In vivo adaptation and persistence of Neisseria meningitidis within the nasopharyngeal mucosa

Abstract

Neisseria meningitidis (Nme) asymptomatically colonizes the human nasopharynx, yet can initiate rapidly-progressing sepsis and meningitis in rare instances. Understanding the meningococcal lifestyle within the nasopharyngeal mucosa, a phase of infection that is prerequisite for disease, has been hampered by the lack of animal models. Herein, we compare mice expressing the four different human carcinoembryonic antigen-related cell adhesion molecules (CEACAMs) that can bind the neisserial Opa protein adhesins, and find that expression of human CEACAM1 is necessary and sufficient to establish intranasal colonization. During infection, in vivo selection for phase variants expressing CEACAM1-specific Opa proteins occurs, allowing mucosal attachment and entry into the subepithelial space. Consistent with an essential role for Opa proteins in this process, Opa-deficient meningococci were unable to colonize the CEACAM1-humanized mice. While simple Opa-mediated attachment triggered an innate response regardless of meningococcal viability within the inoculum, persistence of viable Opa-expressing bacteria within the CEACAM1-humanized mice was required for a protective memory response to be achieved. Parenteral immunization with a capsule-based conjugate vaccine led to the accumulation of protective levels of Nme-specific IgG within the nasal mucus, yet the sterilizing immunity afforded by natural colonization was instead conferred by Nme-specific IgA without detectable IgG. Considered together, this study establishes that the availability of CEACAM1 helps define the exquisite host specificity of this human-restricted pathogen, displays a striking example of in vivo selection for the expression of desirable Opa variants, and provides a novel model in which to consider meningococcal infection and immunity within the nasopharyngeal mucosa.

Figure 1. CEACAM1-humanized mice as model for N. meningitidis colonization.

(A) Staining of mouse nasal sections for human CEACAM1. Overview of nasal cavity structures in low-magnification (5×) image indicates regions of interest for higher-resolution (20×) images. Human CEACAM1 is stained brown and indicated by arrows, nuclei appear in blue. Images are representative for n = 3 animals of each genotype. (B) Expression of CEACAMs analyzed in primary human nasal epithelial cells (HNEPC) in comparison to HeLa cells expressing defined CEACAM family members. Western blot was probed with rabbit polyclonal antibody (CEA-Dako) that recognizes each of these human CEACAMs. (C) Adherence to (left panel) and invasion into (middle panel) HeLa cells that express different CEACAM family members by E. coli expressing Opa proteins cloned from Nme strain MC58 assessed by gentamycin protection assay. Presented is the mean of n = 3 independent experiments ± SEM. Confirmation of Opa expression in E. coli is shown in Fig. S1C. (D) Cohorts of wild-type mice (open circles), CEABAC mice which express human CEACAMs 3, 5, 6, 7 (black circles) or CEACAM1-humanized mice (grey circles) were i.n. infected with 10⁸ CFU Nme strain MC58 and viable meningococci recovered from nasal tissues at indicated time points for quantitative culture. Data were pooled from at least two individual infections for each group. Horizontal bars indicate median. Kruskal-Wallis test applying Dunn's multiple comparison post hoc test was used to analyze the data, with * denoting P<0.05 and ns, not significant. Note that CEACAM1-humanized cohort at day 3 consists of pooled results from four independent experiments and is used as comparator in Fig. 1E, 1E, 5A, 5B and 5C. (E) Comparison of recovered meningococci from cohorts at day 3 after i.n. infection with 10⁸ MC58 at different ages. Note that 6-week old animal cohort is identical to Fig. 1D (day3). Horizontal bars indicate median. (F) Cohorts of mice were i.n. infected with varying CFU numbers of MC58 in the inoculum as indicated and viable bacteria recovered at day 3 post infection. Data were pooled from at least two individual infections for each group. Horizontal bars indicate median. Note that cohort of CEACAM1-humanized mice infected with 10⁸ CFU of MC58 are the same as in Fig. 1D (day 3). (G) Survival curves of wild-type or CEACAM1-humanized mice after intraperitoneal challenge with 10⁶ MC58.

Figure 2. Opa/CEACAM dependency of meningococcal colonization.

(A) Cohorts of mice were i.n. infected with 10⁸ CFU of serogroup B strains MC58 (upper panels) or H44/76 (lower panels) in which Opa protein expression was turned off by phase variation. Viable bacteria were recovered and enumerated at day 3 post infection (left panels). Data were pooled from two individual infections, horizontal bars indicate median. Western blot analysis was performed to monitor Opa expression in recovered clones (right panel). (B) Cohorts of mice were i.n. infected with 10⁸ CFU of strain H44/76 or an isogenic mutant (H44/76Δopa) in which all four opa genes were disrupted. Viable bacteria were recovered and enumerated at day 3 post infection. Data were pooled from two individual infections for each group, horizontal bars indicate median.

Figure 3. Inflammatory responses in nasal tissue after i.n. infection with N. meningitidis.

Cohorts of mice were i.n. infected with 10⁸ CFU of H44/76wt (‘wt’), H44/76Δopa (‘Δopa’), or heat-inactivated H44/76 (‘h.i.’), or left untreated (‘none’) and nasal tissues were harvested 16 h post infection to quantify chemokines KC, MIP-1α, MIP-1 and cytokines TNF-α and IL-1β by ELISA, or myeloperoxidase activity. Group sizes are indicated in graphs. *, **, or *** denote P<0.05, P<0.01, or P<0.001, respectively, in one-way ANOVA applying Tukey's post-hoc comparison of all groups using GraphPad Prism 5.0 software. ns, not significant.

Figure 4. Role of polymorphonuclear cells in mucosal protection against N. meningitidis colonization.

(A) CEACAM1-humanized and wild-tpye mice were either left untreated or were PMN-depleted by anti-Gr-1 antibody clone RB6-8C5 and then left either uninfected, or were i.n. infected with 10⁸ CFU MC58 as indicated. Immunohistological staining of nasal tissues for Gr-1 was performed on tissue samples obtained at day 1 to visualize PMN infiltrates. Overview of nasal cavity structures in low-magnification (5×) image on the top indicate regions of interest for higher-resolution (40×) images below. Gr-1 is stained brown and indicated by arrows, nuclei appear in blue. Images are representative for n = 3 animals of each genotype. (B) Same samples as in (A) were used for immunofluorescence staining. Human CEACAM appears in red, Nme in green and nuclei in blue. Note, that different regions of interest were chosen as compared to (A), but the images are in the same sequence as indicated for the groups above. (C) Cohorts of PMN depleted mice were i.n. infected with 10⁸ CFU of Nme MC58 and viable Nme recovered at indicated time points. Pairwise comparisons of PMN-depleted transgenic versus wild-type mouse cohorts were performed for each time point and analyzed using Mann-Whitney test; ** denotes P<0.01. Horizontal bars indicate median. Note that in the right panel plotting the percentage of colonized mice in each group, cohorts of PMN-sufficient mice from Fig. 1D were included for comparison. (D) Cohorts of untreated or PMN depleted mice were either left uninfected or were i.n. infected with 10⁸ CFU of Nme MC58 and nasal tissues harvested at 16 h post infection to quantify chemokines KC, MIP-1α, MIP-1 and cytokines TNF-α and IL-1β by ELISA, or myeloperoxidase activity. Group sizes are indicated in graphs. *, **, or *** denote P<0.05, P<0.01, or P<0.001, respectively, in one-way ANOVA applying Tukey's post-hoc comparison of all groups using GraphPad Prism 5.0 software. ns, not significant.

Figure 5. Role of complement in mucosal protection against N. meningitidis.

(A): (i): Cohorts of mice were either left untreated or PMN depleted using the antibody clone RB6-8C5 and were subsequently i.n. infected with inoculi of 10⁸ CFU MC58 or MC58ΔsiaD (capsule-deficient mutant). Viable bacteria were recovered and enumerated at day 3 post infection. Bars in graph denote median. * denotes P<0.05, using the Mann-Whitney nonparametric test. CEACAM1-humanized cohort infected with MC58 (encapsulated) is the same as in Fig. 1D (day3) and is shown here for comparison (not included in statistical test). (ii): Cohorts of mice were i.n. infected with 10⁸ CFU H44/76 or H44/76ΔsiaD (capsule-deficient mutant) and viable bacteria recovered at day 3 post infection. Bars in graph denote median. (iii): Cohorts of mice were i.n. infected with 10⁸ CFU of MC58ΔsiaD, H44/76ΔsiaD or 2275, which is naturally non-capsulate because it harbors the capsule null locus 2 (cnl-2), grown in presence of 20 µM cytidine-5′-monophospho-N-acetylneuraminic acid (CMP-NANA). Viable bacteria were recovered and enumerated at day 3 post infection. Horizontal bars in graph represent median. (B) Cohorts of mice were i.n. infected with 10⁸ CFU of MC58 with or without receiving Cobra venom factor (CVF) 4 h prior to infection. Viable bacteria were recovered and enumerated at day 3 post infection. Data were pooled from at least two individual infections for each group, horizontal bars indicate median. Note that untreated control group is the same as in Fig. 1D (day3). (C): (i): CEACAM1-humanized mice were left untreated or received CVF alone, or anti-Gr-1 antibody clone RB6-8C5 alone, or both in combination at 30 h prior to infection with 10⁸ CFU of MC58. Viable bacteria were recovered and enumerated at day 3 post infection. Data were pooled from at least two individual infections for each group, horizontal bars indicate median. Significance levels are *, **, or *** denoting P<0.05, P<0.01, or P<0.001, respectively, using Kruskal-Wallis test applying Dunn's multiple comparison post hoc test. Note that untreated control group is the same as in Fig. 1D (day 3) and shown here to facilitate comparison. (ii): Cohorts of WT or CEACAM1-humanized mice were injected with CVF or left untreated at 24 h prior to i.n. infection with 10⁸ CFU of strain H44/76. Viable bacteria were recovered at day 3 post infection, horizontal bars in graph represent median.

Figure 6. Generation of protective adaptive mucosal immune responses towards N. meningitidis after intranasal colonization.

(A) Schematic representation of experiment. Cohorts of mice were i.n. infected with 10⁵ CFU of Nme H44/76 or either 10⁵ or 10⁸ CFU of H44/76Δopa or heat-inactivated H44/76 at day 0 (first exposure) and again at day 21 (second exposure). Intranasal challenge infections with 10⁸ CFU of H44/76 were performed for some cohorts at day 0 or day 21, for all other cohorts at day 52 and mice were sacrifized after 3 days to recover and enumerate viable bacteria. (B) Resulting recovered viable meningococci after exposure to indicated inoculi of either 10⁵ (left panel) or 10⁸ (right panel), with ‘none’ = no previous exposure, ‘wt’ = H44/76wt, ‘Δopa’ = H44/76Δopa, ‘h.i.’ = heat-inactivated H44/76 as indicated below x-axis. *, **, or *** denote P<0.05, P<0.01, or P<0.001, respectively, in Kruskal-Wallis nonparametric analysis applying Dunn's post hoc test. (C) Nasal lavage fluid was obtained from mice in (B) at day 55 and meningoccoccal (H44/76) specific IgG (upper panel) and IgA (lower panel) was determined by ELISA. Represented are mean values for each group, with group sizes indicated in graph. *, **, or *** denote P<0.05, P<0.01, or P<0.001, respectively, in one-way ANOVA applying Tukey's post-hoc comparison of all groups using GraphPad Prism 5.0 software. (D) Correlation of recovered CFU as in (B) with anti-meningococcal IgA concentrations in nasal lavage fluid as in (C) for each individual mouse exposed twice to 10⁵ CFU of the indicated inoculi before challenge, analyzed by Spearman's rank correlation test. (E) Meningococcal (H44/76) specific serum IgG concentration determined by ELISA in indicated treatment groups of experiment with 10⁵ CFU of the indicated inoculi used for exposures. Represented is the mean ± SEM for each group, with group sizes same as in (B). *, **, or *** denote P<0.05, P<0.01, or P<0.001, respectively, in one-way ANOVA applying Tukey's post-hoc comparison of all groups using GraphPad Prism 5.0 software. ns, not significant. Asterisks above error bars refer to post hoc comparison of CEACAM1-humanized mice and wild-type mice exposed to H44/76wt. Asterisks next to bracket refer to post hoc comparison of either wild-type or CEACAM1-humanized mice (as indicated) with all other groups (infected with H44/76Δopa or heat-inactivated H44/76) at day 55. (F) Meningococcal (H44/76) specific serum IgG concentration at day 55 determined by ELISA in indicated treatment groups of experiment with 10⁸ CFU of the indicated inoculi used for exposures. Group sizes are the same as in (B), right panel. (G) Serum bactericidal antibody (SBA) titres in serum samples taken at day -3 (‘pre’) day 24 or day 55, respectively, of mice exposed to 10⁵ CFU of the indicated inoculi (top panel), or 10⁸ CFU of the indicated inoculi (bottom panel). Each circle represents the result for one individual mouse. Group sizes as well as percentage of SBA-positive mice in each group are indicated in the graph. (H) Correlation of SBA titres as in (G) with meningococcal-specific IgG titres as in (E) for each individual mouse. Where applicable, data were pooled from day 24 and day 55, i.e. up to two data points were obtained from the same mouse. Thresholds indicate minimal IgG concentration above which all samples had positive (≤8) SBA titres. Results of Spearman's rank correlation are shown in each graph.

Figure 7. Polysaccharide conjugate vaccine induces sterile immunity in CEACAM1-humanized mice.

(A) Schematic representation of experiment. Cohorts of mice were vaccinated with serogroup C polysaccharide conjugate vaccine, or alum alone as control, at day 0, 21 and 42 and then i.n. challenged at day 74. (B) Recovered viable CFU at day three after i.n. challenge with 10⁸ CFU of serogroup C strain 90/18311, or serogroup B strain H44/76, as indicated below x-axis. Mouse treatment groups were as follows: ‘naive’ = untreated; ‘alum’ = controls receiving alum alone; ‘vaccine’ = vaccinated mice. (C) Nme-specific (strain 90/18311) Ig titres during the course of vaccination. Mean values ± SEM derived from the same groups as in (B) are plotted. Arrows indicate vaccine administration. (D) Meningococcal-specific (strain 90/18311) Ig titres in nasal wash fluid at day 77. Mean values ± SEM derived from the same groups as in (B) are plotted. *, denotes P<0.05, in one-way ANOVA applying Tukey's post-hoc comparison of all groups using GraphPad Prism 5.0 software. ns, not significant.