Commensal Microbes Affect Host Humoral Immunity to Bordetella pertussis Infection

Abstract

As important players in the host defense system, commensal microbes and the microbiota influence multiple aspects of host physiology. Bordetella pertussis infection is highly contagious among humans. However, the roles of the microbiota in B. pertussis pathogenesis are poorly understood. Here, we show that antibiotic-mediated depletion of the microbiota results in increased susceptibility to B. pertussis infection during the early stage. The increased susceptibility was associated with a marked impairment of the systemic IgG, IgG2a, and IgG1 antibody responses to B. pertussis infection after antibiotic treatment. Furthermore, the microbiota impacted the short-lived plasma cell responses as well as the recall responses of memory B cells to B. pertussis infection. Finally, we found that the dysbiosis caused by antibiotic treatment affects CD4⁺ T cell generation and PD-1 expression on CD4⁺ T cells and thereby perturbs plasma cell differentiation. Our results have revealed the importance of commensal microbes in modulating host immune responses to B. pertussis infection and support the possibility of controlling the severity of B. pertussis infection in humans by manipulating the microbiota.

Keywords: Bordetella pertussis; PD-1; humoral immunity; microbiota.

FIG 1

Effect of broad-spectrum antibiotic treatment on microbial profile in mouse feces and lungs. (A) 3D-PCoA score plot of the gut microbiota in naive mice and mice treated with antibiotics, based on weighted UniFrac metrics. Each dot represents an individual mouse. (B) Phylogenetic profile of bacterial phyla in the feces of mice treated or not treated with antibiotics. Stacked bar charts show the 10 main phyla, identified on the basis of their relative abundance in antibiotic-treated or non-antibiotic-treated mice 3 days after the cessation of antibiotic treatment. (C) Heat map demonstrating the relative abundance of the dominant bacterial phyla in the feces of naive mice and mice treated with antibiotics. (D) Phylogenetic profile of bacterial phyla in the lungs of mice treated or not treated with antibiotics. Stacked bar charts show the 10 main phyla, identified on the basis of their relative abundance in antibiotic-treated or non-antibiotic-treated mice 3 days after the cessation of antibiotic treatment. (E) 3D-PCoA score plot of the lung microbiota in naive mice and mice treated with antibiotics, based on weighted UniFrac metrics. Each dot represents an individual mouse. Ab and A, antibiotic-treated mice.

FIG 2

The microbiota is necessary for early inhibition of B. pertussis colonization in the lungs. The lung colonization profiles of B. pertussis strain BPMM (BP) in mice treated or not treated with antibiotics (Ab) were measured. BALB/c mice were infected intranasally with 5 × 10⁶ CFU of BPMM. The lungs were harvested at the indicated time points and homogenized. Appropriate dilutions of the lung homogenates were plated onto blood agar plates, and the number of CFU was counted after 4 days of incubation at 37°C. Four mice per group per time point were assessed individually. Results are expressed as the mean of the log₁₀ number of CFU per lung ± standard deviations (SD) calculated for each mouse. **, P ≤ 0.01 compared with the values obtained between two groups at the same time points. D, day.

FIG 3

The host microbiota is necessary for early systemic antibody responses to B. pertussis. The systemic anti-B. pertussis antibody responses in antibiotic-treated mice and nontreated naive mice nasally infected with BPMM were analyzed. Groups of 4 adult BALB/c mice were i.n. infected with 5 × 10⁶ CFU of BPMM. Sera were collected at 10 and 17 days postinfection. Systemic anti-BPMM total IgG, IgG2a, IgG1, and IgA titers in serially diluted individual serum samples were measured by ELISA using BPMM whole-cell lysate as the coating antigen. The results are representative of those from three independent experiments. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.

FIG 4

Impaired short-lived plasma cell response to B. pertussis infection in antibiotic-treated mice. B. pertussis-specific antibody-secreting cells (ASCs) in the spleens of mice treated or not treated with antibiotics were measured at 10 days postinfection by an ELISPOT assay. Representative spot formations (A) and total frequencies, expressed as the mean ± SD (B), are presented. Data are representative of those from two independent experiments. *, P ≤ 0.05; **, P ≤ 0.01.

FIG 5

Microbiota dysbiosis caused by antibiotic treatment leads to an impaired recall response of memory B cells against B. pertussis infection. (A) Bacterial carriage in the lungs was measured at 3 h, 3 days, and 7 days after infection and compared. (B and C) B. pertussis-specific IgG, IgG2a, and IgG1 titers (B) as well as the frequencies of B. pertussis-specific IgG ASCs in the spleens (C) were measured in mice treated or not treated with antibiotics and infected twice with B. pertussis. Analysis was performed 14 days after secondary B. pertussis infection. Data are expressed as the mean ± SD. *, P ≤ 0.05; **, P ≤ 0.01.

FIG 6

B. pertussis-specific antibodies restored the colonization resistance to B. pertussis in antibiotic-treated mice. Antibiotic-treated mice and naive mice were i.p. injected with 200 μl of naive or high-titer anti-BPZE1 immune serum before B. pertussis infection. Lung colonization of B. pertussis was measured 3 h, 3 days, and 7 days after infection and compared. Data are expressed as the mean ± SD. *, P ≤ 0.05; **, P ≤ 0.01.

FIG 7

Impaired CD4⁺ T cell generation and PD-1 expression in T cells of mice treated with antibiotics infected with B. pertussis. (B, D, F, and H) Naive mice or mice treated with antibiotics were infected with BPMM, and the frequencies of CD4⁺ T cells (B), GC B cells (D), T_fh cells (F), and PD-1 expression on CD4⁺ T cells in the mouse spleens (H) were measured at the indicated time points following infection. Four mice per group per time point were analyzed individually. (A, C, E and G) Representative flow cytometry analyses. Data are expressed as the mean ± SD. *, P ≤ 0.05; **, P ≤ 0.01.