Introduction Streptococcus pneumoniae (pneumococcus) is an “extracellular” pathogen, generally considered to be killed by phagocytic ingestion, which is facilitated by opsonic antibodies. The success of anti-pneumococcal serum therapy using passive transfer of serotype-specific antibodies [1] and of vaccinations based on purified or conjugated capsular antigens [2],[3] clearly shows that anticapsular antibodies protect humans against pneumococcal colonization and disease. There is good epidemiologic evidence for the importance of such immunity in certain common serotypes [4],[5]. However, we and others have found that factors other than anticapsular antibodies may play a role in the natural development of protection against pneumococcal colonization and disease. First, the reduction in pneumococcal disease incidence after the first year of life occurs simultaneously for both rare and common serotypes, suggesting the acquisition of one rather than many individual immune responses [6]. Similarly, the duration of carriage of many serotypes declines steeply between the first and second birthdays for many serotypes [7]. Since experience with conjugate vaccines has suggested that anticapsular antibodies reduce incidence of carriage but leave duration unaffected [8], this observation also suggests a mechanism of acquired immunity other than anticapsular antibodies. Moreover, the declines in carriage duration and invasive disease incidence precede by several years the detection of naturally-acquired anticapsular antibody in most children [6],[7]. Experimental [9],[10] and observational [4],[11] studies in adults have found little or no evidence that higher anticapsular antibody concentrations are associated with greater protection from colonization. Pneumococci also express non-capsular antigens common among serotypes, and certain of these have been found to elicit antibodies with protective potential in animal models. The role of such antibodies in human immunity has been evaluated [12],[13],[14],[15],[16],[17]. Surprisingly however, recent studies have shown that immunity in mice to pneumococcal colonization acquired from prior exposure to live bacteria [18] or a killed, whole-cell vaccine [WCV, consisting of killed pneumococcal whole cell antigen (WCA) with cholera toxin (CT) as an adjuvant] [19] is independent of antibodies of any specificity, and clearance of longstanding carriage in previously unexposed animals can likewise be antibody-independent [20]. Immunity had been shown to be dependent on the presence of CD4+ T cells at the time of challenge [18],[19], but the co-participation of specific immune factors other than antibody was not ruled out. Here we show that intranasal immunization with the WCV confers protection against experimental pneumococcal colonization via the chemoattractant and neutrophil activating cytokine IL-17A, in a neutrophil-dependent fashion. Methods were devised to assay expression of IL-17A in vitro using peripheral blood samples. IL-17A expression by peripheral blood of WCV-immunized mice is highly correlated with subsequent protection against colonization, and expression by human cells, including those from adults and children, can be shown as well. Finally, we developed a surface phagocytosis assay with which we show that IL-17A enhances pneumococcal killing by human polymorphonuclear cells in the absence as well as presence of opsonins. The data indicate the possibility that IL-17A responses play a role in naturally-acquired immunity to pneumococcus in humans and that assay of this cytokine in vitro may assist in the evaluation of certain candidate pneumococcal vaccines that target mucosal colonization. Results Prior exposure of mice to killed or live pneumococci reduces the duration of experimental pneumococcal carriage The duration of carriage was followed after intranasal challenge with serotype 6B pneumococci 4 weeks post-exposure to WCV. Both WCV-vaccinated and control mice immunized with CT alone were colonized one day after challenge. In mice immunized with WCV however, carriage became significantly reduced after 4 days compared to controls given cholera toxin (CT) adjuvant alone (median density of colonization on day 4 in WCV- vs. CT-immunized mice 251 vs. 3720 cfu/nasal wash, P = 0.029 by Mann-Whitney U test) and was undetectable by day 6 (0/4 WCV-immunized mice had detectable colonies on day 6 vs. 4/4 mice that received CT, P = 0.029 by Fisher's Exact test, Figure 1A). A similar differential was observed in mice that had been repeatedly exposed to live pneumococci vs. saline controls: the density of colonization became significantly different by day 4 after inoculation (Figure 1A). By day 6, similar to what we observed in WCV-immunized mice, 0/4 mice exposed to live pneumococci had detectable colonies compared to 4/4 saline controls (P = 0.029 by Fisher's exact test). When data from all time points were compared, mice immunized with WCV or exposed to live pneumococci had a significantly shorter time to clearance compared to their respective CT or saline controls (P = 0.0001 for comparison of WCV vs. CT and P = 0.004 for comparison of live exposure vs. saline). Thus the protection by prior pneumococcal exposure involves not immediate blockage of colonization but rather an accelerated clearance over days. Subsequent studies compared WCV-vaccinated with control animals 7 days after the intranasal challenge. 10.1371/journal.ppat.1000159.g001 Figure 1 Duration of carriage and effect of adoptive transfer following immunization with killed or live pneumococci. A. Effect of intranasal immunization with WCV or live exposure upon density of pneumococcal colonization in C57BL/6 mice. Density of colonization in mice immunized with WCV vs. CT alone or repeatedly exposed to live pneumococcal strain 0603 vs. saline alone at various time points (n = 4 per time point) following challenge. By day 4, both the incidence and density of carriage were significantly lower in mice immunized with WCV or exposed to live pneumococcus compared to mice immunized with CT or saline, respectively. * P 0.5 vs. CT controls for % colonized mice or density of colonization, Figure 2A). It is noteworthy that IL-17AR-knockout mice in the CT control group had, on average, a ten-fold greater density of colonization than the corresponding IFN-γ or IL-4 deficient mice, suggesting that IL-17A may also be involved in resistance to colonization in naïve mice. 10.1371/journal.ppat.1000159.g002 Figure 2 Role of T-helper-subset-associated cytokines in protection from nasopharyngeal colonization. A. Mice defective in IFN-γ, IL-4 or IL-17A receptor were immunized as described, then challenged with pneumococcal strain 0603. Mice with IFN-γ or IL-4 deficiency were significantly protected by WCV (P 0.5 vs. CT). Dashed line represents the lower limit of detection of bacterial colonization. B. Expression of IL-17A from splenocytes of WCV-immunized mice. Cultured splenocytes from mice immunized with WCV (black columns) or CT alone (white columns) were stimulated for 72 hours with medium alone, Concanavalin A (5 µg/ml), WCA (10 µg dry weight) after which IL-17A production was measured by ELISA. Significantly more IL-17A was expressed following WCA stimulation of WCV-immunized vs. CT-immunized mice, although response to concanavalin A was similar. C. Effect of CD4+ T cell depletion upon IL-17A expression from splenocytes. Splenocytes (without or with CD4+ T cell depletion) from mice immunized with WCV were stimulated for 72 hours with medium alone or WCA after which IL-17A was measured by ELISA. IL-17A expression in splenocytes following WCA stimulation was significantly higher in the presence of CD4+ T cells compared to stimulation with medium alone or when CD4+ T cells were depleted. Repletion of CD4+ T cells restored the response. ** P 90% depletion of neutrophils in most mice, although variability was observed. Because of this variability, peripheral neutrophil counts were determined at the time of euthanasia and correlated with the number of recovered pneumococci from that animal. Measurement of IL-17A secretion by splenocytes Cellular suspensions of splenocytes were obtained by passing spleens from immunized or control mice through a 70-µm cell strainer (BD Biosciences, Bedford, MA). After washing and removal of red blood cells by hemolysis, cells were plated into 24-well tissue culture plates at a concentration of 5×106 cells/well in 500 µl of DMEM/F12 with L-glutamine supplemented with 10% fetal calf-serum, 50 µM 2-mercaptoethanol (Sigma), and 10 µg/ml ciprofloxacin. Following 72-hour stimulation with concanavalin A (5 µg/ml, Sigma) or WCA (equivalent to 106 cfu/ml), supernatants were collected following centrifugation and stored at −80°C until analyzed by ELISA for IL-17A concentration (R&D Systems, Minneapolis, MN). Supernatants were analyzed in duplicate and read against a standard, following directions provided by the manufacturer. For CD4+ T cell depletion, splenocytes were harvested as described above. CD4+ T cells were depleted from half of each spleen by magnetic bead selection (Miltenyi Biotec, Auburn, CA) following instructions by the manufacturer. Flow cytometry confirmed removal of >95% CD4+ T cells (data not shown). Cells were seeded at the same concentration as described above (5×106 cells/well). In some cases, we repleted CD4+ T cells from depleted splenocytes, by adding 106 CD4+ T cells in the relevant wells. Intracellular staining for IL-17A Splenocytes were harvested, seeded, and stimulated with medium or WCA (10 µg/ml) as described above. Twenty-four later, monensin (BD GolgiStop, BD Biosciences) was added as per the manufacturer's instructions and cells were harvested 12 hours later. Cells were washed, stained with anti-CD4+ antibody (antiCD4+-PE, BD Biosciences) in the presence of Fc block, permeabilized with Perm/Wash buffer (BD Biosciences), and incubated with antimouse IL17A Alexa Fluor-647 (eBioscience) for 30 minutes. Intracellular cytokine staining for IL-17A was compared in CD4- or CD4+ cells in medium alone or following stimulation with WCA. Samples were analyzed on a Cytomation MoFlo (Beckman Coulter, Fullerton, CA), and results analyzed with Summit Version 4.3 (Dako, Fort Collins, CO). Measurement of IL-17A secretion by NALT NALT was harvested from immunized and control mice as described [51]. Mice were euthanized humanely, bled via intracardiac puncture to avoid blood contamination, and placed on a dissection board. The mouth was opened wide to expose the palate, which was cut carefully, so that the strips of NALT could be easily peeled off. These strips of cells were collected in medium (DMEM/F12 with L-glutamine supplemented with 10% fetal calf-serum, 50 µM 2-mercaptoethanol (Sigma), and 10 µg/ml ciprofloxacin) on ice. Cells were passed through a 70 µm strainer as described above and plated at 3×105 cells/well in a 96-well tissue culture plate in a total volume of 100 µl. Cells were stimulated with medium with or without added WCA (10 µg/ml) for a total of 3 days, after which supernatants were collected and assayed for IL-17A concentration by ELISA as above. Measurement of IL-17A secretion by whole blood For whole blood assays, blood of mice or humans at a final concentration of 10% was incubated in DMEM/F12 with L-glutamine supplemented with 10% fetal calf-serum, 50 µM 2-mercaptoethanol (Sigma), and 10 µg/ml ciprofloxacin in the absence or presence of killed pneumococcal antigen (corresponding to 107 cfu/ml for mice and 106 cfu/ml for human samples). Supernatants were collected after 6 days and the concentration of IL-17A measured as above for mice and, for human samples, by IL-17A ELISA (eBioscience Inc, San Diego, CA). Human subjects and samples For peripheral blood, samples were obtained at Children's Hospital Boston (for healthy adult volunteers) or from Cambridge Health Alliance, Cambridge, MA (for parturient women or umbilical cord) after written informed consent had been obtained. The studies were approved by the Children's Hospital Boston and Cambridge Health Alliance research ethics committees. For tonsillar specimens, tonsils were obtained from children who were 2 to 12 years old (median age, 5 years), were undergoing tonsillectomy for hypertrophy, and were otherwise healthy at Bristol Royal Hospital for Children, Bristol, United Kingdom. Patients who were immunized against pneumococcus previously, who had received antibiotics within 2 weeks of the operation or steroids, or who had an immunodeficiency or serious infection were excluded. The study was approved by the South Bristol local research ethics committee and written informed consent was obtained in all cases. Agar surface phagocytic killing without opsonins This assay approximates the “surface phagocytosis” described by Smith and Wood [52]. Neutrophils were isolated from heparinized blood by density gradient centrifugation (Histopaque, Sigma) following manufacturer's instructions. Neutrophils were washed extensively then resuspended in Hanks' Balanced Solution (+ Ca2+ and Mg2+) with 0.2% bovine serum albumin (Sigma), then co-incubated for 30 minutes at 37°C with recombinant human IL-17A (R&D Biosystems) at different concentrations. In some experiments, the cells were harvested by centrifugation and the supernatant collected, to examine whether the potentiating effect of IL-17A could be detected with the supernatant alone. Between 8–10 replicates of 10 µl of a bacterial suspension containing on average 100 cfu of strain 0603 were plated onto blood agar and the fluid allowed to adsorb into the agar for 15 min; 15 µl of the neutrophil suspension was overlaid and allowed to adsorb; the plates were incubated at 37°C with 5% C02 overnight after which colonies were counted. Phagocytic killing in suspension with suboptimal opsonization Neutrophils were isolated from whole blood as described above, washed twice with cold Hanks Balanced Salt Solution (HBSS-) (Mediatech, Herndon, VA), and resuspended to a final concentration of 6×106 cells/ml in cold HBSS containing calcium and magnesium (HBSS+) (Cellgro Mediatech, Herndon, VA) then held on ice until used. Cell counts were determined on a standard hemocytometer by counting viable cells (as determined by an absence of blue staining in the presence of Trypan Blue (Cellgro Mediatech, Herndon, VA)). S. pneumoniae (strain 0603 [49]) was diluted in HBSS+ to a final concentration of 5×104 bacteria/ml and incubated with antibodies to pneumococcal polysaccharide (Bacterial Polysaccharide Immune Globulin, BPIG-8, a kind gift of Dr. George Siber, consisting of concentrated IgG obtained from serum of adult volunteers immunized with pneumococcal, Haemophilus and meningococcal polysaccharide vaccines [27]) diluted in HBSS+. The reaction was incubated at 37°C for 15 minutes rotating at 200 RPM to promote bacterial opsonization. After bacterial opsonization, the opsonophagocytic killing reaction was initiated with the addition of baby rabbit complement (Pelfreez Biologicals, Rogers, AR) and neutrophils (ratio of 1∶200 bacteria∶cells) with or without recombinant human IL-17A (R&D Systems, Minneapolis, MN) at 0.01, 0.1 or 1 µg/ml. A 1∶1600 dilution of BPIG was chosen to give sub-optimal bacterial killing (<50% killing when compared to the same conditions without BPIG) in the presence of complement and neutrophils. The opsonophagocytic killing assay was performed in a 96-well round-bottom plate (Thermo Fisher Scientific, Waltham, MA) at 37°C for 90 minutes rotating at 200 RPM. After incubation, the opsonophagocytic reaction was diluted two fold and aliquots of each reaction were plated on blood agar plates then incubated at 37°C with 5% CO2 overnight. Isolation and culture of tonsillar mononuclear cells Mononuclear cells were isolated by using methods described previously [53],[54]. Tonsillar MNC were washed in sterile phosphate-buffered saline (PBS) and resuspended at a concentration of 4×106 cells/ml in RPMI medium containing HEPES, 2 mM glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and 10% fetal bovine serum (Sigma, Dorset, United Kingdom). Cells were cultured in 96-well culture plates (Corning Inc, Corning, NY), and cell culture supernatants were collected at predetermined times and stored at −70°C until assays for human IL-17A were performed by sandwich ELISA (R&D Biosystems). Statistical analysis Incidence of carriage was compared by Fisher's exact test and colonization density in challenged mice was compared by the Mann-Whitney U test. Statistical significance of the difference between time-to-clearance curves was assessed as follows. For each group i (i = WCV, CT, live, or naïve), the proportion of mice cleared at each time point t, pi (t), was calculated. Using the max-min formula for isotonic regression [55], these proportions were smoothed to assure they were nondecreasing in t, yielding smoothed proportions qi (t). Then, a test statistic was calculated to quantify the distance between the smoothed curves for two groups (e.g., WCV vs. CT): . The significance level of this test statistic was estimated by permuting the group identifiers of the cleared mice at each time point, fixing the total number of mice in each group and the total number cleared at each time point. 100,000 replicates of the permuted data were obtained, and T was calculated for each. The p value was calculated as the fraction of these 100,000 permutations having a test statistic strictly less than that calculated for the data. The correlation between neutrophil count or IL-17A concentration and colonization density was determined by Spearman rank correlation. The effect of increasing IL-17A concentrations on enhancing killing of pneumococcus was assessed by Wilcoxon matched pairs test. For all comparisons, P<0.05 was considered to represent a significant difference.
