Surface Expression of PR1 Peptide on Thymic Dendritic Cells Does Not Eliminate PR1-Specific CD8+ Cells In Umbilical Cord Blood

Abstract 959

Positive and negative selection of developing thymocytes is mediated primarily by cortical and medullary epithelial cells (CEC and MEC), respectively, in the thymus. Tolerance to peripheral tissue antigens (PTA) not normally expressed in the thymus can result when PTA expression is induced by AIRE in MEC or when corticomedullary dendritic cells (DC) present endogenous antigens, both mechanisms contributing to deletion of potentially auto-reactive T cells. However, in some instances, negative selection may be incomplete, leading to the release of autoreactive cells into the periphery. Indeed, T cells specific for PR1, the HLA-A2-restricted self-peptide derived from proteinase 3 (P3) and neutrophil elastase (NE), are present in peripheral blood of healthy adults, albeit at extremely low frequency of fewer than 0.0005% of CD8+ T cells. This suggests there is strong thymic central tolerance to PR1. However, PR1-CTL can increase from 0.1 to 2% of peripheral blood CD8 T cells in CML patients treated with interferon or stem cell transplant, and similar levels can be achieved after PR1 peptide vaccination, which suggests that under some circumstances this central tolerance can be reversed. Because of the increasing use of umbilical cord blood (UCB) as an alternative donor source and because incomplete graft-versus-leukemia (GVL) immunity after UCB transplant contributes to relapse, we sought to determine whether PR1-CTL can be expanded from UCB. We hypothesized that PR1-CTL frequency should be low in UCB due to central tolerance. Surprisingly, we found PR1-CTL at a frequency ranging from 0.007 to 0.345% (mean 0.117%) of CD8+ cells in 57 HLA-A2+ cord blood units, similar to what is observed in immunologically responsive vaccine patients and 100- to 1000-fold higher than in healthy adults. The PR1-CTL were predominantly CCR7+CD45+CD28+ and did not efficiently expand ex vivo following peptide stimulation and low dose IL-2, which was consistent with a naive T cell subset. Therefore, this data suggests that central tolerance to PR1 is incomplete. To study whether PR1 is expressed in human thymus, we used the PR1/HLA-A2-specific antibody 8F4 to study PR1 expression. Thymic CEC and MEC expressed no P3, NE, or PR1 by flow cytometry or immunofluorescence imaging of sectioned fetal thymus, which is consistent with previous reports showing absence of P3 or NE induction by AIRE in MEC. Interestingly, by flow cytometric analysis, we found that PR1 is expressed on the surface of thymic dendritic cells (DC) exclusively. This selective expression of PR1 was further confirmed via immunoflourescent staining of sectioned fetal thymus, and the PR1-expressing DC were localized to the corticomedullary junction and medulla. Thus it appears that PR1 expression by DC in the thymus is insufficient for complete central tolerance to PR1. Furthermore, because we have previously shown that PR1-overexpressing CML cells can induce apoptosis of high affinity PR1-CTL, it is possible that peripheral tolerance mechanisms are most critical for preventing autoreactivity to PR1 in humans. These observations suggest possible strategies to overcome tolerance to PR1 by modifying DC uptake and cross-presentation of soluble P3 and NE, when selective autoimmunity may be desirable for leukemia patients or detrimental for patients with vasculitis such as Wegener's granulomatosis.