Oral delivery of engineered Lactococcus lactis microcapsules expressing antioxidant enzymes for inflammatory bowel disease: Mechanisms of intestinal barrier repair and microbiota modulation

Abstract

Inflammatory bowel disease (IBD) is characterized by elevated levels of reactive oxygen species (ROS) and pro-inflammatory cytokines, alongside disrupted gut microbiota. Therefore, eliminating ROS in the inflammatory site by antioxidant enzymes such as catalase (CAT) and superoxide dismutase (SOD) represents a promising therapeutic strategy for IBD. To this end, we genetically engineered Lactococcus lactis to construct an antioxidant probiotic strain, LL-SC, by integrating a fusion gene (SC) encoding SOD and CAT. To further enhance gastrointestinal survival, LL-SC was encapsulated with a composite nanomaterial of mucosal-adhesive chitosan and sodium alginate to produce LL-SC-C₂A₂ with antioxidant enzyme expression capability and a nano-protective coating. Compared to uncoated LL-SC, LL-SC-C₂A₂ exhibited significantly improved gastric acid tolerance with 1.4-fold increases. Protective effects of LL-SC-C₂A₂ were confirmed across cellular and animal models, including H₂O₂-stimulated Caco-2 cells and a DSS-induced murine colitis model. This was achieved through ROS scavenging, pro-inflammatory cytokine reduction, intestinal barrier reinforcement, and restoration of gut microbiota homeostasis. Overall, food-grade LL-SC-C₂A₂ represents a novel approach to probiotic modification, providing a new strategy and experimental evidence supporting further development for IBD therapy.

Graphical abstract

1. Introduction

Inflammatory bowel disease (IBD) is a non-specific gastrointestinal disease characterized by intestinal inflammation, tissue damage, abdominal pain, frequent or persistent diarrhea, and rectal bleeding [1]. Although the mortality rate is relatively low, the disease has a high incidence rate, a high disability rate, and a trend toward younger patients and increases the risk of anal fissures [2], fistulae [3], colorectal cancer [4], and extra-intestinal manifestations [5]. The exact cause of IBD remains unclear, but it is generally believed to be the result of a complex interaction between multiple factors, such as genetic susceptibility, environmental triggers, intestinal dysbiosis, and subsequent aberrant mucosal immune responses [6]. Genome-wide association studies (GWAS) have identified over 240 genetic risk loci associated with IBD. Nevertheless, these genetic predispositions alone are insufficient for disease manifestation. Crucially, they require synergistic interaction with environmental factors, particularly those that induce or exacerbate dysbiosis [7]. A widely accepted framework for IBD pathogenesis posits that the disease arises from a dysregulated gut immune response, driven by the interplay of genetic and environmental factors. Within this framework, intestinal microbiota imbalance is considered a critical element [8,9]. Such dysbiosis can compromise the intestinal barrier, promote immune cell dysregulation, and ultimately lead to a persistent overproduction of pro-inflammatory cytokines that sustains chronic inflammation [10]. Despite the increasing variety of treatment options available, including traditional drugs [11], biological agents [12], natural antioxidants [13], and microbial modulation therapies [14], a substantial proportion of patients still experience primary non-response or secondary loss of response to treatment.

Oxidative stress-driven accumulation of excess reactive oxygen species (ROS) constitutes a central pathological driver in IBD, propagating a self-amplifying vicious cycle through impairment of intestinal barrier integrity, induction of microbial dysbiosis, and activation of inflammatory cascades [15,16]. As a pivotal endogenous antioxidant enzyme, superoxide dismutase (SOD) serves as the first-line antioxidant defense by catalyzing the dismutation of superoxide (O₂⁻) to hydrogen peroxide (H₂O₂). Catalase (CAT) subsequently detoxifies H₂O₂ into water and molecular oxygen [17,18]. This enzymatic cascade can abrogate ROS generation, positioning SOD and CAT as an ideal therapeutic strategy for IBD owing to their target specificity and minimal off-target effects. However, clinical translation of SOD and CAT faces a tripartite delivery challenge: gastrointestinal instability (susceptibility to acidic and enzymatic degradation), non-targeted distribution (poor accumulation at intestinal lesion sites), and rapid systemic clearance (insufficient residence time for therapeutic efficacy). These limitations have led to suboptimal outcomes in oral SOD and CAT clinical trials [19]. Therefore, there is an urgent need to develop advanced delivery systems that can simultaneously maintain enzyme activity and improve intestinal bioavailability.

Engineered probiotics (EPs) represent a promising platform technology uniquely positioned to address these formidable delivery challenges for therapeutic enzymes like SOD and CAT. EPs demonstrate considerable potential for chronic disease treatment through genetic modification enabling high-efficiency expression of exogenous bioactive molecules, combined with inherent gut tropism, food-grade genetic manipulation systems, and oral delivery compatibility [20]. Commonly utilized EPs predominantly include non-pathogenic Escherichia coli, Lactobacilli, and bifidobacteria, whose safety profiles have been extensively validated through long-term applications in the food industry [21]. Within the realm of IBD therapy, EP-mediated intervention strategies primarily encompass including delivery of antimicrobial peptides, delivery of anti-inflammatory cytokines, inactivation of pro-inflammatory cytokines, and delivery of antioxidant enzymes and bioactive compounds [22]. Zhou et al. [23] devised an advanced biocatalytic platform using engineered Escherichia coli Nissle 1917 (EcN-pE (C/A)₂) co-overexpressing SOD and CAT core antioxidant enzymes that synergistically neutralize ROS cascades. The inherent gut tropism of the EcN carrier promotes localization to the diseased site, while the continuous production of enzymes by the colonizing bacteria provides sustained therapeutic action, circumventing rapid clearance. This integrated approach demonstrated efficacy in ameliorating inflammation, repairing epithelial barriers across multiple chemically induced murine IBD models, and modulating gut microbiota by significantly elevating probiotic abundance in dysbiotic intestines. However, despite these promising results, the translational potential of this and similar platforms faces notable constraints, particularly regarding the microbial chassis selection. The use of species with conditional pathogenic associations, such as E. coli, presents challenges for regulatory approval and public acceptance.

Lactococcus lactis was selected as a therapeutic vector to overcome specific limitations of existing platforms. L. lactis MG1363, the first lactic acid bacterium species to undergo complete genome sequencing, offers extensive genetic resources and holds Generally Recognized as Safe (GRAS) status by the Food and Drug Administration (FDA) [24]. This microbial host demonstrates exceptional capacity for producing properly folded and functional heterologous proteins, combining a well-established history of safe use in food production with inherent probiotic properties [25]. The platform offers multiple technical advantages: (1) rapid growth to high cell density under anaerobic conditions, enabling scalable fermentation; (2) Gram-positive character that eliminates endotoxin concerns; (3) efficient secretion of stable recombinant proteins into the culture medium; and (4) versatility through multiple available expression vectors [26,27]. The Nisin-Controlled Gene Expression (NICE) system was employed to enable stable and high-yield protein expression in EPs. This system utilizes complementation genes as selection markers, ensuring that only L. lactis strains successfully incorporating the target gene could proliferate in selective medium containing lactose as the sole carbon source [28]. This innovative approach effectively eliminates the environmental dissemination risks associated with antibiotic selection markers in genetically modified organisms. The resulting integrated engineering strategy establishes a comprehensive therapeutic platform that combines food-grade genetic design with optimized production characteristics, positioning it as an advanced solution for IBD intervention.

Despite the promising therapeutic potential of engineered L. lactis, a critical challenge for its application as an oral live biotherapeutic product lies in the hostile gastrointestinal environment [29]. Therefore, developing effective protective strategies to enhance the persistence and stability of engineered bacteria during intestinal delivery is imperative. Within this context, encapsulation using biocompatible materials presents an attractive solution. Among these materials, polysaccharide-based biopolymers have garnered significant research interest owing to their natural origin, abundance, non-toxicity, exceptional biocompatibility, and favorable thermal and chemical stability. Sodium alginate (SA), an anionic polysaccharide derived from algae, exhibits inherent high porosity that facilitates rapid diffusion of aqueous solutions but concurrently risks premature leakage of encapsulated cargo [30]. Chitosan (CS), a cationic polysaccharide, possesses mucoadhesive properties and pH-responsive solubility [31]. Constructing a CS-SA polyelectrolyte complex through electrostatic interactions would synergistically (1) mitigate the porosity-driven leakage of SA by forming a denser composite membrane and (2) leverage the bioadhesion of CS to prolong intestinal retention, thereby enhancing the gastrointestinal resilience and targeted release of encapsulated engineered L. lactis. This composite colloidal structure maintains structural integrity under the acidic gastric environment, providing a physical barrier that buffers gastric acid and attenuates bile salt impact, thereby enhancing EPs survival rates in the upper gastrointestinal tract. Upon reaching the neutral or weakly alkaline intestinal milieu, the swelling or dissociation properties of the composite facilitate controlled release of the engineered bacteria at target sites [32]. Critically, both SA and CS possess dual GRAS status as food-grade materials, ensuring compliance with safety regulations for pharmaceutical and food applications.

In this study, we engineered an antioxidant probiotic strain by chromosomally integrating the SOD-CAT fusion gene into L. lactis. To improve gastrointestinal survival, the engineered strain was subsequently encapsulated within CS and SA through LBL self-assembly driven by electrostatic interactions. We hypothesized that the dual functional design combining enzymatic ROS scavenging from engineered probiotics and chitosan-sodium alginate nanoencapsulation would synergistically enhance oxidative stress modulation, leading to significantly improved therapeutic outcomes in IBD compared to unencapsulated counterparts. Specifically, treatment with LL-SC-C2A2 attenuated intestinal oxidative stress, suppressed inflammatory responses, promoted intestinal barrier repair, and modulated gut microbiota homeostasis. This integrated strategy combining genetic engineering and nano-encapsulation provides a promising approach for the clinical prevention and treatment of IBD.

2. Materials and methods

2.1. Materials and reagents

Phosphate buffered saline (PBS), Nisin, and a bicinchoninic acid (BCA) assay kit were obtained from Beijing Solarbio Science & Technology Co., Ltd. Glycol chitosan was purchased from Sigma Aldrich, St. Louis, USA. Sodium alginate was provided by Shanghai Aladdin Biochemical Technology Co., Ltd. The CAT, SOD, glutathione peroxidase (GSH-Px), malondialdehyde (MDA), and myeloperoxidase (MPO) assay kits were obtained from Nanjing Jiancheng Biotechnology Co., Ltd. The strains NZ3900 and pNZ8149 plasmids and the empty vector NZ3900-8149 were obtained from Hangzhou Baosai Biotechnology Co., Ltd. The CAT and SOD genes were obtained from the NCBI website and were optimized and synthesized by Suzhou Jinweizhi Biotechnology Co., Ltd. Female C57BL/6 mice (6-8 weeks) were purchased from Harbin Sino-Sines Biotechnology Co., Ltd., and housed in a specific pathogen-free barrier environment (temperature of 22 °C-25 °C, relative humidity of 40%-60%, and lighting of 12 h/12 h cyclic lighting). All animal experiments in this thesis were approved by the Animal Ethics Committee of Northeast Agricultural University (No. NEAUEC-20240435), and all participants were trained and strictly observed the ethical principles of animal experimentation.

2.2. Construction and validation of LL-SC

The pNZ8149-CAT-SOD expression plasmid was constructed by Hangzhou Baosai Biotechnology Co., Ltd., and prepared by adding CAT (GenBank: AB587573.1) and SOD (GenBank: WP_003131560.1) into the polylinker region of pNZ8149, with codon usage adapted to LL and a 6 × His tag fused at the C-terminus of the tandem sequence (Fig. 1 A). Primers for CAT-SOD amplification and vector linearization were designed using Primer 5.0 and synthesized by Shanghai Generay Biotech Co., Ltd. (Table S1). The tandem sequence was cloned into pNZ8149 via In-Fusion HD assembly using primers. Linearized pNZ8149 was generated by inverse PCR (primers 8149-CATSOD-F/R). Competent NZ3900 cells were electroporated (2000 V, 200 Ω, 25 μF) with the recombinant plasmid, resuscitated in SGM17 medium (30 °C, 3 h), and plated on Elliker agar. Recombinant L. lactis were validated by colony PCR and sequencing.

Fig. 1.

(A) Schematic of recombinant Lactococcus lactis (LL-SC) construction; (B) Construction and verification of linearized carriers: (a) Amplification of SOD-CAT fragment analyzed by 1% agarose gel electrophoresis. (b) Reverse PCR amplification of pNZ8149 plasmid analyzed by 1% agarose gel electrophoresis; (C) Identification of the LL-CS colonies by PCR. 1-10 are randomly selected clone numbers; (D) Expression in the sonicated supernatant of the bacterial solution under different conditions; (E) Growth curves of the three strains at 30 °C.

The validated antioxidant enzyme-expressing strain (LL-SC) was cultured in M17 broth, while the empty vector control (LL-pNZ8149) was grown in MRS medium at 30 °C until OD₆₀₀ reached 0.3–0.4. Cultures were induced with 0, 1, or 5 ng/mL Nisin and harvested at OD₆₀₀ when it reached 0.8. Bacterial cells (1.5 mL) were centrifuged and resuspended in 400 μL Tris-HCl buffer (10 mM, pH 7.0) and ultrasonicated (400 W, 9 s pulses with 15 s intervals for 3 min total) to break the cells. The lysate supernatant (40 μL) was mixed with 10 μL 5 × SDS loading buffer and subjected to Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE). Characterization by Western blotting with anti-His mouse monoclonal antibody. L. lactis strains (LL, LL-pNZ8149, and LL-SC) were inoculated into fresh M17 broth at 2% (v/v). For recombinant strains (LL-pNZ8149 and LL-SC), Nisin (10 ng/mL) was supplemented at OD₆₀₀ 0.3–0.4 to induce target gene expression. Bacterial growth was monitored hourly, and growth curves were plotted based on 24 h OD₆₀₀ measurements.

2.3. Preparation and characterization of LL-SC-C₂A₂

Preparation of LL-SC-C₂A₂ was performed according to a previously reported method [33]. Simply put, CS (2 mg/mL) and SA (2 mg/mL) were dissolved in 0.5 M Nacl solution, with the pH adjusted to 6.0 using HCl and NaOH. Prior to LBL assembly, LL-SC was washed three times with 0.5 M NaCl solution. Subsequently, the LL-SC were resuspended in chitosan solution for 30 min to obtain LL-SC-C₁. Following two additional washes with 0.5 M NaCl solution, the resulting product was redispersed in sodium alginate solution for 30 min to prepare LL-SC-C₁A₁. The coating process was repeated to sequentially generate LL-SC-C₂A₁ and LL-SC-C₂A₂. The zeta potential and particle size of the samples were simultaneously measured using a Zetasizer Nano ZS (Malvern Panalytical, UK). To evaluate the impact of LBL coating on bacterial viability and growth kinetics, encapsulated samples were inoculated at 2% (v/v) into M17 broth and cultured for 16 h. Growth curves were subsequently plotted based on the temporal OD₆₀₀ variations observed during the cultivation period (measured every 2 h).

The encapsulation efficacy of LBL assembly was evaluated by transmission electron microscopy (TEM, HT7800, Hitachi, Japan). Bacterial samples were fixed overnight at 4 °C with 2.5% (v/v) glutaraldehyde solution. After fixation, the specimens were rinsed with 0.1 M phosphate buffer (pH 7.2) and post-fixed with 1% (w/v) osmium tetroxide solution. Following thorough rinsing with phosphate buffer, the samples were sequentially dehydrated in graded ethanol solutions (50%, 70%, 90%, and 100%, v/v) for 10 min each. Followed by treatment with a 1:1 (v/v) ethanol-acetone mixture for 10 min and acetone for 5 min. The dehydrated samples were embedded, trimmed, and sectioned into ultrathin slices using an ultramicrotome. Finally, the sections were stained with uranyl acetate and lead citrate prior to TEM imaging. Bacteria following Nisin induction (N-LL-SC-C₂A₂, N-LL-SC, N-pNZ8149) were lysed via ultrasonication (200-600 W, 1-10 min) and centrifuged to collect the supernatant. SOD and CAT activities were quantified using assay kits.

2.4. Gastrointestinal resilience evaluation

Simulated gastric juice (SGJ) was formulated by combining 8.0 mL of 0.2% NaCl (w/v) and 1.0 mL of 0.7% KCl (w/v) with 4000 U of pepsin. The pH was adjusted to 3.0 using 1 M HCl or NaOH, filtered through a 0.22 μm sterile membrane, and immediately used. Simulated intestinal juice (SIJ) was prepared by dissolving 3.4 g KH₂PO₄ in 125 mL distilled water and 95 mL 0.1 M NaOH. The mixture was adjusted to pH 7.5 with 0.1 M NaOH, brought to a final volume of 500 mL with distilled water, autoclave-sterilized (121 °C, 15 min), and supplemented with 3 g bile salts. LL-SC and LL-SC-C₂A₂ (bacteria cultured to the third generation) were inoculated into SGJ (1:9 ratio) and shock incubated at 37 °C (150 rpm, 2 h), then transferred to SIJ for an additional 4 h. Viable cell counts were quantified at 0, 2, 4, and 6 h via the M17 agar pour plate method and expressed as log₁₀ CFU/g.

2.5. In vivo safety evaluation

Sixteen mice were randomly allocated into two groups (n = 8 per group) and orally administered either PBS or LL-SC-C₂A₂ on days 0, 2, 4, and 6. General health status, including fur condition, activity level, motor behavior, and fecal consistency, was monitored and scored daily throughout the experimental period (Table S2). On the tenth day, all mice were euthanized via eyeball removal and blood collection under anesthesia. Serum samples were analyzed for routine blood tests and hepatic function biomarkers, including white blood cell count (WBC), red blood cell count (RBC), hemoglobin (HGB), hematocrit (HCT), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), platelet count (PLT), alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and albumin (ALB). Major organs (liver, spleen, kidney, and colon) were harvested, fixed in 4% paraformaldehyde, and subjected to histopathological evaluation using hematoxylin and eosin (HE) staining.

2.6. In vitro validation of antioxidant functionality in Caco-2 cells

The human colon adenocarcinoma Caco-2 cell line was maintained in Modified Eagle Medium (MEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37 °C in a 5% CO₂ atmosphere. To establish an optimal oxidative stress model, cells were seeded in 96-well plates at a density of 1 × 10⁴ cells per well. After 24 h of adhesion, the cells were exposed to a range of H₂O₂ concentrations (0, 0.05, 0.1, 0.15, 0.2, and 0.25 mM) for 2, 4, and 6 h to determine the appropriate treatment condition. Cell viability was subsequently assessed using the Cell Counting Kit-8 (CCK-8) assay, and the absorbance was measured at 450 nm. For bacterial co-culture, the strains LL-pNZ8149, LL-SC, and LL-SC-C₂A₂ were induced with Nisin (5 ng/mL) during the mid-logarithmic phase. The bacteria were then harvested by centrifugation, washed twice with phosphate-buffered saline (PBS), and resuspended in antibiotic-free MEM. To determine a safe co-culture concentration, Caco-2 cells were incubated with the bacteria at final concentrations of 10⁸ and 10⁹ CFU/mL for 2 h. Cell viability was again measured by the CCK-8 assay.

Following the respective treatments, apoptosis and necrosis were analyzed by Hoechst 33,342 and propidium iodide (PI) double staining. Caco-2 cells were stained with Hoechst 33,342 (5 μg/mL) and PI (5 μg/mL) for 20 min at 4 °C in the dark, then washed with PBS and immediately imaged under a fluorescence microscope (Nikon Eclipse Ti2, Japan). Subsequently, intracellular ROS levels were assessed using the fluorescent probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA). Cells were incubated with 10 μM DCFH-DA in serum-free MEM at 37 °C for 30 min in the dark. After two washes with PBS, the cells were immediately observed and imaged using the fluorescence microscope, followed by quantitative analysis using a fluorescence microplate reader at excitation/emission wavelengths of 488/525 nm to determine intracellular ROS levels.

2.7. DSS-induced murine colitis model

Forty mice were randomly divided into five groups (n = 8 per group) and acclimatized for 7 days. For the final quantitative assessment, we required a complete set of high-quality samples from each animal for all analyses. Accordingly, the data presented in the figures are based on n = 5 mice per group that provided a complete and valid dataset for all endpoints measured. Mice were orally administered PBS, LL-pNZ8149, LL-SC, or LL-SC-C₂A₂ daily for a total of 14 days. This consisted of a 7-day pretreatment period, followed by a 7-day co-administration period during which the mice also received drinking water containing 3% DSS. Nisin (5 ng/mL) was included in the drinking water throughout the entire 14-day administration period to induce gene expression. The normal control (NC) group received only regular drinking water throughout the experiment (Fig. 2 A). The bacterial suspension was administered at a concentration of 10⁹ CFU/mL with a daily gavage volume of 200 μL per mouse. Body weight was recorded daily for all animals. At the endpoint, all mice were euthanized via eyeball removal and blood collection under anesthesia. Colon tissues were rapidly dissected and fixed in 4% paraformaldehyde, and all harvested tissues were stored at −80 °C for further analysis. Thymic indice, splenic indice, and colon length of mice were recorded after euthanasia [34].

Fig. 2.

Preparation, characterization, and biosafety evaluation of LL-SC-C₂A₂. (A) Zeta potential of LL-SC with different chitosan and sodium alginate layer coatings during the preparation of LL-SC-C₂A₂; (B) Growth curves of the strain in M17 broth medium with different coating layers; (C) Morphology of LL-SC and LL-SC-C₂A₂. Scale bar: 500 nm; (D) SOD activity of the strain at different culture times and (E) ultrasound conditions; (F) CAT activity and (G) SOD activity of the strain under different treatments; (H) Survival rate and (I) viable colony count of the strain in simulated gastrointestinal fluids; (J) HE staining of major organs from mice treated with PBS or LL-SC-C₂A₂ (1 × 10⁹ CFU). Scale bar: 100 μm; (K) Safety assessment of LL-SC-C₂A₂ in mice (1 × 10⁹ CFU). Blood cell parameters: WBC (a), RBC (b), HGB (c), HCT (d), MCH (e), MCHC (f), and PLT (g). Liver and kidney function markers: ALT (h), AST (i), BUN (j), and ALB (k). (l) Health status score during the feeding period. (n = 5; data presented as mean ± SD; statistical significance determined by one-way ANOVA: ∗∗∗P < 0.001, ∗∗P < 0.01, ∗P < 0.05, ns: no significance).

2.8. Disease activity index detection

Body weight was recorded daily in all groups starting from the initiation of DSS induction. The disease activity index (DAI) was assessed based on stool consistency, fecal bleeding, and body weight loss, with scoring criteria detailed in Table S3 [34].

2.9. Histological analysis

Colonic tissue sections (4 μm) were baked at 60 °C for 30 min. For HE staining, tissue slides were dewaxed, hydrated, and sequentially stained with hematoxylin for nuclear visualization and eosin for cytoplasmic counterstaining. Subsequent dehydration, clearing, and mounting with neutral resin were performed. For Alcian blue-periodic acid Schiff (AB-PAS) staining, dewaxed and hydrated slides were subjected to sequential treatments: Alcian blue staining for acidic mucins, oxidation, Schiff ‘s reagent incubation for glycogen detection, hematoxylin nuclear staining, differentiation, and bluing. Final dehydration, clearing, and neutral resin mounting completed the procedure [35]. For quantitative assessment of goblet cells, AB-PAS-stained sections were analyzed using ImageJ software (National Institutes of Health, v1.53t).

2.10. Intestinal permeability detection

Mice were fasted for 6 h prior to oral gavage with FITC-dextran (500 mg/kg body weight). Blood samples were collected after 4 h, and serum was isolated by centrifuging whole blood. Fluorescence intensity of serum samples was quantified using a multimode microplate reader (SpectraMax iD3, Molecular Devices, China) with excitation and emission wavelengths set to 492 nm and 525 nm, respectively. Serum FITC-dextran levels in experimental groups were calculated by a linear regression equation derived from the standard curve (R² > 0.99).

2.11. Oxidative stress and inflammatory markers

To detect the production of ROS, the 2′, 7′-Dichlorodihydrofluorescein diacetate (DCFH-DA) fluorescent probe method was employed. The frozen tissue sections were thawed at 25 °C, dried, treated with DCFH-DA staining, washed with PBS, and mounted. Quantitative analysis was performed using flow cytometry (NovoCyte 3110, Agilent, USA). Colon tissues designated for physiological index analysis were cut into small pieces and homogenized with PBS at a 1:9 (w/v) tissue-to-solution ratio. Levels of SOD, CAT, GSH-Px, and MDA in the tissue homogenates were separately determined using assay kits.

The numerical value of MPO is correlated with the degree of infiltration and is generally regarded as an inflammatory marker. The severity of inflammation can be evaluated by measuring the activity of MPO, an inflammatory biomarker. The MPO activity of colon tissues was detected by the MPO assay kit. The expression levels of signature inflammatory cytokines, including IL-10, IL-6, IL-1β, TNF-α, and TGF-β, were measured using enzyme-linked immunosorbent assay (ELISA, Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

2.12. Immunohistochemical analysis

The expression of mucin MUC-2 and tight junction proteins (Claudin-1, Occludin-1, and ZO-1) in colon tissues was analyzed by immunohistochemistry (IHC). Briefly, after paraffin-embedded sections were subjected to antigen retrieval, nonspecific binding was blocked with 5% bovine serum albumin (BSA). Sections were incubated overnight at 4 °C with primary antibodies: Claudin-1, MUC-2 (1:1000 dilution), Occludin, and ZO-1 (1:500 dilution). After washing with buffered saline containing Tween-20 (PBST), sections were treated with goat anti-rabbit IgG antibody (1:200 dilution) for 1 h. Diaminobenzidine (DAB) was used as the chromogenic substrate, followed by dehydration and mounting with neutral resin. Protein expression levels were quantified by measuring the mean optical density (MOD) using ImageJ software. TUNEL staining was also applied to evaluate the apoptosis of colon tissue by a TUNEL assay kit (Elabscience, China).

2.13. Quantitative analysis of tissue mRNA expression

Colon tissue fragments (20-30 mg) were homogenized in 1 mL RNAiso Plus, followed by centrifugation to isolate the aqueous phase. Chloroform (200 μL) was added to the supernatant, mixed vigorously, and centrifuged. The resulting supernatant was combined with 1 mL of isopropanol to precipitate RNA. The RNA pellet was washed twice with 75% ethanol, air-dried, and resuspended in diethylpyrocarbonate (DEPC)-treated water. RNA purity acceptable samples are defined by an OD₂₆₀/OD₂₈₀ ratio of 1.8-2.1. Gene-specific primers were designed using Primer Premier 5.0 software and synthesized by Sangon Biotech (Shanghai) Co., Ltd. The primer sequences for target genes are listed in Table S4. The RT-PCR reaction system for cDNA was prepared according to the TB Green® Premix Ex Taq™ II kit instructions, and the RT-PCR reaction was monitored in real time using the QuantStudio® 3 system. The internal reference gene was β-actin.

2.14. Gut microbiota analysis

Genomic DNA was extracted from murine intestinal contents using the cetyltrimethylammonium bromide (CTAB)-based method. The V3-V4 hypervariable regions of bacterial 16S rDNA were amplified via PCR with the following primers containing the sequences “CCTACGGGNGGCWGCAG” (forward primers) and “GACTACHVGGGTATCTAATCC” (reverse primers). Amplified products were subsequently analyzed using an Agilent 2100 Bioanalyzer system and the Illumina NovaSeq 6000 platform. The levels of short-chain fatty acids (SCFAs; acetic, propionic, butyric, isobutyric, valeric, and isovaleric acids) in colonic contents were quantified using gas chromatography-mass spectrometry (GC-MS) after acidified extraction and appropriate derivatization.

2.15. Statistical analysis

The mean ± standard deviation of data was used to perform statistical analysis through one-way ANOVA: ∗∗∗P < 0.001, ∗∗P < 0.01, ∗P < 0.05, ns: no significance.

3. Results and discussion

3.1. Construction and validation of LL-SC

To confer enhanced ROS scavenging capability on LL, gene-specific primers were designed based on the published CAT and SOD sequences from the NCBI database. Amplification yielded DNA fragments of approximately 2000 bp (Fig. 1B–a). SnapGene (v6.0.5, USA) analysis confirmed that the fused CAT-SOD fragment measured 2359 bp, indicating that the target genes were successfully amplified. The pNZ8149 plasmid was linearized by inverse PCR, generating a 2550 bp fragment as validated by SnapGene and electrophoretic analysis (Fig. 1B–b). The verified recombinant plasmid pNZ8149-SODCAT was electroporated into L. lactis NZ3900 host cells. Positive clones were screened via colony PCR, which detected a 2359 bp target band (Fig. 1C), indicating successful electroporation. Western blot analysis was performed to validate SOD and CAT expression in the engineered strain (LL-SA). A distinct band at 84 kDa was observed in the ultrasonic supernatant of Nisin-induced LL-SA cultures (Fig. 1D), consistent with the expected size of the fusion protein. No bands were detected in controls containing the empty pNZ8149 vector or uninduced LL-SA cultures, confirming Nisin-inducible expression of soluble target proteins in NZ3900. To assess the impact of antioxidant enzyme expression on bacterial growth, growth curves of LL, LL-pNZ8149, and LL-SA were analyzed. LL exhibited the highest growth rate (Fig. 1E), likely attributable to energy diversion toward plasmid replication and heterologous protein expression in engineered strains. Nevertheless, all three strains displayed comparable growth dynamics, indicating that recombinant protein synthesis imposes a negligible metabolic burden on cellular growth.

3.2. Preparation and characterization of LL-SC-C₂A₂

The zeta potential of each coating layer, measured by dynamic light scattering (DLS), was used to monitor encapsulation efficiency. As shown in Fig. 2A, uncoated LL-SC exhibited a negative surface charge (−23.14 ± 0.40 mV), attributed to anionic groups on the bacterial surface. Upon coating with chitosan (cationic polymer), the zeta potential shifted to a weakly positive value (+6.37 ± 0.29 mV, LL-SC-C₁), confirming successful primary layer coating. This inversion is a hallmark of LBL assembly, demonstrating that the positively charged chitosan chains have effectively neutralized and overcompensated for the original negative surface charge, forming a stable primary layer. Subsequent addition of sodium alginate (anionic polymer) reversed the charge to −20.73 ± 0.83 mV (LL-SC-C₁A₁), confirming the successful build-up of a secondary layer through the renewed charge overcompensation [36]. The stepwise growth in particle size (from 1209 nm to 1635 nm; Fig. S1) mirrored the oscillating zeta potential, together offering conclusive evidence for the progressive LbL assembly on the bacterial surface. Tan et al. [37] employed an LBL self-assembly strategy using sodium alginate and chitosan to encapsulate liposomes, which successfully improved their physicochemical stability. To assess the impact of coating layers on bacterial proliferation, growth kinetics of coated bacteria were compared with LL-SC. Both LL-SC-C₁A₁ and LL-SC-C₂A₂ exhibited growth kinetics comparable to uncoated LL-SC. In contrast, the LL-SC-C₃A₃ induced a modest growth lag, likely due to increased diffusion barriers from thicker polymeric matrices (Fig. 2B). Consequently, LL-SC-C₂A₂ was selected for subsequent experiments as it optimally balanced encapsulation stability with preserved bacterial functionality.

To further validate the successful encapsulation of LL-SC with chitosan and sodium alginate, TEM was employed to analyze bacterial morphology. LL-SC exhibited a rough surface with visible cell wall structures, whereas LL-SC-C₂A₂ displayed a smooth, continuous polymeric envelope that obscured the native bacterial surface (Fig. 2C). This distinct morphological transition confirmed the formation of multilayered coatings via electrostatic interactions between chitosan and sodium alginate. Through the electrostatic LBL deposition of chitosan and sodium alginate, a barrier is formed in the acidic environment of the stomach, significantly reducing the rate of hydrogen ion penetration into the interior, thereby greatly improving the survival rate of LL-SC in the gastric acid environment. The natural polysaccharide material confers excellent biocompatibility, safety, and potential prebiotic effects to this coating. Additionally, the chitosan layer improves intestinal mucosal adhesion, prolonging the duration of action of bacteria and enhancing the intestinal colonization efficiency of live bacteria.

3.3. Antioxidant and resistance of the LL-SC-C₂A₂

As the primary defense against superoxide anion radicals, SOD activity directly reflects bacterial responsiveness to initial oxidative stress. The culture supernatant constitutes a complex and variable environment unsuitable for standardized enzyme quantification. In contrast, intracellular extracts provide a controlled and consistent system, yielding more reliable and reproducible data for comparative analysis of the enzymatic activity of engineered strains. Consequently, SOD activity was determined using the supernatant of ultrasonically lysed cells. To elucidate the effects of coating treatment on bacterial antioxidant capacity, the optimal culture time was determined for SOD activity in coated bacteria. There was no significant difference in SOD activity when incubation time was 6 h and 10 h, while the activity at 8 h was significantly reduced (p < 0.01) (Fig. 2D). This dynamic SOD activity pattern likely reflects stage-specific regulatory mechanisms: ROS-induced upregulation during the logarithmic phase (6 h), metabolic downregulation-mediated suppression in the transition phase (8 h), and stress-triggered partial recovery at the early stationary phase (10 h). Although 200 W for 10 min, 400 W for 3 min, and 400 W for 5 min yielded comparable SOD activity (Figs. 2E), 400 W for 3 min was chosen because it offered the highest mean activity with minimal process duration and energy input. Based on these results, the subsequent conditions were selected as an incubation time of 6 h and an ultrasound power of 400 W for 3 min. The effects of inducer supplementation and coating treatment on antioxidant capacity were evaluated. The results show that plasmid-borne antioxidant gene expression was strictly inducer-dependent, with vector-free strains exhibiting minimal antioxidant activity even under induction conditions (Fig. 2F and G). These findings align with the Western blot profiles presented earlier. Furthermore, comparative analysis between LL-SC and LL-SC-C₂A₂ strains revealed no significant differences in enzymatic activity, demonstrating that the coating process does not interfere with protein expression.

The acidic milieu of gastric fluid critically compromises probiotic viability, enzymatic activity, and metabolic function, necessitating robust gastric resistance to ensure biological efficacy in vivo [38]. Carbohydrate-based encapsulation strategies, leveraging the unique physicochemical properties of polysaccharides, offer promising solutions to this challenge. As shown in Fig. 2H, LL-SC-C₂A₂ consistently demonstrated significantly higher survival rates compared to LL-SC across all time points (p < 0.05), with encapsulated cells showing a 1.4-fold increase in survival versus free cells. Further validation under simulated gastrointestinal conditions (Fig. 2I) revealed that encapsulation conferred exceptional protection and exhibited a significant increase in gastrointestinal fluid resistance relative to free cells (p < 0.05). These improvements are likely attributed to the ability of polysaccharide nanocoating to mitigate acid-induced damage through physical shielding, thereby preserving the structural and functional integrity of the encapsulated bacteria [38].

3.4. Biosafety evaluation of LL-SC-C₂A₂

For acute toxicity assessment, mice received daily oral administration of LL-SC and LL-SC-C₂A₂ for 7 consecutive days, respectively. Serum samples were collected for biochemical analysis, and major organs (liver, spleen, kidney, and colon) were harvested for histopathological examination. HE stained tissue sections demonstrated no observable histological damage or morphological differences in key organs between treated and NC groups (Fig. 2J). Hematological parameters in mice, including WBC (Fig. 2K–a), RBC (Fig. 2K–b), HGB (Fig. 2K–c), HCT (Fig. 2K–d), MCH (Fig. 2K–e), MCHC (Fig. 2K–f), and PLT (Fig. 2K–g), remained comparable to those of healthy controls (p > 0.05). Hepato-renal functional indices, including ALT (Fig. 2K–h), AST (Fig. 2K–i), BUN (Fig. 2K–j), and ALB (Fig. 2K–k), also fell within normal physiological ranges, further corroborating the biocompatibility of LL-SC-C₂A₂. Furthermore, behavioral and physiological observations (coat condition, activity levels, locomotor behavior, and fecal consistency) revealed no significant differences between the LL-SC-C₂A₂ and NC groups (Fig. 2K–l). Overall, the antioxidant enzyme delivery system developed in this study demonstrates favorable biocompatibility. It should be noted that this study mainly focused on the acute toxicity assessment of engineered bacteria. In the future, it is necessary to extend the feeding period and conduct 28-day and 90-day subacute toxicity tests, as well as further genetic toxicity and chronic toxicity tests, in order to comprehensively assess their long-term biological safety [39].

3.5. In vitro validation of antioxidant functionality in Caco-2 cells

To establish a reliable platform for evaluating antioxidant efficacy, an H₂O₂-stimulated oxidative stress model was first established in human intestinal epithelial Caco-2 cells. A standardized injury condition of 0.1 mM H₂O₂ for 4 h was selected for subsequent experiments, as this treatment consistently reduced cell viability to 51.86% ± 1.81% (Fig. 3A). Prior to protection assays, safety evaluation confirmed that bacterial co-culture at both 10⁸ and 10⁹ CFU/mL for 2 h induced no detectable cytotoxicity (Fig. 3B), eliminating potential confounding effects from bacterial toxicity. Evaluation of probiotic-mediated protection revealed substantial differences among strains (Fig. 3C). The encapsulated strain LL-SC-C₂A₂ provided superior protection, significantly elevating cell viability to 83.98% ± 2.88%. This protective efficacy markedly exceeded that of both the empty vector control LL-pNZ8149 (55.92% ± 4.02%, p < 0.001) and the unencapsulated engineered strain LL-SC (73.16% ± 2.31%, p < 0.01). These results demonstrate that engineered bacteria effectively mitigate H₂O₂-stimulated cellular injury, with nano-encapsulation substantially augmenting this protective capacity.

Fig. 3.

Cytoprotection and ROS scavenging by LL-SC-C₂A₂ in H₂O₂-stimulated Caco-2 cells. (A) Viability of cells after H₂O₂ treatment, (B) Safety assessment of bacterial co-culture, (C) Bacterial protection against H₂O₂-stimulated cell damage, (D) Fluorescence images of Hoechst 33342/PI stained Caco-2 cells; scale bar: 275 μm, (D) Fluorescence images of Hoechst 33342/DCFH-DA stained ROS in Caco-2 cells; scale bar: 650 μm, (E) Intracellular ROS levels measured by DCFH-DA fluorescence (n = 3; data presented as mean ± SD; statistical significance determined by one-way ANOVA: ∗∗∗P < 0.001, ∗∗P < 0.01, ∗P < 0.05, ns: no significance).

Furthermore, analysis of apoptosis and necrosis via Hoechst 33342/PI double staining confirmed that LL-SC-C₂A₂ effectively intercepted the downstream cell death pathways, significantly reducing the population of PI-positive cells compared to other groups (Fig. 3D). The antioxidant mechanism was further investigated by fluorescence imaging of DCFH-DA staining (Fig. 3E), which confirmed the superior ROS-scavenging capacity of LL-SC-C₂A₂, as evidenced by markedly reduced green fluorescence intensity compared to other treatment groups. Subsequent quantitative analysis (Fig. 3F) demonstrated that the H₂O₂ challenge triggered a pronounced increase in ROS generation, reaching 185.63% ± 3.53% of control levels. Pretreatment with LL-SC-C₂A₂ most effectively suppressed this oxidative burst, reducing intracellular ROS to 122.73% ± 3.40% of control levels, significantly lower than LL-pNZ814 and LL-SC (p < 0.001).

3.6. Alleviating effect of LL-SC-C₂A₂ against DSS-induced colitis model

To recapitulate human ulcerative colitis pathology, an experimental colitis model was established in mice through 3% DSS administration (Fig. 4A). The alleviating effects of engineered bacteria against IBD were quantitatively evaluated via multimodal assessments, including longitudinal body weight monitoring, DAI scoring, thymosplenic indices, colon morphometry, intestinal barrier integrity assays, and histopathological analyses. By the end of modeling, the body weight of the treatment group showed varied degrees of reduction (Fig. 4B). While LL-SC and LL-pNZ8149 demonstrated marginal mitigation of DSS-induced weight loss, LL-SC-C₂A₂ could significantly alleviate the DSS-induced body mass decline (p < 0.001). The DAI scoring system serves as a composite quantitative metric for evaluating disease severity in IBD models (Fig. 4C). The DSS-induced mice exhibited hallmark IBD characteristics, with severe weight loss, liquid stool consistency, and marked hematochezia. LL-SC-C₂A₂ group intervention significantly attenuated DSS-induced colitis severity, restoring normal stool morphology, reducing hematochezia, and improving weight loss [40]. The DAI scores of the LL-SC-C₂A₂ group (1.61 ± 0.33) were significantly lower than the DSS-induced group (2.92 ± 0.17, p < 0.001). Thymic and splenic indices served as critical biomarkers of systemic inflammation in DSS-induced models. The DSS-induced group exhibited thymic atrophy coupled with compensatory splenomegaly, reflecting immunosuppression-driven thymic involution and splenic lymphocyte hyperactivation that exacerbated intestinal inflammation. LL-SC-C₂A₂ treatment alleviated these pathological alterations, restoring thymic index increase to 0.22 ± 0.04 and splenic index reduction to 0.42 ± 0.09, with both parameters approximating NC group levels (Fig. 4D and E). Furthermore, colonic injury due to DSS induction was significantly improved by LL-SC-C₂A₂ (5.07 ± 0.22 cm, p < 0.001) compared to LL-SC (4.23 ± 0.31 cm, p < 0.01) and LL-pNZ8149 (3.93 ± 0.38 cm, p < 0.001) (Fig. 4F). Concurrently, intestinal permeability was assessed by measuring serum FITC-dextran concentrations. As shown in Fig. 4G, concentrations in the DSS group were significantly elevated compared to the NC group (P < 0.001), confirming that DSS induced significant intestinal damage. Following intervention with EPs, FITC-dextran concentrations decreased, indicating protective effects of the engineered antioxidant genes on intestinal mucosa. The most significant reduction was observed in the LL-SC-C₂A₂ group, supporting that the polysaccharide nanocoating, through its protective function, enhanced probiotic delivery and the consequent barrier improvement.

Fig. 4.

Treatment efficacy of LL-SC-C₂A₂ in DSS-induced murine colitis. (A) Schematic of the experimental procedure for the treatment of DSS-induced colitis mice; (B) Body weight changes in mice across different treatment groups; (C) Disease activity index (DAI) scores under different treatments; (D) Thymus index and (E) Spleen index of mice; (F) Quantified colon lengths from mice post-treatment; (G) Intestinal permeability assessed by FITC-dextran assay; (H) Intracellular ROS levels measured by flow cytometry; (I) ROS-positive rate; (J) CAT, (K) SOD, (L) GSH-Px, and (M) MDA levels in mouse colon tissue. Data are presented as mean values ± SD (n = 5 biologically independent samples). Statistical significance was determined by one-way ANOVA (∗∗∗P < 0.001, ∗∗P < 0.01, ∗P < 0.05, ns: no significance).

Flow cytometric analysis revealed significant intergroup variations in colonic ROS levels (Fig. 4H). The ROS level in the NC group measured 4.65% ± 0.05%, whereas DSS-induced colitis significantly increased to 22.63% ± 1.24% (P < 0.001, Fig. 4I). In the intervention groups, ROS levels decreased to 9.46% ± 0.46%, 14.11% ± 0.35%, and 16.10% ± 0.34% in the LL-SC-C₂A₂, LL-SC, and LL-pNZ8149 groups, respectively. Notably, LL-SC-C₂A₂ exhibited superior ROS scavenging capability, achieving 1.5-fold higher than LL-SC. Recent research has also confirmed that the synergistic effect of inulin-trifluoroacetic acid-sodium alginate nanoparticles and probiotics can effectively alleviate IBD through the following mechanisms: reducing oxidative stress, inhibiting inflammatory responses, repairing intestinal barrier function, and promoting probiotic proliferation, thereby reversing intestinal microbiota imbalance [41]. The strategy of employing nano-drug delivery systems to potentiate the efficacy and mitigate the limitations of biologic therapies represents a promising frontier in biomedicine.

In the oxidative stress index assay, the DSS-induced group exhibited decreased SOD (Fig. 4J), CAT (Fig. 4K), and GSH-Px (Fig. 4L) levels and increased lipid peroxidation marker MDA (Fig. 4M) levels compared with the NC group (P < 0.001). Following engineered probiotic intervention, differential amelioration was observed across multiple pathological indicators, and the LL-SC-C₂A₂ group showed the strongest antioxidant repair capacity, with SOD, CAT, and GSH-Px activities reaching 1.56, 1.76, and 1.33-fold higher than the LL-SC group, respectively. To dissect the origin of the antioxidant effect, an H₂O₂ -stimulated Caco-2 cell model was employed. Analysis of the cell culture supernatant revealed that SOD and CAT activities were specifically elevated only upon co-culture with the engineered bacteria (Fig. S2A), providing direct evidence that these functional enzymes are produced and secreted by the engineered probiotics. Furthermore, analysis of Caco-2 cell lysates demonstrated that co-culture with engineered bacteria restored the intracellular activities of all three antioxidant enzymes, which were suppressed by H₂O₂ stimulation (Fig. S2B). This crucial finding, coupled with the observed reduction in intracellular ROS, suggests a synergistic therapeutic mechanism. The engineered probiotics directly supply SOD and CAT to neutralize extracellular ROS. They also mitigate the primary oxidative insult, creating a microenvironment that enables host cells to restore their endogenous antioxidant defenses. This dual mode of action provides a comprehensive explanation for the therapeutic efficacy observed in our murine colitis model.

3.7. LL-SC-C₂A₂ reduced the inflammation

To elucidate the therapeutic mechanisms, we performed histopathological analyses (HE and AB-PAS staining) and apoptosis assessment (TUNEL staining) of colonic tissues. As shown in Fig. 5A, the colonic architecture in the NC group remained intact, with normal lamina propria and muscularis propria thickness and abundant goblet cells. DSS induction resulted in significant pathological alterations in colonic tissues, including epithelial cell necrosis and exfoliation, disrupted and loose tissue architecture, crypt loss, submucosal edema, massive inflammatory cell infiltration, and a severe depletion of goblet cells. Following intervention with LL-SC-C₂A₂, the pathological state of colonic tissues markedly improved, characterized by a compact epithelial structure, only scattered inflammatory cells observed in the mucosal layer, intact glandular architecture, and a near-complete restoration of goblet cell numbers. Quantitative analysis of goblet cells using ImageJ software further confirmed these observations. As shown in Fig. S3, DSS-induced colitis led to severe mucosal erosion and an almost complete loss of goblet cells. In contrast, LL-SC-C₂A₂ treatment significantly restored goblet cell numbers to 243 ± 15, approaching the level observed in the NC group. Consistent with these histological observations, TUNEL staining revealed minimal red fluorescence (indicative of baseline physiological apoptosis) in the NC group. In contrast, the DSS group exhibited a significant increase in red fluorescence, an abnormality attributable to stimulus-induced cellular damage leading to elevated apoptosis rates. Following various interventions, the number of apoptotic cells was reduced to varying degrees. Among these, the LL-SC-C₂A₂ intervention group demonstrated the most pronounced effect, with apoptosis levels approaching those of the NC group, followed by the LL-SC group. These findings indicate that both antioxidant enzymes and polysaccharide nanocoating can confer a degree of protection against DSS-induced cellular damage.

Fig. 5.

Therapeutic effect of LL-SC-C₂A₂ on inflammation in a DSS-induced colitis model. (A) Representative images of H&E, AB-PAS, and TUNEL staining in colon tissues. Scale bar: 200 μm/50 μm; (B–G) Levels of inflammatory markers in colon tissues: (B) IL-6, (C) IL-1β, (D) TNF-α, (E) TGF-β1, (F) IL-10, and (G) MPO in the colon tissues of mice. Data are presented as mean values ± SD (n = 5 biologically independent samples). Statistical significance was determined by one-way ANOVA (∗∗∗P < 0.001, ∗∗P < 0.01, ∗P < 0.05, ns: no significance).

Subsequently, we evaluated the intervention efficacy by measuring key inflammatory cytokines via ELISA. The DSS group exhibited a highly significant elevation (p < 0.001) in pro-inflammatory cytokine levels: IL-6 (73.84 ± 4.20 pg/mL) (Fig. 5B), IL-1β (97.07 ± 5.34 pg/mL) (Fig. 5C), TNF-α (684.55 ± 35.40 pg/mL) (Fig. 5D), and TGF-β (258.76 ± 9.33 pg/mL) (Fig. 5E). Conversely, the level of the anti-inflammatory cytokine IL-10 (44.19 ± 2.50 pg/mL) (Fig. 5F) was significantly reduced. Administration of EPs significantly decreased the levels of all pro-inflammatory cytokines (p < 0.001), with the LL-SC-C₂A₂ group demonstrating the most pronounced inhibitory effect. Specifically, levels in the LL-SC-C₂A₂ group decreased to IL-6 (48.23 ± 3.37 pg/mL), IL-1β (52.00 ± 3.80 pg/mL), TNF-α (416.48 ± 25.83 pg/mL), and TGF-β (161.12 ± 13.13 pg/mL). Concomitantly, IL-10 levels were significantly elevated to 57.37 ± 1.73 pg/mL (p < 0.05) in the LL-SC-C₂A₂ group. Furthermore, compared to other intervention groups, LL-SC-C₂A₂ significantly reduced MPO activity to 0.39 ± 0.05 U/mg tissue (Fig. 5G). In summary, LL-SC-C₂A₂ significantly ameliorated DSS-induced colonic inflammation, as evidenced by a pronounced reduction in pro-inflammatory cytokines (IL-6, IL-1β, TNF-α, TGF-β; p < 0.001), a significant increase in the anti-inflammatory cytokine IL-10 (p < 0.001), and effective suppression of MPO activity. This anti-inflammatory mechanism aligns with research by Liao et al. [42], wherein Bacillus cereus reduced IL-1β, IL-6, and TNF-α concentrations through bile acid-FXR axis activation.

3.8. LL-SC-C₂A₂ repaired intestinal barrier function

3.8.1. Immunohistochemical analysis of barrier components

In addition to upregulated inflammation levels, the impaired intestinal barrier was also an important indicator of colitis [43]. Mucin-2 (MUC-2), the predominant mucin secreted by goblet cells, constitutes the primary chemical barrier of the colonic mucosa. Complementing this protective layer, the tight junction (TJ) proteins maintain physical barrier integrity by regulating paracellular permeability. To assess the restoration of intestinal barrier function in DSS-induced inflammatory mice by the intervention groups, the expression of MUC-2 and TJ proteins was examined. Immunohistochemistry (IHC) analysis (Fig. 6A) showed that DSS induction markedly disrupted intestinal barrier integrity, as indicated by the downregulated expression and structural abnormalities of key proteins, including MUC-2 and TJ proteins (Claudin-1, Occludin-1, and ZO-1). Intervention with LL-SC-C₂A₂ restored tissue architecture and significantly elevated the expression levels of these four proteins to near NC group levels. Quantitative analysis demonstrated significantly reduced (p < 0.001) mean optical density values (MOD) for MUC-2 (Fig. 6B), Claudin-1 (Fig. 6C), Occludin-1 (Fig. 6D), and ZO-1 (Fig. 6E) following DSS administration. Post-intervention, expression levels of the four proteins exhibited varying degrees of recovery, with the LL-SC-C₂A₂ group demonstrating the most pronounced restorative effect. Specifically, LL-SC-C₂A₂ effectively rescued protein expression (MUC-2: 0.1577 ± 0.0033, Claudin-1: 0.1270 ± 0.0085, Occludin-1: 0.0933 ± 0.0049, and ZO-1: 0.0776 ± 0.0040). These results indicate that LL-SC-C₂A₂ intervention mitigates MUC-2 and TJ protein loss in murine colonic tissue, stabilizing the mucosal barrier.

Fig. 6.

Enhancement of intestinal barrier function by LL-SC-C₂A₂ in DSS-induced colitis mice. (A) Representative immunohistochemical (IHC) staining images of MUC-2, Claudin-1, Occludin-1, and ZO-1 in colon tissues. Scale bar: 50 μm; (B–E) Quantification of tight junction protein and mucin expression in colon tissue: (B) MUC-2, (C) Claudin-1, (D) Occludin-1, (E) ZO-1 (mean optical density); (F–I) mRNA expression levels of intestinal barrier-related genes: (F) MUC-2, (G) ZO-1, (H) Occludin-1, (I) Claudin-1; (J–M) mRNA expression levels of related pathway genes (J) Nrf2, (K) Keap1, (L) NQO1, (M) HO-1. Data are presented as mean values ± SD (n = 5 biologically independent samples). Statistical significance was determined by one-way ANOVA (∗∗∗P < 0.001, ∗∗P < 0.01, ∗P < 0.05, ns: no significance).

3.8.2. mRNA expression profiling of barrier-related pathways

Complementing our immunofluorescence findings of protein restitution, we interrogated the transcriptional regulation underlying barrier repair by quantifying mRNA expression of MUC-2 (Fig. 6F), Claudin-1 (Fig. 6G), Occludin-1 (Fig. 6H), and ZO-1 (Fig. 6I). Compared with the DSS group, all intervention groups exhibited varying degrees of increased MUC-2 expression. This restoration of mucin secretion is consistent with previous AB-PAS staining results, confirming its role in maintaining mucosal integrity and protecting goblet cell function. Additionally, DSS administration significantly inhibited the mRNA expression of all three TJ proteins, while LL-SC-C₂A₂ intervention treatment reversed this deficiency, thereby restoring the intestinal mucosal barrier. The superior efficacy of LL-SC-C₂A₂ in upregulating MUC-2 and TJ proteins (p < 0.001) can be attributed to the synergy between the protective delivery afforded by the polysaccharide coating and the sustained local production of antioxidant enzymes by the engineered bacteria.

To investigate the mechanism underlying the dual effects of LL-SC-C₂A₂ on barrier repair and antioxidant properties, we systematically evaluated the Nrf2-Keap1 signaling pathway at the transcriptional level. In DSS-induced colitis mice, we observed a significant suppression of this pathway, characterized by decreased Nrf2 mRNA and increased Keap1 mRNA (Fig. 6J and K). Intervention with LL-SC-C₂A₂ markedly reversed these trends, restoring the expression of both genes to levels comparable to the normal control (p < 0.001). NADPH quinone oxidoreductase 1 (NQO1) and heme oxygenase 1 (HO-1) are downstream effectors of the Nrf2-Keap1 pathway. Following DSS induction, the mRNA expression of the NQO1 (Fig. 6L) and HO-1 (Fig. 6M) genes was significantly suppressed (p < 0.001). LL-SC-C₂A₂ intervention restored NQO1 and HO-1 mRNA levels by fourfold and twofold, respectively, compared with the DSS group. It is important to note that the present study provides evidence at the transcriptional level. A comprehensive validation of Nrf2 pathway activation would require further confirmation of Nrf2 protein nuclear translocation and subsequent upregulation of downstream effector proteins such as HO-1 and NQO1, as demonstrated in previous studies [44]. Nevertheless, the coordinated upregulation of Nrf2, NQO1, and HO-1 mRNA, coupled with the downregulation of Keap1, strongly suggests that the antioxidant effects of LL-SC-C₂A₂ are mediated, at least in part, through the modulation of the Nrf2-Keap1 signaling pathway.

3.9. Regulatory effect of LL-SC-C₂A₂ on the gut microbiome

Gut microbiota homeostasis is closely related to intestinal health. To explore the capability of LL-SC-C₂A₂ to regulate the gut microbiota, the intestinal contents of mice were obtained to perform 16S rDNA analysis. Significant disparities in alpha diversity indices, including Chao1 (Fig. 7A), Simpson (Fig. 7B), and Shannon (Fig. 7C), were observed between the DSS group and the NC group (p < 0.001). After LL-SC-C₂A₂ intervention, the above indices were returned to levels comparable to the NC group, indicating its efficacy in improving the alpha diversity within the gut microbiota. Principal coordinate analysis (PCoA) visualized the spatial distribution of samples based on weighted UniFrac distance (Fig. 7D). The horizontal axis (PCoA1, 55.12%) and vertical axis (PCoA2, 25.71%) percentages represent the explanatory power of each principal coordinate axis for sample differences. The NC group samples exhibited tight intra-group clustering, indicating high structural homogeneity, whereas the DSS group samples occupied divergent quadrants, demonstrating severe dysbiosis. Probiotic interventions partially restored microbial architecture, with LL-SC-C₂A₂ achieving the closest compositional resemblance to the NC group. The non-metric multidimensional scaling (NMDS) results were consistent with PCoA (Fig. 7E), with DSS group samples being farther from the NC group and showing some intra-group dispersion, further confirming that DSS induced changes in mouse gut microbiota composition. After probiotic intervention, microbiota structure changes improved, and diversity tended to recover. The microbiota composition of the LL-SC-C₂A₂ group was similar to the NC group. The beta diversity analyses collectively demonstrate the superior capacity of LL-SC-C₂A₂ in reconstituting gut microbiota integrity.

Fig. 7.

Changes in intestinal flora and short-chain fatty acids in mice. (A–C) Alpha diversity indices of the gut microbiota: (A) Chao1 index, (B) Simpson index, (C) Shannon index; (D–E) Beta diversity analysis: (D) Principal Coordinates Analysis (PCoA) and (E) Non-metric Multidimensional Scaling (NMDS); (F) Relative abundance of the top 10 phyla. (G–H) Relative abundance of (G)Proteobacteria and (H)Firmicutes. (I) Relative abundance of the top 10 genera; (J–K) Relative abundance of (J)Escherichia_Shigella and (K)Akkermansia; (L–Q) Content of SCFAs in intestinal contents of mice (L) acetic acid (M) propionic acid (N) butyric acid (O) isobutyric acid (P) valeric acid, and (Q) isovaleric acid; (R) Intestinal flora and metabolite association analysis. (n = 3; data presented as mean ± SD; statistical significance determined by one-way ANOVA: ∗∗∗P < 0.001, ∗∗P < 0.01, ∗P < 0.05, ns: no significance).

By comparing the taxonomic composition differences of murine gut microbiota across, the impact of EPs intervention was analyzed. Phylum-level analysis (Fig. 7F) revealed that DSS-induced treatment significantly reduced the abundance of the Firmicutes (Fig. 7G) and Bacteroidetes while significantly increasing the abundance of the Proteobacteria (Fig. 7H), Campylobacterota, and Actinobacteria, indicating pathogenic dysbiosis characterized by depletion of dominant phyla and enrichment of pathobionts. EPs intervention moderately increased the abundance of the Firmicutes, Bacteroidetes, and Verrucomicrobia, while significantly reducing the abundance of the Proteobacteria. The ratio of Firmicutes to Proteobacteria in the LL-SC-C₂A₂ group was closest to the NC group. These phylum-level shifts confirmed DSS-induced microbial dysregulation, with LL-SC-C₂A₂ showing particularly prominent effects. Genus-level analysis (Fig. 7I) of gut microbiota demonstrated that DSS induction significantly increased the abundance of pathogenic genera, including Escherichia_Shigella (a proteobacterial pathobiont linked to intestinal inflammation, Fig. 7J), Streptococcus, and Clostridium_sensu_stricto_1, while depleting beneficial taxa, including Akkermansia (promotes mucus layer regeneration, Fig. 7K) [45], Ligilactobacillus (inhibits pathogens and modulates immunity) [46], and Lachnospiraceae_NK4A136_group (produces anti-inflammatory propionate and butyrate) [47]. Following intervention with LL-SC-C₂A₂, the abundance of pathogenic bacterial genera significantly decreased, while the abundance of beneficial bacterial genera such as Akkermansia, Ligilactobacillus, and Muribaculaceae increased. Although LL-SC also exhibited positive regulatory effects, its efficacy was weaker than that of LL-SC-C₂A₂. This further confirms that the polysaccharide nanocoating protects the engineered bacteria, enabling them to effectively release antioxidant enzymes in the gut, which in turn promotes the proliferation of beneficial probiotics and restores microbial community homeostasis.

To further investigate the functional accompanying the structural shifts in the gut microbiota, the levels of short-chain fatty acids (SCFAs) in the colonic contents were quantified. As shown in Fig. 7L–Q, the DSS challenge resulted in a severe depletion of acetic acid, propionic acid, butyric acid, isobutyric acid, valeric acid, and isovaleric acid. Treatment with LL-SC-C₂A₂ most effectively restored the content of these SCFAs. Spearman correlation analysis was performed to directly establish the link between microbial structure and metabolic function (Fig. 7R). The results revealed that the bacterial genera specifically enriched by LL-SC-C₂A₂, notably Ligilactobacillus, exhibited strong and significant positive correlations with acetate and propionate levels. Conversely, pathobionts that were markedly suppressed by LL-SC-C₂A₂, such as Escherichia-Shigella and Streptococcus, showed significant negative correlations with these beneficial SCFAs. It should be noted that the present analysis focused on SCFAs. Future metabolomic profiling should extend beyond SCFAs to include bile acids and other key microbial metabolites.

3.10. Limitations and future perspectives

While this study demonstrates the considerable therapeutic promise of LL-SC-C₂A₂, it is important to acknowledge its limitations to provide a balanced perspective and guide future research. First, the animal experimental design primarily assessed the prophylactic potential of LL-SC-C₂A₂ in a DSS-induced acute colitis model, which reflects certain features of IBD but does not fully represent the full spectrum of the condition, including Crohn's disease. Moreover, the delayed therapeutic efficacy of LL-SC-C₂A₂ in established colitis and its effects in chronic or relapsing ulcerative colitis models remain to be investigated to comprehensively evaluate its long-term effects. Second, as a proof-of-concept study for a novel therapeutic strategy, our work focused on validating the intrinsic efficacy and multi-mechanistic action of LL-SC-C₂A₂ against control groups. Consequently, a direct efficacy comparison with established first-line clinical drugs was not included. Third, while the coordinated mRNA upregulation of key Nrf2-Keap1 pathway components provides strong correlative evidence, our study was limited to transcriptional-level analysis. Fourth, the in vivo kinetics of the engineered L. lactis, including its precise colonization duration, spatial distribution, and metabolic activity, were not fully characterized. A deeper understanding of these dynamics is crucial for rational dosing regimen design. Finally, while we confirmed the expression of the SOD-CAT fusion protein and demonstrated its functional activity in bacterial lysates, a detailed biochemical characterization of the purified protein (purity, stability, and specific antioxidant activity) was not performed. To address these limitations and advance the platform, future studies should focus on: (1) validating the efficacy in broader preclinical models, including assessing its delayed therapeutic effect in colitis and the chronic model of ulcerative colitis, as well as in complementary models such as TNBS-induced colitis, which is characteristic of Crohn's disease; (2) conducting comparative efficacy studies with first-line clinical drugs (5-ASA) to delineate its therapeutic potential and relative advantages; (3) confirming the activation of the key pathway at the protein level; (4) elucidating the in vivo colonization kinetics, spatial distribution, and metabolic activity of the engineered bacteria to inform rational dosing regimens; and (5) performing systematic biochemical characterization of the purified SOD-CAT fusion protein to lay the groundwork for future product quality control.

The clinical translation of this engineered probiotic platform presents specific challenges. Foremost among these is establishing a scalable and reproducible manufacturing process for LL-SC-C₂A₂. This encompasses two main fronts: upstream, the optimization of fermentation processes in large-scale bioreactors to achieve high-density cultivation of the engineered L. lactis; and downstream, the development of a reproducible process for the LBL nano-encapsulation that ensures consistent product quality and high bacterial viability. Beyond manufacturing, systematic long-term biosafety assessments are imperative for clinical advancement. This necessitates extended toxicity studies (subchronic and chronic), detailed evaluations of genetic stability, and the incorporation of specific assays to assess genotoxicity and the risk of horizontal gene transfer.

4. Conclusion

This study employed genetic engineering to integrate SOD and CAT genes into Lactococcus lactis, followed by chitosan-sodium alginate encapsulation, successfully constructing an engineered probiotic LL-SC-C₂A₂ capable of expressing antioxidant enzymes for the treatment of IBD. The coating layer serves as an effective physical barrier, significantly enhancing the resistance of engineered bacteria to the harsh gastrointestinal environment, thereby improving oral delivery efficiency and bioavailability. In a DSS-induced mouse colitis model, oral administration of LL-SC-C₂A₂ significantly improved disease symptoms, manifested as weight recovery, reduced disease activity index, and increased colon length. This protective effect stems from the combined effects of LL-SC-C₂A₂ in effectively scavenging ROS, inhibiting the secretion of pro-inflammatory factors, enhancing anti-inflammatory factor levels, repairing intestinal barrier function, and regulating intestinal microbiota homeostasis, with the coating polysaccharides and engineered bacteria exhibiting significant synergistic effects. This study not only developed an innovative oral delivery platform based on engineered probiotics but also provided a new strategy for synergistic treatment of IBD by integrating multiple mechanisms, including antioxidant therapy, anti-inflammatory protection, barrier repair, and microecological regulation. While this study has certain limitations, as discussed, our future work will be dedicated to advancing its translational path. This includes validating the strategy in additional disease models, resolving key mechanistic and manufacturing questions, and conducting comprehensive long-term biosafety.

CRediT authorship contribution statement

Yue Meng: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft. Chao Miao: Investigation, Methodology. Jiaxin Guo: Formal analysis, Methodology. Wei Zhang: Methodology. Ling Guo: Formal analysis. Yu Zhang: Conceptualization, Funding acquisition, Project administration. Yujun Jiang: Conceptualization, Project administration.

Declaration of competing interest

None.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.