Cathepsin D Processing Differently Affected by Cathepsin B And

Home , Cathepsin A, Cathepsin B, Cathepsin G, CXCL7

Function of Liver Activation-Regulated Chemokine/CC Chemokine Ligand 20 Is Differently Affected by Cathepsin B and Cathepsin D Processing This information is current as of September 28, 2021. Lara Hasan, Luca Mazzucchelli, Mark Liebi, Maddalena Lis, Robert E. Hunger, Angus Tester, Christopher M. Overall and Marlene Wolf J Immunol 2006; 176:6512-6522; ; doi: 10.4049/jimmunol.176.11.6512 Downloaded from http://www.jimmunol.org/content/176/11/6512

References This article cites 58 articles, 27 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/176/11/6512.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Function of Liver Activation-Regulated Chemokine/CC Chemokine Ligand 20 Is Differently Affected by Cathepsin B and Cathepsin D Processing1

Lara Hasan,* Luca Mazzucchelli,*§ Mark Liebi,† Maddalena Lis,* Robert E. Hunger,‡ Angus Tester,†¶ Christopher M. Overall,†¶ and Marlene Wolf2†

Chemokine processing by proteases is emerging as an important regulatory mechanism of leukocyte functions and possibly also of cancer progression. We screened a large panel of chemokines for degradation by cathepsins B and D, two proteases involved in tumor progression. Among the few substrates processed by both proteases, we focused on CCL20, the unique chemokine ligand of CCR6 that is expressed on immature dendritic cells and subtypes of memory lymphocytes. Analysis of the cleavage sites demonstrate that cathepsin B speciﬁcally cleaves off four C-terminally located amino acids and generates a CCL201–66 isoform Downloaded from with full functional activity. By contrast, cathepsin D totally inactivates the chemotactic potency of CCL20 by generating

CCL201–55, CCL201–52, and a 12-aa C-terminal peptide CCL2059–70. Proteolytic cleavage of CCL20 occurs also with chemokine bound to glycosaminoglycans. In addition, we characterized human melanoma cells as a novel CCL20 source and as cathepsin producers. CCL20 production was up-regulated by IL-1␣ and TNF-␣ in all cell lines tested, and in human metastatic melanoma cells. Whereas cathepsin D is secreted in the extracellular milieu, cathepsin B activity is conﬁned to cytosol and cellular membranes. Our studies suggest that CCL20 processing in the extracellular environment of melanoma cells is exclusively mediated by http://www.jimmunol.org/ cathepsin D. Thus, we propose a model where cathepsin D inactivates CCL20 and possibly prevents the establishment of an effective antitumoral immune response in melanomas. The Journal of Immunology, 2006, 176: 6512–6522.

hemokines contribute to a variety of immunoregulatory mokines showing that matrix metalloproteases (MMPs)3 and also and proinflammatory responses by control of leukocyte dipeptidylpeptidase IV/CD26 are potent modifiers of chemokine C recruitment during inflammation (1–3). Beyond this pri- function in addition to their roles in inflammation and wound heal- mary function, chemokines also affect angiogenesis, angiostatic ing (1). MMP activity activates the CXCR2 murine chemokine regulation mechanisms, hemopoiesis, lymphatic organogenesis, CXCL5/LIX in vitro and in vivo (12), inactivates CXCL12 (13, and, through the migration of cancer cells expressing chemokine 14), and generates antagonists from all MCP chemokines (15). by guest on September 28, 2021 receptors to chemokine-secreting organs, tumor metastasis (3, 4). A causal link between chemokine function and organ trophism Recently, several chemokines were described to possess also an- in breast cancer was shown for CXCR4 and its cognate ligand timicrobial activities (5). Chemokine activity is modulated by sev- CXCL12 (16). Thus, the interrelationship between chemokines eral cytokines that directly or indirectly regulate chemokine ex- and members of the cathepsin family, such as cathepsin D pression and/or expression of their receptors (1). An increasing (Cath-D) and cathepsin B (Cath-B), may play an important role in body of evidence suggests that chemokine clearing by proteases tumor biology. Indeed, both cathepsins not only facilitate tumor may also strongly affect their activity (1). In this context, a specific progression (17) but may have a prognostic value in patients with role was recently suggested for cathepsins that cleave and inacti- breast cancer (17, 18) and malignant melanoma (19). We therefore vate CXCL12/stromal cell-derived factor-1␣ (6) and CCL3/MIP- reasoned that the generation of an antitumoral immune response or 1␣, CCL4/MIP-1␤ and CCL21/SLC (7). Conversely, the activity cancer cell motility may be modulated by Cath-D and Cath-B gen- of CXCL5/epithelial cell-derived neutrophil-activating protein-78 erated by the tumor or reactive stroma. Among the novel substrates (8) and CXCL8/IL-8 (9, 10), and CXCL7/neutrophil-activating identified, we focused on CCL20 (also known as liver and activa- peptide 2 (11) is potentiated upon interaction with cathepsins. This tion-regulated chemokine (LARC); MIP-3␣ or exodus-1), which is work builds on initial studies on the effects of proteases and che- a 9-kDa CC-type chemokine expressed constitutively in lung and liver (20), in keratinocytes in the epidermal layer of the skin (21), and in the intestinal mucosa (22–24). Several arguments indicate *Institute of Pathology, †Theodor-Kocher Institute, and ‡Department of Dermatology, CCL20 as a potential important player in the chemokine network. § University of Bern, Bern, Switzerland; Istituto di Patologia, Locarno, Switzerland; First, CCL20 is the unique chemokine ligand of CCR6 and does and ¶Department of Biochemistry and Molecular Biology, Centre for Blood Research and the Canadian Institutes for Health Research Group in Matrix Dynamics, Univer- not show the promiscuity of receptor binding characteristic of sity of British Columbia, Vancouver, British Columbia, Canada other chemokines (25, 26). Second, CCL20 not only triggers an Received for publication May 31, 2005. Accepted for publication March 21, 2006. innate immune response through the attraction of NK cells (27) but The costs of publication of this article were defrayed in part by the payment of page is also involved in adaptative immune responses, primarily by at- charges. This article must therefore be hereby marked advertisement in accordance tracting immature dendritic cells, allowing them to take up foreign with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by grants from the Swiss National Science Foundation (Grant 3100-066804) and the Bernische Krebsliga. 3 Abbreviations used in this paper: MMP, matrix metalloprotease; Cath-D, cathepsin 2 Address correspondence and reprint requests to Dr. Marlene Wolf, Theodor-Kocher D; Cath-B, cathepsin B; Cath-H, cathepsin H; CS, chondroitin sulfate; GAG, glycos- Institute, University of Bern, Freiestrasse 1, 3012 Bern, Switzerland. E-mail address: aminoglycan; EIA/RIA, enzyme immunoassay/radioimmunoassay; pAB, 4-amino- [email protected] benzoic acid; AMC, 7-amino-4-methylcoumarin.

Ags and to mature (28, 29), and by attracting memory T and mem- CCL20 in PBSϩ containing 0.05% Tween 20 and 1% BSA was added to ory B cells to the site of inflammation (20, 30, 31). Furthermore, the GAG-coated wells and incubated for 2 h. After washing, the wells were ␮ CCL20 has antimicrobial activities, similarly to defensins, which incubated overnight at 37°C with 0.25 g/ml Cath-D or Cath-B, and washed again, and CCL20 was probed with biotinylated goat anti-human also bind to CCR6, and can therefore contribute to direct antimi- CCL20. Bound CCL20 was then measured using streptavidin-conjugated crobial defenses (5). Lastly, CCL20 expression has been described alkaline phosphatase and Blue Phos as described above. in several leukemia/lymphoma cell lines, and in a variety of human ϩ Intracellular calcium ([Ca2 ] ) mobilization neoplasms, such as colorectal adenocarcinoma, lung carcinoma, i 2ϩ and malignant glioma (32, 33). In this study, we demonstrate for [Ca ]i changes were measured in CCR6-expressing mouse 300-19 pre-B 6 the first time chemokines as substrates for Cath-B, and we char- cells (35) loaded with 0.1 nmol of fura 2-AM per 10 cells as described previously (36). acterize the processing of CCL20 by Cath-D and Cath-B. We further demonstrate that cleavage of CCL20 by these proteases is Cell migration retained when the chemokine is bound to extracellular matrix pro- Chemotaxis was assayed in 48-well chambers (Neuro Probe) using 3-␮m teins, and we characterized in functional assays the activity of pore size polycarbonate filters (Osmonics). Briefly, chemokines in chemo- degradation products. Using human malignant melanoma as a taxis buffer (RPMI 1640, 20 mM HEPES (pH 7.2), containing 1% pas- 5 model, we propose that neoplastic cells potentially exert autocrine teurized human plasma protein) were added to the lower wells, and 10 cells were resuspended in the same buffer to the upper wells and incubated regulatory mechanisms in the context of chemokine-protease in- for 90 min at 37°C. Cells migrating to the lower side of the filter were teractions by modulating the migratory activities of CCL20 toward stained and counted in five high-power fields. Cells migrating into the responsive cells of the antitumoral immune response. lower wells were also counted. Studies were performed at least three times

with triplicate wells per experiment. Downloaded from Materials and Methods To measure the effect of Cath-D on the activity of immobilized CCL20, Reagents, cell lines, and bacteria the polycarbonate filters were first coated for 75 min at 37°C with 50 nM CCL20 in 50 mM sodium citrate, 50 mM NaCl (pH 4.0), and then incu- Cath-B, Cath-D, cathepsin H (Cath-H), and antisera to human Cath-B or bated overnight with 0.1–1 ␮g/ml Cath-D or Cath-H as control. The filters Cath-D were obtained from Athens Research and Technology. Biotinylated were washed with PBS, air-dried for 30 min, and rehydrated for 20 min in goat anti-human CCL20 was from R&D Systems. IL-1␣, IL-1␤, and fresh chemotaxis buffer. Then, the filters were mounted into the chemotaxis 5 TNF-␣ were obtained from PeproTech. Synthetic chemokines were pre- chamber as indicated above and 10 cells were added to the upper wells. The lower chamber contained chemotaxis buffer with no chemoattractant. http://www.jimmunol.org/ pared as described previously (34). CCL2059–70 was prepared by C. Servis (Protein and Peptide Chemistry Facility, University of Lausanne, Switzer- Chemotaxis was then performed as indicated above. land). Pasteurized human plasma protein was provided by ZLB Bioplasma. Antibacterial activity Chondroitin sulfate (CS), other glycosaminoglycan (GAG) components, benzoylxycarbonyl (Z)-Arg-Arg-AMC hydrochloride (substrate for Cath- To measure bactericidal activities, ϳ1 ϫ 104 exponentially growing E. coli B), leupeptin and pepstatin A were purchased from Sigma-Aldrich. d-F- BL21 cells were incubated with different concentrations of synthetic S(benzoyl)-F-F-A-A-pAB (substrate for Cath-D) was from Bachem. CA- ␮ CCL20 or CCL2059–70 in 100 l of 10 mM potassium phosphate buffer 074, a specific inhibitor for Cath-B, was obtained from Serva. (pH 7.4) supplemented with 1% (v/v) Luria broth medium. The mixture Aminopeptidase M was from Roche Applied Science. Human melanoma was incubated for3hat37°C (250 rpm), diluted in previous buffer, and cell lines A375 (solid tumor metastatic line; ATCC CRL-1619), A2058 plated on Luria broth agar plates as described previously (5). Surviving (lymph node metastatic line; ATCC CRL-11147), and SK-MEL-2 (skin bacteria were quantitated as CFU/milliliter on plates after incubation at by guest on September 28, 2021 metastatic line; ATCC HTB-68) were from American Type Culture Col- 33°C for 14 h. Bactericidal activity was expressed as the ratio of colonies lection. Escherichia coli BL21 was obtained from S. Didichenko (Institute counted to the number of colonies on a control plate (with no peptide). The of Immunology, Bern, Switzerland). LD90 is the concentration of protein that reduces the number of colonies by Chemokine cleavage by cathepsins and characterization of 90%. All results represent mean values of duplicates from at least three independent experiments. cleavage products Patients Five micromolar chemokines were incubated with 50 ng of Cath-D or Cath-B for 90 min at 37°C in 50 ␮l of buffer I (50 mM sodium citrate, 50 The three patients included in this study had histologically confirmed ma- mM NaCl; pH 4.0) or buffer II (PBS containing 4 mM EDTA and 2 mM lignant melanoma with skin metastases removed by surgery. The protocol L-cysteine; pH 6.8), respectively. In control experiments, protease inhibi- was approved by the Medical Ethics Committee of the Canton of Bern, tors were preincubated with the enzyme for 3 min at 37°C. Reactions were Switzerland. All three patients signed informed consent before study entry. stopped by heating to 90°C, and the chemokine cleavage products were separated by SDS-PAGE in 10–20% Tris-Tricine gradient polyacrylamide Melanoma cells gels and stained with Coomassie blue. Melanoma cell lines were seeded into 100 ϫ 20-mm tissue culture plates For cleavage site determination, the proteolytic fragments were acidified (Sarstedt) and fed with fresh culture medium (DMEM supplemented with with trifluoroacetic acid and then loaded onto a reversed-phase C2/C18 15% FCS; Invitrogen) every 2–3 days until 80% confluence. After aspira- column (Amersham Biosciences). Peptides were eluted with a linear gra- tion of culture medium followed by washing with PBS, cells were supplied dient of 0–80% acetonitrile in 0.1% trifluoroacetic acid. The molecular with 6.5 ml of DMEM containing 10% FCS and treated for 1–3 days with masses were determined by MALDI-TOF mass spectrometry using a Voy- IL-1␣ and TNF-␣. When secreted chemokines were further purified, cells ager-DE STR system. The cleaved scissile bonds were then confirmed by were stimulated with DMEM containing the proinflammatory cytokines Edman N-terminal sequencing. and 2% FCS. Chemokine binding to GAG Skin metastases were surgically removed from melanoma patients. After removing adipose and fibrous tissue, the metastasis were cut into small Ninety-six-well enzyme immunoassay/radioimmunoassay (EIA/RIA) pieces and single-cell suspension was prepared by carefully disrupting the plates (Costar) were coated overnight at 4°C with different concentrations tissue pieces between the frosted ends of two microscope slides. Contam-

of GAG. The wells were washed with PBS containing 1 mM CaCl2 and 1 inating lymphocytes and macrophages/monocytes were removed by incu- ϩ mM MgCl2 (PBS ), and then blocked with 5% FCS for2hatroom tem- bating the cell suspension with purified monoclonal mouse anti-CD3 and perature. CCL20 in PBS containing 1% FCS was added and incubated for anti-CD14 Ab (BD Pharmingen) as first-stage reagent and sheep anti- 2 h. After washing, the wells were treated with biotinylated goat anti- mouse IgG-coated Dynabeads (Dynal Biotech) as second-stage reagent. human CCL20 (0.55 ␮g/ml) in PBSϩ containing 1% FCS and 0.05% Depletion was performed according to the manufacturer’s indication. Con- Tween 20 for 1 h, followed by incubation with streptavidin-conjugated tamination of the final melanoma cell preparation by CD3- and CD14- alkaline phosphatase (Promega, Catalys AG). To quantify the reaction, positive cells varied from 0.5 to 12% as assessed by FACScan analysis. Blue Phos (Kirkegaard Perry Laboratories Bioreba) was added, and the Melanoma cells adhered overnight in a 6-well plate (Costar 3506) at a ϫ 6 OD595 was read in a microtiter plate reader. To determine the effect of concentration of 1–2 10 /well in 15% FCS. Adherent and suspended Cath-D or Cath-B on chemokines bound to GAG, the wells were coated cells were stimulated for 48 h with 2 ml DMEM/10% FCS containing with GAG, washed as described above, and then blocked with 5% BSA. IL-1␣ and TNF-␣. 6514 REGULATION OF CCL20 ACTIVITY BY CATHEPSIN PROCESSING

Puriﬁcation of CCL20 from culture medium RNA preparation and RT-PCR A375, A2058, and SK-MEL-2 cells were grown for 24, 48, and 72 h as Total RNA from cultured melanoma cells was extracted with RNeasy Midi described above. CCL20 was measured in the conditioned medium by a kit (Qiagen) according to the manufacturer’s instructions. After treatment sandwich-type ELISA using commercially available kit (R&D Systems). with RNase-free DNase, total RNA (1 ␮g) was reverse transcribed using

The detection range was typically between 20 pg/ml and 1 ng/ml. Super- oligo(dT)17 primer and AMV reverse transcriptase (Promega). Resulting natants were centrifuged for 20 min at 1800 ϫ g and then passed through first-strand DNA was amplified in a final volume of 20 ␮l containing 10 a 0.22-␮m filter before applying to a 1-ml HiTrap heparin column (Am- pmol of each primer and1UofTaq polymerase (Promega). The primers ϩ Ј Ј Ϫ Ј ersham Biosciences) equilibrated with 40 mM KHPO4, 20 mM NaCl, 1 used were the following: 5 -accatgtgctgtaccaagagtttg-3 and 5 -ctaaac mM EDTA, 1 M urea, and 5% glycerol (pH 7.2). Proteins were eluted cctccatgatgtgcaagtga-3Ј for human CCL20; ϩ5Ј-gtcaacggatttggtcgtatt-3Ј stepwise with 0.2, 0.5, and 1 M NaCl. CCL20-containing fractions were and Ϫ5Ј-agtcttctgggtggcagtgat-3Ј for human G3PDH. Amplification con- identified by ELISA and further purified on a reversed-phase C2/C18 col- ditions were denaturation at 94°C for 30 s (5 min for the first cycle), umn (Amersham Biosciences) with a flow rate of 0.1 ml/min and a con- annealing at 60°C for 30 s, and extension at 72°C for 30 s (5 min for the tinuous gradient of 0–80% acetonitrile in 0.1% trifluoroacetic acid. last cycle) for 35 cycles, and 27 cycles for human G3PDH. Amplification CCL20-containing fractions were used for cell migration and Ca2ϩ-releas- products (10 ␮l each) were subjected to electrophoresis on 2% agarose and ing experiments. stained with ethidium bromide.

Melanoma cell membranes In situ hybridization Cells at 80% confluence were washed with PBS and then detached from the Formaldehyde-fixed and paraffin-embedded melanoma tissues from 10 pa- culture dish with trypsin. For the isolation of membranes, cells were cen- tients were retrieved from the archives of the Institute of Pathology, Uni- trifuged for 5 min at 180 ϫ g, washed with PBS, and resuspended in 2 ml versity of Bern (Bern, Switzerland). The human CCL20 cDNA in pBlue- of ice-cold hypotonic buffer containing 10 mM Tris-HCl and 0.5 mM scriptII KS vector was a gift from B. Moser (University of Bern). Digoxigenin-labeled riboprobes (Roche Applied Science) for CCL20 were Downloaded from MgCl2 (pH 7.6). After sonication, 0.25 vol of 10 mM Tris-HCl, 0.5 mM transcribed in vitro from linearized, gel-purified plasmids using T7 poly- MgCl2, and 0.6 M NaCl (pH 7.6) was added to restore tonicity. Cell debris and nuclei were removed by centrifugation at 560 ϫ g; the supernatant was merase (antisense cRNA probe) and T3 polymerase (sense cRNA probe). adjusted to 5 mM EDTA and centrifuged at 150,000 ϫ g. The membrane In situ hybridization was performed as previously described (39), except pellet was resuspended in 100 ␮l of PBS and stored at Ϫ20°C until used that sections were prehybridized (30 min, 50°C) and hybridized (overnight, for CCL20 processing or Cath-B and Cath-D activity assays. Protein con- 50°C) in DIG Easy Rib (Roche Applied Science), and washed at 55°C in ϫ ϫ ϫ centration was determined by the bicinchoninic acid assay (Pierce). 2 SSC, 1 SSC, and 0.1 SSC. http://www.jimmunol.org/ Statistical analysis Cath-D and Cath-B activities All results represent mean values of triplicates. Values of p were calculated Cath-D activity was determined in cells and cellular fractions using a vari- by Student’s t test using Microsoft Excel software. Correlation coefficients ation of the method described by Heylen et al. (37). Briefly, the Cath-D were calculated by linear regression analysis. A value of p Ͻ 0.05 was substrate d-F-S(benzoyl)-F-F-A-A-pAB (1.4 mM) was incubated for4hat considered statistically significant. 37°C with membranes or cytosol (25-␮g protein equivalent) or conditioned medium (30 ␮l) in 0.1 M sodium citrate buffer (pH 4.0). This step was followed by incubation for2hat37°C with aminopeptidase M (7 ␮g/ml). The reaction was stopped by adding TCA (14% v/v final). Released 4-ami- nobenzoic acid (pAB) was measured at 546 nm. The activity of Cath-D is a expressed as picomoles of pAB produced per minute per 105 cells. Cath-D Table 1 Chemokine processing by Cath-D and Cath-B by guest on September 28, 2021 specificity was confirmed using, in parallel assays, 2 ␮M pepstatin A, an inhibitor of aspartic proteases. Cath-B activity in cells and cellular fractions Cath-D Cath-B was determined as described previously (38). Shortly, cells were first washed with PBS, and 0.5 ϫ 106 cells were then transferred to 96-well CXC chemokines plates and resuspended in PBS containing 4 mM EDTA and 2 mM L- CXCL1/GRO␣ ϪϪ cysteine (pH 6.8). Alternatively, 15 ␮g of membrane proteins or 200 ␮gof CXCL6/GCP-2 ϪϪ cytosolic protein or 200 ␮l of conditioned medium were used. The assay CXCL7/NAP-2 ϪϪ was initiated by the addition of 100 ␮M Z-Arg-Arg-AMC (Cath-B sub- CXCL8/IL-8 ϪϪ strate) in the absence or presence of 1 ␮M CA-074 and followed for 50 min CXCL9/Mig Ϫϩ in a fluorescence plate reader with an excitation wavelength of 355 nm and CXCL10/IP-10 Ϫϩ an emission wavelength of 460 nm. To measure intracellular Cath-B ac- CXCL11/I-TAC Ϫϩ tivity, assay buffer was supplemented with 0.1% Triton X-100. Activity CXCL12/SDF-1␣ Ϫϩ inhibited by CA-074 was considered to represent Cath-B activity. The ac- CXCL13/BCA-1 Ϫϩ tivity of Cath-B is expressed as picomoles of 7-amino-4-methylcoumarin CXCL14/BRAK Ϫϩ (AMC) produced per minute per 105 cells. A standard curve of free AMC CC chemokines or pAB was used for quantification. CCL1/I-309 ϪϪ CCL2/MCP-1 ϪϪ Immunoblotting CCL3/MIP-1␣ ϩϪ CCL4/MIP-1␤ ϩϪ Proteins were resolved by SDS-PAGE (12 or 10–20% gradient polyacryl- CCL5/RANTES ϪϪ amide) and electroblotted onto polyvinylidene difluoride membranes. CCL7/MCP-3 ϪϪ CCL20 was detected using polyclonal goat anti-human CCL20 (180 ng/ CCL8/MCP-2 ϪϪ ml), and Cath-B and Cath-D were detected with polyclonal rabbit anti- CCL11/Eotaxin-1 Ϫϩ human Cath-B or Cath-D (1:5000) Ab. CCL13/MCP-4 Ϫϩ CCL19/ELC Ϫϩ Immunohistochemistry CCL20/LARC ϩϩ CCL21/SLC ϩϩ Formalin-fixed, paraffin-embedded tissue samples were cut at 2–3 ␮m, CCL24/Eotaxin-2 ϪϪ dewaxed, rehydrated, and incubated with a polyclonal rabbit-anti-Cath-D CCL27/CTACK ϩϩ Ab (DakoCytomation) in TBS containing 0.5% casein and 5% normal goat serum, for 60 min at room temperature. Sections were then washed with aEach chemokine (5 ␮M) was incubated with 50 ng of Cath-D in buffer I (pH 4.0) TBS. Next, a 1/500 dilution of biotinylated goat-anti-rabbit Ig antiserum or Cath-B in buffer II (pH 6.8) at 37°C for 2 h. The cleavage products were separated (DakoCytomation) was applied for 45 min. The primary Ab was omitted in by SDS-PAGE (10–20% gradient) and stained with Coomassie blue. Abbreviations: BCA-1, B cell-attracting chemokine 1; BRAK, breast- and kidney- control sections. Thereafter, the sections were incubated with streptavidin- expressed chemokine; CTACK, cutaneous T cell-attracting chemokine; GCP-2, gran- conjugated alkaline phosphatase (DakoCytomation). Staining was devel- ulocyte chemotactic protein 2; GRO␣, growth-related oncogene ␣; IP-10, IFN-␥- oped with fuchsin-naphthol AS-BI (Sigma-Aldrich) for 30 min, and coun- inducible protein 10; I-TAC, IFN-inducible T cell ␣ chemoattractant; NAP-2, terstaining was performed with hematoxylin. neutrophil activating peptide 2; SDF-1␣, stromal cell-derived factor 1␣. The Journal of Immunology 6515

Results Inactivation of CCL20 after Cath-D but not Cath-B cleavage Processing of selected chemokines by Cath-D and Cath-B The degradation products of CCL20 were functionally tested by We tested a panel of 24 chemokines as potential substrates for cell migration and intracellular calcium mobilization experiments. Cath-D and Cath-B (Table I). In addition to the previously reported After incubation with Cath-D, CCL20 did not attract CCR6-trans- processing of CCL3, CCL4, and CCL21 (7), we now describe fected pre-B mouse cells. By contrast, exposure to Cath-B had no CCL20 and CCL27 as additional substrates for Cath-D. None of influence on the chemotactic activity of CCL20 (Fig. 2A). Migra- the CXC chemokines was processed by Cath-D. In addition to tion experiments were also performed in the presence of Cath-H, Cath-D, overexpression of Cath-B has also been associated with a protease that does not cleave CCL20. As expected, there were no poor prognosis in cancer. We tested therefore whether chemokines significant differences between experiments with control buffer might also be substrates for this protease. We found that Cath-B, and buffer containing Cath-H (data not shown). We next tested the when measured at pH 6.8, cleaved a large spectrum of chemokines, activity of the purified fragments resulting from Cath-D cleavage, including all non-ELR CXC chemokines (lacking Glu-Leu-Arg CCL201–55 and CCL201–52. CCL201–55 was entirely inactive and sequence preceding the first cysteine) and 6 of the 14 analyzed could neither mobilize intracellular Ca2ϩ (Fig. 2B) nor induce cell members of the CC family. CCL20, CCL21, and CCL27 were the migration (C). Furthermore, we could demonstrate that CCL20 only chemokines that were cleaved by both proteases. 1–55 does not block the receptor or act as an antagonist, because sub- Time courses of CCL20 processing by Cath-D and Cath-B are ϩ sequent addition of intact CCL20 yields a full Ca2 response (Fig. presented in Fig. 1A. Cleavage by both proteases led to stable 2B, right trace). The shorter peptide CCL20 was inactive as products that were not further degraded after an incubation time of 1–52 well (data not shown). By contrast, the product of Cath-B cleav- Downloaded from 6 h and 90 min, respectively. Cleavage of CCL20 by Cath-D oc- age, CCL20 retained full activity, mobilized intracellular Ca2ϩ curred at different sites, and five fragments were isolated by re- 1–66 versed-phase C2/C18 column chromatography (Fig. 1B). Analysis (Fig. 2B) and induced cell migration with similar potency as full- of these peptides by mass spectroscopy and N-terminal sequencing length CCL20 (Fig. 2C). Overall, the above results show that Cath-D inactivates CCL20, whereas removal of four C-terminal led to the identification of three N-terminal peptides, CCL201–19, amino acids by Cath-B has no functional consequences on CCL20 CCL201–52, and CCL201–55, an internal peptide, CCL2023–55,a indicating that the extreme C-terminal part of this chemokine is not http://www.jimmunol.org/ C-terminal peptide of 12 aa, CCL2059–70 and full-length CCL20. essential for proper CCR6 activation. We also assessed whether CCL201–19 and CCL2023–55 were held together by disulfides as demonstrated by MALDI analysis of this column fraction after cleaved CCL20 might use an alternative receptor than CCR6. reducing with DTT. Cleavage by Cath-B was very rapid and also Transfected murine cells expressing CXCR2, CXCR3, CXCR4, complete and generated the distinct cleavage product CCL201–66 CCR1, CCR5, and CCR7 were incubated with the cleavage prod-

(Fig. 1, B and C). ucts, CCL201–55 and CCL201–66, as well as intact CCL20 and by guest on September 28, 2021

FIGURE 1. CCL20 processing by Cath-D and Cath-B. A, Time course of CCL20 processing by Cath-D and Cath-B. CCL20 (5 ␮M) was incubated with 50 ng of Cath-D or Cath-B at 37°C for the indicated times in buffer I (pH 4.0) or buffer II (pH 6.8), respectively. Pep- statin A (2 ␮M) or CA-074 (5 ␮M) were used as inhibitors for Cath-D and Cath-B, respectively (I). The cleavage products were separated by SDS-PAGE and stained with Coomassie blue. B, Reverse phase-HPLC proﬁle of the cleavage products of CCL20 after Cath-D and Cath-B processing. C, CCL20 cleavage sites for Cath-D and Cath-B. The positions of the amino acids located at the cleavage sites are indicated. 6516 REGULATION OF CCL20 ACTIVITY BY CATHEPSIN PROCESSING

the control GAG chondroitin (Fig. 3A). Incubation of GAG with Cath-D did not affect subsequent CCL20 binding (data not shown). When GAG-immobilized CCL20 was treated with 0.25 ␮g/ml Cath-D, the chemokine no longer reacted with the polyclonal CCL20 Ab, suggesting proteolysis. It should be noted that in ELISA or dot-blot assays, polyclonal CCL20 Ab were able to recognize the CCL20 cleavage products after Cath-B but not Cath-D exposure (data not shown). Conversely, treatment with Cath-B, which is known to also have exopeptidase activity at acidic pH (41), did not signiﬁcantly affect the recognition of CCL20 by anti- CCL20 at acidic nor neutral pH (Fig. 3B). This result suggests that the last four residues of CCL20 are not included in the epitopes of the Ab. Incubation of polycarbonate ﬁlters with 50 nM CCL20 Downloaded from

FIGURE 2. Influence of cathepsin treatment on chemotactic and cal- http://www.jimmunol.org/ cium mobilization activity of CCL20. A, Migration of CCR6-transfected cells in response to CCL20 (control) and CCL20 after incubation for 90 min with buffer I (pH 4.0) and Cath-D or Cath-H, or with buffer II (pH 6.8) and Cath-B or Cath-H. Results are expressed as means Ϯ SD of the numbers of cells observed per five high-power fields. No migrated cells were detected in the absence of chemoattractant. Statistically significant decreases in migrated cell number below the control values are indicated by Ͻ 2ϩ ءء asterisks ( , p 0.01). B, [Ca ]i-dependent fluorescence changes in

CCR6-transfected cells by CCL20, CCL201–66, and CCL201–55. C, Migra- by guest on September 28, 2021 tion in response to increasing concentrations of CCL20, CCL201–55, and Ϯ CCL201–66. Results are expressed as means SD of the numbers of cells observed per five high-power fields. monitored for calcium-releasing activity. None of these receptors responded to CCL20 or to its truncated products. It has been demonstrated that CCL20 has potent antimicrobial activity similar to ␤-defensins (40), which is ascribed to positive charges in the C-terminal part (5, 40). We therefore tested whether the C-terminal peptide CCL2059–70 generated by Cath-D cleavage maintains the bactericidal activity of intact CCL20. Using a stan- FIGURE 3. GAG-bound CCL20 is cleaved by Cath-D. A, Binding of dard colony-forming assay with E. coli BL21 bacteria, we could CCL20 to CS-A. EIA/RIA plate wells were coated with 20, 50, or 100 ␮g/ml CS-A, or chondroitin as control. After blocking, increasing concen- show that CCL2059–70 killed E. coli in a dose-dependent manner ␮ ␮ trations of CCL20 (25, 50, 100, or 150 ng/ml) were added to the wells and with a LD50 of 3.2 g/ml, whereas a concentration of 0.53 g/ml full-length CCL20 was able to kill 50% of bacteria (data not incubated for 2 h. The binding of CCL20 to GAG was assessed by ELISA. Results are expressed as the mean of OD Ϯ SD. Statistically significant shown). The activity of CCL20 , the fragment obtained after 1–66 increases above 25 ng/ml CCL20 concentration values are indicated by p Ͻ 0.001). B, Cath-D ,ءءء p Ͻ 0.01; and ,ءء ;p Ͻ 0.05 ,ء) cleavage with Cath-B, was not significantly different from that of asterisks intact chemokine (data not shown). treatment inhibits recognition of CCL20 by anti-CCL20 Ab. EIA/RIA plate wells were coated with 100 ␮g/ml CS-A. After blocking, 150 ng/ml Binding of CCL20 to glycosaminoglycans does not prevent CCL20 was added to the wells, incubated for 2 h, and washed before processing by Cath-D treating with Cath-D or Cath-B (0.25 ␮g/ml) for 16 h at 37°C. CCL20 was Chemokine activity is further regulated by binding to GAGs lo- then assessed by ELISA. Controls with buffer only, are indicated. The cated on endothelial cells and extracellular matrix. GAGs immo- values are expressed as the relative OD of binding to CS-A compared with bilize and enhance local concentrations of CCL20, thereby gener- nonspecific binding to BSA. Statistically significant decreases are indicated p Ͻ 0.05). C, Haptotactic assays with CCR6-transfected ,ء) by an asterisk ating a solid-phase haptotactic concentration gradient. To cells. Polycarbonate filters were coated with 50 nM CCL20 and then in- determine whether binding to GAG may protect CCL20 from pro- cubated overnight with 0.1 ␮g/ml Cath-D, buffer I, or migration buffer teolytic processing by Cath-D, we first established that GAG mol- (control). Results are expressed as means Ϯ SD of the numbers of cells ecules like CS, bind CCL20. Our microwell binding assay showed observed per five high-power fields. No migrated cells were detected in the that CCL20 bound to CS-A-coated wells in a dose-dependent man- absence of chemoattractant. Statistically significant decreases in migrated .(p Ͻ 0.05 ,ء) ner, whereas this specific increase in binding was not observed for cells compared with control are indicated by an asterisk The Journal of Immunology 6517 created a haptotactic chemokine gradient, which was able to attract Based on morphology, the CCL20 transcripts did not appear to be CCR6-transfected cells (Fig. 3C). This gradient was destroyed related to the presence of accumulated leukocytes. when the CCL20-coated filters were incubated with 0.1 ␮g/ml To test whether CCL20 transcripts are translated into a func- Cath-D for 16 h. Conversely, incubation with Cath-H did not in- tional protein, we analyzed CCL20 secretion of human melanoma hibit migration of the CCR6 cells (data not shown). We did not test cell lines and assessed its activity. Resting melanoma cells pro- the effect of Cath-B on the haptotactic gradient of CCL20, because duced relatively low amounts of CCL20. However, treating the CCL20 activity was not affected after cleavage with this protease. cells for 48 h with the proinflammatory cytokines IL-1␣ (10 ng/ml) ␣ As expected, a haptotactic chemokine gradient of CCL201–66, the and TNF- (50 ng/ml) increased CCL20 levels 3- to 50-fold in all product of Cath-B processing, was able to attract CCR6-trans- cell lines (Fig. 5A). CCL20 from conditioned medium of stimu- fected cells, whereas CCL201–55, the product of Cath-D process- lated A375 and SK-MEL-2 cells was partially purified using hep- ing, was inactive (data not shown). Taken together, these experi- arin affinity chromatography and reversed-phase HPLC and tested ments indicate that Cath-D processes not only soluble CCL20 but for biological activity. CCL20 secreted by melanoma cells mobi- also CCL20 attached to GAGs, which is the mostly likely condi- lized intracellular Ca2ϩ (Fig. 5B) and was also able to induce cell tion in vivo. migration with similar potencies as synthetic CCL20 (Fig. 5C). Interestingly, melanoma cell lines did not secrete the skin-specific Expression of CCL20 in melanoma cell lines and human chemokine, CCL27, even upon stimulation with IL-1␣, TNF-␣, melanoma tissue and IL-1␤ either alone or in combination (data not shown). CCL20 is up-regulated in neoplastic human skin diseases, such as In a more physiological setting, we stimulated human metastatic mycosis fungoides or under inflamed conditions, such as psoriasis melanoma cells for 48 h with IL-1␣ and TNF-␣, and we observed Downloaded from and contact dermatitis (41, 42). So far, no expression of CCL20 in CCL20 secretion up to 0.5 nM per one million cells. This findings melanoma has been described; therefore, we investigated whether suggest that melanoma cells may produce sufficient CCL20 to at- expression of this chemokine may also occur in human malignant tract immune cells. melanoma cell lines and primary malignant melanoma of the skin. Cathepsins in melanoma cell lines and human melanoma tissue Total RNA was extracted from three melanoma cell lines, A375,

A2058, and SK-MEL-2, from primary keratinocytes, the keratin- To localize Cath-D and Cath-B activities with respect to the cel- http://www.jimmunol.org/ ocyte cell line NCTC 2544, and the epithelial cell line Caco-2, and lular compartment and the extracellular environment, we deter- subjected to RT-PCR analysis. Expression of CCL20 was evident mined proteolytic activities of these enzymes in cellular mem- in all cells (Fig. 4A). Furthermore, CCL20 transcripts were also branes, cytosol, and conditioned medium of the melanoma cell detected in paraffin-embedded sections of human malignant mel- lines A375, A2058, and SK-MEL-2 using d-F-S(benzoyl)-F-F-A- anoma (Fig. 4B). Keratinocytes of the epidermal layer served as A-pAB as substrate for Cath-D and Z-Arg-Arg-AMC as substrate internal positive control for CCL20 expression. CCL20 mRNA for Cath-B. Cath-D activity was detected in all compartments of was detected in a small proportion of the melanoma cells of 6 of SK-MEL-2 cells, and low activities were found in membranes of the 10 patients studied. Not all melanoma cells expressed CCL20, A2058 cells ( p Ͻ 0.05) (Fig. 6A). However no Cath-D activity was possibly due to the existence of functionally different subsets. found in A375 cells. Immunoblots of membranes, cytosol, and by guest on September 28, 2021 conditioned medium with anti-Cath-D Ab were in line with these findings (Fig. 6B). Coomassie blue staining demonstrated that equal amounts of proteins were loaded (Fig. 6B). As expected, Cath-B activity of melanoma cells was mainly associated with membranes. However, activity was not detected at the cell surface but was associated with intracellular membranes, as demonstrated by experiments with intact and Triton X-100-per- meabilized cells (Fig. 6A). Immunoblots with anti-Cath-B Ab re- vealed that A375 membranes contained a Cath-B doublet at 25/27 kDa (nonglycosylated and glycosylated forms of the H chain protein, respectively). Only the upper band of this doublet was detected in A2058 membranes, whereas SK-MEL-2 cells displayed only the lower band. In addition, the SK-MEL-2 cells contained the 31-kDa isoform both in membranes and cytosol (Fig. 6B). Because Cath-D was found to be secreted into the extracellular milieu of SK-MEL-2 cells and processing of CCL20 by this protease modified its activity, we investigated Cath-D expression in human melanoma tissue by immunohistochemical staining. As shown in Fig. 6C, Cath-D immunostaining (red signals) is detected in cytoplasma of melanoma cells and probably also in the extracellular microenvironment of the tumor. The cells of the epidermis FIGURE 4. Expression of CCL20 in melanoma cell lines and human are negative. malignant melanoma. A, RT-PCR analysis of CCL20 and G3PDH expression in melanoma cell lines A375, A2058, and SK-MEL-2. Positive control Cleavage of CCL20 by membranes of IL-1␣- and TNF-␣- RT-PCR products were obtained from primary keratinocytes (Prim. Ker- stimulated melanoma cells atin.), human keratinocyte cell line (NCTC 2544), and Caco-2 cells. A 100-base DNA ladder (ladder) and a negative control (no cDNA; control We then studied cleavage of exogenously added CCL20 by mem- Ϫ) are shown. B, Localization of CCL20 transcripts in human malignant branes and conditioned medium of three melanoma cell lines and melanoma by in situ hybridization. Hybridization signals of antisense compared it to the cleavage pattern where purified cathepsins were CCL20 riboprobes are identified in keratinocytes and melanoma cells. No used. CCL20 was incubated with membranes or conditioned me- signals were detected in serial sections hybridized with the sense probe. dium of IL-1␣- and TNF-␣-stimulated cells, and the cleavage 6518 REGULATION OF CCL20 ACTIVITY BY CATHEPSIN PROCESSING Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 5. Secretion of active CCL20 by melanoma cell lines. A, CCL20 in culture supernatant of melanoma cell lines after stimulation with IL-1␣ and TNF-␣. Confluent A375, A2058, and SK-MEL-2 melanoma cells were stimulated for 24, 48, or 72 h with 10 ng/ml IL-1␣ and 50 ng/ml TNF-␣ in culture supernatant supplemented with 10% FCS. Culture supernatant from unstimulated cells was used as control. Production of CCL20 was determined by ELISA. The results are the means Ϯ SD of three experiments. Absence of error bar means that the error bar is within FIGURE 6. Cathepsins in melanoma cell lines and human malignant the symbol. Statistically significant increases in CCL20 production are in- melanoma. A375, A2058, and SK-MEL-2 melanoma cells were stimulated p Ͻ 0.01). B, CCL20 from culture ,ءء p Ͻ 0.05; and ,ء) dicated by asterisks with IL-1␣ and TNF-␣ for 48 h before isolation of subcellular fractions. A, supernatant of stimulated A375 cells was partially purified by heparin- Upper panel, Membranes, cytosol, and conditioned medium were incu- Sepharose column and reversed-phase chromatography. [Ca2ϩ] -releasing i bated for 4 h with 30 mM d-F-S(benzoyl)-F-F-A-A-pAB to assess Cath-D activity was analyzed with CCR6-transfected cells. C, CCL20 from culture activity. Results are expressed as picomoles of pAB released per minute supernatant of stimulated A375 cells was partially purified by heparin- per 105 cells. Cath-D activity was also measured in the presence of pep- Sepharose column. CCL20-containing fractions were 350-fold concen- statin A (data not shown). A, Lower panel, Intact cells, cells lysed with trated relative to culture supernatant (Sn) and analyzed for cell migration Triton X-100, membranes, and cytosol were incubated for 50 min with 100 using CCR6-transfected cells (CCR6ϩ; Œ). Nontransfected cells (CCR6Ϫ; ␮M Z-Arg-Arg-AMC to assess Cath-B activity. Results are expressed as E) were used as negative control and synthetic CCL20 (80 nM) as positive nanomoles of AMC released per minute per 105 cells. No Cath-B activity control (CCR6ϩ; f). Results of three experiments are expressed as was detected in conditioned medium or in the presence of CA-074 (data not means Ϯ SD of the numbers of cells observed per five high-power fields. shown). Statistically significant increases in activities above controls in the p Ͻ ,ءء p Ͻ 0.05; and ,ء) presence of inhibitor are indicated by asterisks 0.01). B, Membranes, cytosol, and conditioned medium of A375 (1), products were separated by SDS-PAGE and stained with Coomas- A2058 (2), and SK-MEL-2 (3) cells were immunoblotted with anti-Cath-D sie blue. In the presence of Cath-D-specific buffer, the membranes or anti-Cath-B Ab. Purified Cath-D and Cath-B from human liver were from all melanoma cell lines processed CCL20, but cleavage was used as controls (Cathepsin). A Coomassie blue staining of similar samples only partially inhibited by pepstatin A (Fig. 7A), indicating the is given as a control for equal loading of the samples. C, Cath-D was presence of non-pepstatin A-sensitive protease activity. Notice- immunochemically detected in paraffin sections of human melanoma (left ably, pepstatin A had no influence of CCL20 processing by mem- panel); the negative control was obtained by omitting the primary Ab (right branes of A375 cells, further confirming the results presented in panel). Fig. 6A, which demonstrate that these cells do not express Cath-D. The Journal of Immunology 6519 Downloaded from http://www.jimmunol.org/

FIGURE 7. Cleavage of CCL20 by melanoma cell lines. A, Processing of CCL20 by membranes of A375 (1), A2058 (2), and SK-MEL-2 (3) cells. CCL20 (5 ␮M) was incubated for4hat37°C with membrane equivalents of 3 ϫ 105 cells in 50 mM sodium citrate, 50 mM NaCl (pH 4.0), and pepstatin

A as indicated. The cleavage products were separated by SDS-PAGE and stained with Coomassie blue. The molecular masses of intact CCL201–70 and

CCL201–55 after Cath-D cleavage are 8025 and 6224 Da, respectively (arrows). B, Processing of CCL20 by purified Cath-B (left panel) or Cath-D (middle panel) for 1.5 h or membranes of SK-MEL-2 cells (right panel)for4hatpH4.0inpresence of CA-074 and pepstatin A as indicated. The molecular masses of intact CCL201–70, CCL201–66 after Cath-B cleavage and CCL201–55 after Cath-D cleavage are 8025, 7552, and 6224 Da, respectively (arrows). C,As in A but incubation was with PBS containing 4 mM EDTA, 2 mM L-cysteine, 0.1% Triton X-100 (pH 6.8). D, Processing of CCL20 by conditioned medium by guest on September 28, 2021 of SK-MEL-2 cells. Conditioned medium was concentrated (2.5ϫ) and incubated with CCL20 (5 ␮M) at pH 4.0 for the indicated times and in the presence of pepstatin A (left panel). Coomassie blue staining of CCL20 incubated for 8 h with purified Cath-D or conditioned medium is shown in the middle panel. The right panel shows migration in response to increasing concentrations of CCL20 after incubation for8hinconditioned medium with or without pepstatin A. Results are expressed as means Ϯ SD of the numbers of cells observed per five high-power fields. A representative experiment of three is shown. IB, Immunoblotting; Cond. medium, conditioned medium.

We then tested whether the non-pepstatin A-sensitive protease ac- necessary to induce a maximal cell migration response; 50 nM, tivity would depend on Cath-B, which is known to have exopep- however, was necessary for a maximal response when incubation tidase activity at acidic pH (Fig. 7B, left panel). SK-MEL-2 mem- was performed in conditioned medium with active Cath-D. All of branes were incubated with CCL20 in the presence of CA-074 these results strongly point to Cath-D as the CCL20-inactivating and/or pepstatin A. Results of Coomassie blue staining show that protease in conditioned medium. cleavage of CCL20 was totally blocked when both inhibitors were present (Fig. 7B) and establish that Cath-D (and/or Cath-D-like Discussion proteases) and Cath-B are the only proteases involved in the pro- Chemokine processing by proteases is emerging as an important cessing of CCL20 at pH 4.0. Pepstatin A-sensitive proteases de- regulatory mechanism of chemokine functions apart from tran- graded CCL20 totally, whereas the cleavage by Cath-B was lim- scriptional regulation (43, 44) or modulation of chemokine recep- ited. Such limited proteolysis was also observed when membranes tor expression (45). Proteases involved in this process are mainly of A375, A2058, and SK-MEL-2 cells were incubated at neutral dipeptidylpeptidase IV/CD26, a serine-type protease, several pH (Fig. 7C). Because processing was inhibited by the specific MMPs, but also cathepsin G, elastase, urokinase plasminogen ac- Cath-B inhibitor CA-074, these results suggest Cath-B to be the tivator, and plasmin (1). We previously suggested that cathepsins only protease involved. Conditioned medium of stimulated SK- may represent an additional family of proteolytic enzymes affect- MEL-2 cells processed CCL20 in a pepstatin A-dependent manner ing the chemokine regulatory network (7). In this report, we iden- to the same truncation products as purified Cath-D (Fig. 7D), tified new substrates among the chemokine family for Cath-D, and which was verified by mass spectrometry. The migration experi- we characterized for the first time the processing in vitro of a large ment presented in Fig. 7D, right panel, further supports the view spectrum of chemokines by Cath-B. Both cathepsins are generally that conditioned medium of melanoma cells inactivates CCL20. overexpressed in neoplastic tissues and have been previously ex- CCL20 was incubated for 8 h with conditioned medium of SK- tensively studied as possible prognostic factors in a variety of hu- MEL-2 cells in the presence and absence of pepstatin A, and ali- man cancers including breast cancer (46) and malignant melanoma quots were then used for migration experiments. In the presence of (19, 47). In general, receptor binding and chemokine functionality pepstatin A, no CCL20 degradation took place and only 5 nM was are modulated by epitopes in the N-terminal region of chemokines 6520 REGULATION OF CCL20 ACTIVITY BY CATHEPSIN PROCESSING

(1). Accordingly, N-terminal truncation by proteases is an impor- GAG and cationic charges localized on the C terminus of the che- tant step that may lead to decrease or increase of the chemotactic mokines, such as are present in CCL20. Our data demonstrating potency of chemokines (1). The present data and our previous that Cath-D retain their full proteolytic activity on GAG-bound studies (7), however, indicate that the proteolytic activity of CCL20 strongly suggest that this degradation processes may in- Cath-D and Cath-B is preferentially restricted to the C terminus of deed occur in vivo. chemokines. So far, processing of the C-terminal amino acids has CCL20 contributes to the adaptive immune response by recruit- been described for CXCL7 (48, 49), CCL21 (7), and CXCL10 ing CCR6 expressing immature dendritic cells to the site of in- (50). Notably, for some chemokines the degradation products re- flammation. The finding that epithelial cells and in particular ker- tain inverse agonistic activity or direct antibacterial activity (50). atinocytes are good producers of CCL20 as well as the recently The CCL20 degradation products formed in vitro by Cath-D and described bactericidal properties of CCL20 (5, 40) strongly sug- Cath-B include several stable oligopeptides with 55 and 12 aa, and gest that this chemokine may have important defensive functions 66 aa, respectively. However, because no specific Abs to recognize at barriers between the inner and outer milieu, such as the skin and these peptides are available, we could not identify CCL20 cleavage the intestinal mucosa. Interestingly, we demonstrate that, although products in tissue. Furthermore, it is conceivable that the peptides CCL20 lost its chemotactic potencies after Cath-D cleavage, the are further processed by a combination of the proteases in vivo. 12-aa peptide generated possessed antimicrobial activities. We

Our data demonstrate that in malignant human melanoma Cath-D found that the antibacterial activity of CCL2059–70 is similar to the (and/or Cath-D-like proteases) and Cath-B are responsible for the antibacterial activity of the defensin hBD1 as obtained by Hoover processing of CCL20 in acidic conditions and that Cath-B alone is et al. (40) using E. coli as a read-out system. responsible for the cleavage of CCL20 in neutral conditions. We Here, we demonstrate also that malignant melanocytes produce Downloaded from found that CCL20 is processed and functionally inactivated by significant amounts of CCL20 and that CCL20 release may be Cath-D by cleaving off the last 12 aa, whereas removal of the last stimulated in these cells by proinflammatory cytokines. These 4 aa at the C terminus by Cath-B does not affect its chemotactic findings are indicative for possible other CCL20 functions. In this potency. This latter finding is in agreement with previous reports context, we also demonstrated that certain malignant melanocytes showing that the C-terminally truncated isoform of CCL20 miss- retain Cath-D activity in their membranes and that they produce

ing three residues is equipotent to full-length CCL20 in calcium and release Cath-D, whereas Cath-B activity is predominantly as- http://www.jimmunol.org/ mobilization assays (51), although in that study, no protease re- sociated with the membrane and cytosolic component of cells. sponsible for cleaving off C-terminal amino acids was identified. Both cathepsins were found to process CCL20 at its C terminus. It Our findings suggest that Cath-B may also be directly implicated in the posttranslational modification of CCL20 occurring in some cell populations, such as PBMC (51). Cath-D does not recognize a specific amino acid sequence, but, as observed with other proteases, it cleaves chemokines at different sites, indicating therefore that enzyme activity may be at least in part modulated by the tertiary structure of its substrates. Efficient by guest on September 28, 2021 cleavage of CCL20 by Cath-D was observed at pH 4.0. A pH Ͻ6.0 may be rarely found in the extracellular environment of healthy tissues, but it is well known that the microenvironment of a tumor is acidic (52). Multiple pathways contribute to an acidic microenvironment in tumors, including the finding that breast cancer cells and macrophages within cancerous tissue have a high potential to liberate protons into the extracellular milieu and that anoxia, a characteristic of many cancers, may also promote an acidic environment in tumors (7). Furthermore, Cath-B activity is pH dependent; this protease is an endopeptidase at neutral pH but changes to an exopeptidase below pH 5.5 (53). Hoover et al. (40) investigated the x-ray structure of the human CCL20 monomer. They found that amino acids W54 to L63 form the C-terminal ␣ helix. The C-terminal helix interacts with the ␤-sheet through the side chains of 10 residues including W55, V56, I59, V60, and L63, creating a hydrophobic core spanning the secondary structure elements. Lack of interpretable electron den- sity indicates that the five C-terminal residues, K66–M70, are dis- ordered. Our results concerning the loss of CCL20 activity by processing between residues W55–V56, and the unchanged chemotactic potency of CCL20 after cleavage between residues K66–V67 are therefore in agreement with these structural data. Secreted chemokines generate a chemotactic gradient in vivo FIGURE 8. Model of CCL20 regulation by cathepsin proteolysis. Mel- through immobilization to GAGs of the extracellular matrix. anoma cells produce functional CCL20 forming a haptotactic gradient by binding to GAGs of the extracellular matrix proteins. Dendritic cells mi- GAGs binding may potentiate the chemotactic activity of chemo- grate into tumors and, after taking up Ags, recruit tumor-associated mac- kines by increasing local concentrations, by promoting oligomer- rophages. Cath-D release by the neoplastic cells abrogate the biological ization processes, by protecting chemokines from proteolytic at- activity of CCL20, possibly allowing the melanoma cells to escape anti- tack (54, 55), and by facilitating their presentation to high-affinity tumoral immune responses. A 12-aa peptide resulting from this processing receptors (56–58). The binding between GAGs and chemokines is maintains some antimicrobial activity. ECM, Extracellular matrix; iDC, due to electrostatic interactions between sulfated groups of the immature dendritic cells. The Journal of Immunology 6521 seems reasonable to question the biological relevance of concur- 21. Charbonnier, A. S., N. Kohrgruber, E. Kriehuber, G. Stingl, A. Rot, and ␣ rent expression of cathepsins and their substrate by melanoma D. Maurer. 1999. Macrophage inflammatory protein 3 is involved in the con- stitutive trafficking of epidermal Langerhans cells. J. Exp. Med. 190: 1755–1768. cells. We can exclude a role of CCL20 as factor for tumor prolif- 22. Tanaka, Y., T. Imai, M. Baba, I. Ishikawa, M. Uehira, H. Nomiyama, and eration, because no CCR6 was found on melanoma cells (data not O. Yoshie. 1999. Selective expression of liver and activation-regulated chemokine (LARC) in intestinal epithelium in mice and humans. Eur. J. Immunol. 29: shown). The model in Fig. 8 proposes an autocrine regulatory loop 633–642. where melanoma cells process CCL20 by Cath-D-dependent path- 23. Iwasaki, A., and B. L. Kelsall. 2000. Localization of distinct Peyer’s patch den- way. Because Cath-D in the extracellular milieu may hamper the dritic cell subsets and their recruitment by chemokines macrophage inflammatory protein (MIP)-3␣, MIP-3␤, and secondary lymphoid organ chemokine. J. Exp. recruitment of immature dendritic cells through CCR6, antitumoral Med. 191: 1381–1394. immune responses in melanoma need to be generated in a 24. Cook, D. N., D. M. Prosser, R. Forster, J. Zhang, N. A. Kuklin, S. J. Abbondanzo, CCL20-independent way. X. D. Niu, S. C. Chen, D. J. Manfra, M. T. Wiekowski, et al. 2000. CCR6 mediates dendritic cell localization, lymphocyte homeostasis, and immune responses in mucosal tissue. Immunity 12: 495–503. Acknowledgments 25. Power, C. A., D. J. Church, A. Meyer, S. Alouani, A. E. Proudfoot, We thank Dr. B. Moser for the CCL20 cDNA, Dr. A. Kappeler for im- I. Clark-Lewis, S. Sozzani, A. Mantovani, and T. N. Wells. 1997. Cloning and ␣ munohistochemistry technique, and Dr. M. Heller for help with peptide characterization of a specific receptor for the novel CC chemokine MIP-3 from lung dendritic cells. J. Exp. Med. 186: 825–835. analysis. We also thank Dr. T. Brunner for reviewing the manuscript and 26. Liao, F., R. Alderson, J. Su, S. J. Ullrich, B. L. Kreider, and J. M. Farber. 1997. Dr. C. Mueller and Dr. M. Thelen for valuable comments. STRL22 is a receptor for the CC chemokine MIP-3␣. Biochem. Biophys. Res. Commun. 236: 212–217. Disclosures 27. Al-Aoukaty, A., B. Rolstad, A. Giaid, and A. A. Maghazachi. 1998. MIP-3␣, MIP-3␤ and fractalkine induce the locomotion and the mobilization of intracel- The authors have no financial conflict of interest.

lular calcium, and activate the heterotrimeric G proteins in human natural killer Downloaded from cells. Immunology 95: 618–624. References 28. Campbell, J. J., J. Hedrick, A. Zlotnik, M. A. Siani, D. A. Thompson, and 1. Moser, B., M. Wolf, A. Walz, and P. Loetscher. 2004. Chemokines: multiple E. C. Butcher. 1998. Chemokines and the arrest of lymphocytes rolling under levels of leukocyte migration control. Trends Immunol. 25: 75–84. flow conditions. Science 279: 381–384. 2. Thelen, M. 2001. Dancing to the tune of chemokines. Nat. Immunol. 2: 129–134. 29. Dieu, M. C., B. Vanbervliet, A. Vicari, J. M. Bridon, E. Oldham, S. Ait-Yahia, 3. Rossi, D., and A. Zlotnik. 2000. The biology of chemokines and their receptors. F. Briere, A. Zlotnik, S. Lebecque, and C. Caux. 1998. Selective recruitment of Annu. Rev. Immunol. 18: 217–242. immature and mature dendritic cells by distinct chemokines expressed in different 4. Balkwill, F. 2004. Cancer and the chemokine network. Nat. Rev. Cancer 4: anatomic sites. J. Exp. Med. 188: 373–386. http://www.jimmunol.org/ 540–550. 30. Liao, F., R. L. Rabin, C. S. Smith, G. Sharma, T. B. Nutman, and J. M. Farber. 5. Yang, D., Q. Chen, D. M. Hoover, P. Staley, K. D. Tucker, J. Lubkowski, and 1999. CC-chemokine receptor 6 is expressed on diverse memory subsets of T J. J. Oppenheim. 2003. Many chemokines including CCL20/MIP-3␣ display an- cells and determines responsiveness to macrophage inflammatory protein 3␣. timicrobial activity. J. Leukocyte Biol. 74: 448–455. J. Immunol. 162: 186–194. 6. Delgado, M. B., I. Clark-Lewis, P. Loetscher, H. Langen, M. Thelen, 31. Liao, F., A. K. Shirakawa, J. F. Foley, R. L. Rabin, and J. M. Farber. 2002. M. Baggiolini, and M. Wolf. 2001. Rapid inactivation of stromal cell-derived Human B cells become highly responsive to macrophage-inflammatory protein- factor-1 by cathepsin G associated with lymphocytes. Eur. J. Immunol. 31: 3␣/CC chemokine ligand-20 after cellular activation without changes in CCR6 699–707. expression or ligand binding. J. Immunol. 168: 4871–4880. 7. Wolf, M., I. Clark-Lewis, C. Buri, H. Langen, M. Lis, and L. Mazzucchelli. 2003. 32. Rossi, D. L., A. P. Vicari, K. Franz-Bacon, T. K. McClanahan, and A. Zlotnik. Cathepsin D specifically cleaves the chemokines macrophage inflammatory pro- 1997. Identification through bioinformatics of two new macrophage proinflam- tein-1␣, macrophage inflammatory protein-1␤, and SLC that are expressed in matory human chemokines: MIP-3␣ and MIP-3␤. J. Immunol. 158: 1033–1036.

human breast cancer. Am. J. Pathol. 162: 1183–1190. 33. Kimura, T., H. Takeshima, N. Nomiyama, T. Nishi, T. Kino, M. Kochi, by guest on September 28, 2021 8. Nufer, O., M. Corbett, and A. Walz. 1999. Amino-terminal processing of che- J. I. Kuratsu, and Y. Ushio. 2002. Expression of lymphocyte-specific chemokines mokine ENA-78 regulates biological activity. Biochemistry 38: 636–642. in human malignant glioma: essential role of LARC in cellular immunity of 9. Padrines, M., M. Wolf, A. Walz, and M. Baggiolini. 1994. Interleukin-8 pro- malignant glioma. Int. J. Oncol. 21: 707–715. cessing by neutrophil elastase, cathepsin G and proteinase-3. FEBS Lett. 352: 34. Clark-Lewis, I., L. Vo, P. Owen, and J. Anderson. 1997. Chemical synthesis, 231–235. purification, and folding of C-X-C and C-C chemokines. Methods Enzymol. 287: 10. Ohashi, K., M. Naruto, T. Nakaki, and E. Sano. 2003. Identification of interleu- 233–250. kin-8 converting enzyme as cathepsin L. Biochim. Biophys. Acta 1649: 30–39. 35. Loetscher, M., B. Gerber, P. Loetscher, S. A. Jones, L. Piali, I. Clark-Lewis, 11. Car, B. D., M. Baggiolini, and A. Walz. 1991. Formation of neutrophil-activating M. Baggiolini, and B. Moser. 1996. Chemokine receptor specific for IP10 and peptide 2 from platelet-derived connective-tissue-activating peptide III by differ- mig: structure, function, and expression in activated T lymphocytes. J. Exp. Med. ent tissue proteinases. Biochem. J. 275: 581–584. 184: 963–969. 12. Balbin, M., A. Fueyo, A. M. Tester, A. M. Pendas, A. S. Pitiot, A. Astudillo, 36. von Tscharner, V., B. Prod’hom, M. Baggiolini, and H. Reuter. 1986. Ion chan- C. M. Overall, S. D. Shapiro, and C. Lopez-Otin. 2003. Loss of collagenase-2 nels in human neutrophils activated by a rise in free cytosolic calcium concen- confers increased skin tumor susceptibility to male mice. Nat. Genet. 35: tration. Nature 324: 369–372. 252–257. 37. Heylen, N., L. M. Vincent, V. Devos, V. Dubois, C. Remacle, and A. Trouet. 13. McQuibban, G. A., G. S. Butler, J. H. Gong, L. Bendall, C. Power, 2002. Fibroblasts capture cathepsin D secreted by breast cancer cells: possible I. Clark-Lewis, and C. M. Overall. 2001. Matrix metalloproteinase activity in- role in the regulation of the invasive process. Int. J. Oncol. 20: 761–767. activates the CXC chemokine stromal cell-derived factor-1. J. Biol. Chem. 276: 38. Guo, M., P. A. Mathieu, B. Linebaugh, B. F. Sloane, and J. J. Reiners, Jr. 2002. 43503–43508. Phorbol ester activation of a proteolytic cascade capable of activating latent trans- 14. Zhang, K., G. A. McQuibban, C. Silva, G. S. Butler, J. B. Johnston, J. Holden, forming growth factor-␤L a process initiated by the exocytosis of cathepsin B. I. Clark-Lewis, C. M. Overall, and C. Power. 2003. HIV-induced metallopro- J. Biol. Chem. 277: 14829–14837. teinase processing of the chemokine stromal cell derived factor-1 causes neuro- degeneration. Nat. Neurosci. 6: 1064–1071. 39. Schaerli, P., K. Willimann, A. B. Lang, M. Lipp, P. Loetscher, and B. Moser. 15. McQuibban, G. A., J. H. Gong, E. M. Tam, C. A. McCulloch, I. Clark-Lewis, and 2000. CXC chemokine receptor 5 expression defines follicular homing T cells C. M. Overall. 2000. Inflammation dampened by gelatinase A cleavage of mono- with B cell helper function. J. Exp. Med. 192: 1553–1562. cyte chemoattractant protein-3. Science 289: 1202–1206. 40. Hoover, D. M., C. Boulegue, D. Yang, J. J. Oppenheim, K. Tucker, W. Lu, and 16. Muller, A., B. Homey, H. Soto, N. Ge, D. Catron, M. E. Buchanan, J. Lubkowski. 2002. The structure of human macrophage inflammatory protein- ␣ T. McClanahan, E. Murphy, W. Yuan, S. N. Wagner, et al. 2001. Involvement of 3 /CCL20: linking antimicrobial and CC chemokine receptor-6-binding activi- ␤ chemokine receptors in breast cancer metastasis. Nature 410: 50–56. ties with human -defensins. J. Biol. Chem. 277: 37647–37654. 17. Westley, B. R., and F. E. May. 1996. Cathepsin D and breast cancer. Eur. J. 41. Schmuth, M., S. Neyer, C. Rainer, A. Grassegger, P. Fritsch, N. Romani, and Cancer 32A: 15–24. C. Heufler. 2002. Expression of the C-C chemokine MIP-3␣/CCL20 in human 18. Thorpe, S. M., H. Rochefort, M. Garcia, G. Freiss, I. J. Christensen, S. Khalaf, epidermis with impaired permeability barrier function. Exp. Dermatol. 11: F. Paolucci, B. Pau, B. B. Rasmussen, and C. Rose. 1989. Association between 135–142. high concentrations of Mr 52,000 cathepsin D and poor prognosis in primary 42. Homey, B., H. Alenius, A. Muller, H. Soto, E. P. Bowman, W. Yuan, L. McEvoy, human breast cancer. Cancer Res. 49: 6008–6014. A. I. Lauerma, T. Assmann, E. Bunemann, et al. 2002. CCL27-CCR10 interac- 19. Otto, F. J., T. Goldmann, B. Biess, A. Lippold, L. Suter, and U. Westhoff. 1999. tions regulate T cell-mediated skin inflammation. Nat. Med. 8: 157–165. Prognostic classification of malignant melanomas by combining clinical, histo- 43. Kuschert, G. S., F. Coulin, C. A. Power, A. E. Proudfoot, R. E. Hubbard, logical, and immunohistochemical parameters. Oncology 56: 208–214. A. J. Hoogewerf, and T. N. Wells. 1999. Glycosaminoglycans interact selectively 20. Hieshima, K., T. Imai, G. Opdenakker, J. Van Damme, J. Kusuda, H. Tei, with chemokines and modulate receptor binding and cellular responses. Biochem- Y. Sakaki, K. Takatsuki, R. Miura, O. Yoshie, and H. Nomiyama. 1997. Molec- istry 38: 12959–12968. ular cloning of a novel human CC chemokine liver and activation-regulated che- 44. Culley, F. J., E. J. Fadlon, A. Kirchem, T. J. Williams, P. J. Jose, and J. E. Pease. mokine (LARC) expressed in liver: chemotactic activity for lymphocytes and 2003. Proteoglycans are potent modulators of the biological responses of eosin- gene localization on chromosome 2. J. Biol. Chem. 272: 5846–5853. ophils to chemokines. Eur. J. Immunol. 33: 1302–1310. 6522 REGULATION OF CCL20 ACTIVITY BY CATHEPSIN PROCESSING

45. Bruhl, H., C. D. Cohen, S. Linder, M. Kretzler, D. Schlondorff, and M. Mack. sity of human liver and activation-regulated chemokine/macrophage inflamma- 2003. Post-translational and cell type-specific regulation of CXCR4 expression tory protein-3␣ from normal and transformed cells. J. Immunol. 165: 4470–4477. by cytokines. Eur. J. Immunol. 33: 3028–3037. 52. Tannock, I. F., and D. Rotin. 1989. Acid pH in tumors and its potential for 46. Lah, T. T., M. Cercek, A. Blejec, J. Kos, E. Gorodetsky, R. Somers, and therapeutic exploitation. Cancer Res. 49: 4373–4384. I. Daskal. 2000. Cathepsin B, a prognostic indicator in lymph node-negative 53. Polgar, L., and C. Csoma. 1987. Dissociation of ionizing groups in the binding breast carcinoma patients: comparison with cathepsin D, cathepsin L, and other cleft inversely controls the endo- and exopeptidase activities of cathepsin B. clinical indicators. Clin. Cancer Res. 6: 578–584. J. Biol. Chem. 262: 14448–14453. 47. Kos, J., B. Werle, T. Lah, and N. Brunner. 2000. Cysteine proteinases and their 54. Sadir, R., A. Imberty, F. Baleux, and H. Lortat-Jacob. 2004. Heparan sulfate/ inhibitors in extracellular fluids: markers for diagnosis and prognosis in cancer. heparin oligosaccharides protect stromal cell-derived factor-1 (SDF-1)/CXCL12 Int. J. Biol. Markers 15: 84–89. against proteolysis induced by CD26/dipeptidyl peptidase IV. J. Biol. Chem. 279: 48. Ehlert, J. E., F. Petersen, M. H. Kubbutat, J. Gerdes, H. D. Flad, and E. Brandt. 43854–43860. 1995. Limited and defined truncation at the C terminus enhances receptor binding 55. Webb, L. M., M. U. Ehrengruber, I. Clark-Lewis, M. Baggiolini, and A. Rot. and degranulation activity of the neutrophil-activating peptide 2 (NAP-2): com- 1993. Binding to heparan sulfate or heparin enhances neutrophil responses to parison of native and recombinant NAP-2 variants. J. Biol. Chem. 270: interleukin 8. Proc. Natl. Acad. Sci. USA 90: 7158–7162. 6338–6344. 56. Hoogewerf, A. J., G. S. Kuschert, A. E. Proudfoot, F. Borlat, I. Clark-Lewis, 49. Ehlert, J. E., J. Gerdes, H. D. Flad, and E. Brandt. 1998. Novel C-terminally C. A. Power, and T. N. Wells. 1997. Glycosaminoglycans mediate cell surface truncated isoforms of the CXC chemokine ␤-thromboglobulin and their impact oligomerization of chemokines. Biochemistry 36: 13570–13578. on neutrophil functions. J. Immunol. 161: 4975–4982. 57. Tanaka, Y., D. H. Adams, S. Hubscher, H. Hirano, U. Siebenlist, and S. Shaw. 50. Hensbergen, P. J., D. Verzijl, C. I. Balog, R. Dijkman, R. C. van der Schors, 1993. T-cell adhesion induced by proteoglycan-immobilized cytokine MIP-1␤. E. M. van der Raaij-Helmer, M. J. van der Plas, R. Leurs, A. M. Deelder, Nature 361: 79–82. M. J. Smit, and C. P. Tensen. 2004. Furin is a chemokine-modifying enzyme: in 58. Tanaka, Y., K. Kimata, A. Wake, S. Mine, I. Morimoto, N. Yamakawa, vitro and in vivo processing of CXCL10 generates a C-terminally truncated che- H. Habuchi, S. Ashikari, H. Yamamoto, K. Sakurai, et al. 1996. Heparan sulfate mokine retaining full activity. J. Biol. Chem. 279: 13402–13411. proteoglycan on leukemic cells is primarily involved in integrin triggering and its 51. Schutyser, E., S. Struyf, P. Menten, J. P. Lenaerts, R. Conings, W. Put, A. Wuyts, mediated adhesion to endothelial cells. J. Exp. Med. 184: 1987–1997. Vol. 176 P. Proost, and J. Van Damme. 2000. Regulated production and molecular diver- No. 11 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021