Carrageenan-induced NFkappaB activation depends on distinct pathways mediated by reactive oxygen species and Hsp27 or by Bcl10

Abstract

Carrageenans are highly sulfated polysaccharides that are widely used as food additives due to their ability to improve food texture. They are also widely recognized for their ability to induce inflammation in animal models of colitis. Recently, we reported that carrageenan (CGN) activated a pathway of innate immunity in human colonic epithelial cells mediated by Bcl10 (B-cell CLL/lymphoma 10). However, increases in phospho-IkappaBalpha and Interleukin-8 (IL-8) were not completely inhibited by silencing Bcl10, suggesting that CGN also influenced another mechanism, or mechanisms, of inflammation. In this report, we demonstrate that CGN increases production of reactive oxygen species (ROS) in human colonic epithelial cells. The combination of ROS quenching by the free radical scavenger Tempol and of Bcl10 silencing by siRNA completely inhibited the CGN-induced increases in nuclear NFkappaB (p65), phospho-IkappaBalpha, and secretion of IL-8. The CGN-induced increase in ROS was associated with declines in phosphorylation of MAPK 12 (p38gamma), MAPK 13 (p38delta), and heat-shock protein (Hsp) 27. The CGN-induced decline in phospho-Hsp27 was reversed by co-administration of Tempol (100 nM), but unaffected by silencing Bcl10. Since Hsp27 phosphorylation is inversely associated with phosphorylation of the IkappaBalpha kinase (IKK) signalosome, CGN exposure appears to affect the IKK signalosome by both the catalytic component, mediated by ROS-phospho-Hsp27, and the regulatory component, mediated by Bcl10 interaction with IKKgamma (Nemo). Hence, the CGN-activated inflammatory cascades related to innate immunity and to generation of ROS may be integrated at the level of the IKK signalosome.

Figure 1. Increased ROS production following CGN exposure in IEC

NCM460 cells (1A) and normal primary colonocytes (1B) in culture were treated with CGN (1μg/ml) for 1, 4 and 24 hours and reactive oxygen species (ROS) production was measured by a fluorometric assay using 10 μM hydroethidine (HE) as substrate. CGN stimulated ROS production within 1 hour; 4 and 24 hour values were similar, with levels 8-9 times the control. Differences between the control and treated cells at the same time point are statistically significant (p<0.001, 1-way ANOVA with Tukey's post-test, n=3).

Figure 2. CGN-induced increase in ROS secretion is inhibited by ROS scavenger Tempol in IEC

When NCM460 cells (2A) or normal primary colonocytes (2B) were treated with CGN (1μg/ml) for 24 hours in combination with 100nm Tempol), the production of ROS was markedly diminished, to values less than baseline (p<0.001, 1-way ANOVA with Tukey's post-test, n=3).

Figure 3. Stimulation of IL-8 secretion from NCM460 cells by CGN inhibited by ROS scavengers

NCM460 cells in culture were treated with CGN (1μg/ml) for 1, 4 and 24 hours in presence or absence of 100 nM Tempol (3A) or for 24 hours in the presence or absence of 100 nM Tiron (3B), and secretion of IL-8 in the spent medium was assayed by IL-8 ELISA. CGN increased IL-8 secretion ~100%, but in the presence of Tempol this increase declined ~65%. Similarly, when normal human colonocytes were exposed to CGN (1μg/ml × 24 hours) with or without Tempol (3C), marked changes in the IL-8 secretion occurred. Increases with CGN exposure and declines with ROS scavengers are statistically significant (p<0.01 for comparison between CGN-treated and CGN+Tempol at 1 hour; all other comparisons p<0.001, 1-way ANOVA with Tukey's post-test, n=3).

Figure 4. Increased SOD Activity in CGN-treated Cells

CGN exposure markedly increased the SOD activity in the NCM460 cells. Co-administration of Tempol with CGN inhibited the increase in SOD. Differences between CGN-treated vs. control and CGN-treated vs. CGN+Tempol-treated were statistically significant at both 1 and 24 hours (p<0.001, 1-way ANOVA with Tukey's post-test, n=3).

Figure 5. Increase in Cox-2, but not i-NOS, following exposure to CGN

CGN exposure increased the Cox-2 production in the NCM460 cells, to ~400% baseline, from 2.79(±0.08) ng/mg protein to 11.64 (±0.39) ng/mg protein (5A). The observed difference between the control and treated groups was statistically significant (p<0.0001, unpaired t-test, two-tail, n=3). Prostaglandin E2 was measured by ELISA in the NCM460 cells following exposure to λCGN (1 μg/ml × 24 hours) (5B). Six-fold increase occurred from 632.4(±4.0) pg/mg protein to 388.2 (±20.8) pg/mg protein (p<0.0001, unpaired t-test, two-tail, n=3). iNOS was measured in the control and CGN-treated cells, and showed no change following exposure to CGN for 1 and 24 hours (n=6) (5C). 1U/ml of partially purified human iNOS approximately equals 1ng/ml iNOS.

Figure 6. Decreases in Phosphorylated MAPK12, MAPK13 and Phospho-Hsp27 following CGN exposure

NCM460 cells in culture were treated with CGN for 24 hours, and the cell content of 21 phospho-proteins was analyzed by the Phospho-Proteome Profiler. Densitometry was performed of the control (6A) and CGN-treated (6B) arrays, and statistical significance of the differences in intensity determined. Reduced intensity of the replicate spots for MAPK12 (p38γ), MAPK 13 (p38δ), and Hsp27 is evident and differences in intensity are statistically significant (p<0.001, 1-way ANOVA with Tukey's post-test). Densitometric analysis (6C) confirms the change in intensity of phosphorylated-MAPK 12, MAPK13, and Hsp27 following CGN exposure (p<0.001, 1-way ANOVA with Tukey's post-test, n=3).

Figure 7. CGN reduces and Tempol increases phospho-Hsp27

NCM460 cells in culture were treated with CGN and Tempol, and total and phospho-Hsp27 were measured by ELISA. Total Hsp27 (7A) was not changed following CGN or Tempol, but Tempol increased the phospho-Hsp27 (7B) by 24.3% (±4.6) and CGN reduced the phospho-Hsp27 to 60.9 % (±4.2) of baseline. These observed differences were significant (p <0.001, 1-way Anova with Tukey's post-test, n=3).

Figure 8. PP2A inhibitor okadaic acid prevents CGN-induced increase in IL-8

Co-incubation with OA 5 nM with λCGN 1μg/ml × 24 hours reduced the CGN-induced increase in IL-8 (p<0.001, 1-way ANOVA with Tukey's post-test, n=3) by 62.5%. This finding suggests that the CGN-induced effect is mediated at least in part by PP2A. Baseline IL-8 secretion declined by ~32%.

Figure 9. In contrast to Tempol and CGN, Bcl10 silencing has no impact on Phospho-Hsp27

In the cultured NCM460 cells, Bcl10 silencing by siRNA, CGN, and Tempol do not affect the total Hsp27 (8A). However, in contrast to the effects of Tempol and CGN, Bcl10 siRNA does not modify the phospho-Hsp27 content (8B), consistent with a distinct, non-ROS mediated mechanism for the observed effects associated with Bcl10 silencing.

Figure 10. Combined Bcl10 knockdown and ROS quenching by Tempol inhibit the CGN-induced activation of NFκB, phospho-IκBα, and IL-8

The CGN-induced increases in nuclear NFκB concentration (10A), phospho-IκBα (10B), and secreted IL-8 levels (10C) were reduced to less than baseline by the combination of Tempol and Bcl10 silencing, but neither Tempol nor Bcl10 silencing alone was sufficient to inhibit the CGN-induced increase. Differences from control values following CGN and CGN with Bcl10 siRNA or CGN with Tempol or CGN with Bcl10 siRNA and Tempol are statistically significant (p<0.001, 1-way ANOVA with Tukey's post-test, n=3).

Figure 11. Experimental model of CGN induced IL-8 secretion by in colonic epithelial cells

The schematic drawing indicates two pathways of CGN-induced inflammation, through ROS and through Bcl10, that converge at the level of the IKK signalosome.