and Peripheral Blood Dendritic Cells Lysosomes from Human ...

The Journal of Immunology

Differential Processing of Autoantigens in Lysosomes from Human Monocyte-Derived and Peripheral Blood Dendritic Cells1 Timo Burster,2*§ Alexander Beck,† Eva Tolosa,‡ Petra Schnorrer,‡ Robert Weissert,‡ Michael Reich,*§ Marianne Kraus,*§ Hubert Kalbacher,§ Hans-Ulrich Ha¨ring,† Ekkehard Weber,¶ Herman Overkleeft,储 and Christoph Driessen3*§ Dendritic cells (DC) initiate immunity and maintain tolerance. Although in vitro-generated DC, usually derived from peripheral blood monocytes (MO-DC), serve as prototype DC to analyze the biology and biochemistry of DC, phenotypically distinct primary types of DC, including CD1c-DC, are present in peripheral blood (PB-DC). The composition of lysosomal proteases in PB-DC and the way their MHC class II-associated Ag-processing machinery handles a clinically relevant Ag are unknown. We show that CD1c-DC lack significant amounts of active cathepsins (Cat) S, L, and B as well as the asparagine-specific endopeptidase, the major enzymes believed to mediate MHC class II-associated Ag processing. However, at a functional level, lysosomal extracts from CD1c-DC processed the multiple sclerosis-associated autoantigens myelin basic protein and myelin oligodendrocyte glycoprotein in vitro more effectively than MO-DC. Although processing was dominated by CatS, CatD, and asparagine-specific endopeptidase in MO-DC, it was dominated by CatG in CD1c-DC. Thus, human MO-DC and PB-DC significantly differ with respect to their repertoire of active endocytic proteases, so that both proteolytic machineries process a given autoantigen via different proteolytic pathways The Journal of Immunology, 2005, 175: 5940 –5949.

D

endritic cells (DC)4 are a highly specialized APC. They activate naive T cells and are crucial for initiating an immune response and maintaining tolerance (1, 2). In vitro-generated DC, such as DC generated from peripheral blood monocytes (monocyte-derived DC, MO-DC) or bone marrow precursors, both serve as tools for immunotherapy and a model to understand the basic cell biology and biochemistry of DC (3, 4). The CD1c (BDCA-1) Ag is specifically expressed on the major subset of myeloid DC in the blood (5, 6), which induce a Th1polarized Th response. Days 5–7 MO-DC as well as CD1c-DC show a phenotype of “immature” or “quiescent” DC, i.e., they express relatively low levels of MHC class II and costimulatory molecules on the plasma membrane and are converted into fully mature, activated DC with high expression of MHC class II and

*Department of Medicine II and †Department of Medicine IV, ‡Hertie Institute for Clinical Brain Research and §Natural Sciences Research Centre, University of Tu¨bingen, Tu¨bingen, Germany; ¶Institute of Physiological Chemistry, University of Halle, Halle, Germany; and 储Leiden Institute of Chemistry, University of Leiden, Leiden, The Netherlands Received for publication January 18, 2005. Accepted for publication July 18, 2005. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by grants from the Deutsche Forschungsgemeinschaft (DR378.2-1, DR378.2-3, SFB 685, BU 1822/1-1), the Federal Ministry of Education and Research (Fo¨.01KS9602), and the Interdisciplinary Centre of Clinical Research Tu¨bingen (Interdisziplina¨res Zentrum fu¨r Klinische Forschung). 2 Current address: Department of Pediatrics and Immunology, Stanford University, Stanford CA 94305. 3 Address correspondence and reprint requests to Dr. Christoph Driessen, Department of Hematology, Oncology and Immunology, University of Tu¨bingen, Otfried Mu¨ller Strasse 10, D-72076 Tu¨bingen, Germany. E-mail address: christoph.driessen@ med.uni-tuebingen.de 4 Abbreviations used in this paper: DC, dendritic cell; BLC, B lymphoblastoid cell; Cat: cathepsin; AEP, asparagine-specific endopeptidase; MBP, myelin basic protein; MOG, myelin oligodendrocyte glycoprotein; MS, multiple sclerosis; DAP, diphenyl 1-(N-peptidylamino)alkanephosphonate ester; IAA, iodoacetamide.

Copyright © 2005 by The American Association of Immunologists, Inc.

costimulatory molecules by exogenous danger signals like damaged tissues or microbial products (2). Although fully activated DC induce immunity, immature DC are believed to induce tolerance by killing or paralyzing autoreactive T cells or by induction of regulatory T cells (7–10). The way DC handle exogenous Ag is key to their function. Antigenic material is internalized into the endocytic compartment, where it undergoes proteolytic processing after exposure to reducing and hydrolyzing enzymes, before the proteolytic products form a complex with MHC class II molecules and are routed to the cell surface for the triggering of T cells. Understanding this mechanism of Ag processing can be useful both for vaccine design and toward a selective manipulation of Ag presentation during an autoimmune response. The major endocytic proteases likely to mediate Ag breakdown in APC are the cathepsins (Cat) S, L, V, D, and B, along with the asparagine-specific endopeptidase (AEP) (11). Different types of APC (like DC, B lymphoblastoid cells (BLC), thymic epithelial cells, monocytes) express individual patterns of active endocytic proteases and little is known about the rules that govern proteolysis of Ag in human DC. Studies with BLC have chiefly substantiated the model that one protease or a limited set of proteolytic enzymes dominate Ag processing by initiating the proteolytic cascade (“unlocking protease(s)”), controlling Ag breakdown, the generation or destruction of a given epitope, and thus the triggering of a T cell response (12–14). Depending on the substrate specificity of this unlocking protease and the relationship of its dominant cleavage site to the major immunogenic epitope of the antigenic protein, the unlocking protease can either increase T cell activation by facilitating the processing of the intact Ag or decrease the T cell response by direct destruction of the immunodominant epitope. In BLC, processing of MBP, an autoantigen implicated in the pathogenesis of multiple sclerosis (MS) (15), is initiated by AEP (16), which consequently controls the MBP-specific T cell response (13). AEP, however, is absent from primary 0022-1767/05/$02.00

The Journal of Immunology human B lymphocytes, where CatG controls MBP turnover, resulting in an entirely different proteolytic pattern (17). Potential processing pathways for myelin oligodendrocyte glycoprotein (MOG), another autoantigen implicated in the pathogenesis of MS, have not yet been established (18). In DC generated in vitro, AEP as well as CatS, L, B, H, C, Z, and G are present (3, 19, 20), and exogenous Ag is selectively routed to an endocytic compartment enriched with CatS in vivo (21). However, it is unknown which of these proteases dominates turnover of an intact protein in the endocytic compartment of DC and to what extent expression, activity, and the pathway of MBP or MOG degradation differ between DC generated in vitro and primary peripheral blood DC. We have here compared human MO-DC and primary CD1c-DC with respect to the expression and activity of endocytic proteases and have followed the destruction of MBP and MOG by the lysosomal protease pool isolated from both types of cells.

Materials and Methods

5941 Determination of Cat activity Hydrolysis of Suc-AAPF-AMC (Bachem) that serves as an established substrate to determine CatG activity (26) was quantified fluorometrically exactly as described elsewhere (17). The fraction of the emitted fluorescence that could be inhibited by addition of chymostatin was considered CatG activity. Combined CatB, CatL, and CatS activities were determined using the fluorogenic substrate Z-FR-AMC (0.1 M citrate (pH 5.0), 4 mM DTT, and 4 mM EDTA, as published in Ref. 25).

Affinity labeling of active proteases For labeling of active cysteine proteases, cell lysates (5 ␮g) were incubated with reaction buffer (50 mM citrate/phosphate (pH 5.0), 1 mM EDTA, and 50 mM DTT) in the presence of DCG-0N, a derivative of DCG-04 that shows identical labeling characteristics (27, 28), for 1 h at room temperature. Reactions were terminated by addition of SDS reducing sample buffer and immediate boiling. Samples were resolved by 12.5% SDSPAGE gel, then blotted on a polyvinylidene difluoride membrane and visualized using streptavidin-HRP and the ECL detection kit (29). Active CatG was visualized using the serine protease-restricted biotinylated affinity probe diphenyl 1-(N-peptidylamino)alkanephosphonate ester (DAP) (30) under identical conditions (probe provided by M. Bogyo, Stanford University, Stanford, CA).

Generation and characterization of DC

Abs and Western blot

Human PBMC were isolated from buffy coats of donor blood by density gradient centrifugation. Subpopulations of CD1c-DC and CD14-positive monocytes were positively selected using the MACS technique (Miltenyi Biotec) according to the manufacturer’s protocol with CD1c and CD14 Abs, respectively. MO-DC were generated from CD14-selected monocytes by incubation with GM-CSF and IL-4 for 6 days as described previously (19).

Anti-Cat antisera were generated against recombinant CatS and affinitypurified human CatL, B, H, and D. The anti-AEP Ab was a gift from C. Watts (University of Dundee, U.K.), and anti-cystatin antisera were purchased from Upstate Biotechnology. A polyclonal goat anti-human CatG antiserum was obtained commercially (Santa Cruz Biotechnology). Cells lysed in Nonidet P-40/pH 7.0 lysis buffer (50 mM sodium acetate, 5 mM MgCl2, and 0.5% Nonidet P-40) were adjusted for equal total protein, resolved by SDS-PAGE, and blotted using published conditions (17).

Expression and purification of MBP and MOG

RT-PCR for lysosomal proteases

Human recombinant MBP (18.5 kDa; Swiss-Prot: P02686-5, lacking methionine in pos. 1) was produced in Escherichia coli and purified by ion exchange chromatography as described elsewhere (16). The purified protein was controlled by MALDI-mass spectrometry for both purity (⬎98%) and correct mass. Human recombinant MOG was expressed as previously described (22). Briefly, the DNA sequence encoding the extracellular domain of mature human MOG (including four N-terminal amino acids of the transmembrane domain) was PCR amplified and subcloned into pQ60 (Qiagen). The Histagged fusion protein (recombinant human MOG) was expressed in E. coli and purified under denaturing conditions by metal chelate affinity chromatography on Ni-NTA agarose columns (Qiagen) according to the manufacturer’s guidelines. MOG was solubilized and reduced using 3-bromopropyltrimethyl-ammonium bromide, a method that had been demonstrated not to influence the pattern of proteolytic degradation of a given protein (23, 24).

Cells were lysed in TRIzol (Invitrogen Life Technologies), and the RNA was extracted following the manufacturer’s protocol. First-strand cDNA was prepared by reverse transcription using the Superscript RNase H- Reverse Transcriptase (Invitrogen Life Technologies) and random hexamers. For real-time PCR, gene expression was measured in the ABI Prism 7000 Sequence Detection System (Applied Biosystems). Primer pairs were selected with the Husar Genius software (DKFZ) to span exon-intron junctions and to result in amplicons ⬍150 bp. Amplification of specimens and serial dilutions of the amplicon standards was conducted with 2⫻ SYBR Green Master Mix (Applied Biosystems). Standard curves were generated for each gene with amplification efficiency ⬎90%. Relative quantification of gene expression was determined mathematically as suggested by the manufacturers. Results are normalized with respect to the internal control 18S rRNA and are expressed relative to the levels found in one of the samples. The following oligonucleotide primers (MWG) were used: 18S rRNA forward primer, 5⬘- CGGCTACCACATCCAAGGAA-3⬘ and reverse primer, 5⬘-GCTGGAATTACCGCGGCT-3⬘; human AEP GAAGC CTGTGAGTCTGGGTC, CAGTCCCCCAGGTACGTG; CatB, CTGTG TAT TCGGACTTCCTGC, CCAGGAGTTGGCAACCAG; CatD, AACT GCTGGACATCGCTTG, AGGTACCCGGAGAGGCTG; CatF, ATAT GAGTCAAAGGAAGAAGCCC, GATCAC TGAACTTGGTGACTCC; CatG, CCCCTACATGGCGTATCTTCA, TTGCTTCCC CAGCAATGAG; CatH, ACTGGCTGTTGGGTATGGAG, AGGCCACACATGTTCTTTCC; CatL, ACCAAGTGGAAGGCGATG, TTCCCTTCCCTGTATTCCTG; CatS, ACTCA GAATGTGAATCATGGTG, TTCTTGCCATC CGAATA TATCC; and CatZ, GGGAGGGAGA AGATGATGG, ATGTGGTGTC CTGGTATTCG.

Generation of lysosomal extracts and in vitro processing Lysosomal extracts were generated by differential centrifugation and characterized as published before (17, 25). In brief, crude endocytic compartments were enriched by differential centrifugation from postnuclear supernatants, followed by a short exposure of the membrane pellet to water, which preferentially disrupts the membrane integrity of lysosomes due to their low osmotic resistance. This process yielded highly enriched lysosomal extracts as judged by the distribution of CatD, transferrin receptor, and N-acetyl-glucosaminidase. For in vitro processing, substrate solution (0.04 ␮g/␮l MBP or MOG, 0.1 M citrate (pH 5.0), and 2.5 mM DTT) was incubated with lysosomal fractions at 37°C (0.5 ␮g of total protein).

Identification of processing products from MBP and MOG Processed MBP and MOG were analyzed by liquid chromatography-electrospray-mass spectrometry as published previously (17). Expected masses were calculated as average mass with statistical isotope distribution. The mass accuracy of the detected molecular ions was in the range of ⫾200 ppm. Where proteolytic products were not analyzed by mass spectrometry (see Fig. 3), they were separated by microbore HPLC as published elsewhere (16). CatG inhibitor I was purchased from Calbiochem and used at 100 nM and CatG from human leukocytes was purchased from Sigma-Aldrich.

Results Endocytic proteases from CD1c-DC result in a characteristic processing pattern of MBP in vitro Human peripheral blood DC and CD14-positive human peripheral blood monocytes were directly purified from peripheral blood by the MACS technique using CD1c- or CD14-specific Abs coupled to magnetic beads. MO-DC were generated from MACS-selected CD14⫹ monocytes in vitro using GM-CSF and IL-4 for 6 days. This yielded a ⬎80% pure population of MO-DC (MHC II⫹, CD1a⫹, CD14⫺, and CD 22⫺) and ⬎90% pure populations of CD1c-DC (MHC II⫹, CD1c ⫹, CD22⫺; Fig. 1).

5942

FIGURE 1. DC phenotype. The phenotype of monocyte-derived DC and peripheral blood-derived CD1c-DC, as analyzed by FACS.

To compare the function of the proteolytic machineries between both types of DC (MO-DC generated in vitro vs primary CD1c-DC from peripheral blood), we generated lysosomal extracts from both cell types (17, 25). Equal amounts of total lysosomal protein derived from MO-DC and CD1c-DC were incubated with recombinant MBP in vitro at pH 5, followed by analysis of the respective proteolytic MBP intermediates by HPLC and mass spectrometry (Fig. 2, a and b). Lysosomal proteases derived from CD1c-DC were more effective in degrading intact MBP than the proteolytic apparatus in MO-DC. Although intact MBP was no longer detectable (elution time, 79 min) after a 4-h incubation with CD1c-DCderived lysosomal fractions, it was still present in significant amounts when equivalent amounts of MO-DC-derived lysosomal extracts were used. Turnover of MBP with MO-DC-derived lysosomal proteases yielded major proteolytic cleavage sites at the CatD-sensitive positions 44FF45 and 89FF90 (Fig. 2, a and b and Table I) and at the AEP-sensitive site 92NI93. LPS treatment did not significantly change this pattern, although fragments indicative of CatS-mediated cleavage at 88HF89 and 110SL111 were now also detectable, in agreement with the increase in active CatS observed under these conditions (19). An entirely different pattern of

Ag PROCESSING BY BLOOD DC proteolytic intermediates was observed after incubation of MBP with lysosomal proteases from CD1c-DC: proteolytic breakdown was dominated by two chymotrypsin-like cleavages at 90FK91 and 114FS115 in addition to the CatD-sensitive 44FF45-site. Interestingly, 90FK91 is located in the core region of the major immunogenic MBP epitope MBP85–99. Of note, proteolytic intermediates indicative of AEP- or CatS-mediated processing were not observed. A significant portion of MBP processing products from CD1c-DC, but not from MO-DC, differed at the N terminus while showing identical C termini, suggesting relatively high aminopeptidase activity in CD1c-DC lysosomes. Taken together, these data demonstrate that lysosomal extracts from MO-DC and CD1c-DC yield different processing patterns of MBP. They further suggest that the major proteases implicated in lysosomal Ag processing, CatS and AEP, are functionally dominant proteolytic enzymes in lysosomes from MO-DC, while they appeared to be of little, if any, functional relevance for MBP breakdown in CD1c-DC-derived lysosomes. Strikingly, two of three major MBP processing sites generated with CD1c-DC-derived lysosomes (90FK91 and 114FS115) were identical to cleavage sites reported after treatment of MBP with isolated CatG (17). To further corroborate the different proteolytic processing pathways between lysosomes from either type of DC, recombinant human MOG carrying a C-terminal His-tag was digested with lysosomal extracts and analyzed by mass spectrometry in a similar fashion (Fig. 2c and Table II). Incubation of MOG with lysosomes derived from MO-DC for 5 h resulted in poor substrate turnover with an endoproteolytic cleavage site at 100FF101. LPS treatment of MO-DC led to more efficient proteolytic turnover along a similar proteolytic pathway. The MOG positions 15AL16, 73LK74, and 109AA110 were identified as additional cleavage sites. When CD1c-DC were used as a source of lysosomal enzymes, efficient degradation of MOG was observed after 1 h, and the positions MOG 47VV48 and

FIGURE 2. Processing of MBP and MOG by DC-derived lysosomal proteases. a, Intact MBP was incubated with lysosomal extracts from resting or LPS-stimulated (upper panel) MO-DC or CD1c-DC (lower panel) normalized for total protein. Proteolytic fragments obtained are resolved by HPLC and visualized by absorption at 280 nm. b, Proteolytic MBP fragments are displayed along with full-length MBP and the location of the two major immunogenic epitopes (MBP85–99 and MBP116 –123). Major cleavage sites are marked by arrows and the respective flanking amino acids are indicated. c, Proteolytic fragments similarly obtained after digesting MOG with either type of lysosomal fractions or with purified CatG (1 ␮g/ml) are displayed alongside with the full-length MOG product and the respective immunodominant T cell epitope (MOG99 –107).


5943

Table I. MBP fragments after processing with DC-derived lysosomes, as assessed by mass spectrometrya

Retention Time

56.1–56.9 68.1– 69.4 74.1–75.2 75.0 –76.1 77.8 –78.3 78.4 –79.9

Retention Time

55.0 –56.5 63.0 – 63.7 68.1– 69.4 69.0 –70.7 74.1–75.2 75.0 –76.1 77.8 –78.3 78.4 –79.9

Retention Time

54.5–56.1 59.1– 60.6 68.1– 68.8 68.9 – 69.3 69.9 –70.4 71.3–72.6 74.6 –75.3 77.6 –78.6 a

MO-DC Expected Mass [M ⫹ H]⫹

⌬ Da

MBP Fragment

5,268.7 4,962.6 9,822.8 8,267.3 8,656.7 10,212.3 13,517.0 18,460.5

0.8 0.0 0.2 0.4 0.4 2.0 0.3 1.4

45–92 1– 44 1– 89 93–170 90 –170 1–92 45–170 1–170

[M ⫹ H]⫹

MO-DC ⫹ LPS Expected Mass [M ⫹ H]⫹

⌬ Da

MBP Fragment

5,267.9 3,378.6 4,962.6 9,674.3 9,821.7 8,267.7 8,656.3 10,210.3 13,517.3 18,459.1

5,268.7 3,378.7 4,962.6 9,675.6 9,822.8 8,267.3 8,656.7 10,212.3 13,517.0 18,460.5

0.8 0.1 0.0 1.3 1.1 0.4 0.4 2.0 0.3 1.4

45–92 111–141 1– 44 1– 88 1– 89 93–170 90 –170 1–92 45–170 1–170

[M ⫹ H]⫹

CD1c-DC Expected Mass [M ⫹ H]⫹

⌬ Da

MBP Fragment

2,595.0 5,026.2 4,962.7 5,702.4 3,065.4 3,237.1 2,993.8 5,933.5 3,350.7 4,934.3 8,358.9

2,595.0 5,026.5 4,962.6 5,702.4 3,065.4 3,237.6 2,993.8 5,933.6 3,350.8 4,934.4 8,358.2

0.0 0.3 0.1 0.0 0.0 0.5 0.0 0.1 0.1 0.1 0.7

91–114 45–90 1– 44 39 –90 18 – 44 16 – 44 19 – 44 115–170 15– 44 115–161 15–90

[M ⫹ H]

⫹

5,267.9 4,962.6 9,822.6 8,267.7 8,656.3 10,210.3 13,517.3 18,459.1

See footnote to Table II.

91FS92 were the sites of initial proteolytic attack. The latter corresponded with processing of MBP at 114FS115 in lysosomes from the same type of DC. Further processing of MOG for 5 h revealed 16LV17, 47VV48, 73LK74, 91FS92, and 101FR102 as major proteolytic sites, in contrast to lysosomes from MO-DC or MO-DC stimulated with LPS. Because the dominant 91FS92 cleavage suggested CatG-mediated processing of MOG, we tested purified CatG with respect to MOG processing in vitro. Strikingly, purified CatG reproduced the 16LV17, 73LK74, 91FS92, and 101FR102 processing sites observed with CD1c-DC-derived lysosomes, suggesting a major functional role of CatG in lysosomes from CD1c-DC. Different classes of proteases initiate MBP processing in MO-DC and CD1c-DC To functionally confirm that the initial proteolytic step in processing of intact MBP was indeed performed by different types of unlocking proteases in MO-DC vs CD1c-DC, MBP was digested in the presence or absence of protease inhibitors. Iodoacetamide (IAA) inhibits cysteine proteases including CatS, L, B, C, Z, H, as well as AEP, pepstatin A blocks aspartate proteases like CatD, whereas PMSF inhibits serine proteases like CatG. Resolution of the emerging MBP fragments after 4 h of digest

with MO-DC-derived lysosomal extracts in the absence of protease inhibitors revealed intact MBP eluting at 24 min (Fig. 3), along with a number of processing intermediates at shorter retention times. Addition of IAA and pepstatin A virtually blocked MBP turnover, entirely consistent with the dominant role of AEP and CatD for MBP processing in MO-DC-derived lysosomes observed before. A completely different picture emerged with lysosomes from CD1c-DC: Addition of IAA and pepstatin influenced neither the kinetics of MBP turnover nor the fragmentation pattern observed, excluding that papain-like Cat, CatD, or AEP played a dominant role in controlling the initial steps of MBP destruction by lysosomal proteases from CD1c-DC. Inhibition of serine proteases by PMSF, by contrast, markedly delayed the turnover of intact MBP and resulted in a different pattern of proteolytic fragments. Of note, addition of a selective CatG inhibitor preserved the integrity of the major fraction of MBP. This indicated that a serine protease activity, likely CatG, was the dominant and rate-limiting proteolytic activity for turnover of MBP with CD1c-DC-derived lysosomes, in contrast to MO-DC, where this was controlled by cysteine and aspartate proteases. Thus, different processing pathways are initiated when MBP is degraded by lysosomes from either MO-DC or CD1c-DC.

5944

Ag PROCESSING BY BLOOD DC Table II. MOG fragments after processing with DC-derived lysosomes, as assessed by mass spectrometrya

Retention Time

70.4 –72.4

Retention Time

41.5– 42.3 43.2– 44.7 50.5–51.5 52.9 –53.9 54.3–55.7 64.3– 67.0

Retention Time

67.3– 68.6 68.6 –70.5

Retention Time

47.6 – 48.9 49.2–50.4 50.2–51.6 54.3–55.1 55.6 –56.6 59.9 – 61.1 63.0 – 63.7 67.5– 68.6 68.6 –70.3 70.7–71.7

MO-DC, 5-h Expected Mass [M ⫹ H]⫹

⌬ Da

MOG Fragment

4285.61

0.12

101–135

MO-DC ⫹ LPS, 5-h Expected Mass [M ⫹ H]⫹

⌬ Da

MOG Fragment

1347.80 1418.84 1925.13 2153.84 1725.75 2576.20 4285.61

0.08 0.07 0.10 0.19 0.02 0.26 0.51

4 –14 4 –15 74 –90 92–109 101–114 115–135 101–135

CD1c-DC, 1-h Expected Mass [M ⫹ H]⫹

⌬ Da

MOG Fragment

5085.73 5229.60

0.08 1.02

48 –91 92–135

CD1c-DC, 5-h Expected Mass [M ⫹ H]⫹

⌬ Da

MOG Fragment

2930.44 1578.68 1769.03 3031.31 1788.04 2576.20 2072.20 2287.27 4782.37 4138.44 4887.47

0.0 0.12 0.15 0.31 0.15 0.31 0.13 0.10 0.68 0.36 0.75

49 –73 102–114 74 – 89 48 –73 1–16 115–135 74 –91 1–21 48 – 89 102–135 50 –91

⫹

[M ⫹ H]

4285.73

⫹

[M ⫹ H]

1347.72 1418.91 1925.03 2154.03 1725.77 2575.94 4285.10

⫹

[M ⫹ H]

5085.81 5228.58

⫹

[M ⫹ H]

2930.44 1578.80 1769.18 3031.62 1788.19 2576.51 2072.33 2287.37 4781.69 4138.08 4886.72

a Tables I and II: proteolytic processing of MBP and MOG by DC-derived lysosomes in vitro. Equal amounts of protein from lysosomal extracts generated from MO-DC (resting and stimulated with LPS for 24 h) and CD1c-DC, respectively, were incubated with recombinant MBP (Table I) or MOG (Table II) in vitro. The emerging proteolytic MBP products were resolved and identified by liquid chromatography-ES-mass spectrometry as described. HPLC retention times of the respective fragments, their actual and theoretical masses ([M ⫹ H]⫹, expected mass [M ⫹ H]⫹, respectively), the calculated difference between both (⌬ Da), and the respective proteolytic fragment deferred from this are presented.

Differential expression of lysosomal proteases in different types of human DC To assess the panel of proteases and endogenous protease inhibitors of the MHC class II Ag-processing compartment in both types of DC, cell lysates were prepared from each type of DC and equal amounts of protein were probed for CatD, B, H, L, S, C, and Z and AEP along with cystatin C and B by immunoblot (Fig. 4). Where indicated, maturation/activation was induced by addition of LPS, TNF-␣, or IFN-␥ for the last 24 h of culture. Considerable differences were observed between resting MO-DC and primary CD1cDC: in MO-DC, CatB was abundantly present as the mature 25kDa polypeptide, which was absent from CD1c-DC. In CD1c-DC, the 30-kDa CatB polypeptide was present in larger amounts as compared with the 25-kDa isoform, also in contrast to MO-DC. Treatment with either LPS, TNF-␣, or IFN-␥ did not affect the expression or maturation pattern on CatB in MO-DC. CatC, CatZ, and CatL that were expressed by MO-DC in significant amounts were absent or at the limit of detection in CD1c-DC. The expression of AEP also differed at least by one order of magnitude between MO-DC and CD1c-DC. In agreement with published data, induction of maturation by LPS or TNF-␣ led to increased conversion of pro-AEP into its mature form (3) in MO-DC. CatH was

also markedly reduced in CD1c-DC, as compared with MO-DC. The protease inhibitor cystatin B showed a high expression in CD1c-DC and considerably lower quantities in MO-DC. CatS, CatD, and cystatin C were present in similar amounts at the polypeptide level in all cell types and conditions tested. Expression levels of mRNA for endocytic proteases in DC We next compared the transcriptional activity for endocytic proteases and cystatins between MO-DC and CD1c-DC using quantitative RT-PCR (Fig. 5). Primary monocytes and PBMC served as controls, and the amounts of the respective mRNA detected were expressed relative to its quantities in PBMC. For CatB, CatZ, CatH, and AEP, the quantitative differences in mRNA expression between MO-DC and CD1c-DC1 reflected the differences observed in the protein level, i.e., CD1c-DC1 contained lower levels of mRNA for these proteases. CatL, by contrast, showed significant transcriptional activity in CD1c-DC, although the respective protein was not detected. CD14 monocytes contained sizable amounts of CatL mRNA, as expected. Only moderate differences in the amounts of mRNA between MO-DC and CD1c-DC were observed for CatS and CatD, also roughly in agreement with the results obtained by immunoblot. Cystatins B and C transcribed at


FIGURE 3. Effect of protease inhibitors on MBP processing by DC. Equal amounts of protein from lysosomal extracts generated from MO-DC or CD1c-DC were incubated with MBP for 4 h, either in the absence of protease inhibitors or in the presence of IAA (30 mM), Pepstatin A (Pep, 1 ␮M), PMSF (30 mM), or CatG inhibitor (100 nM), respectively. Proteolytic fragments were resolved by HPLC.

higher levels in MO-DC, in contrast to the data from the protein level. Interestingly, LPS stimulation of MO-DC selectively induced robust transcription of cystatin F. Thus, the fundamental differences between MO-DC and primary human peripheral blood CD1c-DC in the amounts of endocytic proteases were roughly reflected by similar differences in the transcriptional activities of the respective genes. Peripheral blood DC lack significant activity of the major cysteine Cat To assess the major cysteine Cat involved in MHC class II-associated Ag processing on a functional level, we measured the turnover of the fluorogenic substrate Z-FR-AMC, which detects the combined activity of CatS, CatL, and CatB in cell lysates (25). Substrate turnover per microgram of total protein was ⬃20-fold higher in MO-derived DC compared with human peripheral blood DC (Fig. 6a), where it was close to the detection limit. Activitybased affinity labeling using the active site-directed chemical probe DCG-0N was used to differentiate between the individual Cat in this respect. Because this biotinylated probe binds irreversibly to the active site of papain-like cysteine proteases, it allowed us to visualize active CatS, CatL, CatB, CatC, and CatZ polypeptides in an activity-dependent, semiquantitative fashion after SDS-PAGE and HRP blot. Detection of ␤-actin by immunoblot was performed to control for similar amounts of total cellular protein (Fig. 6b). In MO-DC, the active polypeptides visualized with the probe were CatZ (37 kDa), CatB (33 kDa), CatS (28 kDa), and CatL/CatC (both migrating as a single band at 25 kDa) as expected (19, 21, 29, 31). Treatment with LPS, TNF-␣, or IFN-␥ increased the amounts of active CatS detected in MO-DC, but did not significantly change the gross pattern of active polypeptides as reported earlier (19, 20). In contrast to MO-DC, only trace amounts of active CatS were

5945

FIGURE 4. Protease expression in DC assessed by Western blot. Expression of Cat, the asparagine-specific endoprotease as well as its precursor (AEP and proAEP, respectively) and cystatins B and C (Cyst) were assessed in whole lysates from MO-DC, CD1c-DC, and isolated monocytes (CD14). MO-DC were either lysed without activation or were challenged by TNF-␣, IFN-␥ (100 U/ml each), or LPS (1 ␮g/ml) for the last 24 h of culture. Equal amounts of total cellular protein were analyzed by SDS-PAGE and Western blot.

detectable in CD1c-DC. Of note, this did not reflect the situation observed on the polypeptide or mRNA level, where the amounts of CatS were in the same order of magnitude. In agreement with our results from the protein and mRNA levels, active CatB and CatZ were also strongly reduced in peripheral blood DC, as compared with MO-DC generated in vitro. In addition, we failed to detect CatL and/or CatC in CD1c-DC using the affinity-labeling technique, in contrast to MO-DC, where the respective signal was strongly present, also in agreement with our results from the protein level. Taken together, these data indicated that the major endoproteases implicated in Ag processing in the MHC class II compartment (AEP, CatS, CatL) were lacking significant activity in peripheral blood-derived primary DC, in contrast to MO-DC. This suggested that primary human peripheral blood-derived DC might contain as yet unidentified proteolytic enzymes that dominate proteolytic breakdown, resulting in an alternative pathway of Ag processing. CD1c-DC contain dominant CatG activity Because CatG degraded intact MBP at 90FK91 and 114FS115 in primary human B cells (17) and also reproduced the major MOG cleavage sites, we assumed that CatG was the dominant proteolytic enzyme in CD1c-DC-derived lysosomes. We therefore measured turnover of the fluorogenic substrate Suc-AAPF-AMC by equal amounts of protein from cell lysates. This substrate is hydrolyzed by chymotrypsin-like endoprotease activity, including CatG, and therefore represents an established substrate to measure CatG activity (26). As expected, robust substrate turnover was observed with purified CatG and cell lysates from monocytes, while BLCderived lysates served as a negative control (17) (Fig. 7a). Little substrate turnover was observed with lysates from MO-DC, while CD1c-DC contained considerable CatG-like activity. To directly assess the respective cell types with regard to active CatG, the affinity-labeling type of experiments with the CatG-reactive affinity probe DAP (30) were performed (Fig. 7b). A strong positive

5946

Ag PROCESSING BY BLOOD DC

FIGURE 5. mRNA expression of proteases in DC. The relative amounts of RNA for the Cat S, L, H, Z, B, and D along with AEP and the cystatins B, C, and F were analyzed by quantitative RT-PCR and compared between CD1c-DC (CD1c), PBMC, monocytes (CD14), MO-DC, and MO-DC exposed to LPS for 24 h.

activity signal comigrating with purified CatG was obtained from granulocytes as expected. Active CatG was present in CD1c-DC, but not detectable in MO-DC or LPS-stimulated MO-DC, in agreement with what had been observed for hydrolysis of the fluorogenic substrate Suc-AAPF-AMC. The relationship of CatG activity to the amounts of CatG protein and mRNA present in the different types of cells was assessed by Western blot. CD1c-DC contained only relatively low amounts of mature CatG polypeptide compared with MO-DC. Activation of MO-DC with LPS, IFN-␥, or TNF-␣ led to different degrees of down-modulation of CatG, in agreement with recently published data from a monocyte cell line (32). Similarly, the transcrip-

FIGURE 6. Cat activity in DC. a, The combined activity of the Cat CatB, CatL, and CatS was compared between MO-DC, MO-DC stimulated with LPS, TNF-␣, or IFN-␥ and CD1c-DC, respectively, by measuring the turnover of the fluorogenic substrate Z-FR-AMC in cell lysates normalized for total protein. Results are expressed in arbitrary units, mean values of triplicate samples are presented. b, Cell lysates from MO-DC, MO-DC stimulated with LPS, TNF-␣, or IFN-␥, and CD1c-DC and monocytes (CD14) were normalized for total protein content, labeled with the activitybased biotinylated probe DCG-0N, and resolved by SDS-PAGE, followed by semiquantitative visualization of the active papain-like cysteine proteases by streptavidin-HRP blot. A blot against ␤-actin served as control for comparable amounts of total protein analyzed.

tional activity of CatG was higher in MO-DC and CD14 monocytes than in CD1c-DC (Fig. 7, c and d).

Discussion MO-DC generated in vitro serve as a model type of DC to unravel the complex interactions among DC maturation, MHC class II peptide loading, and endocytic transport (3, 33, 34). The emerging picture suggests that endocytic protease activity is regulated by differential activity of the lysosomal ATPase during DC maturation (3). By analyzing lysosomal MBP processing at pH 5.0 in vitro, we have here mimicked the conditions present in the lysosomal compartment of DC in the activated state in vivo. The MHC class II-associated proteolytic machinery is characterized by a hierarchical proteolytic cascade, where the initial step controls the efficiency of Ag processing, peptide presentation, and T cell activation (14). Different types of APC as well as primary cells and immortalized cell lines contain distinct activity patterns of endocytic proteases (11, 17, 31, 35–37) which might result in different processing pathways for a given Ag and hence in different selections of peptides presented. We here demonstrate that the expression of proteolytic enzymes differs significantly between MO-DC and CD1c-DC. As expected, mature polypeptides for CatS, B, L, D, H, and Z as well as AEP were detected in MO-DC (19 –21, 29, 31), and proAEP was converted into its fully mature form upon DC maturation (3). However, CD1c-DC lacked mature CatL, CatB, CatH, CatC, and CatZ, while containing roughly comparable amounts of mature CatS and CatD as well as cystatin C. When assessed on the activity level, MO-DC contained robust amounts of CatB, S, L, C, and Z, while these signals were reduced by at least one order of magnitude in CD1c-DC isolated from peripheral blood. Especially the poor activity signal for CatS was surprising, since peripheral blood DC and MO-DC had shown similar CatS expression on the protein level. This suggests that CatS is tightly controlled at the activity level, e.g., by endogenous inhibitors such as cystatins. Indeed, we found increased expression of the pan-papain protease inhibitor cystatin B (38) in CD1c-DC. The relatively high number of N-terminally truncated MBP-processing intermediates detected in lysosomes from CD1c-DC which might imply a prominent aminopeptidase activity was not observed to a similar degree when MOG processing was analyzed, suggesting that this finding was not a general phenomenon of lysosomal proteolysis in this cell type. Quantitative analysis of the mRNA expression of endocytic cysteine proteases was roughly


5947

FIGURE 7. CatG activity, CatG protein, and mRNA in DC. A, Total amounts of CatG activity were compared between equal amounts of protein from unstimulated MO-DC, MO-DC exposed to TNF-␣/IFN-␥/LPS, respectively, CD1c-DC1, or monocytes (CD14). Purified CatG served as a positive, BLC as a negative control. CatG activity was determined by turnover of the fluorogenic substrate Suc-AAPF-AMC and expressed in arbitrary units. Mean values and SD of three independent experiments are presented. B, Equal amounts of total lysosomal protein from CD1c-DC and MO-DC were probed for active serine proteases using the biotinylated affinity probe DAP alongside purified CatG and enriched granulocytes. C, Total cell lysates from the same cell types as in A were assessed for CatG protein by Western blot. A blot against ␤-actin served as control for equal amounts of total protein analyzed. D, The relative amounts of RNA for CatG were analyzed by quantitative RT-PCR and compared among CD1c-DC (CD1c), PBMC, monocytes (CD14), MO-DC, and MO-DC exposed to LPS for 24 h.

consistent with the data from the protein level, indicating that transcriptional control is also a major regulator for Cat activity in DC. In vitro generation of MO-DC in the presence of IL-4 and GMCSF might, to some extent, explain the differences in the amount of cysteine Cat observed between MO-DC and CD1c-DC, because the proteolytic contents in CD14 control cells closely matched those in primary CD1c-DC. Also, CatG activity has been demonstrated to decrease during DC differentiation in IL-4 and GM-CSF (39). Irrespective of the underlying mechanism that modulates protease activity, MO-DC generated in vitro by this method are frequently being used for immunotherapy approaches and also serve as a model type of DC to assess ways of modifying Ag presentation by applying protease inhibitors. For both types of application, possible differences in the proteolytic machinery between DC generated ex vivo and primary human DC are implicated by our study. In previous work, we have characterized the turnover of MBP by isolated Cat as well as in lysosomal extracts from B cells (16, 17). Using the same experimental system, we here follow the turnover of MBP in lysosomes derived from human DC, including primary peripheral blood-derived DC. As such, this analysis not only allows us to compare MBP processing in vitro between different types of human DC, but in addition between lysosomal extracts from different types of human, freshly isolated peripheral blood APC (CD1c-DC and B lymphocytes). Consistent with the analysis on the protein and activity levels, AEP, CatS, and CatD dominated MBP turnover in MO-DC-derived lysosomes, cleaving at the predicted sites. The cleavage pattern largely resembled that observed with human BLC, where also AEP dominated turnover of the autoantigen and thus controlled immunity by selective destruction of its dominant epitope (13, 16). CD1c-DC, however, contained significantly less CatS activity and AEP was virtually absent from this cell type, while only CatD reached amounts comparable to MO-DC. The molecular pattern of proteolytic degradation of MBP by CD1c-DC-derived lysosomes was entirely consistent with these observations: CatD-mediated cleavage was still observed, whereas cleavage products indicative of CatS or AEP were absent. CD1c-DC-derived lysosomal extracts resulted in a significantly more efficient MBP breakdown than MO-DC-derived lysosomes

and showed the same proteolytic pattern (cleavage at 90FK91 and 114FS115) as primary human B lymphocytes. In human B cells, which also lack AEP activity, this cleavage is dominated by CatG present in amounts comparable to primary monocytes (17). We demonstrate here that CD1c-DC also contain sizable amounts of active CatG, which controls MBP processing in vitro. This is supported not only by blocking MBP turnover with a specific CatG inhibitor, but also by affinity labeling of active CatG and by determining the turnover of an established fluorogenic CatG substrate, although the latter it is not specific for CatG (26). As an additional line of evidence, turnover of MOG by CD1c-DC-derived lysosomal extracts resulted in a proteolytic pattern whose major processing sites were strikingly similar to those obtained when MOG was digested with purified CatG in vitro, also supporting the dominant activity of CatG in this type of DC. There is still a theoretical possibility, however, that not CatG itself, but an as yet unidentified CatG-like protease in CD1c-DC is responsible for the effects observed. This hypothetical protease must then very closely resemble CatG, i.e., have a very similar substrate specificity and molecular mass. If true, this would be not unlike the situation observed with CatS, CatL, and CatV: all three proteases show a very similar substrate specificity, yet are distinct in their structure, tissue distribution, regulation, and function (11, 35–37). However, the bulk of evidence and especially the lack of possible known candidates for such a novel CatG-like protease makes it far more likely that CatG is the responsible enzyme. Given the very low content of papain-like cysteine Cat along with the absence of AEP, in lysosomes from CD1c-DC, any single additional endoprotease, like CatG, might reach functional importance even if expressed at low amounts, because its effects are less likely to be overridden by the activity of the other established processing enzymes. The relatively large amount of CatG protein in the absence of active CatG or CatG-related processing products derived from MBP or MOG suggest the presence of CatG inhibitors in MO-DC. OVA-related serpins comprise a growing class of, to date, 13 known efficient inhibitors of serine proteases including CatG, whose members are present also in monocytes (40). The

5948 expression and function of serpins and other serine protease inhibitors in different types of DC have not been comprehensively addressed so far, but there is evidence for differential expression of the serine protease inhibitor 6 during DC activation (41). The OVA-related serpin SCCA2 is a potent CatG inhibitor and is down-regulated by IL-4 (42). The pathway of MHC class II invariant chain processing by MHC class II-associated endocytic proteases is relatively well characterized in different types of APC. It shows only minor variations in the proteolytic breakdown of MHC class II invariant chain and ultimately the CLIP peptide is generated in all types of APC (11). In contrast, the degradation of autoantigens such as MBP or MOG, as shown here, results in entirely different proteolytic pathways when lysosomes from different types of APC are analyzed in vitro. Two major proteolytic pathways appear to exist: BLC, monocytes, and MO-DC degrade MBP in a fashion dominated by AEP, CatS, and CatD so that processing can be eliminated by inhibition of cysteine and aspartate proteases (e.g., by addition of IAA and pepstatin A), while processing in B lymphocyte- or CD1c-DC-derived lysosomes is not affected by these agents but inhibited by addition of the serine protease inhibitor PMSF (this study and Ref. 17). Although this type of in vitro processing study cannot completely substitute for the analysis in living, intact cells, it has been demonstrated to generate valid predictions regarding the dominant (unlocking) proteolytic enzyme, its cleavage site within the Ag, the consequences of this attack for the generation or destruction of immunogenic epitopes, and hence for the activation of specific T cells (13, 16). It will be important to establish whether these different processing pathways result in changes in the peptide pool presented from a given Ag in vivo and to what extent this can contribute to the induction of immunity vs tolerance in the periphery. Interestingly, lysosomes from primary microglia can result in either proteolytic pathway, depending on whether or not the cells have been exposed to IFN-␥ (T. Buster and C. Driessen, manuscript in preparation). Similar to the role of AEP in the proteolytic pathway of MBP in BLC (13), CatG directly eliminates the integrity of the major immunodominant epitopes of both MBP and MOG in lysosomes from CD1c-DC. At present, it is speculative to interpret this finding with respect to the pathogenesis of MS: the emerging model for self-tolerance includes that quiescent DC that bear self-components are required for the continuous maintenance of self-tolerance (7, 8, 10, 43). The functional and phenotypic relationship between human CD1c-DC and native DC in human lymphoid organs is only poorly understood; however, human CD1c-DC display a phenotype of “immature” or “quiescent” DC (i.e., low for surface class II and costimulatory molecules), similar to the phenotype of the bulk of native DC found in murine and human lymphoid organs (43– 46). Direct destruction of the major immunogenic MBP epitope MBP85–99 or the MOG epitope MOG99 –107 by the dominant CatG-like lysosomal protease in CD1c-DC could result in poor presentation of this region by tolerogenic DC, which might be one reason for the immunodominance of this MBP region during multiple sclerosis in vivo. If this was correct, specific inhibition of this protease in the periphery would be an attractive target to restore MBP tolerance in MS. However, when interpreting this type of data with respect to the situation in vivo, it must be taken into account that both autoantigens analyzed were expressed by bacteria and are therefore likely not to be in their native confirmation. The rules of unfolding and reducing complex autoantigen in the MHC class II compartment are poorly understood, but they might affect the pathway of proteolytic processing and the selection of the sites that are exposed to protease attack.

Ag PROCESSING BY BLOOD DC MO-DC generated in vitro represent a limited model to analyze Ag processing and the cell biology of human DC in general. Studies in well-defined types of primary human DC are warranted to understand the role of different types of DC in the induction of immunity or tolerance.

Acknowledgments We greatly acknowledge C. Watts (University of Dundee, Scotland, U.K.) for providing anti-AEP reagents and Matt Bogyo (Stanford University) for the DAP affinity probe. Steven Verhelst was of significant help for the DAP-labeling procedure and Simone Poeschel for performing the RT-PCR experiments.

Disclosures The authors have no financial conflict of interest.

References 1. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392: 245–252. 2. Liu, Y. J., H. Kanzler, V. Soumelis, and M. Gilliet. 2001. Dendritic cell lineage, plasticity and cross-regulation. Nat Immunol. 2: 585–589. 3. Trombetta, E. S., M. Ebersold, W. Garrett, M. Pypaert, and I. Mellman. 2003. Activation of lysosomal function during dendritic cell maturation. Science 299: 1400 –1403. 4. Kalinski, P., C. M. Hilkens, E. A. Wierenga, and M. L. Kapsenberg. 1999. T-cell priming by type-1 and type-2 polarized dendritic cells: the concept of a third signal. Immunol. Today. 20: 561–567. 5. Grabbe, S., E. Kampgen, and G. Schuler. 2000. Dendritic cells: multi-lineal and multi-functional. Immunol. Today. 21: 431– 433. 6. Dzionek, A., A. Fuchs, P. Schmidt, S. Cremer, M. Zysk, S. Miltenyi, D. W. Buck, and J. Schmitz. 2000. BDCA-2, BDCA-3, and BDCA-4: three markers for distinct subsets of dendritic cells in human peripheral blood. J. Immunol. 165: 6037– 6046. 7. Albert, M. L., M. Jegathesan, and R. B. Darnell. 2001. Dendritic cell maturation is required for the cross-tolerization of CD8⫹ T cells. Nat. Immunol. 2: 1010 –1017. 8. Shortman, K., and W. R. Heath. 2001. Immunity or tolerance? That is the question for dendritic cells. Nat. Immunol. 2: 988 –989. 9. Dhodapkar, M. V., R. M. Steinman, J. Krasovsky, C. Munz, and N. Bhardwaj. 2001. Antigen-specific inhibition of effector T cell function in humans after injection of immature dendritic cells. J. Exp. Med. 193: 233–238. 10. Steinman, R. M., D. Hawiger, K. Liu, L. Bonifaz, D. Bonnyay, K. Mahnke, T. Iyoda, J. Ravetch, M. Dhodapkar, K. Inaba, and M. Nussenzweig. 2003. Dendritic cell function in vivo during the steady state: a role in peripheral tolerance. Ann. NY Acad. Sci. 987: 15–25. 11. Villadangos, J. A., and H. L. Ploegh. 2000. Proteolysis in MHC class II antigen presentation: who’s in charge? Immunity 12: 233–239. 12. Antoniou, A. N., S. L. Blackwood, D. Mazzeo, and C. Watts. 2000. Control of antigen presentation by a single protease cleavage site. Immunity 12: 391–398. 13. Manoury, B., D. Mazzeo, L. Fugger, N. Viner, M. Ponsford, H. Streeter, G. Mazza, D. C. Wraith, and C. Watts. 2002. Destructive processing by asparagine endopeptidase limits presentation of a dominant T cell epitope in MBP. Nat. Immunol. 3: 169 –174. 14. Watts, C., C. X. Moss, D. Mazzeo, M. A. West, S. P. Matthews, D. N. Li, and B. Manoury. 2003. Creation versus destruction of T cell epitopes in the class II MHC pathway. Ann. NY Acad. Sci. 987: 9 –14. 15. Steinman, L. 2001. Multiple sclerosis: a two-stage disease. Nat. Immunol. 2: 762–764. 16. Beck, H., G. Schwarz, C. J. Schroter, M. Deeg, D. Baier, S. Stevanovic, E. Weber, C. Driessen, and H. Kalbacher. 2001. Cathepsin S and an asparaginespecific endoprotease dominate the proteolytic processing of human myelin basic protein in vitro. Eur. J. Immunol. 31: 3726 –3736. 17. Burster, T., A. Beck, E. Tolosa, V. Marin-Esteban, O. Rotzschke, K. Falk, A. Lautwein, M. Reich, J. Brandenburg, G. Schwarz, et al. 2004. Cathepsin G, and not the asparagine-specific endoprotease, controls the processing of myelin basic protein in lysosomes from human B lymphocytes. J. Immunol. 172: 5495–5503. 18. Forsthuber, T. G., C. L. Shive, W. Wienhold, K. de Graaf, E. G. Spack, R. Sublett, A. Melms, J. Kort, M. K. Racke, and R. Weissert. 2001. T cell epitopes of human myelin oligodendrocyte glycoprotein identified in HLA-DR4 (DRB1*0401) transgenic mice are encephalitogenic and are presented by human B cells. J. Immunol. 167: 7119 –7125. 19. Lautwein, A., T. Burster, A. M. Lennon-Dumenil, H. S. Overkleeft, E. Weber, H. Kalbacher, and C. Driessen. 2002. Inflammatory stimuli recruit cathepsin activity to late endosomal compartments in human dendritic cells. Eur. J. Immunol. 32: 3348 –3357. 20. Fiebiger, E., P. Meraner, E. Weber, I. F. Fang, G. Stingl, H. Ploegh, and D. Maurer. 2001. Cytokines regulate proteolysis in major histocompatibility complex class II-dependent antigen presentation by dendritic cells. J. Exp. Med. 193: 881– 892. 21. Lennon-Dumenil, A. M., A. H. Bakker, R. Maehr, E. Fiebiger, H. S. Overkleeft, M. Rosemblatt, H. L. Ploegh, and C. Lagaudriere-Gesbert. 2002. Analysis of


22.

23.

24. 25.

26.

27.

28.

29.

30.

31.

32.

33.

protease activity in live antigen-presenting cells shows regulation of the phagosomal proteolytic contents during dendritic cell activation. J. Exp. Med. 196: 529 –540. Weissert, R., J. Kuhle, K. L. de Graaf, W. Wienhold, M. M. Herrmann, C. Muller, T. G. Forsthuber, K. H. Wiesmuller, and A. Melms. 2002. High immunogenicity of intracellular myelin oligodendrocyte glycoprotein epitopes. J. Immunol. 169: 548 –556. Pluger, E. B., M. Boes, C. Alfonso, C. J. Schroter, H. Kalbacher, H. L. Ploegh, and C. Driessen. 2002. Specific role for cathepsin S in the generation of antigenic peptides in vivo. Eur. J. Immunol. 32: 467– 476. Okazaki, K., H. Yamada, and T. Imoto. 1985. A convenient S-2-aminoethylation of cysteinyl residues in reduced proteins. Anal. Biochem. 149: 516 –520. Schroter, C. J., M. Braun, J. Englert, H. Beck, H. Schmid, and H. Kalbacher. 1999. A rapid method to separate endosomes from lysosomal contents using differential centrifugation and hypotonic lysis of lysosomes. J. Immunol. Methods. 227: 161–168. Owen, C. A., M. A. Campbell, P. L. Sannes, S. S. Boukedes, and E. J. Campbell. 1995. Cell surface-bound elastase and cathepsin G on human neutrophils: a novel, non-oxidative mechanism by which neutrophils focus and preserve catalytic activity of serine proteinases. J. Cell Biol. 131: 775–789. Van Swieten, P. F., R. Maehr, A. M. Van Den Nieuwendijk, B. M. Kessler, M. Reich, C. S. Wong, H. Kalbacher, M. A. Leeuwenburgh, C. Driessen, G. A. Van Der Marel, et al. 2004. Development of an isotope-coded activitybased probe for the quantitative profiling of cysteine proteases. Bioorg. Med. Chem. Lett. 14: 3131–3134. Greenbaum, D., K. F. Medzihradszky, A. Burlingame, and M. Bogyo. 2000. Epoxide electrophiles as activity-dependent cysteine protease profiling and discovery tools. Chem. Biol. 7: 569 –581. Lautwein, A., M. Kraus, M. Reich, T. Burster, J. Brandenburg, H. S. Overkleeft, G. Schwarz, W. Kammer, E. Weber, H. Kalbacher, et al. 2004. Human B lymphoblastoid cells contain distinct patterns of cathepsin activity in endocytic compartments and regulate MHC class II transport in a cathepsin S-independent manner. J. Leukocyte Biol. 75: 844 – 855. Abuelyaman, A. S., D. S. Jackson, D. Hudig, S. L. Woodard, and J. C. Powers. 1997. Synthesis and kinetic studies of diphenyl 1-(N-peptidylamino)alkanephosphonate esters and their biotinylated derivatives as inhibitors of serine proteases and probes for lymphocyte granzymes. Arch. Biochem. Biophys. 344: 271–280. Greiner, A., A. Lautwein, H. S. Overkleeft, E. Weber, and C. Driessen. 2003. Activity and subcellular distribution of cathepsins in primary human monocytes. J. Leukocyte Biol. 73: 235–242. Rivera-Marrero, C. A., J. Stewart, W. M. Shafer, and J. Roman. 2004. The downregulation of cathepsin G in THP-1 monocytes after infection with Mycobacterium tuberculosis is associated with increased intracellular survival of bacilli. Infect. Immun. 72: 5712–5721. Turley, S. J., K. Inaba, W. S. Garrett, M. Ebersold, J. Unternaehrer, R. M. Steinman, and I. Mellman. 2000. Transport of peptide-MHC class II complexes in developing dendritic cells. Science 288: 522–527.

5949 34. Pierre, P., and I. Mellman. 1998. Developmental regulation of invariant chain proteolysis controls MHC class II trafficking in mouse dendritic cells. Cell 93: 1135–1145. 35. Tolosa, E., W. Li, Y. Yasuda, W. Wienhold, L. K. Denzin, A. Lautwein, C. Driessen, P. Schnorrer, E. Weber, S. Stevanovic, et al. 2003. Cathepsin V is involved in the degradation of invariant chain in human thymus and is overexpressed in myasthenia gravis. J. Clin. Invest. 112: 517–526. 36. Nakagawa, T., W. Roth, P. Wong, A. Nelson, A. Farr, J. Deussing, J. A. Villadangos, H. Ploegh, C. Peters, and A. Y. Rudensky. 1998. Cathepsin L: critical role in Ii degradation and CD4 T cell selection in the thymus. Science 280: 450 – 453. 37. Shi, G. P., J. A. Villadangos, G. Dranoff, C. Small, L. Gu, K. J. Haley, R. Riese, H. L. Ploegh, and H. A. Chapman. 1999. Cathepsin S required for normal MHC class II peptide loading and germinal center development. Immunity 10: 197–206. 38. Pol, E., and I. Bjork. 2003. Contributions of individual residues in the N-terminal region of cystatin B (stefin B) to inhibition of cysteine proteinases. Biochim. Biophys. Acta 1645: 105–112. 39. Ishri, R. K., S. Menzies, P. Hersey, and G. M. Halliday. 2004. Rapid downregulation of antigen processing enzymes in ex vivo generated human monocyte derived dendritic cells occur endogenously in extended cultures. Immunol. Cell Biol. 82: 239 –246. 40. Remold-O’Donnell, E., J. Chin, and M. Alberts. 1992. Sequence and molecular characterization of human monocyte/neutrophil elastase inhibitor. Proc. Natl. Acad. Sci. USA 89: 5635–5639. 41. Medema, J. P., D. H. Schuurhuis, D. Rea, J. van Tongeren, J. de Jong, S. A. Bres, S. Laban, R. E. Toes, M. Toebes, T. N. Schumacher, et al. 2001. Expression of the serpin serine protease inhibitor 6 protects dendritic cells from cytotoxic T lymphocyte-induced apoptosis: differential modulation by T helper type 1 and type 2 cells. J. Exp. Med. 194: 657– 667. 42. Yuyama, N., D. E. Davies, M. Akaiwa, K. Matsui, Y. Hamasaki, Y. Suminami, N. L. Yoshida, M. Maeda, A. Pandit, J. L. Lordan, et al. 2002. Analysis of novel disease-related genes in bronchial asthma. Cytokine 19: 287–296. 43. Shortman, K., and Y. J. Liu. 2002. Mouse and human dendritic cell subtypes. Nat. Rev. Immunol. 2: 151–161. 44. Wilson, N. S., D. El-Sukkari, G. T. Belz, C. M. Smith, R. J. Steptoe, W. R. Heath, K. Shortman, and J. A. Villadangos. 2003. Most lymphoid organ dendritic cell types are phenotypically and functionally immature. Blood 102: 2187–2194. 45. Summers, K. L., B. D. Hock, J. L. McKenzie, and D. N. Hart. 2001. Phenotypic characterization of five dendritic cell subsets in human tonsils. Am. J. Pathol. 159: 285–295. 46. McIlroy, D., C. Troadec, F. Grassi, A. Samri, B. Barrou, B. Autran, P. Debre, J. Feuillard, and A. Hosmalin. 2001. Investigation of human spleen dendritic cell phenotype and distribution reveals evidence of in vivo activation in a subset of organ donors. Blood 97: 3470 –3477.