Alpha-tocopherylquinone differentially modulates claudins to enhance intestinal epithelial tight junction barrier via AhR and Nrf2 pathways

Abstract

Defects in intestinal epithelial tight junctions (TJs) allow paracellular permeation of noxious luminal antigens and are important pathogenic factors in inflammatory bowel disease (IBD). We show that alpha-tocopherylquinone (TQ), a quinone-structured oxidation product of vitamin E, consistently enhances the intestinal TJ barrier by increasing barrier-forming claudin-3 (CLDN3) and reducing channel-forming CLDN2 in Caco-2 cell monolayers (in vitro), mouse models (in vivo), and surgically resected human colons (ex vivo). TQ reduces colonic permeability and ameliorates colitis symptoms in multiple colitis models. TQ, bifunctionally, activates both aryl hydrocarbon receptor (AhR) and nuclear factor erythroid 2-related factor 2 (Nrf2) pathways. Genetic deletion studies reveal that TQ-induced AhR activation transcriptionally increases CLDN3 via xenobiotic response element (XRE) in the CLDN3 promoter. Conversely, TQ suppresses CLDN2 expression via Nrf2-mediated STAT3 inhibition. TQ offers a naturally occurring, non-toxic intervention for enhancement of the intestinal TJ barrier and adjunct therapeutics to treat intestinal inflammation.

Keywords: CP: Cell biology; alpha-tocopherylquinone; aryl hydrocarbon receptor; claudin-2; claudin-3; inflammatory bowel disease; intestinal permeability; tight junction.

Figure 1.. TQ enhances the intestinal TJ barrier

(A–C) Caco-2 cell monolayers grown on Transwells were treated with TQ (25 μM) or vehicle for 48 h. TQ increased transepithelial electrical resistance (TER) (A) and reduced paracellular inulin (B) and urea (C) flux. n = 6, *p < 0.05, ***p < 0.005, ****p < 0.001. (D–F) Wild-type C57BL/6J mice (n = 6) were treated with TQ (50 mg/kg/day, oral gavage) or vehicle, and the mice’s colon tissues were mounted in Ussing chambers after 48 h of TQ administration. TQ increased mouse colonic TER (D) and reduced paracellular inulin (E) and urea (F) flux. *p < 0.05; **p < 0.01. (G–I) Surgically resected normal human colonic tissues were stripped of the sero-muscular layer, and the colonic mucosa was cultured using a gelatin sponge for 18 h with TQ (25 mM) or vehicle and mounted in Ussing chambers. TQ treatment increased the human colonic TER (G) and reduced colonic inulin (H) and urea (I) flux. n = 6. *p < 0.05, **p < 0.01. Veh, vehicle; TQ, α-tocopherylquinone.

Figure 2.. TQ differentially modulates TJ claudin-2 and -3 expression

(A–C) Protein lysate of Caco-2 cells treated with TQ (25 μM, 48 h), colonic mucosa of mice treated with TQ (50 mg/kg/day, oral gavage, 48 h), and human colonic mucosa treated with TQ (25 μM, 18 h) or vehicle were studied for candidate claudins protein expression by western blot. TQ consistently increased barrier-forming claudin-3 (CLDN3) and reduced channel-forming CLDN2 protein expression in all tested models (A). The densitometry for CLDN3 (B) and CLDN2 (C) was quantified using ImageJ software. β-Actin (ACTB) was used as a loading control. **p < 0.01, ***p < 0.005, ****p < 0.001. (D) Confocal images of CLDN3 and CLDN2 staining in Caco-2 cells. TQ increased CLDN3 (green) in the Caco-2 cell membrane and reduced CLDN2 (red) in the Caco-2 cell membrane Nuclei are shown in blue in the merged panels. The dotted lines represent optical levels in x-y and x-z planes. White bar: 5 μm. Veh, vehicle; TQ, α-tocopherylquinone. These data are representative of several areas from n = 7.

Figure 3.. TQ reduces the severity of acute DSS colitis

In the acute DSS (Ac DSS) colitis (2.5% DSS, 7 days) model, treatment with TQ (Ac DSS/TQ) (50 mg/kg/day, oral gavage) or vehicle (Ac DSS/Veh) on C57BL/6J mice was started 2 days before DSS treatment and continued throughout the treatment. (A and B) Mouse body weight was measured throughout the experiment. The disease activity index was calculated based on the loss of body weight, ruffled fur, occult blood, and stool consistency. TQ attenuated DSS-induced loss of body weight (A) and the disease activity index (B). *p < 0.05, **p < 0.01. (C–E) After Ac DSS with TQ or vehicle treatment, the unstripped mice colonic tissues were mounted in Ussing chambers. TQ attenuated the DSS-induced decrease in mouse colonic TER (C) and prevented the DSS-induced multi-fold increase in mouse colonic paracellular flux of inulin (D) and urea (E). **p < 0.01, ***p < 0.005, ****p < 0.001. (F and G) The histological score of Ac DSS colitis was reduced by TQ (F). ****p < 0.001. The DSS-induced loss of colonic crypts, epithelial ulceration, and mucosal infiltration of inflammatory cells was attenuated by TQ in histological examination (G). Black bar: 100 μm. Representation of 3 independent experiments (n = 3 per experiment).

Figure 4.. TQ activates AhR and Nrf2

(A) Caco-2 cells were treated with TQ (25 μM) 24 h after transfection with pGL4.43[luc2P/XRE/Hygro] vector and pGL4.74[hRluc/TK] vector. After 48 h of incubation, cells were lysed and analyzed by the luciferase reporter assay. TQ treatment increased luciferase activity, indicating increased XRE/AhR activity compared with that of vehicle control. ***p < 0.005. (B) Caco-2 cells were treated with TQ (25 μM) 24 h after transfection with pGL4.43[luc2P/ARE/Hygro] vector and pGL4.74[hRluc/TK] vector. After 48 h of incubation, cells were lysed and analyzed by the luciferase reporter assay. TQ treatment increased luciferase activity, indicating increased ARE/Nrf2 activity compared with that of vehicle control. ***p < 0.005. (C) Confocal immunofluorescence images of Ahr or Nrf2 staining (red color) in Caco-2 cells. The images showed that TQ treatment (6 h) increased AhR or Nrf2 migration into the nuclei (blue), indicating that TQ activates AhR and Nrf2. The dotted lines represent the optical level for the x-y plane and the x-z plane. White bar: 5 μm. These data are representative of several areas from n = 5. (D and E) Protein lysates from vehicle (Veh) and TQ-treated Caco-2 cells were studied for AhR and its downstream target CYP1A1, along with Nrf2 and its downstream target NQO1, protein expression by western blot. TQ treatment (24 h) increased AhR, CYP1A1, Nrf2, and NQO1 protein levels (D). (E) The densitometry for (D) was quantified using ImageJ software. ACTB was used as a loading control. *p < 0.05, **p < 0.01, ***p < 0.005. (F) The fold changes in mRNA levels of AhR target CYP1A1 and Nrf2 target NQO1 between Veh- and TQ-treated Caco-2 cells were determined by qRT-PCR. TQ treatment (24 h) increased the mRNA levels of both CYP1A1 and NQO1. The mRNA expression is relative to the GAPDH mRNA level. ****p < 0.001.

Figure 5.. TQ transcriptionally regulates CLDN2 and CLDN3 gene expression

(A and B) The fold changes in mRNA levels of CLDN2 and CLDN3 between Veh- and TQ-treated Caco-2 cells (A) and mice colon (B) were determined by qRT-PCR upon TQ treatment. TQ treatment reduced the mRNA levels of CLDN2 and increased the mRNA levels of CLDN3 in both Caco-2 cells and mice colon. The mRNA expression is relative to the GAPDH mRNA level. **p < 0.01. (C–F) Genetic deletion of AhR or Nrf2 impairs TQ-mediated enhancement of TJ barrier. Genetic deletion of AhR (AhRΔ) or Nrf2 (Nrf2Δ) in Caco-2 cells was achieved using CRISPR-Cas9-mediated genome editing technology. (C) Non-target scrambled (Scr), AhRΔ, and Nrf2Δ Caco-2 cell monolayers grown on Transwells were treated with TQ (25 mM) or Veh for 48 h, and the effect of gene knockouts on TQ-mediated increase in TER was studied. The graph represents the percentage increase in TER upon TQ treatment compared with the corresponding Veh-treated control. TQ-mediated increase in TER in Scr cells was inhibited in Δ AhRD cells as well as Nrf2Δ cells. ****p < 0.001. (D–F) Protein lysates from the Veh- and TQ-treated Scr, AhRΔ, and Nrf2Δ were assessed for AhR and Nrf2 deletion and CLDN3 and CLDN2 protein expression by western blot. Western blot shows the efficiency of CRISPR-Cas9-mediated knockout of AhR and Nrf2 in AhRΔ and Nrf2Δ Caco-2 cells (D). TQ treatment increased CLDN3 protein levels in non-target Scr cells and Nrf2Δ cells but not in AhRΔ cells. TQ-mediated reduction of CLDN2 in Scr cells was still observed in AhRΔ but was inhibited in Nrf2Δ cells. The densitometry for CLDN3 (E) and CLDN2 (F) from (B) was quantified using ImageJ software. ACTB was used as a loading control. **p < 0.01, ***p < 0.005, ****p < 0.001, ns, non-significant.

Figure 6.. AhR/XRE-mediated effect of TQ on CLDN3

(A) The effect of TQ treatment on the mRNA levels of CLDN3 was studied in non-target Scr and AhRΔ Caco-2 cells. The graph represents the CLDN3 mRNA fold change to their corresponding Vehtreated control. TQ-mediated increases in CLDN3 mRNA levels were found to be inhibited in AhRΔ cells. ****p < 0.001, ns, non-significant. (B)The large-molecule inulin flux in AhRΔ cells was compared with that in non-target Scr cells upon treatment with TQ. The reduction of inulin flux upon TQ treatment in Scr cells was found to be inhibited in AhRΔ cells. The graph represents the relative inulin flux upon TQ treatment to their corresponding Vehtreated control. ***p < 0.005, ****p < 0.001. (C) Non-target Scr and AhRD Caco-2 cells were transfected with CLDN3 promoter cloned pGL4.26 [luc2P/minP/Hygro] vector and pGL4.74[hRluc/TK] vector. After 24 h incubation, the cells were treated with TQ (25 μM) or Veh for an additional 48 h, lysed, and analyzed by the luciferase reporter assay. TQ treatment increased luciferase activity (CLDN3 promoter activity) in Scr Caco-2 cells, but this increase was significantly inhibited in AhRΔ cells. ***p < 0.005, ns, non-significant. (D) ChIP assays were performed using digestedchromatin from wild-type Caco-2 cells and an AhR antibody to interrogate binding to the CLDN3 promoter. ChIP-PCR amplification demonstrated the increased abundance of AhR binding to the promoter regions of CLDN3 after TQ treatment of Caco-2 cells. The graph represents the signal relative to the corresponding input. ****p < 0.001.

Figure 7.. TQ reduces CLDN2 via STAT3-dependent mechanism

(A) The effect of TQ treatment on the mRNA levels of CLDN2 was studied in Scr and Nrf2Δ Caco-2 cells. The graph represents the CLDN2 mRNA fold change upon TQ treatment compared with their corresponding Veh-treated control. TQ-mediated decrease in CLDN2 mRNA levels was found to be inhibited in Nrf2Δ cells. ****p < 0.001. (B) The effect of TQ treatment on the mRNA levels of SHP was studied in Scr and Nrf2Δ Caco-2 cells. The graph represents the SHP mRNA fold change compared with the corresponding Veh-treated control. TQ-mediated increases in SHP mRNA levels were found to be inhibited in Nrf2Δ cells. ***p < 0.005, ****p < 0.001. C) The small-molecule urea flux in Nrf2Δ cells was compared with that in non-target Scr cells upon treatment with TQ. The reduction of urea flux upon TQ treatment in Scr cells was found to be inhibited in Nrf2Δ cells. The graph represents the relative urea flux upon TQ treatment compared with their corresponding Veh-treated control. *p < 0.05, ***p < 0.005. (D–F) Protein lysates from Veh- and TQ-treated mice colonic mucosa and human colonic mucosa were studied for STAT3, P-STAT3, CDX2, and SHP protein expression by western blot (D). TQ treatment (6 h) reduced P-STAT3 and increased SHP levels in both mouse and human colonic mucosa. No significant difference was observed in CDX2 protein levels upon TQ treatment (D). The densitometry for P-STAT3 (E) and SHP (F) from (A) were quantified using ImageJ software. ACTB was used as a loading control. *p < 0.05, **p < 0.01, ****p < 0.001. (G) ChIP assays were performed using digested chromatin from Scr and Nrf2Δ Caco-2 cells and a STAT3 antibody on the CLDN2 promoter. ChIP-PCR amplification demonstrated the reduced abundance of STAT3 binding to the promoter regions of CLDN2 in Scr cells but not in Nrf2Δ cells. The graph represents the signal relative to the corresponding input. ***p < 0.005, ns, non-significant. (H) Confocal images of P-STAT3 (red) and nuclei (blue) in Caco-2 cells. The images showed that TQ treatment (3 h) reduced P-STAT3 levels as well as its nucleartranslocation, indicating that TQ inhibits STAT3 activation. The dotted lines represent the optical level for the x-y and x-z planes. White bar: 5 μm. These data are representative of n = 5. (I) Non-target Scr and NRF2Δ Caco-2 cells were cotransfected with pLminPLuc2P_RE5 (STAT3 homodimer luciferase reporter) vector and pGL4.74[hRluc/TK] vector. After 24 h incubation, the cells were treated with TQ (25 μM) or Veh for an additional 48 h, lysed, and analyzed by the luciferase reporter assay. TQ treatment decreased luciferase activity (STAT3 dimer) in Scr Caco-2 cells, but this decrease was not observed in the NRF2Δ cells. ***p < 0.001, ns, non-significant.