Abstract Microcapsules composed of synthetic polymeric matrices have attracted considerable attention in delivering oral probiotics. However, existing polymeric microcapsules demonstrate inadequate acid resistance and adaptability, as well as deficiency in the inflamed colon-specificity and uncontrolled release of probiotics therein. Herein, a DNA microcapsule is prepared as a probiotic-transporting micromotor through photo-crosslinking of hyaluronic acid methacrylate and acrydite-modified A-/C-rich oligomers within the microfludically generated droplets in the presence of nitric oxide-cleavable crosslinker and gas donor manganese carbonyl (MnCO). As the microcapsules traverse stomach, duodenum, and ultimately colon, the formation and dissociation of A-motif and i-motif structures instigate a reversible shrinking-swelling transition of microcapsules to preserve probiotic viability. Subsequently, the microcapsules exhibit chemotaxis towards inflamed colon site, driven by a gas-generating reaction between MnCO and elevated reactive oxygen species. Following disintegration of the microcapsules, triggered by endogenous nitric oxide, probiotics are released to reshape the dysbiosis of intestinal microflora. This advanced delivery system offers significant promise for the effective clinical management of inflammatory bowel disease. Subject terms: Biomedical engineering, DNA nanotechnology, Bioinspired materials, Nanobiotechnology __________________________________________________________________ Oral delivery of probiotics faces challenges in survival and targeting. Here, the authors report on acid-resistant chemotactic DNA micromotors which protect probiotics in the gastrointestinal tract, and tissue microenvironmental responsiveness for micromotor targeting to the inflamed colon and triggered release of probiotics. Introduction Inflammatory bowel disease (IBD) is a type of chronic inflammatory disorders, which severely impacts on the daily lives of millions and elevates the risk of bowel cancer^[28]1–[29]3. The primary pathological features of IBD include imbalance of the intestinal microbiota^[30]4,[31]5, dysregulation of immune homeostasis^[32]6–[33]8, and dysfunction of intestinal epithelial barrier^[34]9,[35]10, and they are interlinked and collectively contribute to the exacerbation of IBD. Oral probiotics, a class of biotherapeutics, inhabit the intestinal tract and exert beneficial effects on health by altering the composition of the microbial homeostasis, producing antimicrobial metabolites, and modulating immune responses, which have rapidly gained acceptance as a potent therapeutic modality for IBD^[36]11–[37]13. However, oral probiotics are inherently vulnerable to the hostile acidic environment of gastrointestinal (GI) tract and are deficient in the site specificity within the colon, leading to their diminished bioavailability and limited colonization in the site of inflammation, and ultimately compromising their therapeutic efficacy^[38]14. To improve the acid tolerance of probiotics, various encapsulation strategies have been established to create a physical barrier between the probiotics and the adverse conditions of the GI tract^[39]15–[40]17. Of particular interest is the encapsulation of probiotics within microspheres made of polymeric matrices with a moderate crosslinking degree. Such networks provide a natural extracellular polymeric substance (EPS)-like environment to maintain the bioactivity and proliferation of probiotics, as well as additional possibilities for biofunctionalization^[41]18–[42]20. However, to date, the existing probiotic microspheres exhibit suboptimal probiotic survival rates, primarily due to the inadequate acid resistance and limited acid adaptability of conventional polymeric scaffolds. Although a higher crosslinking degree in the polymeric networks of microcapsules better restricts the intrusion of H^+ ions into the encapsulated probiotics, this persistently high crosslinking degree concurrently hinders essential substance exchange during metabolism and impairs the release rate of probiotics in the colon^[43]21,[44]22. Therefore, there is a significant impetus to develop probiotic microcapsules with an acid-adaptive crosslinking degree. Specifically, these microcapsules achieve a high crosslinking degree exclusively in the acidic GI tract, while then revert to a low degree at the target site of the mildly basic colon—a design that remains unexplored. Furthermore, to confer the inflamed colon-specificity of polymeric microcapsules, it is essential for the microcapsule matrix to exhibit chemotactic behavior, responding to the specific chemical signals (e.g., reactive oxygen species (ROS), inducible nitric oxide synthase (iNOS), etc.) elicited by host cells or immune responses at inflamed sites^[45]23–[46]25. In this context, the development of chemotactic micromotors as probiotic microcapsules is anticipated to facilitate the navigation of probiotics towards the inflamed regions, from which in turn the crosslinked networks ideally disintegrate for on-demand release of probiotics, thereby enabling their effective localization and colonization therein. Interestingly, when the products of reductive gases (e.g., carbon monoxide (CO)^[47]26, hydrogen sulfide (H[2]S)^[48]27, hydrogen gas (H[2])^[49]28 etc.) are utilized as driving forces for chemotaxis, they serve additionally as potent modulators of the immune responses. While previous studies have explored different micro- and nanomotors for the treatment of diseases such as arthritis^[50]29, myocardial ischemia-reperfusion^[51]30, and tumor^[52]31 through the generation of reductive gases, neither the development of chemotactic micromotors with adequate acid resistance or acid-adaptive crosslinking, nor their use as probiotic-transporting microcapsules targeting the inflamed colon, nor the responsive release of probiotics from these micromotors has been explored. Herein, we employed a droplet microfluidic technology^[53]32 to generate uniform water-in-oil (W/O) droplets (Fig. [54]1a). Within these droplets, hyaluronic acid methacrylate (HAMA) and acrydite-modified A-/C-rich oligomers were photo-crosslinked into homogeneous acid-resistant DNA microcapsules. When these microcapsules reached stomach (pH ~ 2) and duodenum (pH ~5) successively, the A- and C-rich sequences formed stable A-motif and i-motif structures, respectively (Fig. [55]1b). In both stomach and duodenum, the additional DNA motifs induced a higher corsslinking degree and apparent microcapsule contraction in response to acidic pHs, creating a dense physical barrier between EcN and GI fluid. In this context, the microcapsules preserved EcN viability until they reached the colon (pH > 7), where the DNA motifs dissociated into individual chains, resulting in the restored corsslinking degree and particle size. During the preparation of DNA microcapsules, we incorporate manganese carbonyl (MnCO) into the gelation mixture during the preparation of DNA microcapsules to confer probiotic microcapsules with micromotor feature of inflamed region-specificity upon reaching the colon (Fig. [56]1c). This site-specificity is achieved through the chemotaxis of microcapsules towards the elevated concentration of reactive oxygen species (ROS) in the inflamed colonic tissue. Notably, during the reaction between MnCO and ROS, the generated CO gas simultaneously acts as a driving force for the chemotactic mobility of microcapsules and a potent anti-inflammatory immunomodulator for M2 macrophage polarization. With a nitric oxide (NO)-cleavable crosslinker (N,N-(2-amino-1,4-phenylene) diacrylamide, APD), the microcapsules then disintegrate for the controlled release of EcN in response to the overproduced NO at the inflamed colonic site, from which EcN reshapes the dysbiosis of the intestinal microflora to facilitate the recovery of the intestinal barrier in dextran sodium sulfate (DSS)-induced colitis in a mouse model. This study presents significant conceptual and practical advancements in the microfluidic production of acid-resistant DNA microcapsules as probiotic-transporting micromotors that feature superior acid resistance and unique acid adaptability in the GI tract, as well as chemotactic migration towards inflamed colonic tissue and stimulus-responsive release of probiotics within that environment. Fig. 1. Schematic illustration of the preparation of acid-resistant probiotic DNA microspheres for chemotactic colitis treatment. [57]Fig. 1 [58]Open in a new tab a Preparation of EcN/CO@HAD microcapsules via a droplet microfludic technology. b Reversible shrinking-swelling transition of EcN/CO@HAD microcapsules induced by formation/dissociation of A-motif and i-motif structures in the GI tract. c Therapeutic mechanisms of EcN/CO@HAD microcapsules in model mice with DSS-induced colitis. Results Microfluidic generation of acid-resistant probiotic DNA microcapsules In order to prepare acid-resistant DNA microspheres, a droplet microfluidic technology based on coaxial electrospinning needle was first employed for generation of uniform W/O droplets. Specifically, the mixture in the droplets contained the polymeric scaffold HAMA, acid-resistant A-/C-rich DNA oligomers (A-rich sequence: 5’-acry-AAA AAA AAA AAA AAA AAA AAA-3’; C-rich sequence: 5’-acry-AAA CCC CAA ACC CC-3’), photo-initiator lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP), and crosslinker APD (Supplementary Fig. [59]1–[60]2). After UV irradiation, the uniform size and monodispersity of DNA microspheres gelled in droplets were confirmed using an optical microscope (Supplementary Fig. [61]3–[62]6). Further image analysis suggests that the size of DNA microspheres could be well tunned by changing the flow rate of the continuous phase (Supplementary Fig. [63]3–[64]4), but not HAMA concentration (Supplementary Fig. [65]5–[66]6). Unless stated otherwise, the mixture containing 2.5 wt% HAMA, 1 wt% APD, 0.5 wt% LAP, 2.6 wt% A-rich DNA oligomer, and 1.4 wt% C-rich DNA oligomer were used to produce DNA microspheres under 1 μL/min continuous phase, and 100 μL/min dispersed phase for further study. For simplicity, such DNA microspheres were denoted as HAD microspheres, from wich “H”, “A”, and “D” standards for the polymer scaffold HAMA, cross-linker agent APD, and DNA oligomer, respectively. Furthermore, probiotic Escherichia coli Nissle 1917 (EcN) and manganese carbonyl (MnCO) could be readily encapsuled into HAD microspheres through including them into the aforementioned gelation mixture, resulting in the formation of unprecedented probiotic microcapsules, EcN- and MnCO-loaded HAD (EcN/CO@HAD). To maximize the viability of EcN, the crosslinking of microcapsules was initiated using relatively low-cytotoxic 405 nm UV irradiation for only 2 min (Supplementary Fig. [67]7). The collected microcapsules were washed with PBS and re-dispersed in PBS or water for further study. In order to better visualize them in aqueous solution, DNA oligonucleotides were stained using GelGreen, a fluorescent nucleic acid dye. The fluorescence microscopy and scanning electron microscopy (SEM) images clearly revealed the uniform size and spherical morphology of microcapsules with an average diameter about 345 μm (Supplementary Fig. [68]8). Importantly, A-rich sequences adopt parallel duplex A-motif structures through the reverse Hoogsteen base pairing and electrostatic interactions under highly acidic conditions (pH <3.0), but A-motif structures dissociate into single-stranded A-rich strands when pH exceeds 4.0 (Fig. [69]2a)^[70]29. Under mildly acidic conditions (pH 4.0-6.0), C-rich sequences form i-motif quadruplex structures, stabilized by hemi-protonated and intercalated cytosine-cytosine base pairs (C:C^+), but the i-motif structures dissociate into single-stranded C-rich strands when the pH is neutral (Fig. [71]2b)^[72]33,[73]34. In this context, probiotic DNA microcapsules went through reversible shrinking-swelling transition under the pH switch between pH 7 and pH 2 (Fig. [74]2c, Supplementary Movie [75]1), or pH 7 and pH 5 (Fig. [76]2d, Supplementary Movie [77]2). The initial average particle size of probiotic DNA microcapsules was about 345 μm at neutral pH (pH 7). Upon adjusting the pH to 2 with HCl, the particle size decreased to 226 μm. When the pH was subsequently adjusted to a neutral level with NaOH, the microcapsule size recovered to 311 μm, which is close to the original size (Fig. [78]2c, Supplementary Movie [79]1). Similarly, reducing the pH from 7 to 5 caused the average particle size of probiotic DNA microcapsules to decrease from 335 μm to 206 μm, and the microcapsule size was recovered to 328 μm upon restoring the pH to a neutral level (Fig. [80]2d and Supplementary Movie [81]2). These reversible size changes were observed in both two pH switch systems across the repeated cycles. Note that, the contraction of particle size under acidic conditions resulted in an enhancement of the Young’s modulus of probiotic DNA microcapsules (Supplementary Fig. [82]9). This size contraction arises from the formation of the additional DNA networks formed under acidic environment, which, in turn, promises to protect the encapsulated EcN against the hostile acidic environment. To validate the hypothesis, EcN/CO@HAD microcapsules, EcN/CO@HA microcapsules (without DNA sequences), and free EcN were added to a solution at pH 2 or 5 and incubated for 5 h, respectively. The survival rates of EcN in different groups were assessed by a plate colony counting method (Fig. [83]2e, f). As expected, in the presence of A-/C-rich sequences, the EcN/CO@HAD group exhibited appreciable EcN survival rates at both pH 2 (87.7%) and pH 5 (99.4%). In sharp contrast, in the absence of DNA sequences, the EcN/CO@HA group displayed quite low EcN survival rates at pH 2 (13.3%) and pH 5 (23.0%), which were only slightly higher than those observed in the free EcN group (1.9% at pH 2; 16.3% at pH 5). Meanwhile, live/dead staining was further used to visualize EcN in different groups (Supplementary Fig. [84]10). Obviously, only a few dead bacteria (stained red) were observed in the EcN/CO@HAD group after exposure to acidic solutions for 5 h. However, a large amount of dead bacteria were detected in both the EcN/CO@HA group and the free EcN group. After extending the incubation of probiotic microcapsules under acidic conditions to 48 h, the EcN/CO@HAD group still maintained a relatively high survival rate of probiotics (67.0% at pH 2, 75.2% at pH 5, Supplementary Fig. [85]11). Taken together, these findings validate that A-/C-rich sequences markedly enhance the acid tolerance of EcN by forming additional DNA networks within the contracted microcapsules. Fig. 2. Characterization of acid resistance of EcN/CO@HAD microcapsules. [86]Fig. 2 [87]Open in a new tab Schematic illustration of molecular structures of (a) A:A base pairing and electrostatic interaction, and (b) hemi-protonated C:C+ base pairing. Representative fluorescence micrographs and size distribution of EcN/CO@HAD microcapsules during the repeated cycles of (c) pH 7 and 2, and (d) pH 7 and 5, respectively. e Digital photographs of bacterial colonies and (f) survival rates of bacteria from free EcN, EcN/CO@HA microcapsules, and EcN/CO@HAD microcapsules after exposure to acidic solutions for 5 h. Exposure of free EcN to PBS was used as a control. Data are presented as mean ± S.D (n = 3 biologically independent samples for f). Statistical significance was calculated via one-way ANOVA with two-tailed LSD multiple-comparisons test in (f) (n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001). Source data are provided as a Source Data file. DNA microcapsules as chemotactic micromotors for site-specific transport and release of probiotics IBD lesions are characterized by oxidative stress^[88]6, which were explored as an attractant for the chemotaxis of EcN/CO@HAD microcapsules. That is to say, the preferred reaction between MnCO in the microcapsules and the surrounding ROS generated CO gas, which acted as a driving force, enabling EcN/CO@HAD microcapsules to function as a micromotor that continuously converge towards areas of high ROS concentration (Fig. [89]3a). With moderate crosslinking density, the HAMA matrix enabled the swift diffusion of H[2]O[2] into the EcN/CO@HAD microcapsules and allowed the reaction product, CO gas, to readily diffuse out (Supplementary Fig. [90]12–[91]13), in addition to ensuring the rapid proliferation of EcN (Supplementary Fig. [92]14). The release kinetics revealed that CO was continuously generated over time, reaching a cumulative concentration of 2.6 mM at equilibrium after 12 h of co-incubation of EcN/CO@HAD (2 mg/mL) with H[2]O[2] (1 mM) (Supplementary Fig. [93]15). Even with this high-concentration combination, the concentration of generated CO remained well below the cytotoxic threshold (~ 3.6 mM)^[94]35. To investigate the motility of EcN/CO@HAD microcapsules, we exposed microcapsules to a homogeneous H[2]O[2] solution and then recorded their motility trajectories using a microscope. Clearly, in the absence of H[2]O[2], the microcapsules displayed negligible displacement, as evidenced by a flat mean square displacement (MSD) curve (Fig. [95]3b, d, and Supplementary Movie [96]3). In contrast, the presence of H[2]O[2] induced significant motility and oscillatory behavior of the microcapsules, which correlated with a parabolic MSD curve (Fig. [97]3c, d, and Supplementary Movie [98]4). The motility characteristics of the micromotors were sustained in the presence of other chemical signal, such as NO, in the inflamed colonic sites within 0.5 h (Supplementary Fig. [99]16, Movie [100]5). After the release of CO gas, the EcN/CO@HAD microcapsules exhibited negligible changes in shape, size, or integrity (Supplementary Fig. [101]17), indicating that the CO generation process had minimal influence on the structural stability of the microcapsules. Furthermore, we exposed microcapsules to a concentration gradient of H[2]O[2] solution to investigate their directional motility. In brief, a Y-shaped channel was filled with PBS, after which agarose solutions without and with H[2]O[2] were placed in reservoirs II and III of the Y-shaped channel, respectively, and allowed to solidify at room temperature (Fig. [102]3e). Subsequently, H[2]O[2] diffused through the Y-shaped channel, creating a spatial chemoattractant concentration gradient over time. Remarkably, microcapsules in reservoir I exhibited directed movement towards and preferential accumulation along the main and bifurcated channels adjacent to reservoir III, where a concentration gradient of H[2]O[2] was present, rather than towards the channels near reservoir II, which lacked H[2]O[2]. Conversely, in the absence of H[2]O[2] within the Y-shaped channel, the microcapsules remained stationary in reservoir I (Supplementary Fig. [103]18). In parallel, the Y-shaped channel was filled with a commercial simulated intestinal fluid containing various enzymes, inorganic salts, phospholipids, and other substances (Fig. [104]3f). The remaining experimental procedures were consistent with those described previously. Over time, the microcapsules preferentially moved along the main channel towards reservoir III and ultimately accumulated in the branch channel of reservoir III (Fig. [105]3f, Supplementary Movie [106]6). These results suggest that the chemotaxis of microcapsules towards ROS is robust and specific, primarily driven by the efficient gas-generating reaction between MnCO and ROS, rather than being influenced by other chemical signals present in the intestinal fluid. Taken together, these findings underscore that EcN/CO@HAD microcapsules function as micromotors by utilizing ROS as the ‘fuel’ and generate gaseous CO as the driving force, which is crucial for their effective accumulation in IBD lesions (discussed below). Given that CO can modulate macrophage polarization towards the anti-inflammatory M2 phenotype^[107]36, EcN/CO@HAD microcapsules represent potent anti-inflammatory micromotors for further mitigating inflammation of IBD lesions (discussed below). Fig. 3. Chemotactic motility of EcN/CO@HAD microcapsules for the site-specific release of probiotics. [108]Fig. 3 [109]Open in a new tab a Schematic illustration of chemotaxis of EcN/CO@HAD microcapsules towards ROS. b, c Movement paths and (d) MSD fitting curves of EcN/CO@HAD microcapsules in the aqueous solutions (b) in the absence or (c) in the presence of H[2]O[2]. Representative micrographs of EcN/CO@HAD microcapsules in the main channel and the forked branches of a Y-shaped channel at different time points after placing microcapsules in reservoir I, when the Y-shaped channel was filled with (e) PBS and (f) simulated intestinal fluid, respectively. g Mechanism of rupture of EcN/CO@HAD microcapsules in response to NO for release of probiotics. h Representative SEM images of EcN/CO@HAD microcapsules after 20 h exposure or non-exposure to NO from 0.2 mM donor. i Micrographs of GelRed stained EcN/CO@HAD microcapsules after exposure or non-exposure to NO from 0.2 mM donor at different time points. Source data are provided as a Source Data file. The close communication between bacteria and their surrounding environments facilitates the bacteriotherapy^[110]37, which calls for the stimulus-responsive release of bacteria in IBD lesions. Since a high concentration of NO is known to be involved in the pathogenesis of IBD^[111]38, EcN/CO@HAD microcapsules bearing NO-cleavable crosslinker APD were subjected to the NO donor, sodium nitroferricyanide(III) dihydrate (SNP) under 405 nm light irradiation (Fig. [112]3g). As expected, EcN/CO@HAD microcapsules maintained the structural integrity (Fig. [113]3h, i) and restricted the EcN release in the absence of gaseous NO (Supplementary Fig. [114]19). Conversely, these microcapsules were broken down into cracked pieces with time in the presence of gaseous NO (Figs. [115]3h, i, Supplementary Fig. [116]20), at a concentration corresponding to or lower than those present at the inflammatory colonic sites^[117]39,[118]40. Meanwhile, the EcN release rate was significantly promoted by microcapsule disintegration triggered by NO, and hence 77.7% of EcN was released from microcapsules at 20 h post treatment (Supplementary Fig. [119]21). Taken together, these findings suggest that EcN/CO@HAD microcapsules hold significant potential for delivery of probiotics to IBD lesions, followed by stimulus-responsive release of probiotics therein. Reactive molecule scavenging and macrophage polarization in vitro and probiotic survival in the colon Prior to assessment of biofunctionalities, the exceptional cytocompatibility of both CO@HAD microcapsules and CO generation was validated using the CCK–8 assay (Supplementary Fig. [120]22). Building on the promising results that demonstrate the responsiveness of CO@HAD microcapsules to ROS and NO (Fig. [121]3), we extended our investigation to assess the efficacy of CO@HAD microcapsules in scavenging ROS and NO at the cellular level (Fig. [122]4a, [123]e). In this context, L929 fibroblast cells were co-incubated with CO@HAD microcapsules and Rosup agent to evaluate the ROS scavenging capability. To visualize the intracellular ROS, the fluorescent probe dichlorofluorescein diacetate (DCFH–DA) was used to stain cells. As shown in Fig. [124]4a, b, the treatment of CO@HAD microcapsules markedly inhibited the rise in intracellular ROS levels and effectively restored ROS levels to baseline, compared to the treatments lacking MnCO. This is in good agreement with the test tube results showing H[2]O[2] scavenging by CO@HAD microcapsules (Supplementary Fig. [125]23). Of note, LPS stimulation elevates the concentration of intracellular ROS^[126]41 that can be eliminated by MnCO to generates CO, a gaseous signaling molecule with potent anti-inflammatory properties^[127]36,[128]42. Therefore, the potential of CO@HAD microcapsules to modulate LPS-stimulated macrophage (RAW264.7 cell) phenotype was explored by immunofluorescence staining and flow cytometry for detection of M1 (iNOS) and M2 (CD206) macrophage markers, respectively. Obviously, iNOS (stained red) in LPS-induced macrophages was significantly downregulated, whereas CD206 (stained green) was considerably upregulated, indicating that macrophages were polarized to the M2 phenotype by CO@HAD microcapsules (Fig. [129]4c). The corresponding flow cytometry results showed that the proportions of CD206-positive and iNOS-positive macrophages in the CO@HAD group were markedly altered to 26.9% and 8.88%, respectively, compared to the PBS (CD206: 0.095%; iNOS: 79.6%), HD (CD206: 0.18%; iNOS: 69.8%), and HAD (CD206: 8.18%; iNOS: 56.0%) groups (Fig. [130]4d). In addition, colonic macrophages isolated from colitis mice were cultivated with the microcapsules for in vitro study of phenotypic transformation. Flow cytometry results revealed that the CO@HAD treatment significantly increased the proportion of CD206-positive macrophages (25.4%), and meanwhile reduced the proportion of iNOS-positive macrophages (7.15%), compared to the PBS (CD206: 5.81%; iNOS: 22.8%), HD (CD206: 5.82%; iNOS: 17.9%), and HAD (CD206: 11.8%; iNOS: 16.3%) groups (Supplementary Fig. [131]24). Taken together, these findings underscore the capacity of CO@HAD microcapsules to effectively modulate the polarization of macrophages from a pro-inflammatory M1 to an anti-inflammatory M2 phenotype. Fig. 4. Reactive molecule scavenging and M2 macrophage polarization in vitro and enhanced probiotic survival in the colon. [132]Fig. 4 [133]Open in a new tab a Representative fluorescence micrographs of the ROS probe and b quantitative assessment of the ROS probe signals in cells after co-incubation with Rosup (1 μg/mL) and PBS, CO@HD, or CO@HAD (1 mg/mL) for 2 h. Cells untreated with Rosup were used as a control. c Representative fluorescence micrographs of immunostained biomarkers of macrophages (red: iNOS; green: CD206). d Flow cytometry analysis of the proportions of M1 and M2 phenotypes in macrophages after treatment with PBS, CO@HD, or CO@HAD (1 mg/mL) for 24 h. e Representative fluorescence micrographs of the NO probe and (f) quantitative assessment of the NO probe signals in cells after co-incubation with DETA-NONOate (0.2 mM) and PBS, CO@HD, or CO@HAD (1 mg/mL) for 6 h. Cells untreated with DETA-NONOate was used as a control. g Representative bioluminescence images (color scale (p/sec/cm^2/sr): Min = 2000; Max = 4500) and (h) quantitative assessment of EcN-mCherry signals in the intestine extracted from mice after gavage of water, EcN-mCherry, EcN-mCherry/CO@HD, or EcN-mCherry/CO@HAD (5×10^8 CFU EcN) for 12 h. Data are presented as mean ± S.D (n  =  3 biologically independent samples for b, f; n  =  4 biologically independent animal samples for h). Statistical significance was calculated via one-way ANOVA with two-tailed LSD multiple-comparisons test in (b, f, h) (n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001). Source data are provided as a Source Data file. Next, L929 fibroblast cells were co-incubated with CO@HAD microcapsules and NO slow-release donor, diethylamine NONOate (DETA-NONOate), to evaluate the NO scavenging capability. For visualization, cells were stained with fluorescence probe diaminofluorescein-FM diacetate (DAF-FM DA). Clearly, the fluorescence signal in the CO@HAD group was significantly lower than that in the CO@HD group (Fig. [134]4e). Further statistical analysis revealed that the fluorescence intensity of NO in the CO@HAD group was about 8.3 times lower than that in the CO@HD group (Fig. [135]4f). Therefore, the incorporation of the NO-responsive crosslinker APD in the microcapsules promise to scavenge excess NO and release EcN in IBD lesions. In addition, the ability to mitigate excessive NO is pivotal in preventing the exacerbation of inflammation and the associated intestinal damage^[136]43,[137]44. Encouraged by the protection effect of DNA microcapsules to EcN in vitro (Fig. [138]2e, f), we subsequently examined their cytoprotective efficacy against the acidic environment of the stomach and duodenum in vivo. To this end, mice were orally administered with EcN-mCherry/CO@HAD microcapsules. After 12 h, the mice were euthanized and the intestines were collected for ex vivo fluorescence imaging. As shown in Fig. [139]4g, h, the EcN-mCherry/CO@HAD group exhibited significantly higher fluorescence intensity of EcN-mCherry in the colon compared to both the free EcN-mCherry group and the EcN-mCherry/CO@HD group. This observation corroborates the enhanced survival and retention of probiotics within the DNA microcapsules in vivo. Therapeutic efficacy against DSS-induced colitis in model mice After validating the multifaceted capabilities of the probiotic DNA microcapsules, including acid resistance, chemotactic motility, and macrophage modulation, we proceeded to evaluate their therapeutic potential in ameliorating DSS-induced colitis, one of the most commonly used IBD models. First, mice were fed 3% DSS for 7 days to construct an experimental model of IBD. Subsequently, mice were randomly divided into seven groups and orally administrated with water, free EcN, HD, HAD, CO@HAD, EcN/CO@HAD, and EcN/CO@HAD, respectively, for 4 times (Fig. [140]5a). Mice without DSS induction were used as a healthy control group. During the experimental period, fecal consistency, fecal occult blood, and weight loss were recorded every two days to determine the severity of colitis. Apparently, the weight of DSS-fed mice decreased significantly on day 7 (Fig. [141]5b). Treatment with EcN/CO@HAD microcapsules resulted in the most substantial improvements in weight, diarrhea, and fecal bleeding in mice, surpassing the effects observed with free EcN, HD, HAD, and EcN/CO@HA (Fig. [142]5b, Supplementary Fig. [143]25). Concurrently, the EcN/CO@HAD group exhibited the most pronounced reduction in the disease activity index (DAI) (Fig. [144]5c). Notably, while each of the individual components within the microcapsules, such as acid-resistant DNA, NO-cleavable APD, CO donor MnCO, and probiotic EcN, contributed to some amelioration of colitis symptoms, their co-existence in the EcN/CO@HAD microcapsules resulted in the most pronounced therapeutic effect. Fig. 5. Therapeutic efficacy of probiotic microcapsules against DSS-induced colitis. [145]Fig. 5 [146]Open in a new tab a Schematic illustration of the experimental design for colitis treatment. b Body weight and (c) disease activity index (DAI) score of the mice with DSS-induced colitis after treatment with water, free EcN, HD microcapsules, HAD microcapsules, CO@HAD microcapsules, EcN/CO@HA microcapsules, or EcN/CO@HAD microcapsules (5 × 10^8 CFU EcN, 2 mg microcapsules, 0.2 mL) for 4 times. d Digital photos of the colons and spleens of mice in different groups. Quantification of (e) colon length and (f) spleen weight in different groups. g Histopathological scoring of the colon tissues in different groups. h Relative ZO-1 mRNA content in colon tissues in different groups. i Representative H&E-stained images of colon tissue. j Representative ZO-1 immunofluorescence-stained images and TUNEL-stained images of the colon tissues in different groups. k Representative fluorescence images of ROS probe in the colon tissues in different groups. Mice without DSS induction were used as a healthy control group. Data are presented as mean ± S.D (n  =  3 biologically independent animal samples for b, c, e, f, h; n  =  4 biologically independent animal samples for g). Statistical significance was calculated via one-way ANOVA with two-tailed LSD multiple-comparisons test in (b, c, e, f, g, h) (n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001). Source data are provided as a Source Data file. On day 15, the mice were euthanized and their colons were collected for photographic and pathological analyses (Fig. [147]5d–k). In the DSS-induced IBD model, the high oxidative stress and inflammation leads to tissue damage and shorter colon length^[148]45. As shown in Fig. [149]5d, e, the mean colon length of mice in the free EcN group (4.1 cm), the HD group (4.8 cm), or the HAD group (4.5 cm) was significantly shorter than that in the healthy group (6.9 cm). The colon length of mice in the CO@HAD group (5.9 cm) and EcN/CO@HA group (6.2 cm) showed some recovery, but the therapeutic effect is still unsatisfactory due to the lack of EcN in the former group, and the inactivation of EcN by the stomach and duodenum in the latter group. In contrast, the mean colon length of mice in the EcN/CO@HAD group (6.7 cm) was restored to the similar level in the healthy mice. In addition, the EcN/CO@HAD microcapsules attenuated splenic hypertrophy, as further evidenced by a reduction in splenic weight (Fig. [150]5d, f), suggesting that the capability of the EcN/CO@HAD microcapsules to reduce systemic inflammation. Given inflammatory cell infiltration and colonic epithelial damage are typical histologic abnormalities of colons in colitis mice, hematoxylin and eosin (H&E) staining was carried out to evaluate the degree of colonic damage by calculating histopathological scores under a specific guideline (Supplementary Table [151]1). H&E-stained images showed partial disappearance of the crypts with concomitant inflammatory cell infiltration in the colon from IBD model after treatment with water only (Fig. [152]5i). Such severe colonic injury was effectively reversed after treatment with EcN/CO@HAD microcapsules, as evidenced by the reduction of inflammatory cell infiltration, restoration of crypts and mucosal layers, and maintenance of colonic epithelial integrity. Therefore, the histopathological scores in the EcN/CO@HAD group were quite similar with that in the healthy control group (Fig. [153]5g). Collectively, these findings underscore that EcN/CO@HAD microcapsules represent an ideal therapeutic candidate for colitis, and the superior efficacy of these microcapsule stems from the synergistic integration of acid-resistant DNA, NO-cleavable APD, CO donor MnCO, and probiotic EcN. Since zonula occludens-1 (ZO–1), a key constituent protein in the intestinal mucosal epithelial cells, plays a crucial role in repairing intestinal mucosal damage and the maintaining intestinal endothelial homeostasis^[154]46,[155]47, the gene expression of ZO–1 in the colon tissue was assessed by reverse transcription-quantitative polymerase chain reaction (RT-qPCR) (Fig. [156]5h) and immunofluorescence staining (Fig. [157]5j, Supplementary Fig. [158]26). Obviously, although the expression levels of ZO-1 mRNA and protein were down-regulated by DSS, they were markedly up-regulated by EcN/CO@HAD microcapsules, as shown by the RT-qPCR data and immunofluorescence images. In addition, the colonic epithelial cells were protected from apoptosis by EcN/CO@HAD microcapsules, as evidenced by the terminal deoxynucleotide transferase dUTP Nick end labeling (TUNEL) staining (Fig. [159]5j, Supplementary Fig. [160]27–[161]28). These results indicate that EcN/CO@HAD microcapsules could facilitate tissue repair by up-regulating ZO-1 expression in intestinal epithelial cells and concurrently safeguarding them against apoptosis. To elucidate the underlying mechanisms, we further investigated the ability of EcN/CO@HAD microcapsules to scavenge ROS and mitigate inflammation in vivo. This is crucial because cell death and tissue necrosis in the compromised intestinal barrier are primarily driven by pathogen-induced production of ROS and pro-inflammatory cytokines. As anticipated, the EcN/CO@HAD group displayed a significant decrease in ROS level in colon tissues (Fig. [162]5k, Supplementary Fig. [163]29–[164]30), consistent with the in vitro ROS scavenging results observed at the cellular level (Fig. [165]4a, b). Impressively, the expression of pro-inflammatory factors, interleukin-6 (IL-6) and tumor necrosis factor (TNF-α), in colon tissues was suppressed (Fig. [166]6a, c, d, Supplementary Fig. [167]31–[168]32), while the expression of anti-inflammatory factors, interleukin-10 (IL-10) and transforming growth factor-β (TGF-β), was elevated to the greatest extent by EcN/CO@HAD microcapsules (Fig. [169]6b, e, f, Supplementary Fig. [170]33–[171]34). Fig. 6. Modulation of immune microenvironment of colon tissue by probiotic microcapsules. [172]Fig. 6 [173]Open in a new tab Immunofluorescence staining of (a) IL-6 and TNF-α and (b) IL-10 and TGF-β in the colon tissues extracted from the mice with DSS-induced colitis after treatment with water, free EcN, HD microcapsules, HAD microcapsules, CO@HAD microcapsules, EcN/CO@HA microcapsules, or EcN/CO@HAD microcapsules (5 × 10^8 CFU EcN, 2 mg microcapsules, 0.2 mL) for 4 times. The mRNA expression levels of (c) IL-6, (d) TNF-α, (e) IL-10, and (f) TGF-β of the colon tissues in different groups. g Flow cytometry analysis of phenotype change of macrophages in the colon tissues in different groups. h Immunofluorescence staining of iNOS (red, M1 marker), CD206 (red, M2 marker), and CD68 (green, macrophage marker) in the colon tissues in different groups. Data are presented as mean ± S.D (n  =  3 biologically independent animal samples for c, d, e, f). Statistical significance was calculated via one-way ANOVA with two-tailed LSD multiple-comparisons test in (c, d, e, f) (n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001). Source data are provided as a Source Data file. In fact, the dysregulation of inflammatory factors is manifested by the M1 macrophage polarization at the IBD lesion^[174]8. Therefore, we further investigated whether EcN/CO@HAD microcapsules could regulate the polarization of macrophage to M2 phenotype in vivo. First, macrophages were extracted from colon tissues after different treatments and the proportions of pro-inflammatory M1 phenotype and anti-inflammatory M2 phenotype were assessed by flow cytometry (Fig. [175]6g, Supplementary Fig. [176]35). For the model mice with DSS-induced colitis, the percentage of CD11b/iNOS-positive M1-type macrophage constituted 56.3% of the total macrophage population in the water group, whereas the percentage of CD11b/CD206-positive M2-type macrophages accounted for only 2.97%, indicating a significant infiltration of pro-inflammatory macrophages within the colon tissue. After treatment with EcN/CO@HAD microcapsules, the percentage of CD11b/CD206-positive M2-type macrophages increased to 37.1%, and the percentage of CD11b/iNOS-positive M1-type macrophages decreased to 19.6%, indicating that EcN/CO@HAD microcapsules favored the polarization of macrophage from M1 phenotype to M2 phenotype. Subsequently, immunofluorescence staining of CD68/iNOS and CD68/CD206 was used to visualize the macrophage M1 phenotype and M2 phenotype, respectively (Fig. [177]6h, Supplementary Figs. [178]36–[179]37). In the EcN/CO@HAD group, the fluorescence signals of iNOS in the colon tissue were quite weak, while those of CD206 were notably strong, consistent with the flow cytometry analysis results. The switching of macrophages from M1 phenotype to M2 phenotype facilitated the reduction of the overexpressed pro-inflammatory factors and promoted the restoration of immune homeostasis. Notably, free EcN, HD microcapsules, and HAD microcapsules showed limited therapeutic efficacy due to their insufficient capability to eliminate oxidative stress and to modulate immune microenvironment. Although CO@HAD microcapsules and EcN/CO@HA microcapsules could eliminate oxidative stress and modulate immune microenvironment effectively, their therapeutic outcomes remained suboptimal in the absence of bioactive probiotics. In stark contrast, among the various treatments evaluated, EcN/CO@HAD microcapsules displayed the superior efficacy through a synergistic approach by eliminating oxidative stress, modulating the immune microenvironment, and preserving the bioactivity of EcN. Regulation of gut microbiome and host genome in the colon tissues Both fundamental and clinical studies have highlighted the pivotal role of the gut microbiome in the IBD progression^[180]11. Encouraged by the excellent biocompatibility (Supplementary Fig. [181]38) and superior therapeutic efficacy of EcN/CO@HAD microcapsules (Figs. [182]5–[183]6), we next investigated whether the microcapsules could rectify the dysbiosis of gut microbiota in colitis models and thus halt the progression of IBD. Therefore, changes in the gut microbiota of DSS-induced colitis mice after treatment with microcapsules were elucidated through 16S rDNA amplicon sequencing. As shown in Fig. [184]7a, the microbial richness, presented as observed operational taxonomic units (OTUs), in the EcN/CO@HAD group was comparable to that in the healthy water group (DSS (-) + water) but significantly higher than that in the colitic water group (DSS (+) + water). In addition, the three indices related to α-diversity, Shannon index (Fig. [185]7b), Simpson index (Fig. [186]7c), and Chao1 index (Fig. [187]7d), decreased in the colitic water group but increased in the EcN/CO@HAD group compared to the healthy water group. Principal component analysis (PCA) of microbial diversity differences showed that data points representing the EcN/CO@HAD group and the healthy water group clustered more closely together and were in closer proximity to each other compared to those representing the colitic water group (Supplementary Fig. [188]39), indicating that the EcN/CO@HAD group and the healthy water group shared similar microbial compositions. Venn diagrams revealed a greater specific overlap of microbial species between the EcN/CO@HAD group and the healthy water group (438) compared to the overlap between the colitic water group and the healthy water group (113) (Fig. [189]7e). These findings indicate that EcN/CO@HAD microcapsules restored the microbial diversity and abundance to a level comparable to the healthy group by virtue of the cytoprotective effect of acid-resistant A-/C-rich sequences. Fig. 7. Bioinformatic analysis of gut microbiome. [190]Fig. 7 [191]Open in a new tab a Observed operational taxonomic unit (OTU) richness of the gut microbiota in different groups. α-diversity analysis of the gut microbiome evaluated by (b) Shannon index, (c) Simpson index, and (d) Chao1 index. e Venn diagram of gut microbiota between different groups, and the values between different circles represent the number of the same species. The relative abundance histogram of gut microbiota at (f) the phylum level and (g) the family level in different groups. h Cladogram based on linear discriminant analysis effect size (LEfSe) analysis showing community composition of the gut microbiota in different groups. Red nodes in the branches represented microbial taxa that play critical roles in the water (with DSS) group, and green nodes represented microbial taxa that play critical roles in the EcN/CO@HAD group. Red nodes in the branches represent microbial taxa that play critical roles in the colitic water group, green nodes represent microbial taxa that play critical roles in the EcN/CO@HAD group, and yellow nodes indicate that taxa are not significantly different. i Clustering heat map of the functional abundance of KEGG pathways for bacterial communities in different groups. The color gradient from blue to red indicates low to high relative abundance. Data are presented as mean ± S.D (n  =  3 biologically independent animal samples for a–i). Statistical significance was calculated via one-way ANOVA with Bonferroni multiple comparisons test in (a, b, c) (n.s. P > 0.05, * P < 0.05, ** P < 0.01, *** P < 0.001). Source data are provided as a Source Data file. Studies at the phylum level revealed that EcN/CO@HAD microcapsules significantly altered the relative abundance of specific microbiota in IBD model (Fig. [192]7f). Briefly, compared to the colitic water group, the proportion of Firmicutes increased and that of Bacteroidetes decreased in the EcN/CO@HAD group, resulting in a higher Firmicutes/Bacteroidetes ratio that plays crucial role in maintaining intestinal homeostasis^[193]45. In addition, the abundance of Proteobacteria (known to mostly consist of harmful bacteria)^[194]48 was also significantly decreased by EcN/CO@HAD microcapsules. The composition change of the major microbiota in the different group was further analyzed at the family level (Fig. [195]7g). Clearly, EcN/CO@HAD treatment significantly increased the relative abundance of Muribaculaceae (known to have anti-inflammatory effects)^[196]49, Akkermansiaceae (known to enhance the intestinal barrier)^[197]50, and Prevotellaceae (known to produce short-chain fatty acids)^[198]51, and simultaneously reduced the relative abundance of Enterobacteriaceae and Desulfovibrionaceae, two typical kinds of pathogens accumulated in the colitis colon^[199]52. Similar patterns were evident in the linear discriminant analysis (LDA) effect size (LEfSe), which highlighted the dominant microbial groups and their impacts across multiple taxonomic levels, from phylum to species (Fig. [200]7h). Analysis at the genus level showed that the predominated microbiota in the water group were Bacteroides, such as Escherichia-Shigella and Desulfovibrio, while probiotics were significantly abundant in the EcN/CO@HAD group, such as Prevotellaceae_UCG_001 and Akkermansia (Supplementary Fig. [201]40). Besides, the clustering heat map was employed to depict the functional abundance of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in bacterial communities (Fig. [202]7i). The results showed that the abundance of metabolic pathways of bacterial communities in the EcN/CO@HAD group was similar with that in the healthy group, but was higher than that in the colitic water group. Notably, the EcN/CO@HAD group exhibited enhanced functions related to amino acid and glycan anabolism, as well as short chain fatty acid synthesis and anabolism, compared to the colitic water group. Collectively, these findings suggest that EcN/CO@HAD microcapsules have a notable capacity to remodel the gut microbiome composition and function in colitis mice, restoring gut microbial homeostasis without inducing aberrant imbalances, thereby offering a promising approach for the effective management of IBD. To better understand the origin of intestinal barrier repair, RNA sequencing was performed to examine the transcriptome of colon tissues of DSS-induced IBD model, and 695 genes were found highly differentially expressed with an absolute fold change greater than 2 between the colitic water group and the EcN/CO@HAD group (Fig. [203]8a). Among the differentially expressed genes (DEGs), 404 genes were up-regulated and 291 genes were down-regulated in the EcN/CO@HAD group compared to the colitic water group. Subsequently, the major biological functions of DEGs were further analyzed via enrichment pathways using Gene Ontology (GO) analysis and KEGG analysis. GO analysis showed that the down-regulated genes in the EcN/CO@HAD group were primarily linked to inflammatory response, immune cell activities (i.e., proliferation, migration, adhesion, activation, etc.) (Fig. [204]8b), consistent with the previously observed reduction in oxidative stress, mitigation of pro-inflammatory response, and enhancement of the intestinal barrier repair (Figs. [205]5, [206]6). In addition, KEGG pathway enrichment analysis showed that the DEGs enriched in the EcN/CO@HAD group were involved in multiple signaling pathways (i.e., NF–κB, PI3K–Akt, HIF–1, TNF, etc.) related to inflammation modulation compared with the colitic water group (Fig. [207]8g)^[208]36. Specifically, microcapsules regulated the gene expression-associated NF-κB signaling pathway, leading to a reduction in the gene expression of pro-inflammatory cytokines (e.g., TNF-α, IL-6), an increase in the gene expression of anti-inflammatory cytokines (e.g., IL-10, TGF-β) (Fig. [209]6a–f), and a suppression in the gene expression of chemokines (e.g., Cxcl9, Cxcl13, Cxcl16) (Supplementary Fig. [210]41). The expression of those genes is likely regulated by CO produced from MnCO within microcapsules, as CO has previously been validated to exert anti-inflammatory effects by acting on NF-κB signaling pathway^[211]53–[212]55. In addition, changes in cytokines are closely related to the colonization of EcN in the intestine, which has been well illustrated. For instance, the amyloid curly fiber of EcN can be recognized by Toll-like receptor 2 (TLR2) of macrophages, promoting the production of anti-inflammatory factor IL-10^[213]56. The flagellin of EcN engages Toll-like receptor 5 (TLR5), facilitating the production of anti-inflammatory factor IL-22 by TLR5^+CD11c^+ cells in colon^[214]57. The outer membrane vesicles secreted by EcN are implicated in intestinal mucosal signal transduction, modulating immune responses and decreasing the expression of pro-inflammatory cytokines such as IL-17, TNF-α, and iNOS^[215]58. Meanwhile, EcN/CO@HAD clearly modulated the macrophage phenotype-associated PI3K-Akt and HIF-1 signaling pathways (Fig. [216]8g), resulting in the polarization of macrophages towards M2 phenotype (Fig. [217]6g, h). Similarly, this effect is likely attributable to that EcN/CO@HAD activated the PI3K-Akt signaling pathway, and subsequently down-regulated IL-6 expression while up-regulated Stat3 expression (Supplementary Fig. [218]41)^[219]59. Concurrently, EcN/CO@HAD inhibited the HIF-1 signaling pathway, thereby suppressing the glycolytic metabolism of M1 macrophages. As a result, the expression levels of several critical rate-limiting enzymes involved in glycolysis, including Hk2, Pkm, and Fbp2, were significantly reduced, facilitating the polarization of M1 macrophage phenotype towards the M2 phenotype^[220]36,[221]60. Protein-protein interaction (PPI) network analysis of Immune cell activities further suggests that EcN/CO@HAD reshaped the immune microenvironment in the inflamed colonic tissue by suppressing immune cell activities (Fig. [222]8c). The aforementioned studies have concluded that EcN, in conjunction with CO, regulated multiple signaling pathways, which is crucial for restoring intestinal immune homeostasis and promoting the recovery of the intestinal epithelial barrier. Further GO analysis not only supported this conclusion but also elucidated the underlying mechanisms associated with cellular biosynthesis, metabolism, and proliferation, because the up-regulated genes in the EcN/CO@HAD group were predominantly related to the division and differentiation of cells, as well as the synthesis and metabolism of nucleotide, amino acids, short-chain fatty acids, and other substances (Fig. [223]8d, e)^[224]61,[225]62. Furthermore, KEGG pathway enrichment analysis revealed that the DEGs enriched in the EcN/CO@HAD group were related to tight junction and gap junction compared to the colitic water group (Fig. [226]8g). In fact, EcN/CO@HAD upregulated the expression levels of proteins including ZO-1 in colonic epithelial cells (Fig. [227]5 h, j, Supplementary Fig. [228]26), enhancing the tight junctions between cells and thereby reducing the permeability of the colonic epithelial barrier^[229]63. Else, EcN/CO@HAD significantly enhanced the expression levels of mucoprotein-2 (MUC-2) and mucoprotein-3 (MUC-3) in goblet cells (Supplementary Fig. [230]41), facilitating the repair of inflammation-induced damage to the mucus barrier^[231]64. PPI network analysis of Metabolism and Cell proliferation further suggests that EcN/CO@HAD treatment restored intestinal homeostasis by regulating intestinal metabolic function and promoting the proliferation and differentiation of intestinal cells (Fig. [232]8f). Taken together, these findings shed light on the origin of intestinal barrier repair through deciphering functions of the genome and identifying changes in relevant signaling pathways. Fig. 8. Bioinformatic analysis of colon tissues. [233]Fig. 8 [234]Open in a new tab a Volcano plot of the differentially expressed genes (DEGs) of the colon tissues in the colitic water group and the EcN/CO@HAD group. b Gene ontology (GO) enrichment analysis of the DEGs in the colitic water group and the EcN/CO@HAD group associated with inflammatory response. c PPI network analysis of immune cell activities. GO enrichment analysis of the DEGs in the colitic water group and the EcN/CO@HAD group associated with (d) cellular biosynthesis and metabolism and (e) cell proliferation. f PPI network analysis of metabolism and cell proliferation. g Kyoto Encyclopedia of Gene and Genomes (KEGG) analysis of the DEGs in the transcriptome map. The horizontal axis represents gene ratio, the vertical axis represents gene term, and the bubble size represents the enriched gene counts. The P values were determined using the negative binomial distribution, and then the Bonferroni was used for multiple-hypothesis testing correction. Source data are provided as a Source Data file. Discussion In summary, an acid-resistant probiotic DNA microcapsule, EcN/CO@HAD, was produced through photo-crosslinking of HAMA and acrydite-modified A-/C-rich oligomers within the microfludically generated droplets in the presence of NO-cleavable crosslinker, APD, and CO donor, MnCO. When these microcapsules reached the stomach (~ pH 2) and the duodenum (~ pH 5) successively, the A- and C-rich sequences formed stable A-motif and i-motif structures, respectively, inducing an apparent microcapsule contraction to create a dense physical barrier between EcN and GI fluid. In this context, the microcapsules preserved the viability of EcN until they reached the colon (pH > 7), where the DNA motifs reverted to individual chains, resulting in the restoration of crosslinking degree and particle size, which is amenable to eliminating the trade-off between increased crosslinking for acid resistance and reduced permeability for nutrient or metabolite exchange in probiotic DNA microcapsules. This reversible shrinking-swelling transition, mediated by acid-adaptive feature of A-/i-motif sequences, improved the protective effect of microcapsules on probiotics, and the survival rates of probiotics were as high as 87.7% and 99.4% under acidic conditions at pH 2 and pH 5, respectively. Subsequently, the microcapsules containing MnCO demonstrated inflamed colonic region-specificity, achieved by the chemotaxis of the microcapsules towards elevated concentration of ROS at the inflamed colonic tissue. Note that during the reaction between MnCO and ROS, the generated CO gas acted as a driving force for the chemotactic mobility of microcapsules and a potent anti-inflammatory immunomodulator for M2 macrophage polarization simultaneously. The large-sized probiotic DNA microcapsules (> 300 μm) are unable to penetrate the gastrointestinal epithelial barrier to access the bloodstream or other organs, thus promising minimal off-target effects^[235]65,[236]66. Eventually, the microcapsules bearing NO-cleavable crosslinker APD disintegrated to enable the controlled release of EcN at the inflamed colonic site with elevated NO levels, so that EcN reshaped the dysbiosis of the intestinal microflora. The effective therapeutic efficacy of EcN/CO@HAD microcapsules against DSS-induced IBD was clearly demonstrated with the overall recovery of body weight, DAI value, as well as colon length and structure through reducing ROS level, altering pro-inflammatory response, and improving richness and diversity of probiotics. The analysis results of DEGs in the colon tissues by high-throughput RNA-sequencing further revealed the origin of intestinal barrier repair. In short, EcN/CO@HAD microcapsules down-regulated genes primarily linked to inflammatory response and immune cell activities, while up-regulated genes predominantly associated with the biosynthesis, metabolism, and proliferation of colonic epithelial cells. Notably, unlike the market-approved probiotic powers and enteric capsules frequently utilized in clinical settings, which often exhibit poor acid resistance, the acid-resistant probiotic DNA microcapsules clearly outperformed these products in terms of cytoprotection (Supplementary Fig. [237]42–[238]43). While previously reported encapsulation systems have shown some improvement in probiotic viability compared to market-approved products, the conventional polymers generally exhibit inadequate acid resistance and limited acid adaptability^[239]14,[240]15,[241]67. Indeed, the acid resistance of conventional microcapsules or polymeric coatings is primarily dominated by the crosslinking degree^[242]68,[243]69. A higher degree of crosslinking in polymeric networks more effectively restricts the penetration of H^+ ions into encapsulated probiotics. However, a persistently high crosslinking degree of conventional systems may impede the bioactivity of probiotics (Supplementary Fig. [244]14a). The probiotic DNA microcapsules developed in this study incorporate A-/C-rich DNA sequences, which exhibit acid-adaptive properties, characterized by a higher corsslinking degree under acidic conditions but a lower corsslinking degree when reverted to neutral conditions (Fig. [245]2a–d). The significance of these characteristics lies not only in its ability to mitigate persistent high crosslinking degree-associated adverse effects, but also in its effectiveness in protecting the encapsulated EcN from the hostile environments (Fig. [246]2e, f). Moreover, the conventional systems lack essential features of intelligent delivery systems, particularly the ability to exhibit chemotaxis towards inflamed colonic sites and to enable responsive release of probiotics. In contrast, our probiotic DNA microcapsules introduces several innovative aspects through addressing these challenges, which can be succinctly summarized as follows: (1) The probiotic DNA microcapsules are engineered with acid-adaptive sequences, enabling reversible crosslinking and shrinking-swelling transitions. This unique design provides enhanced acid resistance, ensuring effective cytoprotection during GI transit. (2) These microcapsules function as intelligent site-specific micromotors, demonstrating chemotactic movement towards inflamed colonic tissues with elevated ROS concentrations. This targeted motility significantly enhances probiotic bioavailability and colonization at inflammatory sites. (3) These microcapsules can disintegrate specifically in response to the overproduction of inflammatory signaling molecules at the colonic inflammation site. This stimulus-responsive disintegration enables targeted probiotic release, facilitating direct intervention in the dysbiotic intestinal microbiota at the precise location of inflammation. Therefore, this study underscores that the development of acid-resistant DNA microcapsules as probiotic-transporting micromotors is a sound and profound strategy for advancing effective clinical management of IBD. Importantly, it offers a paradigm for research in different IBD models sharing common characteristics of elevated levels of imbalance of the intestinal microbiota, dysregulation of immune homeostasis, and dysfunction of intestinal epithelial barrier. This could be accomplished through the direct application of the DNA microcapsules or their use as a foundation for incorporating components of interest to resolve specific pathological parameters. Else, this study is believed to generate significant interest in the creation of a library of adaptive and bioinstructive DNA microstructures for diverse biomedical applications beyond the realm of IBD treatment, primarily facilitated by the integration of DNA microarchitectures with, for example, DNA-protein conjugates^[247]70,[248]71, DNA origami nanostructures^[249]71,[250]72, and DNA-inorganic nanocomposites^[251]73,[252]74. Methods Materials Acryloyl chloride, Tetrahydrofuran (THF), N,N-Diisopropylethylamine (DIPEA) were purchased from Macklin Biochemical Technology Co., Ltd (Shanghai China). Iron powder, manganese carbonyl (MnCO), sodium nitroferricyanide (III) dihydrate (SNP), 2-nitrobenzene-1,4-diamine, lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP), Span80, Cell Counting Kit-8 (CCK-8), DAF-FM DA (nitric oxide fluorescent probe), and mineral oil were purchased from Titan Scientific Co., Ltd. (Shanghai China). Methacrylated hyaluronic acid (HAMA) was purchased from EFL-Tech Co., Ltd. (Suzhou, China). Escherichia coli Nissle 1917 (EcN) was purchased from Testobio Co., Ltd. (Ningbo, China). Hydrogen Peroxide (H[2]O[2]) Content Assay Kit was purchased from Solarbio Science & Technology Co., Ltd. (Beijing, China). GelGreen was purchased from Sang Biotech Co., Ltd. (Shanghai China). Live/Dead Bacterial Viability Kit, Reactive Oxygen Species Assay Kit, RNeasy^TM Animal RNA Isolation Kit with spin column, BeyoRT^TM III First Stand cDNA Synthesis Kit, and lipopolysaccharide (LPS) were purchased from Beyotime Biotechnology Co., Ltd. (shanghai, China). Simulated intestinal fluid was purchased from Yuanye Bio-Technology Co., Ltd (Shanghai, China). Immunostaining blocking solution and DCFH-DA (reactive oxygen species fluorescent probe) was purchased from KeyGEN BioTECH Co., Ltd. (Jiangsu, Shanghai). CD206 recombinant antibody, iNOS recombinant antibody, FITC-labeled secondary antibody, and Cy3-labeled secondary antibody were purchased from Proteintech. Co., Ltd. (Chicago, IL). HPLC-purified DNA oligonucleotides were purchased from Sangon Biotech Co., Ltd. (Shanghai, China). Phusion DNA polymerase was purchased from NEW ENGLAND Biolabs Co., Ltd. (Ipswich, MA). Dihydroethidium was purchased from Sigma-Aldrich Trading Co., Ltd (Shanghai, China). Cell strains and animals Escherichia coli Nissle 1917 (EcN) was purchased from Testobio Co., Ltd. (Ningbo, China). Mouse fibroblasts (L929) and mouse macrophages (RAW 264.7) were purchased from Hysigen Bioscience Co., Ltd. (Shanghai, China). Male C57BL/6 (aged 5–6 weeks) mice were obtained from Shanghai SLAC Laboratory Animal Co, Ltd. Mice were housed in individually ventilated cages with six mice per cage. All the animals were kept at 23 °C and 50% humidity with a 12 h light/dark cycle. Food and water were supplied ad libitum. All animal experiments were performed following the protocols approved by the Animal Care and Use Committee of Tongji University (approval # TJAF00124105) and also in accordance with the policy of National Ministry of Health of China. Cultivation of Escherichia coli Nissle 1917 (EcN) To activate the strain, lysogeny broth (LB) medium (0.5 mL) was mixed with the freeze-dried Escherichia coli Nissle 1917 (EcN) powder (10^9 CFU), then the bacterial suspension was evenly spread on LB plates and incubated overnight at 37 °C. A single colony from LB plate was picked up and inoculated into 10 mL of fresh LB medium for overnight cultivation at 37 °C on the shaker. Afterwards, the bacterial suspension (1 mL) was centrifuged at 2000 g for 10 min, then washed and resuspended using PBS. EcN concentration in the suspension was calculated according to the optical density at 600 nm. Preparation and characterization of EcN/CO@HAD microcapsules To prepare microspheres, a microfluidic droplet technique was employed. Briefly, two inlet ports of a coaxial electrostatic spinning needle (30 G/19 G) were connected to two separate microfluidic syringe pumps via silicone hoses, one of which served as the inlet port for the continuous phase (mineral oil with 10% (w/w) Span 80) and the other one as the inlet port for the dispersed phase. Then, the formed droplets were subjected to irradiation under 405 nm UV light for 2 min. EcN/CO@HAD microcapsules were typically prepared by adding 2.5% (w/w) HAMA, 0.5% (w/w) LAP, 1% (w/w) APD, 1 mg mL^–1 MnCO, 5 × 10^8 CFU/mL, 2.6% (w/w) of A-rich strands (5ʹ-acry-AAA AAA AAA AAA AAA AAA AAA-3ʹ), and 1.4% (w/w) C-rich strands (5ʹ-acry-AAA CCC CAA ACC CC-3ʹ) as the dispersed phase in 1 mL water, The flow rate of the dispersed phase and continuous phase were set at 1 μL/min and 100 μL/min. The collected microspheres were filtered under reduced pressure to remove the mineral oil and washed several times with PBS solution. For the control materials, including HD, HAD, CO/HAD, and EcN/CO@HA, only the corresponding components were not used during preparation. After sprayed with gold, the microstructure of microspheres was observed by a Hitachi S-4800 scanning electron microscope (Tokyo, Japan). The morphology of the microspheres in mineral oil and aqueous solution was observed using the Olympus CKX53 microscope (Tokyo, Japan). For fluorescence imaging, the microspheres were pre-stained with GelGreen in the aqueous solution. To investigate the reversible shrinking-swelling transition of microcapsules, the repeated pH cycles of aqueous suspension were adjusted between 2 and 7 or 5 and 7. The changes of particle size were analyzed using ImageView. Protection of EcN against hostile acid environments Free EcN, EcN/CO@HA, and EcN/CO@HAD were resuspended in PBS (5 × 10^8 CFU, 5 mL), respectively. Then, the pH was adjusted to 2 or 5 using HCl to simulate the acidic environment of stomach or duodenum. All samples were incubated in a shaker at 37 °C for 5 h. At the end of the incubation, all samples were transferred to neutral physiological PBS containing 0.2 mM SNP, followed by illuminating at 405 nm for total 1 h with 4 light on/off cycles. Then, 200 μL of EcN suspension was pipetted from each group and evenly spread on LB agar plates. After 24 h incubation at 37 °C, the colonies were photographed and counted. In addition, the suspensions were stained with Live/Dead Bacterial Viability Kit (acridine orange (AO) and propidium iodide (PI)) and then visualized using a fluorescence microscope. Mobile capture of H[2]O[2] CO@HAD microcapsules (2 mg/mL) were mixed with an aqueous solution containing H[2]O[2] (1 mM) with or without SNP (0.02 mM) and their motion was recorded using an inverted microscope. In the presence of SNP, the mixture was irradiated by 405 nm UV light for 15 min. Then, the trajectories of the microspheres were manually tracked using the ImageJ software, and their mean square displacement (MSD) curves were fitted. After 2 h, the H[2]O[2] scavenging percentage was further determined using a commercial kit. For control, the motion of CO@HAD microcapsules was observed in an aqueous solution without H[2]O[2]. Chemotactic motility towards H[2]O[2] A home-made Y-shaped channel was used to investigate the chemotaxis of CO@HAD towards high concentrations of H[2]O[2]. Three chambers of the open Y channel were set up as follows: chamber I was filled with 100 μL of aqueous solution containing 2 mg/mL CO@HAD microspheres; chamber II was filled with 50 μL of 1% agarose gel; chamber III was filled with 50 μL of 1% agarose gel doped with 1 mM H[2]O[2]. To create a spatial chemoattractant concentration, 500 μL of PBS or commercial simulated intestinal fluid was added to the middle channel. Images of the main trunk and the forked branches were taken at different time points using an inverted microscope. In the control group, H[2]O[2] was added to neither chamber II nor chamber III. NO-triggered microcapsule rupture After mixing EcN/CO@HAD (2 mg/mL) with 0.01–0.2 mM SNP, the mixture was immediately irradiated with 405 nm UV light for 15 min to induce the release of NO. The reaction system was then placed in a metal bath at 37 °C and agitated at 500 g. Morphological changes of EcN/CO@HAD were observed and recorded under a microscope at predetermined time points. At each time point, SNP was replenished and irradiated with 405 nm UV light. As a control, EcN/CO@HAD microcapsules were placed in PBS without SNP. Cell culture Mouse fibroblasts (L929) were cultured in RPMI 1640 medium containing 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37 °C in 5% CO[2] atmosphere. Mouse macrophages (RAW 264.7) were cultured in high sugar DMEM containing 10% FBS and 1% penicillin/streptomycin at 37 °C in 5% CO[2] atmosphere. NO scavenging at the cellular level L929 cells (1 × 10^5 cells) were inoculated in 6-well plates for 12 h at 37 °C. Then, the culture medium was replaced with fresh medium containing DETA-NONOate (0.2 mM), DETA-NONOate (0.2 mM) + CO@HD (1 mg/mL), or DETA-NONOate (0.2 mM) + CO@HAD (1 mg/mL). After 6 h incubation, the cells were incubated with DAF-FM DA (5.0 μM) in 1.0 mL of PBS for 30 min at 37 °C. Then, the fluorescence images were captured using inverted fluorescence microscopy. For control, cells were not exposed to DETA-NONOate. ROS scavenging at the cellular level The extracellular ROS scavenging ability of different samples was determined by Reactive Oxygen Species Assay Kit. Briefly, L929 cells (1×10^5 cells) were inoculated in 6-well plates for 12 h at 37 °C. Then, the medium was replaced with FBS-free medium containing 10.0 μM DCFH-DA fluorescent probe. After 30 min, the cells were incubated with Rosup (1.0 μg/mL), HAD (1 mg/mL) + Rosup (1.0 μg/mL), or CO@HAD (1 mg/mL) + Rosup (1.0 μg/mL) for 2 h at 37 °C. Then, fluorescence images were taken by inverted fluorescence microscopy. For control, cells were not exposed to Rosup. In vitro regulation of macrophage phenotype The ability of microspheres to modulate macrophage phenotype was first assessed using immunofluorescence imaging. In briefly, RAW264.7 cells (5×10^4 cells) were inoculated in confocal dishes and cultured for 12 h. Then, cell culture medium was replaced with fresh medium containing 1 μg/mL LPS for an additional 24 h cultivation prior to co-incubation of cells with HD (1 mg/mL), HAD (1 mg/mL), or CO@HAD (1 mg/mL). After 24 h, cell culture medium was discarded, cells were washed three times with PBS, fixed with 2.5 % glutaraldehyde for 15 min, and blocked by immunostaining blocking solution for 60 min at 37 °C. After that, cells were treated with CD206 primary antibody (rabbit anti-CD206 recombinant antibody, Catalog # 81525-1-RR, diluted 1:250, Proteintech) at 4 °C for 1 h. Then, cells were washed three times with PBS, and stained with secondary antibody (FITC-conjugated goat anti-rabbit IgG(H + L), Catalog # SA00003-2, diluted 1:500, Proteintech) in PBS (1% BSA) for 1 h. iNOS was stained by same procedure (rabbit anti-iNOS recombinant antibody, Catalog # 80517-1-RR, diluted 1:250, Proteintech; Cy3-conjugated goat anti-rabbit IgG(H + L), Catalog # SA00009-2, diluted 1:500, Proteintech). Eventually, cells were washed three times with PBS and their nuclei were counterstained with DAPI. In addition, the macrophage phenotypes were characterized using flow cytometry. To this end, LPS-stimulated RAW264.7 cells (1 × 10^5 cells) were co-incubated with HD (1 mg/mL), HAD (1 mg/mL), and CO@HAD (1 mg/mL) for 24 h, respectively. After washing three times with PBS, cell suspensions were prepared by collecting cells from the well plates. Then, PE-labeled CD206 monoclonal antibody (mAb PE anti-CD206, Catalog # 12-2061-82, 0.125 µg/test, eBioscience) and FITC-labeled iNOS monoclonal antibody (mAb FITC anti-iNOS, Catalog # 14-5920-82, 0.25 µg/test, eBioscience) were added to the above cell suspensions. After 1 h of incubation at 4 °C, the cell suspensions were centrifuged at 1500 × g for 10 min, and resuspended in 200 μL of PBS. A Becton Dickinson LSRFortessa^TM flow cytometer (Franklin Lakes, NJ) was used to analyze the macrophage phenotypes, and its built-in FlowJo software was used to analyze the generated data. For control, cells treated with LPS alone and without any treatment were used. Bioimaging of probiotics in the colon To further evaluate the probiotic survival and retention in the colon, EcN carrying pBBR1MCS2-Tac-mCherry plasmid (EcN-mCherry) was encapsulated in microcapsules. Male C57BL/6 (aged 5–6 weeks) were orally administrated by free EcN-mCherry, EcN-mCherry/CO@HD, or EcN-mCherry/CO@HAD (5 × 10^8 CFU/mL, 0.2 mL). After 12 h, the entire gut was collected and imaged with a Caliper Life Sciences IVIS Spectrum Imaging System (Cambridge, MA). Treatment of DSS-induced IBD mice First, male C57BL/6 (aged 5–6 weeks) mice were randomly divided into 7 groups as follows: (1) healthy group: fed with tap water only (DSS(-) + water); (2) water group: fed with tap water containing 3% DSS (DSS(+) + water); (3) EcN group: fed with tap water containing 3% DSS and orally administrated with EcN (5 × 10^8 CFU, 0.2 mL) (DSS(+) + EcN); (4) HD group: fed with tap water containing 3% DSS and orally administrated with HD (2 mg, 0.2 mL) (DSS(+) + HD); (5) HAD group: fed with tap water containing 3% DSS and orally administrated with HAD (2 mg, 0.2 mL) (DSS(+) + HAD); (6) CO@HAD group: fed with tap water containing 3% DSS and orally administrated with CO@HAD (2 mg, 0.2 mL) (DSS(+) + CO@HAD); (7) EcN/CO@HA group: fed with tap water containing 3% DSS and orally administrated with EcN/CO@HA (5×10^8 CFU, 2 mg, 0.2 mL) (DSS(+) + EcN/CO@HA); (8) EcN/CO@HAD group: fed with tap water containing 3% DSS and orally administrated with EcN/CO@HAD (5×10^8 CFU, 2 mg, 0.2 mL) (DSS(+) + EcN/CO@HAD). Typically, DSS was fed from day 0 to day 7 and from day 10 to day 13, and EcN, HD, HAD, EcN/CO@HA, or EcN/CO@HAD was orally administrated from day 8 to day 14. During this period, stool consistency, fecal occult blood, and weight loss of mice were all daily recorded to determine the disease activity index (DAI) score^[253]45. The DAI score was calculated through summarizing the following parameters: weight loss (grade 0, <1%; grade 1, 1–5%; grade 2, 5-10%; grade 3, 10-15%; grade 4, >15%), fecal bleeding index (grade 0, negative hemoccult; grade 2, positive hemoccult; grade 4, blood traces in stool visible), and the stool consistency index (grade 0, normal; grade 2, soft; grade 4, watery diarrhea). All mice were subjected to euthanasia on day 15, then their colons and spleens were photographed and measured in length and weight, respectively. Therapeutic efficacy against DSS-induced colitis Histopathological analysis of colonic injury was performed according to the standard procedures for paraffin embedding and hematoxylin-eosin (H&E) staining. Briefly, colonic tissues were fixed in 4% paraformaldehyde solution, embedded by paraffin, and sectioned for 4 μm prior to H&E staining. The obtained sections were scanned with 3DHISTECH slide scanning system and the colon images were processed using CaseViewer image processing software. The histological scores of colons were assessed by histological scoring guideline^[254]45 (Table [255]S1). Besides, H&E-stained sections of heart, liver, spleen, lung, and kidney of mice after different treatments were observed using an inverted microscope for biocompatibility assessment. Immunofluorescence staining of colon tissue was further used for analysis of inflammatory factors (e.g., IL-6, IL-10, TNF-α, TGF-β) and constituent protein (e.g., ZO-1). Briefly, the collected colon tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into 4 μm sections. Sections were deparaffinized with xylene and rehydrated in graded ethanol and microwaved in 10 mM sodium citrate, then blocked for 30 min at room temperature by adding immunostaining blocking solution, and then incubated with primary antibody (pAb rabbit anti-IL-10, Catalog # GB11534-100, diluted 1:200, Servicebio; pAb rabbit anti-TNF-α, Catalog # GB115702-100, diluted 1:500, Servicebio; pAb rabbit anti-IL-10, Catalog # GB11534-100, diluted 1:200, Servicebio; pAb rabbit anti-TGF-β, Catalog # GB11179-100, diluted 1:200, Servicebio; pAb rabbit anti-IL-6, Catalog # GB11117-100, diluted 1:200, Servicebio; Recombinant mAb rabbit anti-ZO-1, Catalog # GB151981-100, diluted 1:400, Servicebio) overnight at 4 °C and incubated with secondary antibody (FITC-conjugated donkey anti-rabbit IgG (H + L), Catalog # GB22403, diluted 1:100, Servicebio; Cy3-conjugated goat anti-rabbit IgG (H + L), Catalog # GB21303, diluted 1:300, Servicebio) at 37 °C for 50 min. Immunofluorescence staining sections were observed using an inverted fluorescence microscope. For ROS and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining, the sections were stained with dihydroethidium solution and One-step TUNEL Apoptosis Kit, respectively. In addition, the relative mRNA expression of IL-6, IL-10, TNF-α, TGF-β, and ZO-1 in colon tissues was analyzed by quantitative real-time polymerase chain reaction (RT-qPCR). To this end, fresh colon tissue about 25 mg was firstly taken from each group and washed with PBS. Total RNA from each group colon was extracted using RNeasy^TM Animal RNA Isolation Kit with spin column, and then RNA was transcribed into cDNA using BeyoRT^TM III First Stand cDNA Synthesis Kit. Then, cDNA (2 μL) was mixed with PCR reagents 2 μL of primer (3 μM), 10 μL of SYBR green dye (2×), 6 μL of RNase-free H[2]O in a 96-well plate, and transferred in the Applied Biosystems 7500 instrument (Waltham, MA) to start the reaction. The PCR reaction program was set up as follows: 95 °C for 2 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 30 s. The Ct values of the target gene were obtained based on the PCR curve. Relative gene expression was calculated using the ΔΔCt method and expressed as the ratio of target genes to GAPDH. The primers used in this study were listed in Table [256]S2. In vivo regulation of macrophage phenotype in the colon The modulation of macrophage phenotype in the colon was assessed using flow cytometry and immunofluorescence staining. To isolate the cells of the lamina propria from colon, the fresh colon was rinsed well using pre-cooled DPBS (free of calcium and magnesium) to remove feces. The colon was dissected longitudinally, cut into 1.5-2.0 cm segments, and rinsed with DPBS to remove intestinal mucus. After that, the colon segments were incubated in 7 mL of solution I (40 mL DPBS + 0.5 mM EDTA + 10 mM HEPES + 1 mM DTT + 0.8 mL FBS) at 220 g on a shaker at 37 °C for 20 min. The segments were transferred to a new 50 mL centrifuge tube filled with 5 mL of DPBS at 220 g on a shaker at 37 °C for an additional 20 min incubation to remove the residual EDTA and intestinal epithelium. After blotting out water, the colon tissues were transferred into 15 mL centrifuge tubes containing 3 mL of mixed solution (1% penicillin/streptomycin, 1% 75 mg/mL collagenase IV, 98% RPMI 1640 medium), and the centrifuge tubes were shaken vigorously at 37 °C. After 20 min, 3 mL of complete medium was added to terminate digestion. Then, the tissues were gently crushed for 5 s, and filtered through a 70 μm cell sieve and centrifuge. The obtained cells were redispersed with 4 mL of 40% Percoll in a 15 mL centrifuge tube. Then, 80% Percoll solution was added to the upper layer of centrifuge tube so that a clear liquid level demarcation was formed between 40% Percoll and 80% Percoll. After centrifugation using a vertical centrifuge, 40% Percoll was carefully aspirated, and the middle layer was transferred into a 15 mL centrifuge tube to obtain the colon lamina propria immune cell suspension. The cell precipitate was resuspended using PBS, and then subjected to immunofluorescence staining for flow cytometry. Briefly, CD206 antibody (mAb PE anti-CD206, Catalog # 12-2061-82, 0.125 µg/test, eBioscience) and iNOS antibody (mAb FITC anti-iNOS, Catalog # 14-5920-82, 0.25 µg/test, eBioscience) were added to the above cell suspensions, and the suspensions were incubated at 4 °C for 1 h. Afterwards, cells were collected through centrifuging at 1500 g for 10 min, resuspended in 200 μL of PBS, and analyzed using a flow cytometer. Transcriptome sequencing and analysis Sequencing and analysis of the transcriptome were performed with the help of Shanghai Qianya Jinzhi Biotechnology Co., Ltd. The colons of mice were collected, and then RNA was extracted using TRIzol® Reagent. RNA quality was determined using Agilent 2100 Bioanalyser (Santa Clara, CA) and quantified using the NanoDrop ND-2000 (Wilmington, DE). High-quality RNA samples (OD260/280 = 1.8–2.2, OD260/230 ≥ 2.0, RIN ≥ 6.5, 28S:18S ≥ 1.0, >10 μg) were used to construct sequencing library. RNA-seq transcriptome libraries were prepared following TruSeq^TM RNA sample preparation Kit from Illumina (San Diego, CA), using 1 μg of total RNA. Shortly, messenger RNA was isolated with polyA selection by oligo (dT) beads and fragmented using fragmentation buffer. cDNA synthesis, end repair, A-base addition and ligation of the Illumina-indexed adaptors were performed according to Illumina’s protocol. Libraries were then size selected for cDNA target fragments of 200–300 bp on 2% Low Range Ultra Agarose followed by PCR amplified using Phusion DNA polymerase for 15 PCR cycles. After quantified by TBS380, paired-end libraries were sequenced with the Illumina HiSeq PE 2X151bp read length. The raw paired-end reads were trimmed and quality controlled by Trimmomatic with default parameters. Then, clean reads were separately aligned to reference genome with orientation mode using tophat software^[257]75. To identify differential expression genes (DEGs) between the two different samples, the expression level for each transcript was calculated using the fragments per kilobase of exon per million mapped reads (FRKM) method. Cuffdiff^[258]76 was used for differential expression analysis. The DEGs between two samples were selected using the following criteria: (i) the logarithmic of fold change was greater than 2 and the FDR should be less than 0.05. Gene Ontology (GO) functional enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were plotted by [259]https://www.bioinformatics.com.cn, an online platform for data analysis and visualization^[260]77. 16S rDNA sequencing and analysis 16S rDNA gene sequencing was completed with the help of Shanghai Qianya Jinzhi Biotechnology Co., Ltd. The total DNA in the feces was collected, extracted, and tested according to the manufacturer’s instructions, and then sequenced by PCR amplification, library purification, and library quality inspection, followed by the construction of sequencing libraries on the Illumina Hiseq platform. In order to ensure the accuracy of the results of information analysis, the Raw Data was first filtered and processed to obtain Clean Data. OTUs (Operational Taxonomic Units) were clustered according to Clean Data. Based on the clustering results, species annotation was done for each representative sequence to get the corresponding species classification information and species abundance distribution. Based on the abundance spectrum of species after homogenization, the analysis of abundance and diversity index was carried out. Linear discriminant analysis effect size (LEfSe) analysis and clustering heatmap were completed using the Wekemo Bioincloud (https:// [261]www.bioincloud.tech)^[262]78. Statistics and reproducibility All data are represented by mean ± standard deviation (S.D.). Statistical analysis was evaluated using SPSS 27.0.1. Differences among multiple groups were analyzed using one-way analysis of variance (ANOVA) with two-tailed LSD or Bonferroni multiple-comparisons tests. A value of 0.05 was set as the significance level; the data were marked as (*) P < 0.05, (**) P  <  0.01, and (***) P  <  0.001. The P-value above 0.05 was considered as non-significant (n.s.). The results for Figs. [263]3a–d, [264]4a, b, [265]4e–h were obtained after at least three independent repetitions of the experiment. Data were obtained from one representative of three independent experiments in Figs. [266]2c–f, [267]3e–i, [268]4c, d, [269]5i–k, [270]6a, b, j, h. Reporting summary Further information on research design is available in the [271]Nature Portfolio Reporting Summary linked to this article. Supplementary information [272]Supplementary Information^ (10.7MB, pdf) [273]41467_2025_59172_MOESM2_ESM.pdf^ (115.9KB, pdf) Description of Additional Supplementary Files [274]Supplementary Movie 1^ (16.8MB, mp4) [275]Supplementary Movie 2^ (21.8MB, mp4) [276]Supplementary Movie 3^ (13.4MB, mp4) [277]Supplementary Movie 4^ (16.6MB, mp4) [278]Supplementary Movie 5^ (12MB, mp4) [279]Supplementary Movie 6^ (4.3MB, mp4) [280]Reporting Summary^ (111.6KB, pdf) [281]Transparent Peer Review file^ (2MB, pdf) Source data [282]Source Data^ (974.3KB, xlsx) Acknowledgements