Minimum Essential Medium (MEM) with Earle’s Balanced Salts, fetal bovine serum (FBS), nonessential amino acid (NEAA), antibiotics (penicillin/streptomycin solution), and sinapic acid (3,5-dimethoxy-4-hydroxycinnamic acid) were obtained from Thermo Fisher Scientific (Logan, UT, United States). Twelve-well Transwell and Hanks’ Balanced Salt Solution (HBSS) were purchased from Welgene (Gyeongsan, Gyeongsangbuk-do, Republic of Korea). Antibodies against p65, p-p65, IKKα/β, p-IKKα/β, IκBα, p-IκBα, MLCK, p-MLC, E-cadherin, MKK4/7, p-MKK4/7, P38 MAPK, p-P38 MAPK. SAPK/JNK, p-SAPK/JNK, ERK, p-ERK, and α/β-Tubulin were purchased from Cell Signaling Technology (Danvers, MA, United States). Antibodies against lamin B1, ZO-1, occludin, claudin-4, claudin-2, ATF-2, and p-ATF2 were purchased from Abcam (Cambridge, Cambridgeshire, United Kingdom). The primary antibody against β-actin was purchased from Santa Cruz Biotech (Santa Cruz, CA, United States). Lipopolysaccharide (LPS) derived from Samonella enterica serotype typhimurium was purchased from Sigma-Aldrich (St. Louis, MO, United States), and tumor necrosis factor (TNF)-α was purchased from NKMAX (Seongnam, Gyeonggi-do, Republic of Korea).

Human epithelial colorectal adenocarcinoma Caco-2 cells were purchased from the Korean Cell Line Research Foundation (Seoul, Republic of Korea) and cultured in MEM medium containing 10% FBS, 1% NEAA, and 1% antibiotics (100-U/mL penicillin and 100-μg/mL streptomycin) under a humidified atmosphere of 5% CO₂ at 37°C. The medium was replenished every 2 days. Caco-2 cells were cultured for 21 days, and to ensure complete differentiation, the MEM medium was replaced every 2 days.

Caco-2 cells were seeded at a density of 1 × 10⁶ cells per mL in 96-well plates, incubated overnight, and subsequently treated with sinapic acid (12.5, 25, and 50 μM) for 24 h. Cell viability was measured using 3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) and phenazine methosulfate (PMS) reagents (Promega, Madison, WI, United States) according to the manufacturer’s instructions. PMS + MTS solution (1:20) was added (20 μL per well) and incubated for 1 h. The absorbance was measured at 490 nm using a microplate reader (Bio-Rad Inc., Hercules, CA, United States).

Caco-2 cells were seeded at a density of 1 × 10⁵ cells per mL in the apical chambers of 12-Transwell (Corning, NY, United States) and the MEM medium was placed in the basolateral chambers of 12-Transwell. Caco-2 cells were differentiated. After removing the original medium, it was washed twice with HBSS, incubated for 30 min by adding HBSS again, and subsequently stabilized on the bench for 15 min. TEER values were measured using an epithelial volt-ohm meter (Millicell ERS-2, Millipore, Burlington, MA, United States) according to the manufacturer`s instructions at 0 h. Based on our previous study, we confirmed that stimulation with 20 μg/mL LPS and 20 ng/mL TNF-α is optimal for decreasing TEER values. Therefore, in this study, we used a combination of 20 μg/mL LPS and 20 ng/mL TNF-α (hereafter referred to as “stimulus”). Caco-2 cell monolayers were pretreated with 12.5, 25 and 50 μM sinapic acid. After 1 h, the stimulus was administered. TEER values were measured at 0.5, 1, 3, 6, 12, and 24 h following the administration of the stimulus.

Caco-2 cells were seeded and differentiated in the same way as the 2.4 TEER measurement method, and sinapic acid and stimulus were treated under the same conditions. After 1 h, 0.5 mL of 1 mg/mL 4 kDa FITC-dextran (Sigma-Aldrich, MO, United States) in HBSS was added to the apical chamber, and 1.5 mL of HBSS was added to the basolateral chamber. After 12 or 24 h, HBSS (100 μL) was drawn from the basolateral chamber and placed on the black bottom plate. The absorbance was measured at 480 nm and 520 nm using SPECTRAmax^® GEMINI-XS (Molecular Devices, Sunnyvale, CA, United States).

Caco-2 cells were seeded at a density of 1 × 10⁵ cells per mL in the six-well plates. After 1 h of sinapic acid pretreatment, differentiated Caco-2 cells were treated with the stimulus. The total RNA was extracted from Caco-2 cells according to the specifications of RNAiso Plus (Takara Bio Inc., Kyoto, Japan). Next, cDNA was obtained using the ReverTra Ace^® qPCR RT Kit & Master Mix (Toyobo, Osaka, Japan) based on the manufacturer’s instructions. The produced cDNA was used for qPCR with SYBR^® Green Realtime PCR Master Mix (Toyobo, Osaka, Japan) according to the manufacturer’s instructions. The primer sequences are displayed in Table 1. Relative gene expression levels were normalized using β-actin and the comparative ΔΔCq method using the CFX Maestro Software (Bio-Rad Inc., Hercules, CA, United States).

Caco-2 cells were seeded at a density of 1 × 10⁵ cells per mL in the six-well plates. Differentiated Caco-2 cells were lysed using the M-PER buffer (Thermo Scientific, MA, United States) with a protease and phosphatase inhibitor cocktail (Thermo Scientific, MA, United States). Subsequently, the lysates were centrifuged at 18,000 × g for 10 min at 4°C and supplemented with 10× TNC buffer. After 1 h, the lysates were treated with sinapic acid and digested with pronase (protein-t- pronase ratio, 1:1,000) for 30 min. Digestion was stopped by adding a 5× SDS sample buffer, and proteins were then visualized using western blot assay.

Eight-week-old female C57BL/6 mice were purchased from Joongah Bio (Suwon, Republic of Korea). Mice were housed in an air-conditioned room (23°C ± 2) and exposed to a 12/12 h light/dark cycle with free access to food and water. All animals were treated with care, and the study protocol (KNU-2021–0026) was approved by Kyungpook National University and performed in accordance with its animal use guidelines. Mice were randomly allocated into the following groups: control, 2% DSS, 2% DSS +2 mg/kg/day sinapic acid treatment, and 2% DSS +10 mg/kg/day sinapic acid treatment. During the experiment, mice were treated with sinapic acid (2 or 10 mg/kg) daily for 15 days by oral gavage, using an esophageal catheter. Colitis was induced in mice with 2% DSS (MP Biomedicals Korea, Seoul, Republic of Korea). After 15 days, mice were anesthetized with ketamine and sacrificed using the cervical dislocation method.

For in vitro analysis, Caco-2 cells were seeded at a density of 1 × 10⁵ cells per mL in six-well plates for 3 weeks. After 1 h of pretreatment with 50 μM sinapic acid, the stimulus was administered, and cells were incubated for 0.5, 0.75, 1, and 2 h, washed twice with cold PBS, and then centrifuged at 300 ×g for 15 min. For in vivo analysis, the mice colon tissues washed with a wash solution were placed in 2 mL microcentrifuge tubes containing 1 mL of cell permeabilization buffer and stainless-steel beads and homogenized using Precellys 24 dual tissue homogenizer (Bertin, Montigny-le-Bretonneux, Montigny-le-Bretonneux, France). Subsequently, 500 μL cell permeabilization buffer was added to homogenized tissues. Homogenized tissues were centrifuged at 16,000 ×g for 15 min. The separation of cytoplasmic and membrane proteins of obtained cells and homogenized colon tissues was performed using membrane solubilization buffer and cell permeabilization buffer (Thermo Scientific, Logan, UT, United States) according to the manufacturer’s protocol. Separated proteins were then visualized using a western blot assay.

Caco-2 cells were seeded at a density of 1 × 10⁵ cells per mL on six-well plates. Differentiated cells were pretreated with 50 μM sinapic acid for 1 h and then treated with the stimulus. Cells were washed twice with cold PBS. Caco-2 cells were scraped and collected in 1 mL PBS and then each sample was centrifugated at 500 ×g for 5 min. Following PBS removal, cytoplasmic and nuclear proteins were separated using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, Logan, UT, United States) according to the manufacturer’s protocol. Separated proteins were then visualized using western blot assay (Kim et al., 2022).

For in vitro analysis, Caco-2 cells were seeded at a density of 1 × 10⁵ cells per mL on six-well plates. Differentiated cells were pretreated with 50 μM sinapic acid for 1 h and then treated with the stimulus. Cells were washed twice with cold PBS. Caco-2 cells were scraped and collected in 200 μL of lysis buffer (Cell Signaling Technology, Danvers, MA, United States) containing a protease and phosphatase inhibitor cocktail (Thermo Scientific, MA, United States). For in vivo analysis, mice colon tissues washed with PBS were placed in 2 mL microcentrifuge tubes containing 500 μL of lysis buffer and stainless-steel beads and homogenized using Precellys 24 dual tissue homogenizer. Homogenized tissues were centrifuged at 13,652 ×g for 15 min. Lysed cells and colon tissues were vortexed and maintained on ice for 30 min and subsequently centrifuged at 18,928×g for 15 min. The total protein content was quantified using the DC Protein Assay Kit reader (Bio-Rad Inc., Hercules, CA, United States) and separated by 8%, 10%, and 12% SDS-PAGE gels. The total protein was transferred from the gel to 0.45 μm PVDF membranes (Immobilon-P transfer membrane, Millipore, Burlington, United States), blocked with 5% skim milk in Tris Buffered Saline with Tween ^® 20 (TBST) buffer for 1 h at room temperature, and then hybridized with primary antibodies overnight at 4°C. Following hybridization with horseradish peroxidase (HRP)-conjugated secondary antibody (Thermo Scientific, MA, United States) for 1 h at room temperature, immune-reactive proteins were detected using a chemiluminescence detection kit (ATTO, New York, United States) and GeneGnome XRQ NPC (Syngene, Nuffield Rd, United Kingdom).

For in vitro immunofluorescence analysis, Caco-2 cells were seeded at a density of 1 × 10⁵ cells per mL on an eight-well chamber (ibidi, GmbH, Gräfelfing, Germany) for 3 weeks. Cells were pretreated with 25 and 50 μM sinapic acid for 1 h and then treated with the stimulus. For in vivo immunofluorescence analysis, colon tissues were fixed using an OCT compound (Leica Microsystems, Seoul, Republic of Korea) and cryosectioned into 10 μM sections. Cells and sectioned colon tissues were fixed with 4% formaldehyde for 15 min and washed three times with PBS. Cells and sections were permeabilized with ice-cold 100% methanol and then blocked and incubated overnight at 4°C with the primary antibodies occludin (1:100) and ZO-1 (1:200). After washing, the cells and sections were incubated with a goat anti-rabbit immunoglobulin G H&L conjugated to the Alexa Fluor^® 488 and 647 secondary antibodies (Abcam, Cambridge, United Kingdom) for 1 h. Subsequently the cell nucleus was stained with VECTASHIELD^® Antifade Mounting Medium with DAPI (Vector Laboratories Inc., Burlingame, United States).

Colon tissue sections were stained with filtered 2.9% hematoxylin (Sigma-Aldrich, St. Louis, MO, United States). After rinsing in cool running water for 5 min, sections were dipped in 0.5% eosin (Sigma-Aldrich, St. Louis, MO, United States) five times. After rinsing in water for 5 min, sections were sequentially dehydrated using 50% ethanol, 70% ethanol, and 95% ethanol. Sections were then dipped in xylene (Sigma-Aldrich, St. Louis, MO, United States) for cleaning and mounted with quick-hardening mounting medium (Sigma-Aldrich, St. Louis, MO, United States) for observation. The histological evaluation of colon injury was conducted by three investigators in a blind setting. The criteria for assigning histological evaluation scores of colon injury are described in Table 2 (Fung and Putoczki, 2018).

DNA was extracted from mice fecal samples using the Powerfood Microbial DNA Isolation kit (Mo Bio Laboratories, Inc., Carlsbad, CA, United States) according to the manufacturer’s protocol. Intestinal bacterial community compositions were analyzed using 16S rRNA gene sequencing, which was performed according to a previous report (Ryu et al., 2022). The obtained data were analyzed using Mothur (v. 1.45.3). To visualize differences in the microbiota composition between groups, nonmetric multidimensional scaling plots were generated on the basis of unweighted and weighted UniFrac distances. The α-diversity and relative abundance of OTUs at family and genus levels were analyzed using the MicrobiomeAnalyst web-based platform (Chong et al., 2020).

Data were expressed as mean ± standard deviations of three independent experiments wherever appropriate. Significant differences were calculated using one-way analysis of variance and Tukey’s post hoc multiple comparison test. A p-value of <0.05 was considered statistically significant.

To verify the nutraceutical and pharmaceutical activities of sinapic acid in the treatment of IBD, we evaluated TEER value and FITC as the major indicators for determining intestinal permeability in vitro. The results demonstrated that sinapic acid significantly inhibited stimulus-induced decrease in the TEER value (Figures 1A, B) and FITC-dextran permeability at both 12 and 24 h (Figure 1C) without inducing cell cytotoxicity in Caco-2 cells (Figure 1D). Additionally, sinapic acid significantly suppressed the stimulus-induced mRNA expression of proinflammatory cytokines IL-1β, IL-6, and TNF-α in differentiated Caco-2 cells (Figure 1E). The proper localization of tight junction and adherence junction proteins in the membranes of intestinal epithelial cells plays a critical role in regulating intestinal permeability and maintaining the integrity of the intestinal epithelial barrier (Odenwald and Turner, 2017). Therefore, we analyzed the changes in the localization of tight junction proteins induced by sinapic acid over time. After fractionating the membrane proteins and conducting western blot assays, we found that sinapic acid significantly suppressed the stimulus-induced delocalization of ZO-1 and claudin-2 at 30 min and occludin at 45 min in the membrane fraction of Caco-2 cells (Figure 1F). Immunofluorescence analysis photographs also support these results (Figures 1G, H). Based on these results, we confirmed that sinapic acid regulates abnormal intestinal barrier permeability within a relatively short period, (i.e., 30 and 45 min) by inhibiting the delocalization of tight junction proteins in cell membranes under inflammatory conditions.

Owing to the rapid effect of the stimulus on the delocalization of tight junction proteins, we focused on the impact of sinapic acid on early response inflammatory signaling molecules including NF-κB. NF-κB is a transcriptional factor for mlck, which subsequently phosphorylates MLC, leading to the delocalization of tight junction proteins (Drolia et al., 2018). We found that sinapic acid inhibited stimulus-induced increase in MLCK expression and phosphorylation of MLC at 30 min (Figure 2A) and phosphorylation of p65 and IκBα at 15 min (Figure 2B). Lan et al. reported that sinapic acid suppressed the stimulus-induced translocation of p65 at 30 min by about 30% compared to the stimulus only group (Figure 2C). A recent study`s demonstrated that ATF-2 is a novel transcription factor for mlck expression (Kaminsky et al., 2021). In addition to NF-κB, we found that ATF-2, another transcriptional factor that regulates mlck, was phosphorylated at 15 min and translocated to the nucleus at 30 min following stimulus administration (Figures 3A, B). These alternations were suppressed by sinapic acid (Figures 3A–C).

In our previous study, we have shown that phytochemicals directly target early response kinases and subsequent inflammation and carcinogenesis in vitro and in vivo (Jung et al., 2008). Therefore, we hypothesized that sinapic acid may affect upstream kinase that regulates NF-κB and ATF-2 transcriptional factors. Phosphorylation of MAPK was confirmed under the hypothesis that MAPK may be involved in the phosphorylation of ATF-2 (Desai et al., 2013). Sinapic acid significantly inhibited the stimulus-induced phosphorylation of JNK1/2 and p38 at 15 min of stimulus exposure in differentiated Caco-2 cells but did not affect the phosphorylation of ERK1/2 (Figure 4A). Sinapic acid inhibited the stimulus-induced phosphorylation of MKK4/7 at 10 min in differentiated Caco-2 cells (Figure 4B). Although we initially suspected that TAK1 could be a major regulator of MAKK and NF-κB pathways, we found that sinapic acid did not affect the phosphorylation of TAK1 (Figure 4B). Therefore, we predicted that sinapic acid may directly affect TAK1 by binding to it. The DARTS assay revealed that sinapic acid suppressed pronase-induced TAK1 degradation in differentiated Caco-2 cells (Figure 4C). These results collectively suggest that sinapic acid directly bound to TAK1, which suppressed the activation of TAK1-mediated NF-κB and ATF-2 transcriptional factor.

Although studies have shown that sinapic acid alleviates colitis in TNBS-induced Balb/c female mice (Lee, 2018) and DSS-induced Kunming mice (Qian et al., 2020), the effect of sinapic acid on the homeostasis of gut microbiota has not yet been elucidated. Sinapic acid was administered orally for 15 days from 8 days before DSS intake, and ulcerative colitis was induced via ad libitum intake of 2% DSS for 8 days (Figure 5A). We observed that oral administration of sinapic acid ameliorated DSS-induced colon shortening (Figures 5B, C) and production of inflammatory cytokines, including IL-1β and TNF-α (Figure 5D) in mice. H&E staining and histological scores of mice colon tissue indicated that oral administration of sinapic acid mitigated DSS-induced colon tissue damage, such as loss of crypts and destruction of the mucus layer (Figures 5E, F). Western blot assay using a membrane fraction of mice colon, along with immunofluorescence analysis, showed that sinapic acid effectively suppressed the DSS-induced delocalization of ZO-1, occluding, and claudin-2 (Figures 5G–I). These results confirmed that the in vitro localization of tight junction proteins was also reflected in animal intestinal tissues.

A recent study has summarized current findings that phytochemicals may act as prebiotics for modulating the gut microbiota (Xu and Zhang, 2022). However, current nutraceutical research on gut health is focused on dietary fiber that improves gastrointestinal motility (Makki et al., 2018). Therefore, we aimed to evaluate the effect of oral administration of sinapic acid on changes in intestinal microflora. Shannon diversity, an indicator of microbial richness and evenness, and Chao1 diversity, an indicator microbial richness, are representative α-diversity indexes. Oral administration of sinapic acid significantly prevented DSS-induced decrease in Shannon diversity (Figure 6A). Although there was no significant difference, sinapic acid prevented DSS-induced decrease in Chao1 diversity (Figure 6B). Next, we examined β-diversity distances using weighted and unweighted UniFrac distances. Weighted and unweighted UniFrac distances significantly differed between the control and 2% DSS groups (p < 0.001; Figures 6C, D). Although not significant, the weighted UniFrac distance from sinapic acid groups to the control group was smaller than the distance from the 2% DSS group to control group (Figure 6C). There was no significant difference in the composition of microbiota at the family level between the groups (Figure 6E). However, at the genus level, we found that sinapic acid inhibited DSS-induced declines in the relative abundance of Ligilactobacillus and Limosilactobacillus in feces of C57BL/6J mice (Figures 6F, G). Limosilactobacillus reuteri exerts anti-inflammatory effects, ameliorates hepatic disorders, and prevents infections caused by enterohemorrhagic Escherichia coli and Helicobacter pylori through the production of bacteriocins (Abuqwider et al., 2022). Firmicutes and Bacteroidetes are the major phyla of human gut microbiota, and the Firmicutes to Bacteroidetes ratio (F/B ratio) has been widely studied as an indicator of the correlation between gut microbes and intestinal health (Stojanov et al., 2020). Oral administration of sinapic acid was found to suppress DSS-induced decrease in F/B ratio (Figure 6H). These results demonstrate, for the first time, that sinapic acid can act as a prebiotic.