Mechanisms of Diarrhea in Collagenous Colitis - Gastroenterology

GASTROENTEROLOGY 2002;123:433– 443

Mechanisms of Diarrhea in Collagenous Colitis ¨ RGEL,* CHRISTIAN BOJARSKI,* JOACHIM MANKERTZ,* MARTIN ZEITZ,* NATALIE BU ¨ RG–DIETER SCHULZKE* MICHAEL FROMM,‡ and JO *Departments of Gastroenterology, Infectious Diseases, and Rheumatology and ‡Clinical Physiology, Universita¨tsklinikum Benjamin Franklin, Freie Universita¨t Berlin, Berlin, Germany

Background & Aims: Collagenous colitis is an inflammatory disease of unknown etiology with diarrhea as the leading symptom. The aim of this study was to examine the pathogenic mechanisms of this disease. Methods: Biopsy specimens of the sigmoid colon were obtained endoscopically. Short-circuit current and 22Na and 36Cl fluxes were measured in miniaturized Ussing chambers. Alternating current impedance analysis discriminated epithelial from subepithelial resistance. Tight junction proteins occludin and claudin 1–5 were characterized in membrane fractions by Western blotting. Apoptotic ratio was determined by DAPI and TUNEL staining. Results: In collagenous colitis, net Naⴙ flux decreased from 8.8 ⴞ 1.8 to 0.2 ⴞ 1.5 and net Clⴚ flux from 11.2 ⴞ 3.0 to ⴚ3.0 ⴞ 2.7 ␮mol 䡠 hⴚ1 䡠 cmⴚ2, indicating a pronounced decrease in NaCl absorption. The fact that short-circuit current increased from 1.5 ⴞ 0.4 to 3.9 ⴞ 0.8 ␮mol 䡠 hⴚ1 䡠 cmⴚ2, together with the negative net Clⴚ flux, points to activation of active electrogenic chloride secretion. Subepithelial resistance increased from 7 ⴞ 1 to 18 ⴞ 2 ⍀ 䡠 cm2 due to subepithelial collagenous bands of 48 ⴞ 8 –␮m thickness. Epithelial resistance was diminished from 44 ⴞ 3 to 29 ⴞ 2 ⍀ 䡠 cm2, and this was accompanied by a decrease in occludin and claudin-4 expression. Neither mucosal surface area nor apoptotic ratio was altered in collagenous colitis. Conclusions: Reduced net Naⴙ and Clⴚ absorption is the predominant diarrheal mechanism in collagenous colitis, accompanied by a secretory component of active electrogenic chloride secretion. The subepithelial collagenous band as a significant diffusion barrier is a cofactor. Downregulation of tight junction molecules but not epithelial apoptoses is a structural correlate of barrier dysfunction contributing to diarrhea by a leak flux mechanism.

ollagenous colitis is an inflammatory bowel disease that is clinically characterized by watery diarrhea. It was originally described by Lindstrøm in 1976.1 The mucosa appears normal on colonoscopy, and diagnosis depends entirely on histologic identification of pathognomonic subepithelial collagenous bands that occur in combination with an inflammatory cell infiltrate.

C

The etiology of collagenous colitis is unknown thus far, although several hypotheses exist. From the beneficial effect of long-term small intestinal diversion, it was hypothesized that luminal agents damage the colonic epithelium or initiate inflammation, which, as a reaction or repair mechanism, leads to increased collagen biosynthesis in pericrypt fibroblasts.2– 4 Furthermore, it has been considered that nonsteroidal anti-inflammatory drugs contribute to its onset.5,6 Little is known about the mechanisms of diarrhea in collagenous colitis. Lindstrøm hypothesized that the subepithelial collagenous layers contribute to disturbed colonic electrolyte and water absorption.1 Other reports seem to support this view by showing that stool frequency depends on the thickness of the collagenous band,7–9 although not all subsequent studies could reproduce this finding.10 In general, except for motility-driven diarrhea, all other forms of diarrhea are driven by osmotic forces, including malabsorption, active secretion, and leak flux mechanisms (see Discussion). The Na/H antiporter 3 as a component of active electroneutral NaCl absorption is quantitatively one of the most important absorptive mechanisms in the colon. Therefore, we measured unidirectional 22Na and 36Cl fluxes in the Ussing chamber and quantified active anion secretion in collagenous colitis. To estimate epithelial barrier function, we concentrated on tight junction morphology and apoptoses in the colonic epithelium, because grossly evident epithelial lesions (e.g., ulcers and erosions) are not a typical feature of collagenous colitis. Recently, the transmembrane protein occludin and several members of the claudin transmembrane protein family were identified as strand-formAbbreviations used in this paper: DAPI, 4ⴕ,6-diamidino-2-phenylindole; Isc, short-circuit current; Re, epithelial electrical wall resistance; Rsub, subepithelial electrical wall resistance; Rt, transmural electrical wall resistance; TUNEL, terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling. © 2002 by the American Gastroenterological Association 0016-5085/02/$35.00 doi:10.1053/gast.2002.34784

434

¨ RGEL ET AL. BU

GASTROENTEROLOGY Vol. 123, No. 3

Table 1. Clinical Data of Patients With Collagenous Colitis

Patient

Sex/age (yr)

Duration (yr)

1 2

F/47 M/39

1 2

3 4 5 6 7 8

F/51 F/57 F/50 F/53 M/47 F/51

1 17 2 1 27 2

Other diagnosis — Pulmonary fibrosis. ITP Asthma Schizophrenia — — COLD —

Stools per day

Medication

Thickness of collagenous band (␮m)

8 6

— Prednisolone 7.5 mg once a day

21 87

5 2 4 3 8 5

Azulfidine 1 g 3 times a day Sulpiride 100 mg twice a day — Budesonide 3 mg once a day Amitryptiline 75 mg twice a day Budesonide 3 mg 3 times a day

45 37 36 63 64 28

NOTE. Grade of diarrhea is given as mean frequency of (loose) stools per day during the last week before the biopsy was performed. Patients fasted overnight, and continuous medication was discontinued on the day of the biopsy. In addition to continuous medication, loperamide (up to 4 mg/day) was used by patients 3, 4, and 7 if required (except for the day of the biopsy). Lymphocytic infiltrate in the lamina propria was rather homogeneous in patients with collagenous colitis with moderate to intermediate density, but the thickness of the collagenous band varied considerably from patient to patient. ITP, idiopathic thrombocytopenic purpura; COLD, chronic obstructive lung disease.

ing components of the epithelial tight junction. They are arranged in the junctional complex with other extrinsic tight junction molecules as, for example, are ZO-1 and ZO-2, which anchor occludin to the cytoskeleton (for review, see Anderson and Van Itallie11 and Tsukita and Furuse12). Tight junction permeability has been shown to undergo regulation by a variety of factors, including proinflammatory cytokines13,14 (see Discussion), which may be important mediators in intestinal inflammation. Therefore, the present study for the first time combined electrophysiologic and molecular biological techniques to identify pathomechanisms involved in barrier and transport dysfunction in collagenous colitis.

Solutions and Drugs The bathing solution of flux experiments contained the following (in mmol/L): Na⫹, 140; Cl⫺, 123.8; K⫹, 5.4; Ca2⫹, 1.2; Mg2⫹, 1.2; HPO42⫺, 2.4; H2PO4⫺, 0.6; HCO3⫺, 21, D(⫹)-glucose, 10; ␤-OH-butyrate, 0.5; glutamine, 2.5; and D(⫹)-mannose, 10.17 Amiloride (10 ␮mol/L) was present on both sides of the Ussing chamber throughout the experiments. The solution was gassed with 95% O2 and 5% CO2. The temperature was maintained at 37°C using water-jacketed reservoirs, and pH was 7.4 in all experiments. Antibiotics (50 mg/L azlocillin and 4 mg/L tobramycin) served to prevent bacterial growth and had no effect on short-circuit current (Isc) in the concentration used.18

Isc and Tracer Flux Measurements

Materials and Methods Tissue Preparation Measurements were performed on colonic specimens from 8 patients with histologically confirmed collagenous colitis and from 7 control patients undergoing colonoscopy for polyp surveillance or anemia of uncertain etiology. Clinical data are noted in Table 1. The biopsy specimens were obtained endoscopically from the sigmoid colon at 30 cm ab ano with a 3.9-mm forceps. The mucosa of all patients appeared macroscopically normal. The specimens were spread out under a dissection microscope. A perforated plastic disc with an inner diameter of 2.5 mm was glued on the serosal side of the biopsy specimen using histoacryl tissue glue (B. Braun, Melsungen, Germany). Subsequently, this disc was placed in a microcontainer tightened with silicon rubber seals and mounted in Ussing-type chambers as previously described.15,16 Tissues were transported to the laboratory on ice in oxygenated bathing solution. The time between taking the biopsy specimens and mounting the tissues into the Ussing chambers was about 20 –30 minutes.

Ussing-type experiments were performed as described previously15,18,19 using a computer-controlled voltage clamp device (CVC 6; Fiebig, Berlin, Germany). Epithelial transport was determined by measurements of unidirectional 22Na and 36Cl fluxes (Du Pont, Wilmington, DE). These experiments were performed under short-circuit conditions.19 Isc, opencircuit transepithelial voltage, and transepithelial resistance were recorded to hard disk. Isc values were corrected for bath resistance as described by Tai and Tai.20 Data were given as ␮mol 䡠 h⫺1 䡠 cm⫺2. The exposed tissue area was 0.049 cm2. The tissues were matched for conductance for calculating net fluxes.

Alternating Current Impedance Analysis Alternating current impedance analysis was performed as previously described.19,21–23 This technique can differentiate the epithelial (Re) and subepithelial (Rsub) portion of the transmural wall resistance (Rt) based on the 3-parameter model of the colonic wall.23 In this model, the epithelium is described as an electrical equivalent circuit by a resistor and a capacitor in parallel and the subepithelium by a resistor in series. After application of 48 discrete frequencies of an effec-

August 2002

tive sine-wave alternating current of 35 ␮A 䡠 cm⫺2, ranging from 1.3 Hz to 65 kHz, changes in tissue voltage were detected by phase-sensitive amplifiers (1250 frequency response analyzer and 1286 electrochemical interface; Solartron Schlumberger, Farnborough, Hampshire, England). Complex impedance values were calculated and corrected for the resistance of the bathing solution and the frequency behavior of the measuring device. Then, for each tissue, the impedance locus was plotted in a Nyquist diagram and a circle segment was fitted by least square analysis. Because of the frequency-dependent electrical characteristics of the capacitor, Rt is obtained at low frequencies and Rsub is obtained at high frequencies. Re was obtained from the following: Re ⫽ Rt ⫺ Rsub.

Histology After electrophysiologic measurements in the Ussing chamber, biopsy specimens were cut off the insert of the chamber, fixed in 4% formalin, and embedded in paraffin for morphologic analysis. Serial sections (3 ␮m) were stained with H&E or dewaxed for tenascin (immunostaining) or immunofluorescence detection of epithelial apoptoses. Cellular DNA was either stained with 4⬘,6-diamidino-2-phenylindole (DAPI)24 or terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL) assay (Roche, Mannheim, Germany). In the latter, blunt ends of double-stranded DNA exposed by strand breaks were visualized by means of enzymatic labeling of the free 3⬘-OH termini with fluorescein/deoxyuridine triphosphate. Epithelial apoptosis was determined as the percentage of apoptotic nuclei per visual field (apoptotic ratio).

Mucosal Morphometry Morphometry of the mucosal surface area was performed with slight modification of a previously described protocol.19 Exposed tissues from the sigmoid colon were immediately fixed in 10% buffered formalin after the electrophysiologic experiment. For counting the number of crypts per serosal area, light microscopy was performed in the unstained specimens using a digital camera. Identical specimens were then embedded in paraffin for at least 24 hours and stained with H&E. Crypt length and inner crypt diameter were measured in cross sections. Additionally, for evaluation of the thickness of collagenous bands, the immunohistologic detection of tenascin was performed as previously described.25

Western Blot Analysis To determine tight junction protein expression, Western blot analysis was performed from membrane extracts of colonic biopsy specimens. Tissues were homogenized (by douncing) in iced lysate buffer containing 20 mmol/L Tris, pH 7.4, 5 mmol/L MgCl2, 1 mmol/L EDTA, 0.3 mmol/L ethylene glycol-bis(␤-aminoethyl ether)-N,N,N⬘,N⬘-tetraacetic acid, 1 ␮L/mL aprotinin, 16 ␮g/mL benzamidine HCl, 10 ␮g/mL phenanthroline, 10 ␮g/mL leupeptin, 10 ␮g/mL pepstatin, 1 mmol phenylmethylsulfonyl fluoride, 210 ␮g/mL sodium fluoride, 2.16 mg/mL ␤-glycerophosphate, 18.5 ␮g/mL NaVO4 ,

PATHOGENIC MECHANISMS IN COLLAGENOUS COLITIS

435

and 1 ␮L/mL trypsin inhibitor (all substances obtained from Sigma Chemical Co., St. Louis, MO). Membrane fractions were obtained by freeze-thaw cycles and subsequent passage through a 26 G ⫻ 1⁄2-inch needle. To remove insoluble material, the extract was centrifuged at 200g for 5 minutes at 4°C. The supernatant was then centrifuged at 43,000g for 30 minutes at 4°C. The pellet representing a crude membrane fraction was resuspended in lysate buffer. Protein concentrations were determined by Pierce BCA assay (Pierce, Rockford, IL). Aliquots of 5 ␮g protein were separated by polyacrylamide gel electrophoresis (8.5% for occludin and 12.5% for claudins) and transferred to a polyscreen polyvinylidene difluoride transfer membrane (NEN Life Science Products, Boston, MA). Blots were blocked for 2 hours in 5% milk powder and then overnight in 5% bovine serum albumin (at 4°C) before incubation with primary rabbit polyclonal immunoglobulin G antibodies against claudin 1–5 and occludin. Peroxidase-conjugated goat anti-rabbit immunoglobulin G antibodies and the chemiluminescence detection system Lumi-LightPLUS Western blotting kit (Roche, Mannheim, Germany) were used to detect bound antibodies. Antibodies were provided by Zymed Laboratories (South San Francisco, CA). Synthesis and test of specificity of these antibodies were recently described by Rahner et al.26 Chemiluminescence signals were detected using an LAS-1000 imaging system (Fuji, Tokyo, Japan) and analyzed with the AIDA program package (Raytest, Berlin, Germany).

Statistical Analysis Results are given as means ⫾ SEM. Significance was tested by the 2-tailed Student t test. P ⬍ 0.05 was considered significant.

Results Alternating Current Impedance Analysis Typical impedance locus plots of sigmoid colon from patients with collagenous colitis and controls are shown in Figure 1. Statistical data are presented in Table 2. In controls, Re and Rsub were 44 ⫾ 3 and 7 ⫾ 1 ⍀ 䡠 cm2, respectively. Thus, Re contributed 84% to Rt. At first, a value of 52 ⫾ 3 ⍀ 䡠 cm2 seems to be rather low for the Re of human distal colon, even for biopsy specimens in which only a small part of the subepithelial tissue is preserved; this is especially true when compared with a value of 95 ⫾ 5 ⍀ 䡠 cm2 (n ⫽ 10), which was measured with the same technique in one of our recent studies.27 However, it is important to keep in mind that biopsy specimens are mounted into the Ussing chamber without stretch. Indeed, a systematic study from our group to standardize the miniaturized Ussing chamber showed a significant stretch dependence of the apparent Re, making up a factor of 2.16 This is square of a linear stretch factor of 1.4, which is usually applied in our

436

¨ RGEL ET AL. BU


Figure 1. Original impedance locus plots of human sigmoid colon. (A) Control; (B) collagenous colitis. Zreal gives the ohmic component and Zimaginary the reactive component of the complex impedance. Intersections between the semicircle and x axis at low and high frequencies represent Rt and Rsub, respectively. Rt minus Rsub equals the Re. For a detailed explanation of transepithelial impedance data, see references 21–23.

laboratory when surgical tissue specimens are inserted into the Ussing chamber. Using the same measuring protocol for specimens from controls and patients with collagenous colitis, Re was decreased by 32% to 29 ⫾ 2 ⍀ 䡠 cm2. In parallel, Rsub was increased from 7 ⫾ 1 to 18 ⫾ 2 ⍀ 䡠 cm2 because of the appearance of subepithelial collagenous layers in patients with collagenous colitis. Another important aspect from this analysis is awareness of the correction of active transport rates for Rsub contributions. Whenever significant nonepithelial series resistances are present between the voltage-sensing electrodes in addition to the Re, measured Isc or net fluxes have to be corrected for the contribution of these resistances. This correction is well accepted when applied to bath resistance, but it is also necessary to correct Isc and net fluxes for the Rsub of intestinal preparations. The implications of this correction have been given elsewhere in detail.19 –23 Generally, after correction for the bathing solution, the true active transport rate becomes underestimated by a factor that is given by the ratio of Rt over Re.20,21 This factor was 1.6 ⫾ 0.06 in patients with collagenous colitis and 1.2 ⫾ 0.04 in controls (Table 2). Because these factors significantly differed by about 30%, Isc values and net fluxes of both groups were

compared only after performing this correction (compare rough and corrected data for active transport in Table 3). Naⴙ and Clⴚ Fluxes The data on Na⫹ and Cl⫺ fluxes are given in Table 3. Sigmoid colon specimens from patients with collagenous colitis showed no significant change in Rt when compared with controls (48 ⫾ 6 vs. 47 ⫾ 4 ⍀ 䡠 cm2), because the decrease in Re was compensated for by the increase in Rsub. In controls, significant absorptive net fluxes were present for both sodium and chloride (JNanet was 7.4 ⫾ 1.5 ␮mol 䡠 h⫺1 䡠 cm⫺2 and JClnet was 8.9 ⫾ 2.5 ␮mol 䡠 h⫺1 䡠 cm⫺2). When corrected for Rsub layers between the voltage-sensing electrodes (compare with above), JNanet was 8.8 ⫾ 1.8 ␮mol 䡠 h⫺1 䡠 cm⫺2 and JClnet was 11.2 ⫾ 3.0 ␮mol 䡠 h⫺1 䡠 cm⫺2. In patients with collagenous colitis, net Na⫹ and Cl⫺ fluxes were decreased compared with control (JNanet was 0.2 ⫾ 1.5 ␮mol 䡠 h⫺1 䡠 cm⫺2 and JClnet was ⫺3.0 ⫾ 2.7 ␮mol 䡠 h⫺1 䡠 cm⫺2), indicating impaired absorption of sodium and chloride ions. Isc was 1.5 ⫾ 0.4 ␮mol 䡠 h⫺1 䡠 cm⫺2 in sigmoid colon specimens from controls and increased to 3.9 ⫾ 0.8 ␮mol 䡠 h⫺1 䡠 cm⫺2 in specimens from patients with

Table 2. Re and Rsub Contributions to Rt in Collagenous Colitis

Control Collagenous colitis P

Rt

Re

Rsub

Rt/Re

n

52 ⫾ 3⍀cm2 47 ⫾ 4⍀cm2 NS

44 ⫾ 3⍀cm2 29 ⫾ 2⍀cm2 ⬍0.01

7 ⫾ 1⍀cm2 18 ⫾ 2⍀cm2 ⬍0.001

1.2 ⫾ 0.04 1.6 ⫾ 0.06 ⬍0.001

7 8

NOTE. Re and Rsub were obtained as described in Patients and Methods. Rt ⫽ Re ⫹ Rsub. Rt/Re is the correction factor of active transport rates (Isc or net fluxes) for subepithelial resistance contributions (see Results). As indicated in Table 1, 2 patients received oral budesonide 3 mg once daily and 3 mg 3 times a day, doses that could be therapeutically effective. However, statistical evaluation without these 2 patients was not significantly different (Rt ⫽ 46 ⫾ 4, Re ⫽ 28 ⫾ 3, Rsub ⫽ 18 ⫾ 2 ⍀ 䡠 cm2, n ⫽ 6). All values represent means ⫾ SEM.

August 2002


437

Table 3. Na⫹ and Cl⫺ Fluxes in Collagenous Colitis Raw data

R

Isc

JNanet

JNams

JNasm

JClnet

JClms

JClsm

n

Control Collagenous colitis P Corrected data Control Collagenous colitis P

47 ⫾ 4 48 ⫾ 6 NS

1.3 ⫾ 0.4 2.5 ⫾ 0.5 NS

7.4 ⫾ 1.5 0.1 ⫾ 0.9 ⬍0.01

14.3 ⫾ 1.6 7.7 ⫾ 1.1 ⬍0.01

7.0 ⫾ 0.7 7.6 ⫾ 1.6 NS

8.9 ⫾ 2.5 ⫺1.9 ⫾ 1.7 ⬍0.02

33.4 ⫾ 1.5 11.2 ⫾ 1.2 ⬍0.001

24.4 ⫾ 1.4 13.0 ⫾ 2.0 ⬍0.01

6 7

1.5 ⫾ 0.4 3.9 ⫾ 0.8 ⬍0.05

8.8 ⫾ 1.8 0.2 ⫾ 1.5 ⬍0.01

11.2 ⫾ 3.0 ⫺3.0 ⫾ 2.7 ⬍0.01

6 7

NOTE. Tissue resistance is expressed as ⍀ 䡠 cm2, and Isc as well as Na⫹ and Cl⫺ fluxes are expressed as ␮mol 䡠 h⫺1 䡠 cm⫺2. Isc and net fluxes are given before (raw data) and after (corrected data) correction for subepithelial resistance contributions between the voltage-sensing electrodes as described in Results. All values represent means ⫾ SEM. R, tissue resistance; J, flux; ms, from mucosa to serosa; sm, from serosa to mucosa.

collagenous colitis. This increase was paralleled by a change in Cl⫺ net flux from absorption toward secretion. Epithelial Apoptoses Apoptotic ratio was determined in epithelium from controls and patients with collagenous colitis using TUNEL- and DAPI-stained thin sections. Typical apoptotic changes associated with epithelial apoptosis comprise condensation of chromatin, its compaction along the periphery of the nucleus, and segmentation of the nucleus (Figure 2C and D). In control tissues, apoptotic ratio amounted to 1.2% ⫾ 0.3% (n ⫽ 6; mean of DAPI and TUNEL staining in Table 4). In collagenous colitis tissues, apoptotic ratio was 1.2% ⫾ 0.3% (n ⫽ 6) and not significantly different from control. Mucosal Morphometry Mucosal architecture was not much altered in collagenous colitis compared with normal histology. There was no crypt rarefaction, as is usually observed in ulcerative colitis (Table 5). Epithelial gross lesions such as erosions were also absent in collagenous colitis. Most prominent, subepithelial collagenous bands appeared in collagenous colitis (Figure 3C vs. A) that can be visualized by tenascin staining, an extracellular matrix glycoprotein that is expressed during stromal remodeling (Figure 3D vs. B). Furthermore, an inflammatory cell infiltrate was present in lamina propria and intraepithelially, but its composition was not further investigated in the present study. Table 4. Epithelial Apoptoses in Collagenous Colitis


TUNEL

DAPI

n

1.2% ⫾ 0.3% 1.1% ⫾ 0.3% NS

1.1% ⫾ 0.2% 1.3% ⫾ 0.2% NS

6 6

NOTE. Apoptosis was determined as the percentage of apoptotic nuclei per visual field.

In collagenous colitis, the thickness of the subepithelial collagenous band amounted to 48 ⫾ 8 ␮m (n ⫽ 8; range, 21– 87 ␮m; Table 1). Thus, a 1 ␮m collagenous band contributed 0.26 ⍀ 䡠 cm2 to the Rsub, as had been calculated from the increase in thickness of the collagenous bands from 6 ␮m in controls to 48 ␮m in collagenous colitis and the increase in Rsub from 7 to 18 ⍀ 䡠 cm2. This corresponds to a specific volume resistance of 2619 ⍀ 䡠 cm of the subepithelial collagenous band. Tight Junction Protein Expression The data on tight junction protein expression are given in Figure 4. To further investigate whether altered tight junction structure contributes to the decreased Re in collagenous colitis, immunoblot analyses were performed for the tight junction proteins claudin 1–5 and occludin. Because these proteins are integral membrane proteins forming tight junction strands, crude membrane fractions were used for these investigations. In Western blot analysis and subsequent densitometry, occludin and claudin 4 were decreased and claudin 1, 3, and 5 remained unchanged. In contrast to the other 4 claudins investigated, claudin-2 expression (a pore-forming tight junction protein) was not stable or decreased but elevated in 4 of 5 patients (100%, 122%, 272%, 408%, and 717% of the control group). In addition, the patient with the lowest Re (18 ⍀ 䡠 cm2) showed the highest claudin-2 expression (717%) of the control group. However, because of the pronounced variation of the expression levels, this increase to 324% ⫾ 113% of the control value did not reach statistical significance, and claudin-2 expression did not significantly correlate with Re (Figure 5C). If up-regulation or down-regulation of tight junction molecules is considered, it is important to keep in mind that quantification is performed in reference to the protein content of crude membrane fractions in both groups. Thus, each quantitative comparison has to be drawn with

438

¨ RGEL ET AL. BU


Figure 2. Apoptoses in collagenous colitis. Apoptosis in a thin section of sigmoid colon is shown. Condensed chromatin fragments in nuclei and segmentation of the nuclei indicated epithelial apoptoses (arrows). (A and B) TUNEL and DAPI staining in the sigmoid colon of a control patient. (C and D) TUNEL and DAPI staining in the sigmoid colon of a patient with collagenous colitis (original magnification 20⫻).

Figure 3. Histology of collagenous colitis. Shown is a thin section of a biopsy specimen endoscopically obtained from the sigmoid colon of (A and B) a control and (C and D) a patient with collagenous colitis stained with (A and C) H&E and (B and D) tenascin (original magnification 20⫻).

August 2002


439

Table 5. Mucosal Architecture of Collagenous Colitis


Crypt length (␮m)

Inner crypt diameter (␮m)

Crypts per serosal area (mm⫺2)

Mucosal area per serosal area (mm2 mm⫺2)

304 ⫾ 17 300 ⫾ 17 NS

17 ⫾ 2 16 ⫾ 2 NS

70 ⫾ 7 71 ⫾ 6 NS

1.2 ⫾ 0.1 1.2 ⫾ 0.2 NS

n 5 7

NOTE. All values represent means ⫾ SEM.

caution. Nevertheless, because mucosal surface area parameters were not significantly altered in collagenous colitis and because of the differential changes in tight junction proteins in collagenous colitis, it may be reasonable to conclude that significant tight junction protein composition changes detected in the present study contributed to the decrease in epithelial tightness in the inflamed colonic mucosa. Correlation of Clinical, Functional, and Morphologic Data in Collagenous Colitis To yield further evidence of their importance, functional parameters were correlated to clinical and

morphologic data. There was a highly significant inverse correlation of epithelial barrier function and the degree of the diarrhea in patients with collagenous colitis (Figure 5A). Furthermore, Rsub correlated with the thickness of the subepithelial collagenous band, which further supports the functional relevance of this band in collagenous colitis (Figure 5B). However, the inverse correlation of Re with claudin-2 expression was rather weak and did not reach statistical significance, which does not indicate a prominent role of claudin 2 for the epithelial barrier defect in collagenous colitis (Figure 5C).

Discussion The aim of the present study was to investigate pathomechanisms of disturbed electrolyte and water transport in collagenous colitis and identify factors of etiologic importance. Sodium and Chloride Transport in Collagenous Colitis

Figure 4. Tight junction proteins occludin and claudin 1–5. Expression of tight junction proteins in crude membrane fractions by immunoblotting (samples with 5 ␮g protein). Each lane represents one biopsy. Lanes 1–3 represent samples from the sigmoid colon of controls, and lanes 4 – 6 represent samples from the sigmoid colon of patients with collagenous colitis. A second immunoblot was run on 2 further patients and 2 further controls (data not shown). Statistical data of densitometric analysis represent protein expression of the samples from 5 different patients with collagenous colitis (in percentage of the controls).

As the first main finding of the present study, the microscopically inflamed colon in collagenous colitis showed a marked decrease in absorptive net fluxes of sodium and chloride. The epithelium almost lost its ability to absorb these ions. The molecular mechanisms of this transport have not been further characterized in the present study. However, there is evidence from the literature for expression of the sodium/proton exchanger 3 (NaHE3) and the chloride/bicarbonate antiporter DRA (down-regulated in adenoma) in human distal colon,28,29 which points to electroneutral NaCl absorption as the result of sodium/ proton exchange and chloride/bicarbonate exchange acting in concert. Furthermore, Isc was elevated and there was a shift of the chloride net flux from absorption toward secretion in collagenous colitis, which points to activation of active electrogenic chloride secretion. This resembles the observation of Rask-Madsen et al.30 Even though their report referred to only a single colonic perfusion measurement, the investigators concluded from a chloride net movement against the electrochemical gradient and an increase in (lumen negative) transepithelial potential difference on active anion secretion in

440

¨ RGEL ET AL. BU


collagenous colitis. This does not completely rule out that the increase in potential difference was induced by electrogenic sodium transport as a result of hyperaldosteronism and that the net chloride movement could have been due to a significant epithelial barrier defect in the study, but the data and interpretation fit to the findings of our present study (which were obtained in the presence of 10 ␮mol/L amiloride). Thus, based on the patient in the study by Rask-Madsen et al. and our study group, the epithelium in collagenous colitis does not only lack absorption but even slightly secretes. The fact that Isc (most likely due to rheogenic chloride secretion) was activated in collagenous colitis differs markedly from the situation in ulcerative colitis, even if this secretory component was only moderate. There, voltage clamp and flux measurements in the Ussing chamber did not show any evidence for the induction of rheogenic chloride secretion,27,31 a possible explanation of which could be the higher degree of inflammation with a more pronounced barrier defect in ulcerative colitis because barrier integrity is an important prerequisite of active transport (see below). Epithelial Barrier Function in Collagenous Colitis

Figure 5. Correlation of clinical, functional, and morphologic data in collagenous colitis. (A) Inverse correlation of Re obtained by alternating impedance analysis and the degree of the diarrhea as given by the frequency of (loose) stools per day. (B) Correlation of Rsub obtained by alternating impedance analysis and the thickness of the subepithelial collagenous band. (C) No significant correlation of Re obtained by alternating impedance analysis and claudin-2 expression obtained by densitometry from Western blots could be detected (compare with Figure 4).

The epithelium represents the main barrier between the intestinal lumen and the vascular compartment, because capillaries with their highly permeable endothelium32 are located in near proximity to the enterocytes. However, this is quite different in collagenous colitis, in which collagenous bands are localized directly below the epithelium. To assess the epithelial barrier function in the presence of subepithelial tissue layers and to quantify the resistance of this collagenous band, alternating current impedance analysis was applied. As a second main result of the present study, Re was decreased by 30% in collagenous colitis. Furthermore, this decrease in Re correlated with the degree of the diarrhea (Figure 5A), which most likely reflects that both parameters depend on the degree of the inflammation. Thus, a significant barrier defect exists in collagenous colitis, although it was much less pronounced than that in ulcerative colitis. However, the degree of this barrier defect is comparable to that previously measured with the same technique during Yersinia enterocolitica infection, in which this barrier dysfunction was assumed to represent the predominant pathomechanism.33 Epithelial Apoptoses in Collagenous Colitis In searching for the structural correlate of this epithelial barrier defect in collagenous colitis, epithelial

August 2002

apoptoses were studied. Whether or not epithelial apoptoses impair the integrity of the intestinal barrier has been controversially discussed. Elimination of cells from the normal epithelium is followed by a physiologic rearrangement of tight junctions with maintenance of the macromolecular barrier. Therefore, apoptosis of isolated epithelial cells was assumed to occur without relevant disruption of epithelial integrity.34 However, recent direct measurements of epithelial ion permeability have clearly shown local leaks at the site of apoptosis.35,36 Therefore, apoptotic ratio was studied in addition to tight junction protein composition. Apoptotic ratio was 1.2% in the sigmoid colon of controls, which is within the range of 1.4% described by other groups in normal crypts of mice37 and in human jejunal mucosa.38 Thus, as a third important finding, apoptotic ratio was not increased in collagenous colitis (1.2%). Again, this is in sharp contrast to ulcerative colitis, in which Stra¨ter et al. have described significant up-regulation of epithelial apoptoses.39 Tight Junction Protein Expression in Collagenous Colitis Another possible structural correlate for the decrease in Re in collagenous colitis is a change in tight junction strand-forming molecules. These include occludin40 and several claudins, products of a multigene family with no sequence homology to occludin.41 The exact physiologic role of these integral membrane proteins, however, is far from clear. Very recently, for example, even a pore-forming property was identified for one of these proteins. Claudin-2 transfection resulted in a conversion of MDCK cells from high toward low resistance phenotype.42 A lot of data seem to support a role of occludin for the barrier and fence functions of tight junctions.43– 46 For example Wan et al. showed cleavage of occludin with consecutive barrier breakdown in confluent airway epithelial cells by Der p 1.46 However, other studies indicate that occludin is not a conditio sine qua non for barrier formation.41 For example, intestinal barrier function was not affected in occludin-deficient mice,47 a finding that can only be explained by substitutional redundancy. This means that functional tight junction down-regulation during inflammation or by cytokines requires not only reduced expression of a single tight junction protein (e.g., occludin) but at the same time expression arrest of the other fencing tight junction proteins to prevent their substitutional overexpression. As far as claudins are concerned, a functional role at least of claudin 1 seems to be possible. MDCK cells transfected with claudin 1 exhibited a 4 times higher


441

transepithelial resistance than wild-type MDCK cells.48 Also, claudin 4 was recently shown to have fencing properties.49 Taken together, tight junction strands seem to be composed as heteropolymers of different tight junction proteins that can in part functionally substitute each other. Therefore, we decided to incorporate not only occludin into our analysis but also other tight junction proteins such as claudin 1, claudin 3, claudin 4, and claudin 5 as well as the pore-forming tight junction protein claudin 2. As the fourth important finding of the present study, the decrease in Re in collagenous colitis was accompanied by diminished expression of occludin and claudin 4, while claudin-1, claudin-3, and claudin-5 expression remained unaltered. In addition, there was a tendency for the expression of the pore-forming tight junction molecule claudin 2 to be elevated that, due to a pronounced diversity in expression, failed to reach statistical significance. However, this is further direct support for the interpretation of the functional role of these molecules; on the other hand, it represents the structural correlate for the epithelial barrier defect in collagenous colitis. Mechanisms of the Tight Junctional Barrier Defect Several factors can alter tight junction permeability (e.g., cAMP,50 cytokines,13,14 complement components,51 and hormones52), factors that could also be mediators in collagenous colitis. However, most important in this respect may be that cytokines such as tumor necrosis factor ␣ and interferon gamma can alter tight junction permeability in both cell culture models14 and native intestinal epithelia,53 making them prominent candidates for the induction of this barrier disturbance in collagenous colitis. Mucosal Surface Area and Subepithelial Collagenous Bands Resistance values and transport rates are usually related to the serosal surface area as defined by the opening of the Ussing chamber. However, both depend on the mucosal surface area. Chronic inflammation can lead to alterations in mucosal surface geometry (e.g., due to crypt rarefaction as seen in ulcerative colitis). However, in collagenous colitis, mucosal surface area did not change. Thus, epithelial barrier dysfunction was not due to altered mucosal surface area but may preferentially reflect changes in tight junction structure. As the fifth important point in the present study, the functional contribution of the collagenous band below the epithelium could be quantified for the first time. It turned out that it has a specific volume resistance of 2619

442

¨ RGEL ET AL. BU

⍀ 䡠 cm. For comparison, we measured a specific volume resistance of 980 ⍀ 䡠 cm for noninflamed subepithelial tissue obtained from a 184-␮m reduction in subepithelial tissue layers by mechanical stripping and the concomitant 18 ⍀ 䡠 cm2 decrease in Rsub in rat large intestine.54 This means that the subepithelial band with a mean thickness of 48 ␮m in collagenous colitis represents a significant subepithelial barrier in front of the capillaries, which makes up about one third of the resistance of the epithelium. Thus, absorbed ions would leak back to the intestinal lumen by a 1:3 ratio, reducing absorptive efficiency by 25%. So far, mechanisms responsible for the appearance of collagenous bands are only poorly characterized. Gunther et al. identified collagen 1 and collagen 3 in the subepithelial collagenous band and showed reduced expression of metalloproteinase 1 and up-regulation of tissue inhibitor of metalloproteinase 1 in collagenous colitis.55 As far as expression of tissue inhibitor of metalloproteinase 1 is concerned, its gene expression is stimulated by interleukin 656 and interleukin 1057 activated in reporter gene assays using the functional promoter. Both are therefore potential candidates for this regulation in collagenous colitis. Diarrheal Mechanisms in Collagenous Colitis According to the results of the present study, diarrheal mechanisms in collagenous colitis include (1) malabsorptive diarrhea due to defective transporters and the collagenous bands, (2) secretory diarrhea due to rheogenic anion secretion, and (3) “leak flux–induced diarrhea” due to an impaired epithelial barrier and a passive back leak of ions and water into the intestinal lumen. This latter driving force of diarrhea is important (e.g., in Clostridium difficile infection).58 Normally, the epithelial barrier is a prerequisite for maintaining the “milieu inte´rieur.” If defective, significant amounts of solutes and water will enter the lumen, depending on the electrochemical driving forces and permeability of the paracellular barrier. More direct evidence for this concept was obtained by Fasano et al.,59 who showed that an attenuated Vibrio cholerae strain that is depleted of the chloride secretion–inducing cholera toxin gene still induced watery diarrhea in human volunteers. This was assumed to be due to a second toxin (zonula occludens toxin), leading to a reversible decrease in Re of the ileum by altering tight junction structure. This leak flux mechanism, in concert with the slight active secretory component in collagenous colitis, is also compatible with the finding that diarrhea persists when the patient fasts,28 especially if taking into account the


complete absence of any reserve capacity for NaCl absorption observed in the present study.

References 1. Lindstrøm CG. Collagenous colitis with watery diarrhea: a new entity? Pathol Eur 1976;11:87– 89. 2. Ja¨rnerot G, Tysk C, Bohr J, Eriksson S. Collagenous colitis and fecal stream diversion. Gastroenterology 1995;109:449 – 455. 3. Hwang W, Kelly J, Shaffer E, Hershfield N. Collagenous colitis: a disease of pericryptal fibroblast sheath? J Pathol 1986;149:33– 40. 4. Bohr J. A review of collagenous colitis. Scand J Gastroenterol 1998;33:2–9. 5. Riddell RH, Tanaka M, Mazzoleni G. Non-steroidal anti-inflammatory drugs as a possible cause of collagenous colitis: a case control study. Gut 1992;33:683– 686. 6. Giardiello FM, Hansen C III, Lazenby AJ, Hellman DB, Milligan FD, Bayless TM, Yardley JH. Collagenous colitis in setting of nonsteroidal antiinflammatory drugs and antibiotics. Dig Dis Sci 1990; 35:257–260. 7. Mogensen AM, Olsen JH, Gudmand-Højer. Collagenous colitis. Acta Med Scand 1984;216:535–540. 8. Pieterse AS, Hecker R, Rowland R. Collagenous colitis: a distinctive and potentially reversible disorder. J Clin Pathol 1982;35:338–340. 9. Gledhill A, Cole FM. Significance of basement membrane thickening in the human colon. Gut 1984;25:1085–1088. 10. Lee E, Schiller LR, Vendrell D, Santa Ana CA, Fordtran JS. Subepithelial collagen table thickness in colon specimens from patients with microscopic colitis and collagenous colitis. Gastroenterology 1992;103:1790 –1796. 11. Anderson JM, Van Itallie CM. Molecular structure of tight junctions and their role in epithelial transport. News Physiol Sci 2001;16:126 –130. 12. Tsukita S, Furuse M. Pores in the wall: claudins constitute tight junction strands containing aqueous pores. J Cell Biol 2000;149:13–16. 13. Madara JL, Stafford J. Interferon-␥ directly affects barrier function of cultured intestinal epithelial monolayers. J Clin Invest 1989; 83:724 –727. 14. Schmitz H, Fromm M, Bentzel CJ, Scholz P, Bode H, Epple HJ, Riecken EO, Schulzke JD. Tumor necrosis factor-alpha (TNF-␣) regulates the barrier in the human intestinal cell line HT-29/B6. J Cell Sci 1999;112:137–146. 15. Stockmann M, Fromm M, Schmitz H, Schmidt W, Riecken EO, Schulzke JD. Duodenal biopsies of HIV-infected patients with diarrhoea exhibit epithelial barrier defects but no active secretion. AIDS 1998;12:43–51. 16. Stockmann M, Gitter AH, Sorgenfrei D, Fromm M, Schulzke JD. Low edge damage container insert that adjusts intestinal forceps biopsies into Ussing chamber systems. Pflugers Arch 1999;438:107–112. 17. Fromm M, Schulzke JD, Hegel U. Control of electrogenic Na⫹ absorption in rat late distal colon by nanomolar aldosterone added in vitro. Am J Physiol 1993;264:E68 –E73. 18. Schmitz H, Fromm M, Bode H, Scholz P, Riecken EO, Schulzke JD. Tumor necrosis factor-␣ induces Cl⫺ and K⫹ secretion in human distal colon driven by prostaglandin E2. Am J Physiol 1996;271:G669 –G674. 19. Schulzke JD, Fromm M, Bentzel CJ, Zeitz M, Menge H, Riecken EO. Epithelial ion transport in the experimental short bowel syndrome of the rat. Gastroenterology 1992;102:497–504. 20. Tai YH, Tai CY. The conventional short-circuiting technique undershort-circuits most epithelia. J Membr Biol 1981;59:173–177. 21. Fromm M, Schulzke JD, Hegel U. Epithelial and subepithelial contributions to transmural electrical resistance of intact rat jejunum, in vitro. Pflugers Arch 1985;405:400 – 402. 22. Schulzke JD, Fromm M, Menge H, Riecken EO. Impaired intestinal sodium and chloride transport in the blind loop syndrome of the rat. Gastroenterology 1987;92:693– 698.

August 2002

23. Gitter AH, Schulzke JD, Sorgenfrei D, Fromm M. Ussing chamber for high-frequency transmural impedance analysis of epithelial tissues. J Biochem Biophys Methods 1997;35:81– 88. 24. Kapuscinski J. DAPI: a DNA specific fluorescent probe. Biotech Histochem 1995;70:220 –233. 25. Anagnostopoulos I, Schuppan D, Riecken EO, Gross UM, Stein H. Tenascin labelling in colorectal biopsies: a useful marker in the diagnosis of collagenous colitis. Histopathology 1999;34:425–431. 26. Rahner C, Mitic LL, Anderson JM. Heterogeneity in expression and subcellular localization of claudins 2, 3, 4 and 5 in the rat liver, pancreas, and gut. Gastroenterology 2001;120:411– 422. 27. Schmitz H, Barmeyer C, Fromm M, Runkel N, Foss HD, Bentzel CJ, Riecken EO, Schulzke JD. Altered tight junction structure contributes to the impaired epithelial barrier function in ulcerative colitis. Gastroenterology 1999;116:301–309. 28. Hoogerwerf WA, Tsao SC, Devuyst O, Levine SA, Yun CH, Yip JW, Cohen ME, Wilson PD, Lazenby AJ, Tse CM, Donowitz M. NHE2 and NHE3 are human and rabbit intestinal brush-border proteins. Am J Physiol 1996;270:G29 –G41. 29. Yang H, Jiang W, Furth EE, Wen X, Katz JP, Sellon RK, Silberg DG, Antalis TM, Schweinfest CW, Wu GD. Intestinal inflammation reduces expression of DRA, a transporter responsible for congenital chloride diarrhea. Am J Physiol 1998;275:G1445–1453. 30. Rask-Madsen J, Grove O, Hansen MGJ, Bukhave K, HenrikNielsen R. Colonic transport of water and electrolytes in a patient with secretory diarrhea due to collagenous colitis. Dig Dis Sci 1983;28:1141–1146. 31. Sandle GI, Hayslett JP, Binder HJ. Effect of glucocorticoids on rectal transport in normal subjects and patients with ulcerative colitis. Gut 1986;27:309 –316. 32. Palade GE, Simionescu M, Simionescu N. Structural aspects of the permeability of the microvascular endothelium. Acta Physiol Scand 1979;463(Suppl):11–32. 33. Gogarten W, Ko¨ckerling A, Fromm M, Riecken EO, Schulzke JD. Effect of acute Yersinia enterocolitica infection on intestinal barrier function in the mouse. Gastroenterology 1994;29:814 – 819. 34. Jones BA, Gores GJ. Physiology and pathophysiology of apoptosis in epithelial cells of the liver, pancreas, and intestine. Am J Physiol 1997;273:G1174 –G1188. 35. Gitter AH, Bendfeldt K, Schulzke JD, Fromm M. Leaks in the epithelial barrier caused by spontaneous and TNF-alpha-induced single-cell apoptosis. FASEB J 2000;14:1749 –1753. 36. Bojarski C, Gitter AH, Bendfeldt K, Mankertz J, Schmitz H, Wagner S, Fromm M, Schulzke JD. Permeability of HT-29/B6 colonic epithelium as a function of apoptosis. J Physiol (Lond) 2001; 535:541–552. 37. Fazeli A, Stehen RG, Dickinson SL, Bautista D, Dietrich WF, Bronson RT, Bresalier RS, Lander ES, Costa J, Weinberg RA. Effects of p53 mutations on apoptosis in mouse intestinal and human colonic adenomas. Proc Natl Acad Sci U S A 1997;94:10199–10204. 38. Moss S, Scholes JV, Walters JRF, Holt PR. Increased small intestinal apoptosis in coeliac disease. Gut 1996;39:811– 817. 39. Stra¨ter J, Wellisch I, Riedl S, Walczak H, Koretz K, Tandara A, Krammer PH, Mo¨ller P. CD95 (APO-1/Fas)-mediated apoptosis in colon epithelial cells: a possible role in ulcerative colitis. Gastroenterology 1997;113:160 –167. 40. Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, Tsukita S, Tsukita S. Occludin: a novel integral membrane protein at tight junctions. J Cell Biol 1993;123:1777–1788. 41. Furuse M, Fujita K, Hiiragi T, Fujimoto K, Tsukita Sh. Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin. J Cell Biol 1998; 141:1539 –1550. 42. Furuse M, Furuse K, Sasaki H, Tsukita S. Conversion of Zonulae Occludentes from tight to leaky strand types by introducing claudin-2 into Madin-Darby Canine Kidney I Cells. J Cell Biol 2001; 153:263–272.


443

43. Wong V, Gumbiner BM. A synthetic peptide corresponding to the extracellular domain of occludin perturbs the tight junction permeability barrier. J Cell Biol 1997;136:399 – 409. 44. MacCarthy KM, Skare IB, Stankewich MC, Furuse M, Tsukita S, Rogers RA, Lynch RD, Schneeberger EE. Occludin is a functional component of the tight junction. J Cell Sci 1996;109:2287–2298. 45. Balda MS, Whitney JA, Flores C, Gonzalez S, Cereijido M, Matter K. Functional dissociation of paracellular permeability and transepithelial electrical resistance and disruption of the apical-basolateral intramembrane diffusion barrier by expression of a mutant tight junction membrane protein. J Cell Biol 1996;134:1031–1049. 46. Wan H, Winton HL, Soeller C, Tovey ER, Gruenert DC, Thompson PJ, Stewart GA, Taylor GW, Garrod DR, Cannell MB, Robinson C. Der p 1 facilitates transepithelial allergen delivery by disruption of tight junctions. J Clin Invest 1999;104:123–133. 47. Saitou M, Furuse M, Sasaki H, Schulzke JD, Fromm M, Takano H, Noda T, Tsukita S. Complex phenotype of mice lacking occludin, a component of tight junction strands. Mol Biol Cell 2000;11:4131–4142. 48. Inai T, Kobayashi J, Shibata Y. Claudin-1 contributes to the epithelial barrier function in MDCK cells. Eur J Cell Biol 1999; 78:849 – 855. 49. Van Itallie C, Rahner C, Anderson JM. Regulated expression of claudin-4 paracellular conductance through a selective decrease in sodium permeability. J Clin Invest 2001;107:1319 –1327. 50. Duffey ME, Hainau B, Ho S, Bentzel CJ. Regulation of epithelial tight junction permeability by cAMP. Nature 1981;289:451– 453. 51. Conyers G, Milks LC, Conklyn M, Showell H, Cramer EB. A factor in serum lowers resistance and open tight junctions of MDCK cells. Am J Physiol 1990;259:C577–C585. 52. Lowe PJ, Miyal K, Steinbach JH, Hardison WGM. Hormonal regulation of the hepatocyte tight junction permeability. Am J Physiol 1988;255:G454 –G461. 53. Grotjohann I, Fromm M, Schulzke JD. Effect of tumor necrosis factor ␣ on rat rectum in vitro. Ann N Y Acad Sci 2000;915:282–286. 54. Schulzke JD, Fromm M, Hegel U. Epithelial and subepithelial resistance of rat large intestine: segmental differences, effect of stripping, time course, and action of aldosterone. Pflugers Arch 1986;407:632– 637. 55. Gunther U, Schuppan D, Bauer M, Matthes H, Stallmach A, Schmitt-Graff A, Riecken EO, Herbst H. Fibrogenesis and fibrolysis in collagenous colitis. Am J Pathol 1999;155:493–503. 56. Bugno M, Graeve L, Gatsios P, Koj A, Heinrich PC, Travis J, Kordula T. Identification of the interleukin-6/oncostatin M response element in the rat tissue inhibitor of metalloproteinases-1 (TIMP-1) promotor. Nucleic Acids Res 1995;23:5041–5047. 57. Wang M, Hu Y, Shima I, Stearns ME. Identification of positive and negative regulator elements for the tissue inhibitor of metalloproteinase 1 gene. Oncol Res 1998;10:219 –233. 58. Moore R, Pothoulakis C, LaMont JT, Carlson S, Madara JL. Clostridium difficile toxin A increases intestinal permeability and induces Cl secretion. Am J Physiol 1990;259:G165–G172. 59. Fasano A, Baudry B, Pumplin DW, Wasserman SS, Tall BD. Vibrio cholerae produces a second enterotoxin which affects intestinal tight junctions. Proc Natl Acad Sci U S A 1991;88:5242–5246.

Received November 27, 2001. Accepted April 25, 2002. Address requests for reprints to: Jo ¨rg-D. Schulzke, M.D., Medizinische Klinik I – Gastroenterologie, Infektiologie & Rheumatolgie, Universita ¨tsklinikum Benjamin Franklin, Freie Universita ¨t Berlin, Hindenburgdamm 30, 12200 Berlin, Germany. e-mail: schulzke@ medizin.fu-berlin.de; fax: (49) 30-8445-4493. Supported by grants from Deutsche Forschungsgemeinschaft (DFG Schu 559/6-3 and Schu 559/7-1) and by the Else Kro ¨ner-FreseniusStiftung. The excellent assistance of Anja Fromm and the great support of the electronic engineer Detlef Sorgenfrei is gratefully acknowledged.