Research ArticleInflammatory Bowel Diseases

Food-Grade Bacteria Expressing Elafin Protect Against Inflammation and Restore Colon Homeostasis

See allHide authors and affiliations

Science Translational Medicine  31 Oct 2012:
Vol. 4, Issue 158, pp. 158ra144
DOI: 10.1126/scitranslmed.3004212

Abstract

Elafin, a natural protease inhibitor expressed in healthy intestinal mucosa, has pleiotropic anti-inflammatory properties in vitro and in animal models. We found that mucosal expression of Elafin is diminished in patients with inflammatory bowel disease (IBD). This defect is associated with increased elastolytic activity (elastase-like proteolysis) in colon tissue. We engineered two food-grade strains of lactic acid bacteria (LAB) to express and deliver Elafin to the site of inflammation in the colon to assess the potential therapeutic benefits of the Elafin-expressing LAB. In mouse models of acute and chronic colitis, oral administration of Elafin-expressing LAB decreased elastolytic activity and inflammation and restored intestinal homeostasis. Furthermore, when cultures of human intestinal epithelial cells were treated with LAB secreting Elafin, the inflamed epithelium was protected from increased intestinal permeability and from the release of cytokines and chemokines, both of which are characteristic of intestinal dysfunction associated with IBD. Together, these results suggest that oral delivery of LAB secreting Elafin may be useful for treating IBD in humans.

Introduction

Inflammatory bowel diseases (IBDs), including Crohn’s disease (CD) and ulcerative colitis (UC), are chronic inflammatory disorders of the gut, with a substantial socioeconomic impact worldwide. Current treatments for IBD [glucocorticoids and monoclonal antibody (mAb) therapies] are suboptimal because of their side effects and the high incidence of treatment failure (20 to 40% of patients fail treatment). Thus, alternative approaches to treating IBD focusing on new targets are desirable. Several studies have demonstrated that proteases are crucial for maintenance of chronic inflammatory responses in the gastrointestinal tract (1-4) . Indeed, elevated proteolytic activity has been detected in intestinal tissues from CD and UC patients (5, 6). This increased proteolytic activity could result from the up-regulation of protease expression, from a decrease in the efficacy or expression of endogenous protease inhibitors, or from both. Elafin, a tissue-derived inhibitor of elastase and proteinase 3 that blocks elastolytic activity and has been studied for its anti-inflammatory properties at mucosal surfaces (7, 8), is among the endogenous protease inhibitors expressed in the human intestine (6). We previously demonstrated that transgenic mice expressing human Elafin were protected from colitis in different mouse models of IBD (2). This previous work suggested that endogenous antiproteases such as Elafin could naturally protect against inflammation in the intestine. We thus reasoned that the oral delivery of Elafin might be of value for the treatment of IBD. However, delivery of protease inhibitors to the gut could interfere with digestive functions. To be efficient, protease inhibitors such as Elafin would have to be delivered in small amounts right at the site of injury and would need to be steadily released. To this end, we engineered nonpathogenic lactic acid bacteria (LAB) with GRAS (generally recognized as safe) status (9) to deliver recombinant human Elafin directly to the intestinal mucosa. We used Lactococcus lactis (L. lactis) to deliver therapeutic proteins to the mucosa (10, 11) and Lactobacillus casei (L. casei), which persists longer in the gastrointestinal tract than L. lactis, to secrete active Elafin (12, 13). We demonstrate the therapeutic potential of LAB expressing Elafin in mouse models of acute and chronic IBD and in cultured human epithelial cells exposed to supernatants from cultured biopsy material from IBD patients.

Results

Elastolytic homeostasis is disrupted in IBD patients

Elastolytic activity in colon tissues from IBD and control patients was investigated by in situ zymography and by assaying enzymatic activity in supernatants obtained from organ cultures of tissue biopsies. Strong elastolytic activity was detected in the submucosa of CD patients compared with tissues from healthy individuals (Fig. 1A). Elastolytic activity was increased in culture supernatants of biopsies harvested from both inflamed and non-inflamed areas of the colon of IBD patients compared to supernatants of biopsies from healthy controls (Fig. 1B). In biopsies from IBD patients or healthy controls, fluorescence in situ hybridization (FISH) was used to measure mRNA expression of Elafin, an inhibitor of the proteases elastase and proteinase 3 (Fig. 1C). Elafin mRNA was detected in colon tissues from healthy controls in both the mucosa and the submucosa, with the strongest expression being detected in intestinal epithelium (Fig. 1C, left). In contrast, in tissues from CD and UC patients, Elafin mRNA expression was greatly reduced (Fig. 1C, middle and right panels). Quantification of fluorescence intensity for Elafin mRNA in the mucosa (Fig. 1D) confirmed that Elafin expression was decreased in tissues from CD and UC patients compared to those from controls. Thus, an increase in elastolytic activity in the colon of IBD patients correlates with a reduction in the expression of the endogenous elastolytic inhibitor Elafin. Given the depletion of Elafin observed in colon tissues from IBD patients (Fig. 1B) and the anti-inflammatory properties of Elafin (2), we hypothesized that the oral delivery of Elafin might be able to attenuate the symptoms of colitis.

Fig. 1

Elastolytic homeostasis is disrupted in colon tissue from IBD patients. (A) Representative elastolytic activity detected in colon tissue from an IBD (CD) patient and a healthy control (these images are representative of n = 5 biopsies from different patients in each group) is shown. Color scale bar indicates the intensity of detected activity. (B) Elastolytic activity in supernatants from cultures of colon biopsies taken from IBD patients and healthy controls. (C) Representative images showing mRNA FISH for expression of Elafin in colon mucosa biopsy samples from healthy controls and patients with CD or UC. The limits of the colon mucosa were defined by light microscopy by tracing a dotted line at the border of the lamina propria. Color scale bar indicates the intensity of mRNA detection. (D) Fluorescence in mucosal tissue was quantified above this line with ImageJ software. Data in (B) and (D) are means ± SEM (n = 4 in each group, ***P < 0.001, **P < 0.01, *P < 0.05, using Kruskal-Wallis with Dunn’s post-test).

LAB expressing Elafin are vectors for Elafin delivery

To deliver Elafin to the colon, we engineered two food-grade strains of LAB, L. lactis and L. casei, to produce and secrete human Elafin. The expression of human recombinant Elafin by LAB was driven by a nisin promoter and was confirmed by Western blot analysis (Fig. 2A). Two protein bands at 25 and 15 kD corresponding to the precursor and mature forms of Elafin, respectively, were detected in lysates of recombinant LAB, whereas in supernatants, only one band corresponding to the mature form of Elafin (~15 kD) was detected. No signal was detected in supernatants or homogenates of wild-type LAB strains that did not express Elafin. We quantified Elafin production by enzyme-linked immunosorbent assay (ELISA) and calculated that 5 × 109 Elafin-expressing L. lactis produced 0.5 μg of Elafin protein in 1 hour.

Fig. 2

Recombinant human Elafin expressed by LAB. (A) Detection of recombinant human Elafin in homogenates of 2 × 107 bacteria (H) or supernatants from bacterial cultures (S) of wild-type (WT) L. lactis and L. casei or Elafin-expressing (Elafin) L. lactis and L. casei. Two forms of Elafin are detected in bacterial homogenates: a precursor and a mature form. (B) Representative immunofluorescence images (n = 5 mice in each group) for colon tissue from mice treated with DSS to induce colitis and then given GFP-tagged WT L. lactis or recombinant human Elafin–expressing L. lactis. GFP-tagged WT L. lactis (anti-GFP, red) and Elafin-expressing L. lactis (anti-Elafin, green) are shown; nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; blue). The limits of the lamina propria are defined by a thin dotted line in all images.

We investigated whether such Elafin-producing LAB strains were able to deliver Elafin to inflamed colon tissue. A mouse model of IBD in which colitis was induced by dextran sodium sulfate (DSS) was used. For 7 days after starting DSS regimen, mice were orally administered with L. lactis tagged with green fluorescent protein (GFP). The tagged L. lactis either did or did not (wild-type) express recombinant human Elafin. Elafin protein expression and the presence of GFP-tagged bacteria were then examined in the colon of these mice. Elafin, which has no murine ortholog, was not detected in colon tissue from mice treated with the wild-type L. lactis strain but was detected in the mucosa of mice orally treated with Elafin-expressing L. lactis (Fig. 2B). Elafin expression was found in clusters, colocalizing with GFP-tagged bacteria at the epithelial surface (arrows). Diffuse staining at sites remote from the GFP-tagged bacteria was also observed (arrowheads), suggesting that Elafin is secreted and diffuses within mucosal tissues (Fig. 2B). Elafin expression was also detected in the lamina propria (Fig. 2B).

Elafin-expressing LAB restore colon tissue elastolytic homeostasis

Increased elastolytic activity was observed not only in colon tissue from patients with IBD (Fig. 1A) but also in mouse colon mucosa after induction of colitis by DSS (Fig. 3A). Daily oral administration of wild-type L. lactis to DSS-treated mice did not modify the pattern of elastolytic activity in colon tissue. However, in mice that received Elafin-expressing L. lactis, the elastolytic activity in inflamed colon tissue was markedly reduced and its pattern returned to that observed in the non-inflamed tissues of naïve mice (Fig. 3A). Elastolytic activity was also evaluated in lumenal washes of mouse colon tissues 7 days after exposure to either DSS or vehicle alone (water). In the DSS-treated animals, there was an increase in elastolytic activity compared with the untreated group. Oral administration of L. lactis or L. casei expressing Elafin, but not wild-type strains not expressing Elafin, resulted in a marked reduction (~60%) in the elastolytic activity of DSS-treated mice (Fig. 3, B and C). This suggested that Elafin expression favored a reduction in elastolytic activity and restoration of colon tissue homeostasis.

Fig. 3

Effects of Elafin-expressing LAB on gut elastolytic homeostasis. (A) Representative images show elastolytic activity in colon tissue from mice that drank water (H2O) or water containing DSS (DSS) for 7 days. Some mice were treated orally with vehicle (PBS) or with WT L. lactis or Elafin-expressing L. lactis. Activity intensity is indicated in a pseudocolor gradient ranging from 0 (black) to 255 (white). (B and C) Elastolytic activity in washes of the colon lumen from mice that drank water or water containing DSS (DSS) for 7 days and that were then orally treated for 7 days with WT LAB. (B) WT or Elafin-expressing L. lactis. (C) WT or Elafin-expressing L. casei. Data are shown as means ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, using Kruskal-Wallis with Dunn’s post-test (n = 10 in each group).

Elafin-expressing LAB reduce acute and chronic gut inflammation

To determine whether oral treatment with Elafin-expressing LAB had an impact on acute DSS-induced intestinal inflammation, we analyzed several markers of inflammation after the induction of colitis. As previously described (14), DSS exposure for 7 days resulted in acute inflammation characterized by a high macroscopic damage score, increased colon thickness, and granulocyte infiltration [measured by myeloperoxidase (MPO) activity] (Fig. 4, A to C). These inflammatory markers were reduced in mice treated with Elafin-expressing L. lactis or L. casei (Fig. 4, A to C). The concentrations of the cytokines keratinocyte chemoattractant (KC), interferon-γ (IFN-γ), interleukin-6 (IL-6), IL-17A, and monocyte chemoattractant protein-1 (MCP-1) were all increased after 7 days of DSS exposure (Fig. 4, D to H). The expression of these cytokines was reduced after oral treatment with Elafin-expressing but not wild-type L. casei (Fig. 4, D to H). Only IL-17A production was reduced by treatment with both Elafin-expressing and wild-type L. casei (Fig. 4G). Staining for specific markers of macrophages (F4/80), neutrophils (Ly-6G), and T lymphocytes (CD3) was quantified in mouse colon tissue. Staining for the three markers was increased after 7 days of DSS exposure (Fig. 5, A and B). Staining for macrophage and neutrophil markers in colon tissue from mice with colitis was reduced (≥50%) by oral treatment with Elafin-expressing but not wild-type L. lactis compared with DSS-treated mice that only received vehicle [phosphate-buffered saline (PBS)]. Staining for the T lymphocyte marker was reduced (30%) by treatment with both Elafin-expressing and wild-type L. lactis (Fig. 5).

Fig. 4

Effects of Elafin-expressing LAB on acute inflammation. (A to C) The macroscopic damage score (A), colon wall thickness (B), and MPO activity (C) in mouse colon tissue are shown. Mice were given water or water containing DSS for 7 days and then received daily oral treatments for 7 days with vehicle (PBS) or with WT or Elafin-expressing strains of L. casei or L. lactis. (D to H) Protein concentrations for cytokines detected in colon mucosal tissues from mice given water or water containing DSS (DSS) for 7 days and then daily oral treatments with PBS or with WT or Elafin-expressing L. casei. Data are shown as means ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, using Kruskal-Wallis with Dunn’s post-test (A, B, C, E) or one-way analysis of variance (ANOVA) followed by Bonferroni post-test (D, F, G, H) (n = 10 in each group).

Fig. 5

Effects of Elafin-expressing L. lactis on the number of immune cells during acute inflammation. (A) Representative immunofluorescence staining (n = 8 mice in each group) for macrophages stained with anti-F4/80 (red), neutrophils stained with anti–Ly-6G (red), or T lymphocytes stained with anti-CD3 (red) in mouse colon. Mice were given water or water containing DSS (DSS) for 7 days to induce colitis and also received oral treatment with vehicle (PBS) or with WT or Elafin-expressing L. lactis. Nuclei were stained with DAPI (blue). (B) Quantification of staining intensity in each group was expressed as fold increase over values for control (water-treated) animals. Data are shown as means ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, using one-way ANOVA followed by Bonferroni post-test.

The effects of Elafin-expressing LAB were compared to the effects of IL-10–expressing LAB. Elafin-expressing L. lactis was able to reduce inflammation more effectively than IL-10–expressing L. lactis or wild-type L. lactis, particularly with respect to tissue damage and colon thickness (fig. S1).

Chronic inflammation mimicking what is seen in IBD patients experiencing periods of remission and relapse was reproduced in mice by exposing the animals to successive cycles of DSS in their drinking water. The colon mucosa of mice treated with three cycles of DSS undergoes successive periods of acute inflammation and healing. Seven days after the last DSS treatment (42 days after the first DSS treatment), colon tissues from mice were seen to be regenerating but still showed macroscopic damage and an increase in the thickness of the mucosal wall (Fig. 6, A and B). Colonic washes of mice that were subjected to chronic DSS treatment showed an increase in elastolytic activity (Fig. 6C). Treatment of mice undergoing DSS-induced chronic inflammation with wild-type LAB did not modify the characteristics of chronic inflammation. Treatment with recombinant human Elafin–expressing L. lactis decreased elastolytic activity, colon wall thickness, and macroscopic damage score (Fig. 6, A to C). Treatment with L. casei secreting recombinant human Elafin also decreased the macroscopic damage score but failed to decrease colon thickness and restore elastolytic homeostasis (Fig. 6, A to C).

Fig. 6

Effects of Elafin-expressing LAB on chronic inflammation. (A to E) The macroscopic damage score (A and D), colon wall thickness (B and E), and elastolytic activity (C) in mouse colon tissue are shown. (A to C) Mice were given three cycles of a 7-day treatment with water or water containing DSS to induce chronic inflammation. (D and E) Chronic colitis was induced by intraperitoneal injection of CD45RBhigh T lymphocytes into SCID mice, whereas controls received CD45RBlow T lymphocytes. Mice were treated with vehicle (PBS) or with WT or Elafin-expressing strains of L. casei or L. lactis. Data are shown as means ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, using one-way ANOVA followed by Bonferroni post-test (n = 10 in each group).

Another mouse model of chronic colitis, based on an immune T cell response, was also examined. As previously described (15), signs of colitis (increased macroscopic damage score and mucosal wall thickness) were observed in tissues from mice receiving CD45RBhigh activated T cells compared to mice receiving CD45RBlow T cells. Treatment with Elafin-expressing L. lactis, but not with wild-type LAB, decreased the macroscopic damage score and mucosal wall thickness (Fig. 6, D and E).

Elafin-expressing LAB did not translocate to the blood, spleen, or liver of treated mice as shown by culturing these tissues on plates containing medium that would support LAB growth (fig. S2).

Elafin-expressing LAB protect human colon epithelium from inflammatory changes

To investigate the mechanisms by which Elafin-expressing LAB are able to protect intestinal mucosa from inflammatory damage, we compared the effects of Elafin-expressing and wild-type bacteria on human intestinal epithelial cell monolayers. Exposure of intestinal epithelial cells to tumor necrosis factor–α (TNF-α) was used to induce an inflammatory phenotype in the monolayers. Permeability to macromolecules was increased by TNF-α exposure as was the expression of mRNAs encoding CXCL-8, MCP-1, and IP-10 (IFN-γ–induced protein 10) (Fig. 7, A to D). Co-incubation of intestinal epithelial cells with Elafin-expressing L. lactis decreased TNF-α–induced permeability increases, whereas wild-type L. lactis had no effect (Fig. 7A). TNF-α–induced increases in CXCL-8 and MCP-1 mRNA expression were reduced by both Elafin-expressing L. lactis and wild-type bacteria (Fig. 7, B and C). Neither Elafin-expressing nor wild-type L. lactis was able to inhibit TNF-α–induced IP-10 mRNA expression (Fig. 7D). These results indicate that L. lactis alone can exert some beneficial effects on intestinal epithelial cell expression of the proinflammatory molecules CXCL-8 and MCP-1, but only Elafin-expressing L. lactis was able to restore epithelial barrier function.

Fig. 7

Effects of Elafin-expressing LAB on the response of human intestinal epithelial cells to TNF-α. Cultured monolayers of human intestinal epithelial cells (Caco-2 cells) were exposed to TNF-α and permeability to macromolecules was assessed (A) when cells were exposed to medium alone or to WT or Elafin-expressing L. lactis. (B to D) mRNA expression of CXCL-8 (B), MCP-1 (C), and IP-10 (D) was quantified in human intestinal epithelial cells exposed to TNF-α or medium alone or to WT or Elafin-expressing L. lactis. Data are shown as means ± SEM, *P < 0.05, **P < 0.01, ***P < 0.001, using one-way ANOVA followed by Bonferroni post-test [n = 24 (A), n = 18 (B), n = 12 (C and D)].

To evaluate the effects of Elafin-expressing L. lactis in an IBD environment, we incubated a cultured human intestinal epithelial cell line with the different L. lactis strains (wild-type or Elafin-recombinant). The cells were then exposed to supernatants from cultured colon biopsies from IBD patients. Culture biopsy supernatants harvested from inflamed areas of IBD patients (CD and UC) increased the permeability of intestinal epithelial cell monolayers when compared to vehicle or supernatants from biopsies harvested from healthy controls (Fig. 8A). When epithelial cell monolayers were co-incubated with the IBD biopsy supernatants in the presence of Elafin-expressing L. lactis, there was a marked reduction in the IBD supernatant–induced increases in permeability compared to the effects of wild-type L. lactis (Fig. 8B). CXCL-8 and MCP-1, but not IP-10, mRNA expression was increased in intestinal epithelial monolayers exposed to supernatants from biopsies from IBD patients (Fig. 8C). Increased CXCL-8 and MCP-1 mRNA expression was inhibited by co-incubation with Elafin-expressing L. lactis compared to co-incubation with wild-type L. lactis (Fig. 8, D and E). However, no effect was observed on IP-10 mRNA expression by co-incubation with Elafin-expressing LAB compared to wild-type L. lactis (Fig. 8, C and F).

Fig. 8

Effects of Elafin-expressing LAB on the response of human intestinal epithelial cells to supernatants from IBD patient colon biopsy tissue. Cultured monolayers of human intestinal epithelial cells (Caco-2 cells) were exposed to supernatants from cultures of colon tissue biopsies harvested from IBD patients (open circles and closed triangles) or from healthy subjects (closed circles). Data from exposure to biopsy supernatants harvested from CD patients are represented by closed triangles and from UC patients by open circles. (A and B) Permeability to macromolecules was assessed (A) when intestinal epithelial cells were exposed to medium alone (HBSS) (represented by crosses) or to supernatants from healthy subjects or IBD patients and (B) when intestinal epithelial cells were cocultured with WT or Elafin-expressing L. lactis. (C) mRNA expression of CXCL-8, MCP-1, and IP-10 was quantified in human intestinal epithelial cells exposed to medium alone (HBSS) or to IBD supernatants. (D to F) CXCL-8 mRNA expression (D), MCP-1 mRNA expression (E), and IP-10 mRNA expression (F) were quantified in human intestinal epithelial cells exposed to IBD supernatants and co-incubated with WT or Elafin-expressing L. lactis. Data are shown as means ± SEM, *P < 0.05, **P < 0.01, using one-way ANOVA followed by Bonferroni post-test.

Discussion

The incidence of IBD has been growing over the past decades and new therapeutic options are needed. A number of studies have investigated the possibility of using probiotics as a means to restore and maintain intestinal epithelial health (16). Clinical studies in IBD patients treated with probiotics have shown limited anti-inflammatory effects. The protective properties of probiotics need to be improved, and one way to do this is to genetically engineer these bacteria to express additional anti-inflammatory proteins to increase efficacy. However, the choice of protein to be expressed in probiotics has to be carefully considered and cannot be based solely on the anti-inflammatory properties of the heterologous protein. Given that the protein will be delivered at the mucosal surface, it has to be naturally present in the mucosa. Pleiotropic anti-inflammatory properties for this protein would ensure better efficacy by targeting different signaling pathways. Finally, a deficiency in the expression and function of this protein would have to be demonstrated in IBD patients.

Elafin recapitulates many of these properties. We and others have shown the expression of Elafin at the mucosal surface of human gut epithelia (6). Elafin is a low molecular weight molecule (10 kD) originally discovered as a protease inhibitor in the lung and skin (17, 18) that specifically inhibits neutrophil elastase and proteinase 3 (7). This inhibitory activity against neutrophil proteases accounts largely for its anti-inflammatory effects (2, 7, 8, 19). Indeed, many cell types involved in inflammatory responses, such as monocytes, macrophages, or endothelial cells, are modulated by Elafin through its inhibition of the proinflammatory transcription factors activating protein 1 (AP-1) and nuclear factor κB (NF-κB) (20, 21). Elafin also exhibits antimicrobial activity against a variety of pathogenic microorganisms (bacteria, fungi, and viruses). Elafin also restores barrier function to damaged intestinal epithelia (2). These data indicate that Elafin is pleiotropic in its role as a guardian of mucosal surfaces. Considering the mixed etiology of IBD, where defective immunoregulation, permeability defects, and luminal content (microbiota) are known to play active roles, in situ delivery of a protein such as Elafin might be able to modulate innate and adaptative immunity, regulate tissue repair, and exert antimicrobial properties. We demonstrate here that Elafin concentrations are down-regulated in tissues from IBD patients compared to healthy subjects and that this down-regulation is associated with an increase in elastolytic activity in colon biopsy tissue from those IBD patients. We chose to use probiotic bacteria to deliver Elafin to sites of inflammation instead of delivering other proteins such as the anti-inflammatory cytokine IL-10 (9). L. lactis–associated IL-10 delivery targets only the immune system and was found to be ineffective in clinical trials (22) and, as we have shown here, was less efficient than Elafin-expressing LAB in mouse models of colitis. Indeed, our results demonstrate that probiotics secreting Elafin ameliorated the effects of both acute and chronic colitis in mice. Elafin-expressing LAB protected against mucosal erosion and T cell–mediated damage in these mouse models. Elafin-expressing L. lactis was able to restore the mucosal architecture, modify the profile of proinflammatory cytokine production, and inhibit the recruitment of inflammatory cells in mice. However, could LAB-producing recombinant human Elafin be protective in IBD patients? We established that cultured human intestinal epithelial cells stimulated by an inflammatory milieu in the form of culture medium from fresh biopsy tissue from IBD patients lost their barrier function and secreted the proinflammatory chemokines CXCL-8 and MCP-1. Coculture of these human intestinal epithelial cells with Elafin-expressing L. lactis before treatment with the IBD biopsy supernatants prevented the loss of barrier function and blocked chemokine synthesis. This suggests that treatment with Elafin-expressing L. lactis is effective against intestinal damage and dysfunction induced by the IBD biopsy supernatants.

To be delivered into the colon and to be effective at the injury site after oral administration, recombinant probiotics must survive transit through the gastrointestinal tract. We detected Elafin in the inflamed colon mucosa of mice treated orally with LAB expressing recombinant human Elafin. Clusters of Elafin production were detected at mucosal surfaces, colocalizing with GFP-tagged bacteria. The bacteria were not detected in the lamina propria or deeper in the colon tissue, indicating that they stay at the mucosal surface. However, diffuse staining for Elafin was detected in the lamina propria, and this was not associated with GFP-tagged bacteria. This suggested that Elafin originating from L. lactis is secreted and diffuses deeper into the colon mucosa, where it can exert its anti-inflammatory properties by, for example, protecting against infiltrating immune cells and their inflammatory mediators such as neutrophil elastolytic enzymes.

TNF-α–induced increases in the expression of CXCL-8 and MCP-1 mRNAs in human intestinal epithelial cells were reduced by coculture with both Elafin-expressing and wild-type L. lactis. This demonstrates a protective effect of LAB that is independent of Elafin expression and confirms the previously reported probiotic effects of the L. lactis strain through blocking CXCL-8 production (23). These data also demonstrate that the LAB expressing recombinant human Elafin had the same probiotic effects as the wild-type strain. However, Elafin-expressing L. lactis was more efficient than the wild-type L. lactis at down-regulating CXCL-8 and MCP-1 expression induced by mediators released from IBD patient biopsies and also in restoring the barrier function of human intestinal epithelial cells stimulated by either TNF-α or supernatants from IBD patient biopsies. Together, these data provide evidence that oral treatment with recombinant L. lactis expressing Elafin allows the delivery of Elafin to the epithelial surface of the colon mucosa where it exerts protective effects on epithelial function. Our results also demonstrate that Elafin delivery by L. lactis exerts in vivo protective effects against the recruitment of inflammatory cells, particularly macrophages and neutrophils, in mouse models of colitis. This effect may be mediated by Elafin-induced inhibition of CXCL-8 (or KC in mice) and MCP-1 release by epithelial cells. Indeed, these two chemokines are known to act as chemoattractants for neutrophils and macrophages. However, we cannot rule out a possible effect of Elafin on other cell types involved in leukocyte recruitment, such as endothelial cells or leukocytes themselves. The presence of Elafin-expressing LAB in the colon lumen might also modify the composition of the gut microbiota, which is associated with the development of colitis in the DSS mouse model (24). The antimicrobial properties of Elafin could influence gut microbiota composition, rendering it more protective.

If bacteria expressing Elafin were to be used in the clinic for the treatment of intestinal inflammation, these genetically modified strains would need to be biologically contained as has been done for LAB carrying IL-10 (25, 26). Other proteins of interest involved in maintaining mucosal homeostasis such as epidermal growth factor 8 found in milk fat globules or other endogenous protease inhibitors (27) could also be expressed in recombinant LAB.

Few challenges remain before potentially moving Elafin-expressing LAB into clinical studies. Elafin has been shown to be safe when delivered to humans (8), and recombinant LAB have also been shown to be safe when given orally to humans (11, 22). However, the safety of Elafin-recombinant LAB would have to be tested. Another challenge is to determine the best LAB carrier to be used in patients. Our studies in mice suggest that L. casei and L. lactis may be equally good carriers of Elafin, but this may not necessarily be the case in humans.

We provide evidence that localized delivery of Elafin into the colon, secreted in situ by food-grade bacteria, ameliorates symptoms of inflammation and accelerates mucosal healing. These results suggest that there may be a potential clinical application for Elafin delivered by probiotic bacteria for treating IBD. This approach may offer a cost-effective, long-term treatment for IBD and other disorders of intestinal inflammation.

Methods

Patients and biopsies

Human colonic tissues were obtained from individuals treated at the Centre Hospitalier de Toulouse (France) and at the Digestive Disease Unit, Foothills Hospital (Calgary, Canada) (table S1). Biopsies were collected during colonoscopy procedures aimed at clinically evaluating the disease of established and well-characterized CD and UC patients or done in individuals undergoing colon cancer screening who were otherwise healthy (healthy controls). Written and verbal informed consent was obtained before enrollment in the study, and the Ethics Committee approved the human research protocol. Fresh biopsies were rinsed in isotonic sterile HBSS and were then immediately incubated in 2 ml of Hanks’ balanced salt solution (HBSS) at 37°C for 1 hour. For all experiments, supernatant volumes were standardized to the weight of incubated biopsies. Isolated biopsy specimens were embedded in optimal cutting temperature (OCT) compound (Dako) at −186°C and stored for in situ elastolytic activity.

Animals

C57BL/6 mice (6 to 8 weeks old) were obtained from Janvier. All mice were kept at room temperature with 12-hour light/dark cycles and free access to food and water. All procedures were approved by the Institutional Animal Care Committee and Veterinary Services.

Cloning and expression of Elafin in LAB

Gene encoding for Elafin was polymerase chain reaction (PCR)–amplified from plasmid DK6-Elafin (28). The sequences of the primers used were 5′ forward Elafin (CCAATGCATCAGCAGCTGTCACGGGAGTTCC) and 3′ reverse Elafin (GGACTAGTCCTCACTGGGGAACGAAACA GGCC). Primers were designed to eliminate the first codons of the Elafin region encoding for signal peptide (SP) and was replaced by the SP of Usp45 protein (SPUsp45), the main secreted protein from L. lactis. To that aim, the PCR product was digested, purified, and cloned in pSEC, an L. lactis secretion vector. In the resulting plasmid pSEC-Elafin, Elafin is fused in frame with a DNA fragment encoding for the ribosome-binding site and SPUsp45. Expression of the cassette is controlled by the inducible promoter PnisA, the activity of which depends on the concentration of nisin used. This plasmid was then introduced in an L. lactis strain harboring the nisin regulatory genes nisR and nisK integrated into its chromosome (L. lactis NZ9000), resulting in the recombinant strain NZ(pSEC-Elafin). To visualize in vivo our Elafin-expressing L. lactis strains, we cotransformed L. lactis NZ9000 with both pMV158:GFP (29) and NZ9pSEC:Elafin. The tools used are functional in lactobacilli strains such as L. casei, harboring the gene nisRK integrated into its chromosome.

Induction of Elafin production in LAB

Cultures of L. lactis and L. lactis GFP strains (L. lactis wild-type and L. lactis Elafin) were performed in M17 medium (Oxoid) supplemented with glucose (0.5%) at 30°C without shaking. Cultures of L. casei strains (L. casei wild-type and L. casei Elafin) were performed in Mann-Rogosa-Sharpe medium (BD Difco, Fisher Scientific) containing erythromycin (5 μg/ml) at 37°C without shaking. Exponential growth cultures (OD600 = 0.4–0.6) of L. lactis or L. casei strains were treated for 1 hour with nisin (L. lactis, 1 ng/ml; Sigma) or 3 hours (L. casei, 25 ng/ml) to induce recombinant protein expression. Recombinant strains of L. casei and L. lactis were maintained with chloramphenicol (10 μg/ml) and tetracycline for L. lactis GFP strains (4 μg/ml). Bacteria were centrifuged after nisin induction, washed, and suspended in corresponding volume of sterile PBS to get 5 × 109 colony-forming units (CFU) for oral dosing in mice.

Induction of colitis and study design

Colonic inflammation was induced by treatment with DSS (MP Biomedicals) dissolved in drinking water (3 to 5%, w/v) (30). We confirmed that in vitro incubation of LAB (recombinant or not) with DSS (1, 3, and 5%) for 8 hours had no effect on the growth of the bacteria.

The animals were free to drink the DSS solution for 7 days (no differences were reported for the volume consumed between water and DSS). For the whole period of DSS exposure, mice were orally treated daily with 5 × 109 CFU of recombinant LAB (either L. lactis or L. casei) expressing Elafin, wild-type LAB (5 × 109 CFU), or PBS. On day 7 after adding DSS to their drinking water, mice were sacrificed and colons were harvested for measurement of inflammation parameters (31): macroscopic damage score, bowel thickness, MPO activity, elastolytic activity, and cytokine expression.

Chronic DSS colitis was induced by administration of DSS for three 7-day cycles (3% from days 0 to 6, 2% from days 14 to 20, and 2% from days 28 to 34). Each DSS cycle alternated with a 7-day interval with normal drinking water. Mice were orally treated daily with 5 × 109 CFU (in 100 μl of PBS) of recombinant LAB (either L. lactis or L. casei) expressing Elafin from days 24 to 27 and from days 35 to 42. Wild-type LAB (5 × 109 CFU in 100 μl of PBS) and PBS (100 μl) were used as controls. Body weight was measured daily after the induction of colitis. On day 42 after the beginning of the DSS treatment, mice were sacrificed and colons were harvested for measurement of inflammation parameters: macroscopic damage score, bowel thickness, and elastolytic activity.

Spleen cells isolated from BALB/c mice (Janvier) were stained with PerCP (peridinin chlorophyll protein)–Cy5.5–conjugated anti-CD4 mAb (clone RM4-5, eBioscience), APC (allophycocyanin)–conjugated anti-CD25 mAb (clone PC61.5, eBioscience), and FITC (fluorescein isothiocyanate)–conjugated anti-CD45RB mAb (clone C363.16A, eBioscience). CD4+CD25CD45RBhigh and CD4+CD25+CD45RBlow T lymphocytes were separated by fluorescence-activated cell sorting. Colitis was induced in 7-week-old severe combined immunodeficient (SCID) mice (Charles River Laboratories) by intraperitoneal injection of 4 × 105 naïve CD4+CD25CD45RBhigh T lymphocytes in 100 μl of PBS. SCID mice injected with CD4+CD25+CD45RBlow were used as controls. Development of colitis (body weight and disease activity index) was monitored everyday after the onset of colitis to sacrifice (day 19). Mice were orally treated daily with 5 × 109 CFU (in 100 μl of PBS) of L. lactis wild-type or recombinant L. lactis (L. lactis Elafin) from day 19 to day 39. On day 39, mice were sacrificed and colons were harvested for measurement of inflammation parameters.

In situ elastolytic activity in humans and mice

Frozen OCT sections of colonic tissues from patients and mice (8-μm thickness) were rinsed with a washing solution (2% Tween-20) and incubated at 37°C overnight with Bodipy-FL-Elastin (0.5 μM) by the EnzChek Elastase Assay Kit (Invitrogen) according to a previously published protocol (32). All sections were visualized with the LSM710 microscope (Carl Zeiss France) and were analyzed by observers blinded of the treatments, with Zen 2009 software (Carl Zeiss).

Elastolytic activity in supernatants from human biopsies

Elastolytic activity was measured in supernatants with Bodipy-FL-Elastin (0.5 μM) as substrate. Samples (20 μl) were resuspended in 100 μl of buffer (50 mM tris-HCl, 500 mM NaCl, and 0.1% Triton X-100). The change in fluorescence (excitation, 485 nm; emission, 530 nm) was measured over 30 min at 37°C on a microplate reader NOVOstar (BMG Labtech) to calculate the rate of Elastin degradation. Data were represented as fold increase of activity detected in healthy control supernatants.

Elastolytic activity in colonic washes of mice

The entire colon of mice was excised, and 1 ml of PBS was instilled and washed twice through the lumen. Elastolytic activity was measured in those lumenal washes as previously described (2) with MeOsuc-AAPV-pNa (100 μM; Sigma) and MeOsuc-AAPV-AMC (100 μM; Sigma) as substrates.

Detection of Elafin

In situ detection of Elafin and GFP-tagged L. lactis. During DSS exposure, mice were orally treated daily with 5 × 109 CFU of recombinant LAB (either L. lactis–GFP wild-type or L. lactis–GFP Elafin, n = 5 in each group) for in situ localization of L. lactis bacteria and Elafin released. At day 7, colon tissue was cryoconserved and sections of 5 μm were incubated overnight at 4°C with anti-Elafin antibody (Santa Cruz Biotechnology) and GFP (Rockland). Slides were mounted and nuclei were stained with DAPI fluorescent mounting medium (Invitrogen) and analyzed on a confocal microscope (Zeiss LSM710, Carl Zeiss) by observers unaware of treatments. Images representative of each group were selected.

mRNA FISH. Colonic tissues from humans were paraffin-embedded for the detection of Elafin mRNA by mRNA FISH. Sense and antisense RNA probes against full-length Elafin complementary DNA were fluorescently labeled with RNA-RED kit Alexa Fluor 594 dye (Invitrogen) according to the manufacturer’s protocol. Slides were hybridized overnight at 42°C with RNA probes (400 μg/ml). Slides were mounted, and nuclei were stained with DAPI fluorescent mounting medium (Invitrogen) and imaged on a Zeiss LSM710 confocal microscope. The intensity of fluorescent signal specifically in mucosa was quantified by ImageJ software.

Western blot and ELISA. Proteins from bacterial cells or supernatants were precipitated by trichloroacetic acid/acetone and were separated by SDS–polyacrylamide gel electrophoresis (15%). After transfer, blots were exposed for 2 hours (room temperature) to anti-Elafin antibody (1/250) and visualized by chemiluminescence. ELISA was performed as previously described (7).

Measurement of inflammatory parameters and dosage of cytokines

Macroscopic damage, MPO, and bowel thickness were measured at the time of sacrifice and as previously described (14, 31).

Cytokines were dosed by Cytometric Bead Array (BD Biosciences). Raw values were normalized to tissue weight, and concentrations were quantified from standard curves of each cytokine designed in BD FCAP Array software.

For immunohistochemistry of immune cell markers, colonic tissues from mice were snap-frozen in OCT (Dako). Tissue sections of 6 μm were incubated for 1 hour with primary antibody against F4/80 (AbD Serotec) or Ly-6G (AbD Serotec) or CD3 (AbD Serotec). Slides were mounted and nuclei were stained with DAPI fluorescent mounting medium (Invitrogen) and analyzed on a Zeiss LSM710 confocal microscope. Specific fluorescence intensity was quantified with ImageJ software and was reported per unit surface of tissue delimited manually with DAPI staining. Histograms represented the mean of four different fields per animal (n = 4 animals per group) in two independent experiments. Data were represented as fold increase of signal intensity quantified in control group (water-drinking animals).

Paracellular permeability studies

Caco-2 cells (Molsheim, HTB-37) were grown as monolayers in Transwell plates (2 × 105 cells per well) (33). After 15 days of culture (transepithelial electrical resistance >350 Ω cm2), cells were apically exposed for 24 hours to TNF-α (20 ng/ml), TNF-α + L. lactis (107 CFU per well), and TNF-α + L. lactis Elafin (107 CFU per well) or exposed for 48 hours to HBSS (vehicle), human healthy controls or IBD supernatants (150 μl for 10 mg of biopsy), and supernatants supplemented with L. lactis (107 CFU per well) or L. lactis Elafin (107 CFU per well). After 2 hours of coculture, cells were additionally exposed to gentamicin (100 μg/ml) to limit bacterial development. Paracellular permeability was assessed by apical-to-basal transport of dextran-FITC (molecular weight, 3000; Sigma) after 4 hours of incubation (37°C in 5% CO2) as previously described (2).

Real-time PCR analysis

Caco-2 cells were plated at a density of 2 × 105 cells per well. After 24 hours (TNF-α) or 48 hours (IBD supernatants), monolayers were harvested in TRIzol reagent (Invitrogen). Total RNA extraction was performed according to the manufacturer’s protocol. Quantitative real-time PCR (2 μg of RNA) was performed on a LightCycler 480 (Roche) with gene-specific primers (table S2) and SYBR Green I Master (Roche Diagnostics). Relative expression of targeted genes was compared to hypoxanthine phosphoribosyltransferase expression.

Statistical analysis

Data were expressed as column bars representing the means ± SEM or vertical scatter plots with mean. Data were compared with adequate parametric or nonparametric t tests, one-way ANOVA (post hoc Bonferroni or Dunnett tests), and paired t test. Graphic design and statistical analyses were performed with the GraphPad Prism software version 5 for Windows. A P value less than 0.05 was considered significant.

Supplementary Materials

www.sciencetranslationalmedicine.org/cgi/content/full/4/158/158ra144/DC1

Materials and Methods

Fig. S1. Effects of Elafin-expressing, IL-10–expressing, and wild-type L. lactis on inflammation.

Fig. S2. Presence of L. lactis expressing Elafin in feces and tissue.

Table S1. Characteristics and outcomes of IBD patients.

Table S2. Primers used for quantitative RT-PCR studies.

References and Notes

  1. Acknowledgments: We thank ANEXPLO platforms (UMS 006) animal care facility and histopathology core facility (F. Capilla and A. Marrot), S. Allart and A. Canivet, the cellular imaging core facility (U1043), and G. Fournier and M. D. Hollenberg for helpful discussions. Funding: Supported by grants from Agence Nationale pour la Recherche (to N.V., C.D., and G.D.), the région Midi-Pyrénées (to J.-P.M.), and the Canadian Institute of Health Research (to N.V.). Author contributions: J.-P.M., L.G.B.-H., and C.D. performed experiments, data acquisition, analysis, and interpretation and helped with the drafting of the manuscript; L.M., G.D., J.B., C.R., P.R., K.C., and P.K. performed experiments, data acquisition, analysis, and interpretation; J.-P.V., L.A., and E.M. performed human biopsy collection, data acquisition, and analysis; J.-M.S. provided reagents and edited the manuscript; P.L. and N.V. designed the study, performed analysis and interpretation of data, helped with manuscript drafting and editing, and supervised the study. Competing interests: N.V., L.G.B.-H., P.L., and J.-M.S. hold the following patent associated with this work: Recombinant probiotic bacteria for the prevention and treatment of inflammatory bowel disease (IBD) and irritable bowel syndrome (IBS) [WIPO (World Intellectual Property Organization) Patent Application WO/2011/086172]. The other authors declare that they have no competing interests.
View Abstract

Navigate This Article