Heterogeneous Biological Responses to Low-Level Sodium Saccharin Exposure: An Integrated Computational Toxicology and Mendelian Randomization Study

Abstract

Background

The long-term biological effects of low-level exposure to artificial sweeteners remain debated. We integrated computational toxicology and Mendelian randomization (MR) to explore potential biological responses associated with sodium saccharin across multiple cancer outcomes.

Methods

Public databases were screened to identify saccharin-related targets, followed by protein–protein interaction network construction and enrichment analyses. Molecular docking assessed binding potential. Two-sample MR analyses were conducted using European genome-wide association study (GWAS) summary statistics to evaluate potential causal associations.

Results

Among 191 identified targets, 160 overlapped with genes implicated across 36 cancer types. Hub genes included TP53, IL6, and MAPK3, with docking indicating moderate binding affinity. Enrichment analyses highlighted inflammatory pathways including interleukin-17 signaling. MR analyses suggested inverse associations with non-melanoma skin cancer (odds ratio [OR] = 0.38, P = 0.006) and colon cancer (OR = 0.56, P < 0.001), while a preliminary positive signal was observed for Kaposi sarcoma (OR = 3.73, P = 0.044). No significant associations were detected for most other cancers.

Conclusion

Sodium saccharin may interact with cancer-related molecular networks; however, population-level evidence indicates heterogeneous and context-dependent associations rather than a generalized carcinogenic effect under low-level exposure conditions.

Graphical Abstract

Keywords

sodium saccharin network toxicology molecular docking mendelian randomization cancer risk low-dose exposure

1. Introduction

Cancer remains a leading cause of mortality worldwide, with an estimated 20 million new cases and 9.7 million deaths reported in 2022.¹ Artificial sweeteners are widely used globally as food additives, and their long-term health effects remain a controversial topic in public health. Since its discovery in the late 19th century, saccharin has been added to beverages, processed foods, and table sweeteners due to its high sweetness and zero-calorie profile.² Pharmacokinetic studies indicate that saccharin is not metabolized in humans and is excreted unchanged via the kidneys.³ Although major regulatory agencies, including the Food and Drug Administration and European Food Safety Authority, consider saccharin safe within established intake limits, these assessments do not rule out potential biological activity. Historically, saccharin faced scrutiny regarding carcinogenicity based on animal studies. A recent comprehensive review of animal and mechanistic evidence reaffirmed that high-quality studies have not shown genotoxic or carcinogenic potential for non-sugar sweeteners, noting that saccharin-induced bladder tumors in rats occur via mechanisms irrelevant to human physiology.^4-6 However, most existing epidemiological evidence comes from observational studies, which are susceptible to residual confounding or exposure measurement errors. Mechanistic simulation and causal inference are therefore necessary to clarify the relationship between saccharin exposure and human cancer risk.

Potential health risks of saccharin may arise from interference with cellular signaling networks. Sodium saccharin can activate taste receptors in the ovaries of rats with polycystic ovary syndrome. This activation enhances p38-MAPK/ERK signaling, accompanied by changes in apoptosis markers and the disruption of steroidogenesis. Cellular experiments further demonstrated that saccharin promotes neutrophil chemotaxis and migration via TAS1R2/TAS1R3 sweet taste receptors, upregulating inflammatory chemokines such as CXCL1 and CCL2.⁷ Microbiological studies revealed that saccharin can destabilize bacterial cell membranes and interfere with DNA replication kinetics, suggesting a molecular potential to disrupt basic biological processes.⁸ Furthermore, in vitro and in vivo experiments confirmed that sodium saccharin activates the T1R3 taste receptor, regulating downstream transcription factors via the cAMP-PKA pathway and affecting testosterone synthesis.⁹ Current research relies heavily on in vitro and animal models, making direct extrapolation to complex human systems difficult. Computational toxicology can systematically simulate molecular interference networks to predict potential risks, while Mendelian randomization (MR) assesses causal associations with health outcomes based on human genetic variation, addressing the limitations of traditional experimental evidence regarding extrapolation and causal inference.

Most current studies assessing saccharin risks are limited to single experimental methods or focus on individual cancer types. This approach lacks a system-level perspective that integrates molecular mechanisms with population evidence. Traditional toxicological assessments focus primarily on acute toxicity or clear carcinogenicity at high doses, often overlooking complex effects triggered by long-term low-dose exposure. In this study, “low-level exposure” is operationally defined as chronic dietary intake of sodium saccharin within the acceptable daily intake range (≤9 mg/kg body weight/day) established by the European Food Safety Authority (EFSA).¹⁰ This intake-based definition reflects real-world exposure conditions under habitual consumption and distinguishes physiological exposure from high-dose experimental toxicity settings. The corresponding internal levels measured by metabolomic GWAS are therefore interpreted as proxies of this exposure state. Such effects typically perturb cellular homeostasis through multiple targets and pathways, defined as network toxicity.¹¹ Additionally, it remains unclear whether the association between saccharin and different cancers is specific. Whether the risk presents as a pan-cancer characteristic or varies by cancer type is a critical scientific question. Computational toxicology systematically predicts biological targets and signaling pathways that chemicals may interfere with, simulating mechanisms of multi-target mediated network toxicity. MR uses genetic variants as instrumental variables (IVs) to assess causal relationships between environmental exposures and disease outcomes, largely minimizing confounding bias inherent in observational studies.¹² Integrating these methods helps construct an integrated evidence chain from molecular hypothesis to population-level causal inference, enabling a systematic assessment of the long-term health risks of saccharin.

A multi-level analysis framework integrating computational toxicology and genetic epidemiology was constructed to assess the potential carcinogenic risk of sodium saccharin and clarify heterogeneous association patterns. By establishing a saccharin-target-cancer interaction network, conducting molecular docking validation, and performing two-sample MR analysis, this study aimed to: (1) systematically predict cancer-related targets and signaling pathways potentially disrupted by saccharin; (2) evaluate binding possibilities with key targets structurally; and (3) explore potential causal associations between saccharin and 36 cancer types from a population genetics perspective.

2. Methods

2.1. Sodium Saccharin Target Collection

The data acquisition, computational simulations, and statistical analyses for this study were systematically conducted from September 2025 to January 2026. The overall workflow of this study, integrating network toxicology, molecular docking, and Mendelian randomization, is illustrated in Figure 1. The standard 2D structural formula of sodium saccharin (Molecular Formula: C7H5NO3S) was obtained from the PubChem database.¹³ Targets related to sodium saccharin were searched using the isomeric SMILES in the SwissTargetPrediction¹⁴ and SEA¹⁵ databases. Searches were also conducted in the Comparative Toxicogenomics Database (CTD),¹⁶ STITCH,¹⁷ Super-PRED,¹⁸ GeneCards,¹⁹ and Therapeutic Target Database²⁰ using “Saccharin” as the keyword. The species was limited to Homo sapiens. In SwissTargetPrediction, a Probability > 0.01 was used as the selection criterion. Targets obtained from different databases were merged for subsequent analysis.

Figure 1.

Flowchart of the study design. Flowchart of the integrated computational toxicology and mendelian randomization study design

2.2. Cancer-Related Target Collection

Targets related to 36 cancer keywords (listed in Table 1) were retrieved, focusing on Homo sapiens.²¹ Databases included GeneCards, OMIM, Therapeutic Target Database, and CTD. Selection criteria were: (1) In GeneCards, a “Relevance Score ≥ 15”; (2) In CTD, the top 5,000 genes based on inference scores. Targets for the 36 cancers retrieved from all databases were integrated, duplicates and false positives were removed, and data were standardized using the UniProt database.

Table 1.

Information on the 36 Cancer Types Included in the Study

No.	Anatomical site	Specific cancer name (standard terminology)	ICD-10 code	Cancer category (system/Organ)	Notes
1	Brain,nervous system	Brain cancer	C71	Head	Includes malignant neoplasms of brain and other parts of central nervous system (CNS). Classified under “Head” for anatomical grouping in this study.
2	Lip,oral cavity	Oral cancer	C00-C06	Head	Covers malignancies of lip, tongue, gum, floor of mouth, palate, and other parts of mouth.
3	Salivary gland	Salivary gland cancer	C07-C08	Head	Includes major and minor salivary gland cancers.
4	Nasopharynx	Nasopharyngeal cancer	C11	Head	Carcinoma originating in the nasopharynx, often associated with EBV infection.
5	Thyroid	Thyroid cancer	C73	neck	Classified under “Neck” due to anatomical location; includes both differentiated and medullary types.
6	Larynx	Laryngeal cancer	C32	neck	Includes glottic, supraglottic, and subglottic carcinomas.
7	Oropharynx	Oropharyngeal cancer	C10	neck	Includes tonsillar and base of tongue cancers, often associated with HPV.
8	Hypopharynx	Hypopharyngeal cancer	C13	neck	Often presents at advanced stage due to subtle early symptoms.
9	Lung	Lung cancer	C34	Thoracic organs	Includes both non-small cell and small cell lung carcinoma.
10	Esophagus	Esophageal cancer	C15	Thoracic organs	Includes both squamous cell carcinoma and adenocarcinoma.
11	Breast	Breast cancer	C50	Thoracic organs	Classified under thoracic organs for anatomical grouping; includes both invasive and in situ carcinomas.
12	Stomach	Gastric cancer	C16	Digestive system	Predominantly adenocarcinoma; includes cardia and non-cardia subtypes.
13	Colon	Colon cancer	C18	Digestive system	Excludes rectal cancers (C19-C20).
14	Rectum	Rectal cancer	C19-C20	Digestive system	Analysed separately from colon cancer in this study.
15	Anus	Anal cancer	C21	Digestive glands	Includes both anal canal and margin cancers; often squamous cell carcinoma.
16	Liver	Liver cancer	C22	Digestive glands	Primarily hepatocellular carcinoma (HCC); includes intrahepatic bile duct cancer.
17	Pancreas	Pancreatic cancer	C25	Digestive glands	Predominantly pancreatic ductal adenocarcinoma.
18	Gallbladder	Gallbladder cancer	C23-C24	Digestive glands	Often adenocarcinoma; includes cancers of extrahepatic bile ducts.
19	Kidney	Kidney cancer	C64	Urinary system	Primarily renal cell carcinoma; excludes renal pelvis cancers (C65).
20	Bladder	Bladder cancer	C67	Urinary system	Predominantly urothelial carcinoma.
21	Cervix uteri	Cervical cancer	C53	Female reproductive system	ncludes both squamous cell carcinoma and adenocarcinoma.
22	Corpus uteri	Uterine corpus cancer	C54-C55	Female reproductive system	Primarily endometrial carcinoma.
23	Ovary	Ovarian cancer	C56	Female reproductive system	Includes epithelial, germ cell, and stromal tumors.
24	Vagina	Vaginal cancer	C52	Female reproductive system	Rare; often squamous cell carcinoma.
25	Vulva	Vulvar cancer	C51	Female reproductive system	Primarily squamous cell carcinoma.
26	Prostate	Prostate cancer	C61	Male reproductive system	Almost always adenocarcinoma.
27	Testis	Testicular cancer	C62	Male reproductive system	Primarily germ cell tumors.
28	Penis	Penile cancer	C60	Male reproductive system	Often squamous cell carcinoma.
29	Non-Hodgkin lymphoma	Non-Hodgkin lymphoma	C82-C85	Hematologic and lymphatic system	Heterogeneous group of lymphoid malignancies; analysed as an aggregate in this study.
30	Hodgkin lymphoma	Hodgkin lymphoma	C81	Hematologic and lymphatic system	Characterized by Reed-Sternberg cells.
31	Leukemia	Leukemia	C91-C95	Hematologic and lymphatic system	Includes acute and chronic forms of lymphoid and myeloid leukemias (analysed as a combined phenotype here).
32	Multiple myeloma	Multiple myeloma	C90	Hematologic and lymphatic system	Malignancy of plasma cells.
33	Nonmelanoma of skin	Nonmelanoma skin cancer	C44	Skin and soft tissues	Primarily includes basal cell carcinoma and cutaneous squamous cell carcinoma.
34	Melanoma of skin	Melanoma	C43	Skin and soft tissues	Malignant melanoma of skin only (excludes ocular and mucosal melanomas).
35	Kaposi sarcoma	Kaposi sarcoma	C46	Skin and soft tissues	Vascular tumor associated with HHV-8 infection; may involve skin, mucosa, or viscera.
36	Mesothelioma	Mesothelioma	C45	Skin and soft tissues	Primarily associated with asbestos exposure; most commonly pleural, but includes other sites.

Note. Information on the 36 cancer types analyzed in this study. Cancers were systematically categorized into 10 anatomical systems to facilitate downstream specific target intersection analysis.

2.3. Identification of Saccharin-Cancer Common Targets

A Venn diagram was generated using an online bioinformatics platform to identify common target genes.²² The 36 cancers were classified according to anatomical physiology, and intersections with sodium saccharin targets were determined.

2.4. Construction of the Saccharin-Target-Cancer Network

The intersection targets were used to construct a Saccharin-Target-Cancer interaction network using Cytoscape 3.10.1 software. Topological analysis was performed using the CytoNCA plugin.²³ Nodes were ranked by degree value to screen for key targets. An integrated network containing three levels of nodes was manually constructed: saccharin molecules (compound layer), identified core targets (target layer), and 36 cancer types (disease layer). Edges represented compound-target and target-disease interactions. This network visualized the potential pleiotropic mechanisms by which saccharin affects multiple cancers through multi-target regulation.

2.5. Protein-Protein Interaction (PPI) Network Construction

Saccharin-cancer intersection targets were imported into the STRING database,²⁴ with the organism set to Homo sapiens and a minimum interaction score threshold of 0.400. Isolated nodes were hidden. Interaction data were imported into Cytoscape (version 3.10.1) for visualization and analysis. The built-in NetworkAnalyzer tool calculated topological parameters,²⁵ including degree, betweenness centrality, and closeness centrality. To identify hub genes, algorithms such as Maximal Clique Centrality (MCC) in the CytoHubba plugin were used for ranking.

2.6. GO and KEGG Enrichment Analysis

Statistical analysis was performed using R version 4.3.1. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG)²⁶ enrichment analyses were performed on intersection genes using R packages including clusterProfiler,²⁷ enrichplot,²⁸ org.Hs.eg.db,²⁹ and ggplot2.³⁰ GO analysis covered biological process, cellular component, and molecular function. The significance of KEGG enrichment results was tested using the hypergeometric distribution and corrected for False Discovery Rate. The threshold was set at an adjusted P-value (adj.P.Val) < 0.05. Results were visualized as bar plots, bubble plots, and Sankey diagrams for the top 10 pathways.

2.7. Molecular Docking and Visualization

Molecular docking was employed to predict potential mechanisms. The chemical structure of saccharin was downloaded from PubChem, and target protein structures were obtained from UniProt and the RCSB PDB database.³¹ AutoDockTools 1.5.7 was used to process ligand and protein files into PDBQT format. AutoDock Vina performed docking to obtain binding energy data. PyMOL was used to visualize the docking results between proteins and ligands.

2.8. Mendelian Randomization Analysis

Two-sample MR was employed to investigate the potential causal association between saccharin exposure and the risk of multiple cancers.

2.8.1. Instrumental Variable Selection

Instrumental variables for saccharin exposure were obtained from a large-scale genome-wide association study on human blood metabolites in a European population (N = 7,824; quantified via untargeted mass spectrometry-based metabolomics).³² Given that saccharin is minimally metabolized and predominantly excreted unchanged in urine, circulating metabolite levels can serve as a proxy for internal exposure. However, the genetic variants derived from metabolomic GWAS primarily capture inter-individual differences in absorption, distribution, and excretion processes, rather than direct dietary intake volumes. Therefore, they are utilized as genetic proxies for chronic internal exposure to saccharin, rather than exact measures of absolute dietary consumption. Selection strictly followed core MR assumptions³³: (1) Relevance: Single Nucleotide Polymorphisms (SNPs) significantly associated with blood/urine saccharin levels at the genome-wide level (P < 5 × 10^-6) were selected; (2) Independence: SNPs were clumped using the 1000 Genomes Project European reference panel with linkage disequilibrium r² < 0.001 and a window of > 10,000 kb to ensure independence; (3) Exclusion of weak IVs: The F-statistic for each SNP was calculated, with F > 10 indicating sufficient strength.

2.8.2. Data Sources

Outcome data (incidence risk of 36 cancers) were derived from independent, large-sample GWAS summary statistics. To maintain consistency and feasibility, oropharyngeal and hypopharyngeal cancers were grouped as “pharyngeal cancer” based on anatomical proximity. Nasopharyngeal, vaginal, and penile cancers were excluded due to the lack of publicly available GWAS data meeting IV selection requirements and statistical power. All included outcome data were based on populations of European ancestry to minimize bias from population stratification.

2.8.3. Statistical Analysis

Inverse Variance Weighted (IVW) meta-analysis was used as the primary method for estimating causal effects. Sensitivity analyses were performed to ensure robustness: (1) MR-Egger regression to assess and correct for pleiotropy, determining directional pleiotropy via the intercept term (P < 0.05)³⁴; (2) Weighted Median method, which provides unbiased estimates even if up to 50% of IVs are invalid³⁵; (3) MR-PRESSO to detect and remove outlier SNPs.³⁶ Heterogeneity among IVs was assessed using Cochran’s Q statistic.³⁷ All analyses were implemented using the TwoSampleMR package (version 0.5.8) in R software (version 4.3.1), with P < 0.05 considered statistically significant.

3. Results

3.1. Screening and Identification of Potential Saccharin Targets

Predicted and known interaction data from seven databases were integrated. Analysis was limited to Homo sapiens. Starting with the 2D structure (Figure 2A) and molecular details (Table 2) of sodium saccharin from PubChem, the isomeric SMILES string was used to search SwissTargetPrediction with a probability threshold > 0.01. The SEA method predicted targets based on chemical similarity. Merging and deduplicating results from all databases yielded 191 unique gene targets.

Figure 2.

Identification of common targets. (A) Standard 2D structural formula of sodium saccharin; (B) Venn diagram showing the intersection of 191 saccharin targets and 21,226 cancer targets

Table 2.

Molecular Details of Saccharin (PubChem CID 5143)

Property name	Property value	Reference
Molecular Weight	183.19g/mol	Computed by PubChem 2.2 (PubChem release 2025.04.14)
Molecular Formula	C₇H₅NO₃S	Computed by PubChem 2.2 (PubChem release 2025.04.14)
SMILES	C1=CC=C2C(=C1)C(=O)NS2(=O)=O	Computed by OEChem 2.3.0 (PubChem release 2025.04.14)

Note. Detailed information about saccharin molecules.(National Center for Biotechnology Information (2025). PubChem Compound Summary for CID 5143, Saccharin. Retrieved December 22, 2025 from https://pubchem.ncbi.nlm.nih.gov/compound/Saccharin).

3.2. Acquisition of Targets for 36 Cancers

A dataset of human cancer-related targets was constructed by merging targets for 36 cancers from four databases. After rigorous data cleaning to remove duplicates and false positives, and standardization via UniProt, an integrated dataset containing 21,226 unique gene/protein targets was obtained. This dataset covers genetic, molecular, and therapeutic targets. Intersection analysis identified 20 common targets shared across all 36 cancers (Table 3).

Table 3.

Common Targets Shared Across 36 Cancer Types

Standard gene symbol	UniProt id	Protein name	Known core functions/Relatedpathways
TP53	P04637	Cellular tumor antigen p53	Master tumor suppressor; orchestrates cell cycle arrest, apoptosis, or senescence in response to stress signals such as DNA damage.
MYC	P01106	Myc proto-oncogene protein	Transcription factor governing cell proliferation, growth, metabolism, differentiation, and apoptosis; frequently overexpressed in cancers.
VEGFA	P15692	Vascular endothelial growth factor A	Key angiogenic factor; stimulates endothelial cell proliferation, migration, and vascular permeability to promote tumor blood supply.
IL6	P05231	Interleukin-6	Pleiotropic pro-inflammatory cytokine; modulates immune response, hematopoiesis, inflammation, and the tumor microenvironment.
TNF	P01375	Tumor necrosis factor	Major pro-inflammatory cytokine; central regulator of inflammation, immunity, and programmed cell death (necrosis); dual role in tumorigenesis.
EGFR	P00533	Epidermal growth factor receptor	Receptor tyrosine kinase; activates downstream pathways (e.g., MAPK, PI3K-Akt) to drive cell proliferation, survival, and migration.
MAPK3	P27361	Mitogen-activated protein kinase 3	Core component of the MAPK/ERK cascade; transduces mitogenic signals to regulate cell proliferation, differentiation, and survival.
MAPK1	P28482	Mitogen-activated protein kinase 1	Functions redundantly with MAPK3 as a key effector kinase within the Ras-Raf-MEK-ERK signaling pathway.
HSPA5	P11021	Endoplasmic reticulum chaperone BiP	Master regulator of endoplasmic reticulum stress; facilitates protein folding and the unfolded protein response; promotes tumor cell survival and chemoresistance.
IL1B	P01584	Interleukin-1 beta	Potent pro-inflammatory cytokine; activated by the inflammasome; drives inflammatory responses and influences tumor progression.
JUN	P05412	Transcription factor AP-1	Constituent of the AP-1 transcription complex; responds to growth factors and stress to regulate cell proliferation and transformation.
BCL2	P10415	Apoptosis regulator Bcl-2	Prototypic anti-apoptotic protein; inhibits mitochondrial outer membrane permeabilization to block programmed cell death.
CCL2	P13500	C-C motif chemokine 2	Key chemokine for recruiting monocytes/macrophages to inflammatory and tumor sites; shapes an immunosuppressive microenvironment.
BAX	Q07812	Apoptosis regulator BAX	Pro-apoptotic Bcl-2 family protein; promotes mitochondrial cytochrome c release to initiate the intrinsic apoptosis pathway.
NOS2	P35228	Nitric oxide synthase, inducible	Induced under inflammatory conditions to produce high-output nitric oxide; exhibits immunomodulatory and cytotoxic functions in cancer.
CXCL8	P10145	Interleukin-8	Potent neutrophil chemotactic factor; also promotes angiogenesis and tumor cell migration, strongly linked to cancer aggressiveness.
CASP3	P42574	Caspase-3	Executioner caspase; cleaves numerous cellular substrates upon activation, leading to apoptotic cell death.
SLC2A1	P11166	Solute carrier family 2, facilitated glucose transporter member 1	Principal high-affinity glucose transporter; frequently upregulated in cancer cells to sustain aerobic glycolysis (Warburg effect).
BIRC5	O15392	Baculoviral IAP repeat-containing protein 5	Unique bifunctional protein that inhibits apoptosis and regulates mitotic progression; overexpressed in most human tumors.
EGR1	P18146	Early growth response protein 1	Immediate-early gene-encoded transcription factor; induced by growth factors and stress to modulate proliferation, apoptosis, and immune-related genes.

Note. List of the 20 common core targets shared across all 36 investigated cancer types. These targets primarily converge on pathways related to cell cycle regulation, apoptosis, and immune-inflammatory responses.

3.3. Identification of Common Disease-Compound Targets

Comparing the 191 saccharin targets with the 21,226 cancer targets using a Venn diagram revealed 160 intersection targets (Figure 2B). This high overlap (approximately 83.8%) suggests that saccharin targets and cancer-related targets are closely linked, implying saccharin may participate in core biological processes related to cancer through these shared targets.

To explore the relationship between saccharin targets and specific cancer types, the 36 cancers were classified into 10 anatomical system groups: head, neck, thoracic organs, digestive system, digestive glands, urinary system, female reproductive system, male reproductive system, hematologic/lymphatic system, and skin/soft tissue tumors. Intersection analysis between the 191 saccharin targets and the target sets of these 10 categories revealed the “saccharin-cancer anatomical classification intersection targets” (Figure 3A–J). The anatomical system-based grouping was designed to characterize the organ-level distribution patterns of saccharin targets. By identifying targets consistently shared across multiple anatomical systems, this approach enables the prioritization of genes with systemic pleiotropic relevance, thereby reducing bias introduced by organ-specific noise and strengthening the robustness of downstream core target identification.

Figure 3.

Organ distribution characteristics. Venn diagrams of intersection targets across 10 anatomical cancer groups (A-J), providing a logical basis for downstream core target screening

3.4. Construction and Analysis of the Saccharin-Target-Cancer Multidimensional Interaction Network

A three-layer integrated network was constructed to reveal the potential pleiotropic mechanism of saccharin. Network analysis identified multiple core targets with high degree values located at hub positions. The constructed compound-target-disease network (Figure 4) showed that saccharin acts on 144 core targets forming an extensive association network with various cancer types, indicating a mechanism of cross-cancer effects through multi-target synergistic action.

Figure 4.

Multidimensional saccharin-target-cancer interaction network. The network illustrates the pleiotropic mapping from the compound (saccharin, top right) to 144 core targets (middle layer), and subsequently to the 36 specific cancer types (left layer). Node size and color intensity indicate the topological degree

3.5. PPI Network Construction and Core Hub Target Selection

The PPI network of saccharin-cancer intersection targets (Figure 5A) was constructed to identify key mechanisms. Topological analysis indicated a centralized network structure. Using multiple centrality algorithms, a group of core hub targets was identified, with ALB, IL6, TNF, and MAPK3 exhibiting prominent values for degree and betweenness centrality (Figure 5B). The “Saccharin-Core Target-Cancer” integrated network (Figure 5C) demonstrates that saccharin is widely associated with multiple cancers by regulating a functional module of highly interconnected core targets.

Figure 5.

Protein-protein interaction (PPI) network analysis. (A) Overall PPI network of intersection targets constructed via string database; (B) Bar chart ranking the top 20 core hub targets based on degree centrality; (C) The highly interconnected “Saccharin-core target-cancer” functional module extracted using CytoHubba, highlighting central nodes such as ALB, IL6, TNF, and MAPK3

3.6. GO and KEGG Functional Enrichment Analysis of Core Targets

Functional enrichment analysis of the 144 core targets using the DAVID database revealed 804 significantly enriched biological processes, 21 cellular component terms, and 38 molecular function terms (Figure 6A). Top terms indicated involvement in cytokine-mediated signaling pathways, regulation of apoptosis, and negative regulation of cell proliferation. Targets were primarily located in the cytoplasm and nucleus, with molecular functions concentrated in enzyme binding, transcription factor binding, and cytokine receptor binding. KEGG pathway analysis identified 150 significantly enriched pathways (Figure 6B), including the interleukin-17 (IL-17) signaling pathway, AGE-RAGE signaling pathway, cAMP signaling pathway, lipid and atherosclerosis pathway, and nitrogen metabolism pathway. These pathways may play a central regulatory role in the process by which saccharin affects cancer.

Figure 6.

Functional enrichment analysis of core targets. (A) Top Gene Ontology (GO) terms across Biological Process, Cellular Component, and Molecular Function categories; (B) Sankey bubble plot of the significantly enriched KEGG pathways, illustrating the mapping of specific genes to key pathways

3.7. Molecular Docking Validation and Binding Mode Analysis

Molecular docking validated the interaction between saccharin and 13 core protein targets. A bubble heatmap showed binding energies consistently below -5 kcal/mol (Figure 7A). Visualization of selected lower-energy bindings (Figure 7B–F) revealed specific interactions: saccharin formed hydrogen bonds with residues ARG-257, ALA-261, and SER-287 of ALB (-6.6 kcal/mol); ARG-155 of FOS (-7.0 kcal/mol); DA-31, YS-273, and DG-30 of JUN (-6.8 kcal/mol); PRO-315 and ASN-316 of MAPK3 (-6.4 kcal/mol); and MET-154 and LYS-157 of IL10 (-6.4 kcal/mol). The visualization of docking complexes (Figure 7B–F) suggests that saccharin may interact with functionally relevant regions of these targets. For ALB, saccharin is located within a hydrophobic binding pocket and forms hydrogen bonds with ARG-257, ALA-261, and SER-287. For FOS and JUN, saccharin binds near specific residues (e.g., ARG-155, DA-31) involved in transcriptional regulation. For MAPK3, binding occurs in proximity to the catalytic pocket (PRO-315, ASN-316), which may influence kinase activity. Similarly, interaction with IL10 involves residues located in functionally relevant regions (MET-154, LYS-157). These findings provide preliminary structural clues but require future experimental validation.

Figure 7.

Molecular docking results and visualization. (A) Bubble heatmap of binding energies for core targets; (B-F) Visualization of specific binding sites and hydrogen bond formation for ALB, FOS, IL10, JUN, and MAPK3

3.8. Causal Association Between Saccharin and Cancer via Mendelian Randomization

3.8.1. Instrumental Variables and Outcome Data

From the GWAS summary data, 10 SNPs significantly associated with saccharin levels (P < 5 × 10^-6) were initially screened. After linkage disequilibrium clumping (r² < 0.001, distance > 10,000 kb) and F-statistic checking (>10), 6 independent strong IVs were retained. The mean F-statistic was 22.59, avoiding weak instrument bias. GWAS summary data for 32 cancers were included (Table 4). Oropharyngeal and hypopharyngeal cancers were merged; nasopharyngeal, vaginal, and penile cancers were excluded. All data were based on European ancestry.

Table 4.

GWAS Summary Data Sources for the 32 Cancer Outcomes Analyzed in Mendelian Randomization

No.	Trait/Phenotype	Type	Ancestry	GWAS catalog id	PMID	Notes
1	Saccharin level	Exposure	European	GCST90243535	24816252
2	Brain cancer	Outcome	European	GCST90041899	34737426
3	Oral cancer	Outcome	European	GCST90423171	39528825
4	Salivary gland cancer	Outcome	European	-	-
5	Thyroid cancer	Outcome	European	GCST90475600	39024449
6	Laryngeal cancer	Outcome	European	GCST90435574	30104761
7	Oropharyngeal cancer	Outcome	European	GCST90018898	34594039	coincidence with hypopharyngeal cancer as pharyngeal cancer
8	Hypopharyngeal cancer	Outcome	European	GCST90018898	34594039	coincidence with oropharyngeal cancer as pharyngeal cancer
9	Lung cancer	Outcome	European	GCST90479533	39024449
10	Esophageal cancer	Outcome	European	GCST90477174	39024449
11	Breast cancer	Outcome	European	GCST90551892	40229280
12	Gastric cancer	Outcome	European	GCST90043847	34737426
13	Colon cancer	Outcome	European	GCST90011811	32887889
14	Rectal cancer	Outcome	European	GCST90011810	32887889
15	Anal cancer	Outcome	European	GCST90043915	34737426
16	Liver cancer	Outcome	European	GCST90479784	39024449
17	Pancreatic cancer	Outcome	European	GCST90479786	39024449
18	Gallbladder cancer	Outcome	European	GCST90435586	30104761
19	Kidney cancer	Outcome	European	GCST90011818	32887889
20	Bladder cancer	Outcome	European	GCST90041901	34737426
21	Cervical cancer	Outcome	European	GCST012776	35157032
22	Uterine corpus cancer	Outcome	European	GCST90041848	34737426
23	Ovarian cancer	Outcome	European	GCST90011821	32887889
24	Vulvar cancer	Outcome	European	GCST90043931	34737426
25	Prostate cancer	Outcome	European	GCST90428117	38538606
26	Testicular cancer	Outcome	European	GCST90041906	34737426
27	Non-Hodgkin lymphoma	Outcome	European	GCST90043909	34737426
28	Hodgkin lymphoma	Outcome	European	GCST90043904	34737426
29	Leukemia	Outcome	European	GCST90435647	30104761
30	Multiple myeloma	Outcome	European	GCST90477249	39024449
31	Nonmelanoma skin cancer	Outcome	European	GCST90475188	39024449
32	Melanoma	Outcome	European	GCST90468141	39789286
33	Kaposi sarcoma	Outcome	European	GCST90041920	34737426
34	Mesothelioma	Outcome	European	GCST90043881	34737426

Note. Detailed source information for the GWAS summary statistics used in the Mendelian randomization analysis. All datasets were restricted to populations of European ancestry to minimize population stratification bias.

3.8.2. Causal Association Between Saccharin and Cancer Risk

Primary IVW analysis (Figure 8) showed statistically significant associations between genetically predicted higher saccharin levels and risk in three of the 32 cancers. Saccharin showed a protective association with non-melanoma skin cancer (OR = 0.38, 95% CI: 0.19–0.76, P = 0.006) and colon cancer (OR = 0.56, 95% CI: 0.41–0.78, P < 0.001). Conversely, a positive association was observed with Kaposi sarcoma risk (OR = 3.73, 95% CI: 1.04–13.40, P = 0.044). No statistically significant causal associations were found for the remaining 29 cancers, including common types such as brain, breast, lung, and prostate cancer (P > 0.05).

Figure 8.

Forest plot of Mendelian randomization analysis showing the causal association between genetically predicted sodium saccharin levels and cancer risk. OR: Odds Ratio; CI: Confidence Interval

3.8.3. Sensitivity Analysis

MR-Egger regression detected no significant directional pleiotropy (intercept P > 0.05), with estimate directions consistent with IVW. The Weighted Median method also showed consistent directions and magnitudes. MR-PRESSO did not identify any outlier SNPs. Heterogeneity testing indicated heterogeneity in bladder and gastric cancer analyses (Cochran’s Q P < 0.05); however, associations remained non-significant after correction using a random-effects model. Heterogeneity tests for the three significant findings (Kaposi sarcoma, non-melanoma skin cancer, colon cancer) yielded P > 0.05, supporting the reliability of the estimates.

4. Discussion

Artificial sweeteners are widely used in the global food supply, making their long-term health implications an important issue in toxicology and public health. Sodium saccharin, one of the oldest artificial sweeteners, is considered safe within regulatory frameworks, yet its potential biological effects, particularly regarding cancer risk, remain a subject of scientific debate. This study integrated computational toxicology and genetic epidemiology to systematically explore the multidimensional relationship between sodium saccharin and 36 cancers. Through the construction of a Saccharin-Target-Cancer interaction network, PPI analysis, functional enrichment, molecular docking, and MR analysis, potential molecular targets and pathways were revealed, and causal associations with specific cancers were preliminarily assessed from a genetic perspective.

Integrating data from seven databases, 191 potential saccharin targets were identified. Comparison with 21,226 targets related to 36 cancers revealed 160 intersection targets, an overlap of 83.8%. This high overlap suggests extensive intersection between saccharin targets and key molecules in tumorigenesis. The intersection set included TP53, MYC, EGFR, IL6, and TNF. TP53 is a major tumor suppressor gene mutated or inactivated in over 50% of human cancers.³⁸ MYC is a key transcription factor regulating cell proliferation and metabolism, with abnormal expression closely linked to tumor progression.^39,40 Potential interactions between saccharin and these genes suggest interference with basic biological processes such as cell cycle regulation, apoptotic escape, or immune surveillance. Although this does not directly prove carcinogenicity, the ability of saccharin or its metabolites to bind these key targets may affect their normal function under long-term or high-dose exposure.

Network topology analysis showed that the PPI network of saccharin-cancer intersection targets was highly centralized, with nodes such as ALB, IL6, TNF, and MAPK3 exhibiting prominent centrality. These hub genes are often at the core of signal transduction or regulatory networks, where functional perturbation can propagate and affect multiple downstream physiological processes. Chronic inflammation is a driver of cancer development, promoting tumor progression, immune escape, and therapy resistance via signaling pathways such as NF-κB, JAK-STAT, and MAPK.⁴¹ IL-6 and TNF-α play central roles in the tumor microenvironment. IL-6 is a key driver in breast cancer multi-omics networks and promotes tumor growth via synergy with proteins like CRP.⁴² TNF-α activates pathways such as NF-κB in head and neck squamous cell carcinoma, mediating invasion, metastasis, and shaping the immunosuppressive microenvironment.⁴³ Network analysis also highlighted ALB as a hub gene in clear cell renal cell carcinoma and MAPK3 as significantly associated with immune infiltration in glioma.⁴⁴ Targeting inflammatory factors is currently a focus of clinical research for cancer therapy.

Biologically, these targets were significantly enriched in cytokine-mediated signaling pathways, negative regulation of apoptosis, and negative regulation of cell proliferation, suggesting saccharin may participate in regulating the balance between immune response and cell survival. Pathway analysis showed significant enrichment in IL-17, AGE-RAGE, cAMP, lipid and atherosclerosis, and nitrogen metabolism pathways. The IL-17 pathway links inflammation and tumor progression; abnormal activation promotes proliferation and angiogenesis. IL-17/IL-17RA signaling is essential for tumorigenesis in pancreatic epithelial cells, recruiting inhibitory immune cells via CXCL5 and B7-H4 upregulation.⁴⁵ Interaction between antigen-presenting cells and tumor antigens can promote IL-10 and IL-17 expression, weakening anti-tumor immunity.⁴⁶ The AGE-RAGE pathway involves advanced glycation end productsbinding to RAGE, activating PI3K-Akt, p38 MAPK, and NF-κB pathways to drive chronic inflammation and cancer.⁴⁷ Recent research confirmed that advanced glycation end products induce oxidative stress and apoptosis in intestinal cells via the RAGE/MAPK/NF-κB pathway.⁴⁸ The co-enrichment of these pathways suggests the influence of saccharin is not limited to single molecular events but may involve interference with multiple interconnected signaling networks, indirectly shaping a microenvironment favorable for tumorigenesis.

It is worth noting that the enrichment of cAMP signaling and nitrogen metabolism pathways may reflect potential effects of saccharin on cellular energy metabolism and nitrogen balance. Tumor cells often exhibit the Warburg effect, and cAMP regulates cell proliferation via PKA/CREB pathways. Nitrogen metabolism is tightly linked to amino acid utilization and nucleotide synthesis.⁴⁹ An integrative analysis found that saccharin targets were significantly enriched in Pathways in cancer, PI3K-Akt signaling, and chemical carcinogenesis-receptor activation.⁵⁰ Although direct evidence of saccharin-induced metabolic reprogramming in humans is lacking, potential effects under long-term low-dose exposure warrant further investigation.

Molecular docking validated the interaction between saccharin and 13 core protein targets. Saccharin showed potential binding affinity (energy < -5 kcal/mol) with ALB, FOS, JUN, MAPK3, and IL10.⁵¹ Binding to serum albumin (ALB, -6.6 kcal/mol) via hydrogen bonds suggests specific transport characteristics, as small molecule-albumin interactions determine pharmacokinetics. Furthermore, such systemic transport may alter the pharmacokinetic properties of other endogenous or exogenous compounds.⁵² Glycation can alter the HSA binding pocket.⁵³ Binding near regulatory regions of FOS and JUN raises the possibility of modulating the AP-1 transcriptional complex, which is heavily involved in cellular proliferation, migration, and invasion.⁵⁴ Interaction with MAPK3 (ERK1), a core kinase in the MAPK/ERK pathway, involved residues PRO-315 and ASN-316 near the active pocket, potentially affecting kinase conformation or substrate binding. Previous studies also found saccharin binds stably to MAPK3 (molecular docking < -5.0 kcal/mol; SPR KD = 45.0 μM) and may activate pro-inflammatory pathways.⁵⁵ Furthermore, association with IL10 may indicate potential involvement in immunological homeostasis and inflammation modulation.⁵⁶ However, outside of previously validated targets, these interpretations remain structural hypotheses and should be considered strictly as hypothesis-generating rather than evidence of direct functional effects, warranting rigorous in vivo and in vitro validation in future studies.

Two-sample MR analysis revealed that genetically predicted higher saccharin levels were significantly associated with three cancers: negatively with non-melanoma skin cancer and colon cancer, and positively with Kaposi sarcoma. No significant causal association was found for the other 29 cancers. These results indicate that potential effects are highly heterogeneous across cancer types rather than pan-carcinogenic. Our findings of considerable heterogeneity across cancer types are consistent with emerging evidence. For instance, a recent integrative analysis by Xie et al suggested that artificial sweeteners might increase the risk of kidney, breast, and prostate cancers through multiple signaling pathways.² In contrast, our genetic causality assessment did not support significant associations for these specific cancers. Furthermore, the vehicle of consumption appears critical. Pan et al reported no genetic causal association between overall artificial sweetener consumption and most major cancers.⁵⁷ Notably, however, their stratified analyses revealed that artificial sweeteners added to coffee were associated with ovarian cancer risks, while those in tea were linked to oral and pharyngeal cancers.⁵⁷ Similarly, Zhang et al identified a specific positive association between artificial sweeteners in cereal and respiratory system cancers, whereas perceived aspartame intensity showed a protective effect.¹² The positive association with Kaposi sarcoma (OR = 3.73) in this study, though with a wide confidence interval and marginal significance (P = 0.044), along with protective associations for skin and colon cancers, highlights that health effects may depend heavily on cancer type and exposure context. The positive association with Kaposi sarcoma warrants cautious interpretation. Since Kaposi sarcoma is causally linked to Human Herpesvirus 8 infection, this association might reflect a potential influence of saccharin on immune surveillance pathways, such as the identified IL-6 and TNF signaling, rather than a direct carcinogenic effect. Furthermore, given the limited sample size for this rare cancer type, the possibility of pleiotropic bias cannot be excluded. The protective association observed in colon cancer contrasts with concerns regarding artificial sweeteners and gut microbiota dysbiosis. However, these results rely on genetic predisposition, reflecting lifelong exposure patterns rather than acute intake. Mechanistically, our network analysis indicated interactions between saccharin and inflammatory mediators, including NF-κB and cytokines, which may contribute to an anti-inflammatory microenvironment that counteracts tumorigenesis.

This study constructed a multi-level framework combining network toxicology, molecular docking, and MR. This integration validated the link between molecular mechanism prediction and population causal inference, addressing limitations of traditional toxicology such as high cost and difficulty in establishing causality.

Limitations exist. Network toxicology relies on database predictions, carrying the potential for false positives or negatives. Molecular docking is a static simulation and cannot reflect dynamic processes or concentration-dependence.⁵⁸ MR results suggest causal direction but cannot replace prospective studies.⁵⁹ Furthermore, specifically regarding the MR analysis, a key limitation of this study lies in the nature of the instrumental variables. As saccharin is an exogenous compound, its circulating levels are influenced by multiple factors, including dietary intake, gut microbiota interactions, and renal excretion. The selected SNPs from the metabolomic GWAS therefore represent metabolite quantitative trait loci (mQTLs), which primarily capture genetic variation in internal handling processes rather than direct intake levels. Consequently, these instruments reflect the biological processing and net internal exposure of saccharin rather than exposure per se. Although this limits interpretability in terms of absolute dietary consumption, it remains highly informative for assessing the potential biological consequences of long-term internal exposure. Finally, alongside the inherent limitations of potential pleiotropy and restriction to European populations, broader exposure contexts must also be evaluated. Dose-response relationships in real-world exposure, interactions with other additives, or individual susceptibility differences were not considered. While saccharin is permitted and likely low-risk at normal dietary levels, potential bioactivity, especially for individuals with chronic inflammation or metabolic abnormalities, warrants exploration. Crucially, these findings should be interpreted within the context of established dietary exposure limits, distinct from high-dose toxicity models.

5. Future Directions and Clinical Relevance

Future research should experimentally validate core targets like MAPK3, IL6, and ALB; use animal models or organoids to assess effects on inflammation and metabolism under realistic exposure; and conduct prospective cohort studies combined with biomarkers to clarify long-term effects.

6. Conclusion

In conclusion, our integrated computational and MR study reveals that low-level, physiological sodium saccharin exposure does not causally increase pan-cancer risk across the general population. Instead, its novel biological impact lies in its subtle interaction with specific immune and metabolic targets (such as ALB, FOS, and IL10). These findings shift the paradigm from broad carcinogenic concerns to targeted immunomodulatory surveillance, providing a robust evidence base for modern food safety re-evaluations.

Footnotes

Acknowledgements

We thank the original authors and consortia for sharing the GWAS summary datasets used in this study.

ORCID iDs

Zhao Xin

Wang Yanchun

Ethical Considerations

Not applicable. This study utilizes publicly available aggregate data and does not involve any new animal or human experiments performed by the authors.

Author Contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by the authors. All authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant No. U1804181 to Yanchun Wang); the Collaborative Research Project of Chinese and Western Medicine for Chronic Disease Management (Grant No. CXZH2024029 to Yanchun Wang); the Special Scientific Research Project of Traditional Chinese Medicine of Henan Province (Grant No. 2024ZY2134 to Yanchun Wang); the Top Talent Program of Traditional Chinese Medicine of Henan Provincial Health Commission (Grant No. Yu Wei Zhong Yi Ke Ke Ji [2025] 14 to Yanchun Wang); the Postgraduate Scientific Research Innovation Ability Enhancement Plan Project of Henan University of Chinese Medicine in 2024 (Grant No. 2024KYCX018 to Xin Zhao); and the Science and Technology Key Project of Henan Province (Grant No. 242102311284 to Xuemei Wang).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The datasets presented in this study can be found in online repositories. The names of the repositories and accession numbers can be found in the article. Publicly available GWAS summary statistics were obtained from the GWAS Catalog (https://www.ebi.ac.uk/gwas/) and saccharin structure data from PubChem ().

References

Bray

Laversanne

Sung

, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229-263. doi:10.3322/caac.21834

Xie

Zhu

Yang

Wang

. An integrative analysis reveals cancer risk associated with artificial sweeteners. J Transl Med. 2025;23(1):32. doi:10.1186/s12967-024-06047-0

Magdy

Abdelkader

Zaki

Kamel

. Potential exacerbation of polycystic ovary syndrome by saccharin sodium via taste receptors in a letrozole rat model. Food Chem Toxicol. 2024;191:114874. doi:10.1016/j.fct.2024.114874

Becker

Dreier

Manibusan

Tony Cox

Simon

Bus

. How well can carcinogenicity be predicted by high throughput “characteristics of carcinogens” mechanistic data? Regul Toxicol Pharmacol. 2017;90:185-196. doi:10.1016/j.yrtph.2017.08.021

Smith

Guyton

Kleinstreuer

, et al. The key characteristics of carcinogens: Relationship to the hallmarks of cancer, relevant biomarkers, and assays to measure them. Cancer Epidemiol Biomarkers Prev. 2020;29(10):1887-1903. doi:10.1158/1055-9965.EPI-19-1346

Marchitti

Boon

Jack

Goodman

. Lack of genotoxic and carcinogenic potential for nonsugar sweeteners: A review of animal and mechanistic evidence. Adv Nutr. 2025;16(12):100552. doi:10.1016/j.advnut.2025.100552

Salama

Al-Rashidi

. Saccharin from sweetener to immune orchestrator: Saccharin-induced neutrophil chemotaxis and transmigration via taste receptor signaling. Immunol Res. 2025;73(1):149. doi:10.1007/s12026-025-09709-3

De Dios

Proctor

Maslova

Dzalbe

Rudolph

McCarthy

. Artificial sweeteners inhibit multidrug-resistant pathogen growth and potentiate antibiotic activity. EMBO Mol Med. 2023;15(1):e16397. doi:10.15252/emmm.202216397

Liu

Wang

Gong

. Taste receptor T1R3 regulates testosterone synthesis via the cAMP-PKA-SP1 pathway in testicular leydig cells. Theriogenology. 2025;231:210-221. doi:10.1016/j.theriogenology.2024.10.019

10.

EFSA Panel on Food Additives and Flavourings (FAF) Castle

Andreassen

, et al. Re-evaluation of saccharin and its sodium, potassium and calcium salts (E 954) as food additives. EFSA J. 2024;22(11):e9044. doi:10.2903/j.efsa.2024.9044

11.

Arulanandam

Babu

Soorni

Prathiviraj

. Mutagenicity and carcinogenicity prediction of sugar substitutes: An in silico approach with compound-gene interactions network. Toxicol Res. 2024;14(1):tfaf008. doi:10.1093/toxres/tfaf008

12.

Zhang

Fan

Mao

, et al. Artificial sweetener and respiratory system cancer: A mendelian randomization analysis. Clin Nutr ESPEN. 2024;63:259-266. doi:10.1016/j.clnesp.2024.06.008

13.

Enuwosa

Gautam

King

Chichger

. Saccharin and sucralose protect the glomerular microvasculature in vitro against VEGF-induced permeability. Nutrients. 2021;13(8):2746. doi:10.3390/nu13082746

14.

Daina

Michielin

Zoete

. SwissTargetPrediction: Updated data and new features for efficient prediction of protein targets of small molecules. Nucleic Acids Res. 2019;47(W1):W357-W364. doi:10.1093/nar/gkz382

15.

Keiser

Roth

Armbruster

Ernsberger

Irwin

Shoichet

. Relating protein pharmacology by ligand chemistry. Nat Biotechnol. 2007;25(2):197-206. doi:10.1038/nbt1284

16.

Davis

Wiegers

Johnson

Sciaky

Wiegers

Mattingly

. Comparative toxicogenomics database (CTD): Update 2023. Nucleic Acids Res. 2023;51(D1):D1257-D1262. doi:10.1093/nar/gkac833

17.

Szklarczyk

Santos

von Mering

Jensen

Bork

Kuhn

. STITCH 5: Augmenting protein-chemical interaction networks with tissue and affinity data. Nucleic Acids Res. 2016;44(D1):D380-D384. doi:10.1093/nar/gkv1277

18.

Nickel

Gohlke

Erehman

, et al. SuperPred: Update on drug classification and target prediction. Nucleic Acids Res. 2014;42(W1):W26-W31. doi:10.1093/nar/gku477

19.

Stelzer

Rosen

Plaschkes

, et al. The GeneCards suite: From gene data mining to disease genome sequence analyses. Curr Protoc Bioinforma. 2016;54(1):1, 30.1-1.30.33. doi:10.1002/cpbi.5

20.

Zhou

Zhang

Zhao

, et al. TTD: therapeutic target database describing target druggability information. Nucleic Acids Res. 2024;52(D1):D1465-D1477. doi:10.1093/nar/gkad751

21.

Sung

Ferlay

Siegel

, 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(3):209-249. doi:10.3322/caac.21660

22.

Bardou

Mariette

Escudié

Djemiel

Klopp

. jvenn: An interactive venn diagram viewer. BMC Bioinformatics. 2014;15(1):293. doi:10.1186/1471-2105-15-293

23.

Shen

Wang

Zhang

, et al. Exploring the active components and potential mechanisms of rosa roxburghii tratt in treating type 2 diabetes mellitus based on UPLC-Q-exactive orbitrap/MS and network pharmacology. Chin Med. 2023;18(1):12. doi:10.1186/s13020-023-00713-z

24.

Ullanat

Jing

Sledzieski

Berger

. Learning the language of protein-protein interactions. Nat Commun. 2026;17(1):1199. doi:10.1038/s41467-025-67971-3

25.

Doncheva

Assenov

Domingues

Albrecht

. Topological analysis and interactive visualization of biological networks and protein structures. Nat Protoc. 2012;7(4):670-685. doi:10.1038/nprot.2012.004

26.

Kanehisa

Furumichi

Sato

Kawashima

Ishiguro-Watanabe

. KEGG for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res. 2023;51(D1):D587-D592. doi:10.1093/nar/gkac963

27.

, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The Innovation. 2021;2(3):100141. doi:10.1016/j.xinn.2021.100141

28.

Deng

, et al. Multi-omics analysis reveals prognostic value of tumor mutation burden in hepatocellular carcinoma. Cancer Cell Int. 2021;21(1):342. doi:10.1186/s12935-021-02049-w

29.

Wang

Yang

, et al. Cuproptosis related genes associated with Jab1 shapes tumor microenvironment and pharmacological profile in nasopharyngeal carcinoma. Front Immunol. 2022;13:989286. doi:10.3389/fimmu.2022.989286

30.

Peng

Jiang

Song

, et al. ggalign: Bridging the grammar of graphics and biological multilayered complexity. Adv Sci. 2025;12(44):e07799. doi:10.1002/advs.202507799

31.

Burley

Bhikadiya

, et al. RCSB protein data bank (RCSB.org): Delivery of experimentally-determined PDB structures alongside one million computed structure models of proteins from artificial intelligence/machine learning. Nucleic Acids Res. 2023;51(D1):D488-D508. doi:10.1093/nar/gkac1077

32.

The Multiple Tissue Human Expression Resource (MuTHER) Consortium Shin

Fauman

, et al. An atlas of genetic influences on human blood metabolites. Nat Genet. 2014;46(6):543-550. doi:10.1038/ng.2982

33.

Skrivankova

Richmond

Woolf

BAR

, et al. Strengthening the reporting of observational studies in epidemiology using mendelian randomisation (STROBE-MR): Explanation and elaboration. BMJ. 2021;26:n2233. doi:10.1136/bmj.n2233. Published online October.

34.

Leung

PBM

Liu

Zhong

, et al. Bidirectional two-sample mendelian randomization study of differential white blood cell counts and schizophrenia. Brain Behav Immun. 2024;118:22-30. doi:10.1016/j.bbi.2024.02.015

35.

Galimberti

Levey

Deak

Zhou

Stein

Gelernter

. Genetic influences and causal pathways shared between cannabis use disorder and other substance use traits. Mol Psychiatry. 2024;29(9):2905-2910. doi:10.1038/s41380-024-02548-y

36.

Liu

Wang

Yang

Zou

. Causal associations between type 1 diabetes mellitus and cardiovascular diseases: A mendelian randomization study. Cardiovasc Diabetol. 2023;22(1):236. doi:10.1186/s12933-023-01974-6

37.

Anabtawi

Pasala

Grimshaw

, et al. Low skeletal muscle mass and treatment outcomes among adults with haematologic malignancies: A systematic review and meta-analysis. J Cachexia Sarcopenia Muscle. 2024;15(3):1084-1093. doi:10.1002/jcsm.13446

38.

Zhou

Hao

. Mutant p53 in cancer therapy—the barrier or the path. J Mol Cell Biol. 2019;11(4):293-305. doi:10.1093/jmcb/mjy072

39.

Ding

, et al. PDCD11 stabilizes C-MYC oncoprotein by hindering C-MYC-SKP2 negative feedback loop to facilitate progression of p53-mutant breast and colon malignancies. Adv Sci. 2025;12(17):2502416. doi:10.1002/advs.202502416

40.

Dang

. MYC, metabolism, cell growth, and tumorigenesis. Cold Spring Harb Perspect Med. 2013;3(8):a014217. doi:10.1101/cshperspect.a014217

41.

Zhao

Yan

, et al. Inflammation and tumor progression: Signaling pathways and targeted intervention. Signal Transduct Target Ther. 2021;6(1):263. doi:10.1038/s41392-021-00658-5

42.

Jung

Papp

Pellegrini

Sobel

. Molecular biology networks and key gene regulators for inflammatory biomarkers shared by breast cancer development: Multi-omics systems analysis. Biomolecules. 2021;11(9):1379. doi:10.3390/biom11091379

43.

Long

Huang

Meng

, et al. TNF patterns and tumor microenvironment characterization in head and neck squamous cell carcinoma. Front Immunol. 2021;12:754818. doi:10.3389/fimmu.2021.754818

44.

Ren

Sun

Zhang

Yang

Cao

. MAPK activated protein kinase 3 is a prognostic-related biomarker and associated with immune infiltrates in glioma. Front Oncol. 2021;11:793025. doi:10.3389/fonc.2021.793025

45.

Peixoto

Miranda

Santos

Ferreira

. A roadmap for translational cancer glycoimmunology at single cell resolution. J Exp Clin Cancer Res. 2022;41(1):143. doi:10.1186/s13046-022-02335-z

46.

RodrÍguez

Schetters

STT

Van Kooyk

. The tumour glyco-code as a novel immune checkpoint for immunotherapy. Nat Rev Immunol. 2018;18(3):204-211. doi:10.1038/nri.2018.3

47.

Byun

Yoo

Son

, et al. Advanced glycation end-products produced systemically and by macrophages: A common contributor to inflammation and degenerative diseases. Pharmacol Ther. 2017;177:44-55. doi:10.1016/j.pharmthera.2017.02.030

48.

Zhang

Tian

Xie

. Exploring the protective effects of spinosin against advanced glycation end product-induced cellular damage through modulation RAGE/MAPK/NF-κB pathway. Food Biosci. 2025;64:105951. doi:10.1016/j.fbio.2025.105951

49.

Zou

Wang

Deng

, et al. CREB, another culprit for TIGAR promoter activity and expression. Biochem Biophys Res Commun. 2013;439(4):481-486. doi:10.1016/j.bbrc.2013.08.098

50.

Song

Yang

Wang

Lin

Wang

. Potential disease-related risk of saccharin in ovarian cancer: An integrative multi-evidence analysis. Reprod Toxicol. 2025;138:109044. doi:10.1016/j.reprotox.2025.109044

51.

Liu

Yang

Chen

Shao

Zhang

. Integrating network pharmacology and molecular docking to explore the pharmacological mechanism of tanshinone IIA in improving chronic obstructive pulmonary disease. Medicine (Baltimore). 2025;104(12):e41638. doi:10.1097/MD.0000000000041638

52.

Ashraf

Qaiser

Tariq

, et al. Unraveling the versatility of human serum albumin - a comprehensive review of its biological significance and therapeutic potential. Curr Res Struct Biol. 2023;6:100114. doi:10.1016/j.crstbi.2023.100114

53.

Das

Mandal

. Integrating machine learning and statistical methods for prediction and feature-based network characterization of stress response protein. Comput Biol Chem. 2025;119:108539. doi:10.1016/j.compbiolchem.2025.108539

54.

Song

Lian

Zhang

. The potential of activator protein 1 (AP-1) in cancer targeted therapy. Front Immunol. 2023;14:1224892. doi:10.3389/fimmu.2023.1224892

55.

Yang

Guo

, et al. Saccharin aggravates eosinophilic esophagitis via MAPK3 interaction: A network toxicology and machine learning study with SPR analysis. Food Sci Nutr. 2026;14(1):e71409. doi:10.1002/fsn3.71409

56.

Carlini

Noonan

Abdalalem

, et al. The multifaceted nature of IL-10: Regulation, role in immunological homeostasis and its relevance to cancer, COVID-19 and post-COVID conditions. Front Immunol. 2023;14:1161067. doi:10.3389/fimmu.2023.1161067

57.

Pan

Feng

Zhou

, et al. The causal association between artificial sweeteners and the risk of cancer: A mendelian randomization study. Food Funct. 2024;15(8):4527-4537. doi:10.1039/D3FO05756A

58.

Bruno

. Successes and failures of static aptamer-target 3D docking models. Int J Mol Sci. 2022;23(22):14410. doi:10.3390/ijms232214410

59.

Stender

Gellert-Kristensen

Smith

. Reclaiming mendelian randomization from the deluge of papers and misleading findings. Lipids Health Dis. 2024;23(1):286. doi:10.1186/s12944-024-02284-w