ORIGINAL ARTICLE

Blockade of tumor necrosis factor-a-converting enzyme improvesexperimental small intestinal damage by decreasing matrixmetalloproteinase-3 production in rats

HIROSHI MATSUMOTO1, HIDEKI KOGA1,2, MITSUO IIDA2, KEN-ICHI TARUMI1,

MINORU FUJITA1 & KEN HARUMA1

1Division of Gastroenterology, Department of Internal Medicine, Kawasaki Medical School, Kurashiki, Okayama, Japan,

and 2Department of Medicine and Clinical Science, Graduate School of Medical Sciences, Kyushu University, Fukuoka,

Japan

AbstractObjective. Tumor necrosis factor (TNF)-a-converting enzyme (TACE), which has been purified, regulates maturity ofTNF-a. Matrix metalloproteinases (MMPs) play a key role in various inflammatory conditions. The incidence of intestinaldamage has increased, but the mechanism and treatment have not been well understood. The purpose of this study was toinvestigate the roles of TACE and MMP in indomethacin (Indo)-induced intestinal damage as well as the therapeutic effectsof TACE inhibitor and selective MMP inhibitor (sMMPi) on this intestinal damage in rats. Material and methods. In thefirst experiment, serial changes in intestinal ulcers and the production of MMP were investigated. In the second experiment,we assessed the effect of three TACE and/or MMP inhibitors and the production of TNF-a, TACE, MMP-3, -9 and tissueinhibitor of MMP (TIMP)-1. The rats were divided into five groups: a control group, and four groups that received Indoalone, Indo plus TACE inhibitor (GM6001), Indo plus a selective MMP-3 inhibitor and Indo plus an MMP-9/13 inhibitor,respectively. Results. MMP-3 was overexpressed at 24 h after Indo administration, when intestinal injury was mostprominent macroscopically and microscopically. GM6001 significantly decreased ulcer severity and suppressed MMP-3 in adose-dependent fashion. The selective MMP-3 inhibitor dose-dependently ameliorated intestinal damage to the samedegree as GM6001, but the MMP-9 inhibitor had no effect on the injury. Conclusions. MMP-3 inhibition amelioratesintestinal damage without apparently affecting either TNF-a or TACE production and the dose�response curve suggeststhat the beneficial effect of the so-called TACE inhibitor is actually mainly mediated via MMP-3 inhibition rather thanTNF-a inhibition.

Key Words: Inflammation, matrix metalloproteinase (MMP)-3, matrix metalloproteinase (MMP) inhibitor, non-steroidal

anti-inflammatory drugs (NSAIDs)-induced intestinal damage, tumor necrosis factor (TNF)-a, tumor necrosis factor

(TNF)-a-converting enzyme (TACE, ADAM17, CD156b)

Introduction

It has been suggested that tumor necrosis factor

(TNF)-a is an important mediator in the pathogen-

esis of various inflammatory diseases, including

inflammatory bowel disease (IBD) [1]. Two forms

of TNF-a have been identified: a membrane-bound

precursor 26 kD protein, and a 17 kD mature

secreted form. The cleavage of cell-associated

TNF-a is mediated by metalloproteinases, a disin-

tegrin and metalloproteinase (ADAM) known as

TNF-a-converting enzyme (TACE, ADAM17,

CD156b) [2,3]. TACE shares a significant degree

of structural homology with matrix metalloprotei-

nase (MMP) in enzyme structures, particularly

around zinc-containing active sites. Firm evidence

showing TACE as the major TNF-a convertase has

led to TACE attracting considerable interest as a

specific therapeutic target in diseases known to

benefit from anti-TNF-a treatment [4�6].

MMPs, a family of zinc-containing enzymes that,

as a group, can degrade and remodel essentially all

Correspondence: Hiroshi Matsumoto, MD, PhD, Division of Digestive Diseases, David Geffen School of Medicine at UCLA, MacDonald Research Building,

MRL-2736, 675 Charles E. Young Circle Dr. South, Los Angeles, California 90095, USA. Fax: �/1 310 8255 204. E-mail: HMatsumoto@mednet.ucla.edu

Scandinavian Journal of Gastroenterology, 2006; 41: 1320�1329

(Received 16 December 2005; accepted 3 March 2006)

ISSN 0036-5521 print/ISSN 1502-7708 online # 2006 Taylor & Francis

DOI: 10.1080/00365520600684571

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

extracellular matrix components, have a role in

connective tissue remodeling processes associated

with embryonic development, pregnancy, growth,

and wound healing [7,8]. MMPs are produced

mainly by myofibroblasts, but in intestinal tissue

of human IBD, they are also produced by submu-

cosal mononuclear cells which show positive expres-

sion of CD68 [9]. The production of MMPs is

stimulated by lipopolysaccharide (LPS) and pro-

inflammatory cytokines such as TNF-a and inter-

leukin (IL)-1. Consequent imbalance of MMPs with

a physiologic inhibitor, a tissue inhibitor of MMPs

(TIMPs), can result in degradation of the extra-

cellular matrix.

Inhibition of TACE and/or MMP is one antici-

pated approach in the treatment of inflammatory

disease [5�8]. Previous studies have demonstrated

that some potent broad-spectrum inhibitors of

MMPs, called TACE inhibitors, can also prevent

TNF-a release by inhibiting TACE. During clinical

trials, some inhibitors of TACE/MMP unfortunately

have shown adverse effects in patients with cancer or

rheumatoid arthritis, suggesting that these drugs

have too broad a range of MMPs suppression [10].

Therefore, development of more specific inhibitors

of MMP and TACE is needed, but the question of

which MMP is the most important therapeutic

target in intestinal inflammation remains unclear.

Furthermore, the effect of selective MMP inhibition

against experimental intestinal injury has not been

well defined.

We previously reported on the beneficial effects of

antibiotics, some anti-pro-inflammatory cytokine-

neutralizing antibodies, and immunosuppressive

agents against indomethacin (Indo)-induced intest-

inal injury [11�15]. This model is a suitable experi-

mental model showing small intestinal damage by

chemical agents, and has two immunopathological

characteristics: first, a non-steroidal anti-inflamma-

tory drug (NSAID)-induced small intestinal injury,

and the other is also deemed to be a small intestinal

ulcer model of IBD, because they showed the

T-helper (TH)-1-dominant intestinal inflammation

resembling IBD [16]. The aim of the present study

was to investigate the roles of TACE and MMPs in

such intestinal ulceration as well as the therapeutic

effects of the TACE inhibitor and selective MMP

inhibitor (sMMPi) on this experimental intestinal

injury. Among the whole cohort of MMPs, we

focused on MMP-3 and MMP-9 in this study,

because these MMPs were secreted not only by

mucosal myofibroblasts but also by monocyte/

macrophages whose number was dramatically in-

creased in the TH-1-dominant inflamed intestinal

mucosa.

Materials and methods

Animals

Male Wistar rats (weight 170 to 220 g, age

6�/7 weeks) were purchased from Charles River

Japan (Kanagawa, Japan) for use in this study. The

animals were maintained in a room with a controlled

temperature (mean9/SD, 239/38C) and a light�dark

cycle of 14�10 h. The rats were housed in plastic

cages, with a maximum of three animals per cage.

Standard pelleted laboratory formula and tap-water

were provided ad libitum . All experiments were

approved by the Animal Research Committee of

Kawasaki Medical School (Study no. 03-049, 2003)

and conducted in accordance with the ‘‘Guide for

the Care and Use of Laboratory Animals’’ of

Kawasaki Medical School.

Induction of small intestinal damage

The rats were weighed, randomized to one of four

inhibitor-treated groups or the control group, and

then anesthetized with an intraperitoneal injection of

sodium amobarbital (100 mg/kg). A 6-cm-long plas-

tic catheter was inserted into the colon, and 24 mg/

kg indomethacin (Indo) (Sigma Chemical Co., St.

Louis, Mo., USA) dissolved with 1% sodium car-

boxymethylcellulose was injected via the catheter.

Intrarectal Indo administration was chosen for this

study, because this method consistently produces

ulcerations located predominantly in the small

intestine [15].

Serial change in small intestinal ulcerations and MMP

production during development intestinal injury

The rats were sacrificed at 6 h, 12 h, 24 h, 3 days,

and 7 day after Indo administration. The small

intestines were removed and assessed macroscopi-

cally and histologically by the methods described

below. MMP production and activities were assessed

by Western blotting, zymography, and immunohis-

tochemistry.

Treatment with a TACE or one of two selective MMP

inhibitors

All animals were killed for evaluation of small

intestinal damage and production of TNF-a,

TACE, MMP-3, MMP-9, and TIMP-1 24 h after

receiving indomethacin. The rats were divided into

five groups as follows: no treatment (control); Indo

administration alone; Indo plus 0.2, 1, 10, or 30 mg/

kg of a TACE inhibitor (GM6001, Galardin; Cal-

biochem, Darmstadt, Germany) diluted with di-

methyl sulfoxide (DMSO); Indo plus 0.2, 1, or

5 mg/kg selective MMP-3 inhibitor VI (sMMP3i,

TACE/MMP3 inhibitor improved experimental intestinal damage 1321

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

Calbiochem) diluted with DMSO; Indo plus 0.2, 1,

or 5 mg/kg MMP-9/13 inhibitor II (MMP9i, Calbio-

chem) diluted with DMSO. Inhibitors were injected

intraperitoneally 3 h after Indo administration.

Macroscopic and microscopic assessments of longitudinal

ulceration of the small intestine

The small intestine was removed, opened by long-

itudinal incision, pinned flat on a wax block, washed

with saline, and fixed in 10% formalin for 48 h.

Fixed specimens were examined macroscopically

and microscopically for small intestinal damage.

Longitudinal ulcers were defined as �/10 mm long,

located on the mesenteric side of the lumen. As

previously described [15], macroscopic ulcerations

were evaluated by an ulcer index (UI) defined as the

ratio of the total length of longitudinal ulcers in the

small intestine to the total length of the small

intestine from the pylorus to the ileocecal junction.

Ulcers that were smaller than 10 mm in largest

dimension were not included. Tissue samples were

removed from the ulcers for histologic examination

as paraffin sections by a histopathologist who was

kept unaware of the treatment groups. All sections

were stained with hematoxylin and eosin (H&E). For

histologic assessment, we used a histologic damage

score (HDS) according to the modified criteria

reported by Vilaseca et al. [17]. The width of

ulceration was scored according to the following

criteria: 0, no ulcer; 1, below 3 mm; 2, 3 to 6 mm;

and 3, above 6 mm. The depth of ulceration was

scored as: 0, none; 1, mucosa; 2, submucosa; 3,

muscularis propria; or 4, serosa. Inflammatory cell

infiltration was scored as: 0, none; 1, mild; 2,

moderate; or 3, severe. Thrombi were scored as: 0,

absent; 1, present.

Collection of samples of intestinal ulcers

Tissue samples including ulcerations in the middle

portion of the small intestine were collected for

determination of protein production and activities

using the enzyme-linked immunosorbent assay

(ELISA), Western blotting, zymography, and im-

munohistochemistry. These samples were obtained

immediately after the rats were killed by amobarbital

overdose, and then washed with phosphate-buffered

saline (PBS), frozen in liquid nitrogen, and stored at

�/808C.

ELISA for TNF-a analysis

As described in previous studies [18], tissue was

homogenized in nine volumes of Greenberger lysis

buffer (300 mM NaCl, 15 mM Tris-HCl (pH 7.4),

2 mM MgCl2, 2 mM Triton X-100, Pepstain A

(20 ng/ml), leupeptin (20 ng/ml), and aprotonine

(20 ng/ml)). Tissue was lysed for 30 min on ice

followed by two centrifugations at 14,000 g for

10 min each. Homogenates were stored at �/208Cuntil further use. TNF-a concentrations were mea-

sured by the ELISA (R&D Systems, Abingdon,

England).

Zymography for MMP activity

Zymography was carried out in tissue homogenates

and conditioned media in accordance with the

published methods [19,20]. Tissue samples were

homogenized in sample buffer (20 mM Tris-HCl

(pH 7. 5), 0.01% Briji-35). After centrifugation at

48C and 10,000g for 20 min, supernatants were

collected from the samples. Protein concentrations

were measured using a bicinchoninic acid (BCA)

method. All samples were then adjusted to a uniform

total protein concentration, and added to triple

volumes of sodium dodecyl sulfate (SDS) zymogra-

phy buffer (0.5 M Tris-HCl (pH 6.8), SDS, 70%

glycerol, and 5% bromophenol blue (BPB)). Pro-

teins were separated on 12.5% polyacrylamide gels

(Bio-Rad, Hercules, Calif., USA) under non-redu-

cing, non-denaturing conditions with 0.4% gelatin/

casein as substrate (Sigma). The gels were then

washed in buffer (50 mM Tris-HCl (pH 7.5), 0.1 M

NaCl, 2.5% Triton X-100) for 90 min, and incu-

bated in reaction buffer (1 M Tris-HCl (pH 7.5),

1 M CaCl2, 1 mM ZnCl2, 2% NaN3, 2 mM

o-phenanthroline, and 200 mM phenylmethanesulfo-

nyl fluoride) at 378C for 30 h. Thereafter, they were

fixed and stained in 45% methanol, 10% acetic acid,

and 0.1% Coomassie blue R-250 (Sigma) for 2 h

and finally decolorized. Bands were determined as

resulting from the activity of MMP as opposed to

that of other proteinases by experiments using MMP

inhibitors in reaction buffers.

The relative molecular weights of the bands were

estimated by comparison to both molecular weight

standards and the uniform amount of active human

MMP-2, -3, and -9 (Wako, Osaka, Japan). Band

analysis was performed with a Macintosh (model

G4) computer using the public domain NIH Image

program (developed at the US National Institute of

Health and available via the Internet by anonymous

FTP from zippy.nimh.nih.gov or on a floppy disk

obtained from the National Technical Information

Service (Springfield, Va., USA; part number PB95-

500195GEI). Zymographic activity was measured by

densitometric analysis in 1-D gels using an external

density standard (Kodak, Rochester, N.Y., USA)

and NIH Image 1.62. We defined protein produc-

tion in the control group as 100%. Typical zymo-

1322 H. Matsumoto et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

graphy gel illustrating the MMPs positive control is

shown in Figure 1.

Western immunoblotting

The frozen samples were homogenized in sample

buffer (20 mM Tris-HCl (pH 7.5), and 0.01% Briji-

35). After total protein amounts were measured by

BCA and then made uniform, samples were added

to SDS buffer (125 mM Tris-HCl (pH 6.8), 4%

SDS, 20% glycerol, 10% 2-mercaptoethanol,

0.002% BPB, and 0.01% protein inhibitor cocktail

set V (EDTA-free, Calbiochem), and heated to

1008C for 5 min. For measurement of TACE, the

samples were then added to the buffer (0.2% Nondit

P40, 100 mM 1,10 � phenanthroline, and 1 mM

dithiothreitol), and incubated at 08C for 60 min.

Samples were loaded (10 mg/lane) onto 10�12.5%

polyacrylamide gels. After sodium dodecyl sulfate-

polyacrylamide gel electrophoresis (SDS-PAGE),

proteins were transferred onto a PVDF membrane

(Hybond-P, Amersham, England) and detected

using mouse anti-human MMP-3, -9 (4 mg/ml;

Daiichi Fine Chem., Takaoka, Japan), anti-TACE

(1:1000; Oncogene Res, San Diego, Calif., USA), or

anti TIMP-1 (1:500; Daiichi Fine Chem.); followed

by peroxidase-linked-anti-mouse IgG secondary

antibody. By means of an enhanced chemilumines-

cence (ECL) detection system (Amersham), mem-

branes were exposed to appropriate film (Hyperfilm

ECL; Amersham). Band analysis was performed on

the Macintosh computer using the public domain

NIH Image program. We defined MMP activity in

control group as 100%.

Immunohistochemistry

Deparaffinized 4-mm sections were stained immuno-

histochemically (streptoavidin-biotin method) using

mouse anti- human MMP-3, or -9 antibody (1:800).

These had been recognized as cross-reacting with

corresponding rat MMPs. Endogenous peroxidase

activity was blocked by incubation in 1% H2O2 for

30 min at room temperature. After microwave-oven

antigen retrieval, the specimens were exposed to

anti-MMP antibodies, and then to the secondary

antibody and color development reagents. Nuclei

were counterstained with hematoxylin. Immunor-

eactive cells were counted at six random points at

either the edge or central area of ulceration.

Statistical analysis

All values were expressed as the means9/SD. Com-

parisons between groups were made using Student’s

t-test; p-values of less than 0.05 were considered

significant. Statistical analysis was performed using

Graph Pad Prism (version 4.00 for Macintosh;

GraphPad Software, San Diego, Calif., USA).

Results

Serial change in intestinal ulcerations and MMP

production during development of Indo-induced intestinal

damage

Both the UI and HDS were at their peak 24 h after

Indo administration (Figure 2A, B). MMP-3 pro-

duction and activities, assessed by Western blotting,

zymography, and immunohistochemistry were also

overexpressed at 24 h (Figure 3A, B, C, D), whereas

MMP-9 production and activities were not.

Both GM6001 and sMMP3i decreased the UI and

HDS dose-dependently, whereas MMP9i did not. In

the group treated with Indo alone, the UI and HDS

had respective values of 16.59/1.1 and 7.69/0.2

(Figure 4A, B). In the group treated with TACE

inhibitor, GM6001 significantly and dose-depen-

dently decreased both the UI and HDS. Using

GM6001 at the 30 mg/kg dose, no ulceration

occurred (UI, 0; HDS, 0.259/0.2), and we therefore

did not collect frozen samples or examine protein

production at 30 mg/kg of GM6001. sMMP3i also

had a beneficial effect, to almost the same degree as

GM6001, with a minimum UI of 1.09/0.3 and a

minimum HDS of 1.79/0.6 (5 mg/kg). Regarding

the HDS, the main effect of these two drugs was to

decrease the number of infiltrating inflammatory

cells (Figure 7A, B, C, D). In contrast, MMP9i did

not cause any significant decrease in either the UI or

HDS. Even the minimum UI and HDS levels were

still at 13.79/2.5 and 4.79/0.6.Figure 1. Typical zymography gel illustrating the standard of

matrix metalloproteinase (MMP).

TACE/MMP3 inhibitor improved experimental intestinal damage 1323

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

GM6001 dose-dependently suppressed produc-

tion of both TNF-a and TACE, whereas sMMP3i

did not. Production of both TNF-a and TACE in

the Indo alone group significantly increased beyond

production in the normal group. GM6001 sup-

pressed production in a dose-dependent manner

(Figure 5A, B). Minimum TNF-a and TACE

production rates in the GM6001-treated group

were 35% and 36.5%, respectively, at the 10-mg/kg

dose, when production in the Indo alone group was

defined as 100%. In contrast, sMMP3i did not

significantly suppress production (respective mini-

mum production rates, 68.7% and 90%).

Both GM6001 and sMMP3i down-regulated

MMP-3 activity and production, while up-regulating

TIMP-1 production. MMP-3 activity and production

in the Indo alone group increased significantly

beyond the finding in the normal group (Figure 6A,

Figure 2. Serial change in small intestinal ulcerations after indomethacin (Indo) administration. The ulcer index of the small intestine was

at its peak at 24 h after Indo administration (A). The histological damage score also was at its peak at 24 h after Indo administration (B).

*p B/0.01 versus control.

Figure 3. Serial change in matrix metalloproteinase (MMP) production of the small intestine after indomethacin (Indo) administration.

MMP-3 production assessed by Western blotting (A) and by zymography (B) was at its peak 24 h after Indo administration. In the

immunohistochemical assessment, the numbers of MMP-3 positive cells at the edge of the ulcer (C) and in the central area of the ulcer (D)

were also at their peak 24 h after Indo administration. *p B/0.01 versus MMP-3 control group, $p B/0.01 versus MMP-9 of control group.

1324 H. Matsumoto et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

B, C). The maximum suppression by GM6001 was

99.5% and 99.1%, respectively, at the 10 mg/kg dose.

sMMP3i dose-dependently decreased MMP-3 activ-

ity and production to about the same degree as

GM6001. Maximum suppression was 85.1% and

99.5%, respectively, at the 5-mg/kg dose. GM6001

also dose-dependently down-regulated the produc-

tion and activity of MMP-9 (Figure 6D, E, F). In

contrast, there was no significant suppression with

sMMP3i. Both GM6001 and sMMP3i dose-depen-

dently up-regulated TIMP-1 production, with their

respective maximum values in the groups treated with

GM6001/sMMP3i being 22 and 31 times greater

than those in the normal group (data not shown).

Thus, the inhibitors changed the MMP-3/TIMP-1

balance to the degree that it resembled that in the

normal intestine. In addition, these drugs signifi-

cantly and dose-dependently decreased the number

of MMP-3- but not MMP-9- immunoreactive cells

both in the center and at the edge of the ulcerations.

MMP9i suppresses MMP-9 production, but not that of

TNF-a, TACE, MMP-3, or TIMP-1

MMP9i suppressed MMP-9 production, decreasing

the number of MMP-9-immunoreactive cells in the

center of the ulcerations, but had no effect on other

cytokines and MMP production. These findings do

not resemble those observed after treatment with

TACE inhibitor or sMMP3i (data not shown).

Discussion

Based on three observations made in this study, we

postulate that MMP-3 could be just as important as

the TACE and/or TNF-a protein in treating small

Figure 4. Effects of three inhibitors on indomethacin-induced intestinal ulceration. A. The macroscopic assessment ulcer index (UI); B.

microscopic assessment, and the histologic damage score (HDS). In the group treated with the tumor necrosis factor-a-converting enzyme

inhibitor (TACEi), GM6001 significantly decreased both UI and HDS in a dose-dependent manner. Selective matrix metalloproteinase-3

inhibitor (sMMP3i) ameliorated this enteropathy to a similar degree. Matrix metalloproteinase-9 inhibitor (MMP9i) did not have any

beneficial effect against intestinal ulceration. *p B/0.01 versus control, **p B/0.01 versus Indo alone.

Figure 5. TNF-a- and TNF-a-converting enzyme (TACE) production in intestinal ulcerative tissues treated with GM6001 and sMMP3i.

A. Concentration of TNF-a; B. Production of TACE. GM6001 significantly suppressed production of both TNF-a and TACE in a dose-

dependent manner; sMMP3i did not suppress either TNF-a or TACE. *p B/0.01 versus control, **p B/0.01 versus Indo alone, %%p B/0.05

versus Indo alone.

TACE/MMP3 inhibitor improved experimental intestinal damage 1325

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

intestinal damage with TACE/MMP inhibitors.

First, it was found that MMP-3 was involved in

the development of intestinal damage following Indo

administration. Second, two effective inhibitors

against this experimental injury, sMMP3i and a

TACE inhibitor, showed MMP-3 suppression. We

found that the TACE inhibitor, GM6001, along with

sMMP3i had a beneficial effect against this intestinal

injury, but not with MMP9i. Interestingly, the

efficacy of sMMP3i was equal to that of GM6001

in terms of longitudinal ulcerations of the small

intestine. These two effective drugs, sMMP3i and

the TACE inhibitor, had two effects in common

against experimental injury, but differed with respect

to another two. The common effects in this entero-

pathy were dose-dependent MMP-3 suppression

Figure 6. Matrix metalloproteinase (MMP) production and activities in intestinal ulcerative tissue treated with GM6001 or sMMP3i:

MMP-3 production (A, B, C): MMP-9 production (D, E, F). GM6001 significantly suppressed both production and activities of MMP-3

and MMP-9 in a dose-dependent manner. In contrast, sMMP3i dose-dependently suppressed only MMP-3, not MMP-9. MMP-9

production in the presence of these inhibitors was not decreased from that in the Indo alone group.

Figure 7. Histologic appearance of ulceration of the small intestine (H&E). Indomethacin-induced longitudinal ulceration of the small

intestine (A). Intestinal tissue treated with GM6001 at 30 mg/kg (B). Intestinal tissue treated with sMMP3i at 5 mg/kg (C). Intestinal

ulceration treated with MMP9i at 5 mg/kg (D).

1326 H. Matsumoto et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

and TIMP-1 up-regulation. Differences involved

TNF-a and TACE production; GM6001, the

TACE inhibitor, suppressed the production of both

TNF-a and TACE, while sMMP3i did not suppress

production of either. Finally, an ineffective inhibitor,

MMP9i, show no MMP-3 suppression.

In the near future, small intestinal ulceration could

become a common gastrointestinal disease, because

the incidence of small intestinal damage that could

be induced mainly by NSAIDs and Crohn’s disease

(CD) could clinically increase [21,22]. Interestingly,

the small intestinal damage has been characterized

almost entirely by TH-1-dominant inflammation.

New methods of direct visualization of the small

bowel using a capsule or double-balloon endoscopy

have been developed and could also contribute to

making small intestinal ulceration more common

hereafter [23,24]. However, we have found very few

studies regarding a novel treatment for experimental

small intestinal ulcerations. Furthermore, there have

been no studies comparing the effectiveness of some

TACE/MMP inhibitors against intestinal damage.

The present study presents new information on the

effectiveness of three TACE/MMP inhibitors against

small intestinal damage.

Some studies have suggested that either MMP-3 or

MMP-9 is related to intestinal inflammation

[9,19,20,25,26]. However, there has been no evi-

dence to show MMPs, especially MMP-3 or MMP-

9, to be more important in intestinal damage and a

more effective therapeutic target MMP against in-

testinal damage. However, production of MMP-3

and/or MMP-9 has been reported to be greater in the

inflamed colonic mucosa of patients with CD than in

normal mucosa. In fetal intestinal tissue, the balance

of which resembles the immune balance in CD,

blockade of TNF-a decreased production of MMP-

3 [27]. In a study on IBD, inflammatory cytokines

such as TNF-a and IL-1b up-regulated MMP-3

production by monocytes which increased in number

in the submucosal layer [27]. TNF-a down-regulates

production of TIMP-1 that is a natural inhibitor of

MMP-3 [28]. Heuschkel et al. [19] described colonic

mucosal lesions in human IBD as showing MMP-3/

TIMP-1 imbalance. Baugh et al. [20], on the other

hand, showed that MMP-9 levels were elevated in

human IBD. However, our data might support the

hypothesis that MMP-3 is as important a protein as

TNF-a in treating TH-1-dominant intestinal inflam-

mation, although there is a difference between the

large intestine and small intestine.

In the present study, we used a TACE inhibitor,

GM6001, and two selective MMP inhibitors,

sMMP3i and MMP9i. GM6001 inhibited broad-

spectrum MMPs, MMP-1, -2, -3, -8, and -9, and

blocked the release of TNF-a [29]. Two selective

inhibitors have been shown to produce specificity

inhibition against, respectively with low concentra-

tions [30,31]. They might not have specific without

any cross-inhibition of MMP or TACE.

sMMP3i, 4-(4?-Biphenyl)-4-hydroxyaminobutyric

acid, is a potent and competitive inhibitor of MMP-

3, and no cross-inhibition has ever been reported in

rats. The functional groups, the oxime and carboxyl

groups, have been shown to interact with the active

site zinc of MMP. MMP9i, N-hydroxy-1-(4-

methoxyphenyl) sulfonyl-4-benzyloxycarbonylpiper-

azine-2-carboxamide, is a piperazine-based potent

inhibitor of MMP-9 (IC50�/1.9 nM) and MMP-13

(IC50�/1.3 nM). It inhibits MMP-1 and MMP-3 at

higher concentrations (IC50�/24 nM and 18 nM,

respectively). In this study, we ascertained that both

MMP3 production and TACE production were not

decreased significantly when high dosages of MMP9i

were used against experimental damage. These

results seem to confirm the evidence that MMP3

could be more effective than MMP9, TNF-a, and

TACE.

MMP-3 plays a central role in the MMP cascade

pathway and is a key enzyme that controls other

MMPs [7]. It is impossible to inhibit an MMP

enzyme alone with an MMP inhibitor, because the

MMP family has a cascade pathway that is activated

by any one of its members. MMP-3 suppression

could naturally cause other MMP suppression.

Therefore specificity against MMP-3 is lost in these

TACE/MMP inhibitors at high doses. In other

words, MMP-3 has powerful effects on the MMP

cascade pathway and could be an effective therapeu-

tic target against intestinal ulceration. This study

demonstrated that two effective inhibitors against

experimental injury, sMMP3i and a TACE inhibitor,

showed MMP-3 suppression. To study MMP-3

function, more specifically TACE/MMP production

needs to be evaluated using MMP-3 KO mice. In

addition, it is necessary to evaluate other MMP

production, especially that of MMP-7 and MMP-

12, as MMP-7 plays an immunodefensive role in the

colon, and MMP-12 is also associated with macro-

phages [7].

TACE has been of interest in the treatment of

inflammatory diseases characterized by TH-1-domi-

nant inflammation, where the key cytokine is TNF-a(examples: CD, rheumatoid arthritis) [4�6,28] be-

cause it is responsible for cleavage of transmem-

brane-domain-anchored TNF-a. Transmembrane

TNF, not soluble TNF, could be the critical form

of TNF for TNF signaling in IBD, because etaner-

cept, which binds only soluble TNF, was shown to

be ineffective in the treatment of moderate to severe

CD [32,33]. Transmembrane TNF, which was

mediated by TACE, was a potent inducer of the

TACE/MMP3 inhibitor improved experimental intestinal damage 1327

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

CD4CD45RBhi T-cell transfer model of colitis even

in the absence of its secreted form [34]. Another

TACE inhibitor (BB1101) also had a beneficial

effect in an experimental model of colitis [28,35].

In a clinical study of IBD, TACE activity was shown

to be up-regulated in inflamed colonic mucosal [36].

In our study, while the tested TACE inhibitor was

found to down-regulate TNF release as expected, it

was not anticipated that the inhibitor would also

down-regulate the production of TACE. There is a

possible explanation for this unexpected result;

down-regulation of TNF-a might cause down-reg-

ulation of the TACE protein. While Bzowska et al.

[37] reported that stimulation with TNF-a results in

increased levels of TACE mRNA and protein

expression, they showed that the increase in TACE

was accompanied by increased shedding of TNF

receptor in TNF-a-stimulated endothelial cells.

However, since we lack data on changes in TACE

expression being stimulated by TNF-a in intestinal

cells, future investigation is required.

Recently, Brynskov et al. [36] reported that

ulcerative colitis (TH-2-dominant inflammation)

showed a larger increase in TACE expression in

inflamed human colonic mucosa than that seen in

CD (TH-1-dominant inflammation). They tenta-

tively linked this unexpected result to inactivation

of TACE activity in CD mucosa catalyzed by other

MMPs as part of a self-regulatory mechanism.

Another possible explanation for these differences

in expression could be due to differences in the

production of TIMP-3, which inhibits TACE, in

inflammation. Only TIMP-3, as opposed to other

TIMPs, has been shown to inhibit TACE [38].

Overexpression of TIMP-3 has been reported in

ulcerations of the intestine [39]. Therefore, the

effects of the balance of TACE with that of TIMP-

3 and/or that of MMP-3 with TIMP-3 in inflamma-

tion and ulceration need to be investigated.

Pro-inflammatory cytokines could be involved in

the production of MMPs during the development of

intestinal injury. TNF-a is linked with NSAIDs, for

example Indo-induced intestinal injury in the rat

[14,16,40,41]. We previously reported that admin-

istration of anti-TNF-a, IL-6, and IL-1b neutraliz-

ing antibodies significantly ameliorated this

experimental enteropathy in a dose-dependent fash-

ion [14]. The maximum suppressive rates of UI by

these antibodies were respectively 43.0%, 49.6%,

and 59.7%. In addition, inhibition of enteropathy

with a combination of the three antibodies reached

up to 87.5%. In the present study, GM6001

completely abolished intestinal ulceration in parallel

with both 56% suppression of TNF-a and complete

suppression of MMP-3. Therapy with anti-TNF-ain patients with rheumatoid arthritis resulted in

decreased serum MMP-3 concentrations [42]. How-

ever, since we lack data concerning changes in

MMP-3 in CD patients treated with anti-TNF-aneutralizing antibodies or other therapies, future

investigation is required.

In summary, sMMP3i showed the same degree of

therapeutic effect as TACE inhibitor in Indo-

induced small intestinal damage in the rat, involv-

ing suppression of MMP-3. On the other hand,

MMP9i had no therapeutic effect on this entero-

pathy despite suppression of MMP-9. The TACE-

TNF-a-MMP-3/TIMP-1 network may be an

important inflammatory pathway in this experimen-

tal intestinal inflammation.

Acknowledgements

This study was a supported by a Grant-in-Aid (No.

13670574) for Scientific Research from the Ministry

of Education, Science and Culture, Japan. Further

funding was given through a gastrointestinal disease

grant from the Specially Selected Disease Project of

the Ministry of Health and Welfare of Japan, and a

Project Grant (No. 12-409) from Kawasaki Medical

School.

References

[1] Beutler BA. The role of tumor necrosis factor in health and

disease. J Rheumatol 1999;/Suppl 57:/16�21.

[2] Black RA, Rauch CT, Kozlosky CJ, Peschon JJ, Slack JL,

Wolfson MF, et al. A metalloproteinase disintegrin that

releases tumour-necrosis factor-alpha from cells. Nature

1997;/385:/729�33.

[3] Moss ML, Jin SL, Milla ME, Bickett DM, Burkhart W,

Carter HL, et al. Cloning of disintegrin metalloproteinase

that processes precursor tumour-necrosis factor-a. Nature

1997;/385:/733�5.

[4] Bazzoni F, Beutler B. The tumor necrosis factor ligand and

receptor families. N Engl J Med 1996;/334:/1717�25.

[5] van Deventer SJ. A place of TACE. Gut 2002;/51:/5�6.

[6] Shanahan F. Inflammatory bowel disease: immunodiagnos-

tics, immunotherapeutics, and ecotherapeutics. Gastroenter-

ology 2001;/120:/622�35.

[7] Park WC, Wilson CL, Lopez-Boado YS. Matrix metallo-

proteinases as modulators of inflammation and innate

immunity. Nat Rev Immunol 2004;/4:/617�29.

[8] Schuppan D, Hahn EG. Matrix metalloproteinases in the

gut: inflammation hits the matrix. Gut 2000;/47:/12�4.

[9] von Lampe B, Barthel B, Coupland SE, Riecken EO,

Rosewicz S. Differential expression of matrix metalloprotei-

nases and their tissue inhibitors in colon mucosa of patients

with inflammatory bowel disease. Gut 2000;/47:/63�73.

[10] Coussens LM, Fingleton B, Matrisian LM. Matrix metallo-

proteinase inhibitors and cancer: trials and tribulations.

Science 2002;/295:/2387�92.

[11] Matsumoto T, Iida M, Kuroki F, Hizawa K, Koga H,

Fujishima M. Effect of diet on experimentally induced

intestinal ulcer in rats: morphology and tissue leukotrienes.

Gut 1994;/35:/1058�63.

[12] Matsumoto T, Iida M, Nakamura S, Hizawa K, Kuroki F,

Fujishima M. Preventive effect of immunosuppressive agents

1328 H. Matsumoto et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.

against indomethacin-induced small intestinal ulcers in rats.

Dig Dis Sci 1994;/39:/787�95.

[13] Koga H, Aoyagi K, Matsumoto T, Iida M, Fujishima M.

Experimental enteropathy in athymic and euthymic rats:

synergistic role of lipopolysaccharide and indomethacin. Am

J Physiol (Gastrointest, Liver Physiol) 1999;/276:/G576�82.

[14] Tarumi K. Essential role of macrophages and macrophage-

related proinflammatory cytokines in indomethacin-induced

enteropathy in rats [English abstract in Japanese article].

Kawasaki Igakkaishi 2002;/28:/243�56.

[15] Matsumoto T, Iida M, Nakamura S, Kuroki F, Hizawa K,

Fujishima M. An animal model of longitudinal ulcer in the

small intestine induced by intracolonically administrated

indomethacin in rats. Gastroenterol Jpn 1993;/28:/10�7.

[16] Anthony A, Pounder RE, Dhillon AP, Wakefield AJ.

Similarities between ileal Crohn’s disease and indomethacin

experimental jejunal ulcers in the rat. Aliment Pharmacol

Ther 2000;/14:/241�5.

[17] Vilaseca J, Salas A, Guarner F, Rodriguez R, Malagelada JR.

Participation of thomboxane and other eicosanoid synthesis

in the course of experimental inflammatory colitis. Gastro-

enterology 1990;/98:/269�77.

[18] Lindsay J, Van Montfrans C, Brennan F, Van Devenrer S,

Drillenburg P, Hodgson H, et al. IL-10 gene therapy pre-

vents TNBS-induced colitis. Gene Ther 2002;/9:/1715�21.

[19] Heuschkel RB, MacDonald TT, Montelone G, Bajaj-Elliott

M, Smith JA, Pender SL. Imbalance of stromelysin-1 and

tissue inhibitor of matrix metalloproteinase (TIMP)-1 in the

mucosal lesions of children with inflammatory bowel disease.

Gut 2000;/47:/57�62.

[20] Baugh MD, Perry MJ, Hollander AP, Davies DR, Cross SS,

Lobo AJ, et al. Matrix metalloproteinase levels are elevated

in inflammatory bowel disease. Gastroenterology 1999;/117:/

814�22.

[21] Garcia Rodriguez LA. Variability in risk of gastrointestinal

complications with different nonsteroidal anti-inflammatory

drugs. Am J Med 1998;/104:/30S�4.

[22] Loftus EV Jr. Clinical epidemiology of inflammatory bowel

disease: incidence, prevalence, and environmental influ-

ences. Gastroenterology 2004;/126:/1504�17.

[23] Iddan G, Meron G, Glukhovsky A, Swain P. Wireless

capsule endoscopy. Nature 2000;/405:/417.

[24] Yamamoto H, Yano T, Kita H, Sunada K, Ido K, Sugano K.

New system of double-balloon enteroscopy for diagnosis and

treatment of small intestinal disorders. Gastroenterology

2003;/125:/1556�7.

[25] Kirkegaard T, Hansen A, Bruun E, Bryskov J. Expression

and localization of matrix metalloproteinases and their

natural inhibitors in fistulae of patients with Crohn’s disease.

Gut 2004;/53:/701�9.

[26] McKaig BC, McWilliam D, Watoson SA, Mahida YR.

Expression and regulation of tissue inhibitor of matrix

metalloproteinase-1 and matrix metalloproteinase by intest-

inal myofibroblasts in inflammatory bowel disease. Am J

Pathol 2003;/162:/1355�60.

[27] Pender SL, Tikle SP, Docherty AJ, Howie D, Wathen NC,

MacDonald TT. A major role of matrix metalloproteinases

in T cell injury in the gut. J Immunol 1997;/158:/1582�90.

[28] Colon AL, Menchen LA, Hurtado O, De Cristobal J,

Lizasoain I, Leza JC, et al. Implication of TNF-alpha

convertase (TACE/ADAM17) in inducible nitric oxide

synthase expression and inflammation in an experimental

model of colitis. Cytokine 2001;/16:/220�6.

[29] Galardy RE, Grobelny D, Foellmer HG, Fernandez LA.

Inhibition of angiogenesis by the matrix metalloprotease

inhibitor N-[2R-2-(hydroxamidocarbonymethyl)-4-methyl-

pentanoyl]]-L-tryptophan methylamide. Cancer Res 1994;/

54:/4715�8.

[30] Johnson LL, Bornemeier DA, Janowicz JA, Chen J, Pav-

lovsky AG, Ortwine DF. Effect of species differences on

stromelysin-1 (MMP-3) inhibitor potency. An explanation of

inhibitor selectivity using homology modeling and chimeric

proteins. J Biol Chem 1999;/274:/24881�7.

[31] Cheng M, De B, Almstead NG, Pikul S, Dowty ME,

Dietsch CR, et al. Design and synthesis of piperazine-based,

matrix metalloproteinase inhibitors. J Med Chem 1999;/42:/

2295�314.

[32] van Deventer SJ. Transmembrane TNF-alpha, induction of

apoptosis, and the efficacy of TNF-targeting therapies in

Crohn’s disease. Gastroenterology 2001;/121:/1242�6.

[33] Sandborn WJ, Hanauer SB, Katz S, Safdi M, Wolf DG,

Baerg RD, et al. Etanercept for active Crohn’s disease: a

randomized, double-blind, placebo-controlled trial. Gastro-

enterology 2001;/121:/1088�94.

[34] Corazza N, Brunner T, Buri C, Rihs S, Imboden MA,

Seibold I, et al. Transmembrane tumor necrosis factor is a

potent inducer of colitis even in the absence of its secreted

form. Gastroenterology 2004;/127:/816�25.

[35] Sykes AP, Bhogal R, Brampton C, Chander C, Whelan C,

Parsons ME, et al. The effect of an inhibitor of matrix

metalloproteinases on colonic inflammation in a trinitroben-

zenesulphonic acid rat model of inflammatory bowel disease.

Aliment Pharmacol Ther 1999;/13:/1535�42.

[36] Brynskov J, Foegh P, Pedersen G, Ellervik C, Kirkegaard T,

Bingham A, et al. Tumor necrosis factor-alpha converting

enzyme (TACE) activity in the colonic mucosa of patients

with inflammatory bowel disease. Gut 2002;/51:/37�43.

[37] Bzowska M, Jura N, Lassak A, Black RA, Bereta J. Tumour

necrosis factor-a stimulates expression of TNF-a converting

enzyme in endothelial cells. Eur J Biochem 2004;/271:/2808�20.

[38] Amour A, Slocombe PM, Webster A, Butler M, Knight CG,

Smith BJ. TNF-alpha converting enzyme (TACE) is inhib-

ited by TIMP-3. FEBS Lett 1998;/435:/39�44.

[39] Vaalamo M, Karjalainen-Lindberg ML, Saarialho-Kere U.

Distinct expression profiles of stromelysin-2 (MMP-10),

collagenase-3 (MMP-13), macrophage metalloelastase

(MMP-12), and tissue inhibitor of MMP-3 (TIMP-3) in

intestinal ulceration. Am J Pathol 1998;/152:/1005�14.

[40] Bertrand V, Guimbaud R, Tulliez M, Mauprivez C, Sogni P,

Couturier D, et al. Increase in tumor necrosis factor-aproduction linked to the toxicity of indomethacin for the

rat small intestine. Br J Pharmacol 1998;/124:/138594.

[41] Reuter BK, Wallace JL. Phosphodiesterase inhibitors prevent

NSAID enteropathy independently of effects on TNF-arelease. Am J Physiol (Gastrointest, Liver Physiol) 1999;/40:/

G847�54.

[42] Catrina AI, Lampa J, Ernestam S, af Klint E, Bratt J,

Klarekog L, et al. Anti-tumour necrosis factor (TNF)-alpha

therapy (etanercept) down-regulates serum matrix metallo-

proteinase (MMP)-3 and MMP-1 in rheumatoid arthritis.

Rheumatology 2002;/41:/484�9.

TACE/MMP3 inhibitor improved experimental intestinal damage 1329

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

Lib

rary

Utr

echt

on

04/0

8/13

For

pers

onal

use

onl

y.