Eight-week-old male C57BL/6 mice were purchased from Doo Yeol Biotech, Inc. (Seoul, Korea) and maintained under specific pathogen-free conditions. After a 1-week acclimation period, the mice were randomly divided into non-irradiated (control, Con), fractionated radiation exposure (Fraction), and single radiation exposure (Single) groups (n = 12 per group). Single irradiation was performed using a gamma-ray machine (BIOBEAM8000, 137Cs; dose rate: 2.6 Gy/min), and fractionated irradiation was performed using a different gamma-ray machine (MDI-KIRAMS 137, 137Cs; dose rate: 6.7 mGy/min) to irradiate the entire body of the mice. The single exposure group received a single 1 Gy dose on the final day of fractionated radiation. The fractionated exposure group received radiation 5 days per week for 2 min per day for 15 weeks, accumulating a total dose of 1 Gy (75 fractions, 6.7 mGy/min, 2 min per session). A schematic of the experimental design is shown in Figure 1A. The total dose of 1 Gy was selected based on previous studies investigating low-dose radiation biology (Morgan and Bair, 2013; Azzam et al., 2012). Fractionated exposure over 15 weeks was designed to mimic environmental or occupational low-dose radiation scenarios (Brooks and Couch, 2006; Tanaka et al., 2003; Gapeyev et al., 2015). The single-dose group was included to compare acute versus chronic responses, since prior reports demonstrate distinct biological outcomes depending on radiation delivery mode (Mitchel, 2006). Blood samples were collected from each mouse by cardiac puncture one week after radiation exposure and were immediately transferred into EDTA-coated tubes. Fecal samples were collected on day 7 post-exposure by placing each mouse in a sterile cage. All intestinal tissues were harvested on the day of sacrifice; those for histological analysis were fixed in paraformaldehyde, while those for molecular analyses were snap-frozen and stored at –80°C until use.

All mice were placed under general anesthesia, and blood was collected via cardiac puncture into tubes prefilled with EDTA (K2EDTA, BD Microtainer #365974). A CBC was performed using a hematology analyzer (Abaxis/HM5), measuring white blood cells (WBC), red blood cells (RBC), lymphocytes (LYM), and platelets (PLT).

Radiation-induced DNA damage was assessed by quantifying γ-H2AX foci in mouse peripheral blood mononuclear cells. RBCs were lysed using eBioscience™ RBC lysis buffer (Invitrogen, Waltham, MA, USA) and fixed with 1× Lyse/fix solution (BD Phosflow™; BD Biosciences, San Jose, CA, USA). Fixed cells were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) for 10 min at room temperature (RT) and then stained with Alexa Fluor^® 488 Mouse anti-H2AX (pS139) antibody (clone N1–431, BD Pharmingen™; Becton Dickinson, Franklin Lakes, NJ, USA) for 1 h at RT. After washing with 1× phosphate-buffered saline (PBS), cells were stained with 20 μM DRAQ5™ (Thermo Fisher Scientific, Waltham, MA, USA) for 5 min at RT. γ-H2AX foci were quantified using imaging flow cytometry.

We performed qRT‐PCR to measure mRNA expression levels. Total RNA was extracted from mice intestine tissues and Caco-2 cells using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. cDNA was synthesized using oligo-dT primers and AccuPower RT premix (Bioneer, Daejeon, Korea). The synthesized cDNA was amplified using a LightCycler 480 system (Roche, Basel, Switzerland) with specific primers. The cycling conditions were 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. Data were analyzed using StepOne software version 2.2.2 (Applied Biosystems, Waltham, MA, USA). The mRNA expression levels were normalized to GAPDH, and fold changes were calculated using the 2^-ΔΔCt method. The primers used for qRT-PCR are listed in Table 1.

Fecal DNA was extracted using the DNeasy PowerSoil^® Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. Extracted DNA was quantified using Quant-IT PicoGreen (Invitrogen). For microbiome analysis, the V3–V4 regions of the 16S rRNA gene were amplified according to the Illumina 16S Metagenomic Sequencing Library protocol. A total of 5 ng of genomic DNA (gDNA) was PCR-amplified using a 5× reaction buffer, 1 mM dNTP mix, 500 nM of universal forward and reverse primers, and Herculase II fusion DNA polymerase (Agilent Technologies, Santa Clara, CA, USA). The first PCR cycle conditions were as follows: initial denaturation at 95°C for 3 min, followed by 25 cycles at 95°C for 30s, 55°C for 30s, 72°C for 30s, with a final extension at 72°C for 5 min. PCR products were purified using AMPure beads (Agencourt Bioscience Corporation, Beverly, MA, USA). For final library construction, 2 µL of the first PCR product was subjected to a second PCR using Nextera XT Indexed Primer under the same conditions as the first PCR, except for 10 cycles. The PCR products were purified using AMPure beads. To ensure optimal cluster densities on Illumina sequencing platforms, accurate quantification of DNA library templates was performed using qPCR, following the Illumina qPCR Quantification Protocol Guide (KAPA Library Quantification Kit for Illumina Sequencing Platforms). Library quality was assessed using the TapeStation D1000 ScreenTape (Agilent Technologies). Paired-end sequencing (2 × 300 bp) was performed using a MiSeq™ sequencer (Illumina, San Diego, CA, USA).

Paired-end FASTQ files were converted into QIIME2 artifacts for further analysis. Demultiplexed data were processed using the DADA2 algorithm, which included error correction and removal of rare taxa, to generate representative sequences and a feature table. Microbial classification of each representative sequence was confirmed by blasting against the 16S rRNA gene database. The Q2-Feature classifier, a Naive Bayes classifier trained based on the SILVA reference database (V3–V4 region; https://www.arb-silva.de/), was used to classify the dataset. The resulting feature table was used to generate a phylogenetic tree for downstream alpha and beta analyses. The “core metrics analysis” command was used to calculate Shannon diversity, Pielou’s evenness, observed operational taxonomic units, and Simpson’s index. Analysis of composition of microbiomes was used to verify the differences in feature composition between groups, and the results were visualized.

SCFAs were quantitatively detected using gas chromatography-mass spectrometry (GC-MS). Fecal samples were mixed with 1 N HCl and an internal standard (0.1 M isobutanol), followed by vortexing for 10 min at RT. Extraction was performed using diethyl ether (2:1 ratio), and the mixture was centrifuged at 12,000 rpm for 5 min at 4°C to collect the supernatant. After collecting the supernatant, 1 µL was injected into a gas chromatograph-mass spectrometer (Hewlett Packard Model 7890) equipped with a DB-FATWAX Ultra Inert column. SCFAs were identified and quantified based on the retention time and standard calibration curves.

For histological examination, formalin-fixed, paraffin-embedded small intestine tissues were cut into 3 μM sections, mounted on slides, dewaxed in xylene, and rehydrated in 30–100% ethanol. The sections were stained with hematoxylin and eosin (H&E). Stained slides were scanned using a slide scanner (MoticEasyScan Pro 6).

Human Caco-2 cells, derived from human colorectal adenocarcinoma, were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM; Corning, Corning, NY, USA) supplemented with 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific) and 1% antibiotics. For in vitro studies of the intestinal barrier, we differentiated Caco-2 cells into an epithelial monolayer resembling that of enterocytes. Cells were cultured in a humidified incubator at 37°C with 5% humidity. Caco-2 cells were exposed to either fractionated or single radiation doses. The fractionated group was irradiated with 0.25 Gy/day for 4 days, totaling 1 Gy, whereas the single exposure group was irradiated with a single dose of 1 Gy. Both radiation exposures were performed using a gamma-ray machine (BIOBEAM8000, 137Cs; dose rate: 2.6 Gy/min).

ROS levels were assessed using a H₂DCFDA staining kit (#ab113851; Abcam, Cambridge, UK) according to the manufacturer’s instructions. Caco-2 cells were exposed to radiation and incubated for 24 h before ROS measurement. Cells were treated with H₂DCFDA (20 μM) in serum-free media at 37°C for 30 min in the dark. After staining, cells were washed with PBS, harvested, and resuspended in fresh PBS. The fluorescence intensity of oxidized H₂DCFDA (excitation: 488 nm, emission: 525 nm) was quantified using imaging flow cytometry. Unstained cells served as negative controls for ROS validation.

Caco-2 cells were seeded onto four-well cell culture slides for immunocytochemical analysis. Cells were fixed with 4% paraformaldehyde, blocked with 3% bovine serum albumin (BSA), and permeabilized with 1% BSA and 0.1% Triton X-100 for 30 min at RT. Primary antibody ZO-1 (#339100; Thermo Fisher Scientific) was applied at a 1:1000 dilution in blocking buffer and incubated overnight at 4°C. The following day, cells were washed thrice with PBS and incubated with Alexa Fluor 488-conjugated anti-mouse IgG (Thermo Fisher Scientific) for 1 h at RT in the dark. After three washes with PBS, the cells were counterstained with DAPI for nuclear visualization and mounted using a fluorescence mounting medium. Fluorescent images were acquired using a Carl Zeiss Axioscope 5 fluorescence microscope.

Caco-2 cells were irradiated with a single dose of 1 Gy (2.6 Gy/min) and subsequently treated with exogenous short-chain fatty acids (SCFAs). Cells were incubated with 2 mM sodium acetate (#CAS No. 127-09-3; SAMCHUN, Seoul, Korea), 2 mM sodium propionate(#Cat. No. P1880; Sigma-Aldrich St. Louis, MO, USA), or a mixture of both (2 mM each) for 24 h. Following treatment, cells were harvested for downstream analyses, including ROS measurement (H₂DCFDA assay), qRT-PCR for tight junction genes (Tjp1, OCLN, CLDN1), and immunofluorescence staining of ZO-1.

Data are presented as the mean ± standard deviation. Statistical significance was set at p < 0.05. Comparisons between groups were performed using the Kruskal–Wallis test and one-way analysis of variance. Dunnett’s test was used for multiple comparisons against the control (Con), while Tukey’s post-hoc test was used for pairwise comparisons among all groups (Con, Fraction, and Single). Correlation heatmap analyses were performed using Pearson’s correlation coefficient. Statistical analyses were performed using GraphPad Prism version 10.4.1 (GraphPad Inc., CA, USA), and SPSS Statistics version 30 (IBM Corp., NY, USA).

To examine the differential biological effects of single and fractionated radiation exposures, the mice were assigned to three groups (Con, Fraction, and Single). γ-H2AX fluorescence intensity and foci number, markers of DNA double-strand breaks, were considerably increased in peripheral blood mononuclear cells from the Single group compared to those in the Fraction and Con groups (Figure 1B; Supplementary Figure S1). CBC analysis revealed that WBC, RBC, LYM, and PLT counts were markedly reduced in the Single group compared with those in the Con group. Additionally, WBC and PLT levels were notably lower in the Fraction and Single groups than in the Con group, with the Single group showing a more marked reduction compared to the Fraction group. There were no statistically significant differences in the RBC and PLT counts between the Con and Fraction groups or between the Fraction and Single groups (Figure 1C). Histological analysis of the small intestine showed that the villus length was markedly reduced in the Single group compared to that in the Con group, and there was no significant difference between the Con and Fraction groups or between the Fraction and Single groups. In both radiation exposure groups, the crypt depth considerably increased compared to that in the Con group; in particular, the increase was greater in the Single group than in the Fraction group (Figures 1D, E). Furthermore, the expression levels of inflammatory cytokine (interleukin [IL]-6, monocyte chemoattractant protein-1 [MCP-1], IL-1β, and transforming growth factor-beta 1 [TGF-β1]) mRNAs in mouse intestinal tissues did not exhibit significant differences between the Fraction and Con groups. However, all cytokine mRNAs were upregulated in the Single group compared to those in the Con group, with IL-6 exhibiting more than a six-fold increase (Figure 1F). These results suggest that single radiation exposure induces greater biological damage than fractionated exposure.

Gut microbiota composition was analyzed from fecal samples of mice in each irradiation group using bacterial 16S rRNA gene v3–v4 amplicon sequencing. At the phylum level, Firmicutes, Bacteroidota, and Actinobacteriota were the most abundant taxa (Figure 2A). The Firmicutes/Bacteroidetes (F/B) ratio differed remarkably among the groups, with the Single group showing a three-fold increase compared to the Con group and a two-fold increase compared to the Fraction group. No significant differences were observed between the Con and Fraction groups (Supplementary Figure S2). The relative abundance of Firmicutes was 49% in the Con group, 53% in the Fraction group, and 66% in the Single group, showing a marked increase in the Single group compared to the Con group. The relative abundance of Bacteroidota was 47% in the Con group, 34% in the Fraction group, and 23% in the Single group, showing a marked decrease in the Single group compared to the Con group. At the genus level, Lachnospiraceae, Muribaculaceae, Lactobacillus, Muribaculum, Bacteroides, Alistipes, Turicibacter, Faecalibaculum, and Bifidobacterium were the dominant genera. The relative abundance of Muribaculaceae decreased in the Single group compared to that in the Con group, reaching 15%, 14%, and 10% in the Con, Fraction, and Single groups, respectively. In contrast, Lachnospiraceae abundance increased in the Single group (32%) compared to that in the Con (15%) and Fraction (14%) groups. Krona charts showed more pronounced changes in community composition after single irradiation, whereas the Con and Fraction groups showed similar compositions. Notably, the imbalance in Firumicutes (pink) and Bacteriodota (yellow) was more prominent in the Single group, whereas the Con and Fraction groups showed similar compositions (Figure 2B). The Chao1 index markedly decreased in the Single group compared to that in the Con group, with no significant differences observed between the Con and Fraction groups or between the Fraction and Single groups. The abundance-based coverage estimator (ACE) index showed no significant differences among the groups. Additionally, the Shannon and Simpson diversity indices and richness were lower in the Single group compared to those in the Con and Fraction groups, with no significant differences between the Con and Fraction groups (Figures 2C-F). Furthermore, beta diversity, which indicates the difference in the composition of the intestinal microbial community between groups, showed that the distance between the Con and Fraction groups was similar. Conversely, the Single group exhibited a greater distance than the other groups according to the principal coordinate analysis, indicating a distinct microbial community structure (Figure 2G). These results suggest that single radiation exposure induces microbial dysbiosis, characterized by reduced and altered gut microbiota composition.

Linear discriminant analysis Effect Size (LEfSe)-based Linear discriminant analysis (LDA) was performed to identify the key differences in the gut microbiome composition between the Fraction and Single groups. A taxonomic map of the gut microbiota structure (Figure 3A) highlights the most pronounced taxonomic differences between fractionated and single radiation exposures at the same dose. LEfSe-based LDA score analysis (Figure 3B) confirmed statistically significant differences in microbial abundance between the two groups (p < 0.05). In the Single group, we observed a higher relative abundance of Firmicutes (p.), Lactococcus (spp.), Desulfovibrionaceae (f.), Desulfovibrionales (o.), Desulfovibrionia (c.), and Desulfobacterota (p.). Notably, Desulfovibrionaceae and Desulfobacterota are sulfate-reducing bacteria (SRB) known to be associated with increased gut inflammation and epithelial barrier dysfunction. Conversely, the Fraction group exhibited a higher abundance of Bacteroidota (p.), Bacteroidaceae (f.), Prevotellaceae (f.), Rikenellaceae (f.), Bacteroidia (c.), Bacteroidales (o.), Alistipes (g.), Turicibacter (g.), Gastranaerophilales (o.), Vampirivibrionia (c.), and Cyanobacteria (p.). Many of these taxa, including Prevotellaceae and Bacteroidota, are known to produce SCFAs, which play protective roles in maintaining gut homeostasis and immune regulation. These findings suggest that single radiation exposure may contribute to an increase in inflammation-associated microbial taxa, whereas fractionated radiation exposure may favor the retention of SCFA-producing bacteria, potentially mitigating radiation-induced dysbiosis.

We investigated how radiation-induced changes in the gut microbiota affect the expression of SCFAs and SCFA-sensing receptors. SCFAs, primarily produced through microbial fermentation of dietary fiber, play a crucial role in host metabolism, immune regulation, and intestinal homeostasis, particularly after radiation exposure (Eaton et al., 2022). SCFA analysis revealed a marked reduction in acetic acid and propionic acid levels in the Single group compared to those in the Con group, whereas no significant differences were observed between the Con and Fraction groups. Butyrate levels did not differ significantly between groups (Figure 4A). One of the key biological functions of SCFAs is the activation of G protein-coupled receptors (GPCRs), which are essential for host health and the regulation of cellular signaling (Priyadarshini et al., 2018). The Single group exhibited notably lower expression levels of GPR41 and GPR43 than the Con and Fraction groups, whereas GPR109a expression remained unchanged across the groups (Figure 4B). These findings suggest that single radiation exposure leads to a reduction in SCFAs, accompanied by decreased expression of SCFA-sensing GPCRs, potentially contributing to radiation-induced gut dysfunction.

Correlation heatmaps revealed distinct patterns of the microbiota–metabolite–SCFA receptor signaling axis between the radiation regimens (Figure 5). In the fractionated group, Muribaculaceae abundance correlated strongly with acetate (r = 0.87, p < 0.01), propionate (r = 0.86, p < 0.01), and GPR41 (r = 0.88, p < 0.01) and GPR43 (r = 0.85, p < 0.01) expression. Similarly, Alistipes showed significant positive correlations with acetate (r = 0.84, p < 0.05), propionate (r = 0.83, p < 0.05), and GPR41/43 (r = 0.85 and r = 0.82, respectively). In contrast, in the single-dose group, these correlations were markedly weakened (r values close to 0) or reversed (negative r values), indicating disruption of SCFA-mediated GPR signaling pathways.

To investigate the effects of fractionated and single radiation exposures on the intestinal tight junction barrier, we used in vivo and in vitro models. In the small intestinal tissue of irradiated mice, the mRNA levels of the tight junction molecules Tjp1, CLDN1, CLDN3, and OCLN were considerably reduced in the Single group compared to those in the Con group (Figure 6A). Additionally, mRNA levels of tight junction molecules in the Single group were markedly lower than those in the Fraction group. The mRNA expression levels of CLDN1 were not significantly different between the Con and fraction groups. To replicate the in vivo radiation effects, Caco-2 cells were exposed to either four fractionated irradiations or a single dose of irradiation within one passage. Radiation-induced cellular responses were confirmed (Figure 6B). ROS production was assessed using H₂DCFDA staining and flow cytometry, revealing markedly higher fluorescence intensity in the Single group compared to that in the Fraction group, indicating increased oxidative stress (Figure 6C). Consistent with the in vivo findings, the Single group exhibited markedly reduced mRNA expression of tight junction-related genes (Tjp1, OCLN, and CLDN1), compared to the Fraction group. Additionally, Tjp1 and OCLN expression was markedly reduced in both radiation-exposed groups relative to the Con group, whereas CLDN1 expression did not exhibit a significant difference (Figure 6D).

Immunofluorescence analysis of ZO-1 further revealed a marked decrease in ZO-1 expression following single radiation exposure compared to fractionated exposure (Figure 6E). These findings suggest that single radiation exposure induces oxidative stress and disrupts the tight junction barrier, potentially compromising intestinal epithelial integrity.

Given the marked reduction of acetate and propionate levels observed in the Single group (Figure 4), Caco-2 cells were treated with exogenous acetate or propionate following single-dose irradiation. ROS measurement using H₂DCF-DA staining showed pronounced ROS accumulation in the Single group, which was markedly reduced by acetate and propionate supplementation (Figure 7A). RT-PCR analysis demonstrated that the Single group exhibited decreased expression of Tjp1, OCLN, and CLDN1 to 0.45–0.65 fold of control. Supplementation with acetate or propionate increased these levels by 1.3–1.6 fold relative to the Single group, with the Single+Mix condition restoring Tjp1 to 2.0-fold, comparable to the Single group (Figure 7B). Furthermore, immunofluorescence analysis revealed that ZO-1 expression, disrupted in the Single group, was partially restored along the cell membrane in acetate and propionate-treated groups (Figure 7C). Collectively, these results demonstrate that exogenous SCFA supplementation attenuates radiation-induced tight junction disruption and ROS accumulation. In addition, SCFA treatment partially restored ZO-1–based epithelial barrier integrity.