This study enrolled 39 inflammatory bowel disease (IBD group) patients and 42 age-matched healthy controls (HC group) (age range: 20–70 years) from the Department of Gastroenterology at Shenzhen Guangming District People’s Hospital. The study protocol received ethical approval from the Institutional Review Board of Shenzhen Guangming District People’s Hospital (Approval No. LL-KT-2025006), with written informed consent obtained from all participants in accordance with the Declaration of Helsinki. IBD participants met the following inclusion criteria: (1) Diagnosis confirmed through standard clinical evaluation; (2) No familial history of gastrointestinal disorders; (3) Absence of antibiotic or immunomodulatory therapy within 30 days prior to enrollment. Exclusion criteria included: pregnancy, lactation, gastrointestinal surgery, other autoimmune diseases, and severe allergies. All IBD patients enrolled in the study were in an clinical remission state at the time of sample collection.

Healthy controls were required to: (1) Have no active medical conditions; (2) Refrain from antibiotic/immunotherapy use for 30 days pre-enrollment. Exclusion criteria for all participants included pregnancy, lactation, immunodeficiency disorders, and severe allergies. All patients were asked to follow their regular diet before data and sample collection.

Fecal specimens were collected preprandially using sterile, gas-impermeable anaerobic collection tubes (Oxoid AnaeroGen, Thermo Fisher, USA) that were pre-filled with anaerobic gas-generating sachets to create an oxygen-free environment. The samples were immediately sealed and transported on ice to the laboratory within 30 minutes. Upon arrival, the samples were processed in an anaerobic chamber (Whitley DG250, Don Whitley Scientific, UK; 85% N₂, 10% H₂, 5% CO₂) to minimize oxygen exposure. Aliquots intended for transplantation or sequencing were flash-frozen in liquid nitrogen and stored at −80°C to preserve microbial viability and community integrity (28).

Fecal genomic DNA was extracted using the HiPure Stool DNA Mini Kit (Magen, Guangzhou, China) following the manufacturer’s protocol. The V3-V4 hypervariable regions of bacterial 16S rRNA genes were amplified with specific primers (341F: 5’-CCTACGGGNGGCWGCAG-3’; 805R: 5’-GACTACHVGGGTATCTAATCC-3’) in 30 μL reaction volumes. Thermal cycling parameters included: initial denaturation at 95°C for 3 min; 25 cycles of denaturation (95°C, 30 s), annealing (55°C, 30 s), and extension (72°C, 15 s); followed by final extension at 72°C for 5 min.

PCR products were purified using the Agencourt AMPure XP system (Beckman Coulter, USA) and quality-verified through 1.5% agarose gel electrophoresis. Target fragments were recovered using the AxyPrep DNA Gel Extraction Kit (AP-GX-250G, Corning, USA). Library concentrations were quantified with the Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific, USA) and sequenced on the Illumina MiSeq platform (2×300 bp paired-end) for subsequent bioinformatic analysis.

The minimum read depth for each sample was 50,000 reads and the Q30 value was >80%. Following initial quality assessment with FastQC, raw sequencing reads were processed through the DADA2 pipeline implementing the following parameters: truncation of reads containing ambiguous bases (N) or exceeding estimated error thresholds (>2), succeeded by de novo chimera elimination. High-fidelity sequences were clustered into operational taxonomic units (OTU) at 97% similarity using the UPARSE algorithm. Taxonomic classification was executed against the SILVA 138.1 reference database. α-diversity was quantified using Shannon indices, whereas beta diversity analysis was conducted through the application of Bray-Curtis dissimilarity matrices and Bray-Curtis, with results visualized via principal coordinates analysis (PCoA). Differential taxa identification was conducted via LEfSe with significance thresholds set at LDA score >3.0 and p < 0.05. All statistical analyses and visualizations were implemented in R statistical environment (v4.4.2).

LEfSe was employed to identify discriminative microbial taxa between IBD patients and healthy controls. Health-enriched taxa (LDA score >2.0, p < 0.05) were prioritized for donor stratification. Healthy donors were ranked according to median relative abundance thresholds of these signature taxa, with the top and bottom quartiles designated as high-abundance (H donor) and low-fabundance donor(L donor) cohorts, respectively.

Fecal specimens (50 g) from healthy donors were transferred into an anaerobic chamber and homogenized in sterile, pre-reduced physiological saline (0.9% NaCl) at a 1:3 (w/v) ratio using homogenizers. The resultant slurry was subsequently filtered through sterile nylon mesh under anaerobic conditions to yield bacterial suspensions. These suspensions were then aliquoted into sterile, anaerobic cryovials, rapidly frozen in liquid nitrogen, and cryopreserved at -80°C (28, 29).

Male C57BL/6J mice (6–8 weeks old) were used in this study to ensure consistency and avoid potential confounding effects of hormonal fluctuations that may influence the gut microbiota and immune responses. 66 male C57BL/6J mice were purchased from the germfree animal platform of Huazhong Agricultural University. Mice were maintained under controlled SPF conditions with temperature 25 ± 2°C, relative humidity 45-60%, and 12 h light/dark cycles. After 7-day acclimatization with ad libitum access to food and water, animals were randomly allocated into 11 experimental groups: control group, Model group (DSS), high-abundance donor A group (H-A), high-abundance donor B group (H-B), high-abundance donor C group (H-C), high-abundance donor D group (H-D), low-abundance donor A group (L-A), low-abundance donor B group (L-B), low-abundance donor C group (L-C), low-abundance donor D group (L-D), and the 5-ASA group. This study was approved by the Ethics Committee of Shenzhen Guangming District People’s Hospital (approval number: LL-KT-2024108).

Mice in all groups except the control group drank 3% DSS (MP Biomedicals, Santa Ana, United States) solution, and after 10 days, mice in the 5-ASA group (100 mg/kg, JiaxingSiCheng Chemical Co., Ltd., China) were administered by gavage daily. The other groups were given PBS, H and L donor fecal bacteria (0.2 mL) by gavage on days 5, 7, 9, 11, 13 and 15, respectively. Prior to administration, the fecal bacteria were thawed at room temperature for 30 minutes under anaerobic conditions and were immediately gavaged into germ-free mice within a biosafety cabinet that was continuously flushed with an anaerobic gas mixture to mitigate the risk of oxidation. During the experimental period, mice were scored with DAI, blood was taken from the orbits on day 19, the colon was taken to measure the length and photographed, and the contents of the mouse cecum were collected.

Disease Activity Index (DAI) scores, which assess body weight loss, stool viscosity, and the presence of blood in stool, were evaluated on days 1, 4, 7, 10, 13, 16, and 19. The scoring criteria are detailed Table 1 .

Cecal contents were collected, immediately flash-frozen in liquid nitrogen, and stored at −80°C until further processing. Gut microbiota profiling and subsequent bioinformatic analysis were performed according to the protocols detailed in Methods 2.2 (16S rRNA sequencing) and 2.3 (Bioinformatic analysis).

Statistical analyses were conducted in QIIME2 (v2024.2). α-Diversity was quantified using the Shannon index, with group differences assessed by Kruskal-Wallis rank sum test. Bray-Curtis distances were computed for the Anosim analysis, while Bray-Curtis distances were utilized for principal coordinate analysis (PCoA) to evaluate beta diversity. The significance of the findings was assessed through the application of PerMANOVA. LEfSe analysis was performed using Wilcoxon and Kruskal-Wallis tests for biomarker identification and Benjamini-Hochberg for multiple testing. Wilcoxon rank sum test was used to compare the groups in animal experiments. Data are presented as mean ± standard deviation, with p < 0.05 considered statistically significant.

To investigate gut microbiota disparities between IBD patients and HC, we enrolled 39 treatment-naïve IBD patients and 42 age-matched HC. Fecal samples underwent 16S rRNA gene sequencing using the Illumina NovaSeq platform. α-diversity analysis revealed significantly lower Shannon indices in IBD patients compared to HC (p < 0.001) ( Figure 1A ), indicating reduced microbial diversity. β-diversity analysis is employed to investigate alterations in the overall composition of the gut microbiota. An Anosim analysis was conducted to assess whether the differences between groups were significantly greater than the differences observed within groups. The results demonstrated significant disparities between the IBD group and HC (R = 0.215, p < 0.001, Figure 1B ). PCoA based on Bray-Curtis distances showed clear separation along the first principal coordinate (PCoA1: 41.3% variance explained) ( Figure 1C ), confirming structural divergence of microbial communities.

We conducted taxonomic profiling of the top 10 microbial species in IBD patients. At the phylum level, significant intergroup disparities were observed in Firmicutes, Proteobacteria, Bacteroidota, and Actinobacteriota, with IBD patients exhibiting marked enrichment of Proteobacteria (IBD: 33.8%, HC: 9.5%) ( Figure 1D ). Genus-level analysis revealed differential abundance of Escherichia-Shigella, Bifidobacterium, Prevotella, Faecalibacterium, Bacteroides, Blautia, Megamonas, Klebsiella, Agathobacter, and Ruminococcus. Notably, the gut microbiota of IBD patients exhibited marked enrichment of Escherichia-Shigella (5.5-fold increase), Megamonas (1.8-fold), and Klebsiella (2.2-fold) compared to healthy controls ( Figure 1E ). These genera (particularly Escherichia-Shigella and Klebsiella) are established pro-inflammatory pathogens implicated in mucosal inflammation (30, 31). Conversely, putative beneficial taxa including Bifidobacterium, Faecalibacterium, and Agathobacter showed 12%-53% reductions in relative abundance compared to controls ( Figure 1E ) (32, 33). Collectively, these findings substantiate IBD-associated dysbiosis characterized by depleted microbial diversity and taxonomic restructuring.

To characterize the donor-specific microbial profiles, we employed LEfSe to identify differentially abundant taxa between IBD patients and HC. LEfSe was employed to identify microbial biomarkers between groups. Phylogenetic cladograms generated through LDA scores (>2.0) and LEfSe analysis revealed taxonomic cladogram differences across microbial communities ( Figures 1F, G ). We further identified 50 bacterial genera with significant differences in abundance between IBD patients and HCs based on Lefse analysis ( Figure 2A ). Heatmap visualization revealed distinct enrichment patterns: 36 putative beneficial taxa (e.g., Firmicutes, Bacteroidota, Actinobacteriota, Clostridia) dominated HC donors, whereas 14 potentially pathogenic lineages (Proteobacteria, Gammaproteobacteria, Enterobacterales) were enriched in IBD donors ( Figure 2A ). Inter-donor heterogeneity in microbial stratification was evident ( Figure 2B ). Based on median relative abundance of the 36 beneficial genera, donors were stratified into 4 high-abundance (HC-8, HC-5, HC-7, HC-6) and 4 low-abundance (HC-26, HC-41, HC-22, HC-39) cohorts ( Figure 2C ). Comparative analysis demonstrated marked enrichment of Coprococcus, Dorea, and Butyricicoccus in H donors, with significantly reduced beneficial taxa in L donors ( Figure 2D ). Our findings demonstrate compositional stratification of gut microbiota across donor populations based on the relative abundance of health-associated bacterial genera.

To investigate the therapeutic implications of donor microbial composition, we established a dextran sulfate sodium (DSS)-induced murine colitis model ( Figure 3A ). Mice were randomized to receive FMT from either H or L donors, alongside a control treatment of 5-ASA. DSS administration induced significant colonic mucosal damage, as evidenced by shortened colon length ( Figures 3B, C ). Disease progression was characterized by progressive weight loss, hematochezia, and diarrhea ( Figures 3D, E ). Notably, recipients of FMT exhibited significant attenuation of colonic injury, reversal of weight loss trajectories, and amelioration of hematochezia and diarrhea. Importantly, recipients of H-donor FMT demonstrated superior therapeutic efficacy compared to those receiving L-donor FMT and the 5-ASA treatment group. The results showed that the mice receiving FMT from H donors showed a significantly different trend in weight change from IBD mice on the 7th day, and IBD symptoms such as hematochezia and diarrhea were significantly improved on the 13th day. In contrast, mice receiving FMT from L donors showed significant disease improvement only after multiple treatments ( Figures 3B–E ). These findings collectively demonstrate that donor microbial composition critically determines FMT efficacy in ameliorating IBD pathology.

To assess the impacts of donor microbial composition on microbial ecology, we conducted gut microbiota profiling in IBD mice following FMT. The α-diversity analysis revealed significantly reduced Shannon indices in IBD mice compared to control mice (p = 0.02), indicating dysbiosis ( Figure 4A ). Notably, recipients of H-donor FMT exhibited a marked restoration of α-diversity, approximating healthy controls (p < 0.001 vs model group), whereas L-donor recipients showed intermediate recovery (p = 0.02 vs H-group) ( Figure 4A , Supplementary Figure 1A ). Anosim analysis revealed statistically significant differences in gut microbiota across all groups ((R = 0.388, p < 0.001), ( Figure 4B , Supplementary Figure 1B ). The analysis of β-diversity employing Bray-Curtis-based PCoA revealed distinct microbial community restoration patterns (R (2) = 0.254, p = 0.001), the H group exhibited near-complete restitution of gut microbiota diversity, whereas the L group exhibited significantly lower diversity indices ( Figure 4C ).

Phylum-level analysis revealed significant microbial shifts in IBD mice, characterized by a decrease in the abundance of Firmicutes and Bacteroidota, accompanied by an increase in Verrucomicrobiota, Proteobacteria, Actinobacteriota, and Fusobacteriota ( Figure 4D , Supplementary Figure 1C ). The FMT intervention, particularly from H-donors, restored Firmicutes to near-normal levels, demonstrating superior efficacy compared to 5-aminosalicylic acid (5-ASA) treatment ( Figure 4D ). At the genus level, DSS-induced colitis mice exhibited marked microbial dysbiosis, characterized by an increased relative abundance of Bacteroides (3.4-fold), Lachnospiraceae_NK4A136_group (1.1-fold), Akkermansia (33.8-fold), Streptococcus (1.5-fold), Bifidobacterium (2.1-fold), Fusobacterium (1.9-fold), and Parabacteroides (2.0-fold), concomitant with significant depletion of beneficial taxa including Lactobacillus (0.23-fold) and Dubosiella (0.38-fold). Both FMT and 5-ASA treatments improved gut microbiota dysbiosis in IBD mice. Compared to untreated IBD mice, FMT intervention significantly enriched Lactobacillus (5.9-fold increase) and Dubosiella (1.6-fold increase)—the latter being a murine commensal known for its anti-inflammatory properties (34). Furthermore, FMT normalized the relative abundance of Lachnospiraceae_NK4A136_group, Akkermansia, Turicibacter, and Parabacteroides to near-physiological levels. Notably, H-donor FMT effectively normalized the DSS-induced elevation of Bacteroides ( Figure 4E , Supplementary Figure 1D ). These findings highlight the superior therapeutic efficacy of FMT compared to 5-ASA in microbial restoration, with H-donors achieving optimal ecological recovery.

Subsequently, we performed LEfSe analysis on the gut microbiota of IBD mice and treatment groups. The LDA scores (>3.0) revealed taxonomically enriched biomarkers among the cohorts, indicating a compositional divergence in the microbiota ( Figure 4F ). Phylogenetic cladograms illustrated structural community disparities ( Figure 4G ). Significant differences in microorganisms between control and IBD mice were assessed using LEfSe. At the genus level, a total of 24 distinct genera, including Akkermansia, Bacteroides, Lactobacillus, and Dubosiella, were identified as enriched in the two groups. IBD mice exhibited an enrichment of Helicobacter, Akkermansia, and 11 other genera, while control mice were enriched with a total of 13 genera, including Roseburia and Dubosiella ( Figure 5A ). We conducted a comparative analysis of 24 discriminant genera across treatment cohorts. Heatmap analysis demonstrated that 5-ASA treatment significantly reduced the abundance of [Eubacterium]_fissicatena_group, Ileibacterium, UCG-008, and Akkermansia compared to IBD controls. While L-donor FMT attenuated potentially detrimental taxa, only Lactobacillus, Muribaculum, and Anaerotruncus approached control levels. In contrast, H-donor FMT not only reduced detrimental taxa but also enhanced beneficial genera, including Lactobacillus, [Eubacterium]_xylanophilum_group, Anaerotruncus, Romboutsi, and Dubosiella, to near-physiological ranges ( Figure 5B ). Individual recipient analysis through differential taxa average relative abundance clustering heatmaps demonstrated significantly higher therapeutic response rates compared to both L- donor FMT and 5-ASA interventions, along with enhanced microbial restoration in H-donor FMT cohorts ( Figure 5C ). These findings establish the superior capacity of H-donor FMT to ameliorate IBD-associated dysbiosis.