Introduction

Crohn's disease (CD) is a chronic, relapsing inflammatory disorder that may affect any region of the gastrointestinal tract. It is characterized by transmural inflammation, narrowing of intestinal lumen, fistula formation, granulomas, and fibrosis [1]. Both genetic and environmental factors have been implicated in the pathogenesis of CD. A dysregulated response of the intestinal immune system to intraluminal antigens of bacterial origin predisposes to CD resulting in the activation and release of cytokines, eicosanoids, nitric oxide, and proteolytic enzymes, which initiate a cascade leading to intestinal injury [2, 3].

Matrix metalloproteinases (MMPs) constitute a family of 24 human enzymes collectively capable of degrading all extracellular matrices (ECM) and basement membrane components. However, they also regulate the activation of various growth factors and modify inflammatory reactions [4]. Several MMPs have been implicated in the pathobiology of human inflammatory bowel disease (IBD) being involved in tissue remodeling, angiogenesis, cell migration, and promotion of leukocyte extravasation in the actively inflamed area in the ulcer base in both CD and ulcerative colitis (UC) [5]. MMP-3 (stromelysin-1) is upregulated in CD [6–9] and both MMP-3 and MMP-9 (92 kDa gelatinase) have been associated with mucosal damage and fistulae in this disease [8, 10, 11]. MMP-9 is abundantly expressed in IBD and is pivotal for tissue damage [11]. The selective inhibition of both MMP-3 and MMP-9 has been suggested to be of therapeutic benefit in IBD [5]. MMP-1 (collagenase-1) is increased in the granulation tissue of CD [6, 12], may potentiate the inflammatory response, and take part in remodeling and monocyte migration [5]. MMP-1 is one of the highly upregulated genes in microarray studies on diseased areas of CD patients [13]. MMP-10 (stromelysin-2) is present in macrophages and T cells in CD stroma and in migrating enterocytes [14] and, together with MMP-7 (matrilysin-1), has been implicated to have a role in intestinal wound healing [15]. Instead, MMP-26 (matrilysin-2) is not significantly upregulated in the epithelium or stroma of CD patients [16]. Moreover, immune response can increase MMP-12 (matrix metalloelastase) in lamina propria (LP) macrophages [14, 17] and this proteinase may have a role in macrophage migration and tissue inflammation in CD [17].

A critical balance between MMPs and their tissue inhibitors (TIMPs) determines the outcome of tissue damage in CD. TIMP-1 and TIMP-3 are abundantly expressed in the granulation tissue of the stroma in active CD [6, 14]. TIMP-1 has been associated with fibrotic strictures of CD [8]. TIMP-3 is a potent inhibitor of TNF-α-converting enzyme. Single-nucleotide polymorphisms in genes coding for TIMP-1 and MMP-3 affect CD susceptibility and phenotype [18], further implicating the relevance of MMPs and TIMPs in the pathobiology of CD.

Corticosteroids represent standard treatment for CD but the use of anti-TNF-α agents is emerging in severe cases. The aim of this study was to assess the intestinal MMP profile of CD patients receiving corticosteroids and conventional immunosuppressants (azathioprine or methotrexate) or anti-TNF-α agents before initiation of treatment and after a few months to learn more about the therapeutic pathways for immunosuppressive agents and to identify possible therapeutic targets for selective MMP inhibitors.

Materials and methods

Patients and tissues

All patients gave their written informed consent for participation in this study, approved by the Ethics Committee of the Helsinki University Central Hospital, Finland. Seventeen adult patients with established CD diagnosis referred to ileocolonoscopy were recruited between January 2005 and June 2007 at Helsinki University Central Hospital, Helsinki, Finland. Of these patients, seven were female and ten were male, their age ranging between 19 and 47 years (mean 26.5 years). The disease duration, defined as months since diagnosis by endoscopy, ranged between 0 and 249 months (mean 74.76 months; Table 1). One patient was newly diagnosed. Indications for the baseline endoscopies were an acute flare (n = 10), a chronic active disease (n = 2), and a prolonged need of corticosteroid treatment (n = 5). Endoscopy findings were scored according to Crohn's disease endoscopic index of severity (CDEIS) [19] and clinical activity according to Crohn's disease activity index (CDAI) [20]. In this study, CDAI <150 indicated clinically inactive disease and ≥150 active disease [21]. CDEIS 0 to 9 indicated inactive or mildly active CD and CDEIS >9 indicated moderately or severely active disease [22].

Table 1 Clinical data for the patients
Full size table

After endoscopy, 12 patients received therapy with anti-TNF-α agent, either infliximab infusion 5 mg/kg at week 0 and at week 8 (n = 11) or subcutaneous adalimumab 80 mg followed by 40 mg every other week until week 8 (n = 1; Table 1). Of the total 17 patients, five received steroid induction with introduction of conventional immunosuppressive treatment (n = 1 methotrexate, n = 4 azathioprine). Two of the induction steroid treatments began 2–5 days prior to the first endoscopy. Endoscopic response was assessed in anti-TNF-treated patients around 3 months and in those treated with immunosuppressants 4–6 months after the beginning of the treatment. The ileum was reached in 31 out of 34 (91%) and the cecum in 33 out of 34 (97%) of the endoscopies. At the time of the endoscopies, patients provided blood samples for serum C-reactive protein (CRP, in milligrams per liter) and fecal samples for measurement of calprotectin. All but one patient responded to treatment, ten reaching remission of the disease. However, the patient with no improvement in CDEIS values demonstrated marked decrease in total histological score and CDAI and was thus included in the study.

Histology scoring

During all the endoscopies, four biopsies—targeted to most severely diseased areas—were taken from the ileum, right colon, transverse colon, and left colon, and rectum. Routine histology was performed on specimens stained with hematoxylin and eosin. For this study, the most severely affected colonic segment in the baseline endoscopy was selected for MMP stainings of pretreatment and post-treatment biopsy specimens (five right colon, six transverse colon, and six left colon). Histological findings of the corresponding segment were scored by an experienced gastrointestinal pathologist (P.K.) according to a scoring system for histological abnormalities in CD mucosal biopsy specimens [23]: epithelial damage was classified as normal (score = 0), focal (score = 1) or extensive (score = 2) pathology; architectural changes as normal (score = 0), moderately (<50%) disturbed (score = 1), or severely (>50%) disturbed (score = 2). The infiltration of mononuclear cells in the LP and polymorphonuclear cells in the LP was scored as normal (score = 0), moderate (score = 1), or severe (score = 2). Polymorphonuclear cells in the epithelium were scored 1 if they were in the surface epithelium, 2 in the presence of cryptitis, and 3 in the presence of a crypt abscess. Presence of an erosion and/or an ulcer and presence of a granuloma were both scored “yes” (score = 0) or “no” (score = 1). In addition, the number of biopsy specimens affected were scored none (score = 0), <33% (score = 1), 33% to 66% (score = 2), or >66% (score = 3) [23]. The total histological score for the segment was a sum of these variables (minimum 0, maximum 16).

Fecal calprotectin

Fecal calprotectin was measured by a quantitative enzyme immunoassay (PhiCal Test, Calpro AS, Oslo, Norway; NovaTec Immunodiagnostica, Dietzenbach, Germany). The values quoted as normal in our laboratory were <100 µg/g [24, 25]. In this study, fecal calprotectin ≥200 µg/g indicated active CD [22].

Immunohistochemistry

Immunohistochemistry was performed using streptavidin–biotin–peroxidase complex technique (DakoCytomation, StreptABComplex/HRP Duet, Mouse/Rabbit, Glostrup, Denmark; Elite Goat IgG Vectastain ABC Kit, Vector Laboratories, Burlingame, CA, USA) and the antibody–polymer detection technique (PowerVision, Poly-HRP IHC Kit, ImmunoVision Technologies, Brisbane, CA, USA). Diaminobenzidine (DAB) or 3-amino-9-ethylcarbazole (AEC) was used as chromogenic substrates and Mayer's hematoxylin as counterstain. Monoclonal antibodies were used to stain for MMP-1 (1:500, IM35L, Oncogene Research Products, La Jolla, CA, USA) [26], MMP-7 (1:130, IM40L, Calbiochem, La Jolla, CA, USA) [26–28], MMP-9 (1:100, MS-569-P1, Neomarkers, Fremont, CA, USA), MMP-10 (1:300, 5E4, Novocastra Laboratories, New Castle upon Tyne, UK) [28], TIMP-1 (1:100, IM63, Calbiochem) [26], and TIMP-3 (1:400, IM43L, Calbiochem) [26]. Polyclonal antibodies were used for MMP-26 (1:120, a gift from Prof. K. Isaka, Tokyo Medical University) [29]. MMP-1, MMP-9, and MMP-10 were pretreated with 1% trypsin solution for 30 min at +37°C. MMP-7, MMP-26, TIMP-1, and TIMP-3 were pretreated in a +95°C water bath for 20 min (Dako retrieval solution, pH 6; Dakocytomation). The incubation conditions for the antibodies were: +4°C overnight for MMP-1, MMP-7, MMP-9, TIMP-1, and TIMP-3; 2 h at room temperature for MMP-10; and 1 h at +37°C for MMP-26. For negative controls, parallel sections of the same samples were processed using preimmune sera or normal rabbit or mouse immunoglobulin. Immunohistochemical specimens were graded independently by two different investigators (L.M., U.S.-K.) in a semiquantitative fashion under a light-field microscope at ×100 magnification using a scale marking staining intensity as follows: 0 = less than 20 positive cells; 1 = 20–50 positive cells; 2 = 50–200 positive cells; 3 = over 200 positive cells.

Statistical analysis

Paired and independent samples t tests, nonparametric Wilcox and Mann–Whitney's and Spearman's correlation tests were performed to investigate the significance of results with the Statistical Package for Social Sciences 15.0 for Windows. A p value under 0.05 was considered as significant.

Results

When comparing the patient groups before and after treatment, both fecal and serological inflammation indicators decreased significantly (calprotectin means 2,614 vs. 249; p = 0.000 and CRP 32 vs. 1; p = 0.001), as did endoscopic findings according to CDEIS (14.9 vs. 3.5; p = 0.000), total histological score of biopsies (7 vs. 4; p = 0.004), and symptoms of the disease assessed by CDAI (214 vs. 61; p = 0.000). When looking at pretreatment and post-treatment biopsies of patients treated with anti-TNF-α agents or conventional immunosuppressants separately, no significant differences in MMP or TIMP profiles could be observed. Both groups presented with the general patterns as discussed below.

MMP-1

MMP-1 was detected in the stroma and surface epithelium of all samples (Fig. 1a; Table 2). In the stroma, MMP-1 was expressed by macrophages (Fig. 1b). The expression pattern in both pretreatment and post-treatment biopsies was fairly similar (Table 2).

Fig. 1
figure 1

Stromal and epithelial expression of MMP-1 (a) was detected in all samples and the expression pattern was similar in both sets of samples. b Higher magnification of a with MMP-1 positive macrophages (arrowheads). Epithelial expression of MMP-7 was stronger in initial samples (c) when compared to the ones taken after treatment (d). Arrows depict positive epithelial cells (c). MMP-9 was expressed in the stroma of most samples; however, expression was higher in initial samples (e) and decreased with treatment (g). Macrophages (arrowhead) and neutrophils (arrow) expressed MMP-9 (f). AEC was used as chromogenic substrate in a and b, DAB in c–g. Samples a–c, e, and f were taken before treatment. Samples d and g had been treated with anti-TNF-α agents. Scale bars = 7.5 µm (b, f); 15 µm (a, e, g); 30 µm (c, d)

Full size image
Table 2 The inflammation scores and expression of MMPs and TIMPs in the colon of patients with CD
Full size table

MMP-7

MMP-7 expression was detected in the epithelium in seven out of 17 of the initial samples and two out of 17 of the post-treatment specimens (Fig. 1c, d; Table 2). Stronger epithelial expression in initial samples was also confirmed in semiquantitative scoring of immunohistochemical specimens (means 0.47 vs. 0.12; p = 0.055; Tables 2 and 3A). Expression of epithelial MMP-7 tended to be stronger in the group active for CDAI (≥150) compared to that of the inactive group (means 0.50 vs. 0.18; p = 0.091). Epithelial MMP-7 expression tended to have a positive correlation with total histological score (correlation coefficient = 0.334; p = 0.053; Table 3B) and active disease as indicated by CRP (≥10 mg/l; mean 0.55 vs. 0.17; p = 0.018).

Table 3 The means of semiquantitatively assessed scores for MMPs and TIMPs for baseline and post-treatment biopsies and association of MMP and TIMP expression with the total histological score, calprotectin, CDEIS, CDAI, and CRP values
Full size table

MMP-9

MMP-9 protein was generally absent from the surface epithelium, but stromal expression by macrophages and neutrophils was initially strong (Fig. 1e–g; Table 2). In semiquantitative analysis of neutrophil MMP-9, higher stromal expression was observed in initial samples (means 2.12 vs. 1.29; p = 0.039; Table 3A). Total histological score showed significant correlation with both macrophage (correlation coefficient = 0.439; p = 0.009) and neutrophil (correlation coefficient = 0.577; p = 0.000) MMP-9 (Table 3B). Stronger expression of MMP-9 in neutrophils was also observed in samples that were considered active for CDEIS (>9; mean 2.00 vs. 1.38; p = 0.061) and CRP (≥10; mean 2.27 vs. 1.43; p = 0.020).

MMP-10

Expression of MMP-10 in the surface epithelium was detected in 16 out of 17 initial samples and 14 out of 17 post-treatment samples (Fig. 2a; Table 2). All samples in both groups had positive stromal expression in macrophages and lymphocytes (Fig. 2a, b). Thus, expression in both sets of samples was very similar. Total histological score and fecal calprotectin tended to correlate positively with stromal MMP-10 (correlation coefficient = 0.325; p = 0.060 and correlation coefficient = 0.353; p = 0.040, respectively; Table 3B).

Fig. 2
figure 2

MMP-10 was detected in the epithelium and stroma of most samples (a). MMP-10 was expressed by lymphocytes (thin arrows) and macrophages (arrowheads, b). MMP-26 showed stronger stromal expression in pretreatment samples (c) in neutrophils (arrowhead, c′) and endothelial cells (arrow, c′) when compared to post-treatment samples (d). TIMP-1 was expressed in stroma in macrophages (arrowhead, f) and neutrophils (arrow, f). When compared to pretreatment samples (e), the number of positive macrophages decreased with treatment and none of the post-treatment samples presented TIMP-1 positive neutrophils (g, h). Stromal expression of TIMP-3 was initially strong (i) but decreased with treatment (j). However, epithelial TIMP-3 increased with treatment. AEC was used as chromogenic substrate in a and b, DAB in c–j. Samples a–c, e, f, and i were taken before treatment, samples d, h, and j had undergone treatment with anti-TNF-α agents and sample g with corticosteroids and a conventional immunosuppressant. Scale bars = 7.5 µm (b, c′, f); 15 µm (e, g–j); 30 µm (a, c, d)

Full size image

MMP-26

MMP-26 was not present in the surface epithelium of any sample. In the stroma, MMP-26 was positive in 15 of the pretreatment biopsies and in seven of the post-treatment samples in endothelial cells and neutrophils (Fig. 2c′; Table 2). MMP-26 expression in stromal neutrophils decreased with treatment (means 0.94 vs. 0.53; p = 0.030; Fig. 2c, d; Table 3A). Fecal calprotectin and total histological score showed strong positive correlation with stromal MMP-26 (correlation coefficient = 0.432; p = 0.011 and correlation coefficient = 0.418; p = 0.014; Table 3B). Biopsies considered to represent active disease (calprotectin ≥200 or CDEIS >9) showed higher expression of stromal MMP-26 than biopsies from inactive disease (means for MMP-26 0.91 [calprotectin ≥200] vs. 0.36 [calprotectin <200]; p = 0.013; and 0.94 [CDEIS >9] vs. 0.50 [CDEIS ≤9]; p = 0.017).

TIMP-1

Epithelial expression of TIMP-1 did not statistically differ between pretreatment and post-treatment samples (Table 2). However, expression in stromal macrophages was higher in biopsies taken before treatment than after this (means 1.76 vs. 0.76; p = 0.003; Fig. 2e–g; Table 3A). This pattern was also present in stromal neutrophils, as none of the post-treatment samples had positive neutrophils (means 0.47 vs. 0; p = 0.041; Fig. 2e–g; Table 3A). For macrophage expression, there was significant positive correlation with indicators of inflammation, calprotectin (correlation coefficient = 0.586; p = 0.000), CDEIS (correlation coefficient = 0.451; p = 0.007), and total histological score (correlation coefficient = 0.524; p = 0.001; Table 3B). Biopsies considered active for CDEIS (>9) showed stronger stromal expression of TIMP-1 in macrophages (means 1.78 vs. 0.69; p = 0.002) and neutrophils (means 0.44 vs. 0; p = 0.046) than those considered inactive for CDEIS. Also, stronger expression by macrophages was present in biopsies grouped as active disease according to calprotectin (means 1.61 vs. 0.55; p = 0.005).

TIMP-3

TIMP-3 was detected in the surface epithelium in both groups of biopsies (Table 2), showing a trend for weaker expression in pretreatment samples (means 1.18 vs. 1.71; p = 0.058; Fig. 2h, i; Table 3A). Although all samples had positive stromal macrophages and plasma cells, TIMP-3 expression was significantly stronger in the pretreatment samples (means 2.65 vs. 2.29; p = 0.029; Fig. 2h, i; Table 3A). On the contrary, epithelial TIMP-3 showed negative correlation with inflammation indicators: TIMP-3 had significant inverse associations with CRP and CDEIS (correlation coefficient = −0.486; p = 0.004 and correlation coefficient = −0.462; p = 0.006; Table 3B). Furthermore, inverse trends were present when comparing epithelial TIMP-3 with the fecal calprotectin (correlation coefficient = −0.335; p = 0.053; Table 3B) and the group with low CRP (<10 mg/l; correlation coefficient = −0.404; p = 0.056). Epithelial TIMP-3 expression was lower in biopsies considered active for CDEIS (means 1.11 vs. 1.81; p = 0.011), calprotectin (means 1.22 vs. 1.99; p = 0.019), and CRP (means 1.00 vs. 1.65; p = 0.028). In contrast, stromal TIMP-3 showed positive correlation with calprotectin (correlation coefficient = 0.541; p = 0.001) and CDEIS values (correlation coefficient = 0.469; p = 0.005; Table 3B), as well as when separately looking at samples in the group active according to calprotectin level (correlation coefficient = 0.578; p = 0.004). Stromal expression was higher in the group considered active for CDEIS (means 2.67 vs. 2.25; p = 0.014), calprotectin (means 2.61 vs. 2.18; p = 0.019), and CRP (means 2.73 vs. 2.35; p = 0.039).

Discussion

MMPs participate in T cell-mediated injury in the human intestine [30]. In fact, levels of specific MMPs in the urine may provide clinicians with useful diagnostic and prognostic information [31]. However, some MMPs may play a role in epithelial restitution and tissue remodeling. In this study, treatment with anti-TNF-α agents (infliximab or adalimumab) or with corticosteroids and conventional immunosuppressants (azathioprine or methotrexate) decreased the expression of epithelial MMP-7 and that of MMP-9, MMP-26, and their tissue inhibitors TIMP-1 and TIMP-3 in the stroma. While expression of MMP-1 and MMP-10 was generally not altered, epithelial TIMP-3 increased with treatment.

Treatment with anti-TNF-α agents or conventional immunosuppressants decreased the expression of MMP-9 by neutrophils that associated with a decrease in histological severity score. Similarly, downregulation of MMP-9 production of macrophages was observed to correlate with decreases in histological score and CRP. MMP-9 is pivotal for tissue damage [11] and its selective inhibition has been suggested to be of therapeutic benefit in IBD [5]. In a previous study, infliximab treatment increased MMP-2 and decreased MMP-9 in serum of patients (the latter also in tissue) with CD [32]. The changes were not strictly associated with the response to treatment. Enhanced leukocyte MMP-9 expression in CD seemed to be regulated by TNF-α. Moreover, in intestinal mucosal explants from IBD, infliximab downregulated MMP-9 relative to TIMP-1 and TIMP-2 [33]. MMP-9 has been shown to respond to immunotherapeutic agents. There exist several studies on the effect of infliximab on tissue MMP profile particularly in arthritis and psoriasis. Infliximab decreased MMP-9 in psoriatic skin [34]. In rheumatoid arthritis, infliximab infusions downregulated serum MMP-1, MMP-3, MMP-9, TIMP-1, and TIMP-2 levels [35] and modulation of the MMP/TIMP system by infliximab can contribute to the anti-inflammatory and tissue-remodeling effects of TNF-α blockade in spondylarthropathies [36].

Expression of MMP-26 in neutrophils decreased with treatment, depicting a direct relationship with histological score and calprotectin. MMP-26 may thus be important in tissue destruction or migration of neutrophils as it is able to degrade various ECM and basement components, such as type I gelatin, vitronectin, fibronectin, and type IV collagen [37, 38]. MMP-26 has been previously detected in neutrophils [39] and endothelial cells [40], agreeing with our results. MMP-26 was not present in the epithelium of any sample. It has been previously shown to be present in the cytoplasm of migrating enterocytes in UC but not CD [16].

MMP-7 tended to be less expressed in the epithelium after treatment, correlating with a decrease in histological score in post-treatment samples. This agrees with our previous findings in necrotizing enterocolitis and further strengthens the putative role of this MMP in tissue destruction [40]. Expression of MMP-7 has also been discovered to correlate with the degree of inflammation in UC [41].

The lack of significant changes in MMP-1 is not in complete disagreement with previous studies. MMP-1 transcripts have been shown to be present in CD myofibroblasts, as well as MMP-1 protein, but the latter did not possess any bioactivity [42]. However, transcript levels of MMP-1 have been shown to correlate with CDAI in mucosal biopsy specimens of CD patients [43]. Furthermore, when mucosal biopsies of active CD were studied 8 weeks after initiation of infliximab therapy, MMP-1 levels post-treatment were able to predict patients achieving long-term remission [44]. In intestinal mucosal explants from IBD, infliximab downregulated MMP-1 relative to TIMP-1 and TIMP-2 and decreased MMP-1 and MMP-3 activities [33].

No statistically significant changes in semiquantitatively assessed epithelial or stromal MMP-10 were detected post-treatment. However, it tended to decrease when histological score and calprotectin became lower. In an experimental model of IBD using T cell activation of intestinal explants [30], MMP-10 was detected in areas with most severe injury. This discrepancy may be due to different experimental systems, i.e., in vivo biopsies vs. in vitro cultures.

Expression of TIMP-1 in macrophages and neutrophils was weaker in post-treatment samples. Also, histological score, calprotectin, and CDEIS values decreased together with stromal expression of TIMP-1. Our study thus portrayed a decrease in TIMP-1 after therapy. In rheumatoid arthritis, infliximab downregulated serum levels of TIMP-1 [35], but in colonic myofibroblasts isolated from patients with active CD, infliximab and adalimumab induced a dose-dependent increase TIMP-1 production [45].

Stromal TIMP-3 decreased with treatment, correlating with a decrease in calprotectin and CDEIS values. This agrees with our results in an experimental model in which both TIMPs were expressed abundantly at the beginning of culture but downregulated after T cell activation [30]. However, there was a trend for greater TIMP-3 expression in the epithelium in post-treatment samples, but lower values for calprotectin, CDEIS, and CRP. Thus, epithelial TIMP-3 may participate in epithelial regeneration.

Fecal calprotectin associates with the degree of histological inflammation in the colon [46] and showed a positive correlation with stromal MMP-10, MMP-26, TIMP-1, and TIMP-3. However, a negative correlation was seen with epithelial TIMP-3. When comparing the biopsies from inactive and active disease based on calprotectin values, an association with epithelial MMP-1 and stromal MMP-9 was observed and, for the latter, with stromal TIMP-1 and TIMP-3. Interestingly, calprotectin may affect various pathophysiological processes by competing with MMPs for zinc [47].

CDEIS, the activity index determined in endoscopy, depicted a direct relationship with epithelial and stromal MMP-7 and stromal TIMP-1. However, a negative association was present when comparing CDEIS with epithelial TIMP-3. Also, when looking at samples considered active for CDEIS, an inverse relationship was present between this and stromal MMP-26. When looking at all samples and CDAI, only a negative relationship with epithelial TIMP-3 could be seen, but looking separately at samples inactive and active for CDAI, the latter correlated positively with epithelial MMP-1 and MMP-10. CRP did not present as a good inflammation indicator, as correlation with MMPs and other inflammation indicators varied. Previous studies indicate that calprotectin is indeed a more precise marker for inflammation in CD and a useful surrogate marker for mucosal healing [46].

In conclusion, our results suggest that immunosuppressive drugs modulate disease activity in CD by downregulation of MMP-9 and MMP-26 positive neutrophils and stromal TIMP-1 and TIMP-3. MMP-1 and MMP-10 do not seem to be important players in the mucosal wound-healing response elicited by TNF-α inhibition.