Antibody-mediated elimination of the obligate intracellular bacterial pathogen Ehrlichia chaffeensis during active infection

Abstract

It is generally accepted that cellular, but not humoral immunity, plays an important role in host defense against intracellular bacteria. However, studies of some of these pathogens have provided evidence that antibodies can provide immunity if present during the initiation of infection. Here, we examined immunity against infection by Ehrlichia chaffeensis, an obligate intracellular bacterium that causes human monocytic ehrlichiosis. Studies with mice have demonstrated that immunocompetent strains are resistant to persistent infection but that SCID mice become persistently and fatally infected. Transfer of immune serum or antibodies obtained from immunocompetent C57BL/6 mice to C57BL/6 scid mice provided significant although transient protection from infection. Bacterial clearance was observed when administration occurred at the time of inoculation or well after infection was established. The effect was dose dependent, occurred within 2 days, and persisted for as long as 2 weeks. Weekly serum administration prolonged the survival of susceptible mice. Although cellular immunity is required for complete bacterial clearance, the data show that antibodies can play a significant role in the elimination of this obligate intracellular bacterium during active infection and thus challenge the paradigm that humoral responses are unimportant for immunity to such organisms.

FIG. 1

Administration of immune serum protected mice from ehrlichia infection. (a) Transient bacterial elimination in susceptible SCID mice. C57BL/6 scid mice were infected intraperitoneally with 10⁶ to 2 × 10⁶ E. chaffeensis-infected DH82 cells (day 0), followed by subcutaneous injection of 0.1 ml of PBS, normal C57BL/6 serum, or immune serum on day 0 or of immune serum on day 3 postinfection, as indicated. Liver tissue was harvested on the indicated days postinfection and was analyzed for the presence of E. chaffeensis by QPCR for E. chaffeensis 16S rDNA, as described in Materials and Methods. Immune serum was obtained from C57BL/6 mice that had been inoculated with E. chaffeensis-infected DH82 cells. Similar observations of the immune serum protection were made in at least three independent experiments using a semiquantitative PCR assay (Table 1). The lower bacterial titers observed on day 17 postinfection, compared to those on days 10 and 24, were not observed in other experiments (see also Fig. 4b) and most likely represent experimental variability. In all cases, the immune serum group was significantly different from controls. ∗, bacteria were not detected in the assays. Ec, E. chaffeensis. (b) Bacterial elimination in immunocompetent C57BL/6 mice. The mice were infected as described for panel a, and PBS or immune serum was administered via the peritoneum on day 3 postinfection. Tissues were harvested at the time of serum administration and 1 or 4 days later, and representative mice from each group were analyzed by QPCR. The experiment was performed in triplicate, and semiquantitative PCR analysis, which is more sensitive than QPCR, also failed to detect bacteria in spleen or liver tissue of any of the mice that received immune serum (not shown). P values obtained from the semiquantitative analyses were 0.09 and 0.001 for mice analyzed on day 4 and day 7, respectively. n.d., not determined.

FIG. 2

Bacterial clearance was E. chaffeensis specific and was mediated by antibodies. (a) C57BL/6 scid mice were infected by transfer of E. chaffeensis-infected splenocytes obtained from a SCID mouse 17 days postinfection. Immune serum was obtained from C57BL/6 mice that had been inoculated with E. chaffeensis-infected DH82 cells and was administered at the time of bacterial infection. Bacterial loads were determined by QPCR. ∗, bacteria were not detected in the infected mice. Each histogram bar represents a single mouse. Ec, E. chaffeensis. (b) Mice were infected as described for Fig. 1 and were administered on day 10 postinfection normal mouse serum or serum obtained from C57BL/6 mice that had been inoculated with either uninfected DH82 cells (immune serum-uninfected) or with E. chaffeensis-infected DH82 cells (immune serum-infected). Liver tissue was harvested on day 14 for QPCR analyses. QPCR analyses of representative individual mice are shown. The observations were confirmed in a separate experiment (not shown) where 16 mice (in three groups) were analyzed over a period of 24 days. The semiquantitative data from both experiments where serum was administered on day 14 postinfection were normalized and combined (a total of four mice for each group), and the means and standard deviations were as follows: normal serum, 5.0 ± 0.82; serum from mice inoculated with uninfected cells, 4.3 ± 1.1; immune serum, 1.8 ± 1.8. (c) C57BL/6 scid mice were infected on day 0 with E. chaffeensis-infected DH82 cells, followed by intraperitoneal administration of 0.1 ml of PBS or C57BL/6 immune serum, 200 μg of ammonium sulfate-fractionated immune serum (SAS), or 100 μg of protein A affinity-purified antibodies 3 days postinfection. The presence of E. chaffeensis antibodies in each of the preparations was confirmed by immunofluorescence assay. Mice were harvested 5 and 10 days postinfection, and liver tissue from representative individual mice was analyzed by QPCR. Semiquantitative analyses of a total of 16 mice from two experiments revealed significant differences between the buffer- and antibody-treated mice (mean ± standard deviation): PBS, 3.8 ± 0.83; immune serum, 1.0 ± 1.2; protein A-purified antibodies, 0.5 ± 0.87; ammonium sulfate-purified antibodies, 1.3 ± 1.3. ∗, bacteria were not detected in the QPCR assays.

FIG. 3

E. chaffeensis antibody responses in C57BL/6 mice. (a) C57BL/6 mice were inoculated with 2 × 10⁶ infected DH82 cells, serum was harvested on the indicated days postinfection, and E. chaffeensis antibody titers were determined by immunofluorescence assay using a secondary antibody specific for mouse Ig. Each data point represents the serum titer of one mouse. (b) Western analysis of murine and human E. chaffeensis antisera. Bacterial antigens were obtained from uninfected (u) or E. chaffeensis-infected (i) DH82 cells. The samples were Western blotted and probed with mouse or human antisera, followed by a species-specific horseradish peroxidase-conjugated secondary antibody and chemiluminescence development. ∗, E. chaffeensis molecules that were detected by both the mouse and human antibodies. Molecular mass standards, in kilodaltons, are at the left of the gel.

FIG. 4

Clearance of bacteria during active infection. (a) C57BL/6 scid mice received PBS on day 10 or immune serum on day 10 or day 17 postinfection and were analyzed by QPCR 1 and 2 weeks later. Ec, E. chaffeensis. (b) SCID mice received PBS or serum on day 10 postinfection and were analyzed 1, 3, 7, and 10 days later. (c) SCID mice were administered serum on day 17 postinfection and were analyzed 1 and 3 days later. The control mice shown in panels b and c are identical. (d) Infected SCID mice received 0.1 ml of PBS or dilutions of immune serum on day 3 postinfection and were analyzed 7 or 14 days later. The values in the key to the bars are the reciprocal titers of the immune serum that was administered, as determined by immunofluorescence assay. In all experiments, mice were analyzed only where indicated by the histograms. ∗, bacteria were not detected. The data obtained by semiquantitative analyses of the experiments shown in panels a to c, as well as data from an additional experiment, are shown in Table 2.

FIG. 5

Repeated serum administration results in prolonged immunity. Infected C57BL/6 scid mice received PBS or weekly injections of serum beginning 10 days postinfection. Liver tissue was harvested at 2- and 4-day intervals following each serum administration (shaded histograms) or 28 days after infection of a mouse that received no serum (black histogram). The bacterial loads were determined using semiquantitative PCR. The data indicate the integrated optical densities of ethidium bromide-stained PCR products in agarose gels, as determined by densitometry. Semiquantitative PCR was utilized because bacterial loads in the mice that received the antibodies were below the limit of detection using QPCR. Two additional mice that did not receive serum died and were not analyzed. ∗, PCR products were not detectable in agarose gels. Based on the assumption that the 2 SCID mice that did not survive exhibited levels of bacterial infection similar to that of the surviving mouse, a statistical comparison of 10 treated mice that exhibited low bacterial loads with 3 untreated mice indicated a P value of 0.004 by Fisher's exact test (14). In two additional experiments 13 mice that received two or more injections of immune serum all survived longer than 31 days postinfection (not shown).

FIG. 6

Persistent infection in T-cell-deficient mice. T-cell receptor β-chain-deficient (TCRb), TCRβ/TCRδ-chain deficient (TCRb/d), and C57BL/6 scid (SCID) mice were infected with E. chaffeensis, and bacterial colonization was monitored by semiquantitative PCR on the indicated days postinfection. PCR analyses of C57BL/6 scid mice are shown as a basis for comparison. Normalized data from three experiments are shown and represent the averages of semiquantitative PCR analyses of three to four mice from each of the gene-targeted strains. Error bars indicate standard deviations. Data from the analyses of the SCID mice were obtained from a single animal on the indicated days. Semiquantitative PCR was utilized because it was not possible to detect the bacteria in the T-cell-deficient mice by QPCR. The T-cell-deficient mice were maintained on the C57BL/6 genetic background (strains C57BL/6 tcrb and C57BL/6 tcrb/tcrd). No sign of disease was noted in any of the T-cell-deficient mice. ∗, PCR products were not detectable in agarose gels.