Introduction

Inflammatory bowel disease (IBD), encompassing ulcerative colitis (UC) and Crohn’s disease (CD), is a chronic gastrointestinal inflammatory disorder associated with significant morbidity(1, 2). The intestinal mucosa is continuously exposed to a diverse microbial ecosystem, including fungi, bacteria, and viruses(3). In IBD, the delicate balance between the immune system and the microbiota is disrupted, leading to uncontrolled inflammation(4, 5). This dysregulation is further exacerbated by genetic factors, such as loss-of-function mutations in NOD2, which impair antimicrobial peptide production, increase susceptibility to intestinal infections, and decrease IL-10 production by monocytes(6–8).

Trained immunity, a phenomenon where innate immune cells exhibit enhanced responses to secondary challenges, has emerged as a promising therapeutic avenue for various infectious diseases and cancer(9–13). However, its role in chronic inflammatory conditions like IBD remains largely unexplored. While trained immunity can enhance pathogen clearance, it also involves the upregulation of pro-inflammatory cytokines such as TNF, IL-1β, and IL-6(14), therefore the role of trained immunity in IBD remained elusive.

This study aimed to investigate the therapeutic potential of trained immunity in a mouse model of colitis. We hypothesized that inducing trained immunity could enhance microbial control, thereby preserving intestinal integrity and ultimately mitigating disease severity. To test this, we employed β-glucan (BG), a prototypic trained immunity inducer, and assessed its impact on dextran sulfate sodium (DSS)-induced colitis. Our findings demonstrate that BG pretreatment significantly alleviates colitis in mice. Mechanistically, BG induced central immunity in the hematopoietic compartment, leading to the generation of “trained” monocytes in the periphery. Both bone marrow transplantation from trained donors and adoptive transfer of trained monocytes conferred protection against DSS-induced colitis. Furthermore, single-cell RNA-seq revealed that BG-induced trained immunity promotes the expansion of reparative Cx3cr1 intestinal macrophages derived from Ly6C^hi monocytes, thereby facilitating epithelial regeneration. In summary, our data suggest that harnessing trained immunity represents a promising therapeutic approach for IBD treatment.

Results

β-glucan (BG) pretreatment ameliorates DSS-induced colitis

To investigate the impact of trained immunity on colitis, mice were pretreated with BG, a potent inducer of trained immunity. Successful induction of BG-induced trained immunity (BGTI) was confirmed by the enhanced resistance to Staphylococcus aureus infection in BG-treated mice (Fig S1A and S1B), consistent with previous findings(15). Importantly, BG-pretreatment did not induce spontaneous intestinal inflammation (Fig S1C and S1D). Following BGTI induction, mice were subjected to DSS-induced colitis. (Fig 1A). BG pretreatment significantly ameliorated disease severity, as indicated by the reduced body weight loss (Fig 1B and Fig S1E), colon shortening (Fig 1C and Fig S1F and S1G), and diminished mucosal inflammation observed via endoscopic imaging (Fig 1D). Histological analysis demonstrated lower histopathologic scores (Fig 1E and 1F). Additionally, BG-pretreated mice exhibited enhanced intestinal epithelial barrier integrity, as evidenced by the increased expression of tight junction proteins Zo-1 and Occludin (Fig 1G and 1H), and decreased circulating FITC-dextran levels (Fig 1I).

Fig 1.

Acknowledging that the long-lasting effects of BGTI(16), we further examined its long-term protective effects against DSS-induced colitis. Despite peripheral myeloid cells and monocytes returning to homeostasis four weeks post-BG treatment (Fig S2A and S2B), mice remained protected against DSS-induced colitis as demonstrated by body weight change and colon shortening (Fig 1J and 1K), and improved histological outcomes (Fig S2C and S2D). This BG-induced protection persisted for up to seven weeks (Fig S2E and S2F). These findings indicate that BG pretreatment effectively ameliorates DSS-induced colitis in mice, emphasizing the therapeutic potential of trained immunity in IBD management.

BG ameliorates colitis via enhanced myeloid cell activation independent of adaptive immunity

Given that BGTI has been reported to function independently of adaptive immunity(17, 18), we evaluated the effects of BG pretreatment in Rag1 knockout mice (Fig S3A-S3D). Notably, BG pre-treatment remained effective in Rag1^-/- mice, as evidenced by reduced body weight loss, colon shortening, and improved histological outcomes (Fig S3E-S3H). These findings confirm that BGTI-mediated protection is independent of adaptive immune responses, consistent with previous findings demonstrating that adaptive immunity is dispensable for trained immunity.

To gain a comprehensive understanding of the molecular mechanisms underlying BG-mediated protection, we performed RNA sequencing on colonic samples collected at multiple time points following DSS administration. WGCNA identified the MEturquoise module emerged as the core functional module, containing 5,015 genes whose expression exhibited significant co-regulation following BG intervention (Fig S4A). And differential gene expression of genes within this module was most significant on day 7 post-DSS treatment when comparing the BG-treated mice to PBS controls (Fig S4B). GO analysis revealed significant enrichment in pathways related to chemotaxis and leukocyte migration (Fig 2A). KEGG pathway enrichment analysis demonstrated significant activation of cellular processes involved in phagocytic function (Fig 2B) Additionally, the enrichment observed in WGCNA was further corroborated by whole-genome analysis of day 7 colitis samples, which also identified significant enrichment of pathways associated with chemotaxis, leukocyte migration and phagocytic function (Fig S4C and S4D). Multiple analytical approaches concurrently indicated that BG pretreatment not only enhances the recruitment of immune cells but also promotes their functional activation, particularly with respect to phagocytic capabilities.

Fig 2.

To further investigate how BG pretreatment reprograms of intestinal myeloid cells, we performed single-cell RNA-seq to characterize intestinal leukocytes populations on day 7 post-DSS treatment. Unbiased clustering analysis identified multiple clusters of intestinal immune cells subsets, including monocytes/macrophages (Ly6c2, Ccr2 and Adgre1), dendritic (DC) (Cst3 and H2-Aa), neutrophils (S100a8/a9 and Ly6g), B (Cd79a Cd79b and Cd19), Pre B (Myl4 and Mme), CD4 T (Cd3d and Cd4), CD8 T (Cd3d and Cd8a), natural killer (Nkg7), ILC2 (Gata3 and Il4) and ILC3 (Rorc and Il22) cells (Fig 2C and Fig S4E). Importantly, the ratio of monocytes/macrophages and neutrophils was elevated in the BG group (Fig 2D).

Single-cell RNA-sequencing data further demonstrated enhanced expression of innate pathogen recognition receptors (Syk, Nod1, Nod2, Tlr2 and Tlr4), efferocytosis (Cd300lf, Axl, Mertk, Itgal and Itgb7), cytokines (Il1b, Il1rn, Vegfa and Tgfb1), chemokines (Ccl2, Ccl6, Ccl9, Cxcl9) and chemokine receptors (Ccr1, Ccr2, Cxcr4) in intestinal monocytes/macrophages (Fig 2E). AUC and KEGG analysis demonstrated significant enrichment pathways related to activation of innate immune response and phagocytosis in monocytes/macrophages from BG-treated mice (Fig 2F and 2G). These findings, corroborated by WGCNA analysis, strongly suggest that BG pretreatment enhanced the recruitment and activation of myeloid cells, particularly monocytes/macrophages, in the inflamed colon, thereby improving disease outcomes.

BG trained bone marrow monocytes protected against colitis via the Ccl2-Ccr2 axis

Drawing upon single-cell transcriptome analysis that reveals the proportional expansion and functional reprogramming features of colonic monocyte/macrophage populations, in conjunction with the observation of significantly upregulated peripheral monocytes following BG pretreatment as depicted in Fig S2B. A hypothesis was formulated that BG exerts protective effects against colitis via monocytes. To experimentally validate this hypothesis, we evaluated the role of monocytes in Ccr2 KO mice. Circulating Ly6C^hi monocytes were diminished in Ccr2^-/- mice (Fig S5A and S5B), corroborating the pivotal role of the Ccl2-Ccr2 axis in monocyte egression from the bone marrow into circulation(19, 20). Notably, BG pretreatment failed to protect Ccr2^-/- mice from DSS-induced colitis, as evidenced by the lack of significant differences in body weight loss, colon length, and epithelial permeability (Fig 3A-E).

Fig 3.

In view of the fact that BG induces myelopoiesis through the modulation of hematopoietic stem and progenitors cell compartments in the bone marrow via central trained immunity(21, 22). we assessed whether transplantation of bone marrow cells from BG-pretreated donors could confer resistance to DSS-induced colitis in naïve recipients (Fig 3F). At 6 weeks post-transplantation, circulating myeloid cells were predominately derived from donor mice as indicated by the CD45.1 marker on circulating mononuclear cells (Fig S5C). Following DSS treatment, mice receiving bone marrow cells from BG-pretreated donors exhibited significantly reduced body weight loss (Fig 3G). However, no significant difference in colon length was observed (Fig 3H and Fig S5D). Meanwhile, the percentage of circulating CD11b⁺ myeloid cells, neutrophils and Ly6C^hi monocytes were increased (Fig 3I and J, Fig S5E).

To further assess the role of monocyte, we adoptively transferred bone marrow Ly6C^hi monocytes sorted from control or BG-trained donor mice into Ccr2^-/- recipient mice, which were then subjected to DSS treatment (Fig S5F and S5G). Mice received BG-trained monocytes exhibited reduced weight loss and colon shortening (Fig 3K and 3L). These findings suggest that BG-trained Ly6C^hi Ccr2⁺ monocytes play a pivotal role in the alleviation of colitis.

BG-trained monocytes enhance innate immune activation and microbial control

To explore the heterogeneity and functional roles of monocytes and macrophages, we delved deeper into the scRNA-seq data. By examining the expression of key signature genes, including Ly6c2, Ccr2, H2-Ab1, Runx3, Itgax, Adgre1, Cx3cr1 and Cd209a, we identified eight distinct immune cell subsets: three monocyte subsets (Mono1-3) characterized by high Ly6c2 and Ccr2 expression, four macrophage subsets (Macro1-4) defined by expression of H2-Ab1, Runx3, Itgax, Adgre1, and Cx3cr1, and a dendritic cell subset (Cd209a⁺DC) expressing integrin Itgax and C-type lectin CD209a (Fig 4A and Fig S6A). Among macrophage subsets, Macro1 highly expressed Vegfa, suggesting that may be a mucosal repair-promoting macrophage population. Macro2 was enriched with tissue-resident macrophage signature genes (Runx3, Dtx4, Cx3cr1, Hes1). In contrast, Macro3 and Macro4 deviated from the main cluster in cell clustering analysis, showing significant heterogeneity. Macro3 was characterized by high Ighm expression, while Macro4 showed elevated Cd4 expression. Both Macro2 and Macro4 exhibited higher Il10 expression than other monocyte/macrophage subsets, indicating their inflammation-regulatory functions.

Fig 4.

The BG-treated group exhibited a significant increase in the proportions of Mono1, Mono2, Macro1, and CD209a⁺DCs, whereas higher proportions of Mono3 and Macro3 were observed in the PBS group (Fig 4B). Given the strong enrichment of Mono1 and the reduction of Mono3 following BG treatment, we sought to further define the transcriptomic characteristics that differentiate Mono1 from Mono3. GO enrichment analysis revealed that Mono1 was significantly enriched in pathways regulating the activation of innate immune response (Fig S6B). KEGG pathways analysis further identified enrichment in NOD-like receptor (NLR) signaling, Toll-like receptor (TLR) signaling, and phagocytosis (Fig 4C). AUC analysis further confirmed the TLR and NLR signaling pathways were significantly upregulated in Mono1 and Mono2 subsets of the BG-pretreated group (Fig S6C and S6D). Additionally, antimicrobial humoral response pathway was also upregulated in these two subclusters (Fig S6E). Cooperative activation of the pattern recognition receptor (PRR) signaling axis and antimicrobial humoral response significantly promotes the synthesis of antimicrobial peptides(23, 24). BG pretreatment significantly upregulated genes encoding pattern recognition receptors (Nod1, Nod2, Tlr2, Tlr4) and antimicrobial effector molecules (S100a8, S100a9, Lcn2 and Defb1) in the colon on day 7 of colitis (Fig 4D). Furthermore, antimicrobial genes S100a8 and Defa1 were highly expressed in the BG group as validated by qPCR analysis (Fig S6F). These genes upregulation suggest a potential link between trained immunity and enhanced microbial control. Consistently, gene signatures associated with bacterial defense responses were prominently enriched in the BG-treated group (Fig 4E).

Previous studies have demonstrated that GBPs as intracellular effectors induced by IFNγ and LPS to promote antibacterial defense(25, 26). We found that BG training significantly enriched the IFN-γ response pathway in Mono1 and Mono2 subsets (Fig S6G). Meanwhile, Guanylate-binding proteins (GBP)-related genes, including Gbp2, Gbp3, Gbp5, and Gbp7, were also specifically upregulated in Mono1 and Mono2 (Fig 4F). GBP-deficient mice exhibit significantly increased susceptibility to Salmonella typhimurium infection(27–29). We then performed intestinal Salmonella typhimurium infection to assess whether BG pretreatment enhances microbial control in the gut (Fig 4G). As expected, BG pretreatment significantly protected mice from lethal Salmonella Typhimurium infection, suggesting that BG-induced trained immunity enhances mucosal defenses against gut microbial infections (Fig 4H).

Taken together, these results reinforced our bulk RNA-seq findings, indicating that BG pretreatment induces a specific monocyte subcluster with enhanced innate activation and phagocytic capacity, facilitating more efficient control of the breaching microbiome associated with DSS-induced leaky gut.

BG-mediated reprogramming of myeloid differentiation trajectories balances inflammation and enhances mucosal repair in colitis

While BG-induced trained immunity is characterized by enhanced pro-inflammatory cytokine production and increased bactericidal capacity(14), excessive pro-inflammatory cytokines release may exacerbate colitis progression. To balance heightened cytokine production without inducing overwhelming immunopathology, we hypothesized that BG pretreatment might involve feedback loops that modulate the inflammation while preserving enhanced phagocytic capacity.

To investigate this, we analyzed previously obtained bulk colonic RNA-seq data, focusing on genes associated with anti-inflammatory and regulatory functions. On day 7 of colitis, BG pretreatment significantly upregulated genes encoding immunomodulatory factors (Il4, Il10, Il11 and Il33), signaling inhibitors (Socs1, Socs3), and differentiation regulators (Csf1, Csf1r, Stat3 and Stat6) (Fig S7A) The upregulation of Socs1 and Socs3 was further corroborated by scRNA-seq data, which showed increased expression of these genes in the Mono1 and Mono2 subclusters (Fig S7B). As members of the suppressor of cytokine signaling (SOCS) family, SOCS1 and SOCS3 inhibit excessive JAK-STAT signaling activation via negative feedback regulation(30–32). Moreover, coordinated activation of the monocyte/macrophage differentiation regulatory network (Csf1r, Nr4a2, Irf8, Klf4) suggests that BG may induce the differentiation of regulatory-phenotype macrophage subsets by reprogramming myeloid developmental trajectories, thereby modulating inflammatory responses.

To further investigate whether BG pretreatment reprograms monocyte-to-macrophage differentiation, we analyzed the day7 RNA-seq data revealed significant enrichment of myeloid leukocyte differentiation pathway (Fig S7C). Additionally, Single-cell trajectory analysis further revealed that Macro1 and Macro2 differentiated from Mono2 and Mono3, respectively (Fig 5A). BG training significantly accelerated the differentiation kinetics of Mono3 into Macro1, which may underlie the reduced frequency of the Mono3 subset in the BG group. Coupled with the gene expression signatures in Fig S6A, Macro1 shares similarities with Macro2 in marker gene expression, and the differentiated Macro1 subset could further differentiate into the tissue-resident macrophage Macro2 subset (Fig 5B).

Fig 5.

Furthermore, we observed that the expression levels of Ccr2 and Ly6c2 decreased from Mono1 to Macro2, while H2-Ab1 and Cx3cr1 progressively increased, particularly in macrophages (Fig 5C). Since Cx3cr1 upregulation is a hallmark of circulating monocytes migrating to the intestinal lamina propria and undergoing macrophage differentiation(33), we used CX3CR1-GFP reporter mice to monitor the monocyte-to-macrophage transition in the DSS colitis model. In circulating leukocyte subsets, the total percentage of CD11b⁺ myeloid cells was higher in BG-pretreated mice compared to controls (Fig S8A-C). A similar trend was observed in neutrophils (Fig S8D and S8E). While the initial monocyte percentage was higher in the BG group, no significant difference was observed the two groups on day 7 post-DSS treatment (Fig 5D and S8F), suggesting that circulating monocytes in BG-pretreated mice had extravasated into the intestinal tissue. Meanwhile, the percentage of colonic CD11b⁺ myeloid cells and neutrophils was increased in the BG-pretreated group (Fig S9A-C).

We further defined monocyte/macrophage populations from P1 to P6 based on the expression of CD11b, CX3CR1-GFP, Ly6C, and MHCII. P1 to P3 were gated from CD11b⁺&CX3CR1-GFP^-&Ly6G^- cells. While P1 (Ly6C⁺&MHCII^-) and P3 (Ly6C^-&MHCII⁺) were comparable between group, P2 (Ly6C⁺&MHCII⁺) was increased in BG-pretreated mice (Fig 5E). P4 to P6 are gating from CD11b⁺&CX3CR1-GFP⁺ cells and further separate by the expression of Ly6C and MHCII. P4 (Ly6C⁺&MHCII^-) as infiltrating monocytes gaining CX3CR1 expression. P5 (Ly6C⁺&MHCII⁺) is regarded as intermediate transitory immature macrophage and P6 (Ly6C^-&MHCII⁺) is considered as mature macrophage as it completely loss expression of Ly6C and highly express MHCII. BG pretreatment led to an increase in P5 level (Fig 5F). The increase of P2 and P5 in BG group in line with the decrease of monocytes at day 7 in circulation post DSS treatment, suggesting that BG trained monocytes have higher capacity to infiltrate into the damaged colon and undergo macrophage differentiation.

Previous WGCNA analysis suggested that BG pre-treatment upregulated the focal adhesion pathway, which may enhance tissue repair mechanisms (Fig 2B). Consistent with this, we observed increased expression of genes involved in wound healing and tissue repair, such as Mmp-related genes, Spon1, Notch1, Notch2, Tgfb, Fgf1, Fgf2, Col1a1, Col1a2, and Fn1 in the BG-treated group on day 7, followed by a rapid decline by day 12. In contrast, the PBS group exhibited a much milder induction of these genes, with several genes remaining at comparable levels between day 7 and day 12 (Fig 5G). This led us to hypothesize that the upregulated monocytes and macrophages in the BG-treated group might exhibit enhanced expression of tissue repair-associated genes. As expected, the expression of Spon1 and Mmp14, which promote the mucosal repair was increased in Mono2, Macro1, and Macro2 (Fig 5H and 5I). These findings suggest that BG pretreatment indeed induced monocyte and macrophage subsets with enhanced tissue repair capacity.

Discussion

Trained immunity, a non-antigen-specific form of innate immunological memory, enables rapid host defense mobilization upon pathogen challenge(34, 35). In inflammatory bowel disease, gut microbial dysbiosis manifests as reduced α-diversity, diminished commensal taxa (e.g., Faecalibacterium prausnitzii), and pathobiont expansion (e.g., Enterobacteriaceae, adherent-invasive Escherichia coli), which may compromise intestinal barrier integrity and precipitate microbial translocation(36–38). Additionally, the long-term use of immunosuppressants and biologics in IBD patients elevates infection susceptibility, rendering the microbial-controlling properties of trained immunity a viable therapeutic strategy for IBD(39). However, Excessive or sustained immune activation could trigger chronic inflammatory cascades and provoke autoimmune tissue damage through aberrant immune recognition, such as rheumatoid arthritis and cardiovascular diseases(9, 40–42). However, the functional correlation between trained immunity and colitis remains unclear, and how to apply trained immunity to colitis treatment represents an urgently unresolved scientific question.

Here, we demonstrated that BG-induced trained immunity confers cross-pathogen protective effects using a lethal Staphylococcus aureus infection model, while failing to induce intestinal inflammation under conditions of an intact mucosal barrier. Further investigations revealed that BG effectively attenuated dextran sulfate sodium (DSS)-induced colitis upon secondary challenge. BG pretreatment conferred long-term protection against DSS-induced colitis, with benefits lasting for at least two months after the peripheral myeloid cell populations returned to baseline levels. Bone marrow transplantation experiments confirmed that BG-induced trained immunity persists within hematopoietic progenitors, leading to durable reprogramming of peripheral monocytes. This long-term effect is likely supported by epigenetic modifications in hematopoietic cells, as reported in previous studies(21, 43). These findings highlight the potential of BG-induced trained immunity in establishing long-lasting immune memory, which could be leveraged to achieve sustained therapeutic effects in IBD.

The conventional understanding of IBD pathogenesis has long posited that abnormal immune activation is central to disease development and progression. This paradigm has predominantly focused on the role of adaptive immunity, particularly the dysregulation of Th17 and Treg functions over the past decades(44–46). However, our results suggest that BG-induced protection is independent of adaptive immune responses. Even in Rag1^-/- mice, which lack adaptive immunity, BG pretreatment significantly improved DSS-induced colitis outcomes. This underscores the dominant role of innate immunity.

In DSS-induced colitis, epithelial barrier disruption facilitates microbiota translocation across the mucosal epithelium, triggering inflammatory responses mediated by innate cells(38, 47, 48). We identified monocytes as key mediators of BG-induced protection against DSS-induced colitis, as demonstrated by experiments using Ccr2^-/- mouse models and monocyte transfer studies. BG pretreatment enriched Ly6C^hiCCR2⁺ monocytes, which exhibited enhanced antimicrobial capabilities, as revealed by scRNA-seq. Ly6C^hiCCR2⁺ monocytes displayed higher levels of antimicrobial receptors like NOD2, TLR4 and GBP-related gene. Guanylate-binding protein (GBP) enhances Salmonella clearance by promoting inflammasome activation and intracellular bacterial elimination, particularly through disrupting Salmonella-containing vacuoles and inducting pyroptosis(27, 49). Our finding revealed that BG pretreatment increased resistance to Salmonella typhimurium infection. As inflammation and infection is partially controlled, monocytes can differentiate into macrophage subsets with anti-inflammatory and tissue repair functions, secreting cytokines such as IL-10, TGF-β and VEGF to promote inflammation resolution and tissue repair(50, 51).

Previous studies have underscored the essential role of circulating Ly6C^hi monocytes in replenishing intestinal macrophages(52, 53). During colitis and infection, these monocytes are rapidly recruited to injured colonic tissues, where they differentiate into Cx3cr1⁺ macrophages. We also observed that these monocytes exhibited enhanced migratory capabilities, enabling them to infiltrate inflamed tissues and differentiate into CX3CR1⁺ macrophages. These macrophages contribute to mucosal repair by clearing pathogens such as Salmonella and promoting epithelial regeneration(54). Single-cell trajectory analysis further showed that BG training promoted the differentiation of Macro1 into Macro2, which is consistent with the dynamic process of monocyte-to-mature macrophage conversion observed in CX3CR1-GFP reporter mice. Maintaining a balance between inflammatory monocytes, differentiated macrophages, and tissue-resident macrophages is critical for resolving inflammation, promoting tissue repair(33, 55). These findings suggest that BG not only augments monocyte antimicrobial functions but also facilitates their differentiation into macrophages, enhancing their ability to recognize, phagocytose, and eliminate bacteria while promoting tissue repair. This dual function likely strengthens the intestinal barrier, mitigates inflammation, and supports gut homeostasis.

In the clinical exploration of IBD treatment, novel therapies based on cellular regeneration (intestinal organoid transplantation) and fecal microbiota transplantation (FMT) are emerging as critical avenues to overcome the limitations of conventional pharmacotherapies(56–58). Given the efficacy of BG-induced trained immunity in alleviating DSS-induced colitis, this approach holds significant promise as a novel therapeutic strategy for IBD treatment. We transferred BG-preconditioned monocytes significantly alleviated clinical symptoms, including weight loss and colon shortening, demonstrating the potential of cell-based therapies (Fig S10A-C).

In conclusion, BG confers long-term colitis amelioration through trained immunity by reprogramming bone marrow-derived monocytes. Mechanistically, BG enhances monocyte chemotaxis and antimicrobial functionality activation, enabling efficient microbial containment. Crucially, this trained state establishes self-regulating circuits, engaging in negative feedback regulation of excessive immune responses, while accelerated CX3CR1⁺ macrophage differentiation facilitates mucosal repair. These findings establish a foundation for leveraging trained immunity as a therapeutic strategy for IBD, highlighting the potential of BG in modulating intestinal immunity and restoring gut homeostasis.

Materials and methods

Animals

C57BL/6J wild-type male mice (6-8weeks) were purchased from Xiamen University Laboratory Animal Center. Rag1^-/- and CX3CR1-GFP mice were generously provided by Dr. Kairui Mao and Dr.Xiaofen Chen from Xiamen University, respectively. Ccr2^-/- mice were kindly provided by Dr. Shih-Chin Cheng from Xiamen University. Mice were maintained under specific pathogen-free conditions. All animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Xiamen University.

In vivo beta-glucan (BG) training and DSS colitis model

Mice were i.p. injected with one dose of BG (1 mg) on day -7/-49, Then, Mice received 3% indicated DSS (36,000 to 50,000 molecular weight; MP Biomedicals) in their drinking water for 5 days, followed by 2 days or 7 days of distilled water without DSS. Control animals received distilled water for the entire period. Mice were monitored every day from day 0 for body weight until day 7 or 12.

Histology

Colons were collected from sacrificed mice at the end of each treatment schedule, as described above, colon was fixed with 4% paraformaldehyde and embedded in Paraffin. Tissue sections (5 µm) were prepared, deparaffinized, and stained with hematoxylin and eosin. with minor modifications, by assessing mucosa thickening, inflammatory cells, and submucosa cell infiltration. Each criterion was scored as 0-4, and the sum of each score was defined as the histological score. Histological scores were assigned by experimenters “blinded” to sample identity.

Coloscopy

A coloscopy system (ENDOCAM® Logic HD, RICHARD WOLF) was used to monitor the severity of DSS-induced colitis. After an 8-hour fast, we administered isoflurane (RWD, Co. Ltd.) and applied glycerin to the anus for smooth insertion.

In vivo intestine permeability assay

Age and sex-matched mice were orally administered with 0.6 mg/g body weight of an 100 mg/ml solution of FITC-dextran (FD4, Sigma). 5 hours later, retro-orbital blood was collected from each mouse. Serum was prepared by allowing the blood to clot by leaving it undisturbed overnight at 4°C and then subsequently centrifuged at 5000 rpm for 10 minutes. Dilutions of FITC-dextran in PBS and separately in pooled mouse serum were used as a standard curve. Absorbance of 50 μl serum was measured at microplate reader with excitation and emission filters set at 490 and 530 nm, respectively. Experiments were performed at least two independent times each in triplicate.

RNA extraction and real-time PCR

Total RNA was extracted from colon tissue previously washed from luminal content. Samples were homogenized in TRIZOL. Homogenized tissues were then added with chloroform in order to separate RNA from genomic DNA and proteins. RNA isolation and purification required isopropanol (RNA precipitation) and 70% ethanol (RNA wash). Total RNA was resuspended in Nuclease-free water. Subsequently, 1 µg of RNA was reverse-transcribed using the Hifair® ll lst Strand cDNA Synthesis SuperMix for qPCR (gDNA digester plus) following manufacturer’s instructions. cDNA analysed through quantitative Real-Time PCR.

Isolation of cells from lamina propria

To isolate lamina propria mononuclear cells (lamina propria MCs), extraintestinal fat tissue was carefully removed and colons were then flushed of their luminal content with physiologic solution, opened longitudinally and cut into 1 cm pieces. Epithelial cells and mucus were removed by incubation with 10 mL of D-Hank’s balanced salt solution (containing EDTA and DTT, free of Ca2⁺ and Mg2⁺) for 20 minutes at 37°C at 200 rpm in a constant temperature shaker. Colon pieces were then digested in Hank’s (with Ca2⁺ and Mg2⁺) containing 0.5 mg/ml Collagenase IV and 15 µg/ml DNase I, for 40 min at 37°C shaking at 200 rpm. The remaining cells were centrifuged and resuspended in FACS buffer (1% FBS, 2 mM EDTA in PBS).

Bone Marrow Transplantation

Wild type CD45.2 mice were subjected to 8 Gy irradiation (RAD SOURCE, RS2000). CD45.1 wild type and BG trained mice were sacrificed, and the femurs were used to harvest bone marrow cells. 1×10⁷ bone marrow cells from CD45.1 wild type and BG trained donor mice were transferred to irradiated CD45.2 recipients via tail-vein injection. 2.5% DSS was given at 6 weeks-post bone marrow transplantation. Mice were monitored every day until day 12.

Adoptive transfer of bone marrow monocytes

BG trained mice were sacrificed, and the femurs were used to harvest bone marrow cells, Bone marrow cells were stained with the following panel: Biotin antibody (B220, CD4, CD8, NK1.1, Ly6G), Live/dead (Cat. L34976, Invitrogen), CD45 (Cat. 45-0451-82, Clone. 30-F11, ebioscience), CD11b (Cat. 48-0112-82, Clone. M1/70, ebioscience), Ly6C (Cat. 128016, Clone. HK1.4, Biolegend). Biotin-labeled lineage antibodies were used to stain bone marrow cells at 4℃ for 20 min and then incubated with fluorescence conjugated streptomycin together with other surface protein antibodies for another 20 min. Ly6C^hi monocytes were sorted to > 97% purity using a BD Aria III cell sorter.

RNA-seq

RNA was extracted from colon LPLs of individual mice. RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kit v3-cBot-HS (Illumia) according to the manufacturer’s instructions. The library was sequenced on an Illumina Novaseq platform and 150 bp paired-end reads were generated.

scRNA-seq

Isolated lamina propria mononuclear cells, sample from three mice were pooled together and analyzed per condition. For scRNA-seq of Colon CD45⁺ cells were sorted by BD Arial III. These cells were barcoded with 10×Cellplex oligos before being encapsulated using the 10X Chromium 3′ Reagent Kits v3 according to manufacturer’s instructions.

RNA-seq data processing and analysis

Raw data of fastq format were quality checked through fastqc software. Reference genome and gene model annotation files were downloaded from genome (mm10) website directly. Index of the reference genome was built using Hisat2 v2.1.0 and paired-end clean reads were aligned to the reference genome using Hisat2 v2.1.0. Expression levels across the sample were expressed in FPKM (Fragment per kilobase per million). Weighted correlation network analysis was perform using an R package from horvath.genetics.ucla.edu/html/CoexpressionNetwork/Rpackages/WGCNA/. Differential gene expression was used DESeq2 by a twofold cutoff and pvalue < 0.05. All differentially expressed genes were plotted with the “pheatmap” and “ggplot2” R packages. For KEGG enrichment analysis, p-value < 0.05 was used as a threshold to determine significant enrichment for gene sets with the “clusterProfiler” R package.

scRNA-seq data processing and analysis

Raw sequencing data was processed and aligned mm10 mouse reference genome with CellRanger (10×Genomics) v 7.2.0. Resulting filtrated matrices (count matrices) of molecular counts were used as input for further processing with Seurat package V5.0.1running under R studio. First, quality control was performed to create Seurat object with min features > 200 and removal of cells having < 200 or > 8000 expressed genes or > 5% mitochondrial counts. The total number of recovered cells was mentioned before. Variable features using FindVariableFeatures (using RNA and vst as an assay and selection method as parameters) and normalization using normalization.method = “LogNormalize”, scale.factor = 10000 were performed . This integrated data is scaled using ScaleData function using all genes and “RNA” as assay method.

Doublet cells were filtered by DoubletFinder v3. Principal Component Analysis was performed on variable features using RunPCA and first 30 PCs were chosen. The nearest neighbors using the FindNeighbors(dims=1:12) and FindClusters(resolution=0.8) function. For visualization, the non-linear dimensional reduction such as UMAP analysis was using RunUMAP function from Seurat. To calculate pseudotime based on the Mono/Marco subset data using the monocle3.

Statistical Analysis

Statistical analyses were conducted using GraphPad Prism version 9.0 and R statistical software (version 4.2.2, R Foundation for Statistical Computing). Data were analyzed using a two-tailed Student’s t-test. Multiple-group comparisons were performed using one-way or two-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparisons test. Survival data were assessed using Kaplan-Meier survival plots, followed by the log-rank test. Significant differences were indicated by an asterisk (p < 0.05).

Data availability statement

All data relevant to the study are included in the manuscript or uploaded as online supplemental information. Sequencing data have been deposited to the GEO with accession number GSE285859 (single cell RNA-seq), GSE285860 (RNA-seq).

Supplementary figures

Fig S1.

Fig S2.

Fig S3.

Fig S4.

Fig S5.

Fig S6.

Fig S7.

Fig S8

Fig S9

Fig S10

Acknowledgements

We thank the staff of Xiamen University Laboratory Animal Center. Special thank to Prof. Kairui Mao and Prof. Xiao-fen Chen for generously providing mice. We appreciate the help of Dr. Jia Zhang and other staff from Prof. Shih-Chin Cheng’s laboratory during the study.

Additional information

Author contributions

Y.Y.L. designed and did the experiments, analyzed the data and wrote the paper; Y.Y.F., Q.X.G. and Q.Y. C. analyzed the data, wrote the paper, and supervised the study. Y.Q.H., L.W., H.X.S., E.M.C., Q.Y.X., Y.C., Q.Q.F., L.L.L provided advice and oversaw a portion of the work. J.L.R., S.-C.C., H.Z.X. conceived the idea, secured funding, supervised the study, helped in data interpretations, revised drafts of manuscript; supervised overall project. All authors reviewed and approved the final version of the manuscript.

Ethics approval

All animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Xiamen University.

Patient consent for publication: Not applicable.

Funding

This project was supported by National Key R&D Program of China (2022YFA1304000), National Natural Science Foundation of China (82300630&32161133020), Healthcare System Youth Backbone Talent Training Project of Fujian Province(2023GGB09), Municipal Natural Science Foundation of Xiamen (3502Z20227271), Foundation of State Key Laboratory of Vaccines for Infectious Diseases, Xiang An Biomedicine Laboratory (2023XAKJ0101012), Medical and Health Key Project of Xiamen (3502Z20204007).