Cellular Stress, Inflammation and Barrier Damage in Gut Epithelial Cells Caused by Aspartame

ABSTRACT

Aspartame has been widely used as a sweetener in foods and beverages since the 1980s. In this study, we aimed to assess its effects on gut epithelial cell biology, inflammation and the epithelial barrier. We found that aspartame induces cytotoxic effects, disrupts the epithelial barrier and triggers proinflammatory cytokine release in gastrointestinal epithelial cells, organoids and gut‐on‐a‐chip models at concentrations corresponding to daily consumed doses in food products. Cellular cytotoxicity was observed at doses as low as 1.25 mg/mL. RNA sequencing analysis revealed that aspartame significantly alters the transcriptome in gut‐on‐a‐chip models, with upregulation of pathways involved in the unfolded protein response, pro‐apoptotic and inflammatory processes and downregulation of those related to DNA repair and replication. Aspartame exposure upregulated proinflammatory genes, particularly in the TNF signalling pathway, and induced multiple chemokine responses. It also activated the NF‐κB pathway via oxidative stress, promoting inflammation in NF‐κB reporter monocyte cells and leading to gut epithelial cell death. Additionally, aspartame affected genes involved in tight and adherens junctions, disrupting gut epithelial barrier integrity in a dose‐dependent manner. It further suppressed key DNA repair and replication genes associated with double‐strand break repair, mismatch repair and DNA replication. Overall, our findings indicate that, at commonly consumed levels, aspartame induces cellular stress, inflammation and epithelial barrier damage in gastrointestinal epithelial cells. These results underscore the biological relevance of our study and raise concerns that daily dietary intake of aspartame may pose previously underappreciated risks to gut health.

This study shows that aspartame, at dietary‐relevant doses, induces cytotoxicity, cellular distress associated with endoplasmic reticulum stress, oxidative stress, DNA replication and repair mechanism disturbance culminating in epithelial barrier damage. Aspartame triggers NF‐κB activation via induction of oxidative stress, leading to proinflammatory cytokine and chemokine release. Aspartame metabolites synergistically contribute to cytotoxicity, oxidative stress and epithelial barrier disruption.

Abbreviations

CXCL

X‐C chemokine ligand

NF‐κB

nuclear factor kappa B

RNAseq

RNA sequencing

TEER

transepithelial electrical resistance

TNF

tumour necrosis factor

1. Introduction

Aspartame is an artificial non‐nutritive sweetener composed of the amino acids phenylalanine and aspartic acid linked to a methyl group. It is approximately 200 times sweeter than sucrose, enabling manufacturers to achieve the same level of sweetness while using significantly smaller quantities, thus reducing the overall caloric content of foods and beverages [1]. The use of aspartame in foods and beverages as a sweetener dates to the 1980s. Since then, it has become widely used in tabletop sweeteners, chewing gums, food supplements and artificially sweetened beverages [2]. The acceptable daily intake limit of aspartame is established as 40 mg/kg body weight based on animal studies [2, 3]. According to the EU Commission regulation, the maximum permitted level of aspartame for various food and beverage categories is 6 mg/g (mg/ml). The data allocated from manufacturers show that tabletop sweeteners in powder, tablet, or liquid form could contain up to 500 mg/mL (mg/g) aspartame, about half of the volume [3] (Table 1). According to the Food and Drug Administration (FDA), aspartame is safe to use in food in accordance with good manufacturing practices as a sweetening agent and a flavour enhancer. Therefore, the FDA does not impose a general limitation on aspartame usage. The only limitation applies to baked goods. According to the FDA, when aspartame is used in baked goods or baking mixes, its amount must not exceed 0.5% of the product's total weight, whether in the ready‐to‐bake form or the finished formulation before baking [4].

TABLE 1.

Maximum permitted and reported use levels of aspartame across food categories.

	Max permitted level (mg/L)	Reported use levels (mg/L)
Table‐top sweeteners	quantum satis	8700–36,000
Table‐top sweeteners in tablets	quantum satis	100,000–360,000
Other confectionery including breath refreshing microsweets	1000–6000	100–1000
Flavoured fermented milk products including heat‐treated products	1000	50–1000
Edible ices	800	40
Processed fruit and vegetables	1000	350–1000
Cocoa and chocolate products as covered by Directive 2000/36/EC	2000	500–1000
Food supplements as defined in Directive 2002/46/EC[5] excluding food supplements for infants and young children	5500	—
Flavoured drinks with sweeteners	600	100–600
Alcoholic beverages, including alcohol‐free and low alcohol counterparts	600	100–600
Chewing gum	2500–5500	600–5420

Note: Data compiled from the Scientific Opinion on the Re‐evaluation of Aspartame (E 951) as a Food Additive by the EFSA Panel on Food Additives and Nutrient Sources Added to Food (ANS). ‘Max permitted level’ indicates the regulatory limit in mg/L (or mg/kg, where applicable), while ‘Reported use levels’ represent industry‐submitted usage ranges. ‘Quantum satis’ refers to use levels not specified by numerical limits but governed by good manufacturing practice.

As a dipeptide of L‐phenylalanine methyl ester and L‐aspartic acid, aspartame fully degrades to methanol, phenylalanine and aspartic acid in the human gastrointestinal tract [2]. Even in foods and beverages, it degrades spontaneously to these products [3]. Therefore, the potential contribution of aspartame degradation products to its observed biological effects has been a subject of scientific investigation [2, 3, 5]. On July 14th, 2023, the International Agency for Research on Cancer (IARC), the World Health Organisation (WHO) and the Food and Agriculture Organisation (FAO) Joint Expert Committee on Food Additives (JECFA) jointly reclassified aspartame as ‘possibly carcinogenic to humans’ (Group 2B) specifically in relation to hepatocellular carcinoma [2]. This marks the first time a semi‐independent committee for the WHO has declared an artificial sweetener as ‘possibly carcinogenic to humans’.

Beyond research on the possible hepatocarcinogenic effect of aspartame, studies on its impact on intestinal epithelial cell activation, distress, release of inflammatory cytokines, cell death and changes in the barrier are still scarce [2, 3, 6, 7, 8, 9, 10, 11]. In the context of the epithelial barrier theory, any substance in doses used in human daily life should be devoid of cytotoxicity, noninflammatory, and should not impair the epithelial barrier, ensuring its safety for human consumption [12]. The aim of this study was to assess the effect of aspartame on the biology, inflammation and epithelial barrier of gut epithelial cells.

2. Material and Methods

2.1. Caco‐2 Cell Culture

Caco‐2 cells were purchased from Sigma Aldrich (#86010202) and were used between passages 20 and 45 in all experiments. Cells were cultured as a monolayer in Eagle's minimal essential medium (ATCC) supplemented with 10% FBS (Sigma‐Aldrich), 1% penicillin/streptomycin (Sigma‐Aldrich), 1% nonessential amino acids (Sigma‐Aldrich) and 1% sodium pyruvate (Sigma‐Aldrich) in a 150‐cm2 T‐flask in a humidified incubator at 37°C in 5% CO₂.

2.2. MTT Assay

Caco‐2 cells were seeded in 96‐well plates at a 40,000 cells/well concentration and incubated in a 37°C humidified incubator with 5% CO₂ overnight. Cells were treated with aspartame and its degradation products, phenylalanine, aspartic acid and methanol, with varying doses for 24, 48 and 72 h. For mitigation experiments, cells were pretreated with 4 mM n‐acetylcysteine (NAC) (Sigma‐Aldrich), 200 mM IKK‐16 (Sigma‐Aldrich) and 5 μg/ml anti‐TNF‐α antibody (Sigma‐Aldrich) for 1 h and treated with aspartame and indicated agents for 24 or 48 h. After incubation, cell culture media were removed, and cells were washed with PBS. MTT solution was added with new media and incubated at 37°C for 4 h in a CO₂ incubator. After incubation, DMSO was added to each well and incubated for another 15 min. After incubation, absorbance was measured at 520 nm with a microplate reader (Berthold).

2.3. Thymidine Incorporation Assay

The effect of aspartame on Caco‐2 cell proliferation was monitored during 24 h of exposure. Briefly, cells were seeded in 96‐well plates (20 × 104 cells/well). After 24 h of incubation, the cells were exposed to aspartame at the range of concentrations described above or to the culture medium, followed by incubation with methyl‐H3‐thymidine (Hartmann Analytic GmbH) at 0.1 uCi/μL for 8 h at 37°C and 5% CO₂. After the incubation period, the plates were frozen at −20°C prior to harvesting (TOMTEC Harvester 96 Mach III). The samples were then transferred onto filters for scintillation counting on the MicroBeta TriLux instrument (PerkinElmer) following the manufacturer's instructions.

2.4. Reactive Oxygen Species (ROS) Determination

CM‐H2DCFDA (Thermo Fisher Scientific) was used to detect ROS generation within cells with aspartame treatment. Caco‐2 cells were seeded in 96‐well plates at a 10,000 cells/well concentration and incubated in a 37°C humidified incubator with 5% CO2 overnight. The cells were treated with 20 μM CM‐H2DCFDA for 45 min. After CM‐H2DCFDA treatment, cells were washed and treated with different doses of aspartame and its degradation products, phenylalanine, aspartic acid and methanol for 6 h in PBS supplemented with 10% FCS. At the beginning of the assay and after 6 h of incubation, absorbance was measured using a microplate reader (Berthold) with excitation at 485 nm and emission at 535 nm. As a positive control, 10 mM H₂O₂ was used.

2.5. Detection of Endotoxin Levels in Aspartame

Endotoxin levels in aspartame solutions were assessed using the PyroGene Recombinant Factor C Endpoint Fluorescent Assay (Lonza Bioscience), following the manufacturer's protocol. Several 10‐fold serial dilutions of aspartame were prepared. To validate the assay's performance, each dilution was spiked with a known concentration of endotoxin standard, and the recovery of the spike was calculated. Recovery values between 50% and 200% were considered acceptable, indicating the assay's functionality in detecting endotoxin in aspartame. For dilutions meeting this acceptance criteria, the measured endotoxin levels were reported. The fluorescence was measured at the beginning and the end of the 1‐h incubation in a fluorescent microplate reader (Berthold) using excitation/emission wavelengths of 380/440 nm.

2.6. Exposure to Aspartame and Its Degradation Products, Phenylalanine, Aspartic Acid and Methanol in Gut‐On‐a‐Chip Cultures and Transepithelial Electrical Resistance Measurement

A gut‐on‐a‐chip system was used to assess the effect of aspartame on colon epithelial cells [13, 14]. Gut‐on‐a‐chip cultures were established with OrganoPlate 3‐lane 64 plates (Mimetas BV) following the manufacturer's protocols. The extracellular matrix gel is prepared by combining 1 M HEPES, 37 g/L NaHCO₃ and 5 mg/mL collagen‐I in a ratio of 1:1:8. The gel mixture is added to the gel channel as 2 μL and incubated in a 37°C humidified incubator with 5% CO₂ for 15 min. After incubation, 30 μL PBS was added to the gel channel. After 24 h, Caco‐2 cells were seeded in the right channel as 20,000 cells/channel from the right perfusion inlet in Eagle's minimal essential medium (EMEM, ATCC) supplemented with 10% FBS (Sigma‐Aldrich), 1% penicillin/streptomycin (Sigma‐Aldrich), 1% nonessential amino acids (Sigma‐Aldrich) and 1% sodium pyruvate (Sigma‐Aldrich). To allow the cells to settle on the extracellular matrix gel, the OrganoPlate was incubated on its side for 4 h in a 37°C humidified incubator with 5% CO₂. After incubation, all perfusion inlets were filled with 50 μL of previously described cell culture media. On day 8 of gut‐on‐a‐chip culture, the chips were exposed to 10, 2.5 and 0.63 mg/mL aspartame and equivalent doses of its degradation products dissolved in culture media for 4 days. A transepithelial electrical resistance (TER) measurement was done with the OrganoTEER device (Mimetas) every day. For each condition, at least three chips were used. For mitigation experiments, cells were pretreated with 4 mM n‐acetylcysteine (NAC) (Sigma‐Aldrich) and 5 ng/mL anti‐TNF‐α antibody (Sigma‐Aldrich) for 1 h and treated with aspartame for 4 days together with aspartame and indicated agents.

2.7. RNA Isolation and RNA Sequencing

RNA from the gut‐on‐a‐chip culture was harvested at 24 h of the aspartame exposure. Total RNAs were extracted and purified with the RNeasy Plus Micro Kit (Qiagen) according to the manufacturer's protocol. NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific) and 2200 Tape Station Automated Electrophoresis System (Agilent Technologies) were used to determine the quantity and quality of the isolated RNA.

The samples with RNA integrity numbers 9.4 or more were chosen for sequencing. RNA sequencing (RNAseq) was performed on Illumina NovaSeq 6000 with the TruSeq RNA Stranded Sample Prep Kit (Illumina). Raw sequencing reads were first subjected to quality control checks and trimming using BBDuk, which removes adapter sequences and low‐quality bases. The high‐quality, trimmed reads were then aligned to the human reference genome (GRCh38. p14) using the Subread aligner, ensuring accurate mapping of reads. Gene‐level expression was quantified using featureCounts, part of the Rsubread package. This method counts reads mapping to known gene features annotated in Gencode release 46, providing robust expression metrics.

2.8. Transcription Factor Enrichment Analysis

Transcription factor enrichment analysis was used to predict the activated transcription factors upon aspartame exposure. Transcriptional Regulatory Relationships Unravelled by Sentence Based Text Mining (TRRUST) and all RNAseq and ChIP‐seq sample and signature search (ARCHS4) were used to predict the activated transcription factors.

2.9. Human Intestinal Organoid Culture

Human induced pluripotent stem cell (iPSC)‐derived intestinal organoids were obtained from DefiniGEN (Cambridge, UK) and were used between passages 4 and 6. For organoid culture, advanced DMEM/F12 (Thermo Fisher Scientific) supplemented with HEPES, GlutaMax 100×, B27 supplement, N2 Supplement (Thermo Fisher Scientific, Waltham), 100 U/mL penicillin, 100 μg/mL streptomycin (Sigma‐Aldrich), R‐Spondin1, Noggin (Peprotech), Prostaglandin E2, Laduviglusib, A83‐1 (Selleckhem), EGF (BioTechne) and 10 μM Y27632 (Selleckhem, only on the first 3 days) was used. Organoids were cultured in the above‐mentioned medium and Culturex UltiMatrix (R&D Systems) on 24‐well plates (Thermo Fisher Scientific). Organoid medium was changed every 2–3 days and passaged on day 7. On day 5, organoids were treated with 2.5 and 10 mg/mL aspartame for 2 days.

2.10. Targeted Proteomic Analysis

The gut‐on‐a‐chip, human iPSC‐derived intestinal organoid and THP1‐XBlue (InvivoGen) were used for targeted proteomics analysis. Culture media from gut‐on‐a‐chip and intestinal organoid cultures were used. For aspartame treatment of THP1‐XBlue cells, 24‐h cell culture media was used. The samples were analysed using the proximity extension assay (Olink)–targeted proteomics technology (Uppsala, Sweden). The inflammation panel, which measures 92 immune response‐related proteins, and the immune response panel, which measures 92 inflammation‐related proteins, were used according to the manufacturer's protocol.

2.11. NF‐kB Reporter Assays

For the NF‐κB activation assay, NF‐κB/AP‐1 reporter THP1‐XBlue (InvivoGen) monocytic cells and NF‐κB reporter (Luc)‐HCT‐116 (BPS Bioscience) colon cells were used. THP1‐XBlue cells were routinely cultured in RPMI 1640 medium (Thermo Fisher Scientific) supplemented with fetal calf serum, nonessential amino acids, sodium pyruvate, MEM vitamins, penicillin–streptomycin, normoxin and 4.5 g/L glucose. For selection, 0.2 mg/mL zeocin was used. For the NF‐κB/AP‐1 stimulation assay, 100,000 cells were seeded in the 96‐well plate and treated with aspartame with or without 4 mM N‐acetylcysteine (NAC) (Sigma‐Aldrich) and 5 μg/mL anti‐TNF‐α (R&D Systems) antibody in a 37°C humidified incubator with 5% CO₂ for 24 h. QUANTI‐Blue solution is used according to the manufacturer's recommendations for detecting NF‐κB/AP‐1 stimulation. One hundred and eighty microliters of QUANTI‐Blue solution was added to a 96‐well plate, the 20 μL cell supernatant was added to the plate and incubated for 1–2 h, and the optical density was measured at 655 nm with a microplate reader (Berthold). 50 ng/mL TNF was used as the positive control.

NF‐κB reporter (Luc)‐HCT‐116 cells were routinely cultured in McCoy's medium (Thermo Fisher Scientific) supplemented with foetal calf serum and penicillin–streptomycin. For the NF‐κB stimulation assay, 5000 cells were seeded in the 96‐well plate and treated with aspartame and its degradation products in a 37°C humidified incubator with 5% CO₂ for 24 h. ONE‐Step Luciferase Assay System (BPS Bioscience) is used according to the manufacturer's recommendations for detecting NF‐κB stimulation. An equal amount of luciferase assay working solution as the volume of the cell culture media was added to the 96‐well plates at the end of the treatment. The 96‐well plate was gently rocked for 20 min, and the firefly luminescence was measured with a luminometer (Berthold).

2.12. Immunofluorescence Staining of HIOs for ZO‐1 and Occludin 1 Expression After Aspartame Treatment

HIOs were treated with 10 mg/mL aspartame for 48 h. HIOs were harvested with ice‐cold PBS and fixed with 4% paraformaldehyde (Fluka/Sigma Aldrich, Buchs, Switzerland) for 45 min. Fixed organoids were washed with ice‐cold PBS, then perm‐blocked with 1% DMSO, 1% goat serum, 1% BSA and 0.5% Triton X‐100 for 60 min. After the perm‐block, organoids were washed with 0.1% BSA/PBS and incubated with primary antibodies zonula occludens‐1 (ZO‐1) and occludin (Thermo Fisher Scientific) overnight at 4°C. After incubation, organoids were washed with wash buffer (0.1% BSA and 0.1% Triton X‐100 in PBS) and then incubated with the secondary antibodies (goat anti‐rabbit IgG‐AF488, Life Technologies and goat anti‐mouse IgG‐AF546, Life Technologies) overnight at 4°C. As the final step, the organoids were washed with wash buffer and stored until confocal microscopy assessment in 0.1% BSA/PBS. Samples were analysed with confocal microscopy (Zeiss LSM780, Oberkochen, Germany). At least three organoids were analysed for each condition. From each organoid, 12 images were taken at 1 μM intervals, then combined with maximum intensity projection and processed in ImageJ/Fiji software.

2.13. Data and Statistical Analyses

Differential expression analysis of RNAseq data was performed using DESeq2, which employs statistical techniques to determine significant changes in gene expression across different conditions. Genes with the adjusted p value (BH) less than 0.05 and an absolute log₂ fold change greater than 1 were considered differentially expressed. Over‐representation analysis (ORA) was performed using the goseq package, which accounts for gene length bias. ORA was conducted on significantly differentially expressed genes to identify enriched pathways and biological processes from the Gene Ontology (GO) and KEGG databases. Significantly enriched categories were identified using a hypergeometric test with a Benjamini–Hochberg adjusted p value threshold of 0.05. Gene set enrichment analysis (GSEA) was performed using the clusterProfiler package in R. GSEA identifies whether predefined sets of genes (e.g., pathways or gene signatures) show statistically significant differences in expression between conditions. ClusterProfiler was used for its efficient handling of large datasets and its ability to provide insightful visualisations of enriched pathways, complementing the GO analysis by highlighting broader biological themes. All sequencing datasets are publicly available at the NCBI Gene Expression Omnibus under accession number GSE285905.

The targeted proteomics data were pre‐processed and quality‐controlled using the platform‐specific ‘Olink NPX manager’ software, which background corrects, log2 transforms and normalises all samples to an arbitrary normalised protein expression (NPX) scale. Data are presented as normalised protein expression NPX values.

All experiments were performed at least three times with three to six technical replicates. For transcriptomic and proteomic analysis, an adjusted p value < 0.05 was considered statistically significant. For other assays, statistical analyses were performed with one‐way ANOVA followed by Dunnett correction, and a p value < 0.05 was considered statistically significant.

3. Results

3.1. Aspartame Shows Cytotoxicity to Gut Epithelial Cells and Organoids

The cytotoxicity of aspartame on gastrointestinal epithelial cells was assessed at levels consistent with reported usage (Table 1), which ranges between 1 and 5 mg/mL in confectionary, processed foods and vegetables, food supplements and chewing gums. For the daily usage of tabletop sweeteners, it is quantum satis (as much as), which means there is no maximum permitted level [3]. Accordingly, we first analysed the viability of Caco‐2 cells and human intestinal organoids. Aspartame shows cytotoxicity at 0.625–10 mg/mL (Figure 1A). It also disrupts the 3D structure of human intestinal organoids, causing shrinkage and loss of lumen (Figure 1B,C).

FIGURE 1.

Aspartame induces cytotoxicity in colon and intestinal epithelial cells. (A) Viability of 24, 48 and 72 h of treatment with different concentrations of aspartame was assessed. Viability was measured using the MTT assay on monolayer cultures of Caco‐2 cells. Data are presented as means ± SDs. For statistical analysis, two‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001. (B) Change in human intestinal organoid size after 2 days of exposure to 2.5 and 10 mg/mL aspartame. Data are presented as the relative change in the organoid area after 2 days of culture compared to day 0. (C) Representative images of intestinal organoids exposed to 2.5 and 10 mg/mL aspartame for 2 days. Data are presented as means ± SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001. Scale bar 50 μm.

3.2. Daily Used Doses of Aspartame Induce Cellular Distress, Cell Death and Inflammation Signature in the Transcriptome

Dose‐dependent transcriptomic analyses were performed to assess the effects of aspartame on 3D epithelial cells on gut‐on‐a‐chips after 24‐h exposure to aspartame. Exposure to 2.5 and 10 mg/mL of aspartame resulted in the differential expression (p < 0.05) of 8272 genes. In 10 mg/mL, 4004 genes were upregulated, and 4229 genes were downregulated (Figure 2A). In 2.5 mg/mL, 399 genes were affected, with 262 upregulated and 137 downregulated. Interestingly, 360 of these genes (more than 90%) were overlapping. The direction of regulation of the overlapping affected genes was the same, and a dose–response was observed (Figure 2B).

FIGURE 2.

Aspartame alters the transcriptome of gut epithelial cells. (A) Venn diagram of significantly expressed genes (adjusted p < 0.05) upon 2.5 and 10 mg/mL aspartame exposure on the gut‐on‐a‐chip system for 24 h. (B) Heatmap of significantly expressed genes with overlap between 2.5 and 10 mg/mL aspartame treatments. (C) GSEA analysis of 10 mg/mL aspartame exposure. Gene Ontology Biological Process (GOBP) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases were used. Cytoplasmic translation (GO: 0002181), response to unfolded protein (GO: 0006986), chemical carcinogenesis–reactive oxygen species (hsa05208), positive regulation of apoptotic signalling pathway (GO: 2001235), TNF signalling pathway (hsa04668), response to oxidative stress (GO: 0006979), leucocyte migration (GO: 0050900), DNA‐templated DNA replication (GO: 0006261), double‐strand break repair (GO: 0006302). (D) ORA analysis of significantly expressed genes with overlap between 2.5 and 10 mg/mL aspartame treatments. GOBP database was used. DNA strand elongation is involved in DNA replication (GO: 0006271), cellular response to stress (GO: 0033554), programmed cell death (GO: 0012501) and cellular response to xenobiotic stimulus (GO: 0071466). (E) Scatter plot of normalised counts from RNAseq for key genes involved in the unfolded protein response. Data are presented as means ± SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001.

Gene set enrichment analysis (GSEA) of 10 mg/mL aspartame exposure revealed significant upregulation of cytoplasmic translation, response to unfolded protein, chemical carcinogenesis, oxidative stress, positive regulation of apoptotic signalling, TNF signalling and leucocyte migration pathways. Interestingly, DNA repair mechanisms were downregulated. We observed the downregulation in DNA‐templated DNA replication gene sets and double‐strand break repair (Figure 2C). Four of these nine pathways overlapped and appeared in GSEA analyses of 2.5 mg/mL exposure. They were upregulated in cytoplasmic translation and response to unfolded protein and downregulated in DNA‐templated DNA replication gene sets and double‐strand break repair.

We chose to focus on overlapping genes here because 2.5 mg/mL is a very commonly exposed dose, and the appearance of the alteration of the same genes in 10 mg/mL provides additional evidence for their effects on epithelial cells. Analysis of overlapping genes between 2.5 and 10 mg/mL shows alteration of genes related to response to DNA replication mechanisms, cellular response to stress, programmed cell death and cellular response to xenobiotic stimulus. Within the genes related to cellular response to stress, the upregulated genes include endoplasmic reticulum stress and TNF signalling pathways. However, the downregulated genes mostly relate to DNA repair mechanisms (Figure 2D). Shows overlapping genes in both doses, and Figures S2–S4 depict additional changes in 10 mg/mL doses.

In addition, response to oxidative stress appeared in both the 2.5 and 10 mg/mL doses (Figure S3A,B). Aspartame exposure in gut‐on‐a‐chip models increased PRDX1 and PRDX5 protein values (Figure S6B), which showed a response to oxidative stress consistent with transcriptomic data. Moreover, signature genes for NRF2, an oxidative stress‐induced transcription factor crucial for regulating the antioxidant response [15], are upregulated (Figure S4A). Programmed cell death‐related gene activation was visible in the overlapping group of two doses and became more prominent at 10 mg/mL doses, further underlining its role in cellular distress (Figures 2D and S4B).

Further analysis of transcriptome data showed that aspartame causes cellular distress, as shown by the upregulation of genes related to the response to unfolded protein and cytoplasmic translation, which can also be referred to as ribosomal stress (Figure S2A,B). Aspartame increased endoplasmic reticulum stress‐related genes. In 2.5 mg/mL aspartame, genes related to the response to the unfolded protein pathway, such as ATF3, HSPA8, CHAC1, FBXO6 and CREBRF, showed significant upregulation (Figure 2E).

Aligning with the GSEA, analysis transcriptomic data with two different datasets (TRRUST transcription factors 2019 and ARCHS4 TFs co‐expression) showed increased activation of transcription factors associated with unfolded protein response (ATF3, ATF4, DDIT3 and XBP1), xenobiotic response (ARNT), inflammation (NFKB1, RELA, SMAD3, FOS, FOSB and JUN), antigen presentation (RFXANK), epigenetic regulation (HDAC1) and protein translation (EIF3K, POLR2L) (Figure S1) upon 10 mg/mL aspartame exposure on gut‐on‐a‐chip culture.

3.3. Aspartame Disrupts DNA Repair and Replication Mechanisms

RNAseq analysis demonstrated that the genes involved in the DNA repair mechanisms were notably repressed in both 2.5 and 10 mg/mL doses (Figures 2C and 3A,B). Aspartame significantly suppressed double‐strand break repair and decreased the expression of essential genes such as MCM, CDC, BRIP1, RAD51 and RIF1, which work together in a complex network to maintain genomic stability and contribute to cell cycle regulation. In addition, RNAseq analyses revealed that aspartame significantly downregulated BRCA genes. Notably, BRCA1 and BRCA2 are essential genes involved in double‐strand break repair and are necessary for the homologous recombination pathway, along with many others that contribute to genomic stability through repair mechanisms (Figures 3A and S5A).

FIGURE 3.

Aspartame disrupts the DNA repair and replication mechanisms. (A) Scatter plot of normalized counts from RNAseq for key genes related to double‐strand break repair. (B) Scatter plot of normalized counts from RNAseq for mismatch repair‐related genes. (C) Scatter plot of normalized counts from RNAseq for DNA‐templated DNA replication‐related genes. (D) ³H‐Thymidine incorporation assay was used to assess the proliferation of Caco‐2 cells treated with aspartame. Data are shown as counts per minute. Data are presented as means +/‐ SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001.

Aspartame suppressed the expression of mismatch repair genes, such as XPC and RPA1 (Figure 3B). These genes impact micronuclei formation and chromosome aberrations, indicating their crucial roles in double‐strand break repair and related cell cycle arrest mechanisms.

Aspartame altered the expressions in the DNA replication‐related genes, such as POLA1, PCNA, RFC1, RPA4, TOP1, PRIMPOL and FEN1, which may cause significant reflections in response to cell cycle regulation (Figures 3C and S5B). These genes are crucial for maintaining DNA replication when replication forks encounter DNA damage, as they are involved in repriming DNA synthesis past lesions that stall the forks. Moreover, aspartame exposure downregulated the WNT signalling pathway. The WNT pathway is a crucial signalling cascade for stem cell maintenance, cell proliferation, differentiation and tissue regeneration [16, 17] (Figure S5B). Therefore, we assessed the effect of aspartame on gut epithelial cell proliferation with a [³H]‐thymidine incorporation assay. Aspartame dose‐dependently decreases the proliferative capacity of the gut epithelial cells (Figure 3D). RNAseq results revealed that aspartame significantly downregulated pathways critical to DNA replication mechanisms, potentially causing damage to the organism at the cellular level.

3.4. Aspartame Induces NF‐κB Activation and Proinflammatory Cytokine Release

Our next focus was the in‐depth analyses of the proinflammatory response induced by aspartame. A complex inflammatory response is observed by aspartame with 2.5 and 10 mg/mL. Genes related to the TNF signalling pathway increased with aspartame exposure, including the signalling molecules, transcription factors, chemokines, cytokines, enzymes and inhibitory molecules of TNF signalling (Figure 4A). The increased expression of CXCL1, CXCL2, CXCL3, CXCL10, CXCL16 and CCL20 implies that aspartame exposure caused a strong chemokine response, mainly targeting neutrophils (Figure 4B). Moreover, an increase in the expression of the PTGS2 gene, which encodes cyclooxygenase‐2, a key enzyme in prostaglandin biosynthesis, was observed. Increased IL15 gene expression suggests possible activation of innate lymphoid cells, T cells and NK cells upon aspartame exposure. In addition, increased IL18 mRNA level is observed as a part of the complex innate immune response. Aspartame differentially regulated the gene expression associated with proinflammatory signalling pathways, as indicated by the upregulation of FOS, JUN and JUNB, major AP‐1 transcription factors, and the signalling molecules in the MAPK pathway. Aspartame exposure caused an upregulation of TNF signalling pathway inhibitor genes, such as TNFAIP3 and NFKBIA, suggesting the activation of the NF‐κB pathway.

FIGURE 4.

Aspartame induces NF‐κB activation and inflammation through oxidative stress. (A) Heatmap of the TNF signalling pathway (hsa04668) related genes in RNAseq data from 24‐h‐treated aspartame gut‐on‐a‐chip. (B) Scatter plot of TNF pathway‐related key genes in RNAseq data. Data shown as normalised counts. (C) Scatter plot of TNF pathway‐related proteins in gut‐on‐a‐chip cell culture media treated with aspartame. (D) Heatmap of targeted proteomic analysis of human iPSC‐derived intestinal organoids treated with aspartame. Data are shown as NPX values. (E, F) I NF‐κB activation assay was performed with reporter cell lines THP1‐XBlue and HCT‐116. Cells were treated with the indicated compounds for 24 h. Data are shown as relative NF‐κB stimulation. (H) ROS generation experiment performed with CM‐H2DCFDA after 6 h incubation with aspartame on Caco‐2 cell culture. As a positive control, 1 mM H₂O₂ was used. (G, J) Scatter plot of targeted proteomic analysis of THP1‐XBlue cells treated with indicated compounds. Data are presented as means +/−SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001.

Consistent with the transcriptomic results, targeted proteomic analysis from gut‐on‐chip showed increased secretion of TNF‐α signalling‐related cytokines and chemokines (Figure 4C). Increased CXCL1 has a role in neutrophil chemotaxis. Increased CCL20 suggests the stimulation of lymphocyte chemotaxis. Further, we assessed the proinflammatory mediator increase upon aspartame exposure to intestinal cells with human iPSC‐derived human intestinal organoids. Proteomic analysis of cell culture media of human intestinal organoids showed increased expression of CSF‐1, CXCL1, CXCL11 and CCL4 with 10 mg/mL aspartame. In addition, 2.5 mg/mL of aspartame increased proinflammatory mediators IL8 and IL‐17C. Increased detection of CASP‐8 and ADA in cell culture media shows elevated cell death with aspartame (Figure 4D).

We performed NF‐κB reporter assays with colon (NF‐κB reporter Luc‐HCT‐116 cells) and monocyte (THP1‐XBlue) transfected cell lines to support the findings on TNF‐α secretion and proinflammatory cytokine and chemokine signalling. Aspartame induced a dose‐dependent NF‐κB activation as low as 2.5 mg/mL in the colon and monocyte cell lines (Figure 4E). We investigated the role of TNF‐α in NF‐κB activation. Anti‐TNF‐α treatment abolished the aspartame‐induced NF‐κB activation (Figure 4F). The blocking of TNF‐α decreased the amounts of secreted CXCL9, CXCL10 and CCL19 (Figure 4G). Furthermore, anti‐TNF‐α antibody treatment and NF‐κB pathway antagonism with a synthetic inhibitor reduced the aspartame‐induced cytotoxicity (Figure S6A). In addition, we ruled out possible endotoxin presence in aspartame with the PyroGene Recombinant Factor C assay. Endotoxins were not found at detectable levels (< 0.01 EU/mL) in aspartame (data not shown). Overall, aspartame stimulates an innate immune response, a proinflammatory cytokine and chemokine release in gut epithelial cells.

Next, we hypothesised that aspartame‐induced oxidative stress could play a role in NF‐κB activation. Aspartame caused oxidative stress as low as 2.5 mg/mL (Figure 4H). Accordingly, we used NAC, an antioxidant agent, to suppress the aspartame‐induced oxidative stress. NAC treatment with aspartame decreased the PRDX1 and PRDX5 protein concentrations compared to aspartame only, which shows the clear inhibition of oxidative stress (Figure S6B). Inhibition of oxidative stress via NAC treatment decreased the NF‐κB activation due to aspartame exposure (Figure 4I). NAC treatment reduced the aspartame‐induced TNF‐α, IL12B, IL18, CXCL1, CXCL9, CXCL10 and CSF1 (Figure 4J). However, it did not affect the cytotoxicity of aspartame (Figure S6C). All targeted proteomic results can be found in Figure S7.

The data show the critical role of aspartame‐induced NF‐κB activation due to oxidative stress and TNF release, which leads to proinflammatory cytokine and chemokine production.

3.5. Aspartame Causes Gut Epithelial Barrier Damage

Aspartame influenced one of the key functions of the gut epithelium, namely the barrier. Aspartame significantly affected the genes of tight and adherence junction structures (Figure 5A) and showed a complex impact on the expression of claudins, affecting both the barrier‐forming and pore‐forming types (Figure 5B). Specifically, the expression of barrier‐forming claudin, CLDN4, is increased, while another barrier‐forming claudin, CLDN23, is decreased. In contrast, among the pore‐forming claudins, CLDN2 showed a decrease, while CLDN15 was upregulated. It has been demonstrated that CLDN23 is important in the intestinal epithelial barrier permeability by regulating the distribution of CLDN3 and CLDN4 [18]. In addition, aspartame increases the gene expression of occludin, an important transmembrane protein within tight junctions. Furthermore, myosin heavy chains (MYH) gene expression, MYH9 and MYH10, important molecules in epithelial barrier development, cell migration, cytoplasmic division, intracellular transport and signal transduction, are significantly decreased by aspartame.

FIGURE 5.

Aspartame damages the gut epithelial barrier. (A) Tight junction and adherence junction genes in RNAseq data from 24‐h‐treated aspartame gut‐on‐a‐chip. (B) Scatter plot of tight junction and adherence junction‐related genes in RNAseq data. Data are shown as normalized counts. (C) Assessment of gut epithelial barrier integrity on gut‐on‐a‐chip treated with aspartame. Gut‐on‐a‐chips were treated with three different doses of aspartame for 4 days. Data are shown as a relative change of transepithelial electrical resistance (TEER). Data are presented as means ± SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001.

To investigate the functional properties of these complex changes in barrier gene expressions, we assessed the impact of aspartame on gut epithelial barrier integrity using a gut‐on‐a‐chip model. The results demonstrated that aspartame caused significant epithelial barrier damage in a dose‐dependent manner compared to the control. The highest dose of 10 mg/mL caused damage to the epithelial barrier on day 2, while the lower dose of 2.5 mg/mL resulted in damage on day 4 (Figure 5C). The immunofluorescence staining for ZO‐1 and occludin of human intestinal organoids treated with 10 mg/mL aspartame showed that aspartame disturbs the 3D structure of the tight junction assembly with apparent morphological changes in HIOs (Figures 1 and S8A).

We further investigated the specific roles of aspartame‐induced TNF release and ROS formation on epithelial barrier damage. We treated the gut‐on‐a‐chips with the anti‐TNF‐α antibody and NAC treatment 1 h before the aspartame treatment. Our data show that neither NAC nor anti‐TNF‐α antibody could rescue the gut epithelial barrier damage induced by aspartame (Figure S8A).

3.6. Degradation Products of Aspartame Cause Cytotoxicity, NF‐κB Activation, ROS Generation and Epithelial Barrier Damage

Each mole of aspartame breaks down into 1 mole each of phenylalanine, aspartic acid and methanol [2]. At the highest concentration used in our study (10 mg/mL, equivalent to 34 mM aspartame), the corresponding end concentrations of the degradation products are 5.6 mg/mL phenylalanine, 4.52 mg/mL aspartic acid and 1.09 mg/mL methanol (34 mM each). We tested each degradation product individually as well as in combination to evaluate both their independent and potential synergistic effects. The doses used for each were equivalent to those of aspartame applied in earlier experiments. Due to their differing physical properties, we used molarity (mM) for methanol (a liquid) and mass concentration (mg/mL) for phenylalanine and aspartic acid. We expressed the combined degradation product mixture in terms of aspartame‐equivalent concentrations (mg/mL). Table S1 shows the used doses alone and combined with their equivalent doses for aspartame. The degradation products alone induced cytotoxicity at relatively high concentrations: 1.4 mg/mL for phenylalanine, 4.52 mg/mL for aspartic acid and 4.5 mM for methanol (Figure 6A). However, the combined degradation product mixture caused a significant decrease in cell viability at much lower concentrations, starting from 0.62 mg/mL aspartame equivalent dose. At this dose, the mixture contains approximately fourfold less phenylalanine, 16‐fold less aspartic acid and twofold less methanol than their individual cytotoxic thresholds. This indicates a potential synergistic effect among the degradation products. Notably, the cytotoxicity threshold of the degradation product mixture closely mirrors that of intact aspartame.

FIGURE 6.

Effect of aspartame degradation products on cell viability, NF‐κB activation, ROS generation and gut epithelial barrier integrity. (A) Cell viability after 48 h of treatment with different concentrations of aspartame degradation products and their mixture was assessed by using the MTT assay on monolayer cultures of Caco‐2 cells. Data are presented as mean ± SD. (B) NF‐κB activation assay was performed with reporter cell line HCT‐116. Cells were treated with the indicated compounds for 24 h. Data are shown as relative NF‐κB stimulation. (C) ROS generation experiment performed with CM‐H2DCFDA after 6 h of incubation with the indicated compounds on Caco‐2 cell culture. (D) Assessment of gut epithelial barrier integrity on gut‐on‐a‐chip treated with aspartame degradation products and their mixture. Gut‐on‐a‐chips were treated with three different doses of the indicated compound for 3 days. Data are shown as a relative change of transepithelial electrical resistance (TEER). Data are presented as means ± SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001.

Methanol induced strong NF‐κB activation starting from noncytotoxic concentrations (Figure 6B). Additionally, the degradation product mixture triggered significant NF‐κB activation at concentrations as low as 0.04 mg/mL aspartame‐equivalent, which is markedly lower than the activation thresholds observed for either methanol or intact aspartame alone. Due to severe cytotoxicity in HCT‐116 reporter cells, concentrations above 0.62 mg/mL of the degradation product mixture could not be assessed.

In terms of oxidative stress, aspartic acid and methanol individually did not induce ROS production in gut epithelial cells. In contrast, phenylalanine and the degradation product mixture significantly elevated ROS levels at their highest tested concentrations (Figure 6C).

Regarding barrier integrity, both phenylalanine and the degradation product mixture led to a rapid decline in TEER by day 1, with significant barrier disruption observed at 5.6 mg/mL for phenylalanine and 10 mg/mL aspartame equivalent doses for the mixture. This early‐onset damage contrasts with the gradual TEER reduction seen with intact aspartame.

4. Discussion

Aspartame is widely used in table‐top sweeteners, beverages, packaged foods, cocoa and chocolate products, as well as dietary supplements. Since aspartame was classified as ‘possibly carcinogenic to humans’ (Group 2B), particularly causing hepatocellular carcinoma by the Joint Expert Committee on Food Additives of the IARC, WHO and the FAO) [2], we mainly focused on its cellular toxicity in doses that are similar to daily exposures to human cells and iPSC‐derived human intestinal organoids. Our study revealed that aspartame disrupts DNA replication and repair mechanisms and induces oxidative stress and endoplasmic reticulum stress, leading to inflammation and eventual cell death in gut epithelial cells at daily consumption levels as low as 2.5 mg/mL.

Our data demonstrated that aspartame is cytotoxic at as low as 0.625 mg/mL. It is noteworthy that aspartame has been observed to elicit a complex cellular injury mechanism via inducing DNA damage, oxidative stress, endoplasmic reticulum stress and the production of proinflammatory cytokines and chemokines. Collectively, these effects may contribute to the observed cytotoxicity. The mismatch and double‐strand break repair system is integral to the DNA damage response pathway, which plays a crucial role in cellular defence mechanisms by identifying and eliminating cells with severe DNA damage. The mismatch repair system is essential for initiating cell cycle arrest and programmed cell death in response to specific types of DNA damage [19, 20]. Mismatch repair proteins, especially those involved in the nucleotide excision repair pathways, significantly influence double‐strand break repair [19, 21]. By correcting mismatches during DNA replication and maintaining genomic stability, mismatch repair prevents the accumulation of genetic errors in the short term and acts as a critical defence against cancer development in the long term [19]. In the present study, aspartame causes a defect in DNA repair mechanisms. Timely repair of damaged DNA has a major impact on preventing mutation and cancer development. Here, we demonstrated that aspartame significantly suppresses double‐strand break repair, the homologous recombination pathway and mismatch repair, which are essential to maintaining genomic stability.

Furthermore, our findings indicate that aspartame had an adverse effect on the DNA replication mechanisms. An impairment of the epithelial cell proliferation results in several defects in local defence, namely the decreased barrier function. Additionally, there was a remarkable increase in the transcriptional activity of genes associated with the positive regulation of apoptosis, suggesting an elevated cellular process for eliminating damaged cells. This suggests that aspartame exposure affects cellular integrity and triggers programmed cell death pathways, potentially as a protective mechanism against the propagation of cells with compromised DNA. Indeed, we detected a remarkable increase in gene expression associated with apoptosis. This enhanced pro‐apoptotic signalling could indicate that cells exposed to aspartame are experiencing stress or damage that triggers programmed cell death mechanisms. The affected genes play a key role in removing RNA primers and completing the maturation of Okazaki fragments on the lagging strand [22, 23, 24]. More specifically, TOP1 relieves the torsional strain created ahead of the replication fork by transiently breaking and rejoining one strand of the DNA duplex [25, 26]. Hence, impairment of the molecular systems involving these genes can result in structural and/or functional inadequacies at the DNA level [24, 27, 28]. These findings highlight the dual impact (both DNA repair and replication systems) of aspartame on cellular processes, emphasizing the potential genomic and cytotoxic challenges posed by daily artificial sweetener exposure.

Reactive oxygen species can induce cellular toxicity by interacting with cysteine residues of cell survival proteins, triggering lipid peroxidation and causing DNA damage [29]. Aspartame induces oxidative stress, as shown in rat models and in cell lines, targeting its effects on the nervous system and liver [6, 7, 8, 9, 10]. Cytoplasmic translation, namely ribosomal stress and endoplasmic reticulum stress, represents two additional cellular distress mechanisms induced by aspartame. In the present study, transcriptomic analysis shows elevated key genes such as ATF3, ATF4, DDIT3 and DDIT4 as evidence for the unfolded protein response induced by aspartame.

Oxidative stress and unfolded protein response both activate the NF‐κB pathway [29, 30]. Accordingly, we demonstrated that aspartame induces NF‐κB activation in reporter cells, transcriptomic and proteomic studies. NF‐κB is known to play a crucial role in cytokine and chemokine release, cell survival, proliferation, morphogenesis, differentiation and apoptosis [31]. Our data show that aspartame‐induced NF‐κB activation results in the release of proinflammatory cytokines and chemokines, especially TNF‐α. In epithelial cells, NF‐κB can be activated through cytokine receptors such as the TNF‐α receptor, pattern‐recognition receptors and cellular stress such as unfolded protein response and oxidative stress [29]. Following these findings, we assessed the role of aspartame‐induced oxidative stress in NF‐κB activation. Inhibition of oxidative stress with NAC results in decreased NF‐κB activation, demonstrating that aspartame‐induced oxidative stress activates NF‐κB, releasing proinflammatory cytokines and chemokines, especially TNF‐α. The released chemokines, such as CXCL1, CXCL11 and CCL20, show that aspartame activates leucocyte chemotaxis. We found that TNF‐α contributes to further NF‐κB activation and cell death caused by aspartame, since inhibition of NF‐κB and neutralizing TNF‐α modestly decrease the cytotoxicity of aspartame. However, antioxidant treatment with NAC couldn't overcome the aspartame‐induced cell death. Our study shows that aspartame triggers a multifaceted cellular distress response that collectively culminates in epithelial cell death. These include oxidative stress, ER stress and DNA replication and repair failure, all of which were detectable prior to substantial cell death. Therefore, inhibiting just one pathway is not sufficient to overcome the aspartame‐induced decrease in cell viability. NF‐κB plays a pivotal role in the inflammatory response associated with inflammatory bowel disease (IBD). In patients with IBD, NF‐κB activation is significantly elevated in the inflamed bowel tissue but not in healthy individuals [32, 33]. Higher levels of NF‐κB activation are linked to the severity of intestinal inflammation [32]. These findings suggest that activation of NF‐κB via aspartame can exacerbate the ongoing inflammation in the gastrointestinal system.

We demonstrated that aspartame damages the gut epithelial barrier in gut‐on‐a‐chips. The transcriptomic changes occur at 24 h, before the apparent gut epithelial barrier impairment that takes place at 48 h, which shows these transcriptomic alterations are induced by aspartame‐related cell distress and are the reason for causing epithelial barrier damage. The increased expression of OCLN, CLDN4 and CLDN15 and decrease in CLDN2, CLDN23, MYH9 and MYH10 show a complex regulation of tight junction and adherence junction proteins. The observed increase in pore‐forming CLDN15 and the decrease in sealing CLDN23 indicated a reduction in the epithelial barrier. CLDN23 plays a pivotal role in regulating the permeability of the intestinal epithelial barrier by regulating the distribution of CLDN3 and CLDN4 [18]. This response may include both direct damage to the constituents of the barrier and a subsequent healing response following each other. In this context, the observed increase in OCLN, CLDN4 and the decrease in CLDN2 may be indicative of an epithelial barrier repair response. A reduction in MYH9 expression is of particular significance, as studies have demonstrated that the knockout of MYH9 in mice leads to heightened sensitivity to dextran sulphate sodium, the development of colitis‐like structural changes, and the promotion of colitis‐related adenomas in the colon [34]. Aspartame‐induced molecular toxicity and cellular distress may contribute to the disruption of epithelial barrier damage through multiple mechanisms. In addition to regulating tight junction and adherence junction proteins, aspartame‐induced inflammation, such as TNF release [35, 36] and oxidative stress, may also affect the epithelial barrier integrity. TNF is known to impair gut epithelial barrier integrity. However, the anti‐TNF‐α antibody and antioxidant treatment with NAC could not reverse the aspartame‐induced epithelial barrier damage. Therefore, the aspartame‐induced epithelial barrier damage may be attributed to a complex network of cell distress mechanisms, including cell death and inflammation. Future studies may dissect the contributions of individual pathways via specific inhibitors or gene knockdown models. Given its rapid degradation under in vivo conditions, the effects of aspartame degradation products on gut epithelial cells are highly relevant to physiological exposure. To address this, we investigated the impact of phenylalanine, aspartic acid, methanol and their combination on epithelial cell viability, NF‐κB activation, ROS production and barrier integrity. While all individual degradation products induced cytotoxicity at high concentrations, their combination led to a significant loss of viability at much lower doses, suggesting a potential synergistic effect. Similarly, methanol alone and the degradation product mixture both triggered substantial NF‐κB activation, with the mixture stimulating NF‐κB at even lower concentrations, which again indicates synergistic activity. Notably, the degradation product mixture activated NF‐κB at significantly lower concentrations compared to intact aspartame, pointing to a mechanistic contribution of these breakdown products to aspartame's inflammatory effects.

Further, phenylalanine and the degradation product mixture increased ROS production and disrupted epithelial barrier integrity. The overlapping profiles of cytotoxicity, oxidative stress, NF‐κB activation and barrier damage between aspartame and its degradation products strongly suggest that the biological activity of aspartame may be caused, at least in part, by its degradation products. It is also quite relevant that all the detrimental effects of aspartame are reproducible with its degradation product mixture. Among these molecules, methanol appears to play a particularly notable role, demonstrating both cytotoxic and pro‐inflammatory effects at relatively low concentrations.

An important consideration is that methanol is further metabolised in vivo to formaldehyde and subsequently to formic acid, both of which possess well‐documented cytotoxic properties [37]. Formaldehyde, in particular, is a highly reactive aldehyde capable of crosslinking DNA and proteins, leading to genotoxic stress, while formic acid contributes to metabolic acidosis and mitochondrial inhibition [37]. It should also be noted that in vivo, aspartame‐derived formaldehyde can bind to the protein and DNA of the tissues [38]. Therefore, low‐level chronic methanol exposure could have cumulative or downstream effects via these metabolites.

Our working model for aspartame‐induced cellular effects offers a complex integration of various pathways. The earliest responses include oxidative stress and endoplasmic reticulum (ER) stress, both of which were observed within 6–24 h and at relatively low concentrations. These stress responses are known to lead to the activation of NF‐κB signalling [29, 30], In our study, we showed that antioxidant treatment with NAC leads to diminished NF‐κB activation. In addition, reporter assays and the upregulation of proinflammatory cytokines and chemokines (e.g., TNF‐α, CXCL1, CCL20) demonstrate the NF‐κB activation. Concurrently, we observed downregulation of DNA replication and repair pathways, including double‐strand break repair and mismatch repair components, suggesting a disruption of genomic maintenance. These upstream events collectively contribute to cellular distress, apoptosis and loss of barrier integrity at later time points. Notably, the degradation products recapitulate much of this cascade, implying that these effects are not solely mediated by intact aspartame but also by its breakdown components.

It should be noted that the data presented herein are not without limitations. One of the main limitations of the study is the inability to determine the precise quantity and concentration of aspartame that reaches the gut. Consequently, the concentrations utilised correspond to those reported in foods and beverages [3]. Although it is challenging to estimate the concentration of aspartame in intestinal fluids due to the rapid degradation of aspartame in the upper gastrointestinal tract and the complex fluid dynamics of the gastrointestinal system, we would like to provide some details. The acceptable daily intake limit of aspartame (40 mg/kg/day) would result in 2800 mg of aspartame intake daily for a 70 kg adult. The intake of just 100 g of aspartame‐containing dry foods, such as cocoa and chocolate products, processed fruits and vegetables, in a single meal can result in 10–100 mg of aspartame ingestion. With an empty stomach, which contains approximately 35 ± 7 mL of gastric fluid [39], the aspartame concentration could reach 0.23–3.6 mg/mL. Although this is diluted by hepatic, pancreatic and intestinal secretions in the duodenum, the aspartame will become more concentrated again in the later parts of the intestine, particularly in the colon. In addition, the models studied are analogous to a human in vivo situation; however, they are short‐term and do not represent chronic exposure. It is expected that the multiple mechanisms of cellular distress caused by the daily doses in a relatively short time (in a few days) could reflect chronic exposure to even lower doses. Given the chronic low‐grade exposure to aspartame and the fact that both aspartame and its degradation products exert pathological effects even at lower concentrations in our models, these findings raise important concerns regarding its long‐term impact on gut epithelial health.

5. Conclusion

The data highlight that aspartame triggers cellular stress, inflammation and damage to the gut epithelial barrier, suggesting a potential link to the onset or exacerbation of intestinal diseases characterised by epithelial barrier impairment. The extensive use of aspartame in food products is a matter of concern [3]. Given the cytotoxic and inflammatory properties of aspartame, it is imperative that strict regulation be imposed on its use in foods and beverages, with clearly defined maximum permitted levels across all products.

Author Contributions

Y.P. had primary responsibility for the framework of the study, data analysis and manuscript preparation. Y.P., D.Y., C.Z., H.B., S.A., A.G.‐S., X.B., A.H., O.G.V., L.C., S.S., A.A., C.A., T.J., R.D., I.O. and C.A.A. contributed to the conception and design of the study and interpretation of results, and C.A.A. supervised the project. Y.P., D.Y., C.Z., S.A., A.G.‐S., X.B., A.H. and O.G.V. performed the experiments. H.B. and Y.P. conducted data analysis and interpretation of the results.

Conflicts of Interest

C.A.A. has received research grants from the Swiss National Science Foundation, European Union (EU CURE), Novartis Research Institutes (Basel, Switzerland), Stanford University (Redwood City, Calif) and SciBase (Stockholm, Sweden); he is the Co‐Chair for EAACI Guidelines on Environmental Science in Allergic diseases and Asthma and serves on the Advisory Boards of Sanofi/Regeneron, Novartis, GlaxoSmithKline and SciBase, and is the Editor‐in‐Chief of Allergy. S.S. and C.A. are former employees of Seed Health, T.J. is a current employee at Seed Health and A.A. is a consultant at Seed Health. All other authors have no conflicts to disclose.

Supporting information

Figure S1: Transcription factor enrichment analysis of 10 mg/mL of aspartame. TRRUST Transcription Factors 2019 and ARCHS4 TFs co‐expression databases were used.

Figure S2: Heatmap of significant genes (adjusted p value < 0.05) related to response to unfolded protein (GO: 0006986) and cytoplasmic translation (GO: 0002181) upon 10 mg/mL aspartame exposure.

Figure S3: Heatmap of significant genes (adjusted p value < 0.05) related to response to oxidative stress (GO: 0006979) and chemical carcinogenesis‐ reactive oxygen species (hsa05208) upon 10 mg/mL aspartame exposure.

Figure S4: Heatmap of significant genes (adjusted p value < 0.05) related to Nrf2 pathway and positive regulation of apoptotic signalling pathway (GO: 2001235) upon 10 mg/mL aspartame exposure.

Figure S5: Heatmap of significant genes (adjusted p value < 0.05) related to double‐strand break repair (GO: 0006302) and DNA − templated DNA replication (GO: 0006261) upon 10 mg/mL aspartame exposure.

Figure S6: A, Viability assessment with MTT of aspartame treatment with or without anti‐TNF‐α antibody or IKK‐16 for 48 h on Caco‐2 cell culture. B, Targeted proteomic analysis of THP1‐XBlue cells treated with aspartame and NAC. C, Viability assessment with MTT of aspartame treatment with or without NAC. Data are presented as means +/‐SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. **p < 0.01. ***p < 0.001.

Figure S7: Targeted proteomics analysis of aspartame treatments on different platforms.

Figure S8: A, Gut epithelial barrier mitigation experiment with anti‐TNF‐α antibody and NAC. For the experiment, 5 μg/mL anti‐TNF‐α antibody and 4 mM NAC were used. Data are presented as means +/‐SDs. For statistical analysis, one‐way ANOVA was used. *p < 0.05. ***p < 0.001. B, Immunofluorescence staining for ZO‐1 and occludin of human intestinal organoids treated with 10 mg/mL aspartame and control (n: 3). DAPI (blue), ZO‐1 (green) and occludin (red). Graphical representations of the relative intensities of ZO‐1 and occludin, normalised to DAPI, in human intestinal organoids treated with 10 mg/mL aspartame and control (n: 3) are shown on the right.

Table S1: Dose equivalency of aspartame and its degradation products. This table presents the calculated concentrations of phenylalanine, aspartic acid and methanol corresponding to various doses of aspartame, both as isolated compounds and as part of a simulated degradation mixture. Values are expressed in mg/ml or mM as appropriate.

Data Availability Statement

The data that support the findings of this study are openly available in Gene Expression Omnibus at https://www.ncbi.nlm.nih.gov/gds/, reference number GSE285905.