Pneumolysin induces platelet destruction, not platelet activation, which can be prevented by immunoglobulin preparations in vitro

Abstract

Community-acquired pneumonia by primary or superinfections with Streptococcus pneumoniae can lead to acute respiratory distress requiring mechanical ventilation. The pore-forming toxin pneumolysin alters the alveolar-capillary barrier and causes extravasation of protein-rich fluid into the interstitial pulmonary tissue, which impairs gas exchange. Platelets usually prevent endothelial leakage in inflamed pulmonary tissue by sealing inflammation-induced endothelial gaps. We not only confirm that S pneumoniae induces CD62P expression in platelets, but we also show that, in the presence of pneumolysin, CD62P expression is not associated with platelet activation. Pneumolysin induces pores in the platelet membrane, which allow anti-CD62P antibodies to stain the intracellular CD62P without platelet activation. Pneumolysin treatment also results in calcium efflux, increase in light transmission by platelet lysis (not aggregation), loss of platelet thrombus formation in the flow chamber, and loss of pore-sealing capacity of platelets in the Boyden chamber. Specific anti-pneumolysin monoclonal and polyclonal antibodies inhibit these effects of pneumolysin on platelets as do polyvalent human immunoglobulins. In a post hoc analysis of the prospective randomized phase 2 CIGMA trial, we show that administration of a polyvalent immunoglobulin preparation was associated with a nominally higher platelet count and nominally improved survival in patients with severe S pneumoniae-related community-acquired pneumonia. Although, due to the low number of patients, no definitive conclusion can be made, our findings provide a rationale for investigation of pharmacologic immunoglobulin preparations to target pneumolysin by polyvalent immunoglobulin preparations in severe community-acquired pneumococcal pneumonia, to counteract the risk of these patients becoming ventilation dependent. This trial was registered at www.clinicaltrials.gov as #NCT01420744.

Conflict of interest statement

Conflict-of-interest disclosure: A.G. reports grants and nonfinancial support from Aspen, Boehringer Ingelheim, Merck Sharp & Dohme (MSD), Bristol Myers Squibb (BMS), Bayer Healthcare, and Instrumentation Laboratory; personal fees from Aspen, MSD, Macopharma, BMS, Chromatec, and Instrumentation Laboratory; and nonfinancial support from Portola, Ergomed, and Biokit, outside of the submitted work. Charité (G.N. and M. Witzenrath) receives funding for research from Biotest AG. C.H., S.W., and J.S. are employees of Biotest AG. The remaining authors declare no competing financial interests.

© 2020 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

CD62P expression of platelets by pneumolysin is caused by pore formation. Washed platelets of a defined set of 6 donors were incubated with various concentrations of pneumolysin (Ply). CD62P was detected by flow cytometry using antibodies against CD62P (P-selectin). The data are presented as GMFI of the positive gated events multiplied with the percentage of positive gated events in the dot plots. (A) PBS (gray) and phospholipase C (PLC; gray) from Staphylococcus aureus known to not activate platelets were used as negative controls and 20 µM TRAP-6 (gray) as positive control. (B) Pneumolysin (red; ng/mL) caused CD62P expression and dose-dependently inhibited an additional response to TRAP-6 (black). (C) Polyvalent human immunoglobulins (human IgG; Privigen) neutralized the effect of pneumolysin (pneumolysin plus human IgG = light blue) (to enable comparison with the experiments without immunoglobulins, the data are shown here, although they are presented in the text at the end of “Results”). (D) PneumolysinC428G without lytic activity (brown) did not activate platelets or impaired the response to TRAP-6 and (E) pneumolysinW433F with ∼10% lytic activity (purple) had a very minor effect only at 300 ng/mL. (F) Visualization of pore formation in the platelet membrane by pneumolysin by scanning electron microscopy. Platelets are altered in their shape and formed vesicles but not pseudopodias. At the left side, a platelet with pores can be seen. Inset, a higher magnification of the platelet indicating a pore by an arrow.

Figure 2.

Loss of platelet function due to pneumolysin is prevented by immunoglobulins. (A) Prior to pneumolysin treatment intracellular Ca2+ of washed platelets was labeled with Fluo-4-AM for 30 minutes. After incubation with pneumolysin, the kinetics of Ca2+ release was measured and values are given as fold change compared with NaCl control. Different concentrations of pneumolysin (Ply) are color coded: 300 ng/mL (red); 30 ng/mL (orange); 3.0 ng/mL (light blue). PneumolysinC428G without lytic activity (brown) pneumolysinW433F with ∼10% lytic activity (blue) did not cause Ca2+ release. (B) Platelet aggregation is typically directly proportional to an increase in light transmission. Only pneumolysin 300 ng/mL (red) and 30 ng/mL (orange) induced an increase in light transmission, but platelets where no longer responsive to 20 µM TRAP-6. Light transmission did not change by addition of buffer, pneumolysin 3 ng/mL, or the mutant pneumolysins, but platelets were still responsive to 20 µM TRAP-6. (C-D) Polyvalent human immunoglobulin (human IgG (Privigen); 1 mg/mL; green), polyclonal rabbit anti-pneumolysin (10 µg/mL; orange) and a monoclonal mouse anti-pneumolysin antibody (7.5 µg/mL; blue) prevented the effects of pneumolysin (300 ng/mL; red) in calcium influx (C) and platelet aggregation (D). In the presence of these immunoglobulins platelets became again responsive to 20 µM TRAP-6 (to enable comparison with the experiments without immunoglobulins, the data are shown here, although they are presented in the text at the end of “Results”).

Figure 3.

Fluorescence microscopy of pneumolysin-treated platelets. (A) Pneumolysin-treated platelets were stained for F-actin (green) and CD62P (magenta). Platelets were not permeabilized, with the exception of the Triton X-100 control. Insets, Single platelets at higher magnification and the line used for measuring fluorescence intensities shown in panel B. In the presence of 3.0 and 30 ng/mL pneumolysin, intracellular staining of CD62P and α-tubulin become visible. At 200 ng/mL pneumolysin, vesicles staining strongly for pneumolysin surround the platelets (compare Figure 1F). (B) Staining pattern of CD62P throughout single cells treated with pneumolysin was quantified to distinguish between cytoplasmic and only surface-associated CD62P staining. The pattern indicates that CD62P is stained intracellularly and not extracellularly. The different concentrations of pneumolysin used are color coded: 3.0 ng/mL (blue), 30 ng/mL (orange), 300 ng/mL (green). (C) Orthogonal views of confocal Z-stacks and 3D isosurface rendering of pneumolysin-treated platelets stained for F-actin (green) and CD62P (magenta). It shows distinct intracellular accumulation of anti-CD62P antibody in platelets treated with different concentration of pneumolysin and membrane permeabilization with Triton X-100 and surface expression of CD62P upon TRAP-6 stimulation.

Figure 4.

Pneumolysin induces platelet death. (A) Kinetics of platelet viability. PBS was used as viability control and Triton X-100 to induce platelet death. Pneumolysin in increasing concentrations induced platelet death measured by reduced substrate turnover. (B) Platelet viability was maintained in the presence of polyvalent human immunoglobulin (human Ig [Privigen]; 1 mg/mL), polyclonal rabbit anti-pneumolysin (rabbit pAb; 10 µg/mL), or a monoclonal mouse anti-pneumolysin antibody (mouse mAb; 7.5 µg/mL) despite a high concentration of pneumolysin (300 ng/mL). (C) Thrombus formation on collagen in a flow chamber in the absence of pneumolysin was monitored by image acquisition at an interval of 10 seconds by fluorescence microscopy at a shear stress of 1000 s−1. In the presence of pneumolysin, thrombus formation was impaired. (D) Thrombus formation in the presence of pneumolysin was restored by polyvalent immunoglobulin. Human Ig (Privigen) alone had no effect on thrombus formation (supplemental Figure 10). (E) Quantification of the percentage of surface area covered over time by thrombi in the presence of pneumolysin in different concentrations or nonactive pneumolysin mutants. Different concentrations of pneumolysin (Ply) are color coded: 300 ng/mL (red); 30 ng/mL (orange); 3.0 ng/mL (green). PneumolysinC428G without cytolytic activity (brown); pneumolysinW433F with ∼10% cytolytic activity (blue). (F) Quantification of the effect of polyvalent immunoglobulin (human Ig (Privigen); 1 mg/mL) on restoring thrombus formation in the presence of pneumolysin (300 ng/mL). RLU, relative luminescence unit.

Figure 5.

Sealing of Transwell membranes by platelets is impaired by pneumolysin. Platelets seal Transwell membranes with 3-µm pores. This was impaired in the presence of pneumolysin but not the pneumolysin mutants as indicated by the flow through of BSA-FITC to the lower chamber measured by fluorescence intensity (to enable comparison with the experiments without immunoglobulins, the data are shown here, although they are presented in the text at the end of “Results”). *P < .05; **P < .01; ***P < .001.

Figure 6.

Neutralizaton of pneumolysin by trimodulin and relevance for severe community-acquired pneumonia patients. (A) In vitro pneumolysin neutralization assay measuring free hemoglobin as marker for cell lysis, using erythrocytes and different concentrations of trimodulin as indicated. Shown are results from 3 repeated measurements (mean plus or minus standard deviation.). (B) Platelet counts of patients in the CIGMA study with confirmed S pneumoniae infection were obtained before (pre), during (days 2-5), and after (days 6, 7, and 14) treatment with trimodulin or placebo. Pretreatment values were obtained from n = 15 (trimodulin group) and n = 18 (placebo group) and day 14 values obtained from n = 11 (trimodulin group, 1 missing value) and n = 12 patients (placebo group), respectively. (C) The 28-day mortality rate in patients with severe, confirmed S pneumoniae lung infection (n = 15 in trimodulin group, n = 18 in placebo group) was nominally lower in the trimodulin group compared with the placebo group (2 of 15 [13.3%] vs 7 of 18 [38.9%]). Due to small patient numbers, no statistical analysis has been performed. NC, negative control (PBS); PC, positive control (1% Triton X-100).