Targeted delivery of the probiotic Saccharomyces boulardii to the extracellular matrix enhances gut residence time and recovery in murine colitis

Abstract

Probiotic and engineered microbe-based therapeutics are an emerging class of pharmaceutical agents. They represent a promising strategy for treating various chronic and inflammatory conditions by interacting with the host immune system and/or delivering therapeutic molecules. Here, we engineered a targeted probiotic yeast platform wherein Saccharomyces boulardii is designed to bind to abundant extracellular matrix proteins found within inflammatory lesions of the gastrointestinal tract through tunable antibody surface display. This approach enabled an additional 24-48 h of probiotic gut residence time compared to controls and 100-fold increased probiotic concentrations within the colon in preclinical models of ulcerative colitis in female mice. As a result, pharmacodynamic parameters including colon length, colonic cytokine expression profiles, and histological inflammation scores were robustly improved and restored back to healthy levels. Overall, these studies highlight the potential for targeted microbial therapeutics as a potential oral dosage form for the treatment of inflammatory bowel diseases.

Fig. 1. Genetic engineering of S. boulardii to express mSA handle on the cell surface for attachment of ECM-specific targeting ligands.

A Schematic of engineering steps to enable the binding of S.b. to extracellular matrix proteins. B Schematic of key components of plasmid used to engineer stable expression of monomeric streptavidin (mSA) on the yeast surface. C Representative fluorescent images of S.b. mSA (left) and S.b. (right) attached to a biotin-coated well plate incubated at varying initial cell concentrations (starting OD₆₀₀ = 0.5, 0.25, or 0.125). Scale bars are 100 μm. D Quantification of attached S.b. on the biotin-coated well plate. Images were quantified using ImageJ, n = 3 wells per condition. Data are shown as mean ± SD. Significance was determined using unpaired, two-tailed Student’s t-tests, n = 3 per condition. E Mean fluorescence intensity (MFI) values resulting from labeling S.b. mSA with titrating concentrations of biotinylated antibodies against fibronectin (FN, blue), fibrinogen (FB, purple), or collagen IV (CIV, red) or non-biotinylated antibodies (grey) as measured by flow cytometry (n = 3, points represent mean). F Collection of fold change values in colonic expression of fibronectin (Fn1), fibrinogen (Fgb), or collagen IV (Col4a1) in human subjects with active ulcerative colitis (n = 181) as compared to healthy subjects (n = 82) from 7 whole-genome transcriptional analysis datasets. For violin plots, the mean is represented by the solid line and interquartile ranges represented by dotted lines. Significance assessed by unpaired, two-tailed Student’s t-tests. α = 0.05, *p < 0.05, **p < 0.01, ****p < 0.0001.

Fig. 2. Antibody-labeling enables S. boulardii to bind to corresponding extracellular matrix proteins.

A Schematic of biotinylated antibodies specific to extracellular matrix proteins attached to the S. boulardii cell surface via biotin-streptavidin interactions. B Quantification of attached targeted S.b. (colored bars) or non-targeted S.b. (grey bars) on fibronectin-coated (FN blue), fibrinogen-coated (FB purple), or collagen IV-coated (CIV red) well plates at varying seeding densities. Images were quantified using ImageJ, n = 3 wells per condition. Data are shown as mean ± SD. Significance was determined using ordinary one-way ANOVA with Šídák’s multiple comparisons test, n = 3 per condition. C Representative fluorescent images of Left: S.b. FN or non-specific antibody-labeled S.b. (S.b. NS) attached to a fibronectin-coated well plate; Middle: S.b. FB or S.b. NS attached to a fibrinogen-coated well plate; Right: S.b. CIV or S.b. NS attached to a collagen IV-coated well plate. Scale bars are 100 μm. D Schematic depicting the growth of antibody-labeled S. boulardii, resulting in the dilution of the number of antibodies remaining on each cell surface, and subsequent reduction in the capacity to bind to the corresponding ECM protein. E Anti-fibronectin antibody-labeled S. boulardii growth (red squares, right y-axis), percent antibody remaining on the cell surface as determined by flow cytometry (blue circles, left y-axis), and percent mSA remaining on the cell surface as determined by flow cytometry (grey triangles, left y-axis) over time incubated in supplemented simulated intestinal fluid (n = 3 technical replicates, points represent mean). Data are shown as mean ± SD, n = 3. F Quantification of plate-bound S.b. FN as compared to non-targeted S.b. on a fibronectin-coated well plate following incubation in supplemented simulated intestinal fluid (Images were quantified using ImageJ, n = 3 wells per condition) Data are shown as mean ± SD. G Representative fluorescent images of S.b. FN bound to a fibronectin-coated well plate over time. Scale bars are 100 μm. α = 0.05, ****p < 0.0001.

Fig. 3. Engineered S. boulardii retain probiotic mechanisms of action in vitro.

A Schematic of probiotic mechanisms of action elicited by S. boulardii. B Percent viability after 4 h incubations of control and engineered yeast in media at pH 2.5, pH 4.0, or containing 0.3% or 0.6% OxGall bile salts (OxG). Data are shown as mean ± SD, n = 3 biological replicates. C Secretion of acetate, propionate, and butyrate from cultures of either S.b., S.b. mSA, S.b. FN, S.b. FB, and S.b. CIV, or blank media, n = 4. Solid line: simple regression of the linear portion before saturation (0 to 18 h for acetate, 0 to 12 h for butyrate and propionate. D Secretion rate of acetate, propionate, and butyrate from cultures of either S.b., S.b. mSA, S.b. FN, S.b. FB, and S.b. CIV, or blank media calculated from linear regression in (C), n = 4 biological replicates. E Concentration of murine interleukin-10 (IL-10) in the cell culture supernatant following an 18 h co-incubation of murine bone marrow-derived dendritic cells with controls (PBS or LPS 1 μg/mL) and engineered yeast, n = 3 biological replicates. F. IL-8 concentrations in HT-29 human epithelial cells co-cultured with probiotic yeast strains then stimulated with TNFa (20 mg/ml). Data are shown as mean ± SD, n = 3 biological replicates. Significance was determined using ordinary one-way ANOVA with Tukey’s for panels B, D, E, and F. For panel B, black asterisks indicate significance against S.c. For panels E, and F: black asterisks indicate significance against PBS and grey asterisks indicate significance against S.c., α = 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Fig. 4. Extracellular matrix targeting increases colon residence time of S. boulardii and decreases inflammation in acute DSS-induced murine colitis.

A Schematic depicting time course of in vivo murine colitis study. C57BL6/J mice were administered 2% dextran sulfate sodium (DSS) in the drinking water for 5 days, then administered regular drinking water for the remainder of the study. Mice were dosed on day 5 with 10⁹ colony-forming units (CFU) of nontargeted (S.b. mSA), fibrinogen-targeted (S.b. FB), collagen IV-targeted (S.b. CIV), or fibronectin-targeted (S.b. FN) yeast via oral gavage. Stool was collected at 12, 24, 48, 72 h post-gavage to measure viable yeast by quantitative culture. B S. boulardii CFU in the stool post-gavage, n = 5 independent animals. C S. boulardii CFU in colon tissue 72 h post-gavage, n = 5 independent animals. D Mean percent body weight of mice over the course of the study. N = 5, significance indicates comparisons to average weights of DSS-only mice at each timepoint. Arrow indicates day of DSS removal and yeast dosing. E Mouse colon length at study termination, n = 5 independent animals, lines represent the mean. Black asterisks indicate significance against DSS and grey asterisks indicate significance against S.b. mSA. F Mean relative expression of pro-inflammatory (Tnfa, Ifng, Il6) and anti-inflammatory (Tgfb, Il10) cytokines in distal colon tissue compared to healthy controls, n = 5. Statistical tests between groups for each gene were independently performed then represented in a single heatmap. Black asterisks indicate significance against DSS and grey asterisks indicate significance against S.b. mSA. Significance indicated as compared to mean relative expression values of DSS-only mice. G Representative images of hematoxylin and eosin staining of colon Swiss rolls; black arrows indicate lesions of severe inflammation/ulceration. H Semi-quantitative histological scores of inflammation accounting for extent of mucosal loss, hyperplasia, and erosions, n = 5 independent animals. Black asterisks indicate significance against DSS and grey asterisks indicate significance against S.b. mSA. I Mean relative expression of Fn1, Fgb, and Col4a1 in the distal colon of DSS mice relative to healthy controls, n = 4. No statistical significance was found. J Representative immunofluorescence of DSS-only colons dosed with either S.b. mSA or S.b. FN. Cell nuclei (blue), fibronectin (pink), and S.b. (cyan) imaged with 2x and 10x objectives, n = 3. Dotted lines represent the limit of detection (Panels B and C). Data are shown as mean ± SD. Significance was determined using ordinary one-way ANOVA with Dunnett’s (panels B, C, D) or Tukey’s (panels E, F, H) multiple comparisons test α = 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars are 100 μm.

Fig. 5. S. boulardii targeted to collagen IV exhibits enhanced gut retention time and anti-inflammatory effects in a chronic DSS model of colitis.

A Schematic depicting the time course of in vivo murine colitis study. C57BL6/J mice were administered 2% dextran sulfate sodium (DSS) in the drinking water for 5 days (injury) followed by administration of regular drinking water for 3 days (recovery). This cycle was repeated twice more. Following the first cycle, mice were administered 10⁹ colony-forming units (CFU) of nontargeted (S.b. mSA), fibronectin-targeted (S.b. FN), or collagen IV-targeted (S.b. CIV) yeast via oral gavage every 3 days for the remainder of the study. The stool was collected every 24 h post-gavage to measure viable yeast. B S. boulardii CFU in the stool on the 48-h post-dose timepoints (Days 11, 14, 17, 20, and 23), n = 25. and on the 72-h post-dose timepoints (Days 12, 15, 18, 21, and 24), n = 25, lines represent mean, (5 independent animals x 5 collection timepoints). C Area under the curve (AUC) analysis of CFU in the stool over time plots, n = 5 independent animals. D S. boulardii CFU in the colon tissue at study termination, n = 5 independent animals. E Mean percent body weight of mice over the course of the study. n = 5, significance determined by comparing the area under each curve as compared to that of DSS mice. F Mouse colon length at study termination, n = 5, lines represent the mean. Black asterisks indicate significance against DSS and grey asterisks indicate significance against S.b. mSA. G Mean relative expression of pro-inflammatory (Tnfa, Ifng, Il6) and anti-inflammatory (Tgfb, Il10) cytokines in distal colon tissue compared to healthy controls, n = 5. Black asterisks indicate significance against DSS and grey asterisks indicate significance against S.b. mSA. H Representative images of hematoxylin and eosin staining of colon Swiss rolls at study termination, black arrows indicate lesions of severe inflammation/ulceration. I Semi-quantitative histological scores of inflammation accounting for extent of mucosal loss, hyperplasia, and erosions, n = 5 independent animals. Black asterisks indicate significance against DSS and grey asterisks indicate significance against S.b. mSA. J Mean relative expression of Fn1, Fgb, and Col4a1 in distal colon of DSS mice relative to healthy controls, n = 4. Dotted lines represent the limit of detection and shaded regions below represent any values below the limit of detection (Panels B, C, and D). Data are shown as mean ± SD. Significance was determined using ordinary one-way ANOVA with Dunnett’s (panels B–E) or Tukey’s (panels F, G, I) multiple comparisons test, α = 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars are 100 μm.