Engagement of NKG2D on bystander memory CD8 T cells promotes increased immunopathology following Leishmania major infection

Abstract

One of the hallmarks of adaptive immunity is the development of a long-term pathogen specific memory response. While persistent memory T cells certainly impact the immune response during a secondary challenge, their role in unrelated infections is less clear. To address this issue, we utilized lymphocytic choriomeningitis virus (LCMV) and Listeria monocytogenes immune mice to investigate whether bystander memory T cells influence Leishmania major infection. Despite similar parasite burdens, LCMV and Listeria immune mice exhibited a significant increase in leishmanial lesion size compared to mice infected with L. major alone. This increased lesion size was due to a severe inflammatory response, consisting not only of monocytes and neutrophils, but also significantly more CD8 T cells. Many of the CD8 T cells were LCMV specific and expressed gzmB and NKG2D, but unexpectedly expressed very little IFN-γ. Moreover, if CD8 T cells were depleted in LCMV immune mice prior to challenge with L. major, the increase in lesion size was lost. Strikingly, treating with NKG2D blocking antibodies abrogated the increased immunopathology observed in LCMV immune mice, showing that NKG2D engagement on LCMV specific memory CD8 T cells was required for the observed phenotype. These results indicate that bystander memory CD8 T cells can participate in an unrelated immune response and induce immunopathology through an NKG2D dependent mechanism without providing increased protection.

Figure 1. LCMV memory T cells migrate to leishmanial lesions and upregulate gzmB expression.

B6 mice received CD45.1+ P14 cells (P14 chimeras) and were infected with LCMV for 5 days (effector time point) or 30 days (memory time point). At the indicated time post LCMV infection, splenocytes were harvested from naïve P14 mice or P14 chimeras and the numbers of P14 CD8 T cells were quantified. Equal numbers (5×10⁵) of P14 CD8 T cells were transferred into congenically marked B6 mice that had been infected with L. major for 2 weeks (A). After 36 hours, spleens (B), infected ears (C), and contralateral uninfected ears (D) were harvested from the recipients and P14 frequency was analyzed by flow cytometry (E). P14 CD8 T cells present in the spleen (F) or infected ear (G) were incubated with BFA alone for 5 hours and analyzed for production of gzmB and IFN-γ. The relative mean fluorescence intensity for each was calculated (H). Data are representative of two independent experiments (n = 3–5 per group).

Figure 2. Previous heterologous infection increases leishmanial lesion size with no effect on parasite control.

B6 mice were infected with LCMV Armstrong or left uninfected. After 30 days, mice were infected with L. major and ear thickness was measured weekly (A). Pictures were taken 4 weeks post L. major infection (B). B6 mice were infected with Listeria-OVA or left uninfected and 30 days later boosted again with Listeria-OVA or left uninfected. Thirty days after the boost, all mice were infected with L. major and ear thickness was measure weekly (C). Pictures were taken 5 weeks post L. major infection (D). Infected skin from both groups was taken at various time points post infection and parasite burden was assessed using a limiting dilution assay (E and G). At 4 weeks post infection in the LCMV immune mice, draining lymph nodes were removed and cultured with leishmanial antigen or media alone. After 72 hours, supernatants were removed and analyzed for IFN-γ, IL-4, or IL-17 by ELISA (F). Cells cultured with media alone did not produce any cytokines (data not shown). These data are a compilation of five independent experiments (n = 5–10 mice per group per time point; A, B, E, and F) or representative of two independent experiments (n = 10–16 mice; C, D, and G).

Figure 3. Increased cellular infiltration into the lesion of LCMV immune L. major infected mice.

B6 mice were infected with LCMV Armstrong or left uninfected. After 30 days, mice were infected with L. major. After 4 weeks, infected ears were taken and fixed for histological analysis (A–D) or digested for flow cytometric analysis (E–H). Fixed ear tissue was sectioned and stained with hematoxylin and eosin (H&E) and imaged at 10× magnification or 40× magnification (inlay)(A). A Verhoeff's stain was also done on fixed ear tissue from LCMV immune L. major infected mice to stain for collagen and imaged at 10× magnification (B–D). Scale bars represent 100 µm. Damage unique to the LCMV immune lesions is highlighted by arrows, specifically epidermal thickening (B), cartilage destruction (C), epidermal ulceration (C), and loss of dermal collagen (D). Digested ears were stained for surface markers of monocytes and neutrophils (E), T cells (G), or NK cells (I). Results are shown in the form of representative plots (E, G, and I) and total cell numbers (F, H, and J). Samples were pregated on live, CD45+, TCRβ−, CD11b+ events (E) or on live, CD45+ events (G and I). Histological data are representative of two independent experiments (n = 4–5 mice per group; A–D). Flow cytometric data is a representative of five independent experiments (n = 4–5 per group; E–J). Percentages are shown as mean ± SEM.

Figure 4. CD8 T cells infiltrating the leishmanial lesions express gzmB but low levels of IFN-γ.

B6 mice were infected with LCMV Armstrong or left uninfected. After 30 days, mice were infected with L. major. After 4 weeks, infected ears (A) and peripheral blood (C) were taken for analysis by flow cytometry. Cells from the infected ears were incubated with BFA, monensin and CD107a antibody for 5 hours and then stained for additional cell surface markers and intracellular proteins. Representative dot plots (A) and total cell numbers (B) are shown. Peripheral blood was taken and white blood cells were isolated and stained for cell surface markers and intracellular proteins. Representative dot plots are shown (C). Whole ear tissue was homogenized and supernatants were analyzed for gzmB and IFN-γ by ELISA (D). Flow data are representative of five independent experiments (n = 4–5 mice per group). Ear supernatant data are representative of two independent experiments (n = 4 mice per group). Percentages are shown as mean ± SEM.

Figure 5. LCMV specific CD8 T cells are present in the leishmanial lesion and express gzmB but low levels of IFN-γ.

B6 mice were infected with LCMV Armstrong or left uninfected. After 30 days, mice were infected with L. major. After 4 weeks, L. major infected skin was taken and incubated with a pool of 20 LCMV derived peptides and BFA for 6 hours before staining for surface and intracellular proteins. Representative plots (A) and pooled data (B) are shown. Data are representative to 2 independent experiments (n = 5 per group). P14 T cells were transferred into B6 mice and the next day a group was infected with LCMV. After 30 days, GFP+ OTI T cells were transferred into all groups and mice were infected with L. major-OVA (C). After 4 weeks, L. major infected skin was harvested and incubated with BFA for 5 hours. The samples were analyzed for the presence of P14 cells or OTI cells (D) and numbers of each were calculated (E). The OTI or P14 cells from the same infected ear of an LCMV immune L. major infected mouse were also incubated with BFA, monensin, and CD107a antibody then stained for additional surface and intracellular proteins (F). Data are representative of four independent experiments (n = 4–5 per group). Percentages are shown as mean ± SEM.

Figure 6. Exacerbated immunopathology is lost following depletion of CD8 T cells in LCMV immune mice prior to L. major infection.

B6 mice were infected with LCMV or left uninfected for 45 days. Mice in each group were then treated with CD8-depleting antibody every 3 days for 15 days or left untreated. Blood was taken from animals in each group prior to L. major infection to assess CD8 depletion efficiency (A) and at 4 weeks post L. major infection to monitor reconstitution of the CD8 T cell compartment (B). All mice were with metacyclic L. major and ear thickness was measured weekly (C). Pictures were taken 4 weeks post L. major infection (D). Infected skin was taken at 4 weeks post infection and parasite burden was assessed using a limiting dilution assay (F). Infected skin was analyzed 4 weeks after L. major infection and the number of CD8 T cells present in each group was calculated (E). Data are representative of two independent experiments (n = 10 per group).

Figure 7. CD8 T cells induce immunopathology through engagement of NKG2D.

B6 mice were infected with LCMV Armstrong or left uninfected. After 30 days, mice were infected with L. major. After 4 weeks, infected skin was taken for analysis by flow cytometry (A, B and D) or frozen for immunohistochemical staining (C). Cells were pregated on live, CD45+, CD8β+ (A and D) or live, CD45+, CD11b+ (B). Frozen sections were cut and stained with DAPI (blue) and anti-Rae1-γ (red) and imaged at 40× magnification (C). CD8 T cells from the lesions of LCMV immune L. major infected mice were harvested 5 weeks post infection and enriched by negative selection. RMA cells were labeled with a high dose of CTV and Rae1 expressing RMA cells were labeled with a low dose for use as target cells. Control and Rae1 expressing cells were mixed 1∶1 and incubated with LCMV immune CD8 T cells. Specific lysis was calculated for each group of target cells (E). The ratio of live target cells with or without effector cells is shown (F). Cells from the infected skin were divided and incubated with BFA, monensin, and CD107a antibody ±10 µg/ml of NKG2D blocking antibody and the number of CD107a+ cells calculated (D). B6 mice were infected with LCMV Armstrong or left uninfected. After 30 days, some mice in each group were treated with NKG2D blocking antibody or isotype control antibody. The following day mice were infected with L. major, and antibody treatment continued biweekly for the duration of the experiment. Ear thickness was measured weekly (G). Infected skin was taken 6 weeks post infection and parasite burden was assessed using an LDA (H). Data are representative of two independent experiments (n = 5 per group; A–D and G–H) or a single experiment (n = 5; E and F).