Degraded Carrageenan Causing Colitis in Rats Induces TNF Secretion and ICAM-1 Upregulation in Monocytes through NF-κB Activation Carrageenan Activates Monocyte 

Abstract
Carrageenan (CGN) is a high molecular weight sulphated polysaccharide derived from red seaweeds. In rodents, its degraded forms (dCGN) can induce intestinal inflammation associated with macrophage recruitment and activation. The aim of this study was: 1) to analyze the size-dependent effects of dCGN on colon inflammation in vivo, and 2) to correlate these effects with monocyte/macrophage proliferation, cytokine production and expression of various cell surface antigens including ICAM-1 adhesion molecule. Peripheral blood monocytes (PBM) and THP-1 monocytic cells were cultured in the presence of either 10 or 40 kDa, dCGN. The 40 kDa, but not the 10 kDa dCGN, induced colitis in in vivo. Degraded CGN inhibited THP-1 cell proliferation in vitro, arresting the cells in G1 phase. In addition, dCGN increased ICAM-1 expression in both PBM and THP-1 cells with a major effect seen after 40 kDa dCGN exposure. Also, dCGN stimulated monocyte aggregation in vitro that was prevented by incubation with anti-ICAM-1 antibody. Finally, dCGN stimulated TNF-α expression and secretion by both PBM and THP-1 cells. All these effects were linked to NF-κB activation. These data strongly suggest that the degraded forms of CGN have a pronounced effect on monocytes, characteristic of an inflammatory phenotype.

Introduction
Carrageenan (CGN) is a high molecular weight sulphated polysaccharide (>200 kDa) derived from red algae (Rhodophyceae). Three main forms of CGN have been identified: kappa, iota, and lambda. They differ from each other in sulphation degree and solubility [1], [2]. Native CGN is thought to be harmless and is widely used as a food additive to improve texture. It is also used in cosmetics and pharmaceuticals. However, acid treatment at high temperature (80°C) triggers CGN hydrolysis to lower molecular weight (<50 kDa) compounds known as poligeenan or degraded CGN (dCGN). These dCGNs induce inflammation and have been widely used as models of colitis in several species, including rats [3], rabbits [4] and guinea pigs [5]. The role of dCGN as a tumor-promoting factor remains controversial [4], [6]–[8].
Although the native form is thought to be harmless for human consumption, small amounts of dCGN are probably produced by acid hydrolysis during gastric digestion [9], [10] or interaction with intestinal bacteria [11], [12]. Whereas the effects of native and dCGN on intestinal inflammation have been extensively analyzed in animal models, only few studies have been conducted using human cell lines. Recent studies have shown a link between exposure to native form CGN and IL-8 production by the human intestinal epithelial cell line, NCM460, via Nuclear Factor-κB (NF-κB) activation [13], [14]. NF-κB is a transcription factor that regulates the expression of genes associated with inflammation [15], [16].
Macrophage infiltration and accumulation is a common characteristic of intestinal diseases [17]. Macrophages represent 10% of total lamina propria cells, secrete a wide range of biologically active compounds and express cell-adhesion molecules. The immune cell response to an inflammatory stimulus seems to be amplified or directly generated by cells exposed to sulphated polysaccharides such as carrageenans. Indeed, inflammation induced by dCGN was associated with recruitment of macrophages to inflammation sites [18], [19]. Also, inflammation induced by Dextran Sulphate Sodium (DSS), another sulphated compound, was directly associated with macrophages recruitment [20], since DSS still provoked inflammation after T-lymphocyte and NK cell depletion [20]. Although inflammation can be induced by dCGN, there are no data on human monocyte responses to dCGN exposure. Therefore, to investigate the effects of dCGN on human monocytes, normal Peripheral Blood Monocytes (PBM) and tumoral monocyte/macrophage THP-1 cells were exposed to 10 kDa and 40 kDa dCGN. We found that dCGN inhibited THP-1 cell proliferation in vitro, increased ICAM-1 expression, stimulated ICAM-1-dependent monocyte aggregation, and stimulated TNF-α expression and secretion. These responses were more pronounced after 40 kDa dCGN exposure and were linked to NF-κB activation. In addition, the 40 kDa dCGN, but not the 10 kDa dCGN induced in vivo colitis as shown by the inflammatory response in the rat colon. These results suggest that the degraded forms of CGN have an important effect on monocytes resulting in an inflammatory phenotype.

Materials and Methods

Preparation of Degraded Carrageenan
Two preparations of degraded carrageenan with low, (∼10 kDa; C10), and medium, (∼40 kDa; C40) molecular weight were prepared from native iota-carrageenan extracted from Euchema spinosum (generously provided by Sanofi Biosystems Industry, Boulogne-Billancourt, France). Native carrageenan was dissolved in distilled water (5% w/v) under vigorous stirring and heated to 60°C. Then, the carrageenan solution was submitted to two different treatments to obtain both low and medium molecular weight fractions. Briefly, for the low molecular weight fraction, carrageenan solution was hydrolyzed with 0.3% (v/v) concentrated sulphuric acid for 15 min at 80°C. After neutralization with NaOH 4N, the solution was ultra filtered through a hollow fibre cartridge with MW cut-off 5 kDa, (Amicon Inc, Beverly, USA). For the medium molecular weight fraction, the carrageenan solution was hydrolyzed with 0.3% (v/v) concentrated sulphuric acid for 30 min at 60°C. After neutralization, the supernatant was ultra filtered (MW cut-off 100 kDa). The filtrate was submitted to a second ultra filtration (MW cut-off 5 kDa). Both preparations of dCGN were precipitated with 4 volumes of 95% ethanol, dried at room temperature and ground to small particles (1 mm in diameter). Using gel-permeation chromatography in combination with light scattering measurements (see Viebke et al. [21]), it was confirmed that the low fraction had an average molecular weight of 10 kDa, and the medium fraction of 40 kDa. The sulphate content of polysaccharides in both fractions was measured following the method of Quemener et al. [22]. Finally, the absence of polysaccharide structure modifications in the two fractions was confirmed using 2H-NMR spectroscopy. The absence of LPS contamination in the two fractions was confirmed using the e-Toxate® kit (Sigma, St Quentin Fallavier, France). Before use in cell culture, the two fractions were dissolved in complete medium during 30 min at 56°C.

Animals, Chemicals and Diet
Male Wistar rats (150 g average weight) were housed under standard conditions and fed ad libitum with standard rodent laboratory chow. Degraded iota-carrageenans were administered in the drinking water (5% w/v) for 55 days to 2 groups of six animals each. The first group received the low molecular weight carrageenan (10 kDa dCGN) and the second received the medium molecular weight carrageenan (40 kDa dCGN). An additional group of four rats were maintained on regular tap water (control group). To increase palatability 0.2% sucrose was added to the drinking water of all groups (Van der Waaji et al., [23]). Fresh carrageenan solutions were prepared daily.

Evaluation of Colitis
Body weight, liquid and food consumption, diarrhea and rectal bleeding (detected by eye inspection) were recorded throughout the feeding period. After 55 days, animals were sacrificed by cervical dislocation. The length of the colon was measured as described by Okayashu et al. [24]. Then, each colon was ligated in sections of 2 cm and 1 to 2 ml of 10% formalin was infused into the intestinal lumen. The moderately distended segment was sectioned and fixed in 10% formalin. The following day, the intestinal content was removed by vortexing. The fixed segment was kept in 10% formalin at 4°C until the paraffin embedding procedure. To evaluate the degree of inflammation, this segment of colon was opened longitudinally and macroscopic and histological scores of inflammation were recorded as previously described [25], [26]. The toluidine blue staining was used for identification of sulphated polysaccharides in the intestinal mucosa. On the day of sacrifice, a fresh sample of each colon (50 mg) was collected for myeloperoxidase (MPO) assay according to Krawisz et al., [27]. The level of MPO, mainly expressed by neutrophils, indicates the rate of recruitment of neutrophils to the intestinal mucosa. One unit of MPO activity corresponds to the degradation of 1 µmol of peroxide per minute at 25°C.

Cell Culture
All tissue culture reagents were from Invitrogen (Cergy Pontoise, France). THP-1 human monocytic cells were maintained in RPMI-1640 supplemented with 10% FCS, 2 mM L -glutamine, 50 U/ml penicillin and 50 mg/ml streptomycin at 37°C in a 5% CO2 incubator. Human peripheral blood mononuclear cells were obtained from heparinized blood by Ficoll-Hypaque density gradient. Monocytes were then isolated by adherence to culture flasks as described [28]. For cell aggregation, monocytes were cultured in the presence or absence of C10 or C40 for 72 h. Cell colonies were monitored under an inverted phase contrast microscope coupled through a video camera to a computer. In some wells, neutralizing monoclonal antibody to ICAM-1 (2.5 µg/ml) (Tebu, Le Perray en Yvelines, France) was added.

Cell Cycle Analysis
THP-1 cells in exponential growth phase were exposed to complete medium in the presence or absence of carrageenans for 24 h before being stained with propidium iodide using the DNA-Prep Coulter kit according to the manufacturer's instruction (Beckman-Coulter, Villepinte, France). Cell DNA content was then analyzed by flow cytometry using an EPICS XL2 (Beckman-Coulter). Raw data for the distribution of DNA content of 30,000 cells retrieved from the cytometer were expressed as the percentage of G0/G1 through G2/M populations. Multicycle AV software (Phoenix Flow Systems, San Diego, CA) was used to generate DNA content frequency histograms and facilitate data analysis.

Cell Surface Antigen Expression Analysis
Peripheral Blood Monocytes or THP-1 cells were exposed to complete medium in the presence or absence of carrageenan for 36 h. After two washes in PBS without Ca2+ and Mg2+, cells were incubated in PBS containing 0.1% gelatin and 8% AB human serum to prevent binding to Fc receptors. Then, 5×105 cells were incubated with primary antibodies at 4°C for 30 min. Two other washes in PBS preceded incubation with FITC-conjugated goat antibody anti-mouse IgG diluted 1/1000 at 4°C for 30 min (Tebu). After two additional washes, analysis of stained cells was performed on an EPICS XL2 (Beckman-Coulter). The cell population was gated according to its forward and wide-angle light scattering. Data were expressed as mean relative fluorescence intensity (MFI) of 3000 cells.

TNF Activity Bioassay
Monocytes or THP-1 cells were cultured with or without different concentrations of CGNs or LPS (Salmonella typhosa, Sigma) for 24 h or the indicated time. Biologically active TNF-α/β in tissue culture supernatant was measured using the WEHI 164 clone 13-cell killing assay [29]. TNF concentrations are expressed as pg/ml.

RT-PCR Analysis
Total RNA from monocytes was isolated using TRIzol Reagent™ (Invitrogen). cDNA was generated on 1 µg of total RNA in a reaction volume of 20 µl, using M-MLV reverse transcriptase (Invitrogen). PCR was done in the linear range of amplification (determined for each primer pair-cDNA combination). Standard PCR reactions were performed with 1 µl of the cDNA solution, 50 µM of each primer solution, 10 mM of each dNTP, 25 mM MgCl2, 10X Goldstar DNA polymerase reaction buffer, and 0.5 units of Goldstar DNA polymerase (Eurogentec, Seraing, Belgium). First PCR cycle consisted of 1 min at 92°C, 1 min at 58°C and 1 min at 72°C; then each PCR cycle consisted of 40 sec at 92°C, 40 sec at 58°C and 50 sec at 72°C. cDNA for β-actin was amplified for 28 cycles using the oligos: sense 5′-GGCATCGTGATGGACTCCG-3′ and antisense 5′GCTGGAAGGTGGACAGCGA-3′. cDNA for TNF-α was amplified for 35 cycles using the oligos: sense 5′-AAGCCTGTAGCCCATGTTGT-3′ and antisense 5′-CAGATAGATGGGCTCATACC-3′. cDNA for ICAM-1 was amplified for 35 cycles using the oligos sense 5′-GTAGCAGCCGCAGTCATAATGG-3′ and antisense 5′-A TGCTGTTGTATCTGACTGAGG-3′.

NF-kB Transcription Reporter Gene Assay
The plasmid 3XMHC-luc (a generous gift from Drs. J. Westwick and D.A. Brenner, University of North Carolina, Chapel Hill) contains three copies of NF-κB-responsive element from the MHC class I locus, placed upstream of the luciferase gene. Human monocytic THP-1 cells were transiently transfected as previously described [30], and then cultured for 4 h alone or with increasing concentration of either C10 or C40. Luciferase activity was determined using a luminometer (Monolight 2010 Luminometer, Ann Arbor, MI).

Western Blot Analysis
THP-1 cells were stimulated for various lengths of time with 0.1 mg/ml C10 or C40, or 10 µg/ml LPS. Cells were then pelleted, washed and homogenised in lysis buffer (10 mM Hepes, pH 7.9, 150 mM NaCl, 1 mM EDTA, 0.6% NP-40, and 0.5 mM PMSF) on ice. Homogenates were sonicated, centrifuged at 10,000 rpm to remove cellular debris, and supernatant collected. Protein concentration was determined using the DC Protein Assay (Bio-Rad). Proteins in samples (15 µg total proteins) were resolved in a denaturing 12% polyacrylamide gel and transferred to a nitrocellulose membrane. I-κBα protein was detected using a rabbit polyclonal antibody (Santa Cruz Biotechnology, CA) followed by a horseradish peroxidase-coupled goat polyclonal antibody against rabbit Ig (Caltag Laboratories). Finally, IκB bands were revealed using the ECL™ detection system (Amersham Pharmacia Biotech, Les Ullis, France) according to the manufacturers' instruction. Antibody to α-Tubulin (Santa Cruz) was use as loading control.
For nuclear NF-κB, THP-1 cells were stimulated with 1 mg/ml C10 or C40 for 30 minutes at 37°C. Cells were then pelleted and nuclei separated as described [31]. Nuclei were washed and homogenized directly in loading (Laemli) buffer and heated for 5 minutes at 100°C. Proteins in samples were resolved in a denaturing 8% polyacrylamide gel and transferred to a polyvinylidine fluoride (PVDF) membrane (Immobilon-P; Millipore, Bedford, MA). Membranes were incubated in blocking buffer (1% BSA, in PBS) for two hours at room temperature. Membranes were subsequently probed with the corresponding antibody in blocking buffer, overnight. Rabbit polyclonal antibody anti-NF-κB p50 subunit (# sc-114) or anti-NF-κB p65 subunit (# sc-109) from Santa Cruz Biotechnology were used. Membranes were washed six times in PBS with 0.05% Tween 20, 5 minutes each time, and incubated with a 1/3000 dilution of HRP-conjugated F(ab')2 goat anti-rabbit IgG in 5% nonfat dry milk and 0.05% Tween 20 in PBS for 1 hour at room temperature. After washing six more times in PBS with 0.05% Tween 20, antibody-reactive proteins were detected using a chemiluminescence substrate (SuperSignal; Pierce, Rockford, IL) according to the manufacturer's instructions. To confirm that equivalent amounts of protein were loaded in each line, membranes were also Western blotted for ERK as described [32].

Analysis of NF-κB Activation by Flow Cytometry
Nuclear activation of NF−κΒ by flow cytometry was performed as described [31].

Statistical Analysis
The results were expressed as the mean value ± S.E.M. of individual experiments. The statistical significance of the differences between mean values was assessed by the Student's t-test and analysis of variance (ANOVA).

Results

Degraded CGN Induce Colonic Inflammation
All rats developed diarrhea during degraded carrageenan administration and gross evidence of blood was frequently detected in the stools. Colon length dramatically decreased in all treated rats with a more pronounced effect being observed in the 40 kDa dCGN treated group (Fig. 1A). Furthermore, prolonged exposure to 40 kDa dCGN resulted in high macroscopic and histological scores of inflammation (Fig. 1B, C). Only weak myeloperoxidase activity was detected in both control and dCGN-treated groups (Fig. 1D), indicating that granulocytes did not play a major role in the inflammation at that stage. Histological examination revealed various degrees of mucosal inflammation. Rats treated with 10 kDa dCGN showed edema, epithelium atrophy and slight lymphocyte infiltration (data not shown). These symptoms were totally absent in the colon of control rats (Fig. 1E). More severe mucosal injuries including ulceration, hyperplastic epithelium, crypt distortion and a strong macrophage infiltration, were observed in the 40 kDa dCGN-treated rats (Fig. 1F).