Cathepsins B, L and D in inflammatory bowel disease macrophages and potential therapeutic effects of cathepsin inhibition in vivo

Abstract

The cathepsins D (CTSD), B (CTSB) and L (CTSL) are important for the intracellular degradation of proteins. Increased cathepsin expression is associated with inflammatory diseases. We have shown previously an induction of CTSD expression in intestinal macrophages (IMAC) in inflamed mucosa of patients with inflammatory bowel disease (IBD). Here we investigated the regulation of CTSB and CTSL in IMAC during IBD and effects of CTSD and CTSB/CTSL inhibition in vivo. Human IMAC were isolated from normal and inflamed mucosa. Reverse transcription–polymerase chain reaction (RT–PCR) was performed for CTSB and CTSL mRNA. Immunostaining was used to confirm PCR results. Cathepsin inhibition was investigated in the dextran–sulphate–sodium (DSS) colitis model in mice with application of pepstatin A (CTSD inhibitor), CA-074 (CTSB inhibitor) and Z-Phe-Tyr-aldehyde (CTSL inhibitor). CTSL mRNA was significantly up-regulated in IMAC isolated from IBD mucosa. Up-regulated protein expression was found mainly in areas of mucosal damage by immunostaining. Inhibition of CTSD in mouse DSS colitis was followed by an amelioration of the disease. Inhibitor-treated mice showed a significant lower histological score (HS) and less colon reduction in comparison to controls. Similarly, simultaneous inhibition of CTSB/CTSL was followed by a significant amelioration of colitis. Expression of tissue-degrading cathepsins is increased in IMAC in IBD. Inhibition of CTSD as well as CTSB/CTSL is followed by an amelioration of experimental colitis. The prevention of mucosal damage by cathepsin inhibition could represent a new approach for the therapy of IBD.

Introduction

In recent years the understanding of the pathogenesis of Crohn's disease (CD) and ulcerative colitis (UC) has made enormous progress. The importance of the mucosal barrier for preventing bacterial invasion and subsequent inflammation has been shown in genetic studies identifying risk genes and animals models. The relevance of mucosal damage and mucosal healing has been discussed extensively.

Under normal conditions, the gastrointestinal immune system keeps a delicate balance between the immune response to microbial pathogens and tolerance to the local commensal flora that is mediated partially by down-regulation of surface receptors involved in the recognition or transmission of inflammatory signals [1–3]. In patients with inflammatory bowel disease (IBD) the mucosal barrier may become leaky, leading to uncontrolled uptake of antigens and proinflammatory molecules, including luminal bacteria and bacterial products from the gut lumen. Intestinal macrophages (IMAC) in normal non-inflamed (NI) mucosa show a down-regulation of molecules involved in rapid responses to bacteria and the innate immune system. Among those down-regulated surface molecules are CD14, Toll-like receptors (TLRs) 2 and 4, human leukocyte antigen-D-related (HLA-DR), CD11b/CD18 and CD11c/CD18, which are discussed as alternative binding proteins for Gram-negative bacteria and the T cell co-stimulatorymolecules CD80 (B7-1) and CD86 (B7-2) [2–6]. This indicates a low ability of IMAC to mediate immune responses. Furthermore, reduced oxidative burst activity indicates a ‘desensitizing’ during their differentiation in the intestinal mucosa [7–10]. During mucosal inflammation, however, IMAC express T cell co-stimulatory molecules, indicating that IMAC in inflamed mucosa have regained the ability to mediate immune responses [4,6]. Furthermore, activation of nuclear factor (NF)κB during human mucosal inflammation indicates an involvement in the inflammatory process [11].

When we studied differences in gene expression between normal IMAC and inflammation-associated macrophages by arrays and subtractive hybridization screenings we found a dramatic induction of tissue degrading enzymes such as nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and cathepsin D (CTSD) [8,12]. This led us to investigate other members of the cathepsin family.

The proteolytic and destructive properties of the lysosomal cathepsins have been shown to play a role in degenerative, as well as chronic inflammatory diseases. CTSL was found to be secreted by tumour cell lines and is thought to contribute to degradation of basement membrane components during tissue invasion and metastasis [13]. Its collagenase- and elastase-degrading activity is involved in the pathogenesis of diseases such as artheriosclerosis and lung emphysema [14]. The lysosomal cysteine proteinases cathepsin B (CTSB) and cathepsin L (CTSL) are likely to be involved in osteoporosis, cancer metastasis, rheumatoid arthritis and infectious diseases [15–17]. The usefulness of these cathepsins as a prognostic factor during the disease course, however, is controversial [18]. A correlation has been found between increased CTSB expression and tumour stage [19,20].

Furthermore, proteases of the cathepsin family are involved in the remodelling of extracellular matrix (ECM) proteins [21]. Destruction of elastin-rich tissues during inflammatory responses is associated with local accumulation of macrophages that contain high levels of elastinolytic enzymes, such as CTSB and CTSL [22]. The aspartic proteinase CTSD has the potential to initiate a proteolytic cascade and to degrade and remodel extracellular matrix [23]. CTSD-deficient mice are born normally but die at postnatal day 26 because of massive intestinal necrosis, thromboembolism and lymphopenia [24]. As we had demonstrated an induction of CTSD mRNA and protein expression in IMAC from IBD mucosa, we hypothesized a participation of cathepsins in intestinal inflammation and mucosal tissue damage.

Therefore, we first investigated the expression of CTSB and CTSL in IMAC from IBD patients and found an up-regulation especially in areas of tissue damage and mucosal ulceration. Subsequently we tested the influence of cathepsin inhibition on the extent of inflammation in an experimental colitis model.

Materials and methods

Patients

Surgical specimens were taken from healthy areas of the colonic mucosa of patients undergoing surgery for colorectal carcinoma (control patients) or from the inflamed colonic mucosa of patients with CD or UC. Histological evaluation was performed by an experienced IBD pathologist. IBD patients were treated with 5-aminosalicylic acid (5-ASA) and/or steroids. The therapies had no influence on the results. The study was approved by the University of Regensburg Ethics Committee.

Thirteen samples from colon without detectable inflammation were included in the study as NI. Twelve CD and four UC samples with a moderate to severe inflammation were used. Specimens from UC patients were taken from descending colon or sigmoid colon, specimens from CD patients were taken from ascending colon or sigmoid colon.

Isolation of human lamina propria mononuclear cells (LPMNCs)

Intestinal epithelial cells (IEC) were removed from mucosal specimens by incubation in calcium- and magnesium-free Hanks's balanced salt solution (HBSS) with 1 mmol/l ethylenediamine tetraacetic acid (EDTA) for 30 min at 37°C by gentle shaking. Subsequently, human lamina propria mononuclear cells (LPMNCs) were isolated according to a protocol described recently [25]. Specimens were incubated for 45 min in 10 ml phosphate buffered saline (PBS) with 1 mg/ml collagenase type I (= 336 U/ml), 0·3 mg/ml deoxyribonuclease (DNaseI; Boehringer, Mannheim, Germany) and 0·2 mg/ml hyaluronidase at 37°C. After dispersing and washing with 1·5 ml PBS with 500 µl fetal calf serum (FCS) cells were finally submitted to Ficoll density gradient centrifugation for 20 min at ∼690 g, without brake) for the isolation of mononuclear cells. The interphase was removed carefully and cells were washed with PBS.

Isolation and purification of human IMAC

LPMNCs isolated from normal and IBD mucosa specimens were incubated with MicroBeads armed with a monoclonal mouse anti-human macrophage CD33 antibody (Miltenyi Biotec, Bergisch Gladbach, Germany) and purified twice with AS separation columns (Miltenyi Biotec), as described recently [25]. The magnetically labelled cells were retained in the column. After removal of the column from the magnetic field, the retained fraction was eluted. Eluted cells were passed through a second AS separation column to increase the purity of IMAC [25]. The final purity of > 95% IMAC was confirmed by fluorescence activated cell sorter (FACS) analysis using phycoerythrin (PE)-conjugated goat anti-mouse IgG antibody (Caltag; Medac, Hamburg, Germany).

Reverse transcription–polymerase chain reaction (RT–PCR) for CTSB and CTSL

Poly(A)-RNA was isolated by polyT magnetic beads (Dynal, Oslo, Norway) from CD33⁺ cells according to the manufacturer's protocol. IMAC were resuspended in lysis/binding buffer. The vial was placed in the magnetic field of a permanent magnet (MCP®-E-1; Dynal) for 5 min to remove immunomagnetic MicroBeads armed with CD33 antibody. The suspension was mixed with Dynabeads oligo(dT)₂₅ and rotated for 5 min. The lysate was washed three times. Elution of mRNA from Dynabeads oligo(dT)₂₅ was performed at 65°C for 2 min in 20 µl elution solution.

Poly(A)-RNA was reverse transcribed with the RT–PCR system (Promega, Madison, USA) in a 15-min reaction at 42°C. For PCR, the primers used were as follows: CTSL upstream GACAGGGACTGGAAGAGAG; CTSL downstream GTTTCCCTTCCCTGTATTC; CTSB upstream TCGGATGAGCTGGTCAACTATG; CTSB downstream TCCAAGCTTCAGCAGGATAG.

To test cDNAs for representation and full-length genes RT–PCR with a 5′β-actin, 3′β-actin, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 6K clathrin and 2K clathrin primer set from the Gene Checker™ Kit (Invitrogen, Leek, the Netherlands) was performed. The PCR products were subjected to 1% agarose gel electrophoresis together with a size standard (GeneRuler™; MBI Fermentas, St. Leon-Rot, Germany).

TaqMan®–PCR

Poly(A)-RNA was isolated from CD33⁺ cells and reverse transcribed as described above. Expression levels of CTSB and CTSL were quantified by TaqMan®–PCR as single-tube reactions (20 µl) in 384-well plates. Primers used were as follows: CTSL forward TAGAGGCACAGTGGACCAAGTG; CTSL reverse ACTGCTCTCCTCCATCCTTCTTC; CTSB forward TCAACTATGTCAACAAACGGAATACC; CTSB reverse CAAGTAGCTCATGTCCA CGTTGTA. Probes used (CTSL: 5′-AGGCGATGCACAACAGATTATACGGCA-3′, CTSB: 5′-TGGCAGGCCGGGCACAACTT-3′) were labelled 5′ with 6-carboxy-fluorescein (FAM) and 3′ with 6-carboxy-tetramethylrhodamine (TAMRA). TaqMan®–PCR was performed using 1 µl cDNA, 1 µl upstream and downstream primer each (18 µM), 1 µl probe (5 µM), 10 µl TaqMan Mastermix (Applied Biosystems, Foster City, USA) and made up to the final volume of 20 µl with sterile H₂O. Simultaneous measurement of GAPDH expression using an appropriate GAPDH Kit (Applied Biosystems) served as reference. Cycling was as follows: 50°C for 2 min, 95°C for 10 min followed by 44 repeats: 95°C for 15 s and 60°C for 1 min.

Demasking of paraffin-embedded sections

Paraffin embedded sections were cut (5 µm), floated on demineralized water, placed on slides and baked for 60 min at 60°C. Slides were dewaxed for 10 min with xylene and rehydrated in a graded ethanol series (99%, 95%, 70% ethanol and PBS for 5 min each). For demasking, sections were incubated for 30 min with target retrieval solution (S2031, Dako, Hamburg, Germany) at 95°C. Endogenous peroxidase was quenched for 30 min with 1% hydrogen peroxide in PBS. Slides were washed three times in PBS.

Immunohistochemistry

For the identification of CTSL positive cells in human tissue, specimens were incubated with primary mouse-anti-human CTSL antibody (7 µg/ml; IgG₁, monoclonal antibody (MCA) 2066, Serotec, Duesseldorf, Germany) recognizing human precursor and mature forms of CTSL. Mouse IgG₁ (7 µg/ml; M-5284, mouse IgG₁κ; Sigma, Taufkirchen, Germany) was used as isotype-control. Biotin-conjugated goat-anti-mouse secondary antibody (1/500 dilution; (IgG (H + l), 115-065-062; Jackson ImmunoResearch, Hamburg, Germany) and consecutively the Vectastain ABC-elite-standard-system (#PK-6100; Vector Laboratories, Burlingame, USA) were applied. After washing the tissue was incubated with NovaRED® (AEC, Vector Laboratories) for a red immunostaining.

Double labelling immunohistochemistry

To show that mucosal macrophages express CTSL, slides were incubated for 30 min with 0·3% hydrogen peroxide in PBS to suppress remaining peroxidase. After washing the specimens were incubated with mouse-anti-human CD68/fluorescein isothiocyanate (FITC) (7 µg/ml; F7135, clone KP1, IgG₁, Dako). Purified mouse isotype (7 µg/ml) served as control. Subsequently, sheep-anti-mouse anti-fluorescein-horseradish peroxidase (HRP)-conjugated secondary antibody (1/25 diluted; NEF 710, NEN®; Boston, MA, USA) and the Vectastain ABC-elite-standard-system were applied (isotype: M-5284, mouse IgG₁κ; Sigma). Slides were preincubated for 8 min in 0·01% benzidine dihydrochloride (BDHC; Sigma) with 0·03% sodium nitroprusside (Sigma) followed by incubation in 0·01% BDHC/0·005% hydrogen peroxide/0·03% sodium nitroprusside for a blue immunostaining. The formation of the blue reaction product was monitored under the microscope and interrupted with water. For semiquantitative analysis red, blue and double-stained cells were determined separately by evaluating four high-power fields at magnifications of 200× or 400×.

Immunofluorescence

Immunofluorescence was used for detecting CTSB-positive cells in human tissue. Blocking of endogenous peroxidases was performed as described above. To reduce non-specific binding slides were incubated in 1% bovine serum albumin (BSA)/PBS for 30 min. After washing in PBS a polyclonal rabbit-anti-human antibody (5 µg/ml; IgG, AHP 591; Serotec) against the human precursor and mature forms of CTSB was applied. For negative control the primary antibody was replaced by a rabbit immunglobulin fraction (X0903; Dako) at identical dilutions. The slides were then incubated with goat-anti-rabbit-Alexa488 secondary antibody (1/300 diluted; IgG (H + l), A110-08 (Molecular Probes, Eugene, OR, USA). To show that CTSB-expressing cells are macrophages, the sections were again rinsed in 0·3% hydrogen peroxide in PBS for 30 min to suppress resting peroxidases. After washing the specimens were incubated with mouse-anti-human CD68 antibody (0·5 µg/ml; M0814, clone KP1, IgG₁κ, Dako). Purified mouse isotype (mouse IgG₁κ, M-5284; Sigma) served as negative control. Next the slides were coated with a biotin-conjugated goat-anti-mouse-Alexa546 secondary antibody (1/300 diluted; IgG (H + l), A110-03; Molecular Probes). Nuclei were 4′,6- diamidino-2-phenylindole (DAPI)-stained with a mounting medium (Vectashield®; Vector Laboratories Inc., Burlingame, CA, USA). Immunofluorescent green (Alexa488), red (Alexa546) and/or blue (DAPI)-stained sections were captured separately (four high-power fields each) with a microscope at a 400-fold magnification using fluorescent light (Leitz DM RBE; Leica, Bensheim, Germany).

Animal model

Female Balb/c mice weighting 20–25 g (Charles River, Sulzfeld, Germany) were used. Except for the periods of induction of colitis they had food and water ad libidum. The animal studies were approved by the local Institutional Review Board.

For induction of acute colitis mice received 2·5% dextran–sulphate–sodium (DSS) in drinking water for 7 days. Mice used for experiments were age-matched and had received DSS treatment simultaneously. The control group received normal drinking water.

Cathepsin inhibition in vivo

Inhibitors were diluted in 5% dimethylsulphoxide (DMSO) containing sterile PBS and stored at −20°C until use. CTSB and CTSL inhibition was performed by simultaneous intraperitoneal (i.p.) injection of 5 mg/kg body weight l-trans-Epoxysuccinyl-Ile-Pro-OH propylamide (CA-074) (Bachem, Weil am Rhein, Germany) specific for CTSB and 5 mg/kg body weight Z-Phe-Tyr-aldehyde (Bachem) specific for CTSL daily over 7 days starting on day 3 after induction of colitis. The reason for performing simultaneous inhibition of CTSB and CTSL is the difficulty to validate specific roles for CTSB and CTSL due to conflicting information about potentially overlapping substrate specificities of cysteine-type cathepsins.

CTSD inhibition was performed by i.p. injection of 10 mg/kg body weight synthetic pepstatin A (Chemicon, Hampshire, UK) according to the same protocol. The control groups received PBS with corresponding volumes of DMSO. Each group consisted of five mice.

Assessment of histological score in mice

For the assessment of the histological score 1 cm of the distal third of the colon was removed and scored as described [26]. Mice were scored individually. Each score represents the mean of two sections. The histological examination was performed blinded.

Statistical analysis

Data are expressed as mean ± standard deviation (s.d.) or standard error of the mean (s.e.). Statistical analyses were performed using the SigmaPlot 8·0 t-test and spss for Windows version 12·0. Differences were considered significant at a P-value of < 0·05.

Results

CTSB and CTSL mRNA expression in human IMAC

CTSB and CTSL mRNA expression was analysed in IMAC from normal and inflamed mucosa from IBD patients by RT–PCR. In contrast to CTSD mRNA, which we had found to be expressed only in IMAC isolated from inflamed mucosa, mRNA expression of CTSB and CTSL was found in IMAC from all samples (Fig. 1a).

Fig. 1.

(a) Cathepsins B and L mRNA expression in CD33-positive intestinal macrophages (IMAC) from a patient with non-inflamed (NI) mucosa and two patients with inflammatory bowel disease (IBD) [ulcerative colitis (UC), Crohn's disease (CD)]. Reverse transcription–polymerase chain reaction (RT–PCR) for CTSB, CTSL and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as indicated. Positive control (+): gel extraction of digested CTSB or CTSL DNA fragment; negative control (–): without DNA. This experiment was performed three times independently with IMAC of different patients each. (b) Analysis of CTSB and CTSL expression in IMACs from control patients (NI) and from IBD patients. mRNA was isolated from IMAC of mucosal specimens, reverse transcribed and amplified by Taqman®–PCR. Cathepsin/GAPDH mRNA ratio (arbitrary units) as indicated. Every box consists of mean values of triple mRNA quantification of each patient. The number of patients tested is indicated. Both CTSB and CTSL expression was detected in IMAC from control patients and from patients with IBD. A significant increased expression of CTSL in IBD was demonstrated (*P < 0·05). A tendency for a higher expression of CTSB in IBD was seen.

However, when we quantified mRNA expression using TaqMan®–PCR, expression of CTSL was significantly up-regulated up to 10-fold in IBD IMAC compared to IMAC from control mucosa (Fig. 1b). A trend for an increased mRNA expression of CTSB in IBD mucosa was also found; however, the difference was not significant due to the high variability of CTSB mRNA expression in inflammation-associated IMAC. Together with our findings on CTSD mRNA expression, these results indicate an induction of cathepsin mRNA in IMAC during mucosal inflammation.

Protein expression of CTSL protein in human mucosa

Human colonic specimens without inflammation and with severe inflammation were compared for the protein expression of CTSL by immunohistochemistry. CTSL protein was detected in specimens from patients with histologically non-inflamed mucosa (Fig. 2a,b) and in inflamed mucosal tissue from UC patients (Fig. 2c,d) and CD patients (Fig. 2e). CTSL-positive cells were accumulated subepithelially and located preferentially close to the crypts. In the isotypes no specific staining was detected (Fig. 2f). The number of CTSL-positive cells was higher in areas of mucosal ulcerations and mononuclear cell accumulations (Fig. 2c,e).

Fig. 2.

Detection of cathepsin L (CTSL)-expressing cells in human intestinal mucosa by immunohistochemistry. CTSL-positive cells were visualized with NovaRED (brown) and the macrophage marker CD68 with benzidine dihydrochloride (BDHC) (blue reaction product). (a) Single-stained CTSL-positive cells (brown) in control mucosa. Cells were counterstained with haematoxylin. (b–e) CTSL (brown) could be co-localized with macrophages (blue) in the mucosa of a control patient (b), a patient with ulcerative colitis (UC) (c, d) and a patient with Crohn's disease (CD) (e). Black arrows indicate double-stained cells. CTSL-positive macrophages were accumulated in areas of severe inflammation and tissue damage (c, e). (f) Isotype control (non-inflamed mucosa). (g) Single stained macrophages (blue) in control mucosa. No immunostaining (brown) was detected with CTSL isotype control antibody in the first immunostaining step. Original magnification 200 × (a, b, c, f) and 400 × (d, e, g). The figure is representative of two additional experiments.

To identify the cell types expressing CTSL we applied double-labelling immunohistochemistry. We have shown previously that CTSD is expressed by IMAC from IBD patients [12]. The distribution of CTSL-positive cells (Fig. 2) led us to assume that they were likely to be IMAC. CD68 was used as characteristic macrophage marker and visualized by BDHC as an intracellular, blue deposit (Fig. 2g). Both in non-inflamed mucosal tissue and in the mucosa of IBD patients CTSL was detected in CD68-positive IMAC. In addition, few CTSL-negative IMAC were visible (Fig. 2c–e).

Three isotype stainings were performed: substitution of the primary anti-CTSL antibody, substitution of the anti-CD68 antibody or both, with corresponding isotype antibodies. Only brown immunostaining of CTSL positive cells was detected in the mucosa if instead of anti-CD68 an isotype control antibody was applied. When both primary antibodies were replaced neither CTSL nor CD68 immunostaining were detectable (Fig. 2f). After substitution of the anti-CTSL antibody with an isotype antibody only the blue CD68-immunostaining was detected (Fig. 2g).

Protein expression of CTSB in human mucosa

Colonic specimens without inflammation and with severe inflammation were compared for CTSB protein expression by immunofluorescence staining. CTSB protein was detected in specimens from histologically non-inflamed mucosa (Fig. 3a) and in inflamed mucosa tissue from UC-patients (Fig. 3b) and CD patients (Fig. 3c). By substituting the primary antibody with an isotype control no immunofluorescence was detectable (data not shown). Specimens of three controls and three patients, each with severe inflammation due to CD or UC, were stained for CD68. CTSB protein expression could be located to IMAC both in non-inflamed mucosa (Fig. 3a, merge) and tissue of patients with UC (Fig. 3b, merge) and patients with CD (Fig. 3c, merge). Accumulation of CTSB expressing IMAC could again be found in areas of ulceration and tissue damage.

Fig. 3.

Immunofluorescent detection of cathepsin B (CTSB)-expressing cells in human intestinal mucosa from non-inflamed (NI) patients and patients with inflammatory bowel disease (IBD). Paraffin-embedded sections were fluorescently single- and double-stained for CTSB (green), macrophage-specific CD68 (red) and diamidinophenylindole (DAPI). CTSB expressing intestinal macrophages (IMAC) were detectable in mucosa from a NI patient (a), from a patient with ulcerative colitis (b) and a patient with Crohn's disease (c). White arrows indicate double-stained cells. Original magnification 400 ×. The figure is representative for two additional experiments.

Specific immunostaining was ensured by three isotype controls with substitution of the primary anti-CTSB-antibody, substitution of the anti-CD68-antibody or both with corresponding isotype antibodies (data not shown).

As protein expression of CTSB and CTSL (as well as CTSD, which is not shown here) was associated with inflammation and mucosal damage, we intended to investigate the contribution of cathepsins to the pathophysiology of colitis by inhibiting the proteases in the murine DSS model.

Inhibition of CTSD activity in experimental colitis

The contribution of IMAC CTSD expression to the pathophysiology of colitis and the influence of CTSD inhibition on intestinal inflammation was investigated in the acute DSS-colitis model. Mice received DSS for 7 days and were treated with pepstatin A from day 3 for further 7 days. On day 10 the mice were killed and colon length and histological scores were evaluated. Characteristic colitis symptoms such as inflamed anus and bloody diarrhoea were most prominent in the non-inhibitor treated colitis group, whereas they were virtually absent in the pepstatin A-treated animals.

CTSD inhibition resulted in an amelioration of DSS-induced colitis in all parameters investigated. Control animals receiving water instead of DSS solution did not show significant weight loss during the experiment (Fig. 4). Weight loss was 17·1% in the PBS-treated colitis group versus 9·3% of the pepstatin A-treated group, P < 0·001 (Fig. 4). Colitis is followed typically by a reduction of the colon length. Colon length of non-treated colitis mice (9·6 ± 1·1 cm) was reduced significantly compared to the pepstatin A-treated group (11·0 ± 1·3 cm, P = 0·039; Fig. 5). Mice without DSS-induced colitis showed a normal colon length with 12·5 ± 1·3 cm in the H₂O/PBS group and 12·7 ± 0·9 cm in the H₂O/pepstatin A group (Fig. 5). The histological score (HS) of non-treated colitis mice was 3·6 ± 0·5 versus 2·1 ± 0·8 in the pepstatin A-treated colitis group, P < 0·03 (Fig. 6). As expected, mice without DSS treatment did not develop an inflammation, proved by a histological score of 0·6 ± 0·9 in the PBS group and 0·4 ± 0·4 in the pepstatin A-treated group. Notably, all mice (100%) of the DSS/PBS control group developed a strong inflammation with typical colitis symptoms as described above, while in the inhibitor treated group only two of five mice (40%) developed such signs of inflammation. Tissue sections shown in Fig. 7 show severe inflammation in DSS-induced colitis mice with obvious tissue destruction and mononuclear infiltrate (Fig. 7a), whereas there was only mild inflammation in pepstatin A-treated mice (Fig. 7b) without signs of ulcerations. Mice without DSS treatment did not develop an inflammation (Fig. 7c,d).

Fig. 4.

Weight course of inhibitor treated mice versus PBS treated control mice. Weights are indicated for only the first 7 days of the experiment because of misleading shifts in the weight curves from day 8 resulting from dying or advanced killing of four mice. The arrow indicates the first dose of pepstatin A. Datapoints are mean values of each group ± standard error of the mean (n = 5). The weight course of the dextran–sulphate–sodium/phosphate-buffered saline (DSS/PBS) group showed significant differences versus the DSS/pepstatin A group (variance analysis, P < 0·001) as well as versus the H₂O/pepstatin A group (variance analysis, P = 0·006).

Fig. 5.

Colon lengths of pepstatin A-treated mice (triangles) and phosphate-buffered saline (PBS)-treated mice (circles) with and without experimental colitis. The horizontal line indicates the mean values. Apart from the dextran–sulphate–sodium/PBS group that contained four mice each group consisted of five mice (*P = 0·039).

Fig. 6.

Histological appearance of colitis in mice induced by treatment with dextran–sulphate–sodium. Pepstatin A-treated colitis mice developed an inflammation according to a histological score (HS) of 2·1 ± 0·8 (grey triangles) while the phosphate-buffered saline (PBS)-treated control group showed an inflammation according to 3·6 ± 0·5 (severe inflammation, black circles). Mice without colitis but treated with inhibitor (dark grey triangles) and PBS-treated controls (grey circles) did not develop an inflammation. Horizontal lines indicate mean values of each group. Scoring was performed blinded by an independent person (F. O.) (*P < 0·03).

Fig. 7.

Photomicrographs of adjacent sections of murine colon tissue stained with haematoxylin and eosin (H&E). (a) Colon section of a dextran–sulphate–sodium (DSS)-colitis mouse treated with phosphate-buffered saline (PBS); (b) colon section of a DSS-colitis mouse treated with pepstatin A; (c) colon of a mouse receiving pure water and treated with pepstatin A; (d) colon of a mouse receiving pure water and treated with PBS. A dramatic reduction of mucosal thickening, lymph follicles enlargement and inflammatory cells accumulation is found in pepstatin A-treated mice. Magnification 100 ×.

Inhibition of CTSB and CTSL activity in experimental colitis

To test whether CTSB and CTSL contribute to inflammation and tissue damage of the intestinal mucosa, the therapeutic effect of inhibitors was investigated in DSS-induced colitis. Mice received DSS in the drinking water for 7 days. Treatment with the specific inhibitors CA-074 (for CTSB) and Z-Phe-Tyr-aldehyde (for CTSL) for a combined inhibition of CTSB and CTSL was applied i.p. simultaneously from days 3–9.

Weight was measured daily, the experiment was terminated on day 10 and colon lengths and histological scores were evaluated. The inhibitors reduced the severity of colitis in all parameters tested. Characteristic colitis symptoms such as inflamed anus and bloody diarrhoea were prominent in the non-inhibitor-treated DSS group.

The weight loss after 10 days was 25% in the non-treated DSS group versus 6% in the inhibitor-treated group (Fig. 8; P < 0·01). The colon length was 9·1 ± 1·2 cm in the untreated group versus 10·6 ± 0·8 cm in the treatment group (Fig. 9; P < 0·048). The histological score was 3·5 ± 0·5 in the untreated animals compared to 1·8 ± 1·3 in the inhibitor treated group (Fig. 10a, P = 0·029). In the DSS/PBS control group all mice (100%) developed characteristic colitis symptoms with inflamed anus and bloody diarrhoea. Remarkably, in the inhibitor-treated group (DSS/inhibitors) only one mouse of five (20%) even developed visible intestinal inflammation. Tissue sections showed severe inflammation with a massively destroyed mucosa with ulcerations, large lymph follicles, a thickened submucosa and infiltrating cells in the untreated colitis mice (Fig. 10bi), whereas there was only mild inflammation in the inhibitor-treated group (Fig. 10bii).

Fig. 8.

Weight course of inhibitor-treated mice versus phosphate-buffered saline (PBS)-treated control mice. Inhibition of cathepsin B (CTSB) and CTSL was performed intraperitoneally simultaneously with 5 mg/kg CA-074 for CTSB and Z-Phe-Tyr-aldehyde for CTSL given daily over 7 days. The arrow indicates the first dose of inhibitors. Inhibitor-treated mice (open circles) showed a significantly lower weight loss versus the PBS-treated control group (filled circles) (variance analysis, P < 0·01). Datapoints are mean values of each group ± standard error of the mean (n = 5).

Fig. 9.

Colon lengths of cathepsin B (CTSB)- and CTSL-inhibitor-treated mice (open circles) and phosphate-buffered saline (PBS)-treated control mice (filled circles). The horizontal line indicates the median value of each group (*P = 0·048).

Fig. 10.

(a) Histological scores (HS) of inhibitor treated mice and phosphate-buffered saline (PBS)-treated mice with and without experimental colitis. Inhibitor-treated dextran–sulphate–sodium (DSS)-colitis mice had a HS of 1·8 ± 1·3 while the PBS-treated control group showed a HS of 3·5 ± 0·5 (severe inflammation). Horizontal lines indicate the mean values of each group (n = 5). Scoring was performed blinded by an independent person (F. O.) (*P = 0·029). (b) Photomicrographs of adjacent sections of murine colon tissue stained with haematoxylin and eosin. (i) Colon section of a DSS-colitis mouse treated with PBS. The colitis is characterized by a thickening of the mucosa, large lymph follicles and an influx of inflammatory cells. (ii) colon section of a DSS-colitis mouse treated with inhibitors. Magnification 100 ×; m, intestinal mucosa; f, lymph follicle.

Discussion

In the present paper we show (1) an induction of CTSL (and to a lesser extent CTSB) expression in IMAC during mucosal inflammation and (2) an important pathophysiological role of the IMAC-produced proteases CTSD, CTSB and CTSL, as their inhibition in DSS-induced colitis is followed by a clear amelioration of the disease. To demonstrate the presence of CTSB and CTSL in IMAC we performed double-labelling immunohistochemistry with the macrophage marker CD68 in human tissue samples.

An involvement of cathepsins in inflammatory disease had been shown previously [15–18]. In a mRNA screening by subtractive hybridization and Affimetrix-array analysis we had found an induction of tissue degrading enzymes in general and a specific induction of CTSD in IMAC in IBD versus control mucosa [8,12]. CTSD inhibition with pepstatin A resulted in an amelioration of inflammation in inhibitor-treated mice in comparison to controls demonstrated by less weight loss, reduced colon shortening and low histological scores. Similarly, simultaneous inhibition of CTSB and CTSL with the specific CTSB inhibitor CA-074 and the specific CTSL inhibitor Z-Phe-Tyr-aldehyde led to a decreased extent of inflammation as shown by the same parameters.

The aspartic proteinase CTSD is a member of the family of cysteine proteinases. Unlike other proteinases, which are mainly secretory proteins, Pro-CTSD is sorted to the lysosomes. Under normal physiological conditions neither Pro-CTSD nor CTSD are secreted and are not found extracellularly. The lysosomal cysteine proteinases CTSB and CTSL participate in degradation of extra- and intracellular proteins. CTSB and CTSL are also involved in antigen processing in concerted action with CTSD and the formation of peptide-receptive dimers during major histocompatibility complex (MHC) class II-restricted antigen presentation [27,28]. The proteolytic activity of cathepsins B, D, H and L plays an important role in the generation of antigenic peptides from larger polypeptides presented on class-II molecules to T cells [27,28]. Several classes of proteinases, including cathepsins B, L and D, are thought to be involved in lysosomal turnover of proteins with degradation of the basement membrane and surrounding ECM [29]. CTSB is thought to work as activator of other proteinases which, in turn, mediate degradative processes [30]. Unlike CTSD, cathepsins B and L are secretory proteins. Extracellular accumulation of mature CTSL is up-regulated by inflammatory stimuli, as observed in interferon-γ-treated macrophages and lipopolysaccharide (LPS)-activated dendritic cells (DCs) [31].

Macrophages mobilize proteinases and participate in the pathophysiological remodelling of the ECM in numerous tissue-destructive diseases, e.g. arthritis, bone resorption or metastasis. CTSB and CTSL expressing CD68-positive human mononuclear cells were shown to play important roles in patients with rheumatoid arthritis, where they take part in joint destruction and bone erosion [32]. Remodelling of the ECM by proteases such as human alveolar macrophage–CTSL is suggested to play a role in lung emphysema [33]. Furthermore, during chronic pulmonary inflammation, the activities of cathepsins L, B, H and S in isolated alveolar macrophages were found to be increased strongly [34]. CTSB activity was shown to be stimulated in peritoneal macrophages by induction of an acute peritonitis in mice [35]. Our data also now indicate a pathophysiological role of cathepsins in IBD. As these proteases are expressed by IMAC, this highlights further the role of these cells.

How could cathepsins be involved in IBD pathophysiology? By their proteolytic activity on extra- and intracellular proteins secreted CTSB and CTSL, as well as by macrophage damage or apoptosis, liberated CTSD may destroy ECM components leading to tissue destruction, mucosal damage, barrier defects and bacterial invasion. In healthy individuals the proteolytic action of cathepsins is restricted to controlled protein degradation in cellular processes and metabolism. In patients with IBD, proteolysis by cathepsins may also be triggered by bacterial invasion resulting in tissue destruction and inflammation.

Our data indicate not only that IMAC derived cathepsins may play an important role in the pathophysiology of IBD; they also provide a new target for IBD therapy. Whereas so far most therapeutic strategies aim on a reduction of inflammation by inhibiting T cells (or macrophages), our data support an alternative concept of inhibition of tissue destruction. Cathepsins are an ideal target for therapy. Intracellularly they play an important role, especially for macrophage function. A CTSD knock-out is followed by death shortly after birth [24]. Combined homozygous deficency of CTSB and CTSL is also lethal between the second to fourth week of life. Heterozygous cathepsin knock-out shows no abnormal phenotype and is not lethal [36].

The inhibitors used in our setting, on the other hand, do not pass through the cell membrane and remain extracellularly. This means that in a therapeutic approach the extracellular action of cathepsins can be blocked selectively, whereas the intracellular essential function of cathepsins remains preserved.