CD11c⁺CD8α⁺ Dendritic Cells Promote Protective Immunity to Respiratory Infection with Bordetella pertussis

Animals

BALB/c mice were purchased from Harlan U.K. Animal experiments and maintenance were approved and regulated by the university ethics committee and the Irish Department of Health.

B. pertussis challenge model

Mice were challenged by aerosol exposure to live B. pertussis. Virulent B. pertussis Wellcome 28 was grown at 36°C in Stainer-Scholte liquid medium. Bacteria from a 48-h culture were resuspended at a concentration of ∼2 × 10¹⁰ CFU/ml in physiological saline containing 1% casein. The challenge inoculum was administered to mice over a period of 15 min (1.5 ml/min) with a nebulizer in a sealed container within a class 3 exhaust-protected cabinet. This was followed by a rest period of 10 min. The course of infection was followed by performing B. pertussis CFU on lung homogenates at intervals after infection (8).

Depleting and blocking Abs

Hybridoma cells were obtained from the American Type Culture Collection and secreted IgG that was purified from hybridoma supernatants using a MAbTrap affinity chromatography kit (GE Healthcare). The optimal concentration of Ab required to deplete >95% of the relevant target cells was determined by titration. Briefly, mice were injected i.p. with increasing concentrations of depleting Ab. The spleens of these mice were subsequently investigated for the target cell 24, 48, and 72 h after injection by flow cytometric analysis. The hybridoma cell lines, clones 2.43 and N418, were used to generate rat anti-mouse (IgG1) CD8α and CD11c, respectively. In vivo blocking of CD103 was achieved by injecting mice i.p. with anti-CD103 (clone M290). A rat anti-mouse HRP (HRPN) IgG1 was used as the isotype-matched control (Bioxcell) in all experiments except those using anti-CD8α. Rat anti-mouse IgG2b (Sigma-Aldrich) was used as the isotype-matched control for anti-CD8α (clone 2.43).

Cell isolation and purification

Infiltrating leukocytes were isolated from lungs by digesting lung tissue in PBS containing 1 mg/ml Collagenase D (Sigma-Aldrich) plus 0.2 μM DNase (Sigma-Aldrich) for 1 h at 37°C. T cell and DC subtypes were isolated from various tissues or from cells expanded in culture using either magnetic-assisted cell sorting (MACS; Miltenyi Biotec) or a MoFlo cell sorter (Beckman Coulter). CD11c⁺CD24⁺CD45RA⁻ cells were sorted from FL-generated bone marrow DC using fluorochrome-conjugated Abs directed against the relevant surface Ags (eBioscience).

Flow cytometry

Cells were stained with Abs directed against CD3, CD4, CD8α, CD8β, CD11c, CD40, H-2^K, I-A/I-E, CD1d, CD24, CD45RA, GR1, B220, CD103, CD205, CD80, and CD86 (eBioscience). Labeled cells were acquired using a CyAn flow cytometer (Beckman Coulter). T cell and DC subsets were determined by gating on viable forward vs side scatter gates.

Detection of intracellular cytokines in activated T cells and CD11c⁺ DC

Lung cells were activated with PMA (10 ng/ml) and ionomycin (1 μg/ml) in the presence of irradiated (30 Gy from a ¹³⁷Cs source) spleen cells for 4 h and then incubated overnight at 37°C in the presence of brefeldin A (BFA; Sigma-Aldrich). Alternatively, CD4⁺ T cells were activated with Ag in vivo by injecting mice with 4 × 10⁹ heat-killed B. pertussis i.p. 4 h before removal of lungs. Lung mononuclear cells were incubated with BFA overnight without additional stimulus. Cells were permeabilized and stained with Abs directed against IL-4, IL-10, IL-17, and IFN-γ. Intracellular cytokine expression in DC was determined directly by staining isolated cells with Abs directed against relevant Ags without restimulation or incubation with BFA.

H&E staining of lung tissue sections

Lungs were removed from the animals and fixed overnight in 10% neutral-buffered formalin. Tissues were then processed overnight to paraffin wax and embedded. Blocks were sectioned at 4 μm and were stained with H&E by a Cot-20 automated linear stainer (Midite).

In vitro activation of CD4⁺ helper T cells from B. pertussis-infected mice with DC

CD11c⁺CD8α⁺ DC were sorted from lung homogenates on day 7 after infection using a MoFlo cell sorter. CD4⁺ T cells were purified from spleens of infected mice using CD4 T cell isolation kits (Miltenyi Biotec). In brief, 0.2 × 10⁵ DC and 1 × 10⁵ T cells were cultured in RPMI 1640 medium supplemented with l-glutamine and 10% FCS in wells of a 96-well plate and activated with 2 μg/ml immobilized anti-CD3. After 48 h, supernatants were removed and analyzed for IL-10, IL-12p70, IFN-γ, and IL-4 by ELISA (R&D systems). Proliferation was assessed 72 h after activation by tritiated thymidine incorporation.

Generation of CD11c⁺CD24⁺CD45RA⁻ cells with FL in vitro

Bone marrow cells were isolated from the femurs of BALB/c mice and cultured at a concentration of 2 × 10⁶ cells/well in Stemline dendritic cell maturation medium (Sigma-Aldrich) supplemented with 0.3 g/L l-glutamine and 300 ng/ml recombinant FL (ImmunoTools). Cells were incubated in humidified air for 8 days at 37°C and in 10% CO₂. Additional FL (300 ng/ml) was added on days 3 and 6. These cells were subsequently washed and stained with Abs against CD24, CD45RA, and CD11c. CD11c⁺CD24⁺CD45RA⁻ cells (FL-DC) were purified using a MoFlo cell sorter and counted. In brief, 0.5 × 10⁵ CD11c⁺CD24⁺CD45RA⁻ cells were injected i.p. into mice 30 min after infection.

Uptake of green fluorescent latex bead (FLB) particles by DC

Plain yellow-green fluorescent 0.5-μm latex particles (Polysciences) were diluted 1/25 in PBS for a final administration of ∼3.64 × 10⁸ particles/mouse. Twenty microliters of this solution was administered intranasally (i.n.) 15 min before infection. Control mice received latex particles in the absence of infection.

Statistical analysis

Results, expressed as mean ± SEM were compared using the Student t test or by ANOVA. Values of p < 0.05 were considered statistically significant.

CD11c⁺CD8α⁺ DC infiltrate the lungs and draining lymph nodes during acute phase of infection with B. pertussis

We investigated the infiltration and function of CD11c⁺CD8α⁺ lymphoid DC in the lungs and draining lymph nodes of mice through the course of infection with B. pertussis. Respiratory challenge of naive BALB/c mice by exposure to an aerosol of live B. pertussis resulted in a reproducible nonlethal infection with peak bacterial counts on day 14 and approached complete clearance by 28 days (Fig. 1⇓A). The relative percentage of CD11c⁺CD8α⁺ DC was determined by gating on viable lymphocytes from forward vs side scatter plots in conjunction with the appropriate isotype and compensation controls. The total number of cells expressing various surface Ags was calculated based on the total number of viable cells and the percentage of CD11c⁺CD8α⁺ cells within the viable lymphocyte gate. The numbers of CD11c⁺ and CD11c⁺CD8α⁺ DC were low in the CLN, thoracic lymph nodes (TLN), and lungs of noninfected control mice (Fig. 1⇓, B–D). However, during B. pertussis infection, we consistently found an early influx of CD11c⁺ and CD11c⁺CD8α⁺ DC into the CLN and a later influx into the lungs (Fig. 1⇓, B–D). The numbers of both DC subsets in the CLN increased dramatically between 4 and 24 h of aerosol challenge with live B. pertussis and steadily diminished thereafter to baseline levels (Fig. 1⇓, B and C). Conversely, the number of CD11c⁺ and CD11c⁺CD8α⁺ DC increased in the lung from day 1 and peaked on days 5–7 of infection. The number of DC subsets in the draining lymph nodes and lungs declined to that in control mice after 14–21 days (Fig. 1⇓, B and C). There was a small, but not statistically significant, increase in the frequency of CD11c⁺CD8α⁺ DC in the TLN 3 days after B. pertussis aerosol challenge (Fig. 1⇓C). However, the number had returned to baseline values by day 5 (Fig. 1⇓C) and remained there when examined 7, 14, and 21 days after challenge (data not shown). Our findings suggest that following B. pertussis aerosol challenge, DC are recruited to the CLN and then may migrate to the lungs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Recruitment of CD11c⁺CD8α⁺ DC to the lungs and draining lymph nodes soon after respiratory challenge with B. pertussis. Mice were infected by exposure to an aerosol of live virulent B. pertussis. A, The course of infection was followed by performing CFU counts on lung homogenates at intervals after B. pertussis challenge-infected mice. B–D, The CLN, TLN, and lungs were removed and digested with Collagenase D at various time points after challenge. Cells were counted and stained for CD11c and CD8α. The relative percentage of CD11c⁺CD8α⁺ DC was determined by gating on viable lymphocytes in conjunction with the appropriate isotype and compensation controls. The total number of CD11c⁺ (B) or CD11c⁺CD8α⁺ (C) was calculated from the total number of viable cells and the percentage of CD11c⁺ or CD11c⁺ CD8α⁺ cells within the viable lymphocyte gate. D, Representative FACS profiles for CD11c⁺CD8α⁺ cells in either CLN or lung tissue from naive control mice or B. pertussis-infected mice 1 and 4 days after challenge. These results are representative of four separate experiments. *, p < 0.05 and **, p < 0.01 vs noninfected control mice (day 0).

CD11c⁺CD8α⁺ DC in the CLN acquire particulate Ag from the respiratory tract within 4 h of respiratory challenge with B. pertussis

We have shown that CD11c⁺CD8α⁺ DC with a mature phenotype are present in the CLN of mice within 24 h of B. pertussis challenge. However, it is unclear whether CD11c⁺CD8α⁺ DC migrate from the lung or reside in the CLN and acquire Ag from other migrating cells types. A recent report suggested that CD11c⁺CD8α⁺ DC are resident in the draining lymph nodes and rarely migrate (28). Therefore, we hypothesized that these cells acquire particulate Ag from dying APC that migrate from the lung and in turn present Ag via MHC to T cells within the lymph node. To test this hypothesis, green FLB were administered i.n. to mice 15 min before infection with B. pertussis and the number of FLB⁺ DC were determined in the CLN and lung at different time points after infection. FLB were found in the total leukocyte population and in CD11c⁺CD8α⁺ DC isolated from the CLN 4 h after infection (Fig. 2⇓A). On the other hand, very few total FLB⁺ leukocytes or FLB⁺CD11c⁺CD8α⁺ DC were detected in the CLN or lungs of noninfected mice up to day 3 after infection (Fig. 2⇓B and data not shown). Furthermore, very few FLB⁺ cells were detected in the lungs from either infected or noninfected control mice up to day 7 after challenge (data not shown). These findings suggest that CD11c⁺CD8α⁺ DC in the CLN can acquire particulate Ag from the upper respiratory tract within hours of infection. It remains unclear however whether or not the CD11c⁺CD8α⁺ DC activated in the CLN are the same as those observed in the lung some days later. Based on this data, it would seem that they are not the same cell.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Intranasally delivered green FLB are detectable in CD11c⁺CD8α⁺ DC from the CLN of mice 4 h after B. pertussis challenge. Green FLB were administered i.n. to mice that were infected with B. pertussis 15 min later. CLN from noninfected and B. pertussis-infected mice were examined after 4 h for the presence of total FLB⁺ cells within the viable leukocyte gate and FLB⁺CD11c⁺CD8α⁺ cells. A, Histogram gated on viable leukocyte gate, representing FLB⁺CD11c⁺CD8α⁺ DC from the CLN, 4 h after infection (white histogram, black line). FLB⁺ control (gray histogram) represents cells isolated from the CLN of noninfected animals and incubated for 15 min at room temperature with FLB. FLB only (dashed line) represents cells from the CLN of animals that received FLB in the absence of B. pertussis infection. CLN cells from the CLN of B. pertussis-infected animals that did not receive any FLB⁺CD11c⁺CD8α⁺ (black histogram). B, The total number of FLB⁺ cells or FLB⁺CD11c⁺CD8α⁺ DC in CLN or lungs of mice 4 h after challenge with B. pertussis. These results are representative of two separate experiments. **, p < 0.01 vs noninfected control mice.

CD11c⁺CD8α⁺ DC cells from the CLN and lungs of infected mice are mature and secrete proinflammatory cytokines

We next examined the activation status of CD11c⁺CD8α⁺ DC in the CLN and lungs. Infiltrating CD11c⁺CD8α⁺ DC expressed surface markers indicative of maturation, but at different stages of infection in the CLN and lungs. Expression of MHC class I and CD40 was enhanced on CD11c⁺CD8α⁺ DC in the CLN, with peak expression 1 day after challenge (Fig. 3⇓, A and B). MHC class II expression was also enhanced with peak expression on day 3. A similar pattern was observed on CD11c⁺CD8α⁺ DC in the lung, except that peak expression of these surface Ags was delayed by at least 24 h when compared with the CLN (Fig. 3⇓, A and B). We also investigated whether CD11c⁺CD8α⁺ DC expressed surface CD205, a marker of Ag uptake, during acute infection with B. pertussis. CD205 expression was significantly enhanced in the CLN after 24 h and in lungs 36–48 h after B. pertussis challenge and gradually declined during the first 7 days of infection (Fig. 3⇓, A and B). The number of CD11c⁺CD8α⁺ cells expressing either CD80 or CD86 DC peaked by days 3 and 5 in the CLN and lungs, respectively, in mice after challenge with B. pertussis (Fig. 3⇓A). Finally, the majority of CD11c⁺CD8α⁺ DC from CLN and lungs early in infection expressed CD24 (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

CD11c⁺CD8α⁺ DC cells from the CLN and lungs of B. pertussis-infected mice express a number of activation and maturation markers. Mice were infected by exposure to an aerosol of live virulent B. pertussis. The CLN and lungs were removed and digested with Collagenase D at various time points after infection. Cells were counted and stained for CD11c and CD8α with CD40, MHC class I (MHC-I), MHC class II (MHC-II), CD205, CD80, and CD86. A, The total number of cells expressing various surface Ags was calculated based on the total number of viable cells and the percentage of CD11c⁺CD8α⁺ cells within the viable lymphocyte gate. B, Representative dot plots showing peak expression of costimulatory molecules and activation markers within the CD11c⁺CD8α⁺ gate. Control dot plots represent cells isolated from the CLN of noninfected control mice. CLN and lung dot plots represent peak expression of MHC class I and CD40 on days 1 and 2, CD205 on days 2 and 3, MHC class II on days 3 and 5, and CD80/CD86 expression on days 3 and 5, respectively, after challenge with B. pertussis (B). Results are representative of four separate experiments. *, p < 0.05 and **, p < 0.01 vs noninfected control mice.

CD11c⁺ DC have been shown to secrete IFN-γ, IL-12p70, IL-4, and IL-10 in a variety of murine models of infection and inflammation (29, 30, 31, 32, 33, 34). Therefore, we performed intracellular cytokine staining on CD11c⁺CD8α⁺ DC during the acute stage of infection with B. pertussis. Cells isolated from the CLN or lungs of infected mice were permeabilized and stained directly with anti-cytokine Abs, without reactivation or incubation with BFA in vitro. A significant number of CD11c⁺CD8α⁺ in the CLN secreted IL-4 and IL-10 and this peaked 4 days after infection. A smaller number of CD11c⁺CD8α⁺ DC in the CLN of B. pertussis-infected mice secreted IFN-γ (Fig. 4⇓). In contrast, a high frequency of IFN-γ-secreting CD11c⁺CD8α⁺ DC were detected in the lungs of infected mice, especially 3–5 days after B. pertussis challenge (Fig. 4⇓). IL-4- and IL-10-secreting CD11c⁺CD8α⁺ DC were also detected in the lungs, but at a lower frequency than the IFN-γ-secreting CD11c⁺CD8α⁺ DC (Fig. 4⇓). Although a number of CD11c⁺CD8α⁺ DC expressed both IL-4 and IFN-γ or IL-10 and IFN-γ in the CLN, very few coproducers were detected at any stage during acute infection in the lung (Fig. 4⇓). The percentage of CD11c⁺CD8α⁺IFN-γ⁺ DC was lower than CD11c⁺CD8α⁺IL-4⁺ or CD11c⁺CD8α⁺ IL-10⁺ DC in the CLN on day 4 of infection when compared with noninfected control mice (Fig. 4⇓, B and C). In contrast, the percentage of CD11c⁺CD8α⁺IFN-γ⁺ DC was increased in the lungs of infected animals on day 5 (Fig. 4⇓D). These findings demonstrate that activated CD11c⁺CD8α⁺ DC found in the CLN soon after respiratory challenge with B. pertussis predominantly express IL-4 and IL-10 with a lower frequency expressing IFN-γ, whereas the CD11c⁺CD8α⁺ DC infiltrating the lungs a few days later express predominantly IFN-γ.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

CD11c⁺CD8α⁺ DC from the CLN and lungs of infected mice express IFN-γ, IL-10. and IL-4. Mice were infected by exposure to an aerosol of live virulent B. pertussis, and the CLN and lungs were removed and digested with Collagenase D at various time points after challenge. Cells were permeabilized and stained directly with anti-IL-4, IL-10, and IFN-γ Abs without restimulation or incubation with BFA. A, The total number of cells expressing various cytokines was calculated based on the total number of viable cells and the percentage of CD11c⁺CD8α⁺ cells within the viable lymphocyte gate. Representative dot plots showing cells from the CLN of noninfected control mice (B), CLN (C), and lungs (D) of B. pertussis-infected mice on day 4 after challenge. Results are representative of four separate experiments. *, p < 0.05 and **, p < 0.01 vs noninfected control mice.

CD11c⁺CD8α⁺ DC from infected mice can induce proliferation and proinflammatory cytokine secretion by CD4⁺ T cells in vitro

We next investigated whether CD11c⁺CD8α⁺ DC were as effective as total CD11c⁺ DC at stimulating CD4⁺ T cells and if the activated T cells had a similar cytokine profile. Total CD11c⁺ DC or CD11c⁺CD8α⁺ DC were sorted from the lungs of mice at the peak of infiltration 7 days after challenge with B. pertussis (Fig. 5⇓A) and cocultured in vitro with CD4⁺ T cells purified from the spleens of infected mice. CD11c⁺CD8α⁺ DC were as effective as total CD11c⁺ DC in inducing proliferation of CD4⁺ T cells in response to anti-CD3 (Fig. 5⇓B). IFN-γ (Fig. 5⇓C) and IL-4 (Fig. 5⇓D) were also detected in supernatants of CD4⁺ T cells stimulated with CD11c⁺CD8α⁺ DC as well as CD11c⁺ DC, although the concentrations were higher with the latter (Fig. 5⇓C). IL-10 and IL-12p70 production were undetectable in supernatants from CD4⁺ T cells stimulated with either population of DC (data not shown). The results show that CD11c⁺CD8α⁺ DC from the lungs of mice during the acute stage of infection with B. pertussis can activate CD4⁺ T cells in vitro.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

CD11c⁺CD8α⁺ cells from the lungs of B. pertussis-infected mice can stimulate CD4⁺ T cells in vitro. A, Infiltration of CD11c⁺ or CD11c⁺CD8α⁺ DC into lungs at intervals after B. pertussis challenge. B, CD11c⁺ or CD11c⁺CD8α⁺ DC were sorted from lung homogenates of mice on day 7 of respiratory infection with B. pertussis. Each cell subset was activated with 2 μg/ml immobilized anti-CD3 in the absence or presence of conventional CD4⁺ T cells (purified from the spleens of infected animals). Proliferation was determined by [³H]thymidine incorporation 72 h later. *, p < 0.05 and **, p < 0.01 vs noninfected control mice. The concentrations of IFN-γ (C) and IL-4 (D) in the supernatants were quantified by ELISA 48 h after activation. These results are representative of two separate experiments. Results were compared using Student’s t test. *, p < 0.05 and **, p < 0.01 vs DC or T cells alone.

Depletion of CD8α⁺ cells impairs bacterial clearance

Having shown that functionally active, mature CD11c⁺CD8α⁺ DC infiltrate into the lungs during acute infection with B. pertussis, we examined the hypothesis that they may play a positive role in antibacterial immunity and thereby function in bacterial clearance. To examine the function of CD11c⁺CD8α⁺ DC rather than total CD11c⁺ DC, we depleted CD8α⁺ cells in vivo using an anti-CD8α-depleting Ab. Although this approach will also deplete CD8⁺ T cells, unlike other approaches (35), it will not deplete other DC subsets. CD8α⁺ cells were depleted from mice infected with B. pertussis by injecting anti-CD8α Abs i.p. on days −1, +1, +3, +6, +13, and +19 after B. pertussis challenge. Depletion of CD8α was confirmed by flow cytometric analysis on spleen cells (Fig. 6⇓A). Depleting CD8α⁺ cells throughout infection had a marginal impact on the numbers of total CD11c⁺ cells in the lung (Fig. 6⇓B); however, CD11c⁺CD8α⁺ DC were almost completely absent from the lungs of infected mice treated with anti-CD8α Ab (Fig. 6⇓C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Depleting CD8α⁺ cells delays bacterial clearance in infected mice. Mice were infected by exposure to an aerosol of live virulent B. pertussis. Mice received either 200 μg/mouse anti-CD8α-depleting Ab (anti-CD8α) or irrelevant IgG2b isotype control Ab (200 μg/mouse) i.p. 24 h before infection and on days 1, 3, 7, 14, and 19 thereafter. A, FACS analysis on spleen cells isolated from CD8α-depleted mice 24 h after injection showing absence of CD8α⁺ cells. The number of CD11c⁺ (B) and CD11c⁺CD8α⁺ (C) DC infiltrating the lung in B. pertussis-infected mice, control mice, and mice treated with anti-CD8α 24 h before challenge with B. pertussis and on days 1, 3, 7, 14, and 19 (−1 to +19). D, The course of infection in B. pertussis-infected control mice and in CD8α-depleted infected animals was followed by perfuming CFU counts on lung homogenates at intervals after the initial challenge. One of five animals died from the CD8α-depleted group after challenge with B. pertussis (†). **, p < 0.01 vs nondepleted B. pertussis infected mice. E, Representative dot plots, gated on viable CD3⁺CD4⁺ T cells, showing intracellular cytokine expression in cells isolated from the lungs of mice from control and experimental groups on day 30 after B. pertussis challenge. These results are representative of four separate experiments.

Depletion of CD8α⁺ cells from mice throughout infection with B. pertussis delayed bacterial clearance; the CFU counts were significantly higher (p < 0.01) in depleted animals when compared with B. pertussis-infected control mice 21 days after challenge (Fig. 6⇑D). This was a consistent finding in six experiments. Furthermore, deaths were observed among the group of mice depleted of CD8α⁺ cells, a feature previously observed in mice lacking functional IFN-γ (6). We next investigated the impact of depleting CD8α⁺ cells (throughout infection) on CD4⁺ T cell cytokine production. Cells were isolated from the lungs of infected animals 30 days after challenge and intracellular cytokine staining was performed. IFN-γ-secreting and IL-10-secreting CD4⁺ T cells were detected in the lungs of mice infected with B. pertussis (Fig. 6⇑E). In contrast, IL-17-producing T cells were almost undetectable. The frequency of CD4⁺IFN-γ⁺ T cells infiltrating the lung 30 days after infection was reduced in CD8α-depleted compared with noninfected control and nondepleted B. pertussis-infected control mice (Fig. 6⇑E). Conversely, the percentages of CD4⁺IL-10⁺ T cells and CD4⁺IL-17⁺ T cells were enhanced in CD8α-depleted mice (Fig. 6⇑E). These findings are consistent with the data in Fig. 5⇑ showing that CD8α DC could activate IFN-γ-producing CD4⁺ T cells.

To determine the role of CD8α⁺ cells at different stages of infection, infected mice were injected with anti-CD8α-depleting Ab either on day −1 and + 1, from day −1 through to +19, or from +6 through to +19. Depleting CD8⁺ cells either early in infection or throughout the course of infection resulted in a significant enhancement of bacterial burden in the lungs (p < 0.05) when compared with infected control mice (Fig. 7⇓A). In addition, the number of cells in the perivascular spaces of lung tissue (isolated 30 days after infection) was greater in mice depleted of CD8α⁺ cells either during the early phase or throughout infection when compared with B. pertussis-infected undepleted control animals (Fig. 7⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Depleting CD8α⁺ cells during the first 48 h after B. pertussis challenge attenuates bacterial clearance and delays cellular infiltration. Mice were infected by exposure to an aerosol of live virulent B. pertussis. Mice received either 200 μg/mouse anti-CD8α-depleting Ab (anti-CD8α) or irrelevant IgG2b isotype control Ab (200 μg/mouse) i.p. on days −1 and +1 and on days −1, +1, 3, 7, 14, and 19 (−1 to +19). A, Bacterial colonization in B. pertussis-infected control mice and in CD8α-depleted infected animals was assessed by performing CFU counts on lung homogenates on day 21 after B. pertussis challenge. B, H&E staining of paraffin-embedded lung tissue sections (original magnification, ×40) from B. pertussis-infected control mice or infected mice that were treated on days −1 and +1 or on days −1, +1, 3, 7, 14, and 19 (−1 to +19) with anti-CD8α⁺ depleting Ab. Arrows show greater numbers of immune cells infiltrate the perivascular spaces in anti-CD8α-depleted compared with control B. pertussis-infected mice. C, The number of CD3⁺CD8⁺ T cells infiltrating the lungs of mice up to 28 days after B. pertussis challenge. *, p < 0.05 and **, p < 0.01 vs noninfected control mice. D, Mice were treated with anti-CD8α or irrelevant IgG2b isotype control Ab (200 μg/mouse) i.p. on days +6, 14, and 19 (+6 to +19). Bacterial colonization was assessed by performing CFU counts on lung homogenates on day 21 after B. pertussis challenge. *, p < 0.05 vs control nondepleted B. pertussis-infected mice. Results are representative of four separate experiments.

Since B. pertussis causes a predominant extracellular infection, it was assumed that CD8⁺ T cells would not play a protective role in immunity to B. pertussis, in particular during the acute phase of infection. Indeed, we have previously demonstrated that transfer of CD8⁺ T cells exacerbated infection with B. pertussis (5). Nevertheless, we examined the possibility that the enhanced infection seen in CD8-depleted mice resulted from depletion of CD8⁺ T cells. We first examined the infiltration of CD8⁺ T cells and found that they do infiltrate into the lungs of B. pertussis-infected mice, especially late in infection (Fig. 7⇑C). Depletion of CD8α⁺ cells from day 6 after B. pertussis challenge significantly (p < 0.05) enhanced bacterial clearance when compared with nondepleted infected control mice (Fig. 7⇑D). This observation points to a possible regulatory role for CD8⁺ T cells in immunity to B. pertussis. In contrast, the enhancement of infection observed following depletion of CD8⁺ cells early in infection along with the early appearance of CD8α⁺ DC suggest a protective role for this DC population in immunity to B. pertussis.

Transfer of in vitro-expanded FL-DC into infected mice enhances cellular infiltrate and promotes bacterial clearance from the lung

The results of the depletion experiments suggested that CD11c⁺CD8α⁺ DC contributed to antibacterial immunity in the lungs; we sought to confirm these findings using a cell transfer approach. We used FL to expand CD11c⁺CD24⁺CD45RA⁻ (FL-DC) from freshly isolated bone marrow cells and removed CD11c⁺CD24⁻CD45RA⁺ pDC by cell sorting (36). CD8α is not expressed on these cells in vitro; however, CD8α expression is up-regulated when they are transferred in vivo (37). The CD11c⁺CD24⁺CD45RA⁻ (FL-DC), which were routinely 95% pure (Fig. 8⇓A), were transferred 30 min after B. pertussis challenge. Transfer of FL-DC significantly enhanced bacterial clearance when compared with B. pertussis-infected controls. Mice that received the DC transfer had significantly lower (p < 0.05) CFU counts, particularly on day 14 after challenge (Fig. 8⇓B). To exclude the possibility that transfer of any DC subset could enhance bacterial clearance in B. pertussis-infected animals, experiments were performed where mice received either DC expanded in culture with GM-CSF and IL-4 (GM-DC) or FL-DC 30 min after infection. Only FL-DC transfer significantly accelerated bacterial clearance (p < 0.05) in infected animals (Fig. 8⇓C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Transfer of in vitro-expanded FL-DC before B. pertussis challenge enhances bacterial clearance. CD11c⁺CD24⁺CD45RA⁻ DC (FL-DC) were generated in vitro from freshly isolated bone marrow cells cultured for 8 days in DC medium supplemented with 300 ng/ml FL. A, Representative dot plots showing CD11c⁺CD24⁺CD45RA⁻ DC gated on viable CD11c⁺ cells, which were sorted before transfer in vivo. B, The course of infection in B. pertussis-infected control mice and in infected animals that received 0.5 × 10⁵ FL-DC (30 min before challenge) was followed by performing CFU counts on lung homogenates at fixed time points after the initial challenge. *, p < 0.05 vs control B. pertussis-infected mice. C, The bacterial load in B. pertussis-infected control mice and in infected animals that received either 0.5 × 10⁵ FL-DC or GM-DC before infection was assessed by performing CFU counts on lung homogenates on day 21 after infection. *, p < 0.05 vs control B. pertussis-infected mice. D, The number of CD3⁺CD4⁺, CD3⁺CD8⁺, B220⁺, and GR1⁺ cells in lung homogenates from B. pertussis-infected mice and infected animals that received FL-DC was determined 7, 14, and 21 days after infection. E, The number of CD11c⁺ and CD11c⁺CD8α⁺ DC infiltrating the lungs 7 days after challenge in control-infected mice and mice that received sorted FL-DC. Results are representative of three separate experiments. *, p < 0.05 vs control B. pertussis-infected mice.

Transfer of FL-DC was associated with an enhancement in the number of CD4⁺ T cells and B cells but not neutrophils infiltrating the lungs (Fig. 8⇑D). Although not statistically significant, this was a consistent finding over five separate experiments (data not shown). Finally, the numbers of CD11c⁺ or CD11c⁺CD8α⁺ DC in the lungs of infected mice were not significantly altered by transfer of FL-DC when examined 7 days after challenge (Fig. 8⇑E).

Blocking CD103 function in vivo impairs bacterial clearance

A number of studies have reported that DC, particularly in the gut and lung, express the α₄β_E integrin CD103 whose ligand, E-cadherin, is expressed by epithelial cells on mucosal surfaces (23, 25). Therefore, we examined expression of CD103 on DC in the lungs and CLN of B. pertussis-infected mice. We found a significant increase in the percentage and number of CD11c⁺CD8α⁺CD103⁺ DC in the CLN on day 4 after infection when compared with noninfected control mice (Fig. 9⇓, A and B). In contrast, significant numbers of CD11c⁺CD8α⁺CD103⁺ DC were detected in lung by day 6 after challenge (Fig. 7⇑, A and B). It is possible that CD11c⁺CD8α⁺ DC from the CLN require surface CD103 to track into the lung by interacting with E-cadherin on epithelial cells in the mucosal surface. We next investigated the possible role of CD103 in protective immunity to B. pertussis infection. Mice were injected once with an anti-CD103-blocking Ab 24 h before infection. It was not possible to specifically eliminate CD11c⁺CD8α⁺CD103⁺ cells, as only 40% of the CD103⁺ cells in either CLN or lung, during the acute phase of infection, express CD8α (data not shown). Nevertheless, blocking CD103 function with a single injection of anti-CD103 Ab had a dramatic impact on bacterial clearance, with a significant (p < 0.01) enhancement in CFU counts 21 days after B. pertussis challenge (Fig. 9⇓C). Inhibition of CD103 constrained cellular infiltration into the lungs 21 days after challenge (Fig. 9⇓D). These findings suggest that CD11c⁺CD103⁺ DC infiltrate the CLN and then the lungs in the early stages of infection with B. pertussis where they promote cellular immune responses that mediate clearance of the bacteria from the respiratory tract.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

Blocking CD103 function in vivo impairs bacterial clearance. Mice were infected with 2 × 10¹⁰ live, virulent B. pertussis. Infected lungs and CLN were removed and digested at various time points after infection. The percentage (A), based on the viable lymphocyte gate and total (B) number of CD11c⁺CD8α⁺CD103⁺ DC in CLN from noninfected control mice and in CLN and lungs from infected animals, was determined by flow cytometric analysis daily, up to day 7 after challenge with B. pertussis. *, p < 0.05 and **, p < 0.01 vs uninfected control mice. C and D, Mice were injected with 50 μg of either anti-CD103 or rat anti-HRP IgG1 Ab (control) i.p. 24 h before infection to block CD103 function in vivo. Bacterial clearance (C) and cellular infiltration into the lung (D) was examined on day 21 after B. pertussis infection. Results are representative of two separate experiments. **, p < 0.01 vs nondepleted B. pertussis-infected control mice.