The CD8alpha(+) dendritic cell is responsible for inducing peripheral self-tolerance to tissue-associated antigens

Abstract

We previously described a mechanism for the maintenance of peripheral self-tolerance. This involves the cross-presentation of tissue-associated antigens by a bone marrow-derived cell type that stimulates the proliferation and ultimate deletion of self-reactive CD8 T cells. This process has been referred to as cross-tolerance. Here, we characterize the elusive cell type responsible for inducing cross-tolerance as a CD8alpha(+) dendritic cell (DC). To achieve this aim, transgenic mice were generated expressing yellow fluorescent protein (YFP) linked to CTL epitopes for ovalbumin and glycoprotein B (gB) of herpes simplex virus under the rat insulin promoter (RIP). Although tracking of YFP was inconclusive, the use of a highly sensitive gB-specific hybridoma that produced beta-galactosidase on encounter with antigen, enabled detection of antigen presentation by cells isolated from the pancreatic lymph node. This showed that a CD11c(+)CD8alpha(+) cell was responsible for cross-tolerance, the same DC subset as previously implicated in cross-priming. These data indicate that CD8alpha(+) DCs play a critical role in both tolerance and immunity to cell-associated antigens, providing a potential mechanism by which cytotoxic T lymphocyte can be immunized to viral antigens while maintaining tolerance to self.

Figure 1.

gB and OVA-specific T cells proliferate to YSS in the pancreatic LNs of RIP-YSS mice. 2 × 10⁶ CFSE-labeled recombination activation gene (Rag)-1^−/− OT-I (A and D) or Rag-1^−/− gBT-I (B, C, E, F) CD8⁺ T cells were adoptively transferred into RIP-YSS (A, B, D, E) or B6 (C and F) recipients. 3 d later, pancreatic (A–C), and inguinal (D–F) LNs were analyzed by flow cytometry, gating on CD8⁺CFSE⁺PI⁻ cells (reference 7).

Figure 2.

A bone marrow–derived cell type cross-presents YSS in RIP-YSS mice. Bone marrow chimeric mice were generated using RIP-YSS (A, C, D, F) or nontransgenic littermate (B and E) recipients. Lethally irradiated (2 × 550 cGray, 3 h apart) recipients were engrafted with T cell–depleted RIP-YSS (H-2K^b, B6 background)(A, B, D, E) or H-2K^bm1 mutant bone marrow (C and F). After 10 wk reconstitution, 2 × 10⁶ CFSE-labeled recombination activation gene (Rag)-1^−/− gBT-I CD8⁺ T cells were adoptively transferred into bone marrow chimeric recipients. 3 d later, the pancreatic (A–C) and inguinal (D–F) LNs were analyzed by flow cytometry. Histograms are gated on CD8⁺CFSE⁺PI⁻ cells.

Figure 3.

Cross-presentation of endogenous antigen induces deletion of antigen-specific CD8⁺ T cells. 4 × 10⁶ gBT-I CD8⁺ T cells were adoptively transferred into RIP-YSS transgenic mice or nontransgenic littermates. After 7 wk, the number of gBT-I cells in the LNs and spleen was determined by flow cytometry (reference 7). Individuals (white circles) and means (—) for each group are shown.

Figure 4.

Identification of the cell type responsible for cross-presentation of islet antigens. LN cells were isolated from RIP-YSS mice or nontransgenic B6 controls and subjected to collagenase and DNase digestion to release DCs. The preparations were either left (A) undepleted or (B) depleted of various populations including rat Ig⁺ (negative control), CD11c⁺, CD11b⁺, or CD8α⁺ cells using magnetic beads. These populations were then tested for their ability to stimulate lacZ production by the T cell hybridoma HSV-2.3.2E2, specific for gB (reference 19). The equivalent of one pancreatic LN (5–8 × 10⁵ cells) or 1/4–1/10 of an inguinal LN (4–8 × 10⁶ cells) was placed into each well containing 10⁵ hybridoma cells and cultured for 24 h at 37°C. Error bars indicate SD of triplicate wells. The experiments shown in A and B were performed 6 and 3 times, respectively. Data were collected blind.

Figure 5.

Isolation of the cell type responsible for cross-presentation of islet antigens. Pancreatic LNs were removed from RIP-YSS mice and digested with collagenase and DNase. The preparations were then enriched for DCs by magnetic bead depletion of other cells (Materials and Methods). Finally, DC-enriched populations were stained with anti-CD11c, anti-CD45RA, and anti-CD8α and sorted for CD11c⁺CD45RA⁻ cells that were either CD8α⁺ or CD8α⁻. CD45RA was used to exclude plasmacytoid DCs. The CD8α⁺ and CD8α⁻ DCs were titrated and used to stimulate the gB-specific hybridoma. CD8α⁺ DCs were recovered in small numbers that only allowed analysis of single wells per dilution, whereas CD8α⁻ DCs were titrated in duplicate and the mean count shown. This experiment was performed twice with similar results.