained from GWAS datasets. SNPs with genome-wide significance (P<5×10⁻⁸), independent inheritance (R²<0.001), and no linkage disequilibrium (LD) with each amino acid were selected as IVs. SNPs with an F-statistic below 10, considered ‘weak tools’ were excluded. We employed several MR methods, including inverse variance weighting (IVW), MR Egger, simple mode, weighted mode, and weighted median, to estimate causal effects. The IVW method, combining meta-analysis with the Wald ratio of causal effects for each SNP, offers the most accurate estimates. We assessed heterogeneity and pleiotropy to ensure result reliability and conducted reverse MR to explore the effect of OSCC on circulating amino acid levels. Additionally, MR findings were validated through in vitro and in vivo experiments.

Differentially expressed gene analysis and prognostic analysis

OSCC bulk transcriptome data and clinical prognostic data (GSE85446 and GSE11139) were retrieved from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). The Cancer Genome Atlas (TCGA)-OSCC cohort (tumor =270, normal =19) in this study was selected from the broader TCGA-head and neck squamous cell carcinoma (HNSC) group, which includes cases from various anatomical sites such as base of tongue, bones, joints and articular cartilage of other and unspecified sites, floor of mouth, gum, lip, other and ill-defined sites in lip, oral cavity and pharynx, other and unspecified parts of mouth, other and unspecified parts of tongue and palate. These data were obtained from the TCGA database (https://portal.gdc.cancer.gov/). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

To identify differentially expressed genes (DEGs) between tumor and normal tissue samples, we utilized the limma package in R software. The analysis was conducted with a significance threshold of P<0.05 and a fold change (FC) of |log₂ FC| ≥1. For visual representation of these DEGs, we employed the ggplot2 package to create box plots.

Additionally, we performed single-sample gene set enrichment analysis (ssGSEA) to calculate the GLDC score for each sample. An optimal cutoff point was determined to categorize the samples into GLDC-low and GLDC-high groups. To assess the impact of GLDC scores on overall survival, we employed the Kaplan-Meier survival curve analysis.

Selection of IVs

The genome-wide association study (GWAS) dataset on 20 amino acids and OSCC was sourced from the Integrative Epidemiology Unit (IEU) Genome-wide association study (GWAS) database (https://gwas.mrcieu.ac.uk/), which is restricted to the European population. The dataset for OSCC is identified as ebi-a-GCST012238. Complete information on all amino acids is detailed in Table 1.

In this MR study, SNPs significantly associated with 20 blood amino acids at the genome-wide level (P<5×10⁻⁶) and not in LD with other SNPs (r²<0.001 within a clumping window of 10,000 kb) were used as IVs for these amino acids. Additionally, reverse MR analysis was performed with oral cancer as the exposure and cysteine as the outcome. For this analysis, SNPs associated with oral cancer at genome-wide significance (P<5×10⁻⁶) were selected, ensuring no LD with other SNPs (r²<0.001 within a clumping window of 10,000 kb).

Cell culture

Human OSCC (SCC25, RRID: CVCL_1681) cells were cultured in DMEM/F-12 (1:1) (Gibco, USA, 11320033) medium supplemented with 10% fetal bovine serum (FBS, Biosharp CN, BL205A) and 1% penicillin-streptomycin solution (Gibco, USA, 10378016). Human OSCC (SCC15, RRID: CVCL_1682) cells were cultured in DMEM (Gibco, USA, 11965092) medium supplemented with 10% FBS and 1% penicillin-streptomycin solution. All the cells were incubated at 37 ℃ in a humidified atmosphere containing 5% CO₂.

Cell proliferation

SCC15 and SCC25 cells, in the logarithmic growth phase, were seeded into a 96-well plate at a density of 4,000 cells per well and incubated for 24 hours. Following this, the cells were treated with cysteine (MCE, HY-Y0337, New Jersey, USA) at concentrations of 0, 0.1, 1, 10, and 100 µM, and then returned to the incubator. Absorbance measurements were taken at 24, 48, and 72 hours. For the Cell Counting Kit-8 (CCK-8) assay, the culture medium was removed, and the wells were rinsed with phosphate-buffered saline (PBS). Then, 100 µL of a CCK-8 reagent (Accurate Biology, AG51006) diluted 1:10 in fresh culture medium was added to each well. After a 1-hour incubation at 37 ℃, absorbance was measured at 450 nm. Statistical analysis was performed, and the experiment was repeated three times to ensure reproducibility.

Wound-healing assay

In this study, we inoculated 1×10⁶ SCC15 and SCC25 cells into each well of a 6-well plate and allowed them to incubate overnight. To create a wound-like gap in the cell layers, we gently scraped the surface of the cells with a pipette tip. After scraping, the wells were rinsed with PBS to eliminate any debris. Subsequently, we treated the SCC15 and SCC25 cell lines with 10 and 1 µM concentrations of cysteine, respectively. The cells were then incubated at 37 ℃ for an additional 24 hours. After this treatment period, we captured images of the cell monolayer using a microscope to assess the extent of the wound formed during the scraping process. Wound closure (%) = (initial scratch area − 24-hour scratch area)/initial scratch area × 100%.

Invasion assay

Chambers with an 8-µm pore size were pre-coated with 20 µg/100 µL of EHS Matrigel (Corning, 354277, New York, USA) and allowed to dry. After drying, the chambers were rehydrated by rinsing with 100 µL of serum-free Roswell Park Memorial Institute (RPMI) medium. The cells (1×10⁵ cells/well) were then seeded in serum-free medium into the chambers, which were placed into a 24-well plate. An aliquot of 0.5 mL of medium containing 10% FBS was added to the lower well to act as a chemical attractant. The Matrigel invasion chambers were incubated at 37 ℃ in 0.5% CO₂. Following incubation, the invaded cells were fixed with 1% glutaraldehyde in PBS for 30 minutes, then stained with crystal violet for 30 minutes. Images of the cells in the upper chamber were captured and counted.

RNA sequencing and DEGs analysis

Total RNA was extracted using the TRIzol reagent (Invitrogen, 15596026CN, CA, USA) according to the manufacturer’s protocol. RNA purity and quantification were evaluated using the NanoDrop 2000 spectrophotometer (Thermo Scientific, Massachusetts, USA). RNA integrity was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Then the libraries were constructed using VAHTS Universal V6 RNA-seq Library Prep Kit according to the manufacturer’s instructions. The transcriptome sequencing and analysis were conducted by OE Biotech Co., Ltd. (Shanghai, China). Each experimental group underwent five biological replicates.

The library was sequenced on the Illumina NovaSeq 6000 platform, generating paired-end reads of 150 bp. Raw reads in fastq format were processed using fastp to remove low-quality sequences and obtain clean reads. Hierarchical Indexing for Spliced Alignment of Transcripts 2 (HISAT2) was used to map the clean reads to the reference genome. The fragments per kilobase of transcript per million mapped reads (FPKM) of each gene was calculated, and gene read counts were obtained using HTSeq. Biological repeatability was assessed by performing principal component analysis (PCA) in R (v3.2.0).

Differential expression analysis was conducted using DESeq2, with thresholds set at q-values <0.05 and FC >2 or <0.5 to identify significantly DEGs. Hierarchical clustering of DEGs was performed in R (v3.2.0) to visualize gene expression patterns across different groups and samples. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were carried out on DEGs using R (v3.2.0) based on a hypergeometric distribution to identify significantly enriched terms. Finally, a bar chart of the enriched terms was generated using R (v3.2.0).

Animals experiments

Male BALB/c nude mice (4 weeks old, weighted 18–22 g) were purchased from Beijing Vital River Laboratory Animal Technologies Co. Ltd. All animal experiments were performed under a project license (No. 2021-272) granted by the Animal Care and Use Ethics Committee of Shandong Provincial Hospital Affiliated to Shandong First Medical University, in compliance with the institutional guidelines for the care and use of animals, and conducted according to the AAALAC and the IACUC guidelines.

Establishment and analysis of subcutaneous xenograft tumor model was performed as we described before (9). In brief, in subcutaneous xenograft tumor model, 1×10⁷ SCC25 cells were injected subcutaneously into BALB/c nude mice. Tumor width and length were recorded every 3 days by the following formula: volume = (length × width²)/2. When the tumor sizes were above 50 mm³, nude mice were randomly divided into two groups (n=5 per group) and vector (PBS) and cysteine (100 mg/kg weight) were intraperitoneally injected into each mouse every 2 days, this dosage was based on previous research. Two weeks after injection, the nude mice were all killed, and tumors were removed, weighed and imaged.

Immunohistochemical staining

For immunohistochemical staining, tumor tissues were isolated from tumor-bearing mice, fixed in 37% formalin and embedded in paraffin, and then deparaffinized, rehydrated, and followed by antigen retrieval. The sections were then incubated with primary antibodies GLDC (Abcam, Cambridge, USA, ab97625), Ki-67 (Proteintech, Wuhan, China, 27309-1-AP) and EGFR (Proteintech, 18986-1-AP), respectively. Subsequently, the sections were incubated with an anti-rabbit secondary antibody (ZsBio, Beijing, China, SAP-9100) after being washed with PBS, and staining was then visualized with DAB (ZsBio, ZLI-9017).

Real-time quantitative polymerase chain reaction (RT-qPCR) assay

Total RNA was extracted from SCC15 and SCC25 cells using TRIzol reagent (Invitrogen, Waltham, MA, USA; catalog number 15596018CN), adhering to the manufacturer’s recommended protocol. The extracted RNA was dissolved in RNase-free water, and its purity and concentration were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Massachusetts, USA). Subsequently, 1 µg of total RNA was reverse transcribed into complementary DNA (cDNA) using a PrimeScript RT Reagent Kit (TaKaRa, RR047A, Beijing, China). RT-qPCR was then performed using BeyoFast™ SYBR Green qPCR Mix (2X) (Beyotime, D7260-5ml, Shanghai, China) on a Bio-Rad real-time PCR system. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as the internal reference control for normalization. The relative gene expression levels were calculated using the 2^−ΔΔCT method. For this analysis, SCC15 cells were subjected to six biological replicates, while SCC25 cells were analyzed with three biological replicates. The primer sequences utilized in this study are provided in Table 2.

Western blotting assay

In this study, SCC15 and SCC25 cells were harvested and lysed using ice-cold radioimmunoprecipitation assay (RIPA) buffer (NCM Biotech, WB3100, Suzhou, China), which was supplemented with a protease inhibitor cocktail (MCE, HY-K0010). The lysates were incubated on ice for 30 minutes, followed by centrifugation at 12,000 ×g for 10 minutes at 4 ℃ to separate the protein supernatant. The protein concentrations were measured using a bicinchoninic acid (BCA) kit (Beyotime, P0010). Subsequently, proteins were resolved using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, IPVH00010, Massachusetts, USA). To reduce background noise, the membranes were blocked with a solution of 5% non-fat milk in PBS containing 0.1% Tween-20 (PBST) for one hour at room temperature. Following the blocking step, the membranes were incubated overnight at 4 ℃ with the primary antibody targeting GLDC (Abcam, ab97625). After three washes with PBST to remove unbound antibodies, horseradish peroxidase-conjugated secondary antibodies were applied for one hour at room temperature. The target protein bands were ultimately visualized using an enhanced chemiluminescence (ECL) reagent (NCM Biotech; P10300), allowing for the assessment of protein expression levels.

Statistical analysis

The inverse variance weighted (IVW) method served as the primary approach for MR analysis to examine the causal effects of 20 blood amino acid levels on the risk of OSCC. Additionally, we utilized other methods, including the weighted median and weighted mode, to corroborate our findings. Following the primary analysis, a series of sensitivity analyses were conducted to assess robustness. These included evaluating heterogeneity using Cochran’s Q statistic, detecting horizontal pleiotropy through MR Egger intercept analysis, and performing Leave-One-Out analysis to determine the influence of individual SNPs.

All data processing and statistical analysis performed in this study were based on R software (v.4.3.3). We utilized t-test for data that was normally distributed to determine differences between two groups, and the Mann-Whitney U test was applied to data that did not follow a normal distribution. In correlation analysis, the Pearson test was used to verify the statistical significance. All assays were biologically replicated at least three times. Statistical significance was set with * indicating a P value less than 0.05; **, P<0.01; and ***, P<0.001. In addition, all statistical analyses were performed using SPSS 25.0 for Windows. Results are expressed as mean ± standard deviation (SD).

Results

Two-sample MR analysis of serum amino acids and OSCC

The IVW method was employed for MR analysis to investigate the relationship between blood amino acid levels and the risk of OSCC. The analysis demonstrated a statistically significant causal association between cysteine levels and a risk of OSCC, with a P value less than 0.05. This suggests that cysteine may play a role in the pathogenesis of OSCC. Conversely, no significant causal relationships were identified between other amino acids and OSCC risk, as evidenced by P values exceeding 0.05 (Figure 1A,1B).

Figure 1 Two-sample MR analysis of serum amino acids and the risk of oral cancer. (A) Predicting the association between plasma amino acid levels and oral cancer risk using IVW method based on plasma amino acid data analysis. (B) Circle diagram representing causal effects of genes, predicting the association between plasma amino acids and oral cancer risk using different MR methods. IVW, inverse variance weighting; MR, Mendelian randomization; OR, odds ratio; SNPs, single nucleotide polymorphisms.

Two-sample MR analysis of cysteine and OSCC

We employed five MR techniques to thoroughly investigate the association between cysteine exposure and OSCC risk. The analysis revealed that both the IVW and weighted median methods showed P values less than 0.05 and the weighted mode method produced P values close to 0.05, with an odds ratio (OR) less than 1 (Figure 2A). The examination of cysteine exposure data indicated a negative correlation between cysteine levels and the risk of OSCC (Figure 2B,2C). Furthermore, a leave-one-out analysis demonstrated that no single SNP significantly affected the results (Figure 2D). The funnel plot evaluation did not reveal any significant bias (Figure 2E). Meanwhile, the validation tests for heterogeneity and pleiotropy yielded P values greater than 0.05 (Table S1). This indicates the absence of significant issues with these factors, thereby supporting the stability of causal inferences within the study.

Figure 2 Two-sample MR analysis of cysteine and the risk of oral cancer. (A) Forest plots representing the MR estimates and 95% CI values of the causal effects of cysteine. Scatter plot (B) and forest plot (C) of the causal effect of cysteine on oral cancer risk. (D) Leave-one-out sensitivity analysis for cysteine on oral cancer. (E) Funnel plot assessing SNP bias in MR analysis. CI, confidence interval; MR, Mendelian randomization; OR, odds ratio; SE, standard error; SNPs, single nucleotide polymorphisms.

Reverse MR analysis of cysteine and OSCC

To explore the potential reverse causation between cysteine and OSCC, we conducted an MR analysis with OSCC as the exposure variable and cysteine as the outcome variable. The results indicated no significant relationship between these variables, with a P value greater than 0.05, suggesting no clear effects of OSCC on cysteine levels (Table 3).

Elevated levels of exogenous cysteine inhibit the proliferation, migration, and invasion of OSCC cells

We analyzed the effects of varying extracellular cysteine concentrations on cell proliferation, migration, and invasion using SCC15 and SCC25 cells. When the concentration of cysteine exceeds 10 µM, it can significantly inhibit the growth activity of SCC15 cells (P<0.05); when the concentration of cysteine exceeds 1 µM, it can significantly inhibit the growth activity of SCC25 cells (P<0.05) (Figure 3A,3B). Additionally, the results of the cell scratch assay showed that the average wound closure rates of SCC15 and SCC25 in the drug treatment group were 18.85% and 18.78%, respectively; while those in the control group were 25.15% and 27.10%, respectively. The results of the transwell invasion assay showed that the average number of invasive cells in the drug-treated groups SCC15 and SCC25 were 1,014 and 163, respectively; while in the control groups, the average number of invasive cells in SCC15 and SCC25 were 1,255 and 297, respectively. The results indicated treatment with cysteine reduced both migration and invasion of OSCC cells compared to the vehicle group (P<0.01) (Figure 3C-3E).

Figure 3 Cysteine inhibits tumor cell proliferation, migration, and invasion. (A,B) Proliferation was measured using CCK-8 assay. Cell migration was measured by wound scratch assay. Quantification of margin closure rate between different groups at (C) SCC15 and (D) SCC25. (E) Transwell assay for detecting cell invasion ability and its statistical analysis (Crystal violet stained). Statistically significant differences between groups were indicated by *, P<0.05; **, P<0.01; ***, P<0.001. Scale bar: 200 µm. CCK-8, Cell Counting Kit-8.

Cysteine suppresses OSCC progression by targeting GLDC

To elucidate the molecular mechanisms underlying the anticancer effects of cysteine, we performed transcriptome sequencing. The results revealed that cysteine treatment induced significant alterations in gene expression. Of all involved genes, 173 were upregulated, while 87 were downregulated in cysteine-treated group (Figure 4A,4B). KEGG analysis suggested that DEGs predominantly enriched in metabolic pathways associated with serine, threonine, and valine metabolism (Figure 4C). Among these, glycine decarboxylase (GLDC) emerged as the most critical hub in the protein-protein interaction (PPI) network (Figure S1).

Figure 4 Exploring key genes involved in cysteine’s inhibition of tumor growth through transcriptomes. Volcano plot (A) and heatmap (B) of DEGs between the control group and the cysteine-treated group. (C) KEGG enrichment analyses of DEGs. (D) Differential analysis of GLDC expression levels between the normal and tumor groups in RNA-seq data from the TCGA database. (E) GLDC mRNA expression was measured by qPCR in the control group and the cysteine-treated group. (F) Levels of GLDC proteins in the control group and the cysteine-treated group. Kaplan-Meier survival analysis and log-rank test for stratification by GLDC score in relation to overall patient survival in TCGA-OSCC (G), GSE85446 (H) and GSE111390 (I). *, P<0.05; **, P<0.01. CI, confidence interval; DEGs, differentially expressed genes; FC, fold change; HR, hazard ratio; KEGG, Kyoto Encyclopedia of Genes and Genomes; mRNA, messenger RNA; OSCC, oral squamous cell carcinoma; qPCR, quantitative polymerase chain reaction; RNA-seq, RNA sequencing; TCGA, The Cancer Genome Atlas.

In recent years, an increasing number of studies have shown that GLDC plays an important role in cancer progression. In order to further investigate the clinical significance of GLDC in OSCC, we analyzed gene expression data from OSCC and normal tissue samples obtained from the TCGA database. Compared to adjacent normal tissues, GLDC was found to be highly expressed in OSCC (Figure 4D). qPCR and western blot have shown that cysteine reduces the expression of GLDC in OSCC cells (P<0.05) (Figure 4E,4F). Additionally, we examined gene expression and clinical prognosis data from both the TCGA and GEO databases. Notably, low expression of GLDC was associated with improved survival outcomes in patients (P<0.05), suggesting that GLDC may serve as a potential biomarker for predicting prognosis in OSCC (Figure 4G-4I).

Cysteine inhibits tumor progression in mice

Mice were treated daily with either PBS or cysteine for 15 days (Figure 5A). The results showed that cysteine treatment significantly reduced the growth rate of tumors and the endpoint volume and weight of drug treatment (P<0.05), suggesting that cysteine may inhibit the progression of OSCC in vivo (Figure 5B-5D). Research indicated that cysteine treatment did not cause significant toxicity or side effects in normal tissues, although it induced vacuolar deformation in tumor cells (Figure 5E,5F). Immunohistochemical analysis revealed a significant decrease in the expression of Ki67 and GLDC in cysteine-treated tumor tissues (P<0.05) (Figure 5G-5I). Interestingly, the expression of EGFR was significantly decreased in cysteine-treated cells (P<0.05) (Figure 5G,5J).

Figure 5 Cysteine inhibits tumor growth in nude mice. (A) Flowchart showing the operational procedure of animal experiment. (B) Reduction in the weight of tumors by the daily intraperitoneal injection of cysteine. (C) After subcutaneous inoculation of SCC25 cells into nude mice, cysteine was delivered by intraperitoneal injection once every day for 15 days, then tumor growth of nude mice was detected as indicated. (D) Tumor weight was measured and quantified on day 15. (E) The body weights of mice from each group were measured every three days. (F) H&E staining of tumor heart, liver, spleen, lung and kidney. (G) The IHC staining images of Ki-67, GLDC and EGFR were observed under a microscope with 20 times magnification. (H-J) Lower panels show quantitative data of the IHC, n=3. Scale bar, 200 µm. *, P<0.05. H&E, hematoxylin and eosin; IHC, immunohistochemistry.

Discussion

Oral cancer is an aggressive malignancy with a 5-year survival rate of less than 50%, largely due to its low mutation rates, lack of distinct molecular subtypes, and limited therapeutic targets (10). OSCC is the most common histological form of this cancer. Currently, the tumor, lymph node, and metastasis (TNM) staging system is the primary method for assessing tumor size and lymph node involvement, helping to guide prognosis and treatment decisions (11). Despite significant advances in cancer research, surgery remains the primary treatment option for OSCC, and survival rates have not seen substantial improvement. Consequently, there is an urgent need for novel treatment strategies to enhance the prognosis and outcomes for patients with OSCC.

Tumor metabolism refers to the metabolic pathways affected by cancer cells to support their rapid growth, survival, and ability to spread. This process involves metabolic abnormalities of various substances such as sugars, lipids, amino acids, and nucleotides, all of which are part of interconnected metabolic networks (5). Among these, amino acid metabolism is a defining feature of cancer, playing crucial roles in providing carbon and nitrogen sources, maintaining redox balance, regulating epigenetic changes, and activating cancer-related pathways. Research has shown significant differences in amino acid levels between tumor and normal tissues, highlighting changes in amino acid metabolism in OSCC (6).

MR is a statistical approach that uses genetic variation as an instrumental variable to infer causal relationships between exposure factors and outcomes, which minimizes confounding bias and provides more robust causal evidence for epidemiological studies (12).In this research, MR was applied to analyze the GWAS database, revealing a negative correlation between cysteine levels and the occurrence and progression of OSCC. These findings suggest that cysteine may play a potential anti-cancer role in OSCC. Cysteine is a sulfur-containing amino acid with key functions, including antioxidant defense, reactive oxygen species (ROS) clearance, and the maintenance of cellular redox balance. It also stabilizes protein structure through disulfide bonds and participates in thiometabolism, ferroptosis, and the one-carbon cycle (13-15). Previous studies have shown that cysteine deficiency in the tumor microenvironment can impair T cell function, and supplementation with cysteine or N-acetylcysteine (NAC) may restore their anti-tumor activity, supporting cysteine’s potential anticancer effects (16). In our study, in vitro experiments demonstrated that cysteine inhibits the progression of OSCC, particularly by reducing cell proliferation, migration, and invasion. Additionally, a subcutaneous tumor model in mice further confirmed that cysteine effectively suppressed OSCC progression in vivo.

Cysteine, a non-essential amino acid, has been shown to inhibit the progression of OSCC, however, its exact mechanism of action remains unclear. Thus, we conducted transcriptome sequencing and observed significant changes in gene expression following cysteine treatment. Notably, DEGs were primarily enriched in pathways related to the metabolism of serine, threonine, and valine. Further analysis identified GLDC, a key metabolic gene, as a central hub in the PPI network, suggesting that GLDC may be a target of cysteine’s effects. In vitro molecular biology tests and histological staining from animal experiments confirmed that cysteine treatment influenced GLDC expression. GLDC is a core enzyme in the mitochondrial glycine cleavage system (GCS), which catalyzes the breakdown of glycine into CO₂, NH₃, and 5,10-methylenetetrahydrofolate, playing a critical role in one-carbon metabolism and the folate cycle. Recent studies have shown that GLDC acts as an oncogene, promoting tumorigenesis and cell transformation. It is abnormally overexpressed in various cancers, including small cell lung cancer and colorectal cancer, where it drives tumor growth and metastasis through metabolic reprogramming (17-19). Furthermore, GLDC expression is closely linked to cancer prognosis and may present a potential therapeutic target. In OSCC, upregulation of GLDC has been shown to promote tumorigenesis and metastasis by inhibiting the p53 signaling pathway (20). These findings support our hypothesis that cysteine may inhibit OSCC progression by modulating GLDC expression.

Given cysteine’s inhibitory effect on the progression of OSCC, supplementing or regulating its metabolism has become an important strategy. The most straightforward approach is direct cysteine supplementation, which can be administered through oral or intravenous forms of cysteine or NAC. Additionally, increasing dietary intake of cysteine-rich foods, such as eggs, soybeans, and broccoli, can be beneficial. Another approach is to indirectly promote cysteine synthesis by supplementing vitamin B6, which regulates the sulfur conversion pathway and facilitates the conversion of homocysteine to cysteine (21). This study suggests that nutritional therapies involving cysteine supplementation or modulation of its metabolism could offer potential strategies for preventing or treating OSCC.

This study has certain limitations. First, although MR provides genetic evidence supporting a causal link between cysteine and OSCC, the choice of IVs and potential biases remain uncertain and require further clinical validation. Second, the mechanism by which cysteine influences GLDC expression is not fully understood. Future studies using approaches such as gene knockout and transcription factor binding assays are needed to clarify this pathway. Thirdly, the caveat that subcutaneous models may not accurately recapitulate OSCC pathobiology in humans, and further clinical verification is needed in the future.

Conclusions

In summary, reprogramming of amino acid metabolism plays a crucial role in tumor development. This study demonstrates that plasma cysteine level is negatively associated with the prognosis of OSCC. Cysteine inhibits the progression of OSCC cells by regulating the expression of GLDC, suggesting that it may serve as a novel dietary intervention for improving the prognosis of OSCC. We propose that increasing cysteine intake through dietary strategies could offer a novel and promising approach for preventing and treating OSCC. However, further clinical validation is required to confirm its clinical safety and efficacy.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the ARRIVE and MDAR reporting checklists. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0839/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0839/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0839/prf

Funding: This work was supported by the Youth Project of Shandong Provincial Natural Science Foundation (No. ZR2022QH395), the Shandong Provincial Natural Science Foundation (No. ZR2021MH353), and the Shandong Province Medical and Health Technology Project (No. 202408020564).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-0839/coif). The authors have no conflicts of interest to declare.

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. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All animal experiments were performed under a project license (No. 2021-272) granted by the Animal Care and Use Ethics Committee of Shandong Provincial Hospital Affiliated to Shandong First Medical University, in compliance with the institutional guidelines for the care and use of animals, and conducted according to the AAALAC and the IACUC guidelines.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Chiewwit P, Khovidhunkit SP, Tantipoj C, et al. A study of risk factors associated with the presence of oral potentially malignant disorders: a community-based study from Northeastern Thailand. BMC Oral Health 2024;24:932. [Crossref] [PubMed]
2. Bai XX, Zhang J, Wei L. Analysis of primary oral and oropharyngeal squamous cell carcinoma in inhabitants of Beijing, China-a 10-year continuous single-center study. BMC Oral Health 2020;20:208. [Crossref] [PubMed]
3. Tajmirriahi N, Razavi SM, Shirani S, et al. Evaluation of metastasis and 5-year survival in oral squamous cell carcinoma patients in Isfahan (2001-2015). Dent Res J (Isfahan) 2019;16:117-21. [Crossref] [PubMed]
4. Faubert B, Solmonson A, DeBerardinis RJ. Metabolic reprogramming and cancer progression. Science 2020;368:eaaw5473. [Crossref] [PubMed]
5. Lieu EL, Nguyen T, Rhyne S, et al. Amino acids in cancer. Exp Mol Med 2020;52:15-30. [Crossref] [PubMed]
6. Musharraf SG, Shahid N, Naqvi SMA, et al. Metabolite Profiling of Preneoplastic and Neoplastic Lesions of Oral Cavity Tissue Samples Revealed a Biomarker Pattern. Sci Rep 2016;6:38985. [Crossref] [PubMed]
7. Emdin CA, Khera AV, Kathiresan S. Mendelian Randomization. JAMA 2017;318:1925-6. [Crossref] [PubMed]
8. Davies NM, Holmes MV, Davey Smith G. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ 2018;362:k601. [Crossref] [PubMed]
9. Xu L, Li Q, Wang Y, et al. m(6)A methyltransferase METTL3 promotes oral squamous cell carcinoma progression through enhancement of IGF2BP2-mediated SLC7A11 mRNA stability. Am J Cancer Res 2021;11:5282-98. [PubMed]
10. Lee MR, Kang S, Lee J, et al. Organoid morphology-guided classification for oral cancer reveals prognosis. Cell Rep Med 2025;6:102129. [Crossref] [PubMed]
11. Johnson DE, Burtness B, Leemans CR, et al. Head and neck squamous cell carcinoma. Nat Rev Dis Primers 2020;6:92. [Crossref] [PubMed]
12. Sanderson E, Glymour MM, Holmes MV, et al. Mendelian randomization. Nat Rev Methods Primers 2022;2:6. [Crossref] [PubMed]
13. Kularatne RN, Bulumulla C, Catchpole T, et al. Protection of human retinal pigment epithelial cells from oxidative damage using cysteine prodrugs. Free Radic Biol Med 2020;152:386-94. [Crossref] [PubMed]
14. Han X, Zhai Z, Yang X, et al. A FRET-based ratiometric fluorescent probe to detect cysteine metabolism in mitochondria. Org Biomol Chem 2020;18:1487-92. [Crossref] [PubMed]
15. Su AL, Harris SM, Elkin ER, et al. Trichloroethylene Metabolite S-(1,2-Dichlorovinyl)-l-cysteine Stimulates Changes in Energy Metabolites and Amino Acids in the BeWo Human Placental Trophoblast Model during Syncytialization. Chem Res Toxicol 2023;36:882-99. [Crossref] [PubMed]
16. Levring TB, Hansen AK, Nielsen BL, et al. Activated human CD4+ T cells express transporters for both cysteine and cystine. Sci Rep 2012;2:266. [Crossref] [PubMed]
17. Liu R, Zeng LW, Gong R, et al. mTORC1 activity regulates post-translational modifications of glycine decarboxylase to modulate glycine metabolism and tumorigenesis. Nat Commun 2021;12:4227. [Crossref] [PubMed]
18. Zhang WC, Shyh-Chang N, Yang H, et al. Glycine decarboxylase activity drives non-small cell lung cancer tumor-initiating cells and tumorigenesis. Cell 2012;148:259-72. [Crossref] [PubMed]
19. Yu H, Hu X, Zhang Y, et al. GLDC promotes colorectal cancer metastasis through epithelial-mesenchymal transition mediated by Hippo signaling pathway. Med Oncol 2023;40:293. [Crossref] [PubMed]
20. Xu C, Xu Q, Yang H. H3K27 Acetylation-Activated GLDC Accelerated the Advancement of Oral Squamous Cell Carcinoma by Suppressing the p53 Signaling Pathway. Environ Toxicol 2025;40:140-51. [Crossref] [PubMed]
21. Selhub J, Bagley LC, Miller J, et al. B vitamins, homocysteine, and neurocognitive function in the elderly. Am J Clin Nutr 2000;71:614S-20S. [Crossref] [PubMed]