Six-week-old and 2-year-old male and female F344/NSIc rats were obtained from Orient Bio (Seoul, Korea) and maintained under specific pathogen-free conditions. Animals were housed at 23 °C with a 12-h light/dark cycle, and all animal procedures were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee of Seoul National University Bundang Hospital (approval no. BA-2305-367-005-01). Rats were assigned to one of two diets: a chow diet (3.85 kcal/g) as the control (CON), or a high-fat diet (5.24 kcal/g, 60% fat; Raonbio, D12492) and 20% fructose solution of water (Sigma-Aldrich, F0127) (HFHFD) (Lee et al., 2018; Choi et al., 2024). Eight experimental groups were established based on age, sex, and diet: 6-week-old males (CON n = 6; HFHFD n = 6), 6-week-old females (CON n = 6; HFHFD n = 6), 2-year-old males (CON n = 5; HFHFD n = 6), and 2-year-old females (CON n = 6; HFHFD n = 6). One aged male control rat died during the trial, resulting in a final sample size of n = 5 for the aged male CON group. This can occur in long-term studies using aged rodents. Over 8 weeks, caloric intake and body weight were recorded weekly. Specifically, body weight was measured once per week, and feed was replaced weekly. Food intake was calculated by subtracting the remaining food from the total amount provided at the beginning of the week. Similarly, water intake was recorded twice a week by measuring the remaining water, and a 20% fructose solution was replenished at each measurement. The caloric contribution of the fructose solution was calculated by multiplying the volume of solution consumed by the caloric content (4 kcal/g for fructose solution), as described in Ha et al. (2025). The caloric intake from the chow diet (3.85 kcal/g) and high-fat diet (5.24 kcal/g) was also calculated by multiplying the amount of food and water consumed by the corresponding caloric values. The total caloric intake from both the diets and the fructose solution was summed and divided by 7 to obtain a daily intake. At study completion, all rats were euthanized by CO₂ inhalation using a gradual-fill method in a dedicated chamber (approximately 5 L; CO₂ displacement rate ~30–40% chamber volume/min [1.5–2.0 L/min]). Death was confirmed by cessation of respiration, and jejunal tissues were promptly collected and frozen at −80 °C for downstream analyses.

The collected jejunal tissues were placed in a petri dish and processed under cold conditions. Each sample was divided into two or three segments and gently rinsed with 1 mL of ice-cold Dulbecco’s Phosphate Buffered Saline (DPBS) (Biowest, L0615-500) using a syringe (Hecking et al., 2023). The wash solution was collected with a pipette and transferred to a sterile 2 mL tube. The tissue was then incised longitudinally, rinsed again with 1 mL of DPBS, and the wash was collected in a new 2 mL tube. This tube was centrifuged at 4 °C and 2000 rpm for 5 min, and the resulting supernatant was combined with the previously collected sample.

Genomic DNA was extracted from jejunal contents using the QIAamp DNA Mini Kit (Qiagen, 51,306) according to the manufacturer’s instructions. Subsequent amplification of the 16S rRNA V3–V4 region and metagenomic sequencing were performed as described previously (Choi et al., 2024; Song et al., 2025). Metagenome sequencing was conducted on the purified PCR product from the 16S rRNA amplicon at CJ Bioscience, Inc. (Seoul, South Korea). Primer sequences were trimmed using the CJ Bioscience in-house program, and nonspecific amplicons were identified using the HMMER program.

Data processing and analysis were performed by CJ Bioscience following a previously described procedure (Song et al., 2025). Paired-end reads were assembled and subjected to quality control using an in-house bioinformatics pipeline that removed nonspecific, non-target, and chimeric amplicons. Sequencing quality metrics, including total raw reads, valid reads, and read length statistics, are provided in Supplementary Dataset S1 for further details. Taxonomic classification was performed against the EzBioCloud 16S rRNA database using USEARCH (version 8.1.1861_i86linux32), and operational taxonomic units (OTUs) were generated using UCLUST. For downstream analyses, reads were normalized to 28,783 reads per sample, and rarefaction curves were generated using EzBioCloud. Alpha- and beta-diversity analyses were performed to evaluate microbial community composition. Beta diversity was assessed using principal coordinate analysis (PCoA) based on the generalized UniFrac distance (α = 0.5) at the species level, a phylogeny-informed metric that balances the contributions of rare and abundant lineages. In generalized UniFrac, the parameter α controls the weight placed on abundant lineages (α = 1 corresponds to weighted UniFrac, while lower α values reduce dominance by highly abundant taxa); we used α = 0.5 as a commonly used robust setting (Chen et al., 2012). The relative abundance of the taxa was calculated to compare the abundance ratios among the groups. Graphical visualization of the selected taxa was performed using Prism software (GraphPad Software).

Linear discriminant analysis (LDA) effect size (LEfSe) was used to identify taxonomic biomarkers differentiating the microbiota composition among the diet groups at the species level. Group differences were assessed using the Kruskal–Wallis test (p < 0.05), and taxa that consistently differed among subgroups were confirmed using the pairwise Wilcoxon test. Features with a logarithmic LDA score > 2.0 were considered significant. Pathway analysis was performed using MinPath based on the KEGG orthology to predict the functional potential of the microbial community. Graphical visualization of the selected taxa was performed using Prism and RStudio software.

Human normal small intestinal epithelial cells (HIEC-6; ATCC, CRL-3266) were cultured in Opti-MEM (Gibco, #31985070) supplemented with 4% fetal bovine serum (FBS; Gibco, #16000044), GlutaMAX (Gibco, #35050061), 1 mM HEPES (Gibco, #15630080), and 10 ng/mL epidermal growth factor (EGF; Corning, #354052), and maintained at 37 °C in a humidified incubator with 5% CO₂. To model high-fat diet (HFD) conditions in vitro, palmitic acid (PA) (Sigma-Aldrich, P0500-10G) was used. PA was first dissolved in ethanol to prepare a 250 mM stock solution at 70 °C, then conjugated to 10% bovine serum albumin (BSA) to obtain an 8 mM PA–BSA solution by incubation at 37 °C overnight, followed by sterile filtration through a 0.22 μm filter (Knight et al., 2021). L. intestinalis (KCTC, 5052) was cultured in MRS broth (Difco, #288130) at 37 °C in a shaking incubator with 5% CO₂. For treatment, bacteria were prepared at 1 × 10⁸ CFU/mL in serum-free Opti-MEM (SFM). Conditioned medium (CM) was prepared by incubating bacteria in SFM at the same density (1 × 10⁸ CFU/mL), and the resulting CM was sterile filtered through a 0.22 μm filter.

HIEC-6 cells were seeded into 96-well plates at 1 × 10⁴ cells/well and allowed to attach for 24 h. Cells were then pretreated for 24 h with either live L. intestinalis (1 × 10⁸ CFU/mL) (Choi et al., 2024) or CM. To induce lipotoxic stress, PA was added in the presence of bacteria or CM at a final concentration of 400 μM for 12 h (Fatima et al., 2019; Ouyang et al., 2022). After PA exposure, the medium was removed, and cells were treated again with the corresponding bacteria or CM for an additional 24 h. Media were aspirated, cells were gently rinsed with DPBS (Biowest, L0615-500), and 100 μL/well of SFM was added. Cell viability was assessed using the CCK-8 assay (Dojindo, CK-04-11) according to the manufacturer’s instructions, and the absorbance was measured at 450 nm after 4 h of incubation. These experiments were repeated three times to check the reproducibility.

Statistical analyses were performed using SPSS software (version 18.0.0; IBM Corp., Armonk, NY, USA), except for the beta-diversity analysis. All analyses were conducted using nonparametric procedures that do not require equal group sizes, and the exact sample size used for each analysis is indicated in the corresponding Results and figure legends. Differences in microbial abundance ratio and alpha-diversity among groups were analyzed using the Kruskal–Wallis, followed by pairwise comparisons using the Mann–Whitney U-test. Correlations between microbial abundance ratios in jejunal contents and pathological phenotypes (jejunal inflammation score and hepatic steatosis [%]) obtained in our previous study (Ha et al., 2025) were determined using Spearman’s rank correlation (pairwise complete observations). Statistical significance was set at p < 0.05.

All the datasets generated in this study can be found in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) Database (PRJNA1393281).

At the phylum level, both male and female rats in the HFHFD group showed a decrease in Firmicutes and an increase in Bacteroidetes and Verrucomicrobia compared to the CON group (Figures 1A,B). When each group was analyzed separately, Firmicutes was significantly decreased in the young female HFHFD group, and its abundance ratio was reduced in both the male and female HFHFD groups (Figure 1C). In the young group, Firmicutes abundance was significantly higher in females than in males under both CON and HFHFD conditions (Figure 1C; Supplementary Figure 1B). Bacteroidetes abundance was significantly increased in both males and females in the aged HFHFD group (Figure 1D). In the young groups, Bacteroidetes abundance was significantly higher in male CON than in female CON rats (Figure 1D). Among females, the aged HFHFD group showed a markedly higher abundance than the young HFHFD group (Figure 1D). When the sexes were combined and the jejunal microbiota were compared according to diet and age, a consistent pattern was observed: Firmicutes decreased, whereas Bacteroidetes and Verrucomicrobia increased in the HFHFD groups (Supplementary Figure 1A). Proteobacteria showed an overall low relative abundance but tended to increase with advancing age within the same sex and dietary conditions (Figures 1A,B; Supplementary Figure 1E). Verrucomicrobia increased in response to HFHFD, showing a significant increase in the young female and aged male HFHFD groups (Supplementary Figure 1F). Moreover, young male CON rats exhibited a significantly higher Verrucomicrobia abundance than aged male CON rats and young female CON rats (Supplementary Figure 1F).

The Firmicutes/Bacteroidetes (F/B) ratio decreased across all groups fed the HFHFD, with significant reductions observed in the young female and aged male and female HFHFD groups (Figure 1E). Although the young males in the HFHFD group showed a similar decrease, the difference did not reach statistical significance, likely because of the lower baseline ratio observed in the young males in the CON group. In the young CON group, females exhibited a significantly higher F/B ratio than males, indicating sex-dependent differences (Figure 1E). When sex was excluded as a variable and only diet and age were considered, the F/B ratio remained significantly lower in the HFHFD-fed rats than in the CON group (Figure 1F; Supplementary Figure 1C). Correlation analyses between the F/B ratio and previously measured pathological phenotypes (Ha et al., 2025) revealed a negative correlation, indicating that higher degrees of jejunal inflammation and hepatic steatosis, both elevated in the HFHFD group, were associated with a lower F/B ratio (Figures 1G,H).

At the genus level, changes in the abundance were mainly observed in Lactobacillus, Bacteroides, and Akkermansia. Lactobacillus showed a decreasing trend, with a significant reduction observed only in the aged female HFHFD group (Figure 2A). Bacteroides abundance increased significantly in female HFHFD rats compared to that in CON rats, regardless of age, and in the young groups, both male CON and HFHFD rats had significantly higher Bacteroides levels than their female counterparts (Figure 2B). In our data, the genus Akkermansia includes Akkermansia muciniphila and one unclassified species, with the abundance ratio of the unclassified species being very low. Therefore, the data for the genus Akkermansia and Akkermansia muciniphila are highlysimilar, leading to similar p-values and figures (Supplementary Dataset S2). Akkermansia tended to increase following HFHFD feeding and was significantly higher in young female HFHFD rats and in both sexes in the aged HFHFD group (Figure 2C). In both age groups, male CON rats showed significantly higher Akkermansia abundance than females, and the young male HFHFD group also exhibited higher Akkermansia abundance than the young female HFHFD group (Figure 2C).

Similar patterns were observed at the species level. Lactobacillus intestinalis, a commensal lactic acid bacterium residing in the small intestine (Wang et al., 2023), showed a decreasing tendency under HFHFD feeding, with significant reductions in young male HFHFD and old female HFHFD groups, indicating age-, sex-, and diet-dependent variations (Figure 2D). Bacteroides caccae, reported as an opportunistic pathogen (Schwenger et al., 2025), was significantly higher in female HFHFD rats than in the CON group, irrespective of age, whereas in the young HFHFD group, males exhibited significantly lower levels than females (Figure 2E). Akkermansia muciniphila, a beneficial commensal bacterium (Zhao et al., 2024), tended to increase in the HFHFD-fed rats, with significant elevations in young female HFHFD rats and in both sexes in the aged HFHFD groups (Figure 2F). Independent of age, male CON rats had a significantly higher A. muciniphila abundance than females, and young male HFHFD rats also showed higher levels than young female HFHFD rats (Figure 2F). The Lactobacillus reuteri group, known to be commensal (Lee et al., 2025), abundance significantly decreased in HFHFD-fed rats of both sexes at a young age, with male HFHFD rats showing lower levels than female rats, whereas no marked changes were observed in the aged groups (Figure 2G). Bacteroides vulgatus, an opportunistic pathogen (Bahitham et al., 2025), was notably more abundant in young male rats, with the young-male CON group displaying significantly higher levels than the female CON and aged male CON group. The effect of HFHFD was evident only in the aged groups, where B. vulgatus was significantly increased in both male and female HFHFD rats compared to their respective CON groups (Figure 2H). Finally, to provide a comprehensive overview of the jejunal microbial landscape, we analyzed the overall community composition at the genus and species levels (Figures 2I,J). A total of 483 genera and 1,624 species were identified, reflecting a highly diverse community. While the ‘ETC’ category (taxa with <1% average abundance) appeared relatively large due to this extensive diversity, the stacked bar plots clearly illustrate the dominant shifts in the community structure. These holistic profiles confirm that the individual taxonomic changes described above collectively contribute to a distinct remodeling of the rat jejunal microbiota, driven by the complex interplay of age, sex, and diet.

Spearman’s rank correlation analysis was performed to determine whether the abundance ratio at the species level was associated with HFHFD-induced pathological phenotypes, including jejunal inflammation and hepatic steatosis (Ha et al., 2025). L. intestinalis was negatively correlated with jejunal inflammation, indicating that a lower abundance of L. intestinalis was associated with higher inflammatory scores (Figure 3A). Similarly, L. intestinalis negatively correlated with liver steatosis (Figure 3B). In contrast, B. caccae positively correlated with both jejunal inflammation and hepatic steatosis, with a higher B. caccae abundance corresponding to more severe pathological features (Figures 3C,D). A. muciniphila positively correlated with jejunal inflammation and hepatic steatosis, with increased abundance and progressive pathological phenotypes (Figures 3E,F). L. reuteri group showed patterns similar to those of L. intestinalis, with negative correlations indicating a reduced abundance under jejunal inflammation and steatosis (Figures 3G,H). B. vulgatus was positively correlated with hepatic steatosis (Figure 3I).

At the genus level, Lactobacillus negatively correlated with jejunal inflammation and liver steatosis (Supplementary Figures 2A,B). In contrast, Bacteroides positively correlated with these pathological phenotypes (Supplementary Figures 2C,D). The genus Akkermansia was positively correlated with jejunal inflammation and steatosis (Supplementary Figures 2E,F).

Beta-diversity analysis using principal coordinate analysis (PCoA) revealed age- and sex-dependent separation patterns in the jejunal microbiota. In the young groups, all comparisons showed significant separation, except for the male CON and male HFHFD groups (p = 0.058, q = 0.058), which did not differ from each other (Figure 4A). Young female rats showed a clear separation between the CON and HFHFD groups (p = 0.013, q = 0.020), as well as from male CON rats (6-week-old F CON vs. M CON, p = 0.019, q = 0.023), indicating distinct effects of both diet and sex on the microbial community structure (Figure 4A). In contrast, the sex-dependent separation observed in young rats was diminished in aged rats (Figure 4B). Although both male and female rats showed a significant separation between the CON and HFHFD conditions (2-year-old M CON vs. M HFHFD, p = 0.002, q = 0.012; 2-year-old F CON vs. FHFHFD, p = 0.005, q = 0.014), the difference between male CON and female CON was no longer significant (p = 0.139, q = 0.167), and no separation was observed between the male and female HFHFD groups (p = 0.610, q = 0.610, Figure 4B). These findings indicate that sex-dependent differences in the microbiota structure present at a young age are attenuated with age.

When sex was analyzed independently, male rats showed no significant differences between the young CON and young HFHFD groups (p = 0.067, q = 0.067), whereas all other comparisons showed significant differences (Supplementary Figure 3A). In female rats, both diet and age contributed to a significant separation across all group comparisons, suggesting a stronger response of the female microbiota to HFHFD and aging (Supplementary Figure 3B). Alpha-diversity indices, including OTUs, Chao1, Shannon index, and phylogenetic diversity, did not show significant differences among the groups, as the values across the groups were highly variable (Supplementary Figures 2C–F).

Using LEfSe analysis, we identified taxonomic biomarkers that were differentially enriched according to diet with clear age- and sex-dependent patterns. A. muciniphila was enriched in all HFHFD groups except young male rats, showing a consistently higher abundance in young female HFHFD rats and in both sexes of the aged HFHFD groups. In the young CON groups, both males and females showed significantly higher levels of L. reuteri than their HFHFD counterparts, indicating that this species is a shared biomarker of the CON condition (Figures 5A,B; Table 1). In the aged CON group, Bacillus smithii, Lactobacillus jensenii group, and Geobacillus toebii were enriched, representing age-associated biomarkers present only under non-HFHFD conditions (Figures 5C,D; Table 1). Conversely, among the HFHFD-fed rats, B. vulgatus was consistently enriched in both males and females in the aged groups. In addition, an unclassified species, PAC001065_s, was uniformly enriched across all HFHFD groups, indicating a diet-driven increase in specific, potentially opportunistic taxa (Table 1). Several additional biomarkers appeared in an age- or sex-specific manner. However, these species were restricted to individual comparisons and did not form consistent patterns across groups.

To investigate the sex-dependent differences, we performed pairwise comparisons within each age and diet group. In the young male CON, young male HFHFD, and aged male CON groups, A. muciniphila, B. vulgatus, and the two unclassified species were significantly enriched compared to their female counterparts (Figures 6A–C). In addition, both young male CON and HFHFD rats showed a higher abundance of 21 overlapping unclassified species than females (Table 2). Conversely, young female rats exhibited higher levels of Muribaculum intestinale and the four unclassified species than male rats (Figures 6A,B; Table 2). Notably, the Clostridium celatum group displayed a sex- and age-dependent pattern, being enriched in young female CON rats and aged male CON rats (Figures 6A,C).

To assess age-related differences, we compared young and aged rats irrespective of sex. In the CON group, AB626927_s was specifically enriched in young rats, whereas Bacillus smithii and eight unclassified species were enriched in aged CON rats (Table 3). Among the HFHFD-fed rats, Streptococcus acidominimus group, a known pathogen (Wu et al., 2014), was consistently enriched in aged rats, regardless of sex (Figures 7B,D; Table 3). Across all age groups, B. caccae, Veillonella caviae, and four unclassified species were elevated (Figures 7A–D; Table 3). Furthermore, young male rats showed a higher abundance of A. muciniphila, B. vulgatus, Monoglobus pectinilyticus, and 11 unclassified species independent of diet (Figures 7A,B; Table 3), whereas aged male rats showed higher levels of M. intestinale and nine unclassified species (Figures 7A,B; Table 3).

To further characterize taxonomic biomarkers that were not shared across comparisons, non-overlapping taxa were classified according to their reported ecological roles as commensals, opportunistic pathogens, or functionally uncharacterized pathogens. Commensal species are described as beneficial members of the gut microbiota, contributing to maintaining intestinal homeostasis. Opportunistic pathogens are now identified as species that may cause disease under certain conditions, such as dysbiosis or when the host’s immune system is compromised. In comparison by sex, in the young CON group, male rats showed a higher abundance of commensal A. muciniphila and one opportunistic pathogen, B. vulgatus, whereas female rats exhibited enrichment of three commensal species (Muribaculum intestinale, Clostridium celatum, Romboutsia timonensis) together with one opportunistic pathogen, Mycoplasma wenyonii (Figure 6A). In the young HFHFD group, males were characterized by the enrichment of two commensal species (A. muciniphila, Monoglobus pectinilyticus) and one opportunistic pathogen (B. vulgatus), whereas females showed a higher abundance of four commensal species (Lactobacillus murinus, Flintibacter butyricus, M. intestinale, L. reuteri) and one opportunistic pathogen (B. caccae) (Figure 6B). In the aged CON group, males exhibited an enrichment of two commensal species (A. muciniphila, C. celatum) along with one opportunistic pathogen (B. vulgatus). In contrast, in the aged HFHFD groups, females showed a higher abundance of the commensal species Bacteroides uniformis (Figures 6C,D). Taken together, the sex-dependent taxonomic biomarkers indicated reduced commensal abundance in young males, whereas in aged rats, commensal enrichment was relatively reduced in females.

Predictive functional profiling was performed based on KEGG pathways, and representative functional differences were visualized using heatmaps and scatter plots. All comparisons were conducted in a pairwise manner, where deeper red indicated higher functional abundance and deeper blue indicated lower abundance. In young males, the predicted pathway abundances of tyrosine metabolism (ko00350), the pentose phosphate pathway (ko00030), sulfur metabolism (ko00920), and cysteine and methionine metabolism (ko00270) were significantly lower in the HFHFD group than in the CON group (Figure 8A). Young females exhibited similar reductions in ko00270 and ko00030, and d-alanine metabolism (ko00473) also showed decreased predicted pathway abundance (Figure 8B). In the non-metabolism (“others”) category, Staphylococcus aureus infection (ko05150), ribosome biogenesis in eukaryotes (ko03008), and fluid shear stress and atherosclerosis (ko05418) were significantly higher in CON in both young males and females (Figures 8A,B). In the aged group, both males and females showed a higher predicted pathway abundance of D-alanine metabolism (ko00473) and non-ribosomal peptide structures (ko01054) in the CON group than in the HFHFD group (Figures 8C,D). In the other category, the phosphotransferase system (ko02060) was commonly reduced under HFHFD in both sexes. Aged males demonstrated a significant reduction in tryptophan metabolism (ko00380) under CON conditions (Figure 8C).

To evaluate sex-dependent differences, comparisons within the young groups revealed that males exhibited higher predicted pathway abundances for valine, isoleucine, and leucine biosynthesis (ko00290), steroid hormone biosynthesis (ko00140), and lipoic acid metabolism (ko00785) than females in both the CON and HFHFD groups (Figures 8E,F). Tryptophan metabolism (ko00380) was lower in young CON females, whereas nicotinate metabolism (ko00760) was significantly lower in young HFHFD females (Figures 8E,F). In the other category, apoptosis (ko04214) and biosynthesis of terpenoids and steroids (ko01062) were consistently lower in females. In young HFHFD rats, the predicted pathway abundance of the insulin signaling pathway (ko04910) and proteoglycans in cancer (ko05205) was also significantly lower in females. Interestingly, no sex-dependent differences were observed in either the CON or HFHFD groups at 2 years of age, indicating that sex-specific functional distinctions were diminished in aged rats.

Next, age-dependent differences were assessed. In the metabolism category, peptidoglycan biosynthesis (ko00550), ko00785, ko00260, ko00140, and ko00290 were significantly lower in aged male CON rats than in young male CON rats (Figure 8G). In the other category, legionellosis (ko05134), DNA replication (ko03030), and nucleotide excision repair (ko03420) exhibited higher predicted pathway abundance in young males, whereas flagellar assembly (ko02040) was lower in young males (Figure 8G). In female CON rats, no age-related differences were observed in metabolism-related pathways. However, in the other categories, lysosomes (ko04142), Mitogen-Activated Protein Kinase signaling pathway (ko04011), and Huntington’s disease (ko05016) were significantly higher in the aged group (Figure 8H). In HFHFD male rats, age-dependent changes in metabolism pathways were similar to those observed in male CON rats: ko00785, ko00140, and ko00290 were significantly lower in aged rats (Figure 8I). Additionally, Oxidative phosphorylation (ko00190) was significantly higher in the young group (Figure 8I). In the other category, ko05134 and ko03030 remained higher in young male rats, consistent with CON rats, and thyroid hormone synthesis (ko04918), mismatch repair (ko03430), ko05205, and ko01062 levels were significantly elevated in young HFHFD male rats (Figure 8I). Metabolic pathways showed no age-dependent differences in female HFHFD rats. However, nucleotide excision repair (ko03420), sulfur relay system (ko04122), biosynthesis of secondary metabolites (ko01110), and ko03430 were significantly more abundant in young females than in aged females (Figure 8J).

HIEC-6 cells were exposed to 400 μM PA to establish a high-fat diet mimicking condition, as schematically illustrated in Figure 9A. Cells were treated with live L. intestinalis or CM to evaluate their potential protective effects against PA-induced lipotoxic stress. PA treatment alone significantly reduced cell viability compared to the control group (Figure 9B). Treatment with L. intestinalis significantly restored cell viability relative to that in the PA group, whereas treatment with CM did not show a consistent protective effect (Figure 9B). These in vitro experiments were repeated for three times.