点击显示 收起
1 From the Division of Gastroenterology, Hepatology, and Nutrition, Centre de Recherche, Sainte Justine Hospital, and the Departments of Nutrition (FC, SB, MLM, and EL), Pediatrics (EGS), and Biochemistry (ED), Université de Montréal, Montreal, Canada, and the Department of Anatomy and Cell Biology, Université de Sherbrooke, Sherbrooke, Canada (CA).
2 Supported by grants from the Canadian Crohn’s and Colitis Foundation, Natural Sciences & Engineering Research Council of Canada, and the Dairy Farmers of Canada, as well as by Research Scholarship Awards from Le Fonds de la Recherche en Santé du Québec (to EGS and EL) and a Research Chair in Inflammatory Bowel Diseases from the Canadian Institutes of Health Research and the Crohn’s and Colitis Foundation of Canada (to EGS). 3 Reprints not available. Address correspondence to E Levy, Division of Gastroenterology, Hepatology, and Nutrition, Hôpital Sainte-Justine, 3175 Côte Sainte-Catherine Road, Montreal, Quebec H3T 1C5, Canada. E-mail: levye{at}justine.umontreal.ca.
ABSTRACT
Background: Membrane lipid peroxidation may play a role in immune-mediated bowel diseases.
Objective: We examined the effects of lipopolysaccharide (LPS), a ubiquitous endotoxin mediator of gram-negative bacteria, alone and in combination with iron-ascorbate, on enterocyte function. Furthermore, we assessed the antioxidant capacity of butylated hydroxytoluene (BHT) and butyric acid, which are known to play a significant role in the welfare of intestinal mucosa.
Design: Differentiated intestinal Caco-2 cells were used to study the induction of membrane peroxidation by LPS (100 µg/mL) and iron-ascorbate (0.2 and 2 mmol/L, respectively) and to examine the beneficial effects of BHT and butyric acid.
Results: A significant dose-dependent increase in malondialdehyde, accompanied by lower apical membrane fluidity and significantly decreased sucrase activity, was observed when Caco-2 cells were incubated with LPS. LPS also augmented paracellular permeability ([14C]polyethylene glycol flux), prostaglandin E2 production, and cyclooxygenase-2 (EC 1.14.99.1) expression. These abnormalities were exacerbated by the coadministration of iron-ascorbate, but most of them were suppressed by butyric acid and BHT.
Conclusion: Bacterial endotoxin and prooxidants may overwhelm antioxidant defenses and become deleterious to enterocyte function, whereas butyric acid and BHT may provide antioxidant protection.
Key Words: Iron-ascorbate • membrane fluidity • permeability • cyclooxygenase-2 • prostaglandin E2 • PGE2 • lipopolysaccharide • butylated hydroxytoluene • BHT • butyric acid • Caco-2 cells
INTRODUCTION
In addition to providing nutrient transport, the single layer of epithelial cells lining the mucosal surface imparts a functional barrier that protects against many deleterious microbial and chemical agents in the gut lumen (1). Enterocytes are also active participants in the mucosal immune response (2–7). Disruption of the host’s mucosal defense system may contribute to immune-mediated bowel disorders, such as chronic inflammatory bowel diseases (IBD) and multiple organ failure (8, 9). Tissue peroxidation is a major component of the aberrant immune response in IBD (8, 10–12). Oxygen radicals elicit membrane destabilization, alterations in DNA, inactivation of proteolytic enzymes associated with the detoxification of reactive oxygen metabolites, and induction of oxidative stress via interaction with plasma and cell factors (11). The resulting chronic intestinal inflammation promotes an imbalance between oxidant and antioxidant mechanisms at the tissue level (12), compromising circulating antioxidant reserves (13).
Excessive intestinal permeability and enteric bacterial flora have been implicated in the pathogenesis of IBD (8, 14). Lipopolysaccharide (LPS), a ubiquitous endotoxin mediator of gram-negative bacteria, can facilitate microbial translocation by a mechanism involving physical perturbation of the gut mucosal barrier (15, 16). The first aim of the present study was therefore to examine the effect of the potential interaction between free radical–induced lipid peroxidation and LPS on the integrity of the gastrointestinal mucosa. We determined the effect of LPS on intestinal epithelial permeability, transepithelial electrical resistance, fluidity, lipid peroxidation, and sucrase activity. The gastrointestinal mucosa is constantly exposed to luminal oxidants from ingested foods, desquamated cells, and bacteria (17–19). Consequently, our second aim was to evaluate the combined effects of LPS and an additional source of lipid peroxidation on the aforementioned membrane variables by using iron-ascorbate (20, 21). Short-chain fatty acids derived from dairy products and the anaerobic fermentation of carbohydrates by endogenous bacterial flora are key elements in the maintenance of human mucosal health and have potential clinical benefits in the treatment of IBD (22). Butyric acid, a short-chain fatty acid that is thought to play a significant role in the maintenance of mucosal health (23), was tested in terms of its capacity to scavenge reactive oxygen species generated by LPS or iron-ascorbate.
MATERIALS AND METHODS
Cell culture
Caco-2 cells (American Type Culture Collection, Rockville, MD) were grown at 37 °C in MEM (Gibco-BRL, Grand Island, NY) containing 1% (by vol) penicillin:streptomycin and 1% (by vol) MEM nonessential amino acids (Gibco-BRL) and supplemented with 10% (by vol) decomplemented fetal bovine serum (Flow, McLean, VA). Cells (passages 30–40) were maintained with 5% (by vol) CO2 in 75-cm2 flasks (T-75; Corning Inc, Acton, MA). When the cultures reached 70–90% confluence, each culture was split into 6 flasks containing 0.05% (by vol) trypsin and 0.5 mmol EDTA/L (Gibco-BRL). For individual experiments, cells were plated at a density of 1 x 106 cells/well on 24.5mm polycarbonate Transwell filter inserts with 0.4-µm pores (Costar, Cambridge, MA) in MEM (as above) supplemented with 5% (by vol) fetal bovine serum. The inserts were placed into 6-well culture plates, permitting separate access to the upper and lower compartments of the monolayers. Cultures were maintained for 20 d, a period in which we have observed that cells are differentiated and suitable for studies on enterocyte function, including lipid synthesis and lipoprotein secretion (24, 25). The medium was refreshed every second day.
Estimation of lipid peroxidation
Caco-2 cells were cultured in the presence or absence of LPS and Fe2+-ascorbate, which were added to the medium in the apical compartment. The amount of free malondialdehyde (MDA) formed during the reaction was determined by HPLC as described previously (26). Proteins were first precipitated with a 10% (by vol) sodium tungstate solution (Aldrich Chemical Co, Milwaukee). The protein-free supernatant fluids were then reacted with an equivalent volume of 0.5% (wt:vol) thiobarbituric acid solution (TBA; Sigma, St Louis) at 90 °C for 30 min. After cooling to room temperature, the pink chromogene [(TBA)2-MDA] was extracted with 1-butanol and dried over a stream of nitrogen at 37 °C. The dry extract was then resuspended in a potassium dihydrogen phosphate:methanol mobile phase (70 parts:30 parts, pH 7.0) before MDA detection by HPLC.
Effect of antioxidants
To determine whether lipid peroxidation was responsible for the alterations caused by LPS or iron-ascorbate, butylated hydroxytoluene (BHT) was added to the apical compartment for 1 h before incubation with these factors. BHT was dissolved in alcohol (1% in the medium). Control wells received an equal volume of the alcohol carrier. The efficiency of butyric acid to prevent or reduce lipid peroxidation induced by iron-ascorbate or LPS was also tested after preincubation for 20 h.
Membrane integrity and viability of Caco-2 cells
The transepithelial membrane resistance of Caco-2 cells was assessed after exposure to LPS or Fe2+-ascorbate with a Millicell-ERS apparatus (Millipore Corp, Bedford, MA). Apical membrane sucrase activity was also measured as described previously (25, 27). [14C]Polyethylene glycol flux was evaluated after the addition of the labeled probe (1 µCi/well) to the apical compartment (28). Samples were then mixed with Ready Safe Counting fluid (Beckman, Fullerton, CA), and radioactivity was quantified by scintillation counting (Beckman LS 5000 TD; Beckman, Mississauga, Ontario, Canada). Cellular protein content was determined according to the method of Lowry et al (29). Cell viability was assessed by trypan blue exclusion (30).
Isolation of brush border membranes
Brush border membranes were purified from Caco-2 cell homogenates according to the method of Schmitz et al (31). Briefly, the culture medium was removed, and the cells attached to filters were rinsed twice with phosphate-buffered saline (Gibco-BRL). The cells were then scraped off and homogenized in tris-HCl mannitol [2 mmol HCl/L, 50 mmol mannitol/L (pH 7.0)]. After the addition of a 10-mmol MgCl2/L solution, the homogenates were centrifuged for 15 min at 7700 x g and 4 °C. The supernatant fluids were subsequently centrifuged for 30 min at 20 000 x g and 4 °C, and the resulting pellets were used for the determination of fluidity.
Cell membrane fluidity
Aliquots of brush border membranes were diluted with a solution containing 300 mmol mannitol/L, 10 mmol trisHEPES/L, and 100 mmol KCl/L (pH 7.5). Fluidity was estimated by incorporating the fluorescent probe 1-(4-trimethylammoniumphenyl)-6-phenyl-1,3,5-hexatriene p-toluene-sulfonate and subsequently measuring polarization (26). The final molar probe-to-lipid ratio was 1:1000. Fluorescence was measured in a spectrofluorometer at 25 °C with polarization filters parallel and perpendicular to the excitation beam. Excitation was measured at 360 nm, and emission was measured at 420 nm.
Prostaglandin E2 determination
Caco-2 cells were preincubated with sodium butyrate or BHT and exposed to Fe2+-ascorbate or LPS. Cellular prostaglandin E2 (PGE2) was measured by enzyme-linked immunosorbent assay (Neogen Corp, Lansing, MI).
Immunoblot analysis of cyclooxygenase-2
After the incubation with the antioxidants and prooxidants, Caco-2 cells were washed twice with phosphate-buffered saline and scraped in 0.5 mL lysis buffer [50 mmol tris-HCl/L (pH 7.5), 150 mmol NaCl/L, 5 mmol EDTA/L, 0.1% (by vol) sodium dodecyl sulfate, 0.5% (by vol) sodium desoxycholate, 1% (by vol) Triton (Sigma, Oakville, Ontario, Canada), 1 mmol phenylmethylsulfonyl fluoride/L, 1 mmol BHT/L, and 1 mmol pepstatin/L] as described previously (32). Cells were sonicated and aliquots of cell homogenates were mixed with loading buffer (Bio-Rad Corp, Hercules, CA). Cyclooxygenase-2 (COX-2; EC 1.14.99.1) values were determined by sodium dodecyl sulfate-polyacrylamide and transfer onto Hybond nitrocellulose membranes (Amersham, Baie d’Urfé, Québec, Canada), followed by detection with an enhanced chemiluminescence system for antigen-antibody complexes (33). Quantification was carried out by densitometry (Scion Image; Scion Corp, Frederick, MD).
Statistical analysis
All values are expressed as means ± SEMs. Data were analyzed by using a one-way analysis of variance and the two-tailed Student’s t test. The Tukey test was used for comparisons among several groups. A P value < 0.05 was considered significant. Results were analyzed by using BMDP NEW SYSTEM for WINDOWS (version 1.0; BMDP Statistical Software, Inc, Los Angeles).
RESULTS
Membrane lipid peroxidation
LPS facilitates the binding of bacteria to cell surfaces and the delivery of bacterial toxins. Short-chain fatty acids are major byproducts of anaerobic bacterial metabolism that are able to diffuse across biological membranes (22). Thus, differentiated Caco-2 cells were used to study the induction of membrane peroxidation by LPS and the potential protective role of butyric acid. As shown in Figure 1, when Caco-2 cells were incubated with various concentrations of LPS for 24 h, the MDA content in the apical compartment and in the cells was significantly higher than when cells were not incubated with LPS. The exposure of Caco-2 cells to butyric acid did not significantly alter MDA production. The combination of LPS (100 µg/mL) and iron-ascorbate (0.2 and 2 mmol/L, respectively), an established oxygen free radical–generating system, significantly augmented MDA concentrations (Figure 2). The addition of butyric acid significantly decreased LPS and iron-ascorbate–mediated lipid peroxidation, which was totally suppressed in Caco-2 cells and significantly reduced in the apical compartment by the administration of BHT (200 µmol/L).
FIGURE 1. . Mean (± SEM) malondialdehyde (MDA) concentrations (for 3 different experiments, each of which was carried out in triplicate) in Caco-2 cells (A and B) and in the apical membrane compartment (C and D) after incubation of differentiated Caco-2 cells with increasing concentrations of lipopolysaccharide (LPS) or butyric acid for 24 h at 37 °C. *Significantly different from control wells in which LPS or butyric acid was not added, P < 0.05.
FIGURE 2. . Mean (± SEM) malondialdehyde (MDA) concentrations (for 3 different experiments, each of which was carried out in triplicate) in Caco-2 cells (A) and in the apical membrane compartment (B) after differentiated Caco-2 cells were preincubated with butyric acid, butylated hydroxytoluene (BHT), or both for 24 h before being incubated with lipopolysaccharide (LPS; 100 µg/mL), iron-ascorbate (Fe-Asc; 0.2 and 2 mmol/L, respectively), or both for an additional 24 h. a,bSignificantly different from control wells: aP < 0.05, bP < 0.01. c,dSignificantly different from LPS: cP < 0.05, dP < 0.001. eSignificantly different from LPS+Fe-Asc, P < 0.05.
Functional integrity and viability of Caco-2 cells
The effect of LPS on Caco-2 membrane integrity and cell viability was assessed by monolayer transepithelial resistance, sucrase activity, protein content, and trypan blue exclusion after an incubation period of 24 h (data not shown). Transepithelial resistance and protein content remained unchanged with increasing concentrations of LPS. Similarly, cell viability was unaffected by the addition of LPS (data not shown). Only sucrase activity moderately, but significantly, decreased in response to LPS, at concentrations between 25 and 200 µg/mL. Butyric acid did not result in significant alterations in any of these variables (data not shown). Further experiments showed that the combination of LPS and iron-ascorbate did not amplify the derangements observed with the administration of LPS alone. However, butyric acid neutralized the decrease in sucrase activity elicited by LPS or LPS with iron-ascorbate (data not shown).
Membrane fluidity
Experiments were carried out to determine the effects of LPS, iron-ascorbate, and butyric acid on membrane fluidity, indicative of the relative motional freedom of lipid molecules in the membrane bilayer. The degree of fluidity of surface membrane lipids can be estimated by fluorescence polarization analyses with the probe 1,6-diphenyl-1,3,5-hexatriene. Fluidity was thus determined as the reciprocal of 1,6-diphenyl-1,3,5-hexatriene fluorescence polarization in apical membranes isolated from Caco-2 cells exposed to LPS, iron-ascorbate, or butyric acid (data not shown). A trend toward a reduction in membrane fluidity was noted with LPS alone. Iron-ascorbate alone, or combined with LPS, significantly decreased membrane fluidity. Butyric acid and BHT restored membrane fluidity values to those of the control cells in the presence of LPS and iron-ascorbate.
Paracellular membrane permeability
Polyethylene glycol uptake is considered to be a sensitive tool to estimate the paracellular permeability of the intestinal epithelium. To determine whether LPS or iron-ascorbate affects tight junctions, [14C]polyethylene glycol was added to the apical compartment of Caco-2 cell cultures. Caco-2 cells incubated with LPS or iron-ascorbate alone showed a trend toward higher [14C]polyethylene glycol uptake than that in cells not incubated with LPS or iron-ascorbate (Figure 3). The combination of these 2 factors resulted in a significant increase in permeation flux. This alteration decreased with the administration of butyric acid, BHT, or the mixture of butyric acid and BHT.
FIGURE 3. . Mean (± SEM) [14C]polyethylene glycol (PEG) permeation (for 3 different experiments) in differentiated Caco-2 cells that were or were not pretreated with butyric acid (2.5mmol/L), butylated hydroxytoluene (BHT; 200 µmol/L), or both for 24 h before incubation with [14C]PEG in the presence or absence of lipopolysaccharide (LPS; 100 µg/mL), iron-ascorbate (Fe-Asc; 0.2 and 2 mmol/L, respectively), or both for an additional 24 h at 37 °C. a,bSignificantly different from control wells: aP < 0.04, bP < 0.01.
Prostaglandin E2 production and cyclooxygenase protein expression
COX-2 is a highly regulated enzyme that catalyzes the production of prostaglandins under physiologic and pathologic conditions (34). As shown in Figure 4, Caco-2 cells that were incubated with LPS or iron-ascorbate had significantly higher PGE2 concentrations than did those that were not incubated with LPS or iron-ascorbate. The most significant change was obtained with the combination of LPS and iron-ascorbate. PGE2 concentrations returned to normal with the coadministration of butyric acid and BHT. COX-2 protein concentrations were examined by Western blot (Figure 5). Caco-2 cells that were incubated with LPS, iron-ascorbate, or the 2 prooxidants combined had higher COX-2 values than did those that were not incubated. Neither butyric acid nor BHT alone decreased COX-2 protein values after incubation of Caco-2 cells with LPS and iron-ascorbate. However, decreased COX-2 values were noted with the coadministration of butyric acid and BHT.
FIGURE 4. . Mean (± SEM) prostaglandin E2 (PGE2) concentrations measured by enzyme-linked immunosorbent assay (for 3 different experiments) in differentiated Caco-2 cells that were or were not pretreated with butyric acid (2.5 mmol/L), butylated hydroxytoluene (BHT; 200 µmol/L), or both for 24 h before incubation with lipopolysaccharide (LPS; 100 µg/mL), iron-ascorbate (Fe-Asc; 0.2 and 2 mmol/L, respectively), or both for an additional 24 h at 37 °C. a–cSignificantly different from control wells: aP < 0.04, bP < 0.01, cP < 0.005.
FIGURE 5. . Cyclooxygenase-2 (COX-2) expression analyzed by Western blot (of 3 culture experiments) in differentiated Caco-2 cells that were or were not pretreated with butyric acid (2.5 mmol/L), butylated hydroxytoluene (BHT; 200 µmol/L), or both for 24 h before incubation with lipopolysaccharide (LPS; 100 µg/mL), iron-ascorbate (Fe-Asc; 0.2 and 2 mmol/L, respectively), or both for an additional 24 h at 37 °C. Mean (± SD) scanning values (in arbitrary units) were as follows: control, 7438 ± 668; Fe-Asc, 8310 ± 845; LPS, 9692 ± 600; Fe-Asc+LPS, 12 405 ± 256; butyric acid, 7616 ± 128; Fe-Asc+LPS+butyric acid, 11 686 ± 234; butyric acid+BHT, 7910 ± 44; and Fe-Asc+LPS+butyric acid+BHT, 7979 ± 44.
DISCUSSION
Intestinal epithelial cells are a key component to the physiologic barrier between the myriad of toxins, microbes, and antigens present in the gut lumen and the underlying mucosal immune system (1, 2). The activated immune cells residing in the lamina propria and interspersed between epithelial cells have traditionally been considered to be the key effector cellular elements in the host’s mucosal immune response. However, there is increasing evidence that enterocytes are actively involved as full participants in generating an immune response at the mucosal level and are not just the targets of inflammation (35). The major purpose of the present set of experiments was to ascertain how microbial and peroxidative factors alter intestinal epithelial cell barrier function. Exposure of Caco-2 cells to LPS or Fe2+-ascorbate resulted in marked changes in MDA concentrations, membrane lipid fluidity, sucrase activity, and paracellular permeability. Furthermore, these agents induced changes in PGE2 concentrations and COX-2 protein expression. Butyric acid and BHT restored mucosal barrier function and the activity and protein expression of COX-2 to normal values.
Gram-negative bacterial species constitute the main source of endotoxin in the gut microflora (34). LPS is a potent stimulus of the host’s immune response via its capacity to induce the proinflammatory cytokine cascade. LPS characteristically activates neutrophils, monocytes, and macrophages via its CD14 membrane receptor (34). An increasing body of evidence points to the direct effect of LPS on intestinal epithelial cells via toll-like receptors (36). LPS can induce cell proliferation (37), modulate energy metabolism (38), and augment the production of inflammatory cytokines by enterocytes (39). The results outlined in the present report provide evidence that LPS is also able to induce enterocyte lipid peroxidation and concomitantly alter the membrane lipid fluidity, paracellular permeability, and sucrase activity of enterocytes.
Zareie et al (40) described an in vitro model of inflammation in which coculture of confluent monolayers of human T84 intestinal epithelial cells with LPS-activated monocytes resulted in significant abnormalities in epithelial ion transport and barrier function. Because the antioxidant BHT prevented these effects in our experiments, it is reasonable to suggest that impairment of membrane integrity is caused by LPS-mediated lipid peroxidation. A potential mechanism may involve peroxidative attack on polyunsaturated fatty acids in the membrane, altering the membrane phospholipid environment, which, in turn, may influence membrane function. No matter what the initiating event in inflammatory bowel diseases may be, the driving oxidative force provided by luminal microorganisms may damage the intestinal mucosa by means of the microorganisms’ products, such as LPS. The resulting breach of epithelial barrier function would facilitate the translocation of bacterial antigens and the production of proinflammatory mediators by the host’s mucosal cells.
An important issue is whether the luminal endotoxins released from bacteria belonging to the normal gut flora or from enteropathogenic microorganisms interact with ingested prooxidants to amplify epithelial barrier perturbations. An extensive body of literature documents the ability of combined iron and ascorbate molecules to initiate substantial lipid peroxidation in native and reconstituted biological membrane systems. Our previous studies (20, 21) showed the oxidative potential of iron-ascorbate to modify the membrane compartment of enterocytes and hepatocytes, thereby impairing their lipid transport and metabolism. In the present set of experiments, Fe2+-ascorbate effectively induced Caco-2 cell lipid peroxidation, as evidenced by high concentrations of MDA, a commonly used indicator of lipid peroxidation. Moreover, the prooxidant combination of LPS and Fe2+-ascorbate caused synergistic derangements to membrane integrity. Our data suggest that further clinical trials should be undertaken to interfere with the production of peroxidative stress in patients with IBD (41).
The therapeutic role of short-chain fatty acids in various intestinal disorders is still under evaluation. Their use is well established in diversion colitis, providing evidence that the shortage of intraluminal short-chain fatty acids is associated with long-standing inflammation (42, 43). In keeping with this, our findings indicate that butyric acid may help overcome oxidative stress induced by LPS, thus preventing gut injury. Additional studies are needed to elucidate the mechanisms involved in butyric acid–induced repair processes, in view of its multiple effects on the intestinal mucosa. In addition to being an important nutrient, butyric acid also acts as a cofactor for the regulation of transcription proteins involved in gene expression and interacts with G-proteins at the membrane-cytoplasm interface (44). Furthermore, butyric acid has a marked trophic effect on normal colonic mucosa, stimulates epithelial proliferation through neural and hormonal mechanisms, and augments mucosal weight, DNA content, and crypt length in normal bowel. The opposite effects are induced in colonic neoplastic cell lines, with arrest of the cells in the G1 phase and inhibition of DNA synthesis.
Cyclooxygenase exists as 2 isoenzymes, COX-1 and COX-2. Compelling evidence suggests that COX-1 synthesizes prostaglandins involved in the regulation of normal cell activity, whereas COX-2 seems to primarily produce prostaglandins at sites of inflammation (45). Studies showed that oxidative status correlates with increases in the COX-derived prostanoid biosynthetic process (46). In the present study, we found that LPS enhanced COX-2 activity and protein values, which were accompanied by higher MDA concentrations and abnormalities in membrane integrity. Interestingly, butyric acid and BHT reduced lipid peroxidation and simultaneously suppressed COX-2 activity and protein values. Taken together, our data suggest that the prooxidant and antioxidant effects of LPS and butyric acid, respectively, may directly influence COX-2.
REFERENCES