Abstract
Dendritic cells (DC) initiatiate immunity and maintain tolerance. They internalize exogenous antigen and convert it into immunogenic peptides by lysosomal proteolytic degradation, ultimately followed by presentation to CD4 T cells. Monocyte-derived DC (MO-DC) generated in vitro with GM-CSF and IL-4 serve as prototype DC to analyse the cellular biology and biochemistry of DC. However, different types of primary DC, whose functional role in vivo and relationship to MO-DC generated in vitro is unclear, reside in human tissue as well as peripheral blood. The composition of lysosomal proteases in these primary human DC1 and DC2-cells and the way they handle a clinically relevant antigen are unknown, and there is no comparison of the lysosomal processing of antigen by these primary DC to that in primary human B cells or MO-DC generated ex vivo. We have isolated human peripheral blood (PB) DC1 and DC2 cells as well as primary B lymphocytes by magnetic separation and isolated lysosomal compartments from these cells, as well as from MO-DC. Expression and activity of endocytic proteases were assessed by western blot and active site-restricted affinity labelling using a synthetic probe that selectively binds to the active centre of cysteine proteases and allows a simultaneous semiquantitaive assessment and identification of multiple active protease species. In this analysis, PB-DC1 and DC2-cells lacked significant active Cathepsins (Cat) S, L and B as well as asparagine-specific endoprotease AEP, the major enzymes involved in antigen processing in the MHC II-compartment. Surprisingly, lysosomal extracts from PB-DC1 were by far more effective than MO-DC in processing the muliple sclerosis-associated autoantigen myelin basic protein (MBP) in vitro. When analyzed on a molecular scale using mass spectrometry, MBP processing was dominated by CatS, CatD and AEP in MO-DC, as expected, similar to B-lymphoblastoid cells (BLC). PB-DC, however, did not generate proteolytic processing intermediates indicative of CatS or AEP activity but showed the same pattern as primary B-lymphocyte-derived lysosomes, i.e. processing was performed by two cleavage sites that can be reproduced by purified CatG in vitro, suggesting a CatG like dominant lysosomal protease. While active CatG was present in primary human B cells, PB-DC1 cells lacked CatG protein by western blot, suggesting the presence of an as yet unknown dominant endoprotease with CatG-like activity in PB-DC1. By cleaving MBP after pos F90 and F114, this protease directly eliminates the integrity of the major immunodominant MBP epitope MBP85-99. This might lead to poor presentation of this epitope to regulatory T cells resulting in inefficient silencing of MBP-autoreactive T cells during the development of autoimmunity. Our results emphasize the need to apply state-of-the-art biochemical tools to primary human types of APC for the understanding of antigen processing and the rational design of tolerogenic or immunotherapy approaches towards human malignant and autoimmune disorders.
Author notes
Corresponding author