Role of snail activation in alcohol-induced iNOS-mediated disruption of intestinal epithelial cell permeability

Abstract

Background: Chronic alcohol use results in many pathological effects including alcoholic liver disease (ALD). ALD pathogenesis requires endotoxemia. Our previous studies showed that increased intestinal permeability is the major cause of endotoxemia, and that this gut leakiness is dependent on alcohol stimulation of inducible nitric oxide synthase (iNOS) in both alcoholic subjects and rodent models of alcoholic steatohepatitis. The mechanism of the alcohol-induced, iNOS-mediated disruption of the intestinal barrier function is not known. We have recently shown that alcohol stimulates activation of the transcription factor Snail and biomarkers of epithelial mesenchymal transition. As activated Snail disrupts tight junctional proteins, we hypothesized that activation of Snail by iNOS might be one of the key signaling pathways mediating alcohol-stimulated intestinal epithelial cell hyperpermeability.

Methods: We measured intestinal permeability in alcohol-fed C57BL/6 control and iNOS knockout (KO) mice, and measured Snail protein expression in the intestines of these mice. We then examined intestinal epithelial permeability using the Caco-2 cell model of the intestinal barrier ± small interfering RNA (siRNA) inhibition of Snail. We assessed Snail activation by alcohol in Caco-2 cells ± inhibition of iNOS with L-NIL or siRNA. Finally, we assessed Snail activation by alcohol ± inhibition with siRNA for p21-activated kinase (PAK1).

Results: Our data show that chronic alcohol feeding promotes intestinal hyperpermeability in wild-type BL/6, but not in iNOS KO mice. Snail protein expression was increased in the intestines of alcohol-treated wild-type mice, but not in iNOS KO mice. siRNA inhibition of Snail significantly inhibited alcohol-induced hyperpermeability in Caco-2 cell monolayers. Alcohol stimulation of Snail(pS246) activation was blocked by inhibition of iNOS with L-NIL or with siRNA. siRNA inhibition of PAK1 significantly inhibited alcohol-mediated activation of Snail in Caco-2 cells.

Conclusions: Our data confirmed our prior results and further demonstrated that alcohol-induced gut leakiness in rodents and intestinal epithelial cell monolayers is iNOS dependent. Our data also support a novel role for Snail activation in alcohol-induced, iNOS-mediated intestinal hyperpermeability and that PAK1 is responsible for activation of Snail at Ser246 with alcohol stimulation. Identification of these mechanisms for alcohol-induced intestinal hyperpermeability may provide new therapeutic targets for prevention and treatment of alcohol-induced leaky gut, endotoxemia, and endotoxin-associated complications of alcoholism such as ALD.

Figure 1. Alcohol fed BL/6 mice and not iNOS KO mice exhibit increased intestinal permeability (L/M ratio)

WT BL/6 and iNOS KO mice were fed a liquid diet containing 4.5% alcohol (28% calories) for 4 weeks as described in Methods and then intestinal permeability was determined by gavage of a sugar solution described in Methods followed by analysis of excreted urine sugar levels by GC using 6h urine samples. Data are means ± SE for N=6 mice each condition and are presented as percent change from initial respective control baseline (set as 100%) of the L/M ratio (lactulose/mannitol ratio) with * p<.05 for BL/6 mice given alcohol compared to BL/6 mice without alcohol (300% increase) while the iNOS KO mice showed no significant difference ± alcohol (only 80% increase).

Fig. 2. Chronic alcohol feeding results in increased intestinal Snail protein expression in WT BL/6 but not in iNOS KO mice

WT BL/6 and iNOS KO mice were fed a liquid diet containing 4.5% alcohol (28% calories) for 4 weeks as described in Methods and then levels of Snail (30 kDa) were assessed in proximal colon intestinal tissue by western blotting of tissue lysates equalized for total protein with Ab to total Snail. Upper panel shows representative blots from single mice, histogram densitometry data are means ± SE for all mice (N=6 each condition).* p < 0.05 for alcohol fed BL/6 mice vs. corresponding BL/6 control. # p < 0.05 for alcohol fed BL/6 mice vs. iNOS KO alcohol fed mice. Blots were stripped and reprobed with Ab to actin (42 kDa) to further control for protein loading.

Fig. 3. The iNOS specific inhibitor L-NIL blocks alcohol-induced activation of Snail^pS246

Caco-2 intestinal epithelial cells grown on culture inserts were assessed for activation of Snail ± 0.2% alcohol (EtOH) treatment for 2h using western blotting of protein equalized nuclear extracts and Ab to Snail^pS246.Some cells in lane three were pretreated for 1h with 200µM L-NIL iNOS inhibitor. Blots were stripped and reprobed with Ab to the nuclear protein Histone H3 (17 kDa) to demonstrate equal loading. Blot data are from representative wells from 4 separate experiments in triplicate (N=12). Histogram densitometry data are summarized means ± SE for N=12 for each condition. * p< .05 EtOH+ L-NIL vs. EtOH treatment alone.

Fig. 4. Immunofluorescent staining of Caco-2 cells reveals iNOS inhibition prevents alcohol stimulated nuclear Snail^pS246 localization

Caco-2 cells grown on glass coverslips in 6-well plates were treated with alcohol (EtOH, 0.2%) for 2h and then fixed and stained with Ab to Snail^pS246+FITC 2° Ab. Nuclei were stained with DAPI. Cells for 4C were also pretreated with the iNOS inhibitor L-NIL (200µM, 1h).Images are representative from triplicate coverslips from three separate experiments (N=9). Scale Bar = 10 microns. Images were obtained using Zeiss Axiovision imaging as described in Methods with a 40× oil immersion objective (4A&C) and a 63× objective for greater detail in 4B.

Fig. 5. Knocking down iNOS with siRNA prevents EtOH-induced Snail^pS246 in Caco-2 cells

Caco-2 cells grown on inserts were assessed for Snail protein phosphorylation at S246 by western blotting of whole cell lysates with lysis buffer containing 0.2%SDS. Lysates were also assessed for levels of iNOS protein to determine knockdown efficiency and β-actin as a loading control. Some cells were pretreated as described with either control siRNA or siRNA specific for iNOS. Designated cells were treated with 0.2% (43mM) alcohol (EtOH) for 2h. Blots are from representative wells from triplicate wells in one of four separate experiments (N=12). Lower histogram densitometry data are summarized means ± SE for all N=12 wells for each condition for Snail^pS246 (dark bars) and iNOS protein (130 kDa) (light bars). **/++ p< .05 for alcohol vs. media treated control. * p< .05 EtOH + control siRNA compared to control siRNA alone treated cells for the respective protein.⁺ p<.05 iNOS siRNA + EtOH compared to control siRNA + EtOH treated cells.

Figure 6. Inhibition of PAK1 expression with siRNA prevents alcohol-induced activation of Snail^pS246

Caco-2 cells grown on inserts were assessed for Snail phosphorylation at S246 by western blotting of whole cell lysates with lysis buffer containing 0.2%SDS. Lysates were also assessed for levels of PAK1 to determine knockdown efficiency and β-actin as a loading control. Some cells were pretreated as described with either control siRNA or siRNA specific for PAK1. Designated cells were treated with 0.2% (43mM) alcohol (EtOH) for 2h. Data are representative wells from triplicate wells in one of four separate experiments (N=12). Lower densitometry histograms are for means ± SE for all N=12 wells for each condition for Snail^pS246 (dark bars) and PAK1 (light bars). **/++ p< .05 for alcohol vs. media treated control. * p< .05 EtOH + control siRNA compared to control siRNA alone treated cells for the respective protein.⁺ p<.05 PAK1 siRNA + EtOH compared to control siRNA + EtOH treated cells.

Fig. 7. SiRNA knockdown of Snail expression significantly inhibits alcohol-induced permeability in Caco-2 monolayers

Caco-2 cell monolayers grown on permeable inserts were assessed for permeability at 2h after treatment of some wells with 0.2% alcohol (EtOH).Fig. 7A shows permeability data assessed by TER. Histogram data are means ± SE of triplicate wells in 4 separate experiments (total N=12 for each condition).Fig. 7B (upper) shows permeability at 2h assessed using flux of FSA from the apical to basal chambers as described in Methods. Data are means ± SE of triplicate wells from 4 separate experiments (total N=12 for each condition). Fig. 7B (lower) shows western blotting data from representative wells from each condition in the upper panel assessed for total Snail protein expression with β-actin as a loading control, mean Snail knockdown was >70%. * p< .05 EtOH alone treated cells or EtOH +control siRNA vs. media control cells.⁺⁺ p<.05 for Snail siRNA + EtOH treated cells vs. control siRNA + EtOH treated cells.

Fig. 7. SiRNA knockdown of Snail expression significantly inhibits alcohol-induced permeability in Caco-2 monolayers

Caco-2 cell monolayers grown on permeable inserts were assessed for permeability at 2h after treatment of some wells with 0.2% alcohol (EtOH).Fig. 7A shows permeability data assessed by TER. Histogram data are means ± SE of triplicate wells in 4 separate experiments (total N=12 for each condition).Fig. 7B (upper) shows permeability at 2h assessed using flux of FSA from the apical to basal chambers as described in Methods. Data are means ± SE of triplicate wells from 4 separate experiments (total N=12 for each condition). Fig. 7B (lower) shows western blotting data from representative wells from each condition in the upper panel assessed for total Snail protein expression with β-actin as a loading control, mean Snail knockdown was >70%. * p< .05 EtOH alone treated cells or EtOH +control siRNA vs. media control cells.⁺⁺ p<.05 for Snail siRNA + EtOH treated cells vs. control siRNA + EtOH treated cells.