Published online before print January 2, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
Departments of
* Internal Medicine I and
Pathology, Robert Bosch Hospital, Stuttgart, Germany;
Dr. Margarete Fischer-Bosch Institute of Clinical Pharmacology, Stuttgart, Germany; and
Department of Dermatology, University of Kiel, Kiel, Germany
1 Correspondence: Dr. Margarete Fischer-Bosch Institut für Klinische Pharmakologie, Auerbachstr. 112, 70376 Stuttgart, Germany. E-mail: jan.wehkamp{at}ikp-stuttgart.de
ABSTRACT
Elafin (or skin-derived antileukoprotease) and secretory leukocyte protease inhibitor (SLPI) are serine antiproteases antagonizing human neutrophil elastase (HNE), thereby preventing tissue injury from excessive release of proteolytic enzymes by inflammatory cells. Furthermore, elafin and SLPI are "defensin-like" molecules with broad antimicrobial activity. The balance between proteases and antagonists may critically determine inflammatory processes in Crohn’s disease (CD) and ulcerative colitis (UC). Real-time PCR was performed to quantitate colonic, proinflammatory cytokine IL-8, protease (HNE), and antiprotease mRNA (elafin and SLPI) in a total of 340 biopsies from 117 patients (47 CD, 45 UC, 25 controls). Histological inflammation was scored, and HNE, elafin, and SLPI were localized and semiquantified by immunostaining in 51 colonic paraffin sections (23 CD, 11 UC, 17 controls). Proinflammatory IL-8, degree of histological inflammation, and granulocyte content were similar in UC and CD. Elafin stained predominantly in the epithelium and SLPI in mucosal inflammatory cells. HNE mRNA levels and immunostaining were increased equally in both forms of inflammatory bowel disease. Levels of mRNA and immunostaining of the antiproteases elafin and SLPI were enhanced strongly in inflamed versus noninflamed UC. It is surprising that comparing inflamed versus noninflamed CD, this increase was significantly less pronounced for elafin and even lacking for SLPI. Despite comparable degrees of inflammation and protease levels, the induction of both antiproteases was attenuated in CD. This could contribute to the transmural depth of tissue destruction in CD. Elafin and SLPI may be added to the list of defensin-like peptides with diminished induction in CD versus UC.
Key Words: inflammatory bowel disease • antimicrobial peptides
INTRODUCTION
Crohn’s disease (CD) and ulcerative colitis (UC) are the two major groups of inflammatory bowel diseases (IBD) characterized by a chronic inflammation of the intestine. One of the main histopathologic differences between CD and UC is the microscopic extension of inflammation, which in UC, is localized to the mucosa and submucosa, whereas CD is characterized by a transmural inflammation [1 ]. As a result, in CD, deep fissuring ulcers may penetrate through the muscle layer, resulting in abscesses or fistulas between involved segments or adjacent organs [1 ]. About 35% of CD patients are affected by fistulae [2 ]. The reason for the transmural extension of the inflammation and the fistulizing character in CD is still unknown.
Proteases are involved in multiple biological processes including inflammation and tissue injury; the balance between these proteases and their inhibitors seems to be essential in maintaining tissue integrity [3 4 5 ]. Serine proteases, such as human neutrophil elastase (HNE), are produced and secreted by inflammatory cells and play a key role in various inflammatory disorders, contributing to tissue destruction through their proteolytic activity [3 4 5 ]. To minimize tissue damage, serine protease inhibitors are secreted to neutralize aggressive proteases [3 4 5 ]. Elafin, also called skin-derived antileukoprotease, and secretory leukocyte protease inhibitor (SLPI) are two important members of the serine antiprotease family [3 4 5 ]. Various epithelial and inflammatory cells produce and secrete these protective peptides locally at the site of injury [3 4 5 ]. In contrast, the systemic concentration of these antiproteases in the serum is comparably low [6 ]. The main proteolytic activity for elafin and SLPI is directed against HNE, thus, antagonizing excessive elastase release in inflamed tissue. Furthermore, both molecules are capable of antagonizing various other proteases: Elafin inhibits porcine pancreas elastase and proteinase-3, whereas SLPI is also a potent inhibitor of trpsin, chymotrypsin, tryptase, chymase, and cathepsin G [3 , 5 , 7 , 8 ]. In addition to their function as antiproteases, they are effective antimicrobial peptides, active against Gram-positive and Gram-negative bacteria, fungi, and viruses [4 , 9 10 11 12 ]. Both molecules are cationic peptides with a low molecular mass [5 ].
These rather unique features are also a key characteristic of defensins, an important class of antimicrobial peptides. These striking similarities explain why elafin and SLPI are also called "defensin-like" molecules as part of innate immune defense [4 ]. Antimicrobial peptides seem to play an important role in IBD [13 14 15 16 17 18 ], which prompted us to look at the related molecules elafin and SLPI in this context. The aim of this study was to investigate the expression profiles of the serine protease HNE and their inhibitors elafin and SLPI in colonic tissue of IBD patients. A shifted balance between proteases and its inhibitors may play a central role in transmural extension of the inflammation in CD, as opposed to the merely mucosal process in UC.
MATERIALS AND METHODS
Patients
A total of 117 patients underwent routine colonoscopy for anemia, cancer prevention, postsurgical follow-up, diarrhea, and acute IBD flares or IBD cancer surveillance. Twenty-five served as noninflammatory controls, 47 had CD, and 45 patients had UC. The diagnosis of IBD was based on standard radiological and endoscopic findings. Patients were treated according to their clinical status with corticosteroids, azathioprine, and/or aminosalicylates. Up to four biopsy specimens were taken from random areas of the colon in controls. In the case of IBD, samples were taken from macroscopically inflamed and not inflamed mucosa, if possible. Samples were snap-frozen immediately in liquid nitrogen. The median age was 43 years in CD and 40 years in UC. The median duration of disease was 5 years in CD and 10 years in UC, and 33% of patients with CD and 50% with UC were female. Furthermore, 51 colonic, paraffin-embedded tissue sections obtained by surgery were collected by the Department of Pathology, Robert Bosch Hospital (Stuttgart, Germany). Twenty-three of these patients had CD, 11 had UC, and 17 patients with colonic diverticulosis served as controls. All patients gave their written, informed consent, and the local ethical committee approved the study.
RNA preparation and RT
RNA isolation and RT were performed as described previously [17
]. In brief, frozen biopsies were disrupted mechanically, and total RNA was isolated using TRIzol reagent (Gibco-BRL, Grand Island, NY, USA), according to the supplier’s protocol. RNA quality and quantity were determined by gel electrophoresis and photometry. Subsequently, 2 µg total RNA was reverse-transcribed with oligo dT primers and 200 U Superscript (Gibco-BRL), according to routine procedure. These RNA preparations from each single biopsy were used for the assay of all proteases, antiproteases, villin, and cytokines tested.
Real-time RT-PCR
Real-time RT-PCR analyses were performed in a fluorescence temperature cycler (LightCycler, Roche Molecular Biochemicals, Mannheim, Germany) as described previously [17
, 18
]. This technique continuously monitors the cycle-by-cycle accumulation of fluorescent-labeled PCR product. Briefly, cDNA, corresponding to 20 ng RNA, served as a template in a 20-µl reaction containing 4 mmol/L MgCl2, 0.5 µmol/L each primer, and 1x LightCycler-FastStart DNA Master SYBR Green I mix (Roche Molecular Biochemicals). The sense and antisense sequences and the length of the PCR products are listed in Table 1
. The temperature profile for elafin was 95°C for 10 s, 62°C for 5 s, and 72°C for 15 s; for SLPI, 95°C for 10 s, 60°C for 5 s, and 72°C for 15 s; for HNE, 95°C for 10 s, 63°C for 5 s, and 72°C for 15 s; and for IL-8, 95°C for 10 s, 66–60°C (step size 1°C) for 5 s, and 72°C for 15 s. At the end of each run, melting curve profiles were produced by cooling the sample to 65°C for 15 s and then heating slowly at 0.20°C/s up to 95°C with continuous measurement of fluorescence to confirm amplification of specific transcripts. Cycle-to-cycle fluorescence emission readings were monitored and analyzed using LightCycler software (Roche Molecular Biochemicals). The specificity of the amplification products was verified further by subjecting the amplification products to electrophoresis on a 2% agarose gel. The fragments were visualized by ethidium bromide staining, and the specificity of PCR products was verified by sequencing of representative samples. Standard curves were obtained for each primer set with serial dilutions of cDNA, and the transcript levels were normalized to the amount of total RNA, as recommended previously [19
]. To determine whether there was a correlation between proteases and antiproteases with the grade of inflammation, we measured IL-8 [18
]. To assess possible damage of epithelial cells, which could result in secondary changes of epithelial products, we measured expression of villin, which is expressed at the basolateral side of enterocytes.
|
View this table: [in a new window] |
Table 1. Primer Sequences and Length of PCR Products for IL-8, HNE, Elafin, and SLPI
|
DNA isolation and detection of nucleotide-binding oligomerization domain protein 2 (NOD2) mutations
Genotyping of genomic DNA for the functionally relevant NOD2 frameshift mutation [single nucleotide polymorphism 13 (SNP13)] was performed in colonic samples using TaqMan technology (Applied Biosystems, Foster City, CA, USA), as described previously [13
]. Biopsies from inflamed and noninflamed areas from a total of six patients with NOD2 SNP13 mutation were selected and compared with a randomly selected group of wild-type NOD2 patients.
Statistics
All statistical analyses and graphs were performed using Prism 3.0 software. For comparison of grouped real-time PCR as well as for grouped immunohistochemical data, we perfomed the Mann-Whitney test. The Spearman’s rank correlation was used for nonparametric correlation between the different subgroups. Values of P < 0.05 were considered statistically significant.
RESULTS
Mucosal Inflammation and epithelial damage
As expected, IL-8 mRNA was induced in macroscopically inflamed versus noninflamed CD and UC without showing any significant differences between inflamed CD and UC (Fig. 1
). This molecular marker was paralleled by the histological grade of inflammation and the content of granulocytes: Confirming the macroscopic assessment, there was more inflammation as well as a higher content of granulocytes in inflamed IBD samples but no significant difference when comparing CD and UC (Table 2)
. The epithelial marker villin, an indicator of epithelial damage, was decreased in inflamed samples, but there were no differences between CD and UC (Fig. 1B)
.
 View larger version (62K): [in a new window] |
Figure 1. Grade of inflammation in IBD patients and controls. Expression of (A) IL-8 mRNA and (B) Villin mRNA in colonic biopsies from controls, CD, and UC patients: IL-8 mRNA is induced significantly in inflamed versus noninflamed CD as well as in inflamed versus noninflamed UC. Villin expression is almost equal in each group tested. (C) Representative samples of hematoxylin stainings of colonic, paraffin-embedded tissue sections: no inflammation in a control slide (left) and strong inflammation in a CD (middle) and UC (right) slide. Data about systemic, histological comparisons are shown in Table 2
. P < 0.05 is statistically significant.
|
|
View this table: [in a new window] |
Table 2. Histological Inflammation Score and Granulocyte Content in Colonic Tissue Sections in Percent of the Total Group
|
 View larger version (62K): [in a new window] |
Figure 2. HNE in IBD patients and controls. (A) Expression of HNE mRNA in colonic biopsies from controls, CD, and UC patients: No significant differences were found between inflamed and noninflamed IBD samples. (B) Immunohistochemical localization of HNE in inflammatory cells of a control (left), CD (middle), and UC (right) patient. Data about systemic, histological comparisons are shown in Table 3
.
|
|
View this table: [in a new window] |
Table 3A. Immunohistochemical Detection of Elafin, SLPI, and HNE in the Epithelium of Colonic Tissue Sections in Percent of the Total Group
|
|
View this table: [in a new window] |
Table 3B. Immunohistochemical Detection of Elafin, SLPI, and HNE in Inflammatory Cells of Colonic Tissue Sections in Percent of the Total Group
|
 View larger version (70K): [in a new window] |
Figure 3. Elafin in IBD patients and controls. (A) Expression of elafin mRNA in colonic biopsies from controls, CD, and UC patients: Elafin mRNA is induced significantly in inflamed versus noninflamed CD as well as in inflamed versus noninflamed UC. Elafin expression was increased significantly in inflamed UC versus inflamed CD. (B) Representative immunohistochemical staining of elafin in controls (left), CD (middle), and UC (right). (C) A higher resolution showing positive cell types is shown. Data about systemic, histological comparisons are shown in Table 3
Table 3
. P < 0.05 is statistically significant (*, P<0.05; **P>0.01; ***, P<0.001).
|
 View larger version (70K): [in a new window] |
Figure 4. SLPI in IBD patients and controls. (A) Expression of HNE mRNA in colonic biopsies from controls, CD, and UC patients: SLPI mRNA is induced significantly in inflamed versus noninflamed UC but not in inflamed versus noninflamed CD. SLPI expression was increased significantly in inflamed UC versus inflamed CD. (B) Representative immunohistochemical staining of SLPI in controls (left), CD (middle), and UC (right). (C) A higher resolution showing positive cell types is shown. Data about systemic, histological comparisons are shown in Table 3
Table 3
. P < 0.05 is statistically significant (ns, not significant; *, P<0.05; **, P<0.01; ***, P<0.001).
|
 View larger version (20K): [in a new window] |
Figure 5. Correlation of IL-8 and antiproteases in CD (•) and UC ( ): elafin mRNA versus IL-8 mRNA (top), SLPI mRNA versus IL-8 mRNA (middle), SLPI mRNA versus elafin mRNA (bottom). Correlation statistics are given in the text.
|
DISCUSSION
The data presented here show an attenuated induction of antiproteases elafin and SLPI in inflamed versus noninflamed CD compared with UC. As the proinflammatory cytokine IL-8, histological degree of inflammation, and granulocyte content, as well as the corresponding protease HNE were similar in inflamed CD and UC, these data indicate a shifted protease/antiprotease balance, which might predispose to tissue disruption and deeper inflammation in CD.
Historically, SLPI and elafin were first isolated from sputum secretions of patients with chronic obstructive pulmonary disease and cystic fibrosis [20 , 21 ], but more recent work demonstrated the presence and functional importance in various other tissues (especially in the skin) and body fluids [22 23 24 25 26 ]. Furthermore, both antiproteases were shown to be expressed by different inflammatory cells, such as neutrophils and macrophages [27 ]. Based on their antiprotease properties, elafin and SLPI are able to protect local tissue against tissue destruction during inflammatory conditions [3 4 5 , 22 , 26 , 28 ]. In addition,these peptides are directly involved in tissue repair and wound healing [29 , 30 ] and potently recruit inflammatory cells by chemotactic properties [3 ]. In the colon, SLPI and elafin have been found in the normal epithelium [31 , 32 ], and elafin has shown to be increased in UC [33 ]. In this paper, we show that elafin predominated in the epithelium, whereas SLPI was preferentially found in inflammatory cells. Conversely, HNE was detected exclusively in inflammatory cells. Similar to serine proteases and antiproteases, the large family of matrix metalloproteinases (MMP) and their corresponding tissue inhibitors of metalloproteases (TIMP) exhibits a differential expression in IBD: The aggressive proteases MMP-2 and MMP-3 are strongly induced in colonic CD but not in UC, whereas the expression of inhibitors (TIMP-2 and TIMP-3) did not show any differences [34 ]. Looking at intestinal fistula formations, which are characteristic for CD, Kirkegaard et al. [35 ] observed a marked up-regulation of MMP-3 and MMP-9 without increased expression of the corresponding antiproteases and concluded that metalloproteases may contribute to fistula formation through degradation of the extracellular matrix. Taken together, these studies and our data indicate that there is a dysbalance between tissue-destructive proteases and their inhibitors in the colon of CD patients. Serine and metalloproteases are apparently increased in UC, but this induction is attenuated or missing in the case of CD.
Based on their biochemical properties such as low molecular mass, disulfide bonds, and cationic load, elafin and SLPI are part of the large family of antimicrobial peptides such as defensins, and it has been demonstrated that they kill a variety of microbes including gram-positive and gram-negative bacteria, fungi, and selected viruses (e.g., HIV) [4
, 9
10
11
12
]. We have suggested that CD is at least in part a result of a disturbed mucosal antibacterial barrier [16
, 36
]. CD of the small intestine is characterized by a specific decrease of Paneth cell 
-defensins and diminished antimicrobial activity [13
, 18
], whereas CD of the colon is asscociated with an attenuated induction of different human ß-defensins (HBD), such as HBD-2 and HBD-3, but also other antimicrobial peptides, such as LL37 [17
, 37
]. In the case of HBD-2, we recently identified the molecular mechanism, which is a gene copy number polymorphism. Patients with colonic CD have one copy less, which explains why the transcript levels are induced in inflamed UC but much less in inflamed CD [38
]. Finding the mechanism for the attenuated induction of elafin and SLPI will be the aim of future investigations.
In summary, our data imply that the balance between two important serine antiproteases and HNE expression is drastically shifted toward the protease in CD compared with UC. This imbalance may lead to a more aggressive tissue destruction by HNE as well as other proteases in the colon of patients with CD and contribute to transmural and maybe also a fistulizing character of disease. As a result of their antimicrobial action, elafin and SLPI may be added to the list of defensin-like peptides with diminished induction in CD versus UC. These findings could have an important impact on future therapeutic strategies and finally help to prevent or alleviate tissue destruction.
ACKNOWLEDGEMENTS
This work was supported by the Robert Bosch Foundation (Stuttgart, Germany). In addition, J. W. is supported by the Ernst Jung Foundation (Hamburg, Germany). We thank Kathleen Siegel and Dagmar Biegger for excellent technical assistance.
Received September 20, 2006; revised November 10, 2006; accepted November 28, 2006.
REFERENCES
-defensin expression Gut 53,1658-1664-defensins in ileal Crohn’s disease Proc. Natl. Acad. Sci. USA 102,18129-18134This article has been cited by other articles:
 |
|
 |
|
  E. Kolaczkowska, W. Grzybek, N. van Rooijen, H. Piccard, B. Plytycz, B. Arnold, and G. Opdenakker Neutrophil elastase activity compensates for a genetic lack of matrix metalloproteinase-9 (MMP-9) in leukocyte infiltration in a model of experimental peritonitis J. Leukoc. Biol., March 1, 2009; 85(3): 374 - 381. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-F. Colombel, A. J M Watson, and M. F Neurath The 10 remaining mysteries of inflammatory bowel disease Gut, April 1, 2008; 57(4): 429 - 433. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Harder, R. Glaser, and J.-M. Schroder Review: Human antimicrobial proteins effectors of innate immunity Innate Immunity, December 1, 2007; 13(6): 317 - 338. [Abstract] [PDF]
|
|
|
|
|
|
S. Nuding, K. Fellermann, J. Wehkamp, and E. F Stange Reduced mucosal antimicrobial activity in Crohn's disease of the colon Gut, September 1, 2007; 56(9): 1240 - 1247. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
