Introduction

Intestinal bacteria are the predominant microbial inhabitants of the gut and play a pivotal role in maintaining physical health. The disruption of gut homeostasis is closely associated with the occurrence and development of various diseases1,2,3,4. Intestinal pathogen infections can induce inflammatory responses and severely disrupt the balance of intestinal microbiota. Traditional antibiotic therapy indiscriminately eliminates both pathogenic and commensal/probiotic bacteria, further exacerbating the gut microbiota dysbiosis and influencing metabolic transformations5. Moreover, the rapid emergence and global dissemination of antibiotic-resistant bacteria pose a severe threat to public health6. These challenges underscore the urgent need for innovative therapeutic strategies, such as gut microbiota editing, to precisely modulate levels of specific bacterial species in the gut microbial community. However, the inherent complexity, diversity, and dynamic nature of the intestinal ecosystem pose substantial barriers to elucidating the etiology of dysbiosis and developing targeted intervention strategies. Current intervention approaches, such as probiotics, fecal microbiota transplants, and dietary changes, are limited by transient effects, safety concerns, and unstable outcomes7,8,9. These limitations highlight the necessity of advancing emerging, precise, and reliable approaches for the therapeutic modulation of the gut microbiota.

Bacteriophages (phages), which are bacterial viruses characterized by high host specificity, have a unique ability to precisely target and manipulate specific microbes within the microbiota. Beyond their host specificity, phages exhibit several desirable attributes, including rapid proliferation, abundant resources, and ease of physicochemical modification10. These features make phages promising tools for the precise modulation of the gut microbiome11. Phages have been extensively studied for treating multidrug-resistant infections, including those of the lungs, urinary tract, bones, and joints, and severe systemic conditions such as sepsis12. Despite their potential, phage-based interventions targeting the gastrointestinal tract and gut microbiota remain largely empirical and often lead to variable outcomes13. Case studies on phage therapy underscore that therapeutic efficacy depends on the bactericidal activity, specific biological properties of the phages employed, and their effective concentration at the infection site14,15. Although bicarbonate (e.g., sodium bicarbonate) has been used to facilitate the stomach passage of acid-sensitive phages, the multistep administration process, variable dosage, and potential side effects pose significant limitations to its practical application in phage therapy16,17. More importantly, successful gut microbiota editing requires additional considerations, including oral safety, intestinal retention time, and the intricate interactions between phages, target bacterial species, and other microbiota members. These factors are critical to achieve precise modulation and stable ecological outcomes.

Recently, various delivery carriers, including emulsion, liposomes, polymer particles and hydrogel microspheres, have been developed to enhance the biological function of phages during oral administration18. Among these carriers, hydrogel microspheres have garnered widespread attention owing to their prominent merits, such as high encapsulation efficiency, strong adhesion ability, excellent biocompatibility, and favorable mechanical properties. Although several studies have successfully enhanced the acid stability of encapsulated phages using hydrogel microspheres in vitro, translational research on microsphere-based carriers for phage therapy and gut microbiota editing remains limited, particularly in vivo studies19,20,21. To date, only one study reported an engineered lysogenic phage λ expressing a programmable dCas9 that precisely repressed a targeted E. coli gene in vivo using an oral microsphere carrier22. Electrohydrodynamic spraying (electrospray) offers a simple and innovative method to prepare hydrogel microspheres ranging from the nano- to micrometer scale through the cross-linking and droplet-curing of atomized precursor solution23. Notably, electrospray does not involve organic solvents or thermal processes, making it well suited for encapsulating fragile entities (e.g., probiotics and phages). Additionally, the microsphere size can be tailored by adjusting the process parameters such as the solution concentration and applied voltage.

In this study, to control phage dosage and extend intestinal retention time, we introduce a simple and green preparation method based on edible materials for encapsulating and delivering phages (Fig. 1). First, we evaluate the antibacterial properties of several Salmonella phages and identify an optimal phage combination for therapeutic application. We then use sodium alginate (SA), hyaluronic acid (HA), and Eudragit S100 (ES), all FDA-approved and generally recognized as safe (GRAS), to fabricate SA/HA/ES hydrogel microspheres for targeted phage delivery. Employing the electrospray strategy allows for flexible control over microsphere size (ranging from 100 to 900 μm), enabling customization to meet the diverse requirements of animal models in preclinical trials. SA/HA/ES hydrogel microspheres (HMs) effectively enhance the biological stability of phages and enable their responsive release into the intestine, ensuring stable and targeted phage delivery. Using a Salmonella Typhimurium (STm)-induced dysbiosis mouse model, we demonstrate that HMs-encapsulated Salmonella phages (HMs-Phages) significantly reduce intestinal STm burden and improve the physical condition of treated mice compared to free phages administered empirically. To highlight the therapeutic value of targeted phage delivery, ciprofloxacin (CIP)—a clinically used broad-spectrum antibiotic—is evaluated in parallel as a representative of indiscriminate antibacterial approaches. Importantly, our targeted approach circumvents the severe gut dysbiosis and diarrhea commonly associated with antibiotic treatment, achieving in situ gut microbiota editing. By concurrently enabling targeted pathogen eradication and commensal microbiota conservation, this study paves the way for further investigation into the etiology and therapeutics of microbiota-associated disorders.

Fig. 1: Schematic diagram of the preparation of SA/HA/ES hydrogel microspheres loaded with phages (HMs-Phages) and therapeutic application in a mouse model of Salmonella Typhimurium-induced colitis.

SA/HA/ES hydrogel microspheres were fabricated to load Salmonella phages via electrohydrodynamic spraying and calcium ion crosslinking. The microspheres shielded phages from the harsh gastric environment and enabled pH-responsive release in the intestine, leading to efficient pathogen clearance, reduced proinflammatory cytokines, restoration of gut microbiota homeostasis, and alleviation of colitis symptoms. SA sodium alginate. HA hyaluronic acid. ES Eudragit S100.

Results

Screening of efficient phage combinations

Three Salmonella phages, LPST83, LPST94, and LPST153, were selected for this study based on our previous studies, which demonstrated well-defined genetic backgrounds, broad lytic ranges, and strong antibacterial abilities against Salmonella strains24,25,26. The antimicrobial activity of the phages against the model Salmonella strain was assessed using inhibition curves (Fig. 2a). Transmission electron microscopy (TEM) revealed that all three phages had typical icosahedral heads and tails (Fig. 2b), classifying them within the Caudoviricetes class27. Phylogenetic analysis identified LPST83, LPST94, and LPST153 as members of the genera Segzyvirus, Kuttervirus, and Berlinvirus, respectively24,25,26. These phages displayed differences in tail structure (long and non-contractile tail, contractile tail, and short tail) and plaque morphology (Fig. 2b and Supplementary Fig. 1). The antibacterial results indicated that when used individually, none of the three phages effectively suppressed bacterial growth, regardless of the multiplicity of infection (MOI) (Fig. 2c-e). In contrast, the phage combinations exhibited significantly improved antibacterial activity. Specifically, the combination of LPST94 and LPST153 completely inhibited bacterial growth for 24 h under high MOI conditions (MOI ≥ 10) (Fig. 2f). Similarly, the combination of all three phages effectively controlled pathogen growth for 24 h at an MOI ≥ 100 (Supplementary Fig. 2).

Fig. 2: Morphology characterization and antibacterial activity of Salmonella phages.

a Schematic diagram of the method for evaluating the antibacterial activity of phages against STm. b Transmission electron microscopy images of LPST83, LPST94, and LPST153 phages (n = 3 independent samples per phage). Scale bar, 50 nm. c–f Antimicrobial activity of single phage (LPST83, LPST94, LPST153) or phage combination (LPST94 and LPST153) against STm at different MOI. Data are presented as mean ± standard deviation (n = 3 biological replicates). g Essential conditions for phage application in pathogen control. MOI, multiplicity of infection. Source data are provided as a Source Data file.

The above findings are consistent with those of previous literatures, highlighting that the antibacterial efficacy and therapeutic outcomes of phages are closely dependent on the MOI at the infection site28. Moreover, the combination of multiple phages demonstrated superior efficacy in inhibiting pathogens compared with the use of a single phage. This enhanced performance was possibly due to each phage targeting different bacterial receptors, thus minimizing phage resistance25,26. Using fewer phage participants simplifies the operational and preparative procedures, thereby reducing application complexity. Based on these considerations, the LPST94 and LPST153 combination was selected for subsequent experiments. To further evaluate their efficacy, in vitro phage-resistance of STm was examined. The results showed that LPST94-resistant STm remained sensitive to LPST153 (Supplementary Fig. 3). These in vitro results highlighted three critical conditions for phage application: selecting effective phage participants, achieving a high phage titer, and employing a phage combination (Fig. 2g).

Fabrication and characterization of HMs-Phages

Oral administration is an effective approach to deliver phages to the intestine. However, many acid-sensitive phages are rapidly inactivated in the stomach because of the harsh acidic environment22,29. To address this issue, SA/HA/ES hydrogel microspheres were fabricated using a portable electrospraying platform to protect phages (Fig. 3a). Microscopic images and the microsphere size of the hydrogel microspheres with phages (HMs-Phages) are shown in Fig. 3b and Supplementary Fig. 4. The results showed that microspheres sizes ranged from 133 ± 19 to 890 ± 25 μm (SA/HA/ES-1 to SA/HA/ES-6) depending on the precursor solution concentration (1% to 6%), demonstrating a facile and feasible strategy for tailoring microsphere size. The increase in microsphere size was likely attributable to the higher viscosity of the polymer solution at elevated polymer concentrations, which enhanced polymer chain entanglement and led to the formation of larger atomized microdroplets, ultimately producing larger microspheres (Supplementary Fig. 5). As a comparison, non-electrosprayed SA/HA/ES microspheres were also fabricated using a conventional extrusion-dripping method, exhibiting larger particle sizes and narrow size range (1,744 ± 76 μm to 2,216 ± 25 μm) under corresponding concentrations (Supplementary Fig. 6 and Supplementary Fig. 7). This large physical size significantly limits their suitability for precision gastrointestinal delivery applications in preclinical animal models.

Fig. 3: Preparation and physicochemical characterization of HMs-Phages.

a Schematic diagram of the preparation platform for phage encapsulation in hydrogel microspheres. b Microscopy of different HMs-Phages (n = 3 independent samples with similar results). Scale bar, 500 μm. c SEM images of different HMs-Phages (n = 3 independent samples with similar results). Upper scale bar, 100 μm; lower scale bar, 10 μm. d Elemental mapping of HMs-Phages (SA/HA/ES-3) (n = 3 independent samples with similar results). Scale bar, 100 μm. e Fourier transform infrared spectroscopy of HMs-Phages (SA/HA/ES-3). f Encapsulation efficiency of different hydrogel microspheres for phages. Data are presented as mean ± standard deviation (n = 3 biological replicates). Phages LPST94 and LPST153 were combined for microsphere preparation and subsequent evaluations. SA sodium alginate. HA hyaluronic acid. ES Eudragit S100. HMs hydrogel microspheres. Source data are provided as a Source Data file.

Scanning electron microscopy (SEM) analysis was performed to further examine the surface morphology of the electrosprayed microspheres (Fig. 3c). A more compact and denser surface was observed with increasing precursor solution concentrations, possibly providing better protection for the phages against the harsh environments. A similar phenomenon was also observed in the SEM imaging of non-electrosprayed microspheres (Supplementary Fig. 8). To evaluate the compositional homogeneity within microspheres, cross-sectional transmission electron microscopy was performed. SA/HA/ES-3 was selected as the representative sample due to its regular morphology and uniform size distribution compared to the other groups (Fig. 3b and Supplementary Fig. 4). The TEM micrographs confirmed homogeneous dispersion of the constituent polymers (SA, HA, and ES) throughout the hydrogel network (Supplementary Fig. 9). Furthermore, fluorescence imaging revealed that Cy3-labeled HA (red fluorescence) and AF 488-conjugated ES (green fluorescence) were uniformly distributed within the microsphere matrix (Supplementary Fig. 10).

To demonstrate the successful encapsulation of phages within the microsphere, elemental mapping analysis was conducted. The results showed that the phages were uniformly dispersed in the microspheres, as evidenced by the presence of phosphorus, sulfur, and nitrogen elements (Fig. 3d). In addition, Fourier transform infrared (FTIR) spectroscopy further confirmed phage loading, as indicated by the appearance of an amide II peak at 1545 cm-1, a characteristic marker for phage capsid proteins (Fig. 3e). Consistent with these findings, all SA/HA/ES hydrogel microspheres exhibited high encapsulation efficiencies of approximately 90% (Fig. 3f), indicating a gentle and efficient method for phage loading. Moreover, the number of phages encapsulated in each microsphere rose with increasing microsphere size (Supplementary Fig. 11). Specifically, the phage content for SA/HA/ES-1, SA/HA/ES-2, SA/HA/ES-3, SA/HA/ES-4, SA/HA/ES-5, and SA/HA/ES-6 was 2.8×104, 1.0×105, 4.8×105, 7.9×105, 1.0×106, and 1.3×106 PFU per microsphere, respectively. This controllability in phage load is crucial for determining the appropriate dosage for practical applications.

Protection and release performance of HMs-Phages

Although there are no clear standards specifying the exact phage dosages for practical applications, it is generally accepted that delivering phages in sufficient quantities increases the probability of phages capturing host bacteria, thereby improving therapeutic outcomes30. Thus, it is essential to protect the phage activity from adverse conditions and ensure targeted delivery (Fig. 4a). Free phages (unencapsulated) were completely inactivated after 5 min of incubation in simulated gastric fluid (SGF) at pH 2.5 (Supplementary Fig. 12a). The microsphere composition significantly affects their protective efficacy for encapsulated cargos. To investigate this, phage-loaded microspheres with varying compositions were fabricated using a standardized precursor concentration (1% w/v), and their protective ability against internal phages in simulated gastric fluid was evaluated. Quantitative analysis revealed that after 2 h of incubation, encapsulated phages in pure SA-1 microspheres were completely inactivated (Supplementary Fig. 13). In contrast, SA/HA-1 microspheres retained approximately 2 × 103 PFU/g of viable phages, and SA/ES-1 microspheres preserved 4 × 104 PFU/g. Notably, SA/HA/ES-1 microspheres exhibited the highest phage survival, retaining around 1 × 106 PFU/g, indicating the most effective protective effect on phages. The preserved high phage viability directly correlated with antibacterial efficacy at target sites. Specifically, encapsulated phages in SA/HA/ES-1 microspheres demonstrated superior antimicrobial performance in simulated intestinal fluid (SIF) following SGF pretreatment (Supplementary Fig. 14). Based on these findings, the SA/HA/ES was identified as optimal formulation for further experiments.

Fig. 4: Protective effect and responsive release of HMs-Phages.

a Schematic diagram of the biological function of HMs-Phages in the gastrointestinal tract. b Survival of encapsulated phages (LPST94 and LST153) in microspheres after exposure to SGF (pH 2.5) supplemented with pepsin (0.32%) at 37 °C. Data are presented as mean ± standard deviation (n = 3 biological replicates). c Cumulative release of phages from microspheres in SIF (pH 7.2) at 37 °C after pre-incubation in SGF for 2 h. Data are presented as mean ± standard deviation (n = 3 biological replicates). Phages were quantified by plating 100 µL of diluted solution with host bacteria on double-layer agar plates and incubating at 37 °C for 12 h. d Fluorescence intensity of simulated digestive fluid after incubation with SA/HA/ES-3 microspheres loaded with rhodamine B-labeled phages. Data are presented as mean ± standard deviation (n = 3 biological replicates). e Microscopy of SA/HA/ES-3 hydrogel microspheres loaded with rhodamine B-labeled phages after treatment with SGF and SIF (n = 3 independent samples with similar results). Scale bar, 500 μm. SA sodium alginate. HA hyaluronic acid. ES Eudragit S100. SGF simulated gastric fluid. SIF simulated intestinal fluid. PFU plaque-forming units of phage. Source data are provided as a Source Data file.

To further evaluate the impact of precursor solution concentration on microsphere performance under gastric conditions, SA/HA/ES microspheres prepared at varying solution concentrations were systematically evaluated. As shown in Fig. 4b, after 2 h of incubation in SGF, SA/HA/ES-3, SA/HA/ES-4, SA/HA/ES-5, and SA/HA/ES-6 microspheres exhibited higher phage survival rates (negligible loss) than SA/HA/ES-1 (2-log loss) and SA/HA/ES-2 (1-log loss), demonstrating their good protective capability against the harsh gastric environment. In SGF with a lower pH of 1.2, the survival rate of encapsulated phages decreased over time (Supplementary Fig. 15). Microspheres formed from precursor solutions with higher concentrations exhibited enhanced protective effects, which is consistent with the trends observed at pH 2.5. In addition, the acid resistance of non-electrosprayed SA/HA/ES microspheres was also performed. A similar concentration-dependent protective effect was observed (Supplementary Fig. 16), consistent with that of the electrosprayed microspheres. This enhanced protection may be related to the denser gel network formed by the solution with increased polymer concentration, which provides the interior phages with an enhanced physical barrier, preventing or reducing their exposure to acidic media.

To evaluate phage release from the hydrogel microspheres, the viability of free phages in SIF containing trypsin and bile salt was assessed. Free phages remained stable in SIF for up to 12 h without significant changes in the titer (Supplementary Fig. 12b). As illustrated in Fig. 4c, 60, 44, 38, 33, 27, and 24% of the phages were released from microspheres in different groups (SA/HA/ES-1 to SA/HA/ES-6) after 1 h of incubation in SIF, followed by sustained release over 6–10 h, contingent on the microsphere formulation. Most phages were released within the first 6 h, and a faster release behavior was observed in microspheres formulated with a low solution concentration. This behavior was likely due to the difference in the internal gel-network structure of the hydrogel microspheres.

To visualize the release profile and swelling behavior of microspheres in simulated gastrointestinal environments, rhodamine B-labeled phages were encapsulated in SA/HA/ES-3 microspheres with uniform size distribution and tracked using fluorescence imaging. Fluorescence images showed that rhodamine B-labeled phages were effectively released from microspheres in SIF, confirming successful intestine-targeted delivery (Fig. 4d). Also, the physical integrity of the SA/HA/ES-3 microspheres remained intact in stomach conditions (Fig. 4e). Upon transfer to SIF, the microspheres absorbed water and swelled rapidly within the first 4 h, and then gradually disintegrated, suggesting a pH-responsive release behavior. Phage release from the hydrogel network was primarily governed by a swelling-dissolution process. Under neutral or alkaline SIF conditions, the gradual dissolution of ES and HA, along with Ca2+ dissociation from alginate, led to the swelling and eventual disintegration of the microspheres (Supplementary Fig. 17). This highlights the flexibility of controlling the phage release by adjusting the microsphere size and precursor solution concentration.

Biosafety evaluation of HMs-Phages

Many traditional antimicrobial materials exhibit significant cytotoxicity or systemic toxicity, which seriously hinders their practical application31,32. To evaluate the clinical potential of HMs-Phages, in vitro biocompatibility experiments (MTT assay and SYTO9/PI staining) were conducted. The SA/HA/ES-1 hydrogel microspheres were selected based on their optimized size distribution (133 ± 19 μm), enabling compatibility with in vivo delivery protocols. Two types of intestinal epithelial cells (Caco-2 and HT-29) were incubated with varying concentrations of HMs-Phages (Fig. 5a). The MTT assay revealed no significant toxicity of HMs-Phages toward either Caco-2 or HT-29 cells, with over 90% cell viability maintained across all tested concentrations (Fig. 5b). Consistently, the cell staining results demonstrated no significant reduction in cell viability at any microsphere concentration compared with those of the control group, further confirming excellent cytocompatibility (Fig. 5c). The low cytotoxicity of HMs-Phages can be primarily attributed to the safety features of SA, HA, and ES, which serve as microsphere components.

Fig. 5: Biosafety and intestinal retention of HMs-Phages.

a Schematic diagram of the co-culture system of microspheres and epithelial cells for biosafety assessment. b Cytotoxicity of HMs-Phages (SA/HA/ES-1) in Caco-2 and HT-29 cells after 24 h of co-incubation. Data are presented as mean ± standard deviation (n = 5 biological replicates). c Live/Dead staining of Caco-2 and HT-29 cells after co-culture with different concentrations of HMs-Phages (SA/HA/ES-1) for 24 h (n = 3 biological replicates with similar results). Live cells were stained green, and dead cells were stained red. Scale bar, 100 μm. d Representative histological images of major organs (heart, liver, spleen, lung, and kidney) stained with hematoxylin-eosin (H&E) in the Health and HMs-Phages groups (n = 3 biological replicates with similar results). Scale bar, 100 μm. e Retention of rhodamine B-labeled free phages and HMs-Phages (SA/HA/ES-1) in the murine gastrointestinal tract at different time points (n = 3 biological replicates per time point with similar results). HMs-Phages, hydrogel microspheres with phages. Source data are provided as a Source Data file.

To further assess the biosafety of SA/HA/ES-1 hydrogel microspheres, in vivo toxicity assay was conducted. As shown in Supplementary Fig. 18, no significant differences in body weight were observed between the healthy control and oral microsphere-treated groups over the 7-day administration period. In addition, histological analysis revealed no evident signs of inflammation in the examined tissues after 7 days of microsphere administration compared with that in the healthy control group (Fig. 5d). These results demonstrated that HMs-Phages possessed excellent biocompatibility and metabolic activity, highlighting their potential as promising hydrogel-based carriers for oral phage delivery.

In vivo intestinal retention of HMs-Phages

The gastrointestinal retention of HMs-Phages was assessed in mice using a fluorescence imaging system. After the oral administration of SA/HA/ES-1 hydrogel microspheres containing rhodamine B-labeled phages, fluorescence signals were observed to track intestinal transport and retention time. As shown in Fig. 5e, mice administered with free phages displayed high fluorescence intensity in the jejunum at 1 h and in the ileum at 4 h, with fluorescence completely vanishing at 8 h because of metabolism and intestinal excretion. In contrast, HMs-Phages showed similar fluorescence signals in the stomach and small intestine at 1 h and 4 h, respectively. However, fluorescence signals in the small intestine and colon persisted for longer durations, remaining detectable at 8 and 12 h before disappearing by 24 h, indicating prolonged retention of the hydrogel microspheres in the gut. These results indicate that the SA/HA/ES hydrogel microspheres possess strong intestinal retention ability, making them ideal carriers for targeted intestinal delivery.

HMs-Phages efficiently treat Salmonella-induced colitis

A murine model of STm-induced colitis was established to investigate the therapeutic efficacy of HMs-Phages (Fig. 6a). Intestinal colonization by STm typically causes severe local or systemic infections and inflammatory responses, leading to a constant decline in body weight and food intake in mice (Fig. 6b, c). The free phage treatment group (Phages) failed to alleviate weight loss or improve food intake. In contrast, treatment with SA/HA/ES-1 microspheres with phages (HMs-Phages group) exhibited excellent therapeutic efficacy in increasing body weight and food intake, comparable to ciprofloxacin administration (CIP group). Diarrhea is a hallmark of bacterial infection and enteric inflammation. The feces of the uninfected healthy group remained firm, well-formed, and dark-brown throughout the experiment, with a fecal score of 5 points (Fig. 6d, e). In contrast, after STm infection (PBS group), the feces changed from normal to watery and the color shifted from dark-brown to yellowish-brown on day 5 (fecal score of 1 point), indicating severe diarrhea. Fecal samples from the Phages group showed marginal improvements in hardness, shape, and color on day 3 compared to the PBS group. After day 5, however, these improvements deteriorated progressively, ultimately reaching parity with the PBS group. Both the HMs-Phages and CIP interventions effectively alleviated diarrhea, with higher fecal scores. Notably, mild diarrhea reoccurred in the CIP group starting from day 5, which is presumed to be associated with antibiotic-induced gut dysbiosis. In addition, dynamic changes in bacterial and phage numbers in the mouse intestine were monitored. The Phages group exhibited minor variations in fecal bacteria numbers during treatment, with phage counts ranging from 104 to 105 PFU/g of feces (Fig. 6f, g). Furthermore, the individual LPST94 number in the murine gut of the Phages group was higher than that of LPST153, possibly due to its stronger acid tolerance than that of LPST153 (Supplementary Fig. 19). The HMs-Phages group maintained consistently high phage titers (Fig. 6g), and exhibited an approximately 2000-fold reduction in bacterial counts compared to the PBS group, leading to 100% survival protection in mice (Fig. 6f and Supplementary Fig. 20). Moreover, the individual numbers of LPST94 and LPST153 in the HMs-Phages group remained at equivalent levels, primarily due to the stable intestinal phage delivery provided by the microsphere carriers (Supplementary Fig. 19). Based on the observed findings, this sustained reduction in fecal STm numbers was principally attributed to sufficient gastric protection of the phages by the microspheres, prolonged phage retention, and targeted phage release, thereby ensuring high phage concentration and long antibacterial time at the site of action.

Fig. 6: Therapeutic effect of HMs-Phages on STm-induced colitis in mice.

a Establishment of the bacterial colitis model and therapeutic interventions in mice, including streptomycin pretreatment and Salmonella infection, followed by different treatments [PBS (100 μL/day, 0 PFU/mL), HMs (100 mg/day, 0 PFU/g), Phages (100 μL/day, 109 PFU/mL), HMs-Phages (100 mg/day, 109 PFU/g), and CIP (100 μL/day, 2 mg/mL)]. b Body weight and (c) cumulative food intake in the different groups. Data are presented as mean ± standard deviation (n = 6 mice per group). d Representative fecal images and (e) stool scores of mice in the different groups on days 1, 3, 5, 7 and 9. Data are presented as mean ± standard deviation (n = 6 mice per group). f Dynamic changes in bacterial number and (g) phage titers per gram of feces in the different treatment groups (n = 6 mice per group). The central line of box plots represents the median, box bounds indicate the 25th and 75th percentiles, and whiskers extend to the minimum and maximum values. HMs hydrogel microspheres. HMs-Phages, hydrogel microspheres with phages. CIP, ciprofloxacin. Source data are provided as a Source Data file.

Colon shortening is a distinct feature of STm-induced colitis. Both the HMs-Phages and CIP groups showed a positive effect in the recovery of colon length (Fig. 7a, b). Splenomegaly is another visible characteristic of STm infection due to the activation, proliferation, and infiltration of immune cells in the spleen (Fig. 7c, d). Compared with other groups, the CIP group exhibited the strongest suppression of splenomegaly, followed by the HMs-Phages group, with no statistically significant difference between these two treatments (p > 0.05). Notably, intestinally colonized STm can invade other parenteral organs such as the spleen and liver. As shown in Fig. 7e–g, STm burdens in the spleen, liver, and colon were significantly decreased by 3-log units in the CIP group compared with the PBS group, and the HMs-Phages group exhibited comparable antibacterial efficacy. Regarding in vivo phage resistance, STm in both the Phages and HMs-Phages groups remained sensitive to LPST94 and LPST153 throughout the treatment period, indicating no phage resistance development (Supplementary Fig. 21).

Fig. 7: Antibacterial efficacy of HMs-Phages on STm-induced colitis in mice.

Colon tissues were imaged (a) and measured for length (b), and spleens were imaged (c) and recorded for weight (d) (n = 6 mice per group). Scale bar, 10 mm. Bacterial burden was determined in spleen (e), liver (f), and colon (g) (*n *= 6 mice per group). h–j Proinflammatory cytokine levels in colon tissues were analyzed (n = 6 mice per group). k Representative histological images of colon tissue from different groups (n = 3 biological replicates with similar results). Scale bar, 200 μm. Statistical significance was determined by one-way ANOVA with Tukey’s post hoc test (panel b, d, h–j) or unpaired two-tailed Student’s t test (panel e-g) (*p < 0.05; **p < 0.01; ***p < 0.001; ns, no significance). Data are presented as mean ± standard deviation. Exact *p-*values: panel (b): Health vs. PBS, p < 0.0001; PBS vs. Phages, p = 0.1076; PBS vs. HMs-Phages, p < 0.0001; Phages vs. HMs-Phages, p = 0.0070; HMs-Phages vs. CIP, p = 0.9713; panel (d): Health vs. PBS, p < 0.0001; PBS vs. Phages, p = 0.8776; PBS vs. HMs-Phages, p = 0.0015; Phages vs. HMs-Phages, p = 0.0258; HMs-Phages vs. CIP, p = 0.8499; panel (e): PBS vs. Phages, p = 0.4972; Phages vs. HMs-Phages, p = 0.0085; HMs-Phages vs. CIP, p = 0.0464; panel (f): PBS vs. Phages, p = 0.6731; Phages vs. HMs-Phages, p = 0.0144; HMs-Phages vs. CIP, p = 0.0473; panel (g): PBS vs. Phages, p = 0.1759; Phages vs. HMs-Phages, p = 0.0246; HMs-Phages vs. CIP, p = 0.1613; panel (h): Health vs. PBS, p < 0.0001; PBS vs. HMs-Phages, p < 0.0001; HMs-Phages vs. CIP, p = 0.1026; panel (i): Health vs. PBS, p < 0.0001; PBS vs. HMs-Phages, p < 0.0001; HMs-Phages vs. CIP, p = 0.0522; panel (j): Health vs. PBS, p < 0.0001; PBS vs. HMs-Phages, p < 0.0001; HMs-Phages vs. CIP, p = 0.0179. HMs, hydrogel microspheres. HMs-Phages, hydrogel microspheres with phages. CIP, ciprofloxacin. Source data are provided as a Source Data file.

Proinflammatory cytokine levels serve as key indicators of inflammatory response severity. As presented in Fig. 7h–j, STm infection significantly elevated the levels of inflammatory cytokines (TNF-α, IL-6, and IL-1β) compared to the healthy group. Oral CIP administration significantly reduced proinflammatory cytokine levels relative to the PBS group, decreasing TNF-α by 71.7%, IL-6 by 74.1%, and IL-1β by 68.9%. Remarkably, the HMs-Phages group exhibited a comparable anti-inflammatory effect, with reductions of 60.8% for TNF-α, 62.8% for IL-6, and 54.6% for IL-1β. The alleviation of the inflammatory response by HMs-Phages treatment was mainly attributed to targeted phage delivery and effective pathogen clearance. Colonic histological analysis revealed severe tissue injury in the PBS group, which was characterized by crypt damage, reduced goblet cell numbers, and mucosal lesions (Fig. 7k). In contrast, these structural damages were significantly improved in both the HMs-Phages and CIP groups. The histopathology scores of the HMs-Phages treatment group were substantially lower than Phages groups, showing an 81.1% reduction compared with that of the PBS group (Supplementary Fig. 22).

HMs-Phages precisely edit the gut microbiota

Maintaining a balanced gut microbiota is crucial for nutritional metabolism, intestinal barrier function, and immune function. Prior to experimental interventions, the baseline gut microbiota of mice in different groups was analyzed. Taxonomic profiling revealed a similar microbial composition and abundance across all groups (Supplementary Fig. 23). β-diversity analysis further indicated no significant intergroup variation, with no apparent clustering in PCoA space, suggesting a similar baseline gut microbiota structure among experimental cohorts (Supplementary Fig. 24).

Following STm infection and different treatments, significant alterations in microbial communities were observed. Compared to the healthy group, STm infection (PBS group) drastically decreased the observed operational taxonomic units (OTUs), Shannon index, and Simpson index, suggesting a marked reduction in microbial richness and diversity (Fig. 8a–c). Interestingly, the CIP group also exhibited reduced microbial diversity, likely due to its broad-spectrum antibacterial effect. Notably, oral administration of HMs-Phages significantly restored the richness and diversity of the gut microbiota to levels approaching healthy controls. In addition, pathogen colonization significantly disrupted the gut microbiota structure, causing severe dysbiosis (Fig. 8d, e). At the family level, STm colonization significantly reduced the relative abundance of beneficial taxa such as Bacteroidaceae, Lachnospiraceae, and Prevotellaceae relative to healthy controls (Supplementary Fig. 25). These taxa are known to play key roles in dietary fiber degradation and short-chain fatty acid production33. Following treatment, HMs-Phages substantially restored the relative abundance of these beneficial microbes toward levels observed in healthy controls. At the genus level, differential abundance analysis revealed pronounced alterations in the bacterial community of the PBS group, with 23.11% of genera significantly upregulated and 17.47% significantly downregulated relative to healthy group (Supplementary Fig. 26). Compared to the PBS group, the Phages group exhibited a largely similar microbial community structure, with 99.21% of genera showing no significant abundance changes (Fig. 8f). Conversely, the HMs-Phages group exhibited a markedly distinct microbial community compared to that of PBS group, with