Serum resistance in an invasive, nontypeable Haemophilus influenzae strain

Abstract

A common feature of many different organisms causing bacteremia is the ability to avoid the bactericidal effects of normal human serum. In Haemophilus influenzae encapsulated strains are particularly serum resistant; however, we found that a nonencapsulated strain (R2866) isolated from the blood of an immunocompetent child with meningitis who had been successfully immunized with H. influenzae type b conjugate vaccine was serum resistant. Since serum resistance usually involves circumventing the action of the complement system, we defined the deposition of various complement components on the surfaces of this H. influenzae strain (R2866), a nonencapsulated avirulent laboratory strain (Rd), and a virulent type b encapsulated strain (Eagan). Membrane attack complex (MAC) accumulation correlated with the loss of bacterial viability; correspondingly, the rates of MAC deposition on the serum-sensitive strain Rd and the serum-resistant strains differed. Analysis of cell-associated immunoglobulin G (IgG), C1q, C3b, and C5b indicated that serum-resistant H. influenzae prevents MAC accumulation by delaying the synthesis of C3b through the classical pathway. Among the initiators of the classical pathway, IgG deposition contributes most of the C3 convertase activity necessary to start the cascade ending with MAC deposition. Despite similar IgG binding, strain R2866 delays C3 convertase activity compared to strain Rd. We conclude that strain R2866 can persist in the bloodstream, in part by inhibiting or delaying C3 deposition on the cell surface, escaping complement mediated killing.

FIG. 1

Bactericidal kinetic assay. Bacteria were inoculated in 40% serum incubated at 37°C, sampled at the times indicated, and quantitated using serial dilution and plating. im, immediate. Strains: ♦, R2866; ▴, Ela; ×, Rd.

FIG. 2

PCR-based typing confirms nontypeable genotype of R2866. Primers specific to all five capsular types and bexA failed to amplify genes needed for capsule expression in strains R2866 and Rd (note that Rd is a derivative of a type d that has lost most of its capsule genes). The molecular size marker is the 100-bp marker from Gibco BRL (Gaithersburg, Md.).

FIG. 3

Southern blot of R2866 chromosomal DNA with a probe for capsule genes. The 18-kb capsulation locus of strain Ela (BamHI fragment of pUO38) was used as a probe against EcoRI-digested chromosomal DNAs from strains R2866, Rd, and Ela and the reference strains representing the five capsular types.

FIG. 4

Outer membrane proteins show similar affinities to human serum immunoglobulins. Outer membrane proteins of strains R2866, Ela, and Rd were electrophoresed in a Tris-glycine–10% polyacrylamide gel and transferred to PVDF. After incubation in NHS for 1 h, the blot was probed for human IgG binding. Approximate molecular masses in kilodaltons are indicated.

FIG. 5

EGTA treatment of sera. Bactericidal kinetic assay as in Fig. 1 except that the serum was treated with 5 mM EGTA and contained 9 mM MgCl₂. im, immediate. strains: ♦, R2866; ▴, Ela; ×, Rd.

FIG. 6

Flow cytometry analysis of C3 deposition on strain Rd. y axis, counts per channel; x axis, fluorescence channel; insets, forward scatter versus side scatter. (A) Rd, no serum, no antibody; (B) Rd incubated in serum for 45 min, no antibody; (C) same as panel B with goat anti-C3 (αC3) only; (D) Same as panel B with anti-goat (FITC conjugated) only; (E) Same as panel B with anti-C3 and anti-goat (FITC conjugated).

FIG. 7

Flow cytometry analysis of the complement cascade on H. influenzae strains R2866, Ela, and Rd. (A, C, E, G, and I) Percentage of cells in population in the positive gate (Fig. 4D); (B, D, F, H, and J) geometric mean of cells gated positive. im, immediate; time is in minutes. Strains: ♦, R2866; ▴, Ela; ×, Rd.

FIG. 8

Complement deposition and cell death over time with strains R2866 (A) and Rd (B). CFU per milliliter are from Fig. 1. geometric mean fluorescence of the positive gated population are from Fig. 7. im, immediate; time is in minutes.

FIG. 9

ChoP decoration of LOS is not responsible for the serum resistance of strain R2866. (A) Same as Fig. 1 except that the serum was depleted of CRP; (B) M318 ChoP positive; (C) M319 (M318 isotype) ChoP negative; (D) R2866; (E) Rd.

FIG. 10

Complement reactivity of LOS. (A) LOS after electrophoresis and silver staining on a Tris-glycine–15% polyacrylamide gel. (B) Same as panel A except that after electrophoresis the LOS was transferred to PVDF and reacted with NHS, followed by probing with anti-human C3.

FIG. 11

Flow cytometry analysis of factor H binding on strains R2866 and Rd. (A) Percentage of the population positive for factor H binding over time; (B) geometric mean fluorescence of the positive cells counted in panel A. im, immediate; time is in minutes. Strains: ♦, R2866; ×, Rd.