Colonization of dermal arterioles by Neisseria meningitidis provides a safe haven from neutrophils

Abstract

The human pathogen Neisseria meningitidis can cause meningitis and fatal systemic disease. The bacteria colonize blood vessels and rapidly cause vascular damage, despite a neutrophil-rich inflammatory infiltrate. Here, we use a humanized mouse model to show that vascular colonization leads to the recruitment of neutrophils, which partially reduce bacterial burden and vascular damage. This partial effect is due to the ability of bacteria to colonize capillaries, venules and arterioles, as observed in human samples. In venules, potent neutrophil recruitment allows efficient bacterial phagocytosis. In contrast, in infected capillaries and arterioles, adhesion molecules such as E-Selectin are not expressed on the endothelium, and intravascular neutrophil recruitment is minimal. Our results indicate that the colonization of capillaries and arterioles by N. meningitidis creates an intravascular niche that precludes the action of neutrophils, resulting in immune escape and progression of the infection.

Fig. 1. Recruitment of neutrophils to sites of infection.

a Representative time sequence (maximum intensity z-projection) of vascular colonization by iRFP-expressing Neisseria meningitidis (green). The human and murine endothelia were distinguished using UEA-1 lectin (dashed lines) and mouse-specific anti-CD31 (red) staining, respectively. Time, hh:min. Pictures shown are representatives of N = 6 mice. Scale bar, 50 µm. b Representative images (maximum intensity z-projection) of non-infected (NI) and infected human graft tissue sections at indicated times post infection (p.i.) with GFP-expressing Neisseria meningitidis (green) and stained for the human endothelium (UEA-1 lectin, gray) and neutrophils (GR-1, magenta). Pictures shown are representative of N = 3 mice per condition. Scale bar, 10 µm. c Quantification of neutrophil numbers in human xenografts of non-infected (−, gray circles) and infected (+, purple circles) mice at the indicated times post infection (p.i.). Neutrophil numbers in contralateral mouse skin at 16 h post infection is shown on the right (white circles). Two-tailed Mann–Whitney test (3 h, 6 h, 24 h) and two-tailed Kruskal–Wallis test with Dunn’s correction (16 h). n = 7 (16 h p.i.), 8 (3 h, 16 h non-infected; 6 h p.i.), 9 (3 h p.i., 6 h non-infected), 10 (mouse skin, 16 h p.i.), 18 (24 h non-infected), and 21 (24 h p.i.) mice, pooled from N = 3 (3 h, 6 h, 16 h p.i.) and 5 (24 h p.i.) independent experiments. d Movies obtained from intravital imaging were used to quantify the numbers of neutrophils per square millimeter of endothelium during the first 6 h of the infection. Quantifications were performed on n = 56 vessels, pooled from N = 7 infected mice imaged independently. c, d Data are shown as the mean ± SEM. e Intravital imaging (maximum intensity z-projection) of neutrophil (LysM^GFP/+, magenta) recruitment at 16 h post infection with iRFP-expressing Neisseria meningitidis (green) in grafted Rag₂^−/−γ_c^−/−LysM^GFP/+ mice. Dashed lines delineate the human endothelium, stained using the UEA-1 lectin. The left panel shows an important swarm of neutrophils around the infected vessel. Time, hh:min. Scale bars, 50 µm. The right panel shows the directed migration of neutrophils towards the infected vessel as revealed by cell tracking over 30 min. Pictures shown are representative of N = 3 mice. Scale bar, 50 µm.

Fig. 2. Neutrophil recruitment to infected vessels relies on type IV pili-mediated bacterial adhesion and aggregation.

a Schematic representation of the bacterial strains used to assess the impact of bacterial adhesion on neutrophil recruitment. b Bacterial colony-forming unit (CFU) counts from blood of mice infected for 5 min with the indicated bacterial strains. Two-tailed Kruskal–Wallis test with Dunn’s correction. n = 13, 6, 3 mice for WT, pilC1, pilD, respectively, pooled from N = 3 independent experiments (except for pilD N = 1). c Bacterial CFU counts from dissociated xenografts collected from mice infected for 24 h with the indicated bacterial strains. Two-tailed Kruskal–Wallis test with Dunn’s correction. n = 12, 6, 3 mice for WT, pilC1, pilD, respectively, pooled from N = 3 independent experiments (except for pilD N = 1). d Quantification of neutrophil numbers in human xenografts harvested from non-infected (NI) mice and mice infected for 24 h with the indicated bacterial strains. Two-tailed Kruskal–Wallis test with Dunn’s correction. n = 7, 9, 9, 3 mice for non-infected, WT, pilC1, pilD, respectively, pooled from N = 3 independent experiments (except for pilD N = 1). b–d Data are shown as the mean ± SEM. e Schematic representation of the bacterial strains used to assess the impact of bacterial aggregation on neutrophil recruitment. f Bacterial CFU counts from blood of mice infected for the indicated times with SB (gray circles) or SA (green circles) pilin variant-expressing bacteria. Two-tailed Mann–Whitney test per time point. n = 8 mice per bacterial strain (5 min and 24 h p.i.) and 7 mice per bacterial strain (6 h p.i.), pooled from N = 3 independent experiments. g Bacterial CFU counts from dissociated xenografts collected from mice infected for 24 h with SB or SA pilin variant-expressing bacteria. Two-tailed Mann–Whitney test. n = 8 (SB) and 9 (SA) mice, pooled from N = 3 independent experiments. h Quantification of neutrophil numbers in xenografts harvested from mice infected for 24 h with SB or SA pilin variant-expressing bacteria. Two-tailed Mann–Whitney test. n = 4 (SB) and 5 mice (SA), pooled from N = 2 independent experiments. f–h Data are shown as the mean ± SEM.

Fig. 3. Neutrophils limit vascular colonization by meningococci and vessel damages.

a Schematic of the approach used to deplete neutrophils using the neutrophil-depleting antibody anti-GR-1, clone RB6-8C5, and infect mice. b Numbers of blood circulating neutrophils in non-infected mice pre-treated with the isotype control (−, gray circles) or the neutrophil-depleting (+, purple circles) antibody. Two-tailed Mann–Whitney test per time point. n = 14 (control) and 11 (depletion) mice per time point, pooled from N = 4 independent experiments. c Bacterial CFU counts from blood of mice pre-treated with the isotype control (−, gray circles) or the neutrophil-depleting (+, green circles) antibody and infected for the indicated time points. Two-tailed Mann–Whitney test per time point. n = 12 (control, 5 min and 24 h p.i.), 14 and 11 (depletion, 5 min and 24 h p.i., respectively) mice, pooled from N = 4 independent experiments. d Neutrophil numbers in the xenografts of mice pre-treated with the isotype control (−, gray circles) or the neutrophil-depleting (+, purple circles) antibody and infected for 24 h. Two-tailed Mann–Whitney test per time point. n = 14 (control) and 12 (depletion) mice, pooled from N = 4 independent experiments. e Bacterial CFU counts from dissociated xenografts collected from mice pre-treated with the isotype control (−, gray circles) or the neutrophil-depleting (+, green circles) antibody and infected for 24 h. Two-tailed Mann–Whitney test. n = 7 (control) and 10 (depletion) mice, pooled from N = 3 independent experiments. f Representative images of hematoxylin and eosin-stained histological sections of infected xenografts and quantification of vascular damage in infected vessels. Quantifications were performed on n = 500 vessels, pooled from N = 5 (non-infected), 4 (3 h p.i.) and 3 (6 h, 16 h, and 24 h p.i.) mice. Scale bar, 10 µm. g Quantification of vascular damage upon mouse infection and neutrophil depletion. Two-tailed one-way ANOVA with Holm–Sidak’s correction. Quantifications were performed on n = 200 vessels, pooled from N = 3 mice per group. h Quantification of vascular permeability upon mouse infection and neutrophil depletion with the anti-Ly-6G antibody (clone 1A8) using Evans blue. Two-tailed one-way ANOVA with Holm–Sidak’s correction. n = 6 (non-infected and Infected with neutrophil depletion) and 7 (infected without neutrophil depletion) mice, pooled from N = 4 independent experiments. b–h Data are shown as the mean ± SEM.

Fig. 4. Different types of vessels are infected with different levels of neutrophil recruitment.

a Bacterial CFU counts from dissociated xenografts collected from intravenously (gray circles) or intradermally (green circles) infected mice for 3 h. Two-tailed Mann–Whitney test. n = 7 mice per condition, pooled from N = 3 independent experiments. b Representative images of Haematoxylin and Eosin stain of xenografts harvested from mice 3 h following intravenous or intradermal injection of PBS (control) or Neisseria meningitidis (3 h p.i.). Pictures shown are representative of N = 3 mice per condition. Scale bar, 50 μm. The dashed lines delineate the epidermis/dermis border. Arrowheads indicate blood vessels. V vessel. c Neutrophil numbers in xenografts of control (gray circles) and infected (3 h p.i., purple circles) mice by intravenous or intradermal injection. Two-tailed Kruskal–Wallis test with Dunn’s correction. n = 6 mice (i.v. PBS and i.d. PBS), 7 mice (i.v. Infected), and  8 mice (i.d. infected), pooled from N = 3 independent experiments. d Representative images (maximum intensity z-projection) of vascular colonization 2h30 post infection of the different vascular beds: capillaries, venules, and arterioles . The pie chart highlights the proportion and number of each vessel type. Scale bar, 20 µm. e Quantification of capillary, venule, and arteriole diameter. Two-tailed Kruskal–Wallis test with Dunn’s correction. f Immunohistochemistry analysis of a human case of meningococcal septic shock showing vascular colonization by N. meningitidis in both venules and arterioles in the liver (left) and the choroid plexus (right). Bacteria were labeled using a polyclonal antibody directed against the strain cultured from the blood of this patient, as previously described. Pictures shown are representative of one single human donor. g Percentage of vessels recruiting at least one neutrophil following infection according to the different vascular beds: capillaries (gray circles), venules (light-blue circles), and arterioles (light-red circles). h Numbers of neutrophils per square millimeter of endothelium according to the different vascular beds during the first 6 h of the infection. a, c, e, h Data are shown as the mean ± SEM. d, e, g, h Quantifications were performed on n = 56 vessels (21 capillaries, 11 venules, and 24 arterioles), pooled from N = 7 infected mice imaged independently.

Fig. 5. Vascular colonization of venules leads to efficient neutrophil recruitment and function.

a Intravital imaging (maximum intensity z-projection) of neutrophil (Ly-6G, magenta) recruitment to infected venules with iRFP-expressing Neisseria meningitidis (green). The human vessels are shown in gray (UEA-1 lectin) and dashed lines. Scale bars, 30 µm. b Percentage of venules recruiting only intraluminal (white circles), only perivascular (red circles), or both intraluminal and perivascular (gray circles) neutrophils. c Numbers of intraluminal and perivascular neutrophils per square millimeter of venular endothelium during the first 6 h of the infection. Data are shown as the mean ± SEM. b, c Quantifications were performed on n = 11 vessels, in total, pooled from N = 7 infected mice imaged independently. d 3D-rendering of intravital imaging showing neutrophil (Ly-6G, magenta) migration toward and engulfing an adherent bacterial aggregate within a venule 3 h post infection with iRFP-expressing Neisseria meningitidis (green). The human venule is shown in gray (UEA-1 lectin). Scale bar, 10 µm.

Fig. 6. Vascular colonization of arterioles leads to limited neutrophil recruitment and efficiency.

a Intravital imaging (maximum intensity z-projection) of intraluminal (top panel) or perivascular (bottom panel) recruitment of neutrophils (Ly-6G, magenta) to arterioles 3–5 h post infection with iRFP-expressing Neisseria meningitidis (green). The human vessels are shown in gray (UEA-1 lectin) and dashed lines. Time, hh:min. Scale bar, 30 µm. b Percentage of arterioles recruiting only intraluminal (white circles), only perivascular (red circles), or both intraluminal and perivascular (gray circles) neutrophils. c Numbers of intraluminal and perivascular neutrophils per square millimeter of arteriolar endothelium during the first 6 h of the infection. Data are shown as the mean ± SEM. b, c Quantifications were performed on n = 24 vessels, in total, pooled from N = 7 infected mice imaged independently. d Intravital imaging (maximum intensity z-projection) of neutrophil (Ly-6G, magenta) entrapped within bacterial aggregates within an infected arteriole 7 h post infection with iRFP-expressing Neisseria meningitidis (green). The human vessels are shown in gray (UEA-1 lectin) and dashed lines. Scale bar, 30 µm.

Fig. 7. E-selectin endothelium surface expression is differentially upregulated according to the vascular bed upon infection.

a Flow cytometry analysis of cell surface expression of ICAM-1, VCAM-1, and E-selectin (CD62E) on HUVEC cells under resting conditions (Basal, open histograms) or following infection with Neisseria meningitidis for 2 h (Nm_2h, red histograms) or 5 h (Nm_5h, dark-red histograms) or an overnight incubation with 20 ng ml⁻¹ TNFα gray histograms). Data are representatives of N = 3 independent experiments. The percentages of positive cells (above the dashed lines) are shown per condition and marker. b In vivo expression of E-selectin at the surface of different human vessel types (capillary, arteriole, and venule) following two (2 h p.i.) and four (4 h p.i.) hours of infection. Images (maximum intensity z-projection) are representative of N = 3 infected mice imaged independently. GFP-expressing Neisseria meningitidis appears in green and E-selectin in red following in vivo labeling by i.v. injection of PE-labeled anti-CD62E monoclonal antibody. Dashed lines delineate human vessels (UEA-1 lectin). Scale bar, 20 µm. c High-magnification view (yellow dashed square in panel b) of a bacterial microcolony at 4 h p.i. on the venular endothelium surface and the local upregulation of E-selectin expression. Scale bar, 5 µm. d Movies obtained from intravital imaging were used to quantify the numbers of neutrophils per square millimeter of venular endothelium during the first 4 h of the infection in presence of anti-E-selectin blocking antibody (red circles) or isotype control (white circles). Data are shown as the mean ± SEM. Quantifications were performed on n = 22 and 8 venules for E-selectin blocking and isotype control conditions, respectively, pooled from N = 3 infected mice imaged independently per condition. Two-tailed two-way ANOVA. e Representative image (maximum intensity z-projection) of venular colonization 4 h post infection by Neisseria meningitidis (green) in presence of the anti-E-selectin blocking antibody. The human endothelium was labeled using UEA-1 lectin (gray). The picture shown is representative of N = 3 mice imaged independently. Scale bar, 50 µm.