Interaction of Klebsiella pneumoniae with tissue macrophages in a mouse infection model and ex-vivo pig organ perfusions: an exploratory investigation

Abstract

Background: Hypervirulent Klebsiella pneumoniae (hvKp) strains of capsule type K1 and K2 cause invasive infections associated with hepatic abscesses, which can be difficult to treat and are frequently associated with relapsing infections. Other K pneumoniae strains (non-hvKp), including lineages that have acquired carbapenem resistance, do not manifest this pathology. In this work we aimed to test the hypothesis that within-macrophage replication is a key mechanism underpinning abscess formation in hvKp infections.

Methods: In this exploratory investigation, to study the pathophysiology of abscess formation, mice were intravenously infected with 10⁶ colony forming units (CFU) of either hvKp isolates (six strains) or non-hvKp isolates (seven strains). Intracellular bacterial replication and neutrophil influx in liver and spleen was quantified by fluorescence microscopy of sliced cryopreserved organs of mice collected 30 min, 6 h, and 24 h after infection with the aim to provide data of bacterial association to Kupffer cells in the liver and to the different tissue macrophages in the spleen. Microbiological and microscopy analysis of an ex-vivo model of pig liver and spleen infection were used to confirm within-macrophage replication. Pig organs were perfused with heparinised, autologous pig's blood and injected with 6·5 × 10⁷ CFU of hvKp K2 sequence type 25 strain GMR151. Blood and tissue biopsies collected before infection and 30 min, 1 h, 2 h, 3 h, 4 h, and 5 h after infection were used to measure bacterial counts and to identify the subcellular localisation of bacteria by immunohistochemistry analysis.

Findings: We show that hvKp resisted phagocyte-mediated clearance and replicated in mouse liver macrophages to form clusters 6 h after infection, with a mean of 7·0 bacteria per Kupffer cell (SD 6·2); however, non-hvKp were efficiently cleared (mean 1·5 bacteria per cell [SD 1·1]). HvKp infection promoted neutrophil recruitment to sites of infection, which in the liver resulted in histopathological signs of abscess formation as early as 24 h post-infection. Experiments in pig organs which share a high functional and anatomical resemblance to human organs, provided strong evidence for the propensity of hvKp to replicate within the hepatic macrophages.

Interpretation: These findings show subversion of innate immune processes in the liver by K pneumoniae and resistance to Kupffer cell mediated clearance as an explanation for the propensity of hvKp strains to cause hepatic abscesses.

Funding: University of Oxford and a Royal Society Wolfson grant funded biosafety facility.

Figure 1

Tissue tropism and intracellular localisation of Klebsiella pneumoniae in mouse livers 6 h after intravenous infection (A) Quantitative distribution of K pneumoniae in F4/80⁺ and F4/80⁻ cells in the mouse liver. Data are representative of two entire tissue sections of approximately 2 × 2 cm per mouse. (B) Size of bacterial clusters associated with Kupffer cells at 6 h after infection. 30 random fields of view were analysed with a 60 × magnification with the Olympus confocal microscope. Circles represent individual bacterial clusters. (C) 3D reconstruction using Imaris 3D V9.4 of a K1 foci of infection Z-stack acquired on the confocal microscope showing the top of the cell layer, and a single cut into the cell. DAPI stain indicated in blue, F4/80 in red, and K1 ST23 in green. (D) K1 and K2 co-infected tissue section stained with purified IgG raised against K1 (green) capsule, or K2 (red) indicating two monochrome bacterial foci. A higher magnification of each monochrome focus is shown in insets. Lower magnification co-localisation images for all serotypes, in addition to confocal Z-stacks are shown in the appendix (pp 7–8). DAPI=4′,6-diamidino-2-phenylindole. hvKp=hypervirulent K pneumoniae.

Figure 2

Tissue tropism and intracellular localisation of Klebsiella pneumoniae in mouse spleens following intravenous infection (A) Quantitative distribution of K pneumoniae between CD169⁺, F4/80⁺, MARCO⁺ macrophages, and the white pulp in the mouse spleen. Data are representative of two entire tissue sections of approximately 1 × 2 cm per mouse. (B) Size of bacterial clusters associated with host cells at 6 h after infection. Statistical significance was determined by one-way ANOVA. 30 random fields of view were analysed with 60 × magnification from the Olympus confocal microscope. Symbols represent individual bacterial clusters. (C) 3D reconstruction of a K1 focus of infection Z-stack acquired by confocal microscopy showing the top of the cell layer, and a single cut into the cell. DAPI stain indicated in blue, CD169 in red, and K1 ST23 in green. (D) K1 and K2 co-infected tissue section stained with purified IgG raised against K1 (green) capsule, or K2 (red). Lower magnification co-localisation images for all serotypes, in addition to confocal Z-stacks are shown in the appendix (pp 10–11). DAPI=4′,6-diamidino-2-phenylindole. hvKp=hypervirulent K pneumoniae.

Figure 3

Time course analysis of K1 hypervirulent Klebsiella pneumoniae and K17 non-hypervirulent K pneumoniae growth and tissue distribution. (A) Bacterial counts are shown at 30 min, 6 h, and 24 h post-infection. The limit of detection is shown as a dotted black line. The median values are indicated by horizontal bars. Spleen counts are shown in the appendix (pp 12–13). (B) Size of host cell-associated bacterial clusters, of K1 and K17 K pneumoniae throughout the infection in the liver as determined from confocal microscopy. Data are representative of three mice per time point, per organ. Error bars indicate the standard error of the mean. (C) Representative images of the liver after infection with K1 and K17 K pneumoniae are shown for the 30 min and 24 h timepoints after infection. Nuclei are stained with DAPI in blue, K pneumoniae are green, and actin is stained with phalloidin in purple. CFU=colony forming unit. DAPI=4′,6-diamidino-2-phenylindole.

Figure 4

In-vitro characterisation of Klebsiella pneumoniae growth within macrophages (A) Gentamicin protection assay following infection of J774A.1 macrophages with K1 K pneumoniae strains NTUH-K2044 and SGH10 (red), K17 strain KPC157 (green) and K107 strain KPC58 (purple) at a multiplicity of infection of 10. (B) 3D reconstruction of confocal microscopy Z-stack of 4 h infected J774A.1 macrophages with K pneumoniae SGH10. Phalloidin are shown in red, K pneumoniae in green, and nuclei in blue. A single slice into the cell is shown indicating the bacteria's intracellular localisation. (C) Quantification of the number of bacteria associated with single cells infected as in part D in at least 15 fields of a 60 × magnification confocal microscope image. (D) Gentamicin protection assay done on liver cells taken ex vivo from mice intravenously infected with K pneumoniae SGH10. Error bars show the standard error of the mean. The dotted line indicates the limit of detection. CFU=colony forming unit. HvKp=hypervirulent Klebsiella pneumoniae.

Figure 5

Neutrophil recruitment to bacterial foci in the liver leads to the formation of microabsesses (A) Quantification of neutrophil recruitment to infectious foci. Immunostained liver sections from 3 K1 Klebsiella pneumoniae-infected mice were quantified for neutrophil-associated fluorescence (relative fluorescence units) in a 25 μm radius around bacteria at 30 min, 6 h, and 24 h post-infection. Error bars indicate the standard deviation. (B) Representative image showing neutrophil recruitment to K1 foci at 30 min, 6 h, and 24 h after infection. Nuclei are stained with DAPI in blue, K pneumoniae are green, and neutrophils are pink. (C) Representative haematoxylin and eosin stain of mouse liver 48 h after infection with K1 K pneumoniae. (D) Neutrophil bactericidal assay of hypervirulent K pneumoniae (K1 and K2) and non-hypervirulent K pneumoniae (K107 and K17) showing bacterial counts before (filled bars) and after (open bars) a 2-h incubation with the neutrophils at a multiplicity of infection of 10. Error bars indicate the standard deviation. CFU=colony forming unit. DAPI=4′,6-diamidino-2-phenylindole.

Figure 6

Ex-vivo infection of three independent perfused pig livers and spleens (A) Bacterial concentrations in pig blood, liver and spleen biopsies after infection of the perfusion circuit with 6·5 × 10⁷ CFUs of Klebsiella pneumoniae GMR151 K2-ST25. Data are representative of three independent perfusion experiments with individual data in the appendix (pp 18–19). Dotted line is the limit of detection. Statistical significance was determined using a t test. (B) Representative images of foci associated with CD169⁺ cells in the pig liver at 30 min and 5 h after infection. Nuclei are stained with DAPI (blue), CD169 cells are shown in red, K pneumoniae are shown in green, and arrows indicate bacteria. (C) Size of infectious foci (number of bacteria per macrophage) at 30 min and 5 h after infection in the pig liver (circles) and spleen (squares). Data were counted from confocal images of at least 20 macrophages per group. (D) Boxplots showing neutrophil signal area within a 50 μm radius of infected (red) or random non-infected (blue) macrophages, at 30 min and 5 h post-infection of the pig liver. Data are representative of entire tissue sections (approximately 2 cm²) from three replicate organs and were imaged at a consistent resolution. CFU=colony forming unit. DAPI=4′,6-diamidino-2-phenylindole.