Crafting probiotics a self-polymerized smart outfit for effective ulcerative colitis treatment Authors: Zixin Wang, Yanwen Zhang, Yujia Lu, Zicheng Cong, Fan Rong, Zhao Li, Junsheng Chen, Yin Wang Published: 2026 ABSTRACT Probiotic therapy has recently emerged as a promising approach for ulcerative colitis (UC) treatment. However, oral probiotic therapy is often hindered by challenges such as inferior colonization at lesion sites, where elevated intestinal hydrogen sulfide (H2S) level exacerbates epithelial barrier disruption. Moreover, without adjunct drugs, probiotics alone struggle to mitigate oxidative stress and persistent inflammation effectively. To address these issues, an alkali-labile polymerized ferulic acid (pFA) outfit was developed through a self-polymerization strategy, designed to both shield probiotics and relieve intestinal symptoms. Zinc ions were further incorporated as an H2S scavenger, enhancing the probiotic colonization and promoting gut microbiota restoration. Thus, the synthesized probiotics (designated as Lac@FZ) notably ameliorated UC severity, including relieving inflammation, alleviating oxidative stress, and restoring intestinal barrier and microbiota. This study paves a novel way for designing a boosted probiotic therapeutic for UC treatments, while also inspiring the development of other polyphenol-based functional nanoplatforms via self-polymerization.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s12951-026-04650-5. Introduction Ulcerative colitis (UC) is a chronic inflammatory bowel disorder, in which dysregulated intestinal barrier function and compromised epithelial integrity perpetuate persistent mucosal inflammation [1, 2]. In particular, the barrier impairment is closely linked to microbiota dysbiosis and aberrant immune activation [3]. Although various drugs such as 5-aminosalicylic acid, glucocorticoids, and immunosuppressants are available for UC treatment, they primarily target inflammatory pathways, offering limited modulation of the gut microbiota, and are often associated with safety issues such as nephrotoxicity and infection risks [4–6]. Therefore, novel pathogenesis-targeted therapies with improved efficacy are urgently needed. Oral probiotic therapy has recently emerged as a promising modality for UC treatment, as probiotics could inhabit the intestinal tracts and exert beneficial effects by regulating microbial homeostasis [7–9]. However, bare probiotics are highly susceptible to the harsh acidic conditions of gastrointestinal (GI) tracts when administered orally [10, 11]. This susceptibility results in poor bioavailability and low gut colonization efficiency, ultimately undermining their therapeutic outcome. To enhance the acid tolerance, various strategies aiming at shielding probiotics have been established. The most commonly developed bacterial protective coatings are composed of polysaccharides [12–15], metal-phenolic networks [16, 17], polymeric matrices [18, 19], and among others [20, 21]. For instance, She et al. encapsulated Akkermansia muciniphila within gelatin-based porous microgels, which exhibited acid-responsive pore-sealing behavior under gastric conditions [17]. Zan et al. armored Escherichia coli Nissle 1917 with a hyaluronic acid-functionalized procyanidin-Fe3+ coating, enhancing its resistance against external stressors [16]. Despite these advances, current coatings for probiotic protection are still faced with suboptimal efficacy for UC treatment. The limitations are mainly reflected in: (1) Polymeric coatings usually present a trade-off: while a higher crosslinking degree effectively blocks H⁺ diffusion, it simultaneously restricts nutrient exchange and delays probiotic release in the colon [22]. (2) Typical probiotic coatings (e.g., chitosan) usually provide a singular protective function, necessitating co-administration of anti-inflammatory drugs and complicating the fabrication process. Thus, a simple yet multifunctional coating strategy is highly desirable. Ideally, the probiotic coating should concurrently resist gastric acid and regulate inflammation, while retaining synthetic simplicity. However, to date, few coating materials have been reported to simultaneously integrate all these desirable functions without compromising probiotic viability. Pleasingly, we found that ferulic acid (FA), a naturally occurring lignin precursor, could be easily polymerized via a mild one-step self-polymerization strategy adopted by other groups and our group as well [23–25]. The polymerized-FA nanoparticles (pFA NPs) are suitable for oral delivery because they keep stable under acidic conditions while dissociating in alkaline environments. More importantly, the pFA coating embodies a “coating-as-drug” design paradigm, integrating gastric protection, colon-targeted release, intrinsic antioxidant activity, and mucus adhesion into a single material, eliminating the need for additional drug loading [26, 27]. Given these characteristics, we proposed that the pFA NPs represent a promising coating candidate for both shielding probiotics from harsh GI tracts and relieving intestinal symptoms. Besides, an overlooked but critical barrier to probiotic efficacy in UC is the excessive accumulation of hydrogen sulfide (H2S) at lesion sites. In the intestinal lumen, H2S produced by sulfate-reducing bacteria, disrupts the disulfide bonds of mucin glycoproteins within the mucus layer [28]. The degradation would reduce mucus viscosity while heightening permeability, which impairs probiotic colonization and hinders mucosal repair in UC [29]. Hence, incorporating an adjuvant that is capable of eliminating H2S to restore a healthy intestinal environment would greatly enhance the probiotic efficacy. To fulfill the aforementioned requirements, we first developed the pFA outfit using a self-polymerization approach to encapsulate Lactobacillus plantarum (Lac). Next, we screened various metal ions for their H₂S scavenging capacity through the formation of insoluble sulfide precipitation. Particularly, zinc ion (Zn2+), which could be incorporated into the pFA outfits by metal-phenol complexation, was finally selected as the optimal adjuvant due to its dual functions in preserving Lac viability and promoting tight junction repair [30]. The resultant formulation, termed as Lac@FZ (Fig. 1), outperformed in the following aspects: (1) pFA outlayers provided robust GI protection and significantly augmented probiotic survival rates when being exposed to gastric acid, bile salts, and reactive oxygen species (ROS) like H2O2. (2) The outfits improved intestinal retention of Lac and promoted tissue repair through phenolic hydroxyl structures derived from pFA. (3) Excessive H₂S at the lesion site was consumed by Zn2+ via sulfide salt formation, enabling Lac@FZ to effectively regulate sulfide homeostasis and concomitantly recover intestinal barrier integrity. Fig. 1Scheme for the fabrication of Lac@FZ and therapeutic mechanisms for UC treatment. (a) Preparation of pFA NPs and Lac@FZ via a self-polymerization strategy. (b) Therapeutic mechanisms of Lac@FZ in a mouse model with dextran sulfate sodium (DSS)-induced UC Scheme for the fabrication of Lac@FZ and therapeutic mechanisms for UC treatment. (a) Preparation of pFA NPs and Lac@FZ via a self-polymerization strategy. (b) Therapeutic mechanisms of Lac@FZ in a mouse model with dextran sulfate sodium (DSS)-induced UC Results and discussion Preparation of pFA and the underlying mechanism pFA was fabricated via the self-polymerization from free FA using sodium periodate (NaIO4), and the solution underwent a color change from colorless to red during the reaction process (Figure S1). Monodisperse spheres with an average diameter of 66.2 ± 3 nm gradually formed within 30 min, as tracked by dynamic light scattering (DLS) and transmission electron microscopy (TEM) (Fig. 2a, b). To explore the polymerization mechanism of FA, we conducted in situ Fourier transform infrared spectroscopy (FTIR) measurement. It demonstrated that FA polymerization intensified peaks between 1500 cm⁻¹ and 1700 cm⁻¹, reflecting enhanced conjugation of C = C bonds (benzene rings) and C = O bonds (ester moieties) (Fig. 2c, d). Additionally, peak shifts in the 1390–1430 cm⁻¹ range confirmed perturbations in symmetrical stretching vibration of carboxylic acid (-COO⁻) caused by polymerization (Fig. 2e, f). UV–vis spectra analysis also confirmed the dynamic structural changes associated with mono-catechol derivatives, exhibiting variable absorption with shoulder peaks at 282 nm and 314 nm, as well as a pronounced peak at 213 nm (Figure S2). Notably, the spectrum of pFA at the final reaction time point showed a decrease in absorption intensity compared with free FA, which might be attributed to the quenching effect induced by π-π interactions or electronic coupling within the formed NPs [31]. Zeta potential analysis revealed that pFA possessed a negative surface charge (− 20 to − 50 mV) in the pH range of 1–3. A high absolute zeta potential value (typically > ± 30 mV) is widely recognized to indicate strong electrostatic repulsion between nanoparticles, which prevents aggregation and confers good colloidal stability [32]. Under stronger acidic conditions, the zeta potential rose to around zero, possibly due to protonation of carboxyl and phenolic groups, whereas it sharply decreased in alkaline media (Fig. 2g). This pH-dependent zeta potential transition prompted us to investigate the pFA stability under different pHs. As validated by TEM and DLS in Fig. 2h, pFA NPs degraded into fragments when dispersed in buffer at pH 7.8 but kept stable in buffer at pH 2.0. Such pH-responsiveness would likely ensure pFA NPs from premature degradation in the strongly acidic gastric environment (pH 0.9 − 3.5), thereby enhancing delivery efficiency to the target intestinal site. Inspired by this observation, we further investigated the degradation kinetics of pFA under various simulated digestive fluids. Very little FA (4.79%) was released in the simulated gastric fluid (SGF, pH 2.0) even after incubation for 12 h. Similarly, a slow-release pattern was observed in the simulated intestinal fluid (SIF, pH 6.0). In sharp contrast, a rapid release behavior occurred in the simulated colonic fluid (SCF, pH 7.8) (Fig. 2i, j). Notably, the continuous GI digestion simulation over time further revealed the alkali-responsive degradation behavior of pFA NPs (Fig. 2k). Additionally, pFA NPs could exert potent radical-scavenging capacity along with their degradation in alkaline environment. As depicted in Fig. 2l, the total antioxidant capacity (T-AOC) under alkaline condition (pH 7.8) exhibited a 3.74-fold enhancement compared to that under acidic condition (pH 2.0), suggesting targeted delivery could potentially improve the efficacy for colitis therapy. Collectively, the self-polymerization of FA successfully yielded pFA NPs with alkaline-responsive degradation and radical-scavenging properties. Fig. 2Characterizations of pFA NPs. (a) Changes of the hydrodynamic size during the polymerization of FA into pFA. (b) Representative TEM image of pFA NPs. Scale bar = 200 nm. (c) In situ FTIR spectra (1450 cm− 1 ~ 1750 cm− 1) detecting the polymerization of FA into pFA. (d) In situ detection of peak changes in FTIR (1521, 1540, 1559, 1652, 1700, and 1775 cm− 1) weighted by relevant heights. (e) In situ FTIR spectra (1390 cm− 1 ~ 1430 cm− 1) detecting the polymerization of FA into pFA. (f) In situ detection of peak changes in FTIR (1402, 1410, and 1417 cm− 1) weighted by relevant heights. (g) Zeta potentials of pFA NPs incubated in buffers at different pHs. (h) Hydrodynamic size distributions, simulated images (inset), and representative TEM images (inset) of pFA NPs incubated in PBS (pH 2.0 or 7.8). Scale bars = 200 nm. (i) Released ratios of FA from pFA incubated in SGF (pH 2.0), SIF (pH 6.0), or SCF (pH 7.8). (j) UV-vis absorption spectra detecting the release of FA from pFA incubated in SCF (pH 7.8). (k) The relative cumulative release ratio of FA from pFA in simulated digestion tests. SSF refers to the simulated saliva fluid. (l) The relative total antioxidant capacity (compared with Fe2+) of pFA at different pHs (n = 3). ****p < 0.0001, statistical analysis of (l) was calculated via Student’s t-test Characterizations of pFA NPs. (a) Changes of the hydrodynamic size during the polymerization of FA into pFA. (b) Representative TEM image of pFA NPs. Scale bar = 200 nm. (c) In situ FTIR spectra (1450 cm− 1 ~ 1750 cm− 1) detecting the polymerization of FA into pFA. (d) In situ detection of peak changes in FTIR (1521, 1540, 1559, 1652, 1700, and 1775 cm− 1) weighted by relevant heights. (e) In situ FTIR spectra (1390 cm− 1 ~ 1430 cm− 1) detecting the polymerization of FA into pFA. (f) In situ detection of peak changes in FTIR (1402, 1410, and 1417 cm− 1) weighted by relevant heights. (g) Zeta potentials of pFA NPs incubated in buffers at different pHs. (h) Hydrodynamic size distributions, simulated images (inset), and representative TEM images (inset) of pFA NPs incubated in PBS (pH 2.0 or 7.8). Scale bars = 200 nm. (i) Released ratios of FA from pFA incubated in SGF (pH 2.0), SIF (pH 6.0), or SCF (pH 7.8). (j) UV-vis absorption spectra detecting the release of FA from pFA incubated in SCF (pH 7.8). (k) The relative cumulative release ratio of FA from pFA in simulated digestion tests. SSF refers to the simulated saliva fluid. (l) The relative total antioxidant capacity (compared with Fe2+) of pFA at different pHs (n = 3). ****p < 0.0001, statistical analysis of (l) was calculated via Student’s t-test Design and characterization of Lac@FZ By leveraging the unique merits of pFA NPs, we engineered the probiotic Lac with the pFA outfit, designated Lac@pFA, to improve its gastric survival and enable targeted delivery to inflamed colon sites. However, excessive H2S at these sites degrades the mucus layer by breaking mucin disulfide bonds, reducing viscosity but increasing permeability [29, 33, 34]. Considering that the disrupted mucus layer compromises probiotic colonization and impedes UC recovery, we envisioned that introducing an H2S scavenger would restore this barrier and boost the probiotics’ efficacy. Metal ions (Cu2+, Zn2+, Mn2+, Fe2+, Ga3+, Bi3+) were selected for their dual abilities to coordinate with phenolics and eliminate H2S by formation of precipitated sulfide salts (Fig. 3a) [35, 36]. To begin with, their influence on Lac viability was evaluated. The growth trajectories of probiotics were assessed by monitoring the optical density at 600 nm (OD600). As shown in Fig. 3b, c, significant differences were found among these tested metal ions. While Cu2+ and Bi3+ exhibited strong toxicity to Lac, Fe2+, Zn2+, and Mn2+ showed relatively mild influence. Nevertheless, delivery of redox-active transition metals (i.e., Fe2+/Mn2+) to UC-affected regions with elevated oxidative stress may induce Fenton/Fenton-like reactions, potentially exacerbating ROS-mediated damage. Therefore, Zn2+ was ultimately chosen and incorporated as an H₂S scavenger into the pFA-coated probiotics (termed Lac@FZ) for sustained UC management. As expected, both zeta potential changes and UV-vis spectroscopic analysis confirmed the successful preparation of Lac@FZ (Fig. 3d, e). SEM and TEM images revealed a continuous particle-like coating on the surface of Lac@FZ, in contrast to the smooth surface of bare Lac (Fig. 3f). Furthermore, we labeled the formulation with Rhodamine B (Rho). Flow cytometry histograms demonstrated significantly higher fluorescence intensity in Lac@FZ compared to bare Lac (Fig. 3g), which was corroborated with the results of complementary biomolecular scanning and fluorescence imaging (Figure S3 and S4). Subsequently, the bacterial viability of decorated-Lac was studied. After incubation for 10 h, both Lac@pFA and Lac@FZ showed comparable OD600 values to that of parental Lac, implying that surface coating would not affect its growth (Fig. 3h). The CCK-8 assay further validated the coating showed negligible impact on the probiotic growth (Fig. 3i). Given the intrinsic adhesiveness of polyphenols (i.e., pFA in this study), we then investigated whether the motility of coated Lac would be restricted by recording their mobility trajectories. Interestingly, the motility of Lac and Lac@FZ was found to be similar via the confocal laser scanning microscope (CLSM) dynamic tracking, with slight variations in accumulated path lengths and effective diffusivity (Deff) (Fig. 3j, k and Figure S5). Moreover, mean square displacement (MSD) versus time interval (Δt) profiles, derived from 2D trajectory coordinates (x, y), displayed a linear relationship indicative of unhindered Brownian diffusion (Figure S6). Collectively, these findings confirmed that the FZ outfits would not impair probiotic viability and motility. Next, energy-dispersive X-ray spectroscopy (EDS) and XPS analysis were employed to confirm that Zn2+ was successfully loaded onto Lac (Fig. 3l, m and Figure S7). Inductively coupled plasma-mass spectrometry (ICP-MS) quantified the Zn2+ mass fraction as 20.3% in Lac@FZ (Figure S8). Similarly, Lac@FZ displayed a pH-responsive release behavior analogous to that of pFA NPs: it remained stable under acidic conditions (SGF), whereas sustained release of Zn2+ and FA was observed in SCF (Figure S9). Moreover, the released Zn2+ was evidenced to effectively scavenge H2S: fluorescence detection with Washington State probe-5 (WSP-5) showed a 57.9% scavenging efficiency toward 400 µM NaHS (Fig. 3n and Figure S10), which was further validated and quantified by the gold-standard methylene blue colorimetric assay (73.4 ± 0.2% scavenging efficiency within 30 min, Figure S11a). Next, XPS analysis was conducted to verify the chemical reaction between Zn2+ and H2S. Characteristic Zn 2p and S 2p peaks were clearly detected in the precipitate collected from the Lac@FZ and NaHS incubation mixture, with negligible corresponding signals in the supernatant, verifying the formation of insoluble ZnS precipitates (Figure S11b-c). To exert beneficial effects, probiotics must survive during oral administration and reach the intestines in sufficient quantities. However, the GI tract poses significant physiological challenges, particularly (1) extreme gastric acidity, and (2) intestinal bile salts that would disrupt bacterial membranes and induce death [37]. Then, we assessed whether the Lac@FZ could enhance Lac’s resilience. When exposed to SGF, pristine Lac underwent a rapid viability decline, with near-complete inactivation within 30 min. In contrast, Lac@FZ exhibited significantly enhanced acid resistance, even surpassing the commercial enteric coating EUDRAGIT® L100 (Fig. 3o). The improved protection likely stems from the excellent ROS-scavenging and cytoprotective ability of the pFA outlayer. Similarly, Lac@FZ also maintained robust viability under bile salts exposure, showing similar protective efficacy to that of Lac@L100 after 4-hour incubation (Fig. 3p, q). Fig. 3Characterizations of Lac@FZ. (a) The digital photo depicting various metal sulfide precipitates. (b) Growth curves of Lac in the MRS medium containing different metal ions at 37 °C monitored by OD600 values, and (c) OD600 values after incubation for 9 h. (d) Zeta potentials of different formulations (n = 3). (e) UV-vis absorption spectra of different formulations. (f) Representative TEM and SEM images of Lac and Lac@FZ. The surfaces of Lac and Lac@FZ were magnified and highlighted. Scale bar = 1 μm. (g) Flow cytometry detecting the fluorescence intensities of Lac and Lac-Rho@FZ. (h) The growth curves of Lac, Lac@pFA, and Lac@FZ in the MRS medium at 37 °C monitored by OD600 at 1 h intervals (n = 3). (i) Bacterial viabilities of Lac, Lac@pFA, and Lac@FZ evaluated by CCK-8 at 0, 1, 2, and 4 h (n = 3). (j) Motion trajectories of Lac and Lac@FZ monitored by CLSM, and (k) the accumulated path lengths within 15 s (n = 10). (l) Representative high-angle annular dark-field (HAADF) image and elemental mapping images of Lac@FZ. Scale bar = 1 μm. (m) XPS spectra of Zn2p in Lac@FZ. (n) Fluorescence intensity spectra of H2S (S2−) incubated with different formulations detected by WSP-5. (o) Counts of Lac or coated-Lac after incubation in SGF for different time (n = 3). (p) Counts of viable Lac or coated-Lac after incubation in bile salts for different time (n = 3). (q) The representative digital photos of viable Lac or coated-Lac colonies in MRS plates retrieved after SGF (0.5 h) or bile salts (4 h) treatments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, while non-significant results are indicated as ns. Statistical analysis of (k) was processed by Student’s t-test, and statistical analyses of (c), (d), (i), (o), and (p) were processed by one-way analysis of variance (ANOVA) Characterizations of Lac@FZ. (a) The digital photo depicting various metal sulfide precipitates. (b) Growth curves of Lac in the MRS medium containing different metal ions at 37 °C monitored by OD600 values, and (c) OD600 values after incubation for 9 h. (d) Zeta potentials of different formulations (n = 3). (e) UV-vis absorption spectra of different formulations. (f) Representative TEM and SEM images of Lac and Lac@FZ. The surfaces of Lac and Lac@FZ were magnified and highlighted. Scale bar = 1 μm. (g) Flow cytometry detecting the fluorescence intensities of Lac and Lac-Rho@FZ. (h) The growth curves of Lac, Lac@pFA, and Lac@FZ in the MRS medium at 37 °C monitored by OD600 at 1 h intervals (n = 3). (i) Bacterial viabilities of Lac, Lac@pFA, and Lac@FZ evaluated by CCK-8 at 0, 1, 2, and 4 h (n = 3). (j) Motion trajectories of Lac and Lac@FZ monitored by CLSM, and (k) the accumulated path lengths within 15 s (n = 10). (l) Representative high-angle annular dark-field (HAADF) image and elemental mapping images of Lac@FZ. Scale bar = 1 μm. (m) XPS spectra of Zn2p in Lac@FZ. (n) Fluorescence intensity spectra of H2S (S2−) incubated with different formulations detected by WSP-5. (o) Counts of Lac or coated-Lac after incubation in SGF for different time (n = 3). (p) Counts of viable Lac or coated-Lac after incubation in bile salts for different time (n = 3). (q) The representative digital photos of viable Lac or coated-Lac colonies in MRS plates retrieved after SGF (0.5 h) or bile salts (4 h) treatments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, while non-significant results are indicated as ns. Statistical analysis of (k) was processed by Student’s t-test, and statistical analyses of (c), (d), (i), (o), and (p) were processed by one-way analysis of variance (ANOVA) Cytoprotective and re-polarization effects of Lac@FZ Before evaluating the therapeutic efficacy of Lac@FZ, we first assessed its cytotoxicity on human colon adenocarcinoma cell lines (HT-29) and murine leukemia monocyte-macrophage cell lines (RAW264.7). The cell viability exceeded 90% in both cell lines up to 160 µg mL− 1, confirming its good biosafety (Figure S12). Given that the UC microenvironment is characterized by elevated oxidative stress, we next investigated the ROS-scavenging and cytoprotective abilities of Lac@FZ. HT-29 cells were treated with gradient concentrations of H2O2, and 0.75 mM H2O2 was chosen to establish the oxidative stress model, as this dosage demonstrated balanced effects in both inducing intracellular oxidative stress while keeping a controlled cytotoxicity (Fig. 4a). Notably, DCFH-DA-based fluorescence imaging revealed that Lac alone induced a modest disruption of ROS homeostasis (Fig. 4b). This may stem from the intrusion of Lac altering the original redox microenvironment of cells, causing a mild elevation of ROS. Conversely, the probiotics with outfit decorations (i.e., Lac@pFA and Lac@FZ) were capable of neutralizing this perturbation and mitigating oxidative stress. These results were also supported by corresponding flow cytometry analysis (Fig. 4c and Figure S13). Furthermore, treatments with probiotic platforms (Lac@pFA or Lac@FZ) significantly attenuated H2O2-induced cellular damage, attributed to the cytoprotective properties of the pFA shell (Fig. 4d). Given that ROS are critical pro-inflammatory mediators driving cytokine amplification, we then evaluated the immunomodulatory capacity of Lac@FZ using LPS-stimulated RAW264.7 cells as an M1-polarized inflammatory model. Flow cytometry revealed that Lac@FZ significantly repolarized macrophages from M1- to M2-phenotype—an effect primarily mediated by pFA, as evidenced by reductions in the M1/M2 ratio to 0.61 (pFA) and 0.40 (Lac@FZ), respectively (Fig. 4e, f). In contrast, free Lac exerted a negligible effect on macrophage polarization. The discrepancy with the aforementioned ROS perturbation observed in HT-29 cells was likely due to the differential cellular sensitivity to probiotic stimuli. Furthermore, enzyme-linked immunosorbent assay (ELISA) identified that the intervention with Lac@FZ decreased the secretion of M1-phenotype marker tumor necrosis factor-α (TNF-α). Meanwhile, the levels of two representative anti-inflammatory cytokines transforming growth factor-β1 (TGF-β1) and interleukin-10 (IL-10), almost returned to normal (Fig. 4g-i). Lac@FZ also suppressed the liberation of high levels of NO which elicit damage and inflammation (Fig. 4j). Finally, we investigated the cellular H2S removal ability of Lac@FZ by staining RAW264.7 cells with the fluorescent probe WSP-5. The fluorescence signal sharply decreased from 90.9% to 52.7% in the groups where cells were co-incubated with NaHS and Lac@FZ, indicating successful removal of cellular H2S (Fig. 4k, l). This finding was further validated by WSP-5-based fluorescence imaging, with the group treated with Lac@FZ showing dimmed green fluorescence (Fig. 4m). Taken together, the pFA layer outside the probiotics showed potent ROS-scavenging and immunomodulatory capabilities. And the incorporation of Zn2+ into the pFA layer further facilitated the formation of ZnS, enabling efficient H2S clearance. Fig. 4ROS-scavenging and cytoprotective capabilities of Lac@FZ. (a) Cell viability of HT-29 cells treated with different concentrations of H2O2 (n = 3). (b) Representative DCFH-DA fluorescence images of HT-29 cells after different treatments. Scale bar = 50 μm. (c) Intracellular ROS levels in HT-29 cells treated with different formulations detected by flow cytometry. (d) Live/dead cell staining assay of HT-29 cells treated with different formulations for 24 h. Scale bar = 100 μm. (e) Populations of M1- (CD86+/CD206−) and M2-phenotype (CD86−/CD206+) RAW264.7 cells analyzed by flow cytometry after different treatments. (f) Quantitative analyses of flow cytometry for populations of M1- and M2-phenotype RAW264.7 cells (n = 3). (g-i) Protein secretion levels of the pro-inflammatory cytokines TNF-α, and anti-inflammatory cytokines including TGF-β1 and IL-10 after different treatments. (n = 3). (j) NO released by RAW264.7 cells after different treatments detected by Griess assays (n = 3). (k) Flow cytometry results of the cellular H2S levels incubated with different formulations detected by the probe WSP-5, (l) quantitative analysis (n = 3), and (m) representative fluorescence images stained by WSP-5. Scale bar = 50 μm. (n) Experimental group assignments for Fig. 4. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, while non-significant results were indicated as ns. Statistical analyses were processed by one-way ANOVA ROS-scavenging and cytoprotective capabilities of Lac@FZ. (a) Cell viability of HT-29 cells treated with different concentrations of H2O2 (n = 3). (b) Representative DCFH-DA fluorescence images of HT-29 cells after different treatments. Scale bar = 50 μm. (c) Intracellular ROS levels in HT-29 cells treated with different formulations detected by flow cytometry. (d) Live/dead cell staining assay of HT-29 cells treated with different formulations for 24 h. Scale bar = 100 μm. (e) Populations of M1- (CD86+/CD206−) and M2-phenotype (CD86−/CD206+) RAW264.7 cells analyzed by flow cytometry after different treatments. (f) Quantitative analyses of flow cytometry for populations of M1- and M2-phenotype RAW264.7 cells (n = 3). (g-i) Protein secretion levels of the pro-inflammatory cytokines TNF-α, and anti-inflammatory cytokines including TGF-β1 and IL-10 after different treatments. (n = 3). (j) NO released by RAW264.7 cells after different treatments detected by Griess assays (n = 3). (k) Flow cytometry results of the cellular H2S levels incubated with different formulations detected by the probe WSP-5, (l) quantitative analysis (n = 3), and (m) representative fluorescence images stained by WSP-5. Scale bar = 50 μm. (n) Experimental group assignments for Fig. 4. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, while non-significant results were indicated as ns. Statistical analyses were processed by one-way ANOVA Mucosal-adhesive property of Lac@FZ Efficient gut colonization of probiotics is essential for exerting therapeutic efficacy. Considering the outstanding adhesive capacity of polyphenols, we hypothesized that Lac@FZ could achieve robust adhesion via interactions between FA-derived phenolic hydroxyl/carboxyl groups and the mucus layer. To confirm this, we initially examined the in vitro adhesive property of Lac@FZ to two characteristic proteins, transferrin and mucin, which served as mimics of inflamed and healthy intestinal epithelium, respectively [38]. As depicted in Fig. 5a, confocal dishes were pre-coated with these two proteins, then incubated with mCherry-engineered probiotics for 1 h, and the adhered probiotics were finally quantified by CLSM. As for mucin, we found that Lac@FZ group displayed around 1.9-fold enhancement of adhered probiotics than the bare Lac one (Fig. 5b, c), indicating that the hydrogen bonds and Van der Waals interactions between the pFA layer and mucin aided adhesion. Similarly, the Lac@FZ group exhibited significantly higher adhesion on transferrin-coated dishes, further supporting the role of the FZ coating in promoting bacterial retention. Next, equal amounts of mCherry-engineered Lac or Lac@FZ were orally delivered to colitic mice, followed by the whole-body imaging using an in vivo imaging system (IVIS). As depicted in Fig. 5d-f, mice that received Lac@FZ exhibited significantly stronger abdominal fluorescence signals than those in the Lac group within 48 h post-administration. After 48 h, the mice were euthanized for ex vivo imaging. Consistent with in vivo observations, the GI tracts of mice treated with Lac@FZ exhibited robust mCherry signals, with fluorescence intensities 1.8-fold higher across the GI tract and 3.3-fold higher specifically in the colon compared to the unmodified Lac (Fig. 5g-i). By contrast, fluorescence signals in both groups were barely detectable in other organs, with only faint signals observed in the livers—presumably due to the metabolic elimination over 2 days (Figure S14). To evaluate oral bacterial survival, intestinal contents (from the stomach, small intestine, cecum, and colon) were collected and quantified via erythromycin-resistant colony counting. As shown in Fig. 5j, k, bacterial viability in the GI tracts was significantly enhanced in the Lac@FZ-treated group compared to controls, and the protective effect was particularly pronounced in the cecum and colon regions. These results collectively demonstrated that Lac@FZ not only shielded probiotics from harsh GI conditions but also enhanced their adhesion to inflammatory lesions. Fig. 5Mucosal-adhesive property of Lac@FZ. (a) Schematic illustration of the adhesion test performed on mucin- or transferrin-coated confocal dishes. (b) Representative CLSM images of mCherry-engineered probiotics adhered to bare dishes (Blank), mucin or transferrin. Scale bar = 10 μm. (c) Bacterial counts from the protein adhesion tests (n = 3). (d) In vivo fluorescence images of mice treated with Lac or Lac@FZ for 48 h, and (e-f) quantitative analyses of fluorescence intensities at different time points calculated by IVIS (n = 3). (g) Ex vivo fluorescence images of mouse intestinal tracts in different groups, and (h-i) quantitative analyses of fluorescence intensities across the entire GI tract or specifically in the colon (n = 3). (j) Representative digital photos of viable Lac or Lac@FZ colonies on erythromycin-supplemented agar plates retrieved from the mouse colon 48 h post-oral gavage, and (k) statistics of the colonies (n = 3). Groups are assigned to be: (1) stomach, (2) small intestine, (3) cecum, and (4) colon. *p < 0.05, **p < 0.01. Statistical analyses were processed by Student’s t-test Mucosal-adhesive property of Lac@FZ. (a) Schematic illustration of the adhesion test performed on mucin- or transferrin-coated confocal dishes. (b) Representative CLSM images of mCherry-engineered probiotics adhered to bare dishes (Blank), mucin or transferrin. Scale bar = 10 μm. (c) Bacterial counts from the protein adhesion tests (n = 3). (d) In vivo fluorescence images of mice treated with Lac or Lac@FZ for 48 h, and (e-f) quantitative analyses of fluorescence intensities at different time points calculated by IVIS (n = 3). (g) Ex vivo fluorescence images of mouse intestinal tracts in different groups, and (h-i) quantitative analyses of fluorescence intensities across the entire GI tract or specifically in the colon (n = 3). (j) Representative digital photos of viable Lac or Lac@FZ colonies on erythromycin-supplemented agar plates retrieved from the mouse colon 48 h post-oral gavage, and (k) statistics of the colonies (n = 3). Groups are assigned to be: (1) stomach, (2) small intestine, (3) cecum, and (4) colon. *p < 0.05, **p < 0.01. Statistical analyses were processed by Student’s t-test Therapeutic efficacy of Lac@FZ in mice model with DSS-induced UC Having confirmed the enhanced survival and muco-adhesion properties of Lac@FZ under pathological conditions, we further evaluated its therapeutic potential for ameliorating UC. A mouse model of UC was established by administering drinking water with 3% DSS for 7 days (Fig. 6a). Subsequently, DSS-treated mice were randomly divided into five groups and administered (2) water, (3) pFA, (4) Lac, (5) Lac@pFA, or (6) Lac@FZ four times at pre-determined time. Healthy mice (1) served as the untreated control (Fig. 6b). As depicted in Fig. 6c, d, the model group exhibited the most severe weight loss (90.4 ± 1.1% of baseline) and highest disease activity index (DAI) (7.6 ± 0.9) by day 14. The pristine probiotics group displayed no noticeable therapeutic effect, likely due to the degradation by gastric acid during oral administration. By contrast, Lac@pFA-treated mice demonstrated markedly improved outcomes compared to both the Lac and pFA groups. Notably, Lac@FZ treatment further enhanced the therapeutic efficacy, where Zn2+ served as an effective H₂S scavenger, enabling complete weight recovery (104 ± 1.3% of baseline) by day 14. These findings were corroborated with colon length and spleen index measurements (Fig. 6e, f and Figure S15). Next, we detected the concentrations of pro- and anti-inflammatory cytokines expressed in colons. As expected, the secretion of the proinflammatory cytokines myeloperoxidase (MPO) and interferon-γ (IFN-γ) were reduced while the secretion of the anti-inflammatory cytokine IL-10 increased after Lac@FZ treatment (Fig. 6g and Figure S16), suggesting the potent inflammatory modulation capacity of the decorated probiotics. In DSS-induced UC models, colonic inflammation is characterized by inflammatory cell infiltration and epithelial damage, which could be evaluated by hematoxylin and eosin (H&E) staining. As depicted in Fig. 6h, 3% DSS treatment severely disrupted colonic epithelial integrity and promoted infiltration of inflammatory cells (indicated by black arrows), along with crypt atrophy (blue dashed circles) and mucosal edema (red lines). Notably, the Lac@FZ group demonstrated the most effective restoration of epithelial integrity, as exemplified by the minimal H&E score of 2.33 (Fig. 6i). Reduced infiltration of inflammatory cells (black arrows) was also observed in H&E staining of spleens (Figure S17). To further assess the muco-protective effects of Lac@FZ, we performed Alcian blue/periodic acid-Schiff (AB/PAS) staining. The results revealed that Lac@FZ significantly repaired DSS-induced mucus layer damage, restoring both structural integrity and goblet cell density (Fig. 6j). Compared to the model group, Lac@FZ increased goblet cell numbers by 11.2-fold (Fig. 6k). Next, to avoid the influence of mucus loss due to the fecal excretion, we further analyzed colon samples with feces via AB-PAS staining. It was found that the mucus layer in the model group exhibited disorganized fibers, fragments, and reduced thickness (Fig. 6j). In contrast, the mice administered with Lac@FZ maintained the mucus layer thickness comparable to the healthy group and outperformed the Lac@pFA group. We speculated that these beneficial effects arose from Zn2+-mediated H₂S scavenging, as being supported by the significant reduction in colonic H₂S level (Fig. 6m), ultimately counteracting H₂S-induced epithelial barrier disruption and promoting mucosal recovery. The mucosal barrier repair was also validated by an augmentation in the expression level of occludin, a tight junction protein critical for intestinal barrier function (Fig. 6l and Figure S18). Taken together, our findings demonstrated that the administration of Lac@FZ significantly augmented the therapeutic efficacy of probiotics in DSS-induced UC model. Fig. 6Therapeutic efficacy of Lac@FZ in mice model with DSS-induced UC. (a) Schematic representation of the experimental procedure for DSS-induced UC treatment, and (b) experimental group assignments for Fig. 6. (c) Percentage changes in body weights of mice during the treatment period (n = 5). (d) DAI scores of mice in different groups (n = 5). (e) Quantified colon lengths, and (f) spleen indexes (spleen weight/body weight) of mice on day 14 (n = 5). (g) Heatmap illustrating the normalized ELISA results of colonic MPO, IFN-γ, and IL-10 after different treatments. Raw data are presented in Figure S16. (h) Representative H&E staining images of the colon tissues. Scale bar (upper) = 300 μm, scale bar (lower) = 100 μm. (i) Histology scores of colon tissues determined by H&E (n = 3). (j) Representative AB-PAS staining (with or without feces) images of the colon tissues. (k) Goblet cell abundance of the colons evaluated by AB-PAS (n = 3). (l) Representative occludin staining images of the colon tissues. (m) H2S levels in colonic feces treated with different formulations (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, while non-significant results are indicated as ns. Statistical analyses were processed by one-way ANOVA Therapeutic efficacy of Lac@FZ in mice model with DSS-induced UC. (a) Schematic representation of the experimental procedure for DSS-induced UC treatment, and (b) experimental group assignments for Fig. 6. (c) Percentage changes in body weights of mice during the treatment period (n = 5). (d) DAI scores of mice in different groups (n = 5). (e) Quantified colon lengths, and (f) spleen indexes (spleen weight/body weight) of mice on day 14 (n = 5). (g) Heatmap illustrating the normalized ELISA results of colonic MPO, IFN-γ, and IL-10 after different treatments. Raw data are presented in Figure S16. (h) Representative H&E staining images of the colon tissues. Scale bar (upper) = 300 μm, scale bar (lower) = 100 μm. (i) Histology scores of colon tissues determined by H&E (n = 3). (j) Representative AB-PAS staining (with or without feces) images of the colon tissues. (k) Goblet cell abundance of the colons evaluated by AB-PAS (n = 3). (l) Representative occludin staining images of the colon tissues. (m) H2S levels in colonic feces treated with different formulations (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, while non-significant results are indicated as ns. Statistical analyses were processed by one-way ANOVA Modulation of gut microbiota by Lac@FZ The gut microbiota plays a crucial role in maintaining intestinal barrier integrity, and its imbalance is closely linked to UC. We therefore proceeded to explore whether Lac@FZ could restore gut microbiota balance in colitic mice using 16 S ribosomal RNA gene sequencing. DSS administration (the model group) triggered rapid and profound alterations in the gut microbiota, manifesting as reduced observable microbial diversity (Fig. 7a) and a significant decline in α-diversity (Chao1 index, Fig. 7b). Strikingly, Lac@FZ treatment restored gut microbiota richness to levels comparable to those in healthy controls. Venn diagrams revealed higher species similarity between the Lac@FZ-treated group and the healthy group compared to the overlap between the model group and the healthy one (Fig. 7c). Consistent with this, non-metric multidimensional scaling (NMDS) revealed distinct clustering patterns among the treatment groups (Fig. 7d), indicating compositional reorganization of the microbiota. To be specific, at the phylum level the relative abundance of Firmicutes was markedly reduced in the model mice [39], along with an elevated relative abundance of Proteobacteria [40, 41], which are microbial signatures of dysbiosis in UC patients (Fig. 7e) [42]. Nevertheless, intervention with Lac@FZ not only reversed these perturbations by boosting Firmicutes and suppressing Desulfobacteria, but also selectively enriched Lactobacillaceae (Fig. 7f-h). The bacterial composition of UC mice after different treatments was further analyzed at the family level. The ternary plot demonstrated that Lac@FZ reduced the relative abundance of potentially harmful Proteobacteria, including the Enterobacteriaceae family which could thrive in an inflamed gut [43], as well as Desulfovibrionaceae family which impairs colon mucosa integrity (Fig. 7i) [44]. The Lactobacillaceae family, to which the Lac strain used in this study belongs, also had a considerable increase in relative abundance. These results were further confirmed by linear discriminant analysis (LDA) effect size (LEfSe) analysis (Fig. 7j). Taken together, these results suggest that Lac@FZ restored microbial diversity and abundance to healthy levels through the synergistic effects from pFA (as a protective armor) and Zn2+ (as an H₂S scavenger). Fig. 7Modulation of gut microbiota by Lac@FZ. (a) Numbers of observed features, and (b) Chao1 index of the gut microbiota in mice after different treatments (n = 5). (c) Venn diagram of the intestinal bacterial species detected in different groups (n = 5). (d) Assessment of β-diversity of the microbial community in different groups analyzed by NMDS (n = 5). (e) Relative abundance of gut microbiota at the phylum level, and relative abundance of selected taxa including (f) Desulfobacteria, (g) Firmicutes, and (h) Lactobacillaceae (n = 5). (i) Effects of Lac@FZ on gut microbiota at the family level using ternary plot method. (j) Distribution histogram based on LDA. LDA score higher than 4.0 indicates the significantly abundant taxa in the corresponding group than in other groups. *p < 0.05, **p < 0.01, while non-significant results are indicated as ns. Statistical analyses were processed by one-way ANOVA Modulation of gut microbiota by Lac@FZ. (a) Numbers of observed features, and (b) Chao1 index of the gut microbiota in mice after different treatments (n = 5). (c) Venn diagram of the intestinal bacterial species detected in different groups (n = 5). (d) Assessment of β-diversity of the microbial community in different groups analyzed by NMDS (n = 5). (e) Relative abundance of gut microbiota at the phylum level, and relative abundance of selected taxa including (f) Desulfobacteria, (g) Firmicutes, and (h) Lactobacillaceae (n = 5). (i) Effects of Lac@FZ on gut microbiota at the family level using ternary plot method. (j) Distribution histogram based on LDA. LDA score higher than 4.0 indicates the significantly abundant taxa in the corresponding group than in other groups. *p < 0.05, **p < 0.01, while non-significant results are indicated as ns. Statistical analyses were processed by one-way ANOVA Conclusion In this study, we have successfully developed an artificial polyphenol nanoparticle through a straightforward self-polymerization strategy. The obtained pFA is an ideal candidate for oral drug delivery since it is stable under acidic conditions while decomposing under basic conditions. Given this characteristic, we shielded Lac with pFA as a protective outfit against the harsh GI environment. The coating well preserved the viability of Lac until they reached the colon (pH > 7), where the pFA degraded into fragments, released Lac and exerted anti-oxidative and inflammation-modulating activities. Considering that excessive H2S produced at the inflamed sites could deplete the mucus layer and adversely affect the probiotic colonization, Zn2+ ions were introduced as an H2S scavenger after careful screening. Mechanistically, H₂S scavenging by Zn2+ relieves an upstream pathological brake, enabling pFA to exert its anti-inflammatory effects and allowing Lac to achieve durable colonization and microbiota modulation. As noted in the aforementioned studies, the designed Lac@FZ was able to shift the microbial composition profile from dysbiosis toward homeostasis, efficiently modulating the GI milieu by removing H2S, alleviating inflammation, and remodeling the intestinal microbiota in a synergistic manner. Compared with current treatments for UC, our research outperforms for its facile preparation and excellent therapeutic efficacy. Nonetheless, there are still several issues to be explored: First, while pFA was successfully designed and prepared as uniform nanoparticles, the specific polymerization mechanisms and structure of final product need to be further investigated in future studies. Second, research on the detection of H2S in the intestinal tract needs to be further deepened. The gaseous and chemically active nature of H2S poses difficulties in accurately monitoring its concentration in intestinal lumen by commonly employing H2S detection assays, such as fluorescence probes and the methylene blue colorimetric method. We consider that the in situ monitoring systems of H2S at the animal level should be improved, and more attention should be paid to the dynamic change of H2S concentrations at target sites. Nevertheless, the pFA obtained via the self-polymerization strategy has been proven to be an ideal acid-resistant oral drug adjuvant. Given that the monomeric phenol could be replaced with other analogs, we believe that the strategy developed in this study could be expanded to prepare other functional materials and useful platforms for different applications. Methods Preparation of pFA NPs and coated probiotics. 96 mg of FA was dissolved in 100 µL of DMSO. Then, 32 µL of FA solution was added into 40 mL of ultrapure water. Next, 4.24 mg of NaIO4 was added quickly, and the reaction mixture was vigorously stirred at room temperature to form pFA NPs. To prepare Lac@FZ, 5 × 108 colony-forming units (CFU) of Lac was suspended in a 10 mM sodium acetate buffer solution containing 30.72 mg of FA. 2.86 mg of zinc acetate was added to the mixture and stirred for 15 min in an ice bath. Afterwards, 4.24 mg of NaIO4 was added to form Lac@FZ. Products were collected by centrifugation at 4000 rpm for 5 min and purified by washing with sodium acetate buffer twice. Lac@pFA was prepared following the same protocol without the addition of Zn2+. Resistance of Lac@FZ against simulated GI environment. The protective effect of pFA on Lac against GI tract conditions, including SGF and bile salts (300 µg mL− 1) was evaluated. Briefly, 500 µL of Lac@L100, Lac@pFA, or Lac@FZ (OD = 0.5) was incubated with 500 µL of SGF or bile salts at 37 °C. At predetermined time points, the coated-probiotics were collected, washed twice with buffers, and plated on MRS agar. CFUs were counted after incubation for 24 h, and images were recorded. In vitro H2S scavenging ability of Lac@FZ. The H₂S scavenging activity of Lac@FZ was evaluated using both fluorescent and colorimetric methods. First, a NaHS solution (400 µM) was mixed with 200 µg mL− 1 of Lac@pFA or Lac@FZ at 37 °C for 30 min. Zinc acetate solution was set as the positive control. Next, WSP-5 probe (50 µM WSP-5 + 200 µM CTAB) was applied to detect H2S. Fluorescence spectra were recorded using a FL6500 fluorescence spectrometer (PerkinElmer, MA) with λex/λem = 480/525 nm. The scavenging efficiency was further quantified by the methylene blue colorimetric assay. Briefly, NaHS solution was incubated with zinc acetate, Lac@pFA, or Lac@FZ under identical conditions, followed by centrifugation (15,000 rpm, 10 min). The supernatant was collected and the residual H2S concentration was measured using a methylene blue detection kit. Additionally, both the supernatant and precipitate from the Lac@FZ group were separately collected for XPS analysis to confirm the chemical reaction between Zn2+ and H2S. For intracellular H2S detection, HT-29 cells were first incubated with 400 µM NaHS for 6 h, then treated with the WSP-5 probe and kept in the dark for 30 min. After washing twice with PBS, cells were imaged by CLSM (LSM900, ZEISS). Cytoprotective effect of Lac@FZ against ROS-induced damage. The cytoprotective effect of Lac@FZ against ROS-induced damage was evaluated on HT-29 cells using a live-dead cell staining assay and DCFH-DA ROS detection assay. HT-29 cells (except for the Control group) were seeded in confocal dishes and incubated with 750 µM H2O2 for 12 h. Afterwards, cells were incubated with different formulations: (2) Model (replaced with fresh cell culture medium), (3) pFA, (4) Lac, (5) Lac@pFA, (6) Lac@FZ (pFA: 8 µg mL− 1, Lac: 40 µg mL− 1 at a final concentration) for 12 h, followed by staining with DCFH-DA or Calcein-AM/PI in serum-free Dulbecco’s modified eagle medium (DMEM) for 30 min. Cells were then washed twice with PBS and imaged using the CLSM (LSM900, ZEISS). In vitro anti-inflammatory effect of Lac@FZ. RAW264.7 cells were seeded in 6-well plates and cultured overnight. Cells were then stimulated with 200 ng mL− 1 LPS for 12 h to induce M1-phenotype polarization (except the Control group). After LPS treatment, cells were incubated for 24 h with the following formulations: (2) Model (replaced with fresh cell culture medium), (3) pFA, (4) Lac, (5) Lac@pFA, (6) Lac@FZ (pFA: 8 µg mL− 1, Lac: 40 µg mL− 1 at a final concentration). The M1- and M2-phenotype macrophages were labeled with CD86 and CD206 antibodies, and analyzed by flow cytometry (LSRFortessa X-20, BD Biosciences). For measurements of anti-inflammatory and pro-inflammatory indicators, RAW264.7 cells were seeded into 24-well plates and exposed to 200 ng mL− 1 of LPS for 12 h. After treatment with the indicated formulations, nitrite levels in the culture medium, which are the reflective index of NO production, were quantified using the Griess assay. Similarly, the concentrations of IL-10, TNF-α, and TGF-β1 were measured using ELISA kits. Therapeutic effect of Lac@FZ against DSS-induced UC. First, female SPF C57BL/6 mice (20 g, 6–8 weeks) were acclimated for 7 days. The mice were then divided into six groups: (1) Control, (2) Model, (3) pFA, (4) Lac, (5) Lac@pFA, (6) Lac@FZ (pFA: 4 mg kg− 1, Lac: 1 × 108 CFU at a final concentration). The UC mouse model was established following a previously reported protocol [18]. Mice except for the Control group were treated with drinking water containing 3% DSS for 7 days. Then, different formulations were administered by gavage on day 7, 8, 10, and 12. The body weight and DAI of mice were recorded daily during the experiment. At day 14, mice were euthanized for measurements of colon lengths and spleen indexes. The intestinal tissues and major organs, including heart, liver, spleen, lung, and kidney, were collected and fixed in 4% paraformaldehyde for histological and ELISA analyses. In addition, mouse feces were collected for intestinal flora analysis and H2S level evaluation. Statistical analyses. The statistical analyses were conducted using GraphPad Prism 10. In the case of multiple group comparisons, one-way ANOVA was employed. In the case of two group comparisons, Student’s t-test was employed. All results are presented as the mean ± SD, as indicated. Statistical significance is denoted by *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, while non-significant results are indicated as ns. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1