Listeriolysin o is strongly immunogenic independently of its cytotoxic activity

Abstract

The presentation of microbial protein antigens by Major Histocompatibility Complex (MHC) molecules is essential for the development of acquired immunity to infections. However, most biochemical studies of antigen processing and presentation deal with a few relatively inert non-microbial model antigens. The bacterial pore-forming toxin listeriolysin O (LLO) is paradoxical in that it is cytotoxic at nanomolar concentrations as well as being the source of dominant CD4 and CD8 T cell epitopes following infection with Listeria monocytogenes. Here, we examined the relationship of LLO toxicity to its antigenicity and immunogenicity. LLO offered to antigen presenting cells (APC) as a soluble protein, was presented to CD4 T cells at picomolar to femtomolar concentrations- doses 3000-7000-fold lower than free peptide. This presentation required a dose of LLO below the cytotoxic level. Mutations of two key tryptophan residues reduced LLO toxicity by 10-100-fold but had no effect on its presentation to CD4 T cells. Thus there was a clear dissociation between the cytotoxic properties of LLO and its very high antigenicity. Presentation of LLO to CD8 T cells was not as robust as that seen in CD4 T cells, but still occurred in the nanomolar range. APC rapidly bound and internalized LLO, then disrupted endosomal compartments within 4 hours of treatment, allowing endosomal contents to access the cytosol. LLO was also immunogenic after in vivo administration into mice. Our results demonstrate the strength of LLO as an immunogen to both CD4 and CD8 T cells.

Figure 1. Reduced toxicity of LLO tryptophan mutants.

A. LLOWW was tested for hemolytic activity on human red blood cells as in . LLO was added at the indicated concentrations for 60 min. at 37°C then, hemoglobin release was measured at OD570. B. C3.F6 cell line was incubated with the indicated concentrations of LLOWT or LLOWW for 1 hr. at 37°C. Cells were stained with trypan blue and percentage of live cells (trypan negative) was determined. C. An ovalbumin-responsive primary T cell line (5×10⁴) was incubated for 16 hr. with BMDC (5×10⁴) and the indicated concentrations of ovalbumin (antigen) and LLO (toxin). T cell responses were measured by IL-2 production. D-H. BMDC (1×10⁶) were treated with the indicated LLO variants at the given concentrations for 6 hours. Cell death was measured by (D–E) Annexin V/7-AAD staining, (F) JC-1 dye, (G) Red-VAD, or (H) Red-DEVD. All cell death analyses were performed by flow cytometry. Cells were gated for CD11c⁺ events then the percentage of affected CD11c⁺ cells was plotted. Results in (A–C) are representative of two independent experiments. Bars in (D–H–F) represent the mean±S.D. of at least 2 independent experiments performed in duplicate or triplicate (n = 5–7).

Figure 2. Release of vesicular dextrans into the cytosol.

A–I. BMM (1.25×10⁵) were adhered to 12 mm coverslips and treated with FITC-dextran for 30 min then either left untreated (D,G) or treated with LLOWT (E,H) or LLOWW (F,I) for 4 hours. A. The number of cells containing punctate staining was determined by counting ∼100 individual cells for each treatment. A cell was considered to have puncta if it had any visible discrete spots, as in panel G. Cells were considered to have no puncta if the green staining was homogeneous, as in panel H. B–C. ImageJ analysis of the number of particles and size of particles per field was determined as detailed in the materials and methods. Bars represent the mean+/−S.D. particles per field (B) or pixel size of particles (C) for three independent fields with 10–24 cells per field. D–F. Representative epiluminescence images taken following treatment with FITC-Dextran and then either left untreated (D) or treated with 1 nM LLOWT (E) or 10 nM LLOWW (F). The nucleus (DAPI) was false colored blue and FITC-Dextran was false colored green. G–H. Representative laser scanning micrographs of FITC-dextran loaded BMM either untreated (G) or treated with 1 nM LLOWT (H) or 10 nM LLOWW (I). Scale bars in represent 50 µm (D–F) or 10 µm (G–I). Microscopy is representative of 2 independent experiments performed in duplicate.

Figure 3. Enhanced presentation of an LLO protein-derived CD4 T cell epitope.

A–B. 5×10⁴ I-A^b-restricted LLO(190–201)-specific hybridomas 57-3 (A) or LL73 (B) were incubated for 16 h with 10⁵ PEC (A) or 5×10⁴ BMDC (B) from C57BL/6 mice and various concentrations of either LLOWT, LLO(190–201), or LLO(187–205). C. BMDC (10⁵) were infected with the indicated MOI of L. monocytogenes EGD strain as in and 16 hrs. later the response of 57-3 (5×10⁴) was determined. D. LLO was treated with endoproteinase AspN or buffer for ∼16 hr. at 37°C then incubated with for ∼16 hr. with 10⁵ irradiated C57BL/6 splenocytes and 57-3 (5×10⁴). E, Tlr4^−/−, (F) Myd88^−/− (KO), or (E–F) C57BL/6J (WT) irradiated splenocytes (10⁵) were treated with LLOWT or LLO(190–201) and (E) 57-11 or (F) 57-3 T cell hybridomas (5×10⁴). Following all incubations, T cell hybridoma activation was determined by CTLL-2 assay. Points indicate the mean±S.D.

Figure 4. Antigenicity of LLO to CD4 T cells is independent of cytotoxicity.

A–B. C57BL/6 BMDC (5×10⁴) and the LLO-specific hybridomas (5×10⁴) (A) 57-3 or (B) LL73 were incubated with the indicated concentrations of LLO variants or the peptides for ∼16 hrs and T cell activation was determined. C. The half-maximal dose of antigen required to activate T cell hybridomas in 62 independent assays was calculated using non-linear regression fitting of the dose-response plots. The plot shows the log₁₀ of the computed half-maximal activation value. Each dot represents and individual T cell assay. The geometric mean is given for each column. p values were calculated by student T test following validation of normal distribution through D'Agostino & Pearson omnibus normality test. There was a statistical significance in the reduced dose requirement for presentation of LLOWT, LLOW492A, and LLOWW when compared to LLO(190–201) or LLO(185–207) peptides (p<0.0001). D. B10.BR PEC (10⁵) and the 3A9 T cell hybridoma specific for HEL(48–62) (5×10⁴) were incubated with either HEL(48–62) peptide or LLO(HEL48–65) fusion protein for ∼16 hr. and tested for T cell activation by CTLL-2 assay. For (A), (B), and (D) dots represent the mean±S.D. for triplicate values at each antigen dose.

Figure 5. Binding, uptake and catabolism of radiolabeled LLO.

A. BMM (10⁶) were treated with ∼150 ng of ¹²⁵I-LLO or HEL for 30 min. on ice. The amount of CPM bound to cells or found in the supernatant was measured and a percent of total bound to cells was calculated. B. ¹²⁵I-LLO, LLOW492A, or HEL (∼50 ng) were given to BMM for 15, 30, or 60 min. and the amount of incorporation of radioactivity was measured. Values are plotted as a percent of total input radioactivity that was taken by the cells. C–D. BMM (10⁶) were treated with 50 ng ¹²⁵I labeled LLOWT (C) or LLOW492A (D) for the indicated times. Then, TCA precipitation was performed and the amount of total radioactivity present in the precipitate (ppt) or supernatant (sol) was measured. The ppt fraction contains large polypeptides and represents the non-catabolized fraction and the sol fraction contains small peptides and free amino acids and represents the catabolized fraction. Similar uptake and catabolism results were obtained with BMDC. Each experiment in (A–D) is representative of at least two independent experiments performed in triplicate. E–F. BMM (10⁵) were incubated with LLO(190–201 peptide), LLOWT, or LLOW492A at the indicated antigen concentrations for 15 (E), 30 (F), 60 (G) or 120 (H) min. BMM were fixed with PFA, quenched with lysine, washed and then LLO specific 57-3 T cell hybridoma (5×10⁴) was added for ∼16 hr. T cell activation was determined by CTLL-2 assay. Experiment is representative of 2 independent experiments performed in triplicate. Similar results were obtained with BMDC.

Figure 6. Antigenicity of LLO to a CD8 T cell was not dependent on cytotoxicity.

A–B. H-2K^d-restricted LLO(91–99)-specific hybridoma 206.15 (5×10⁴) were incubated for ∼16 hr. with 10⁵ PEC (A) or 5×10⁴ BMDC (B) from BALB/c mice and various concentrations of either LLOWT, LLOW492A, LLOWW, or LLO(91–99). C. 206.15 (5×10⁴) was incubated with the BMDC (10⁵) infected at the indicated MOI of L. monocytogenes EGD strain . Following all incubations, T cell hybridoma activation was determined by CTLL-2 assay. All T cell activation assays are representative of at least two independent experiments performed in triplicate. Points indicate the mean±S.D.

Figure 7. ELISpot response following immunization with LLO or LLO peptide.

A–B. C57BL/6 mice were immunized in the footpad with 250 picomoles of either LLOWT, LLOWW, or LLO(190–201) peptide emulsified in incomplete Freund's adjuvant. After 7 days, draining lymph nodes were isolated and the number of IFNγ (A) and IL-2 (B) producing cells was determined by ELISpot. Recall antigen was either LLOWW at 10 nM or no antigen. Bars represent the mean+/−SEM for 2–6 mice per group over three independent experiments. Analysis of the data using D'Agostino & Pearson omnibus normality test followed by student t test revealed that the increase in IFNγ production following immunization with LLOWW was significant when compared to both LLOWT or LLO(190–201) immunization (p = 0.0012 for LLOWT vs. LLOWW and p<0.0001 for LLOWW vs. LLO(190–201).