Orally administered biomimetic nanovesicles engineered with FGF2 orchestrate mucosal healing and microbiome remodeling in inflammatory bowel disease

Abstract

Inflammatory bowel disease (IBD) is characterized by chronic intestinal inflammation and profound microbial dysbiosis, presenting significant therapeutic challenges. While fibroblast growth factor 2 (FGF2) possesses potent regenerative capabilities, its oral administration is severely hindered by rapid gastrointestinal degradation. To overcome these delivery barriers, this study investigates a novel, targeted therapeutic strategy utilizing nanoscale outer membrane vesicles (OMV/FGF2) naturally secreted during the normal growth of FGF2-engineered Gram-negative bacteria. The isolated OMV/FGF2 (120.1 nm, −13.7 mV) maintained robust structural integrity in simulated gastric fluid and demonstrated highly favorable cytocompatibility. In a dextran sulfate sodium (DSS)-induced murine colitis model, orally administered OMV/FGF2 significantly attenuated disease severity, mitigating weight loss and colon shortening. Mechanistically, OMV/FGF2 actively restored the intestinal physicochemical barrier by upregulating tight junction proteins (Occludin, ZO-1) and promoting mucus hypersecretion. Furthermore, 16S rRNA analysis revealed that OMV/FGF2 reversed microbial dysbiosis, enhancing α-diversity and enriching beneficial commensals (e.g., Bacteroides, Lactobacillus) while suppressing pathogenic populations. Ultimately, OMV/FGF2 ameliorates intestinal inflammation through a synergistic dual mechanism of fortifying the epithelial barrier and remodeling the gut microbiome. This engineered nanoplatform provides a promising, orally bioavailable therapy for IBD.

Graphical abstract

1. Introduction

Inflammatory bowel disease (IBD), primarily encompassing ulcerative colitis (UC) and Crohn's disease (CD), is a chronic, relapsing inflammatory disorder of the gastrointestinal tract that affects millions of individuals worldwide [1,2]. The pathogenesis of IBD is multifactorial, involving a complex interplay of genetic susceptibility, environmental triggers, and immune dysregulation, all of which culminate in profound intestinal epithelial barrier damage and microbial dysbiosis [[3], [4], [5]]. Chronic mucosal inflammation not only disrupts the physical-chemical barrier but also triggers a detrimental feedback loop where aberrant microbial metabolites further exacerbate immune hyperactivation [6]. Despite the availability of conventional therapies, including aminosalicylates, corticosteroids, and immunosuppressants, many patients suffer from inadequate clinical remission or severe systemic side effects, such as hepatotoxicity and secondary infections [[7], [8], [9]]. Consequently, there is an urgent clinical mandate for targeted, locally acting therapeutic platforms that can restore intestinal homeostasis without the risks associated with systemic exposure.

Fibroblast growth factor 2 (FGF2) has emerged as a potent morphogen critical for organogenesis and tissue regeneration, demonstrating significant potential in promoting epithelial cell proliferation and mucosal healing [10]. However, the clinical translation of FGF2 for IBD via oral administration is severely hindered by its inherent pharmacological limitations. Growth factors generally possess a remarkably short biological half-life and are highly susceptible to rapid proteolytic degradation in the harsh, acidic environment of the gastrointestinal tract [11]. To overcome these delivery barriers, recent research has pivoted toward nanoparticle-based drug delivery systems; however, synthetic nanocarriers often encounter challenges regarding long-term in vivo toxicity and high manufacturing costs [[12], [13], [14]].

In contrast, bacterial outer membrane vesicles (OMVs) represent a promising, biomimetic oral delivery platform for targeted colitis therapy. These nanoscale vesicular structures are naturally secreted by Gram-negative bacteria during their normal growth and inherit a diverse array of parental components, including proteins and lipids, which facilitate inherent biocompatibility and mucosal penetration. By leveraging this innate biological process, we genetically engineered Escherichia coli DH5α to serve as biological "nanofactories," yielding nanoscale outer membrane vesicles (OMV/FGF2) naturally secreted during the normal growth of these FGF2-engineered bacteria. This bioengineering strategy ensures that the therapeutic growth factor is integrated with the OMV architecture, protecting its bioactivity from gastric degradation while facilitating direct delivery to the inflamed colonic mucosa.

In this study, we rigorously characterized the physicochemical properties and biosafety of OMV/FGF2 and evaluated its therapeutic efficacy in a murine model of dextran sulfate sodium (DSS)-induced colitis. Our results demonstrate that orally administered OMV/FGF2 effectively alleviates intestinal inflammation through a synergistic dual mechanism of fortifying the epithelial barrier and remodeling the gut microbiome. Specifically, OMV/FGF2 treatment upregulates tight junction proteins (Occludin and ZO-1) to seal the damaged epithelial barrier and reshapes the microbial landscape by enriching beneficial, short-chain fatty acid-producing commensals. This study not only highlights the potential of engineered OMVs as a versatile tool for targeted intervention in IBD but also paves a new technical pathway for the oral delivery of therapeutic growth factors.

2. Results

2.1. Preparation and physicochemical characterization of OMV/FGF2

To establish the biological source of the nanovesicles, an engineered Escherichia coli (E. coli, DH5α) strain harboring the human FGF2 gene and an ampicillin resistance cassette was utilized (constructed by Wenzhou Medical University, China). Wild-type E. coli DH5α served as the control strain for conventional OMV isolation. Utilizing the natural secretion process of these Gram-negative bacteria during normal growth, FGF2-engineered bacterial outer membrane vesicles (OMV/FGF2) were subsequently isolated and purified from the culture supernatant using a customized ultracentrifugation protocol.

Nanoparticle tracking analysis (NTA) demonstrated that the purified OMV/FGF2 possessed an average hydrodynamic diameter of 120.1 nm, with the particle size distribution peaking at 112.5 nm (Fig. 1a). The particle concentration reached approximately 2 × 10¹¹ particles/mL. Dynamic light scattering (DLS) analysis recorded a zeta potential of −13.7 mV (Fig. 1b), indicative of favorable colloidal stability. Furthermore, quantitative ELISA confirmed robust protein loading, with approximately 400 ng of human FGF2 detected per 1 mg of total OMV protein (Fig. 1c).

Fig. 1.

Preparation and physicochemical characterization of FGF2-engineered bacterial outer membrane vesicles (OMV/FGF2).

(a) Hydrodynamic size distribution and absolute particle concentration of OMV/FGF2 determined by nanoparticle tracking analysis (NTA).

(b) Surface zeta potential measurement of OMV/FGF2 via dynamic light scattering (DLS).

(c) Quantitative ELISA analysis of the specific human FGF2 loading capacity within the engineered nanovesicles (analyzed following Triton X-100 membrane lysis to liberate encapsulated contents).

(d) Representative transmission electron microscopy (TEM) and immunogold labeling analyses: (i) unmodified OMVs derived from wild-type E. coli DH5α displaying classic vesicular architecture, (ii) engineered OMV/FGF2 maintaining equivalent structural integrity, (iii) immunogold-labeled wild-type OMVs (negative control confirming antibody specificity), and (iv) immunogold-labeled OMV/FGF2 demonstrating the structural integration of human FGF2 (indicated by localized 10-nm gold nanoparticles). Scale bars: 200 nm (i-iii) and 500 nm (iv).

(e-f) Western blot assay and corresponding densitometric quantification confirming the robust expression of human FGF2 in the engineered nanovesicles.

(g-h) Western blot analysis of outer membrane protein markers OmpF and OmpA in E. coli whole-cell lysates, bacterial culture supernatants, and purified OMV preparations from three independent batches, confirming the specific enrichment of OmpF/OmpA in purified OMV fractions (indicated by the red dashed box). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Morphological assessment via transmission electron microscopy (TEM) with phosphotungstic acid negative staining revealed that OMV/FGF2 maintained a classic spherical vesicular architecture (Fig. 1d–ii), structurally comparable to conventional OMVs derived from the control E. coli DH5α (Fig. 1d–i). Crucially, immunogold labeling which utilizes a rabbit anti-FGF2 primary antibody coupled with a 10-nm gold-conjugated secondary antibody revealed distinct gold particle attachment associated with the OMV/FGF2 (Fig. 1d–iv), whereas conventional OMVs exhibited no such labeling (Fig. 1d–iii). This visually confirms the successful functionalization and structural integration of FGF2 with the OMVs.

Furthermore, Western blot analysis corroborated the robust expression and presence of human FGF2 within the nanovesicles derived from the engineered bacteria (Fig. 1e and f). Importantly, while the presence of endogenous cellular proteins like GAPDH can be natively entrapped within the intra-vesicular matrix during outer membrane blebbing, the loading and normalization of our Western blot tracks were established via the gold standard of equal total protein calibration (40 μg per lane quantified by standard BCA assay) to ensure strict quantitative accuracy. To further validate the molecular identity and purity of our prokaryotic platforms in accordance with standard bacterial vesicle guidelines, Western blotting was extended to profile specific bacterial outer membrane proteins. The data revealed a distinctive enrichment of Outer Membrane Protein F (OmpF) and Outer Membrane Protein A (OmpA), two of the most abundant structural porins natively embedded in the E. coli outer membrane bilayer, across independent production batches, confirming the structural integrity of the isolated vesicles (Fig. 1g and h).

2.2. In vitro cytocompatibility and biosafety profiling of OMV/FGF2

To comprehensively evaluate the cytocompatibility of OMV/FGF2, three representative cell lines were utilized: mouse embryonic fibroblasts (NIH 3T3), human immortalized keratinocytes (HaCaT), and rat small intestinal crypt epithelial cells (IEC-6). Cell Counting Kit-8 (CCK-8) assays revealed that OMV/FGF2, at concentrations exceeding 10 μg/mL, significantly stimulated the proliferation of all three cell lines (Fig. 2a–c). This universal proliferative enhancement indicates both excellent biocompatibility and the successful retention of FGF2 mitogenic bioactivity. Furthermore, in vitro wound healing scratch assays employing NIH 3T3 cells demonstrated that OMV/FGF2 treatment markedly accelerated cell migration compared to both the vehicle control and free FGF2-treated groups (Fig. 2d and e). Hemocompatibility assessments confirmed that OMV/FGF2 induced negligible hemolysis across all tested concentrations (Fig. 2f and g). Importantly, for a 100 μg preparation of OMV/FGF2, the endotoxin concentration was quantified at 4.86 ± 1.23 ng/mL (Fig. 2h), a remarkably low level that falls safely within the acceptable regulatory threshold for oral administration. When calibrated using standard endotoxin conversion metrics, this mass corresponds to approximately 48.6 Endotoxin Units (EU) per dose, which falls safely within the acceptable regulatory limits defined by the U.S. Food and Drug Administration (FDA) Guidance for Industry and United States Pharmacopeia for oral drug administration, where mucosal compartmentalization and oral tolerance margins are drastically wider than parenteral routes.

Fig. 2.

In vitro cytocompatibility, mucosal restitution capacity, and biosafety profiling of OMV/FGF2.

(a-c) Cell viability and proliferation assessments of (a) NIH 3T3 embryonic fibroblasts, (b) human immortalized keratinocytes (HaCaT), and (c) rat intestinal crypt epithelial cells (IEC-6) following exposure to varying concentrations of OMV/FGF2 for respective durations. The black dashed line designates the 100% baseline viability of the vehicle control group.

(d-e) In vitro wound healing (scratch) assay modeling epithelial barrier repair. (d) Representative phase-contrast images depicting the migratory dynamics of IEC-6 epithelial cells following treatment with free FGF2 (50 ng/mL), empty OMVs (10 μg/mL), or OMV/FGF2 (10 μg/mL). (e) Corresponding quantitative analysis of the wound closure rates across different treatment cohorts. Scale bar: 500 μm.

(f-g) Hemocompatibility evaluation. (f) Quantitative hemolysis rates and (g) representative macroscopic images of isolated mouse erythrocytes (RBCs) incubated with varying concentrations of OMV/FGF2 for 2 h at 37 °C.

(h) Absolute endotoxin quantification across independent batches of OMV/FGF2 preparations, determined via a chromogenic LAL assay, confirming adherence to biomedical safety thresholds.

2.3. In vitro stability and in vivo gastrointestinal transit profile of OMV/FGF2

To rigorously evaluate structural stability, OMV/FGF2 was incubated in simulated gastric fluid (SGF, pH 1.2) at 37 °C for 4 h. Nanoparticle tracking analysis (NTA) revealed no significant alterations in either hydrodynamic diameter or particle concentration post-incubation (Fig. 3a–c), demonstrating the robust structural integrity of the nanovesicles under harsh gastric conditions. Subsequent functional assays compared the bioactivity retention of free FGF2 versus OMV-formulated FGF2 across various physiologically relevant conditions (pH 7.4 at 4 °C, pH 7.4 at 37 °C, and a slightly acidic colonic microenvironment simulation at pH 6.5 at 37 °C). After a 24-h incubation, free FGF2 exhibited a near-complete loss of bioactivity; in stark contrast, FGF2 shielded within the OMV architecture remarkably retained its biological activity across all tested conditions (Fig. 3d).

Fig. 3.

In vitro stability, in vivo gastrointestinal transit, and mucosal adherence profiling of orally administered OMV/FGF2.

(a-c) Structural stability of OMV/FGF2 in a highly acidic gastric environment. Equal concentrations of OMV/FGF2 were incubated at 37 °C in simulated gastric fluid (SGF, pH 1.2) or PBS (vehicle control). Temporal monitoring of (a) mean hydrodynamic diameter and (b) absolute particle concentration via nanoparticle tracking analysis (NTA) at indicated time points. (c) Dynamic particle size distribution profiles over the 4-h incubation period.

(d) Functional bioactivity preservation of the FGF2 payload. Comparative ELISA analysis of residual mitogenic activity between free FGF2 and OMV/FGF2 following incubation under physiological storage (pH 7.4, 4 °C), systemic (pH 7.4, 37 °C), and simulated colonic microenvironmental (pH 6.5, 37 °C) conditions for varying durations.

(e-f) Spatiotemporal gastrointestinal biodistribution. (e) Representative ex vivo fluorescence images tracking the gastrointestinal transit of equivalent doses of free Cy5 dye and Cy5-labeled OMV/FGF2 at predetermined time points post-oral gavage, demonstrating rapid systemic clearance of free Cy5 within 8 h and sustained colonic accumulation of intact OMV/FGF2 up to 24 h, and (f) corresponding quantitative temporal analysis of the localized fluorescence intensity of OMV/FGF2.

(g) In situ mucosal targeting. Representative confocal laser scanning micrographs of distal colonic tissue sections harvested 4 h post-oral administration, demonstrating the robust adherence of Cy5-labeled OMV/FGF2 (red) to the colonic epithelium. Nuclei were counterstained with DAPI (blue). Scale bar: 50 μm.

(h) In vitro cellular association. Representative confocal fluorescence images of IEC-6 intestinal crypt epithelial cells following a 4-h co-incubation with Cy5-labeled OMV/FGF2 (red), visualizing the intimate spatial association and adherence of the nanovesicles to the target cells. Nuclei were counterstained with DAPI (blue). Scale bar: 10 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

To delineate the spatiotemporal biodistribution of the nanoplatform, OMV/FGF2 was covalently labeled with a Cy5-NHS ester fluorophore. In vivo fluorescence imaging following oral administration revealed dynamic gastrointestinal transit. The signal initially localized in the stomach at 2 h, transitioned to the small intestine and cecum by 4 h, and exhibited prolonged, specific retention in the colon at 24 h post-administration (Fig. 3e and f). To definitively confirm that the colonic near-infrared signal originated from structurally intact nanovesicles rather than free dissociated dye, an equivalent dose of free Cy5 dye was administered via oral gavage as a rigorous control group. In stark contrast to the persistent 24-h colonic accumulation seen in the OMV/FGF2-Cy5 cohort, the free Cy5 control was rapidly reabsorbed in the upper gastrointestinal tract and eliminated via renal/hepatic clearance within 8 h, showing negligible signal retention in the colon (Fig. 3e). Furthermore, high-resolution confocal microscopy of colonic tissues confirmed that the Cy5-labeled OMVs successfully adhered to the apical surface of the intestinal epithelium and were subsequently internalized by epithelial cells (Fig. 3g and h). This distinct, punctuated intracellular co-localization pattern provides microstructural evidence of energy-dependent vesicle endocytosis, further ruling out the passive molecular diffusion of free, unbound dye components.

2.4. Therapeutic efficacy of OMV/FGF2 in DSS-induced colitis

To evaluate the in vivo translational potential of the nanoplatform, a murine model of acute colitis was established via the administration of 2.5% DSS. Following a 7-day disease induction phase characterized by progressive weight loss, severe hematochezia, and drastically elevated disease activity index (DAI) scores, the mice underwent a 6-day therapeutic regimen. Subjects were stratified into five cohorts: healthy vehicle control, DSS model (PBS), OMV/FGF2 treatment, empty OMV intervention, and free FGF2 intervention. Remarkably, oral administration of OMV/FGF2 spurred rapid macroscopic recovery, evidenced by significant body weight rebound and a steep reduction in DAI scores compared to the DSS-challenged cohort (Fig. 4b–e). Furthermore, colon shortening, a hallmark macroscopic indicator of sustained colonic inflammation, was effectively mitigated in the OMV/FGF2 group, with colon lengths closely approximating those of the healthy controls (Fig. 4a–d). Crucially, intervention with either free FGF2 or empty OMVs yielded substantially inferior therapeutic outcomes. To establish a conservative and highly rigorous benchmark, the free FGF2 control cohort was administered an absolute dose of 100 ng, which is 2.5-fold higher than the absolute payload of structurally integrated FGF2 (∼40 ng) delivered via the 100 μg OMV/FGF2 treatment. The observation that unshielded free FGF2 failed to elicit comparable therapeutic protection, despite being delivered at a 250% higher protein mass, conclusively rules out simple dose-dependent artifacts. This comparative disparity provides definitive proof that free growth factors are biologically neutralized by harsh gastrointestinal transit, and underscores that the robust therapeutic success of our platform relies entirely on the biomimetic protective encapsulation provided by the OMV carrier.

Fig. 4.

In vivo therapeutic efficacy of orally administered OMV/FGF2 in the DSS-induced acute murine colitis model.

(a) Representative macroscopic images of excised whole colons harvested from the respective experimental cohorts on Day 13, visually demonstrating inflammation-induced colon shortening and therapeutic rescue.

(b) Dynamic progression of relative body weight over the entire 12-day experimental duration (comprising a 7-day DSS induction phase followed by a 6-day targeted intervention phase).

(c) Quantitative analysis of final relative body weights across all cohorts on Day 13.

(d) Quantitative measurement of macroscopic colon lengths on Day 13, serving as a primary macroscopic indicator of sustained colonic inflammation.

(e) Temporal monitoring of Disease Activity Index (DAI) scores across all cohorts, systematically calculated based on weight loss, stool consistency, and hematochezia.

(f) Representative hematoxylin and eosin (H&E)-stained microscopic images of distal colonic sections harvested on Day 13, illustrating the extent of mucosal destruction in the vehicle model and robust epithelial restitution following targeted OMV/FGF2 therapy. Scale bar: top row 200 μm,bottom row 20 μm.

(g) Rigorous quantitative histopathological scoring of colonic tissue damage, evaluated based on four standardized criteria: transmural inflammatory cell infiltration (0-3), lesion depth (0-3), crypt structural damage (0-4), and overall lesion extent (1-4).

These macroscopic observations were strongly corroborated by rigorous histopathological evaluation. Hematoxylin and eosin (H&E) staining of the DSS model group unveiled severe colonic mucosal damage, hallmarked by total crypt ablation, extensive epithelial erosion, and massive transmural inflammatory cell infiltration (Fig. 4f). Conversely, treatment with OMV/FGF2 substantially blunted these pathological alterations, promoting robust re-epithelialization, prominent restoration of the glandular crypt architecture, and a marked attenuation of inflammatory infiltrates (Fig. 4f). This profound tissue-level protection was quantitatively reflected in the significantly reduced histological scores for the OMV/FGF2-treated mice (Fig. 4g).

2.5. Restoration of the intestinal physicochemical barrier by OMV/FGF2

The intestinal barrier relies on a highly coordinated defense system comprising a biochemical mucus layer and a physical epithelial network. Covering the apical surface of the intestinal epithelium, the mucus layer serves as the primary frontline defense to fortify gut barrier integrity and is tightly regulated by commensal microbiota. Goblet cells are the predominant mucus-secreting epithelial constituents, reaching their maximum density within the colon. In the PBS-treated DSS model group, histological evaluation revealed severe goblet cell depletion and pathological crypt cavitation. In stark contrast, oral administration of OMV/FGF2 actively reversed this damage, restoring goblet cell populations and stimulating robust mucin secretion, which culminated in a visibly reconstituted and thickened luminal mucus layer.

Beneath this protective biochemical shield, tight junction complexes function as essential gatekeepers regulating paracellular permeability. Occludin serves as a major transmembrane component whose depletion is directly proportional to barrier dysfunction, whereas the scaffolding protein ZO-1 anchors these complexes to the underlying actin cytoskeleton, playing a crucial role in maintaining epithelial homeostasis. Quantitative immunofluorescence analysis demonstrated that OMV/FGF2 treatment significantly rescued the expression of both Occludin and ZO-1 in distal colonic epithelial cells. These expression levels closely paralleled those of the healthy control group and were markedly superior to those observed in the free FGF2 and empty OMV intervention cohorts (Fig. 5a–d). Rather than tight junctions driving proliferation, this robust up-regulation indicates that OMV-delivered FGF2 successfully stimulates mucosal restitution and epithelial proliferation, which subsequently drives the reassembly and consolidation of the tight junction network essential for barrier repair.

Fig. 5.

OMV/FGF2-mediated restoration of the intestinal epithelial physical barrier in DSS-induced colitis.

(a) Representative immunofluorescence (IF) images illustrating the in situ expression and spatial continuous localization of the tight junction transmembrane protein Occludin (red) within the distal colonic epithelium across different experimental cohorts. Nuclei were counterstained with DAPI (blue). Scale bar: 200 μm.

(b) Quantitative densitometric analysis of the mean fluorescence intensity of Occludin, demonstrating significant barrier rescue following OMV/FGF2 intervention.

(c) Representative IF images demonstrating the structural reassembly and expression rescue of the cytoplasmic scaffolding tight junction protein ZO-1 (green) in the colonic mucosal network. Nuclei were counterstained with DAPI (blue). Scale bar: 200 μm.

(d) Corresponding quantitative evaluation of the mean fluorescence intensity of ZO-1. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

2.6. Attenuation of local pro-inflammatory signaling and mucosal immune modulation

Immunohistochemical (IHC) analysis was conducted to evaluate the in situ expression profiles of canonical pro-inflammatory cytokines, such as TNF-α, IL-1β, and IL-6, within the colonic architecture. In the PBS-treated DSS model cohort, intense positive staining for all three mediators was prominently localized across the mucosal and submucosal layers, indicative of a severe, acute localized cytokine storm. Conversely, targeted therapeutic intervention with OMV/FGF2 markedly abrogated both the intensity and spatial distribution of these pro-inflammatory signals, restoring expression patterns to basal levels akin to those of the healthy control group (Fig. 6a and b).

Fig. 6.

Attenuation of local pro-inflammatory cytokine expression and mucosal immune modulation following oral OMV/FGF2 therapy.

(a) Representative immunohistochemical (IHC) images demonstrating the in situ expression profiles and spatial localization of canonical pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) within the distal colonic mucosal tissues across the respective experimental cohorts on Day 13. Brown staining indicates positive target expression; nuclei were counterstained with hematoxylin (blue). Scale bar: 20 μm.

(b) Corresponding quantitative densitometric analysis evaluating the relative IHC staining intensity (Average Optical Density, AOD) for TNF-α, IL-1β, and IL-6, confirming the significant abrogation of localized cytokine storms by OMV-delivered FGF2. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Rigorous quantitative analysis of the IHC signals corroborated these histological observations, confirming that OMV/FGF2 administration yielded significantly lower cytokine expression compared to the DSS vehicle group, as well as the free FGF2 and empty OMV treatment cohorts (Fig. 6b). The stark inability of orally administered free FGF2 or empty OMVs to effectively suppress this inflammatory cascade further validates our underlying therapeutic rationale: robust mucosal immune modulation is strictly contingent upon the OMV-shielded delivery of biologically active FGF2 to the inflamed colonic microenvironment.

2.7. Remodeling of gut microbiota architecture and reversal of dysbiosis

To elucidate whether the OMV/FGF2-mediated restoration of the mucosal barrier translates into the re-establishment of the intestinal microecology, 16S rRNA gene sequencing was performed on colonic contents. Species accumulation curves rapidly plateaued, validating that the sequencing depth was fully adequate to capture the microbial diversity within the samples (Fig. 7a). Subsequent alpha diversity analysis demonstrated that targeted OMV/FGF2 therapy successfully restored gut microbiota richness (Chao1 and Ace indices) to healthy control levels, while maintaining overall community diversity (Shannon and Simpson indices). This alignment suggests that as the physical-chemical barrier returns to a healthy baseline, the local microenvironment synchronously guides the gut microbiota back toward homeostatic equilibrium. (Fig. 7b).

Fig. 7.

OMV/FGF2-mediated structural remodeling of the gut microbiota and reversal of DSS-induced microbial dysbiosis.

(a) Species accumulation curves based on 16S rRNA gene sequencing of colonic contents, validating that the sequencing depth was fully sufficient and saturated to capture the comprehensive microbial diversity within all samples.

(b) Comprehensive Alpha diversity analysis evaluating gut microbial community richness (Ace and Chao1 indices) and species diversity/evenness (Shannon and Simpson indices) across the respective experimental cohorts.

Taxonomic profiling further revealed that OMV/FGF2 administration profoundly reprogrammed the microbial community structure. Notably, the DSS-induced expansion of Proteobacteria, a hallmark phylum canonically associated with mucosal inflammation and dysbiosis, was significantly curtailed following OMV/FGF2 treatment (Fig. 8a and b).

Fig. 8.

Taxonomic profiling and identification of differential microbial biomarkers following OMV/FGF2 intervention.

(a-b) Structural distribution and relative abundance of the colonic microbial communities evaluated at the (a) phylum and (b) genus taxonomic levels across the different experimental cohorts.

(c-d) Linear discriminant analysis Effect Size (LEfSe) analysis characterizing the microbial shifts. (c) LDA score bar plot identifying distinct microbial biomarkers significantly enriched in each respective group (threshold: LDA score >2.0) (d) Taxonomic cladogram illustrating the phylogenetic distribution and hierarchical relationship of these differentially abundant bacterial lineages.

(e-h) Quantitative comparison of the relative abundances of key therapeutic and pathogenic taxa validated by the LEfSe analysis: (e) the SCFA-producing order Lachnospirales, (f) the family Butyricococcaceae, (g) opportunistic pathogenic family Tannerellaceae, and (h) colitis-associated order Coriobacteriales.

Conversely, the therapeutic regimen orchestrated a robust enrichment of beneficial, commensal taxa. Compared to the DSS model group, OMV/FGF2 significantly increased the relative abundance of core beneficial genera, including Bacteroides, Alistipes, Lactobacillus, and Dubosiella (Fig. 8c and d). At the species level, the populations of Bacteroides acidifaciens and Bacteroides thetaiotaomicron were distinctively elevated. Furthermore, OMV/FGF2 intervention promoted the proliferation of critical short-chain fatty acid (SCFA)-producing taxa, specifically the order Lachnospirales (Fig. 8e) and the family Butyricococcaceae (Fig. 8f). Concurrently, the abundance of opportunistic pathogenic family Tannerellaceae and colitis-associated order Coriobacteriales were markedly diminished in the treated mice (Fig. 8g and h). Collectively, these dynamic compositional shifts indicate that the robust epithelial repair driven by OMV/FGF2 reshapes the local microenvironment, effectively reversing microbial dysbiosis by suppressing opportunistic pathogens and fostering a homeostatic, SCFA-enriched ecosystem. Additionally, to investigate the reciprocal feedback loop existing within the gut microenvironment, we evaluated whether SCFA commensal metabolites could cooperatively enhance the therapeutic performance of OMV/FGF2, because FGF2 does not directly act on bacteria. Supplementary in vitro CCK-8 proliferation and wound healing scratch assays were conducted on serum-starved IEC-6 cells across four cohorts: PBS control, SCFA monotherapy (0.5 mM sodium butyrate), OMV/FGF2 monotherapy (10 μg/mL), and OMV/FGF2 + SCFA combination treatment (Fig. S1). Notably, the OMV/FGF2 + SCFA combination group exhibited significantly higher cell proliferation rates and accelerated scratch migration velocities compared to any of the single-treatment groups. This robust synergistic response demonstrates that while the nanoplatform drives mucosal repair to bloom SCFA-producing taxa in vivo, the resulting butyrate subsidy acts back on the host epithelium to provide the essential energetic fuel and auxiliary signals required to optimize FGF2/FGFR-mediated tissue restitution (Fig. S1).

3. Discussion

Inflammatory bowel diseases is a refractory inflammatory disorder characterized by severe epithelial barrier disruption and progressive microbial dysbiosis [[15], [16], [17]]. While the regenerative potential of fibroblast growth factors (e.g., FGF2) in promoting mucosal healing is well documented [10], their clinical translation for gastrointestinal pathologies is severely bottlenecked by rapid proteolytic degradation and short physiological half-lives during oral transit [11,18]. In this study, we engineered a biomimetic nanotherapeutic platform, OMV/FGF2 that cleverly repurposes the innate biological secretion processes of Gram-negative bacteria [19]. By utilizing naturally derived OMVs as structural shields [20,21], this strategy successfully circumvents the gastrointestinal delivery barriers, achieving targeted mucosal restitution and secondary microecological remodeling in a murine model of DSS-induced colitis. This approach aligns with a broader contemporary paradigm shift in biomaterial science that leverages multi-target hybrid networks and advanced signaling modulation to treat inflammatory bowel diseases. For instance, recent pioneering strategies have successfully combined prebiotic matrices with therapeutic gas-releasing components to re-engineer the colonic niche, such as oral dual-targeting inulin-based nanomicelles leveraging H₂S-mediated immunomodulation [22], as well as polyoxometalate and manganese oxide-loaded nanozymes optimized for reactive oxygen species (ROS) scavenging and controlled carbon monoxide (CO) gas release [23,24]. In parallel, the field has increasingly recognized the structural and immunomodulatory prominence of utilizing biological "nanofactories" over conventional synthetic alternatives, exemplified by engineered probiotic E. coli-derived OMVs carrying antioxidant payloads [25]. Our OMV/FGF2 system builds complementarily upon these advanced frameworks, focusing heavily on biological morphogen-mediated tissue regeneration to secondary drive upstream microenvironmental success.

The fundamental advantage of the OMV/FGF2 platform lies in its highly evolved structural composition. Unlike synthetic lipid nanoparticles or polymeric nanocarriers, which frequently encounter issues related to long-term toxicity, low encapsulation efficiency, and complex scale-up manufacturing [[12], [13], [14]], OMVs are naturally endowed with a robust, highly stable lipid bilayer [26,27]. Our in vitro simulated gastric fluid assays confirm that this vesicular architecture provides formidable shielding, effectively preserving the mitogenic bioactivity of the FGF2 payload against the highly acidic and proteolytic gastric environment. Upon reaching the colon, the innate mucoadhesive properties of OMVs facilitate prolonged retention and subsequent internalization by the damaged intestinal epithelium [28], ensuring maximal localized therapeutic concentration while precluding the adverse effects typically associated with systemic growth factor exposure. In mucosal biomaterial science, it is widely established that the intestinal mucus layer is highly negatively charged due to the abundance of sialic acid and sulfate groups on mucin glycans. Consequently, the net surface charge of a nanocarrier represents a critical determinant governing its transport kinetics. While a strong positive surface charge induces severe polyvalent electrostatic trapping within the superficial luminal mucus layer leading to rapid peristaltic clearance, our engineered OMV/FGF2 exhibits a moderate negative surface potential of −13.7 mV. In modern mucosal fluid dynamics, this weak negative window acts as a lubrication effect, generating a subtle electrostatic repulsive force against the mucin network that is strong enough to prevent immobilizing adhesive entrapment. This property allows the nanovesicles to efficiently slip through the steric pores of the macro-mucus gel via rapid Brownian diffusion to achieve deep penetration. Once through, its weak negative charge facilitates selective electrostatic binding to exposed positive pathological protein clusters (such as transferrin and cell-debris debris) that characteristically accumulate at the ulcerated epithelial boundary in acute colitis, validating the robust colonic retention observed in our assays.

Mechanistically, our findings correct the often-oversimplified paradigm of barrier repair. Rather than directly acting on tight junction proteins, the targeted delivery of biologically active FGF2 effectively engages local epithelial fibroblast growth factor receptors (FGFRs), stimulating rapid cellular proliferation and wound closure [29]. It is this primary re-epithelialization that subsequently drives the robust reassembly of the structural tight junction network (characterized by upregulated Occludin and ZO-1) and the functional restoration of goblet cell-mediated mucin secretion [[30], [31], [32], [33], [34], [35]].

Crucially, this structural and biochemical mucosal recovery serves as the fundamental prerequisite for reversing microbial dysbiosis through a mechanism we define as microenvironment-driven ecological niche remodeling [36,37]. OMV/FGF2 does not function as a direct antimicrobial agent; instead, by sealing the leaky gut and restoring the protective mucin layer, it systematically manipulates the physical and biochemical ecological factors of the gut lumen. First, severe colitis causes intense tissue erosion and capillary exposure, causing abnormal oxygen leakage into the colonic lumen. This pathological "oxygen drift" drives the blooms of facultative anaerobic pathogens (such as Proteobacteria). By triggering intense re-epithelialization, OMV/FGF2 seals the tissue and restores the strictly hypoxic microenvironment, excluding oxygen-tolerant pathogens and allowing strictly anaerobic commensals to repopulate. Second, our platform strongly promotes goblet cell recovery and mucin hypersecretion. Host mucus O-glycans serve as a highly precise ecological nutrient source for specific beneficial taxa. For instance, Bacteroides thetaiotaomicron which was profoundly and specifically enriched in our treated group possesses an expansive genome uniquely encoding specialized Starch Utilization Systems (Sus) and carbohydrate-active enzymes dedicated to harvesting and cleaving host mucus glycans. The robust reconstitution of the biochemical mucus barrier thus creates a highly selective nutrient niche, providing a distinct competitive colonization advantage that fuels the specific bloom of B. thetaiotaomicron.

Importantly, our data confirmed that this ecological succession triggers a highly coordinated, non-linear reciprocal feedback loop. The enriched commensal cohorts, specifically within the order Lachnospirales and family Butyricococcaceae, are well-documented producers of SCFAs, primarily butyrate which may nourish colonocytes and promote regulatory T-cell differentiation, thereby reinforcing the resolution of localized inflammation [38,39]. In addition, Luminal butyrate serves as the primary energetic fuel for colonic epithelial cells, driving mitochondrial beta-oxidation. Through in vitro CCK-8 and wound healing scratch assays, we demonstrated that the concurrent administration of OMV/FGF2 and sodium butyrate induced significantly higher cell proliferation rates and accelerated scratch migration velocities in IEC-6 cells compared to any monotherapy. This empirical synergy provides absolute logical closure to the host-microbiome interplay: while OMV/FGF2 restores the physical-chemical tissue matrix to expand the niche for SCFA-producing commensals, the resulting bacterial butyrate subsidy acts reciprocally back on the host epithelium. This metabolic subsidy provides the immense cellular energy and auxiliary signaling required to sustain high-velocity, FGF2/FGFR-mediated mucosal restitution and barrier consolidation.

Despite the compelling in vivo efficacy demonstrated in this study, several critical limitations and translational hurdles warrant rigorous future investigation. First, the precise spatiotemporal kinetics of OMV disassembly and the molecular pathways governing FGF2 release and subsequent FGFR activation within the colonic crypts remain to be comprehensively mapped at the ultrastructural level. Second, because these bioengineered nanovesicles are natively derived from a Gram-negative bacterial cloning strain, they inherently retain LPS within their outer membranes. In the microenvironment of acute colitis where the epithelial barrier is compromised and leaky, introducing exogenous endotoxins presents a potential risk of aggravating localized TLR4 signaling. However, our safety profile and net therapeutic anti-inflammatory outcome are supported by a rigorous risk-benefit equilibrium. Our optimized multi-step ultrafiltration and diafiltration isolation protocol strictly controlled the endotoxin concentration at an ultra-low level, which stands orders of magnitude below the endogenous colonic endotoxin load naturally shed by the trillions of Gram-negative commensals residing in the gut lumen. The oral route of administration leverages unique mucosal compartmentalization and oral tolerance margins that restrict systemic transmigration. The E. coli DH5α strain possesses a modified, low-immunogenicity Lipid A profile with attenuated endotoxic activity compared to pathogenic strains. Most crucially, the high-velocity kinetic benefit of OMV/FGF2 in rapidly sealing the leaky gut via re-epithelialization and tight junction consolidation systematically minimizes the translocation windows for both endogenous luminal pathogens and its own vesicle-bound traces, meaning the structural benefit of barrier restoration vastly outweighs transient localized endotoxin exposure.

Furthermore, a primary biosafety concern for laboratory-scale bioengineering is that our expression host relies on an ampicillin resistance cassette for selection. The potential risk of horizontal gene transfer (HGT) of antibiotic resistance genes to the native gut microbiota represents a significant clinical barrier that must be eliminated prior to human translation. To achieve safe clinical scale-up, our future directions will focus on transitioning to antibiotic-free plasmid maintenance systems, such as essential gene complementation systems. Alternatively, we will employ new recombineering technologies to stably integrate the human FGF2 expression cassette directly into a safe-harbor locus on the host bacterial chromosome. This permanent genomic integration eliminates episomal plasmid shedding and completely aborts the structural pathways of horizontal gene transfer via plasmid dissemination, satisfying stringent clinical safety benchmarks. Ultimately, before this promising bioengineered platform can transition to clinical trials, it is imperative to validate its pharmacokinetic profile, long-term biosafety, and therapeutic efficacy in rigorous large animal models, such as porcine models, which possess gastrointestinal anatomies and immunological responses far more analogous to those of humans.

In conclusion, OMV/FGF2 represents a highly translatable, orally bioavailable nanoplatform that bridges synthetic biology and mucosal immunology. By providing a structural sanctuary for FGF2, this system initiates a synergistic therapeutic cascade to drive profound epithelial regeneration, which in turn orchestrates the restorative remodeling of the gut microbiome. This dual-mechanism strategy not only addresses the core pathophysiological defects of IBD but also establishes a versatile foundational technology for the targeted oral delivery of highly labile therapeutic biologics.

4. Methods

4.1. Materials and reagents

Dextran sulfate sodium (DSS, MW 36,000-50,000 Da) was procured from MP Biomedicals (USA). The human FGF2 ELISA kit and the chromogenic Limulus amebocyte lysate (LAL) endotoxin assay kit were obtained from Solarbio (China). The BCA protein assay and Cell Counting Kit-8 (CCK-8) were purchased from Beyotime (China) and GlpBio (USA), respectively. Simulated gastric fluid (SGF, pH 1.2) was prepared in accordance with standard pharmacopeial protocols. The Cy5-NHS ester fluorophore was sourced from Macklin (China). Primary antibodies against FGF2 (Bioss, China; Cat. No. bs-0217R), Occludin(Affinity Biosciences, USA; Cat. No. AF1556)and ZO-1 (Affinity Biosciences, USA; Cat. No. AF5145), alongside TNF-α, IL-1β, and IL-6 (Abcam, UK; Cat. No. ab9739 No.ab9722 No.ab6672), were utilized. Corresponding HRP-conjugated goat anti-rabbit IgG and Alexa Fluor® (488/594)-conjugated secondary antibodies were provided by Abcam. All auxiliary chemicals were of analytical reagent grade.

4.2. Bacterial strains and animal husbandry

The engineered E. coli DH5α strain, harboring a recombinant plasmid encoding human FGF2 and an ampicillin resistance cassette, was custom-constructed by the National Engineering Research Center for Protein Therapeutics (Wenzhou Medical University, China) [40]. Wild-type E. coli DH5α served as the isogenic control for unmodified OMV isolation. Male C57BL/6J mice (6-7 weeks, ∼20 g) were procured from Vital River Laboratory Animal Technology Co., Ltd. (Zhejiang, China). Animals were housed in a specific pathogen-free (SPF) facility (23 ± 2 °C, 55 ± 10% humidity, 12 h light/dark cycle) with ad libitum access to standard chow and water. All in vivo protocols were rigorously reviewed and approved by the Animal Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University.

4.3. Isolation and purification of OMV/FGF2

Engineered E. coli DH5α was inoculated into 10 mL of Luria-Bertani (LB) broth (supplemented with 100 μg/mL ampicillin) and cultured overnight at 37 °C under 200 rpm agitation. The seed culture was subsequently expanded (1:20 dilution) in 1 L of fresh selective LB broth for 16 h. Intact bacterial cells were pelleted via dual low-speed centrifugation steps (6000 × g, 20 min, 4 °C). The resulting supernatant was clarified through a 0.45 μm vacuum filtration membrane to deplete residual cellular debris. The clarified filtrate was concentrated utilizing 100 kDa molecular weight cut-off (MWCO) centrifugal filter units (Millipore, USA) at 4000 × g for 15 min, followed by exhaustive diafiltration against cold PBS (pH 7.4) to remove soluble low-molecular-weight contaminants. The retentate was then subjected to two cycles of ultracentrifugation (150,000 × g, 2 h, 4 °C) to precipitate the OMVs. The highly purified OMV/FGF2 pellet was resuspended in PBS, terminally sterilized through a 0.22 μm filter, aliquoted into single-use fractions, and stored at −80 °C. Empty OMVs were identically prepared from the wild-type strain.

4.4. Physicochemical characterization

Hydrodynamic diameter and particle concentration were quantitatively evaluated using Nanoparticle Tracking Analysis (NTA) on a ZetaView analyzer (Particle Metrix, Germany; 520 nm laser, shutter 10, sensitivity 80) following a 1:1000 sample dilution in PBS. Surface zeta potential was determined via dynamic light scattering (DLS) utilizing a Zetasizer Nano ZS (Malvern Instruments, UK). For FGF2 loading quantification, OMV/FGF2 vesicles were completely lysed with 0.1% Triton X-100 to release internal contents, followed by ELISA analysis normalized to total BCA protein content. Ultrastructural morphology was assessed via transmission electron microscopy (TEM; Talos L120C, Thermo Fisher, 80 kV) using 1% (w/v) phosphotungstic acid negative staining. For immunogold mapping, OMV-adsorbed grids were sequentially probed with anti-FGF2 primary antibody (1:100, 2 h) and a 10 nm gold-conjugated secondary antibody (1:50, 1 h). Protein expression was validated via Western blotting: denatured samples (40 μg total protein) were resolved on 12% SDS-PAGE, transferred to 0.45 μm PVDF membranes, blocked in 5% non-fat milk, and probed with anti-FGF2 (1:1000, overnight, 4 °C) and HRP-secondary antibody (1:5000, 1 h). Bands were visualized utilizing ECL substrate (ChemiDoc, Bio-Rad) and quantified via ImageJ. Crucially, to comply with standardized quantitative practices for extracellular vesicles, Western blot normalization was achieved using the gold standard method of Equal Total Protein Loading. Prior to electrophoresis, total protein content across all intact OMV formulations was systematically quantified via the BCA assay, and an identical mass of 40 μg total protein was loaded into each respective lane, ensuring that payload identification remained entirely independent of housekeeping protein variations.

For the identification of specific prokaryotic outer membrane markers, E.coli whole-cell samples were lysed using Iron Hammer Super Bacterial Lysis Buffer (ACE Cat No. BR0005-02). Protein samples from purified OMVs and bacterial culture supernatants were used directly without additional lysis treatment. Protein concentration was quantified via BCA assay, and equal amounts of protein (20 μg per lane) were separated by 10% SDS-PAGE and electrotransferred onto PVDF membranes. After blocking with 5% skim milk for 1.5 h at room temperature, membranes were incubated with primary antibodies against OmpF (biorbyt, Cat No.orb308741, 1:1000) and OmpA (biorbyt, Cat No.orb862303, 1:1000) overnight at 4 °C. Membranes were then washed and incubated with HRP-conjugated secondary antibodies (1:5000) for 1 h at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) reagent.

4.5. In vitro cytocompatibility and biosafety

NIH 3T3, HaCaT, and IEC-6 cell lines (ATCC, USA) were maintained in high-glucose DMEM supplemented with 10% FBS and 1% penicillin-streptomycin (37 °C, 5% CO₂). Cell proliferation was evaluated by exposing seeded cells (4 × 10⁴ cells/mL) to OMV/FGF2 (0-20 μg/mL) for 24-72 h, followed by CCK-8 absorbance measurements at 450 nm. Mucosal restitution was modeled using an in vitro scratch assay: confluent IEC-6 monolayers (in Ibidi inserts) were wounded and treated with free FGF2 (50 ng/mL), empty OMVs (10 μg/mL), or OMV/FGF2 (10 μg/mL), with migratory closure quantified over 24 h via ImageJ. Hemocompatibility was ascertained by incubating purified mouse erythrocytes (2% v/v in PBS, isolated at 800 × g) with OMV/FGF2 (5-200 μg/mL) for 2 h at 37 °C, followed by spectrophotometric quantification of supernatant hemoglobin at 545 nm. Preparation endotoxin levels were rigorously verified utilizing the LAL chromogenic assay.

4.6. Stability and in vivo gastrointestinal biodistribution

Gastric structural stability was confirmed by incubating OMV/FGF2 (100 μg/mL) in SGF (pH 1.2) or PBS at 37 °C, with NTA monitoring at 0, 1, 2, and 4 h. Functional bioactivity was assessed by incubating free FGF2 (50 ng/mL) or equivalent OMV/FGF2 at physiological storage (pH 7.4, 4 °C), systemic (pH 7.4, 37 °C), and simulated colonic (pH 6.5, 37 °C) conditions for up to 24 h, followed by ELISA quantification. For biodistribution, OMVs were covalently labeled with Cy5-NHS ester (1:5 M ratio, 0.1 M NaHCO₃, pH 8.5) and extensively dialyzed (MWCO 3500 Da). Mice (n = 3/timepoint) were orally gavaged with Cy5-labeled OMV/FGF2 (100 μg protein) or an equivalent dose of free Cy5 dye, and gastrointestinal transit was spatiotemporally mapped utilizing an IVIS Lumina system (PerkinElmer) under isoflurane anesthesia. Epithelial adherence was verified by incubating IEC-6 cells or ex vivo colonic sections with Cy5-OMVs (4 h), counterstaining with DAPI, and capturing target interactions via confocal microscopy (Nikon).

4.7. DSS-induced colitis model and histopathological evaluation

Acute colitis was induced in C57BL/6J mice (n = 6/cohort) via 2.5% (w/v) DSS in drinking water for 7 days. During the subsequent 6 days therapeutic phase (standard water), cohorts received daily oral gavages (200 μL) of PBS, OMV/FGF2 (100 μg protein), empty OMVs (100 μg protein), or free FGF2 (100 ng). The 100 ng dose of free FGF2 was purposely chosen to establish a rigorous, conservative baseline benchmark, representing a 2.5-fold higher absolute protein payload than the structurally integrated FGF2 (∼40 ng) delivered via the 100 μg OMV/FGF2 complex. Disease Activity Index (DAI) was comprehensively scored based on weight loss, stool consistency, and hematochezia. On Day 13, excised distal colons were fixed, paraffin-embedded, and sectioned (4 μm). H&E-stained sections underwent blinded histological scoring (0-14 scale) evaluating inflammation, lesion depth, and crypt damage. For in situ target validation, immunohistochemistry (TNF-α, IL-1β, IL-6; 1:200) and immunofluorescence (Occludin, ZO-1; 1:200) were executed utilizing heat-induced antigen retrieval, 5% BSA blocking, and appropriate HRP/DAB or Alexa Fluor® 594 detection systems. To satisfy quantitative reporting rigor, relative expression profiles for Occludin and ZO-1 immunofluorescence were evaluated by measuring Mean Fluorescence Intensity (MFI) via ImageJ software (NIH, USA), with baseline signals normalized to the total surface area of the intestinal epithelium. Sampling stringency was enforced by collecting and processing 5 random, non-overlapping high-magnification fields per animal under uniform, fixed confocal laser and acquisition parameters (n = 6 mice per group, totaling 30 fields analyzed per experimental cohort).

4.8. 16S rRNA microbiome profiling

Total genomic DNA was extracted from sterile colonic contents utilizing the QIAamp PowerFecal Pro Kit (Qiagen). The V3-V4 hypervariable region was amplified (primers 338F/806R) and sequenced on an Illumina NovaSeq 6000 platform. Bioinformatic processing entailed FLASH-mediated sequence merging, UCHIME chimera removal, and UPARSE operational taxonomic unit (OTU) clustering at a 97% similarity threshold. Taxonomy was assigned via the Silva database (v138). Community richness and diversity were evaluated using QIIME2 (Ace, Chao1, Shannon, Simpson), and distinct microbial biomarkers were statistically identified utilizing Linear Discriminant Analysis Effect Size (LEfSe).

4.9. In vitro synergistic mucosal restitution and cell proliferation assays

To evaluate the reciprocal feedback loop between bacterial metabolites and the nanoplatform, the synergistic therapeutic efficacy of SCFAs and OMV/FGF2 was assessed via wound healing scratch and CCK-8 proliferation assays. Rat small IEC-6 were utilized as the in vitro model. For the epithelial migration assay, IEC-6 cells were seeded into 6-well plates and cultured in complete high-glucose DMEM until reaching a confluent monolayer. A cell-free clearance zone was introduced in each well by creating a uniform scratch with a sterile 200 μL pipette tip, followed by washing twice with sterile PBS to systematically remove cellular debris. To eliminate confounding effects from serum-induced mitogenesis, the medium was replaced with a low-serum formulation (1% FBS DMEM), and cells were randomly allocated into four distinct experimental cohorts: a Control Group treated with an equivalent volume of vehicle PBS, an SCFA Monotherapy Group treated with 0.5 mM sodium butyrate (AmBeed, Cat. No. A239433), an OMV/FGF2 Monotherapy Group treated with 10 μg/mL of OMV/FGF2, and an OMV/FGF2 + SCFA Combination Group treated concurrently with 10 μg/mL OMV/FGF2 and 0.5 mM sodium butyrate. Bright-field micrographs of the identical denuded scratch regions were captured at 0, 6, 12, and 24 h post-scratching using an inverted phase-contrast microscope, and quantitative image analysis of the remaining denuded area was executed using ImageJ software (NIH, USA) to calculate the migration rate utilizing the following formula: (initial scratch area − remaining scratch area at each timepoint)/initial scratch area × 100%. For the cell proliferation and viability analysis, IEC-6 cells were seeded in 96-well plates at an initial density of 5 × 10³ cells per well, allowed to adhere overnight, and subjected to serum starvation for 12 h prior to receiving the identical four treatment regimens (n = 5 independent replicate wells per group). At designated temporal intervals (0, 6, 12, and 24 h), 10 μL of CCK-8 reagent (GlpBio, USA) was added to each target well, followed by an incubation period of 1.5 h at 37 °C in the dark. The optical density (OD) was measured at an absorbance wavelength of 450 nm using a microplate reader, and absolute cell viability percentages were calculated and normalized against the 0-h baseline vehicle control group.

4.10. Statistical analysis

Quantitative data (n ≥ 3 independent replicates) are presented as mean ± standard deviation (SD) and were analyzed utilizing GraphPad Prism 9.0. Pairwise and multiple comparisons were evaluated using an unpaired, two-tailed Student's t-test and one-way ANOVA followed by Tukey's post-hoc test, respectively. Statistical significance was designated as ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001.

CRediT authorship contribution statement

Meilin Yi: Methodology, Writing – original draft. Jiayi Luo: Data curation, Methodology, Writing – original draft. Ebrahim Abdo: Methodology, Writing – original draft. Zuyao Lu: Methodology. Xin Pan: Investigation, Validation. Xu Han: Investigation, Validation. Xinyi Sun: Methodology. Yuting Xia: Methodology. Juqin Dai: Methodology. Keqing Shi: Conceptualization, Funding acquisition, Investigation, Writing – review & editing. Zimiao Chen: Conceptualization, Funding acquisition, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

List of abbreviations

CCK-8

Cell Counting Kit-8l

DSS

Dextran Sulfate Sodium

E. coli

Escherichia coli

FGF2

Fibroblast Growth Factor 2

IL-1β

Interleukin-1 beta

IL-6

Interleukin-6

OMV

Outer Membrane Vesicles

TEM

Transmission Electron Microscopy

TNF-α

Tumor Necrosis Factor-alpha

ZO-1

Zonula Occludens-1.

Appendix A. Supplementary data

The following is the supplementary data to this article:

Fig. S1.

Short-chain fatty acid (SCFA, sodium butyrate) reciprocally enhances the reparative efficacy of FGF2-engineered outer membrane vesicles (OMV/FGF2) in intestinal epithelial cells

(a) Representative bright-field images of wound healing scratch assay in IEC-6 intestinal crypt epithelial cells, displaying the dynamic scratch closure process over 24 h across four experimental groups: PBS (blank control), SCFA (sodium butyrate) monotherapy, OMV/FGF2 monotherapy, and OMV/FGF2 + SCFA combination therapy.

(b) Quantitative temporal analysis of cell migration ratio, demonstrating accelerated wound closure in the OMV/FGF2 + SCFA combination group compared to all single-treatment groups, confirming synergistic pro-migratory effects.

(c) Cell viability measured by CCK-8 assay at 0, 6, 12, and 24 h post-treatment, showing that OMV/FGF2 + SCFA co-treatment induces significantly higher intestinal epithelial cell proliferation than individual treatments.

Data availability

Data will be made available on request.