Bifidobacterium longum RAPO Attenuates Dermal and Pulmonary Fibrosis in a Mouse Model of Systemic Sclerosis through Macrophage Modulation and Growth of Short-Chain Fatty Acid Producers Authors: Hee Jin Park, Dakyum Yu, Seong-Tshool Hong, Juyeon Lee, Sang-Jun Park, Myeong Soo Park, Hanna Lee, Mingyo Kim, Yun-Hong Cheon, Seung-Geun Lee, Dong Hyun Sohn, Jae-Bum Jun, Suhee Kim, Sang-Il Lee Published: 2024 ABSTRACT AbstractSystemic sclerosis (SSc) is a complex autoimmune disease with an unclear etiology and no effective treatments. Recent research has suggested involvement of the microbiome in SSc pathogenesis. This study aimed to identify specific microbial species associated with SSc and explore their therapeutic potential. Serum Abs against 384 intestinal microbial species revealed a significant depletion in Abs against Bifidobacterium longum in patients with SSc compared to healthy controls. In a bleomycin-induced SSc mouse model, oral administration of B. longum strain RAPO attenuated skin and lung fibrosis, accompanied by reduced infiltration of inflammatory monocytes/macrophages and downregulation of pro-inflammatory cytokines and chemoattractant Ccl2 genes in lymph nodes and fibrotic tissues. B. longum RAPO treatment restored fecal microbial diversity and augmented short-chain fatty acid (SCFA)-producing bacteria in the gut, leading to increased fecal butyrate levels and upregulated SCFA receptor Gpr41 in the mesenteric lymph node. In vitro, B. longum RAPO and its culture supernatant suppressed the expressions of pro-inflammatory cytokine genes in macrophages and inhibited myofibroblast differentiation in fibroblasts. These findings highlight the probiotic potential of B. longum RAPO in preventing tissue fibrosis by modulating macrophage activity and promoting the growth of SCFA-producing bacteria, underscoring the therapeutic potential of microbial modulation in SSc.Keywords: Systemic sclerosis, Bacterial antibodies, Bifidobacterium longum, Fibrosis, Short-chain fatty acid INTRODUCTION Systemic sclerosis (SSc) is a complex autoimmune disease affecting multiple organs, characterized by vascular injury, immune dysregulation, and fibrosis (1,2). Despite extensive research on the genetic, environmental, and immunological aspects, the exact etiology of SSc remains unclear, and effective treatments are still lacking (3,4). Therefore, elucidating the underlying mechanisms would aid in further developing targeted and efficacious therapies in SSc. Gut microbiota potentially contributes to the development of SSc (5). Individuals with SSc often exhibit significant gastrointestinal tract involvement, with marked alterations in gut microbial composition, including an overrepresentation of pathogenic bacteria and a reduction in commensal bacteria (6). Moreover, changes in the gut microbiota, such as the reduced abundance of butyrate-producing bacteria and proliferation of Th17-activating bacteria, suggest the potential of microbiome-driven inflammation in SSc pathogenesis (7). Animal studies have further supported the link between the gut microbiome and SSc pathologies, demonstrating that gut microbial dysbiosis in early-life aggravates fibrosis in SSc mice, while administering bacterial metabolites, such as sodium butyrate, ameliorates SSc-associated skin and lung fibrosis in mice (8,9). Although the precise mechanisms underpinning the microbiota-SSc connection are not fully elucidated, it is hypothesized that alterations in the gut microbiota may trigger immune dysregulation, contributing to the onset of SSc (10). Furthermore, probiotics containing Bifidobacterium and Lactobacillus species have shown promising immunomodulatory effects in patients with SSc, suggesting their potential therapeutic value (11). These findings offer a rationale for exploring microbiome-targeted interventions in SSc (12). Recent research has highlighted the importance of identifying disease-specific microbes at the species or even strain levels for successful microbiome-targeted therapies (5), considering the divergent effects observed among bacteria within the same genus (13,14). However, traditional genomic approaches for gut microbiome profiling may have limited taxonomic resolution, particularly for identifying species or lower taxa (15). In contrast, immunology-based microarrays can offer a sensitive and efficient alternative, enabling the identification and differentiation of bacterial species or strains with simultaneous analysis of numerous interactions using minimal sample volumes (16). Several studies have demonstrated the efficacy of detecting serum Abs responsive to bacterial species (17,18), since bacterial cell surface antigens interact with the host immune system, eliciting Ab production (19). Distinct bacterial surface components can act as markers even among bacterial strains with similar cell surface architectures due to significant structural variations (20). Detecting bacteria-specific Ab responses is a promising approach for profiling targeted bacterial species in limited and extensive human serum samples. Therefore, this study aimed to identify SSc-associated microbial species using a serum-based microbial microarray, explore potential microbial therapeutic candidates, and elucidate the underlying mechanisms of microbiome-mediated treatment in SSc. This research would provide valuable insights regarding involvement of the gut microbiota in SSc and facilitate the development of effective microbiome-targeted therapies. MATERIALS AND METHODS Patients This study included 76 patients diagnosed with SSc and 50 healthy controls (HC). Serum samples were obtained from Pusan National, Hanyang, and Gyeongsang National University Hospital (GNUH)-Korea Biobank in Korea between 2014 and 2018. This study was approved by the Institutional Review Board of GNUH (No. GNUH 2014-02-013). Microbial microarray analysis IgM levels against specific intestinal microbes were evaluated using a previously described method (21). Briefly, lysates of 384 intestinal microbial species were spotted onto a nylon membrane using a microarray spotter (CapitalBio, Beijing, China) and incubated with serum samples, followed by treatment with Alexa Fluor 647-conjugated goat anti-human IgM secondary Ab (Thermo Fisher Scientific, Waltham, MA, USA). Fluorescence intensity, indicating IgM levels, was quantified using Luxscan (CapitalBio). Richness, evenness, and compositional diversity were assessed. Linear discriminant analysis (LDA) effect size (LEfSe) and heatmap visualization were utilized to identify Abs against differentially enriched taxa (p<0.05, LDA score>3.0) and microbial species (p<0.01, Z-ratio>1.5), respectively. Canonical correspondence analysis (CCA) was conducted to investigate potential relationships between significantly different antimicrobial Abs and clinical parameters in patients with SSc. Induction of SSc in mice and treatment with Bifidobacterial strains Six-week-old male C57BL/6 mice (Koatech, Seoul, Korea) were acclimatized for 2 wk before initiating the experiments. Mice were randomly assigned to the following 3 experimental groups: 1) PBS-treated control (normal), 2) Bleomycin (BLM)-induced SSc (BLM), 3) Bifidobacterium-treated SSc (BLM/Bifidobacterium). SSc was induced following a previously described protocol (9). Briefly, mice received subcutaneous injections of 100 µg BLM (Dong-A ST, Seoul, Korea) at a single dorsal location, 5 times a week for 2 wk. The Bifidobacterium strains used in this study were Bifidobacterium longum BORI, B. longum RAPO, and B. bifidum BGN4, provided by BIFIDO Co., Ltd., Hongcheon, Korea. Live Bifidobacterium strains were orally administered at a dose of approximately 1–5×109 colony-forming units (CFUs) per mouse, 5 times a week for 4 wk, beginning 2 wk before BLM injections. All animal experiments were conducted in accordance with the guidelines of Gyeongsang National University (GNU) and approved by the Institutional Animal Care and Use Committee in Korea (approval ID: GNU-221212-M0179). Histological examination Formalin-fixed, paraffin-embedded skin and lung sections (5 μm) from mice were stained with Masson’s trichrome and examined under a Nikon Eclipse Ni-U microscope (Nikon Instruments, Tokyo, Japan). Dermal thickness was measured at 6 sites within each skin section to assess skin fibrosis. Fibrotic changes in lung tissues were evaluated using the Ashcroft score (22), along with collagen and alveolar area percentages at 9 sites. Scores were averaged from multiple sites. Immunofluorescence and immunohistochemistry (IHC) staining Skin and lung tissue sections were deparaffinized before staining. For immunofluorescence, sections were incubated for 16 h with primary Abs against alpha smooth muscle actin (α-SMA, Abcam, Cambridge, UK), CD11b (Santa Cruz Biotechnology, Dallas, TX, USA), and CX3CR1 (Santa Cruz Biotechnology) at 4°C, followed by treatment with a fluorescence-conjugated secondary Ab (anti-mouse Alexa Fluor 488, anti-rat Alexa Fluor 488, or anti-mouse Alexa Fluor 594, Thermo Fisher Scientific) at 20°C–25°C for 2 h. Slides were mounted using an antifade reagent containing DAPI (Thermo Fisher Scientific). For IHC, sections were stained using a VECTASTAIN Elite ABC Peroxidase Kit (Vector Laboratories, Burlingame, CA, USA). α-SMA Ab (Abcam) was applied, followed by incubation with a biotinylated secondary Ab (Vector Laboratories), and color development using the 3,3′-diaminobenzidine systems (Sigma-Aldrich, St. Louis, MO, USA). Images were captured using a Nikon microscope imaging system at 400× magnification. Quantification of fluorescent signals and α-SMA positivity was performed using the ImageJ software (National Institutes of Health, Bethesda, MD, USA). Digital polymerase chain reaction B. longum and its strain RAPO were quantified using the QIAcuity digital PCR system (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Briefly, total bacterial genomic DNA was extracted from mice fecal samples using the MagMAX™ Microbiome Ultra Nucleic Acid Isolation Kit (Thermo Fisher Scientific). Reaction mixtures, containing QIAcuity Probe Master Mix (Qiagen), primer/probe mix (Supplementary Table 1), and template DNA, were transferred to QIAcuity Nanoplate 96-well (Qiagen). The nanoplate was then loaded onto the QIAcuity One 5plex system (Qiagen) for automated partitioning and PCR amplification (95°C for 2 min, followed by 50 cycles of 95°C for 15 s and 60°C for 30 s). After thermocycling, fluorescence from each well was analyzed using QIAcuity Software Suite (Qiagen). Western blot analysis Tissue and cell proteins were extracted, and western blotting for α-SMA was processed as previously described (9). Signal quantification was performed using Image Lab software (Bio-Rad, Hercules, CA, USA), normalized to β-actin. Cell isolation and flow cytometric analysis Bronchoalveolar lavage (BAL) fluid, mesenteric lymph node (MLN), and spleen cells were obtained from mice according to a previously described protocol (9). Subsequently, cells were stained with fluorochrome-conjugated Abs (Supplementary Table 2) and analyzed using an LSRFortessa™ X-20 flow cytometer (BD Biosciences, San Jose, CA, USA). FlowJo software (Tree Star Inc., Ashland, OR, USA) was used for further data analysis. Differential cell counts and measurements of total protein and cytokine concentrations in BAL fluid Slides were prepared by cytocentrifugation of BAL cells and stained with Diff-Quik (BioGnost, Zagreb, Croatia). Differential cell counts of leukocyte subsets were determined by examining a minimum of 300 cells per slide under a light microscope. The total protein concentration of BAL fluid was determined using the Bradford assay (Bio-Rad). Levels of IL-6, TNF-α, and IL-1β in the BAL fluid were measured using DuoSet ELISA kits (R&D System, Minneapolis, MN, USA), following the manufacturer’s instructions. Cytokine production of MLN and spleen cells MLN and spleen cells isolated from experimental mouse groups were seeded into a 96-well plate at a density of 5×105 cells per well. The cells were stimulated with 1 µg/ml of LPS (Sigma-Aldrich) at 37°C with 5% CO2. After 16 h, the supernatants were collected and the levels of IL-6 and TNF-α were measured. IL-10 levels were measured in the supernatants collected after 40 h using the OptEIA™ Mouse IL-10 ELISA Set (BD Biosciences). Quantitative real-time PCR Total RNA extraction and subsequent quantitative real-time PCR were conducted as previously described (9). Relative expression levels of target genes were determined using the 2−ΔΔCt comparative method, normalized to GAPDH. Details on TaqMan assays and SYBR primers can be found in Supplementary Table 3. 16S rRNA gene sequencing and analysis Bacterial genomic DNA was extracted from mouse fecal samples, and sequencing was conducted as previously described (21). Raw sequence reads were analyzed using QIIME2 software. All amplicon sequence variants were classified using the SILVA 138 99% database. Microbial richness, evenness diversity indices, and principal coordinate analysis (PCoA) were performed using the QIIME2 pipeline and R software (R Foundation for Statistical Computing, Vienna, Austria). All sequencing raw data are available at the NCBI Sequence Read Archive repository, accession number, PRJNA1108506 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1108506). Short-chain fatty acid (SCFA) analysis The concentration of SCFAs in mouse fecal samples was measured using gas chromatography/mass spectrometry (GC-2010 Plus, GCMS-TQ 8030, Shimazu, Tokyo, Japan). Cell culture and treatment The murine macrophage cell line RAW 264.7 (1×105 cells/well) was cultured for 24 h in a 24-well plate using DMEM, supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin (Gibco, Carlsbad, CA, USA) at 37°C with 5% CO2. The cells were then co-cultured with different doses of heat-killed (HK)-B. longum RAPO (multiplicity of infection: 0, 0.1, or 1) or cell-free culture supernatant (CS) derived from live B. longum RAPO (final concentrations of 0, 0.01, or 0.1%) for 20 h, followed by stimulation with 1 µg/ml LPS for 4 h. Detailed information on HK-bacteria and CS preparation is provided in Supplementary Data 1. Primary normal human dermal fibroblasts (NHDFs, ATCC, Gaithersburg, MD, USA) were seeded in a 6-well plate (2×105 cells/well) using DMEM with supplements. After 24 h of culture, NHDFs were treated with 10 ng/ml of human TGF-beta1 (hTGF-β1, R&D System) with or without B. longum RAPO CS (2%) for 48 and 72 h. Normal lung fibroblasts (normal-LFs, LL24, ATCC) and LFs diagnosed with idiopathic pulmonary fibrosis (IPF-LFs, LL29, ATCC) were cultured in Ham’s F-12 medium (Gibco) with supplements. The fibroblasts were stimulated with 5 ng/ml of hTGF-β1 with or without B. longum RAPO CS (2%) for 72 h. Statistical analyses Data analysis and visualization were conducted using IBM SPSS (Chicago, IL, USA) or GraphPad Prism (San Diego, CA, USA). Statistical significance was assessed using analysis of variance or Kruskal-Wallis test, followed by post hoc tests. Permutational multivariate analysis of variance was used to evaluate differences in the microbial community. Results are presented as mean ± SD unless otherwise specified, and p<0.05 was considered significant. RESULTS Alterations in serum antimicrobial Ab levels in patients with SSc Our study involved the use of a serum-based bacterial microarray to investigate microbial alterations and identify potential therapeutic bacterial targets specific to SSc. Serum samples from patients with SSc and HC were analyzed for IgM Ab levels against 384 gut microbial species. Detailed patient characteristics are summarized in Supplementary Table 4. Although no significant differences were observed in the α-diversity of serum antimicrobial Abs (Fig. 1A), the β-diversity clustering pattern differed considerably between patients with SSc and HC (Fig. 1B). Through LEfSe analysis, we identified core antimicrobial Abs distinguishing patients with SSc from HC, including enrichment in anti-Actinobacteria and anti-Bacteroidetes Abs in HC and anti-Proteobacteria Abs in patients with SSc at the phylum level (Fig. 1C). At lower taxonomic levels, HC exhibited enriched antimicrobial Abs in classes Actinobacteria, Coriobacteriia, and Bacteroidia, as well as orders Bifidobacteriales, Corynebacteriales, Coriobacteriales, Bacteroidales, and Bacillales, spanning 4 families and 4 genera. Conversely, patients with SSc demonstrated enrichment in antimicrobial Abs against classes Clostridia and Tissierellia, encompassing 4 orders, 4 families, and one genus. Comparison of serum antimicrobial Ab profile between HC (n=50) and patients with SSc (n=76). (A) Alpha-diversity assessment of antimicrobial Ab abundances (Mann-Whitney test). (B) PCoA plot based on Bray-Curtis dissimilarity of antimicrobial Ab communities (PERMANOVA, p<0.001). Each symbol represents an individual. (C) Cladogram generated using LEfSe analysis. The phylogenetic tree shows taxa enriched in HC and SSc represented in green and red, respectively (pairwise Wilcoxon test, p<0.05, LDA score>3.0). (D) Heatmap illustrating differential antimicrobial Ab abundances between HC and patients with SSc (t- and z-tests, p<0.01, Z-ratio>1.5). Colors represent abundance levels, ranging from blue (low abundance) to red (high abundance). (E) CCA diagram visualizing the relationship between differential antimicrobial Abs (circles) and clinical phenotypes (arrows) in patients with SSc. (F) Comparative levels of significantly different antimicrobial Abs against Bifidobacterium species among HC and patients with SSc, stratified according to clinical phenotypes. lc, limited cutaneous; Mi, Mild; Mo, Moderate; mRSS, Modified Rodnan skin score. One-way analysis of variance, Sidak post hoc test, **p<0.01 vs. HC. Our investigation revealed 28 significantly altered antimicrobial Abs at the species level in patients with SSc, compared to those in HC, among which 12 exhibited reduced levels and 16 displayed increased levels (Fig. 1D). Using CCA and Pearson correlation analysis, we found potential associations between specific antimicrobial Abs and clinical phenotypes of SSc (Fig. 1E, Supplementary Table 5). Notably, Abs against Corynebacterium diphtheriae, Enorma timonensis, Escherichia vulneris, Citrobacter werkmanii, Pseudomonas fluorescens, and Lactobacillus animalis correlated with the presence of autoantibodies in patients with SSc (Fig. 1E). Particularly, anti-C. diphtheriae, anti-P. fluorescens, and anti-L. animalis Abs positively correlated with the presence of anti-Scl-70, while anti-E. vulneris and anti-C. werkmanii Abs correlated positively with anti-centromere autoantibody presence (Supplementary Table 5). Moreover, anti-Bacillus thermoamylovorans Ab correlated with cutaneous type and lung involvement in patients with SSc (Fig. 1E), with negative correlations observed with diffused cutaneous type (dc) and lung involvement (Supplementary Table 5). Among the decreased serum antimicrobial Abs in patients with SSc, we focused on those targeting Bifidobacterium adolescentis and B. longum. These well-known probiotics are recognized for promoting the growth of butyrate-producing bacteria by supplying acetate as a substrate (23), potentially contributing to SSc suppression through butyrate formation (9). Remarkably, Ab levels against both these Bifidobacterium species were diminished in patients with SSc compared to those in HC, irrespective of the clinical phenotypes of SSc (Fig. 1F). Specifically, B. longum was the predominant probiotic species in HC compared to those in patients with SSc, as indicated by the LDA score (Supplementary Fig. 1). Subsequently, we comprehensively investigated the therapeutic potential of B. longum as a promising microbial target for treating SSc. B. longum RAPO ameliorates dermal and lung fibrosis in BLM-induced SSc mice To determine the therapeutic potential of B. longum in SSc, we explored the antifibrotic efficacy of the B. longum strains, BORI and RAPO, along with that of B. bifidum BGN4, in a BLM-induced murine SSc model. SSc was induced in mice by subcutaneous injections of 100 µg BLM at the dorsal lesion site, administered 5 times a week for 2 wk. The Bifidobacterium strains were orally administered 5 times a week for 4 wk, starting 2 wk before BLM injections (Fig. 2A). Among SSc mice orally administered Bifidobacterium strains at a dose of 2×109 CFUs per mouse, significant improvements in both dermal and lung fibrosis were exclusively observed in the B. longum RAPO-treated group (Fig. 2B-D). B. longum RAPO notably reduced dermal thickness (Fig. 2C) and attenuated lung fibrotic changes indicated by the Ashcroft score, collagen area, and alveolar area (Fig. 2D). The B. longum BORI-treated group partially ameliorated lung fibrosis, manifested by an increase in alveolar area (Fig. 2D). In an additional experimental study involving various doses of administered Bifidobacterium strains, 2×109 CFUs of B. longum RAPO effectively mitigated both dermal and lung fibrosis (Fig. 2E-G). Neither B. longum BORI nor B. bifidum BGN4 suppressed dermal fibrosis at the tested doses ranging from 1×109 to 5×109 CFUs (Supplementary Fig. 2). Effects of Bifidobacterium strains on dermal and lung fibrosis in BLM-induced SSc mice. (A) Experimental scheme: SSc was induced in mice through subcutaneous injections of BLM to the dorsal skin, 5 times per week for 2 wk. Oral administration of Bifidobacterium strains commenced 2 wk before BLM injection and continued for the subsequent 4 wk. Skin and lung tissues were then obtained from mice in each experimental group to evaluate fibrosis. (B-D) Representative images of Masson’s trichrome stain illustrating skin and lung fibrosis (B), evaluation of skin fibrosis by dermal thickness (C), and assessment of lung fibrosis using the Ashcroft score and the percentages of collagen and alveolar area (D) in skin and lung tissues obtained from SSc mice treated with PBS (n=6), Bifidobacterium longum BORI (n=7), B. longum RAPO (n=6), or B. bifidum BGN4 (n=4). Arrow length indicates dermal thickness in skin tissues. *p<0.05, **p<0.01 vs. PBS. (E-G) Representative images of Masson’s trichrome stain (E) and evaluation of skin (F) and lung fibrosis (G) in skin and lung tissues obtained from normal (n=5) and SSc mice treated with B. longum RAPO at doses of 0 (n=13), 1 (n=9), 2 (n=14), or 5×109 CFUs (n=7). *p<0.05, **p<0.01. Scale bars=100 µm. Mean ± SD. Consistently reproducible suppression of BLM-induced dermal and lung fibrosis was observed with 2×109 CFUs of B. longum RAPO in independent in vivo experiments (Fig. 3A-C). Successful colonization of the administered B. longum RAPO in BLM mice was confirmed, demonstrating a notable elevation in the fecal levels of B. longum RAPO following its administration (Fig. 3D). Immunofluorescence and IHC staining revealed a marked reduction in the expression of α-SMA, a myofibroblast marker, in the skin and lung tissues of BLM mice following B. longum RAPO administration (Fig. 3E). Moreover, B. longum RAPO administration decreased α-SMA protein expression in the skin of BLM mice, although the observed effect on lung tissue was not significant (Fig. 3F). In summary, B. longum RAPO, particularly at a dose of 2×109 CFUs, effectively attenuated dermal and lung fibrosis in BLM-induced SSc mice by impeding structural fibrotic changes and myofibroblast differentiation. Antifibrotic effects of B. longum RAPO on dermal and lung fibrosis in BLM-induced SSc mice. SSc was induced in mice through subcutaneous injections of BLM to the dorsal skin for 2 wk. Oral administration of B. longum RAPO, at a dose of 2×109 CFUs, commenced 2 wk before BLM injection and continued for 4 wk. Skin, lung, and fecal samples were then obtained from mice in each experimental group. (A-C) Representative images of Masson’s trichrome stain showing skin and lung fibrosis (A), accompanied by quantitative analysis of fibrosis in skin (B) and lung tissues (C) obtained from mice of normal (n=6), BLM (n=16), and BLM/RAPO (n=19) groups. Scale bars=100 µm. (D) Quantification of B. longum and its strain RAPO in fecal samples using digital PCR (n=5–6/group). (E) Representative images of immunofluorescence (green) and IHC (brown) staining for α-SMA in skin and lung tissues, respectively, and quantification of α-SMA expression (n=3–5 for normal, 5–8 for BLM, and 5–7 for BLM/RAPO group). Scale bars=50 µm. (F) Expression of α-SMA protein in skin and lung tissues (n=3/group). Mean ± SD, *p<0.05, **p<0.01. B. longum RAPO inhibits pro-inflammatory macrophage infiltration into fibrotic lesions in BLM-induced SSc mice In our previous investigation, we noted significant macrophage infiltration in the fibrotic skin and lung tissues of SSc mice (9). To elucidate the mechanistic basis of B. longum RAPO’s therapeutic efficacy in SSc, we examined macrophage populations in the skin and lungs of B. longum RAPO-treated versus untreated BLM mice (Fig. 4). Immunofluorescence analysis revealed a notable reduction in the percentage of CD11b+CX3CR1+ macrophage areas in the skin of BLM mice treated with B. longum RAPO (Fig. 4A). Morphological evaluation of BAL cells revealed that macrophages were the predominant cell type in the BAL fluid of BLM mice, exhibiting an activated phenotype characterized by enlarged and vacuolated cytoplasm, which was alleviated by B. longum RAPO treatment (Fig. 4B). Furthermore, B. longum RAPO effectively diminished the infiltration of total immune cells, particularly macrophages, in the BAL fluid of BLM mice. Anti-inflammatory effects of B. longum RAPO on dermal and lung fibrosis in BLM-induced SSc mice. SSc was induced in mice through subcutaneous injections of BLM to the dorsal skin for 2 wk. Oral administration of B. longum RAPO commenced 2 wk before BLM injection and continued for 4 wk. Skin, lung, and BAL fluid were then obtained from mice of normal (n=3–4), BLM (n=4–7), and BLM/RAPO (n=5–6) groups. (A) Immunofluorescent images and quantification showing expression of CD11b+CX3CR1+ macrophages in skin tissues stained for DAPI (blue), CD11b (green), and CX3CR1 (red). Scale bars=50 µm. (B) Differential cell counts of BAL cells stained with Diff-Quik. Scale bars=25 µm. (C) Flow cytometric analysis of macrophage populations in BAL cells. Total macrophages (MPs, CD45+CD64+) were assessed, and macrophage subsets were further divided into alveolar macrophages (CD11c+CD11b−) and exudate macrophages (exMac, CD11c+CD11b+). (D, E) Quantification of total protein (D) and pro-inflammatory cytokine concentrations in BAL fluid (E). (F) Expression of genes for profibrotic and pro-inflammatory cytokines and chemokine in the skin and lung tissues. Mean ± SD, *p<0.05, **p<0.01. Flow cytometric analysis of BAL cells (Supplementary Fig. 3), demonstrated a significant reduction in the number of total macrophages (CD45+CD64+), particularly alveolar macrophage (CD45+CD64+CD11c+CD11b−), in B. longum RAPO-treated BLM mice compared to that in untreated BLM mice (Fig. 4C). Moreover, B. longum RAPO administration decreased the accumulation of inflammatory components, such as total protein and IL-6, in BAL fluid of BLM mice (Fig. 4D and E). Gene expression analysis revealed the downregulation of profibrotic (Acta2, Ctgf, and Tgfb1) and pro-inflammatory cytokines (Il6, Il1b, and Tnfa) and monocyte/macrophage chemoattractant (Ccl2) genes in the skin and lungs of B. longum RAPO-treated BLM mice (Fig. 4F, Supplementary Fig. 4). In summary, B. longum RAPO treatment not only reduced macrophage recruitment to the skin and lungs but also inhibited macrophage activation and tissue inflammation within the fibrotic tissues in BLM-induced SSc mice. B. longum RAPO modulates inflammatory monocyte recruitment and immune response in the lymphoid organs of BLM-induced SSc mice In addition to macrophage infiltration observed in fibrotic skin and lung tissues, our previous study identified inflammatory monocyte recruitment to the MLN and spleen in SSc mice (9). This recruitment represents a critical early inflammatory event, followed by their differentiation into activated macrophages and subsequent downstream inflammatory responses (24). Considering the potential pathological mechanisms underlying SSc, we investigated the modulatory effect of B. longum RAPO on inflammatory monocyte recruitment to lymphoid organs. Our analysis revealed a significant inhibition of Ly6C+ inflammatory monocyte recruitment to the spleen in B. longum RAPO-treated BLM mice compared to that in untreated BLM mice, while no differences were observed in MLN between the 2 groups (Fig. 5A). Immunomodulatory impacts of B. longum RAPO on the MLNs and spleen in BLM-induced SSc mice. Mice were subjected to subcutaneous injections of BLM to the dorsal skin for 2 wk and treated with or without B. longum RAPO 2 wk before BLM injection. Cells from MLN and spleen were isolated from normal (n=3–4), BLM-treated (n=4–5), and BLM/RAPO-treated (n=4–6) mice. (A) Flow cytometric analysis characterized macrophage populations in MLN and spleen cells. The numbers and percentages of Ly6C+ inflammatory monocytes, including newly recruited (Ly6C+MHCII−) and maturing monocytes (Ly6C+MHCII+), were assessed among total macrophages (CD45+CD64+CD11b+CX3CR1+). (B) Expression levels of pro-inflammatory cytokines and chemokine genes in MLN and spleen cells. (C) Immune responses of MLN and spleen cells. Cytokine levels secreted upon stimulation with LPS were measured in CSs from MLN and spleen cells. Mean ± SD, *p<0.05, **p<0.01. To further elucidate B. longum RAPO’s mechanism in suppressing fibrosis, we assessed the cytokine and chemokine gene expressions of MLN and spleen cells isolated from mice (Fig. 5B, Supplementary Fig. 4). We found upregulated Ccl2 expression, crucial for Ly6Chi monocyte migration and macrophage activation, in MLN and spleen cells of BLM mice. This upregulation was significantly suppressed in both MLN and spleen cells of B. longum RAPO-treated BLM mice (Fig. 5B). Moreover, we noted reduced TNF-α production in the MLN cells of B. longum RAPO-treated BLM mice, while IL-10 production was enhanced in both MLN and spleen cells of B. longum RAPO-treated BLM mice compared to that in untreated BLM mice (Fig. 5C). In conclusion, the inhibitory effect of B. longum RAPO on fibrosis may be attributed to its ability to suppress inflammatory monocyte recruitment to the spleen via Ccl2 downregulation and its anti-inflammatory action in gut-draining MLN. B. longum RAPO improves gut microbial diversity and enhances the expansion of SCFA-producing bacteria in BLM-induced SSc mice To explore the impact of B. longum RAPO on gut microbiota modulation in SSc, we performed 16S rRNA gene sequencing analysis on fecal samples from experimental mice. Alpha-diversity analysis showed that B. longum RAPO treatment tended to restore both the richness and evenness of the gut microbial community disrupted in BLM-induced SSc mice (Fig. 6A). Specifically, significant enhancement in gut microbial evenness, as indicated by the Shannon index, was observed in B. longum RAPO-treated BLM mice compared to that in untreated BLM mice. Considerable differences were observed in the microbial composition among the experimental groups, as illustrated by significant differences in the overall fecal microbial community composition shown in PCoA plots (Fig. 6B). Influences of B. longum RAPO on gut microbiota in BLM-induced SSc mice. Mice were subjected to subcutaneous injections of BLM to the dorsal skin for 2 wk and treated with or without B. longum RAPO 2 wk before BLM injection. Fecal samples were collected from normal (n=5), BLM-treated (n=5), and BLM/RAPO-treated (n=6) mice, and 16S rRNA gene sequencing was conducted. (A) Alpha-diversity analysis of gut microbiota. (B) Beta-diversity assessed using PCoA based on Weighted UniFrac distance matrix. (C) Relative abundances of gut microbiota at the phylum level. (D) Heatmap illustrating significantly differential genera between groups (p<0.05). Colors represent abundance levels, ranging from blue (low abundance) to red (high abundance). (E) Measurement of SCFA levels in fecal samples from normal, BLM-treated, and BLM/RAPO-treated mice (n=4/group). (F) Expression levels of GPCR genes in MLN cells isolated from normal (n=3), BLM-treated (n=5), and BLM/RAPO-treated (n=6) mice, as determined using quantitative PCR. Mean ± SD, *p<0.05, **p<0.01. At the phylum level, Firmicutes, Actinobacteria, and Patescibacteria were predominant in normal, untreated BLM, and B. longum RAPO-treated BLM mice, respectively (Fig. 6C). Noteworthy changes at the genus level, particularly in SCFA-producing genera, were observed between groups, as depicted in a heatmap (Fig. 6D). Fecal samples from normal and B. longum RAPO-treated BLM mice were notably abundant in SCFA-producing genera, including Lachnospiraceae NK4A136 group, Roseburia, Lachnospiraceae ASF356, Oscillospiraceae, Lachnospiraceae, Eubacterium ruminantium group, Ruminococcaceae Incertae Sedis, Coriobacteriaceae UCG-002, Allobaculum, Prevotellaceae Ga6A1 group, Prevotellaceae, Candidatus Saccharimonas, Prevotellaceae NK3B31 group, and Clostridia UCG-014 (Fig. 6D, Supplementary Fig. 5). To further elucidate the potential mechanism underlying B. longum RAPO’s mitigation of skin and lung fibrosis through SCFA production, we measured the levels of SCFAs, including acetate, propionate, and butyrate, in fecal samples and examined the activation of SCFA-sensing G protein-coupled receptors (GPCRs) in MLN cells. Our results demonstrated a significant increase in fecal butyrate levels in B. longum RAPO-treated BLM mice compared to untreated BLM mice (Fig. 6E), along with upregulation of Gpr41 expression in the MLN cells from B. longum RAPO-treated BLM mice relative to untreated normal and BLM mice (Fig. 6F, Supplementary Fig. 6). These findings suggest that butyrate-mediated Gpr41 activation plays a pivotal role in the therapeutic efficacy of B. longum RAPO in the SSc mouse model. In summary, B. longum RAPO treatment effectively restores fecal microbial diversity, enhances the abundance of SCFA-producing gut bacteria, increases fecal butyrate production, and activates the SCFA receptor Gpr41 in gut-draining lymphoid cells, thereby potentially exerting a suppressive effect on the disease. B. longum RAPO and its CS suppress the expression of pro-inflammatory and profibrotic mediators in macrophages and fibroblasts To investigate the direct modulation of macrophage activation by B. longum RAPO, its secreted components, or metabolites under inflammatory conditions, we assessed pro-inflammatory gene expression induced by LPS in RAW 264.7 macrophages cultured with or without HK-B. longum RAPO or its CS. Doses of HK-B. longum RAPO or its CS were determined based on cytotoxicity evaluation (Supplementary Fig. 7A). Both the samples treated with HK-B. longum RAPO and its CS exhibited significant dose-dependent downregulation of LPS-induced genes, including Il6, Il1b, Tnfa, and Ccl2 in RAW 264.7 cells (Fig. 7A). Modulating ability of B. longum RAPO and its derived components on macrophages and fibroblasts activation in vitro. (A) Pro-inflammatory gene expression of RAW 264.7 macrophage cells stimulated with LPS in the absence or presence of HK-B. longum RAPO or its CS (n=3–4/group). (B, C) α-SMA protein expression (B) and profibrotic and pro-inflammatory gene expression (C) of primary NHDFs stimulated with hTGF-β1 in the absence or presence of B. longum RAPO CS. (D, E) α-SMA protein expression (D) and profibrotic and pro-inflammatory gene expression (E) of normal-LFs (LL24) and IPF-LFs (LL29) stimulated with hTGF-β1 in the absence or presence of B. longum RAPO CS (n=3/group). Mean ± SEM, *p<0.05, **p<0.01. To further explore the potential anti-fibrotic mechanism of B. longum RAPO mediated by soluble molecules, we evaluated the trans-differentiation of fibroblasts into myofibroblasts induced by hTGF-β1 while treating dermal and LFs with or without of B. longum RAPO CS. The concentrations of B. longum RAPO CS were selected based on cytotoxicity evaluation (Supplementary Fig. 7B). B. longum RAPO CS effectively inhibited hTGF-β1-induced myofibroblast transition in NHDFs, as evidenced by reduced α-SMA protein expression (Fig. 7B). Consistently, B. longum RAPO CS suppressed the expression of profibrotic genes such as ACTA2 and CTGF, as well as pro-inflammatory genes including IL6 and IL1B, during co-culture with NHDFs (Fig. 7C). Furthermore, B. longum RAPO CS significantly inhibited hTGF-β1-induced α-SMA protein expression in both normal-LF (LL24) and IPF-LF (LL29) cell lines (Fig. 7D). In normal-LFs, B. longum RAPO CS partially downregulated the expression of hTGF-β1-induced CTGF and IL6. Notably, in IPF-LFs, the CS exhibited a pronounced ability to suppress the expression of hTGF-β1-induced pro-inflammatory genes (IL6, IL1B, and TNFA) rather than profibrotic genes (Fig. 7E). Targeted metabolomic analysis of SCFAs revealed a predominant increase in acetate levels in B. longum RAPO CS compared to that in control media (Supplementary Fig. 7C), suggesting the involvement of B. longum RAPO-derived acetate and its downstream metabolite, butyrate, in regulating the activation of macrophages and fibroblasts in vivo. Additionally, co-culture experiments demonstrated reduced α-SMA protein expression in NHDFs co-cultured with B. longum RAPO-treated THP-1 macrophages compared to NHDFs co-cultured with untreated THP-1 macrophages (Supplementary Fig. 8). Overall, these findings suggest that B. longum RAPO may suppress fibrosis development by directly or indirectly modulating macrophages and fibroblasts through either its intrinsic properties or secreted molecules. DISCUSSION This study unveiled significant alterations in antimicrobial Abs in patients with SSc, identified by changes in serum IgM Abs targeting gut microbial species. Particularly striking was the marked depletion of Abs against B. longum in patients with SSc. Moreover, the oral administration of B. longum strain RAPO effectively attenuated skin and lung fibrosis in SSc mice. These therapeutic effects were linked to the enhanced growth of SCFA-producing gut bacteria, increased fecal butyrate production, Gpr41 activation in MLN cells, and reduced infiltration of inflammatory macrophages/monocytes, along with Ccl2 downregulation in the fibrotic tissues and spleen. These findings provide initial evidence supporting the potential of B. longum RAPO as a promising probiotic for protecting against tissue fibrosis associated with SSc. Genomic methodologies are indispensable for assessing microbiome diversity; however, accurately identifying microbial species and comprehending host-microbiota interactions are still challenging, necessitating innovative strategies (25). Here, we introduced a microarray to scrutinize distinct microbial species and therapeutic targets in SSc by analyzing IgM reactions against a specifically designed gut microbiota panel in human serum samples. Systemic reactivity to microbial components reflects impaired mucosal immunity, with clinical implications (26). Elevated serum microbial Abs further emphasize the gut microbiota’s role in disease pathogenesis (27). Circulating Abs against microbial peptides have been implicated in conditions such as inflammatory bowel disease, where they are thought to arise from immune responses to luminal microbiota. Notably, these Abs correlated with more severe disease phenotypes, with a higher prevalence observed in patients with more severe Crohn’s disease compared to milder forms (28). Moreover, translocated microbial ligands can activate both immune and nonimmune cells, including fibroblasts, via toll-like receptors, thereby driving fibrosis in multiple organs (27). Collectively, these findings suggest that immune responses to microbial antigens contribute to the development of fibrosis and the production of antimicrobial Abs. Although serum Ab disturbances to gut microbiota have been documented in various diseases (29,30), specific Ab reactions against microbial antigens in SSc remain unexplored. However, fecal microbiota transplantation in SSc patients has been shown to modify IgA and IgM coating patterns and extent (31), suggesting the potential existence of gut microbiota-specific Ab responses or repertoire in the disease. While the current study could not establish a definitive relationship among antimicrobial Abs, microbial abundance, and disease severity, our findings indicate that altered Ab profiles may reflect immune responses to microbial changes and offer clues on the associations between specific microbial species and clinical symptoms. Previous studies have highlighted gut microbial disturbances in SSc, primarily concerning gastrointestinal symptoms. Patients with mild symptoms exhibited increased levels of Clostridium, Blautia, and Bacteroides fragilis, while those with severe symptoms had higher Fusobacterium levels (6,32). Microbial profiles also varied by constipation severity, with Lactobacillus elevated in mild cases, Parabacteroides in severe cases (6). Akkermansia muciniphila was associated with severe diarrhea and malnutrition (33). Beyond the gut, skin microbiome alterations included elevated Rhodotorula glutinis in early dcSSc, decreased Propionibacterium and Malassezia, and increased gram-negative bacteria like Burkholderia, Citrobacter, and Vibrio in lcSSc or dcSSc patients (34,35). Furthermore, decreased microbial diversity correlated with reduced lung function and higher mortality (36). These findings underscore the value of a serum-based approach for characterizing microbiome alterations across organ systems in SSc, a disease with heterogenous phenotypes. Robust antigen-specific Abs responsive to microbiota may provide unique insights into exploring specific microbiota members associated with SSc clinical phenotypes and their role in pathophysiology. While IgA is the dominant Ab at mucosal sites (37), IgM also holds significance in maintaining local tolerance and influencing microbiota composition in mucosal surfaces (38). IgM can swiftly respond to external triggers as the primary humoral Ab, closely mirroring disease-relevant processes or pathogen-specific responses (39). The pentameric structure of IgM enables efficient recognition of microbial antigens through multivalent binding (40), effectively pinpointing bacteria-specific reactions at early and current disease stages. However, considering the overlaps and distinctions in the reactivities of Ab isotypes against microbiota members (41), further investigation of microbiota-specific IgA and IgG along with IgM may offer comprehensive insights into microbial alterations associated with disease progression. In this study, we selected B. longum as an initial candidate for SSc treatment due to its SCFA production, which is implicated in SSc therapy (7,9), and its highest-ranked probiotic species in our bacterial microarray analysis. Recent studies have highlighted B. longum’s therapeutic potential in improving skin health (42) and reducing lung inflammation (43). Among 2 B. longum strains tested in an SSc mouse model only B. longum RAPO significantly mitigated both dermal and lung fibrosis. To determine whether these effects were due to general bifidobacterial mechanisms or strain-specific properties, we compared the 2 B. longum strains with another probiotic Bifidobacterium species, B. bifidum BGN4, which exhibits distinct prevalence and metabolic profiles. B. bifidum is less abundant in adults, whereas B. longum is one of the most prevalent species in healthy adults, potentially reflecting differences in metabolic activity and cell adhesion properties (44,45). Notably, our findings indicated that B. bifidum BGN4 lacked therapeutic efficacy in SSc mice. Although the current study did not elucidate the mechanistic differences underlying the differential effects of these strains in the SSc model, our findings underscore that therapeutic attributes vary significantly among probiotic strains, necessitating careful selection. Future research is required to investigate strain-specific metabolic activities, immunomodulatory capabilities, and host-microbiome interactions to better understand the therapeutic potential of B. longum and other Bifidobacterium strains in SSc. The anti-inflammatory effects of B. longum RAPO appear to depend on both species- and strain-specific properties. Among the 32 species within the Bifidobacterium genus, B. longum has demonstrated superior anti-inflammatory efficacy (46). Strain-specific comparisons, including B. longum, B. adolescentis, and B. bifidum, revealed that B. longum RAPO most effectively suppressed proinflammatory cytokines (47), highlighting the importance of strain-specific differences in reducing dermal and lung fibrosis in SSc mice. Additionally, subcellular components of bifidobacteria, such as cytoplasmic extracts, surface proteins, and supernatants, are known to elicit strain-dependent immune responses (48). For instance, B. longum strains exhibit varying capacities to induce IL-10 and regulate TNF-α and INF-γ, while certain B. bifidum strains are linked to CD8+ T cell activation and Th17 polarization. The lack of efficacy observed with B. bifidum BGN4 in this study likely reflects these differences in immune modulation. Based on the known predominant infiltration of inflammatory monocytes/macrophages across multiple organs in humans and mice with SSc (9,49,50), we explored the potential of B. longum RAPO treatment to ameliorate skin and lung fibrosis in SSc mice by targeting these cells. Our findings revealed that B. longum RAPO administration reduced the recruitment/accumulation of inflammatory monocytes/macrophages at affected sites in SSc mice. Since monocytes/macrophages are known to drive inflammatory and fibrotic processes in SSc (51), the observed inhibition of their infiltration by B. longum RAPO suggests a potential mechanism for alleviating disease progression in SSc. Moreover, CCL2 likely plays a pivotal role in recruiting inflammatory monocytes/macrophages to damaged lesions in SSc. Elevated CCL2 expression has been observed in the affected skin of patients with SSc (52), with serum levels correlating with disease activity and poor survival (53). The expression of CCR2 by inflammatory monocytes/macrophages further implicates their involvement in SSc pathogenesis (24). This CCL2-CCR2 axis strongly indicates the attraction of these cells to the fibrotic niche in SSc (51), where they differentiate into effector cells and activate T cells and fibroblasts (54). Our study demonstrates that B. longum RAPO effectively inhibited Ccl2 expression in fibrotic tissues and secondary lymphoid organs of SSc mice, thereby mitigating tissue fibrosis by impeding the recruitment of monocytes/macrophages to affected lesions. Furthermore, our in vitro experiments revealed that B. longum RAPO can dampen macrophage activation under inflammatory conditions by downregulating the expressions of LPS-induced Ccl2 and pro-inflammatory cytokine genes. These findings support the notion that B. longum RAPO may influence disease suppression by modulating the CCL2-CCR2 loop in affected organs, involving monocytes/macrophages. The therapeutic effects of B. longum RAPO in SSc may be closely linked to its acetate production, a primary end-product of Bifidobacterium metabolism. Acetate serves as substrate for butyrate-producing bacteria, facilitating butyrate-mediated anti-inflammatory and immunoregulatory effects in immune cells, such as macrophages (55). Our findings revealed an increased abundance of butyrate-producing genera, including E. ruminantium, Ruminococcaceae, Allobaculum, and Prevotellaceae, as well as elevated butyrate levels in the feces of B. longum RAPO-treated SSc mice. These observations suggest that butyrate-mediated anti-inflammatory mechanisms play a critical role in the therapeutic effects of B. longum RAPO. SCFAs modulate immunity through histone deacetylases inhibition or GPCR activation (56,57). While acetylated histone H3, linked to butyrate’s anti-fibrotic effects (9), was not detected in this study, we observed upregulated Gpr41 expression in MLN cells of SSc mice treated B. longum RAPO. GPR41, which preferentially binds propionate and butyrate (58), has shown anti-inflammatory effects via butyrate-mediated activation in obesity treated with B. longum subsp. infantis (59), supporting its potential therapeutic role in the butyrate-GPR41 axis. However, in this study, Gpr41 activation was confined to MLN, with no significant changes in GPCR gene expression in fibrotic skin and lung tissues. This suggests that B. longum RAPO may indirectly modulate systemic fibrotic process through immune regulation in gut-associated lymphoid tissues. Although the precise role of GPR41 in fibrosis remains unclear (60), this study highlights the enrichment of butyrate-producing gut bacteria and increased fecal butyrate production following B. longum RAPO treatment. These changes, coupled with Gpr41 activation in MLN cells, suggest a mechanism by which B. longum RAPO influences extra-intestinal organs in SSc. Despite the known therapeutic potential of probiotics, their mechanisms of action on extra-intestinal organs are still unclear. Our previous study in SSc mice suggested microbial metabolites influencing distant organs (9). Our current in vitro findings showed that B. longum RAPO CS suppressed pro-inflammatory gene expression in macrophages and inhibited myofibroblast transition in fibroblast cell lines, suggesting potential disease prevention in distant organs by mediating components or metabolites secreted by B. longum RAPO. This is consistent with prior research demonstrating the immune-modulatory effects of CS from probiotic strains (61,62). Although the precise components derived from B. longum RAPO and their in vivo kinetics were not investigated in the present study, we propose its potential to inhibit SSc development through interactions with various cell types in circulation. This study has several limitations. First, the serum microbial array may not directly reflect the gut microbiota abundance or activity due to potential antigen cross-reactivity, requiring further validation using fecal samples in the same individuals. Second, the small sample size within each SSc clinical category limited robust correlations between clinical features and microbial species. Third, we were unable to investigate the distinct responses of B. longum RAPO compared to other Bifidobacterium species or B. longum strains. Reliable techniques need to be developed to assess strain-specific differences in metabolic and immunological activities in vitro. Fourth, while a dose of 2×109 CFU B. longum RAPO was identified as optimal for mitigating both skin and lung fibrosis in SSc mice, clear dose-dependent effects were not observed in treating these fibrotic conditions. Further studies are necessary to refine dosing strategies to better understand the mechanisms underlying tissue-specific responses and the non-linear effects of B. longum RAPO. Lastly, although this study proposed butyrate-mediated Gpr41 activation as a potential mechanism underlying the therapeutic effects of B. longum RAPO administration, the precise role of Gpr41 was not fully investigated. Future research is needed to identify the specific cell types expressing Gpr41 and clarify its functional roles in mediating the therapeutic benefits in this disease model. In conclusion, our study has identified B. longum RAPO as a novel probiotic with protective effects against BLM-induced dermal and lung fibrosis. These effects are attributed to the modulation of inflammatory monocytes/macrophages and the augmentation of SCFA-producing gut bacteria. Moreover, our findings elucidate potential mechanisms underlying the efficacy of B. longum RAPO against SSc, suggesting its influence on extra-intestinal tissues. We propose B. longum RAPO as a promising probiotic candidate for SSc treatment and present a valuable strategy for targeting microbiota using a serum-based microbial microarray. ACKNOWLEDGEMENTS This work was supported by the National Research Foundation of Korea (NRF) grants (NRF-2020R1I1A3071922, NRF-2022R1F1A1071114) funded by the Korea government (http://www.nrf.re.kr/index). Abbreviations alpha smooth muscle actin bronchoalveolar lavage bleomycin canonical correspondence analysis colony-forming unit culture supernatant diffused cutaneous type Gyeongsang National University Gyeongsang National University Hospital G protein-coupled receptor healthy controls heat-killed human transforming growth factor-beta1 lung fibroblast diagnosed with idiopathic pulmonary fibrosis immunohistochemistry linear discriminant analysis linear discriminant analysis effect size lung fibroblast mesenteric lymph node normal human dermal fibroblast principal coordinate analysis short-chain fatty acid systemic sclerosis Footnotes Conflict of Interest: The authors declare no potential conflicts of interest. Author Contributions: Conceptualization: Hong ST, Lee SI.Data curation: Park HJ, Yu D, Lee J, Park SJ, Kim S.Formal analysis: Park HJ, Yu D, Hong ST, Park SJ, Kim S.Funding acquisition: Kim S, Lee SI.Investigation: Park HJ, Lee SG, Sohn DH, Jun JB, Kim, S.Methodology: Hong ST, Park MS, Lee SI.Project administration: Lee SI.Resources: Hong ST, Park MS, Lee H, Kim M, Cheon YH, Lee SG, Sohn DH, Jun JB.Supervision: Lee SI.Validation: Hong ST, Park MS, Kim S, Lee SI.Visualization: Park HJ, Park SJ, Kim S.Writing - original draft preparation: Park HJ, Park SJ, Kim S.Writing - review & editing: Hong ST, Park MS, Kim S, Lee SI. Conceptualization: Hong ST, Lee SI. Conceptualization: Hong ST, Lee SI. Data curation: Park HJ, Yu D, Lee J, Park SJ, Kim S. Data curation: Park HJ, Yu D, Lee J, Park SJ, Kim S. Formal analysis: Park HJ, Yu D, Hong ST, Park SJ, Kim S. Formal analysis: Park HJ, Yu D, Hong ST, Park SJ, Kim S. Funding acquisition: Kim S, Lee SI. Funding acquisition: Kim S, Lee SI. Investigation: Park HJ, Lee SG, Sohn DH, Jun JB, Kim, S. Investigation: Park HJ, Lee SG, Sohn DH, Jun JB, Kim, S. Methodology: Hong ST, Park MS, Lee SI. Methodology: Hong ST, Park MS, Lee SI. Project administration: Lee SI. Project administration: Lee SI. Resources: Hong ST, Park MS, Lee H, Kim M, Cheon YH, Lee SG, Sohn DH, Jun JB. Resources: Hong ST, Park MS, Lee H, Kim M, Cheon YH, Lee SG, Sohn DH, Jun JB. Supervision: Lee SI. Supervision: Lee SI. Validation: Hong ST, Park MS, Kim S, Lee SI. Validation: Hong ST, Park MS, Kim S, Lee SI. Visualization: Park HJ, Park SJ, Kim S. Visualization: Park HJ, Park SJ, Kim S. Writing - original draft preparation: Park HJ, Park SJ, Kim S. Writing - original draft preparation: Park HJ, Park SJ, Kim S. Writing - review & editing: Hong ST, Park MS, Kim S, Lee SI. Writing - review & editing: Hong ST, Park MS, Kim S, Lee SI. SUPPLEMENTARY MATERIALS Supplementary Data 1 Methods Supplementary Table 1 Primer and probe sequences for digital PCR Supplementary Table 2 Cell staining Abs for flow cytometry analysis Supplementary Table 3 TaqMan assays and SBYR primer sequences for quantitative real-time PCR Supplementary Table 4 Clinical characteristics of healthy individuals and patients with SSc Supplementary Table 5 Pearson correlation coefficients between Ab levels against bacterial species and clinical parameters in patients with SSc Supplementary Figure 1 LEfSe analysis of serum anti-microbial Ab profile comparing patients with SSc (n=76) to HC (n=50). The histogram illustrates LDA scores of species enriched in HC and SSc in green and red, respectively (p<0.01, LDA score >2.0). Supplementary Figure 2 Effects of Bifidobacterium strains on dermal fibrosis in BLM-induced SSc mice. Mice received subcutaneous injections of BLM to the dorsal skin, administered 5 times a week for 2 wk. Various doses of B. longum BORI and B. bifidum BGN4 were orally administered 2 wk before BLM injection and subsequently continued for 4 wk. Dermal thickness was evaluated using Masson’s trichrome staining of skin tissues obtained from mice in each experimental group: normal mice (n=6), non-treated SSc mice (n=13), SSc mice treated with B. longum BORI or B. bifidum BGN4 at doses of 1, 2, or 5×109 CFUs (n=4–7/each treated). Median ± min to max. Supplementary Figure 3 Flow cytometric gating strategy for macrophage (MP) subsets in bronchoalveolar lavage (BAL) cells. CD45+CD64+ MPs were subdivided into alveolar macrophages (AM, CD11c+CD11b−) and exudate macrophages (exMac, CD11c+CD11b+). Supplementary Figure 4 Expression levels of genes associated with T cell cytokines in skin, lung, MLN, and spleen cells collected from normal (n=5), BLM-treated (n=5), and BLM/RAPO-treated (n=6) mice. Supplementary Figure 5 Analysis of fecal samples collected from normal (n=5), BLM-induced SSc (n=5), and BLM/RAPO-treated (n=6) mice using 16S rRNA gene sequencing. The relative abundance of gut microbiota was compared between groups at the genus level and significantly different genera were visualized. Supplementary Figure 6 Expression levels of GPCR genes in skin and lung tissues from normal (n=5), BLM-treated (n=5), and BLM/RAPO-treated (n=6) mice. Supplementary Figure 7 (A, B) Cytotoxicity evaluation of B. longum RAPO or its CS on RAW 264.7 macrophages (A) and NHDFs (B) using the MTT assay. n=4–8/group for RAW cells. n=2–4/group for NHDFs. Median and interquartile range (1st –3rd quartile). #p<0.05 vs. group not exposed to LPS, HK RAPO, and its CS, *p<0.05. (C) Quantification of SCFAs in B. longum RAPO CS. Supplementary Figure 8 Modulation of fibroblast activation by B. longum RAPO through macrophage interactions in vitro. THP-1 cells were differentiated into macrophages using phorbol 12-myristate 13-acetate treatment and then treated with or without HK-B. longum RAPO (HK RAPO). The macrophages were subsequently co-cultured with NHDFs under LPS stimulation. Alpha-SMA protein expression was analyzed in the NHDFs. References Associated Data Supplementary Materials Supplementary Data 1 Methods Supplementary Table 1 Primer and probe sequences for digital PCR Supplementary Table 2 Cell staining Abs for flow cytometry analysis Supplementary Table 3 TaqMan assays and SBYR primer sequences for quantitative real-time PCR Supplementary Table 4 Clinical characteristics of healthy individuals and patients with SSc Supplementary Table 5 Pearson correlation coefficients between Ab levels against bacterial species and clinical parameters in patients with SSc Supplementary Figure 1 LEfSe analysis of serum anti-microbial Ab profile comparing patients with SSc (n=76) to HC (n=50). The histogram illustrates LDA scores of species enriched in HC and SSc in green and red, respectively (p<0.01, LDA score >2.0). Supplementary Figure 2 Effects of Bifidobacterium strains on dermal fibrosis in BLM-induced SSc mice. Mice received subcutaneous injections of BLM to the dorsal skin, administered 5 times a week for 2 wk. Various doses of B. longum BORI and B. bifidum BGN4 were orally administered 2 wk before BLM injection and subsequently continued for 4 wk. Dermal thickness was evaluated using Masson’s trichrome staining of skin tissues obtained from mice in each experimental group: normal mice (n=6), non-treated SSc mice (n=13), SSc mice treated with B. longum BORI or B. bifidum BGN4 at doses of 1, 2, or 5×109 CFUs (n=4–7/each treated). Median ± min to max. Supplementary Figure 3 Flow cytometric gating strategy for macrophage (MP) subsets in bronchoalveolar lavage (BAL) cells. CD45+CD64+ MPs were subdivided into alveolar macrophages (AM, CD11c+CD11b−) and exudate macrophages (exMac, CD11c+CD11b+). Supplementary Figure 4 Expression levels of genes associated with T cell cytokines in skin, lung, MLN, and spleen cells collected from normal (n=5), BLM-treated (n=5), and BLM/RAPO-treated (n=6) mice. Supplementary Figure 5 Analysis of fecal samples collected from normal (n=5), BLM-induced SSc (n=5), and BLM/RAPO-treated (n=6) mice using 16S rRNA gene sequencing. The relative abundance of gut microbiota was compared between groups at the genus level and significantly different genera were visualized. Supplementary Figure 6 Expression levels of GPCR genes in skin and lung tissues from normal (n=5), BLM-treated (n=5), and BLM/RAPO-treated (n=6) mice. Supplementary Figure 7 (A, B) Cytotoxicity evaluation of B. longum RAPO or its CS on RAW 264.7 macrophages (A) and NHDFs (B) using the MTT assay. n=4–8/group for RAW cells. n=2–4/group for NHDFs. Median and interquartile range (1st –3rd quartile). #p<0.05 vs. group not exposed to LPS, HK RAPO, and its CS, *p<0.05. (C) Quantification of SCFAs in B. longum RAPO CS. Supplementary Figure 8 Modulation of fibroblast activation by B. longum RAPO through macrophage interactions in vitro. THP-1 cells were differentiated into macrophages using phorbol 12-myristate 13-acetate treatment and then treated with or without HK-B. longum RAPO (HK RAPO). The macrophages were subsequently co-cultured with NHDFs under LPS stimulation. Alpha-SMA protein expression was analyzed in the NHDFs.