Rhinoscleroma pathogenesis: The type K3 capsule of Klebsiella rhinoscleromatis is a virulence factor not involved in Mikulicz cells formation

Abstract

Rhinoscleroma is a human specific chronic granulomatous infection of the nose and upper airways caused by the Gram-negative bacterium Klebsiella pneumoniae subsp. rhinoscleromatis. Although considered a rare disease, it is endemic in low-income countries where hygienic conditions are poor. A hallmark of this pathology is the appearance of atypical foamy monocytes called Mikulicz cells. However, the pathogenesis of rhinoscleroma remains poorly investigated. Capsule polysaccharide (CPS) is a prominent virulence factor in bacteria. All K. rhinoscleromatis strains are of K3 serotype, suggesting that CPS can be an important driver of rhinoscleroma disease. In this study, we describe the creation of the first mutant of K. rhinoscleromatis, inactivated in its capsule export machinery. Using a murine model recapitulating the formation of Mikulicz cells in lungs, we observed that a K. rhinoscleromatis CPS mutant (KR cps-) is strongly attenuated and that mice infected with a high dose of KR cps- are still able to induce Mikulicz cells formation, unlike a K. pneumoniae capsule mutant, and to partially recapitulate the characteristic strong production of IL-10. Altogether, the results of this study show that CPS is a virulence factor of K. rhinoscleromatis not involved in the specific appearance of Mikulicz cells.

Fig 1. Characterization of the K. rhinoscleromatis capsule mutant strain (KR cps^-).

(A) Schematic representation of the wild-type K. rhinoscleromatis (top) and KR cps^- (bottom) capsule export region of the capsule locus. The insertion of the suicide plasmid pDS132 occurred at the level of the wzb gene. The asterisk indicates the base deletion in the sacB gene leading to aberrant protein production. (B) Morphology of the colonies of KR WT and KR cps^- strains after overnight culture on LB agar plates. (C) Quantification of capsule expressed as amount of uronic acid / 10⁶ bacteria in KR WT and KR cps^-. (D) Mice survival after pulmonary infection with KR WT or KR cps^- strains. BALB/c mice were infected with 2.10⁷ KR WT, 2.10⁷ KR cps^-, 4.10⁸ KR cps^- or 10⁹ KR cps^-. Survival was followed overtime. Data are representative of 10 mice per group from two independent experiments.

Fig 2. Bioluminescence imaging and bacterial loads quantification in mice infected by wild-type K. rhinoscleromatis and KR cps^-.

(A) BALB/c mice were infected with bioluminescent 2.10⁷ KR WT, 2.10⁷ KR cps^-, 4.10⁸ KR cps^-, and 10⁹ KR cps^-. The bioluminescent signal in lungs was measured 6, 24, 48, 72, and 96 hours post-infection using the IVIS Imaging System. All images are shown using the same bioluminescence signal intensity scale (in photons/sec/cm²/sr). (B) Quantification of bioluminescent signal detected from six mice per group. Means are indicated as line. The dotted line indicates background level, and the dashed line shows the minimal signal shown in (A). (C) Bacterial load in lungs of mice infected with 2.10⁷ KR WT (left), 2.10⁷ KR cps^- (centre) or 4.10⁸ KR cps^- (right). Data for bacterial loads are shown as log CFU per organ from 6 to 12 mice from two to five independent experiments. Means are indicated as line.

Fig 3. Histology of BALB/c lungs infected by wild-type K. rhinoscleromatis and KR cps^-.

(A) Lungs of mice infected with 2.10⁷ KR WT, 2.10⁷ KR cps^-, 4.10⁸ KR cps^-, and 10⁹ KR cps^- were resected 4 days post-infection and examined by histology. Lungs infected with 2.10⁷ KR WT (left) presented the classical pattern characterized by many alveoli, with an intact epithelial layer, filled with Mikulicz cells. On the contrary, lungs infected with 2.10⁷ KR cps^- (middle left) showed dense inflammatory infiltrate containing numerous polymorphonuclear cells and absence of Mikulicz cells. Lungs of mice infected with 4.10⁸ KR cps^- (middle right) or 10⁹ KR (right) showed the presence of Mikulicz cells similarly to wild-type infection. Insets show magnification of representative zones of the lungs. Scale bars are 1 mm (top), 200 μm (middle row) and 50 μm (bottom). Images are representative of 4 (KR WT), 4 (2.10⁷ KR cps^-), 4 (4.10⁸ KR cps^-) or 5 (10⁹ KR cps^-) mice from 2 to 3 independent experiments. (B) Correlation between the area covered by Mikulicz cells and number of bacteria spots in lungs sections of mice at 96 hours after infection with 2.10⁷ KR WT, 2.10⁷ KR cps^-, 4.10⁸ KR cps^-, 10⁹ KR cps^-.

Fig 4. Production of IL-1β and IL-10 in lungs of BALB/c mice infected by wild-type K. rhinoscleromatis, KR cps^-, Kp52Δwzc or Kp52145.

BALB/c mice were infected with 2.10⁷ KR WT, 2.10⁷ KR cps^- or 4.10⁸ KR cps^- or saline-injected for 6, 24, 48, 72 or 96 hours. Lungs were then homogenized and the pro-inflammatory IL-1β (A) and the anti-inflammatory IL-10 (B) cytokines were measured by ELISA. Data are mean from 6 to 10 mice from three independent experiments. Correlation between production of IL-1β (C) or IL-10 (D) and the amount of bacteria recovered from lungs of mice at 96 hours after injection with saline (control) or infection with 2.10⁷ KR WT, 2.10⁷ KR cps^-, 4.10⁸ KR cps^-, 10⁹ KR cps^-, 10⁹ Kp52Δwzc or 2.10⁴ Kp52145.

Fig 5. The capsule mutant Kp52Δwzc is not able to induce the formation of Mikulicz cells.

(A) Bacterial load in lungs of BALB/c mice after infection with 2.10⁷ KR WT, 10⁹ KR cps^- or 10⁹ Kp52Δwzc. Data show CFU in whole lungs after 96 hours post-infection from 5 to 11 mice from one to five independent experiments. Means are indicated as line. (B) Production of IL-1β and IL-10 in the lungs of BALB/c mice infected with 2.10⁷ KR WT, 10⁹ KR cps^- or 10⁹ Kp52Δwzc. Cytokines were measured 96 hours post-infection by ELISA. Data are mean from 5 to 9 mice from two independent experiments. (C) Histology, representative example of lung from mice infected with 10⁹ Kp52Δwzc. Zones of dense inflammation can be observed with absence of Mikulicz cells. Scale bars are 1 mm (left), 200 μm (middle) and 50 μm (right).