Endotoxin, but not platelet-activating factor, activates nuclear factor-kappaB and increases IkappaBalpha and IkappaBbeta turnover in enterocytes

Abstract

Bacterial endotoxin (lipopolysaccharide; LPS) and platelet-activating factor (PAF) are important triggers of bowel inflammation and injury. We have previously shown that LPS activates the transcription factor nuclear factor (NF)-kappaB in the intestine, which up-regulates many pro-inflammatory genes. This effect partly depends on neutrophils and endogenous PAF. However, whether LPS and PAF directly activate NF-kappaB in enterocytes remains controversial. In this study, we first investigated the effect of LPS and PAF on NF-kappaB activation in IEC-6 (a non-transformed rat small intestinal crypt cell line) cells, by electrophoresis mobility shift assay and supershift, and found that LPS, but not PAF, activates NF-kappaB mostly as p50-p65 heterodimers. The effect was slower than tumour necrosis factor (TNF). Both LPS and TNF induce the expression of the NF-kappaB-dependent gene inducible nitric oxide synthase (iNOS), which occurs subsequent to NF-kappaB activation. We then examined the effect of LPS and TNF on the inhibitory molecules IkappaBalpha and IkappaBbeta. We found that TNF causes rapid degradation of IkappaBalpha and IkappaBbeta. In contrast, LPS did not change the levels of IkappaBalpha and IkappaBbeta up to 4 hr (by Western blot). However, in the presence of cycloheximide, there was a slow reduction of IkappaBalpha and IkappaBbeta, which disappeared almost completely at 4 hr. These observations suggest that LPS causes slow degradation and synthesis of IkappaBalpha and IkappaBbeta and therefore activates NF-kappaBeta via at least two mechanisms: initially, through an IkappaB-independent mechanism, and later, via an increased turnover of the inhibitor IkappaB. NF-kappaBeta activation precedes the gene expression of iNOS (assayed by reverse transcription-polymerase chain reaction), suggesting that LPS up-regulates iNOS via this transcription factor.

Figure 1

Dose–response of LPS-induced NF-κB activation. IEC-6 cells (4×10⁶ cells) were incubated with either 0·2, 1, 5 or 10 µg/ml of LPS for 1 hr and the NF-κB activity of the cell nuclear extracts was assessed by EMSA. NF-κB activation is seen with the lowest dose of LPS (0·2 µg/ml) used and the effect is dose dependent. Supershift experiments were done. –, no added; p50, anti-p50Ab added; p65, anti-p65Ab added. (Similar results were obtained in three independent experiments.)

Figure 2

(a) Time-course and subunit composition of LPS-induced NF-κB activation in IEC-6 cells: comparison with TNF. IEC-6 cells were incubated with 10 ng/ml of TNF or 1 µg/ml of LPS for various time periods, and NF-κB activation detected at 0, 10, 20, 30 and 45 min after TNF (left panel) and at 0, 30, 60, 90 min after LPS stimulation (right panel) (arrows, including p50–p50 and p50–p65). In the right panel, a supershift experiment shows that the LPS-induced NF-κB complex could be supershifted with anti-p50 and anti-p65 antibodies. Similar results were obtained in three independent experiments. (b) LPS and TNF induce iNOS gene expression in IEC-6 cells. IEC-6 cells were incubated for various times with LPS (upper panel) or TNF (lower panel) (two samples for each time-point). RNA was extracted, RT–PCR performed using primers specific for iNOS or actin. PCR products were run on a 1·5% agarose gel and stained with SYBR green I.

Figure 3

LPS-induced NF-κB activation in IEC-6 cells is not enhanced by PMN and is not inhibited by PAF receptor antagonists. (a) EMSA showing the lack of difference in the NF-κB activity between cells treated for 30 min with LPS (1 µg/ml) alone and LPS plus PMN (1×10⁶/ml). (b) A typical gel with supershift experiments showing no difference in the subunit composition of LPS- and LPS/PMN-treated cells. (c) WEB2170 (0·5 µg/ml), a PAF receptor antagonist, had no effect on LPS induced NF-κB activation. Here, a typical supershift experiment is presented. Each experiment was performed at least three times and showed consistent results.

Figure 4

PAF does not activate NF-κB in IEC-6 cells and does not potentiate LPS-induced NF-κB activation. EMSA with supershift experiments showing lack of stimulatory effect of PAF (30 min). (a) Lanes 1–3: untreated cells. Lanes 4–6: cells treated with carbamyl-PAF, 0·5 µg/ml. Lanes 7–9: PAF, 0·5 µg/ml. Note a low level of constitutively present p50 in unstimulated cells. A typical supershift experiment showing the lack of synergistic effect between PMN and PAF (b) and between cPAF and LPS (c). The same results were obtained in at least five experiments.

Figure 5

Western blot analysis showing TNF-induced degradation of IκBα and IκBβ in IEC-6 cells. Cells were treated with: TNF (10 ng/ml) or TNF plus cycloheximide (50 µg/ml) for various time lengths. At the end of the experiment, total cell lysates were prepared and subjected to immunoblot analysis with anti-IκBα or anti-IκBβ antibodies. Note the rapid degradation of IκBα and the slow degradation of IκBβ induced by TNF with resynthesis of both molecules. Cycloheximide (50 µg/ml) completely blocked the TNF-induced IκBα and IκBβ resynthesis.

Figure 6

Western blot analysis showing increased turnover of IκBα and IκBβ in IEC-6 cells after LPS stimulation. Cells were treated with: LPS (1 µg/ml), LPS plus cycloheximide (50 µg/ml), or cycloheximide alone, for various time lengths. At the end of the experiment, total cell lysates were prepared and subjected to immunoblot analysis with anti-IκBα or anti-IκBβ antibodies. Note the absence of IκB degradation with LPS alone (LPS), and the obvious diminishing of IκB when cycloheximide was added together with LPS (LPS + CHX). Cycloheximide alone had no effect (CHX) (N=3).