Critical role of bacterial dissemination in an infant rabbit model of bacillary dysentery

Abstract

The bacterial pathogen Shigella flexneri causes 270 million cases of bacillary dysentery (blood in stool) worldwide every year, resulting in more than 200,000 deaths. A major challenge in combating bacillary dysentery is the lack of a small-animal model that recapitulates the symptoms observed in infected individuals, including bloody diarrhea. Here, we show that similar to humans, infant rabbits infected with S. flexneri experience severe inflammation, massive ulceration of the colonic mucosa, and bloody diarrhea. T3SS-dependent invasion of epithelial cells is necessary and sufficient for mediating immune cell infiltration and vascular lesions. However, massive ulceration of the colonic mucosa, bloody diarrhea, and dramatic weight loss are strictly contingent on the ability of the bacteria to spread from cell to cell. The infant rabbit model features bacterial dissemination as a critical determinant of S. flexneri pathogenesis and provides a unique small-animal model for research and development of therapeutic interventions.

Fig. 1

Symptoms and histopathology in infant rabbits infected with Shigella flexneri. a, b Representative images of animals (top) and colon (bottom) inoculated with phosphate-buffered saline (PBS) (a) or S. flexneri (b). Scale bar, 2 cm. c, d Representative images of hematoxylin- and eosin-stained colonic sections from animals inoculated with PBS (c) or S. flexneri (d). Dotted lines delineate the colon. Double arrows indicate the width of the colon section with massive swelling in d. Brackets indicate the width of connective tissues with edema (E). VL indicates vascular lesions with red blood cells. Insets: high magnification of epithelial cells (EC) with fenestration in d. Scale bars, 100 μm. e Histopathology scores of VLs, edema (E), and percentage of epithelial fenestration (EF). Statistical analysis, unpaired t test. Shigella flexneri (wild type (WT)) vs. mock (PBS): VL, ****P < 0.0001; E, ****P < 0.0001; EF, ****P < 0.0001. Source data are provided as a Source Data file

Fig. 2

Pro-inflammatory immune responses in the colon of infant rabbits infected with Shigella flexneri. a, b Representative images of hematoxylin- and eosin-stained colonic sections from animals inoculated with phosphate-buffered saline (PBS) (a) or S. flexneri (b). Scale bars, 100 μm. Top, ×10 images; bottom, ×40 images of the boxed area in top images. Arrows and arrowheads indicate heterophils. Insets, zoom-in on the heterophil indicated by arrowheads. c Graph showing histopathology score of heterophil infiltration. Statistical analysis, unpaired Student’s t test. Wild-type (WT) vs. mock, ****P < 0.0001. d Chemokine (interleukin-8 (IL-8) and C-X-C motif chemokine 10 (CXCL10)) and cytokine (IL-1β, IL-6, and tumor necrosis factor-α (TNFα)) gene induction (black bars) in the colon 8 h post inoculation (pi) with S. flexneri with respect to mock-treated animals (PBS, white bars). Overlay dot plots represent individual data points. Error bars represent standard deviation of the mean. Source data are provided as a Source Data file

Fig. 3

Shigella flexneri invades epithelia cells in a T3SS-dependent manner. a, b Representative images of colonic sections immuno-stained for E-cadherin and S. flexneri. a Merge: E-cadherin, green; S. flexneri, red. b S. flexneri only. Arrows indicate intracellular S. flexneri. Scale bars, 50 μm. c Zoom-in on the boxed area in a. d Graph showing counts of colony-forming units in the distal colon of animals infected with S. flexneri (wild-type (WT)) or the T3SS mutant (ΔmxiG). Statistical analysis, unpaired t test. ΔmxiG vs. WT, ***P < 0.0005. Source data are provided as a Source Data file

Fig. 4

Symptoms and histopathology in absence of intracellular invasion. a Representative images of animals (top) and colon (bottom) inoculated with the ΔmixG mutant. Scale bar, 2 cm. b Representative image of hematoxylin- and eosin-stained colonic sections from animals inoculated with the ΔT3SS mutant. Dotted lines delineate the colon. Inset: high-magnification of epithelial cells (ECs). Scale bars, 100 μm. c Percentage of epithelial fenestration (EF), and histopthology scores of edema (E), vascular lesions (VLs), and heterophil infiltration (HI). Statistical analysis, unpaired t test. ΔmxiG vs. mock (phosphate-buffered saline (PBS)): EF, *P < 0.05; E, *P < 0.05; VL, ns; HI, ns (not significant). Source data are provided as a Source Data file

Fig. 5

Shigella flexneri spreads from cell to cell in an IcsA-dependent manner. a, b Representative images of colon sections infected with S. flexneri (a) and the ΔicsA mutant (b). Scale bars, 50 μm. E-cadherin, green; S. flexneri, red. c, d Zoom-in on the boxed area in a, b, respectively. e Zoom-in on the area indicated by the arrowhead in a. f Graph showing counts of colony-forming units in the distal colon of animals infected with S. flexneri (wild-type (WT)) or the ΔicsA mutant. Statistical analysis: unpaired t test. ΔicsA vs. S. flexneri (WT), *P < 0.05. Source data are provided as a Source Data file

Fig. 6

Symptoms and histopathology in absence of cell-to-cell spread. a Representative images of animals (top) and colon (bottom) inoculated with the ΔicsA mutant. Arrows indicate tarnished fur. Scale bar, 2 cm. b Representative image of hematoxylin- and eosin-stained colonic sections from animals inoculated with the ΔicsA mutant. Dotted lines delineate the colon. E, edema; VLs, vascular lesions. Insets: (top) high magnification of red blood cells (RBCs) indicative of VLs and (bottom) high magnification of epithelial cells (ECs). Scale bar, 100 μm. c Histopathology scores of VLs) and heterophil infiltration (HI). Statistical analysis: unpaired Student’s t test. ΔicsA vs. mock (phosphate-buffered saline (PBS): VLs, **P < 0.01; ***HI, P < 0.05. d Graphs showing the percentage of epithelial fenestration (EF) and weight loss (WL). Statistical analysis, unpaired t test. ΔicsA vs. S. flexneri (wild type (WT)): EF, ****P < 0.0001; WL, ****P < 0.0001. Source data are provided as a Source Data file

Fig. 7

Model of bacillary dysentery. a In the absence of intracellular invasion (i.e. ΔT3SS mutant) of epithelial cells (green), S. flexneri (red) is present in the lumen, and red blood cells (tangerine) are confined to blood vessels (black line). Infected animals do not display any symptoms (no bloody diarrhea). b In the absence of cell-to-cell spread (i.e., ΔicsA mutant), S. flexneri grows as micro-colony in primarily infected cells, but does not get access to adjacent cells. Epithelial cell production of signaling molecules (not shown) and vascular lesions (doted black line) lead to the infiltration of immune cells (neutrophils, purple) and red blood cells, respectively. In the absence of cell-to-cell spread, infected epithelial cells may undergo cell death (pale green) and be extruded from the epithelium. Fenestration of the epithelial layer is minimal and animals do not display any symptoms (no bloody diarrhea). c Wild-type S. flexneri invades epithelial cells and spreads from cell to cell. Similar to b, intracellular invasion leads to immune cell infiltration and vascular lesions. In addition, cell-to-cell spread leads to massive epithelial cell fenestration, which correlates with bacillary dysentery symptoms (bloody diarrhea)