The NADPH oxidase 1 (Nox1) is a gp91phox homologue preferentially expressed in the colon. We have established primary cultures of guinea pig large intestinal epithelial cells giving 90% purity of surface mucous cells. These cells spontaneously released superoxide anion (O2−) of 160 nmol/mg protein/h and expressed the Nox1, p22phox, p67phox, and Rac1 mRNAs, but not the gp91phox, Nox4, p47phox, p40phox, and Rac2 mRNAs. They also expressed novel homologues of p47phox and p67phox (p41nox and p51nox, respectively). Human colon cancer cell lines (T84 and Caco2 cells) expressed the Nox1, p22phox, p51nox, and Rac1 mRNAs, but not the other NADPH component mRNAs, and secreted only small amounts of O2− (<2 nmol/mg protein/h). Cotransfection of p41nox and p51nox cDNAs in T84 cells enhanced PMA-stimulated O2− release 5-fold. Treatment of the transfected T84 cells with recombinant flagellin (rFliC) from Salmonella enteritidis further augmented the O2− release in association with the induction of Nox1 protein. The enhanced O2− production by cotransfection of p41nox and p51nox vectors further augmented the rFliC-stimulated IL-8 release from T84 cells. T84 cells expressed the Toll-like receptor 5, and rFliC rapidly phosphorylated TGF-β-activated kinase 1 and TGF-β-activated kinase 1-binding protein 1. A potent inhibitor for NF-κB (pyrrolidine dithiocarbamate) significantly blocked the rFliC-primed increase in O2− production and induction of Nox1 protein. These results suggest that p41nox and p51nox are involved in the Nox1 activation in surface mucous cells of the colon, and besides that, epithelial cells discern pathogenicities among bacteria to appropriately operate Nox1 for the host defense.
Reactive oxygen species (ROS),3 notably superoxide anion (O2−) and hydrogen peroxide, operate on a variety of physiological processes, including host defense, gene expression, oxygen sensing, regulation of vascular tone, bone resorption, apoptosis, cell growth, and transformation (for reviews, see Refs. 1, 2, 3). The best-known O2−-producing enzyme is the phagocyte respiratory burst oxidase that plays a crucial role in a process of killing microorganisms. The catalytic core of this oxidase is the membrane-integrated flavocytochrome b558 composed of p22phox and gp91phox subunits, the latter having binding sites for heme, flavin adenine dinucleotide (FAD), and NADPH, and transfers an electron from NADPH to molecular oxygen to generate O2− (4). Recently, two families of gp91phox homologues have been identified: NADPH oxidase (Nox) and dual oxidase (Duox) families (5). The Nox family comprises Nox1 (initially termed Mox1 or NOH-1), Nox2 (renamed gp91phox), Nox3, Nox4 (Renox), and Nox5 (6, 7, 8, 9, 10). These homologues conserve binding sites for heme, FAD, and NADPH of Nox2 (5), and are preferentially expressed in nonphagocytic cells. The Duox family members are Duox1 and Duox2 (initially termed ThOX1 and ThOX2, respectively), which have a peroxidase homology domain plus two EF-hand motifs, as well as binding sites for heme, FAD, and NADPH (5). Of these family members, the Nox1 mRNA is predominantly expressed in human colon tissue and a carcinoma cell line, Caco2 cells (6, 7). However, physiological roles of Nox1 in large intestinal epithelial cells (LIEC) are not fully understood.
We previously reported that primary cultures of guinea pig gastric pit cells expressed Nox1 and spontaneously secreted O2− (11, 12). ROS derived from Nox1 in the pit cells were essential for their growth at least in vitro (12). Furthermore, Helicobacter pylori LPS stimulated the Toll-like receptor (TLR) 4 signaling and activated Nox1 (13, 14). The increased O2− production and enhanced activation of NF-κB resulted in the induction of the TNF-α and cyclooxygenase II mRNAs in pit cells themselves (12), suggesting a potential role of Nox1 in inflammatory and immune responses against H. pylori.
The flavocytochrome b558 requires cytosolic proteins, p67phox, p47phox, and a small GTPase Rac for electron-transfer reactions to form O2− (4). P40phox associates with p67phox and enhances membrane translocation of p67phox and p47phox in stimulated phagocytes (15). However, it remains to be elucidated whether these cytosolic factors are necessary for activation of the other Nox and Duox family members.
In this study, we have established primary cultures of guinea pig LIEC with 90% purity of surface mucous cells and have found that these cells also produce O2− even at a higher rate than that of gastric pit cells. Guinea pig LIEC expressed novel genes encoding p41nox (a p47phox homologue) and p51nox (a p67phox homologue), which have recently cloned in mouse and human, being named as the NOX organizer 1 and NOX activator 1 (16, 17, 18), respectively. Cotransfection of these two homologues was shown to up-regulate O2−-producing capability of Nox1, and in situ hybridization demonstrated that the p41nox and p51nox transcripts were expressed in epithelial cells of mouse colonic mucosa (16, 17, 18), suggesting their crucial roles for Nox1 activity. Using guinea pig LIEC and human colon cancer cell lines, we molecularly and functionally characterized Nox1 expressed in LIEC.
Materials and Methods
A recombinant structural protein of flagella filament (rFliC) of Salmonella enteritidis was prepared, as previously described (19). Staphylococcus aureus peptidoglycan and LPS from Escherichia coli K-235 were purchased from Fluka Chemie AG (Buchs, Switzerland) and Sigma-Aldrich (St. Louis, MO), respectively. Pyrrolidine dithiocarbamate (PDTC) was obtained from Calbiochem (San Diego, CA). Phosphorothioate-stabilized CpG oligodeoxynucleotide (CpG DNA) (TCCATGACGTTCCTGATGCT) (20) was purchased from Hokkaido System Science (Sapporo, Japan).
Preparation and culture of cells
The present study was approved by the Animal Care Committee of University of Tokushima. Male guinea pigs weighing ∼250 g were purchased from Shizuoka Laboratory Animal Center (Shizuoka, Japan). Under general anesthesia with pentobarbital, ascending to sigmoid portions of the guinea pig colon were resected and extensively washed with PBS. Colonic mucosa was scraped with a sterile glass slide and finely minced with sterile surgical blades. The minced pieces were then incubated in DMEM-Ham’s F-12 (1:1) (DF) medium (Life Technologies, Grand Island, NY) containing 0.03% collagenase S-1 (Nitta Gelatin, Osaka, Japan) and 0.2% BSA for 30 min at 37°C. The digested tissues were next washed three times with DF medium by centrifugation. The resulting pellets were resuspended in DF medium containing 10% FBS (PAA Lab., Linz, Austria) and incubated at 37°C for 2 h under 5% CO2. Attached cells were positive for a specific mAb against macrophages (HAM56 clone; Enzo Diagnostic, Farmingdale, NY) and used as a macrophage population derived from colonic mucosa. Collected nonadherent cells were washed with DF medium and then cultured for 24 h in 35-mm-diameter culture dishes that had been coated with type IV collagen (Nitta Gelatin). The growing cells were used as primarily cultured guinea pig LIEC. These cells were completely replaced by vimentin-positive fibroblasts after a 2-wk cultivation. Guinea pig and human PBL were prepared, as previously described (14). T84 and Caco2 cells were maintained in DMEM medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin.
The amount of O2− release was measured by the superoxide dismutase (SOD)-inhibitable reduction of cytochrome c, and O2−-producing cells were cytochemically visualized by detecting blue formazan precipitates of nitroblue tetrazolium (NBT), as previously described (11).
Anti-Nox1 Abs were made by immunizing rabbits with synthetic peptides corresponding to the 480–493 (Nox1-C1) and 544–556 (Nox1-C2) aa residues of human Nox1 (GenBank accession AF166327). The epitopes for the two anti-Nox1 Abs were designed not to overlap the amino acid sequences of the other Nox homologues (GenBank accession NM000397, AF190122, NM016931, and NM024505). Nox1-C1- or Nox1-C2-immunized serum was further purified by affinity chromatography using the Ag peptide-conjugated agarose (Amersham Pharmacia Biotech, Piscataway, NJ). The amino acid sequence of Nox1-C1 has 93% homology between human and guinea pig (GenBank, AB099629), and anti-Nox1-C1 Ab recognized both human and guinea pig Nox1 proteins. In contrast, anti-Nox1-C2 Ab recognized only human Nox1, because guinea pig Nox1 does not share this motif. Anti-TLR5 Ab was made by immunizing rabbits with synthetic peptides corresponding to the 836–849 aa residues of human TLR5. Polyclonal Abs against synthetic peptide of human p22phox (residues 177–195), human p47phox (residues 376–390), human p40phox (residues 1–15), and recombinant human p67phox were provided, as previously described (13). Membrane, cytosolic, and whole cell fractions were prepared from cultured cells, as previously described (11). Each sample of 20 μg protein per lane was separated by SDS-PAGE using an 8% polyacrylamide gel and transferred to a polyvinylidene difluoride filter (PVDF). After blocking nonspecific binding sites with 4% purified milk casein, the PVDF was incubated for 1 h at room temperature with one of the above primary Abs at a 1/1000 dilution. After being washed with PBS containing 0.05% Tween 20, bound Abs were detected by an ECL Western blotting detection system (Amersham Pharmacia Biotech). Bound Abs were then removed, and the PVDF was reblotted with a mAb against β-actin (Oncogene Research Products, Cambridge, MA). Phosphorylation of TGF-β-activated kinase 1 (TAK1) and TAK1-binding protein 1 (TAB 1) was assayed, as previously described (14).
Cytochemical and immunohistochemical stainings
Mucous granule-containing cells were visualized by the periodic acid-Schiff (PAS) reaction. For immunohistochemical analysis, growing LIEC on the dishes were fixed with 4% paraformaldehyde in PBS for 20 min. They were then incubated in PBS containing 0.03% Triton X-100 for 2 min on ice and blocked with 4% purified milk casein. These cells were incubated with a 1/500 dilution of anti-Nox1-C1 Ab for 1 h. After washing, they were treated with a 1/500 dilution of biotin-linked goat Ab against rabbit IgG (Amersham Pharmacia Biotech) for 1 h, and were then incubated with a 1/500 dilution of streptavidin-conjugated FITC probe for 30 min at room temperature. The cells were mounted with Vectashield mounting medium (Vector Laboratories, Burlingame, CA). Fluorescence was viewed using a confocal laser-scanning microscopy (model Axiovert 25CFL; Leica, Heiderberg, Germany). Fibroblasts and macrophages were immunocytochemically identified with mAbs against vimentin (Santa Cruz Biotechnology, Santa Cruz, CA) and a macrophage Ag (HAM56 clone); vimentin-positive fibroblasts and HAM56-positive macrophages were visualized by diaminobenzidine streptavidin-biotin HRP and Vector red alkaline phosphatase methods (Vector Laboratories), respectively.
Paraffin-embedded guinea pig colon tissues were cut into 3-μm-thickness sections and deparaffinized. After blocking nonspecific binding sites with 4% purified milk casein, they were incubated with a 1/1000 dilution of anti-Nox1-C1 Ab in TBS containing 0.1% Tween 20 and 2% BSA overnight at 4°C. After washing, bound Abs were visualized using Vector red alkaline phosphatase kit (Vector Laboratories). Finally, specimens were counterstained with hematoxylin and mounted with Aquatex (Merck, Darmstadt, Germany).
Total RNA was prepared from the indicated cells or tissues with an acid guanidium-thiocynate-phenol chloroform mixture (14). RT-PCR was performed using the following specific PCR primer sets: Nox1-A, 5′-ATGGGAAACTGGGTGGTTA-3′ and 5′-TAGCTGAAGTTACCATGAGAA-3′; Nox1-B, 5′-TTCTTGGCTAAATCCCATCCA-3′ and 5′-TTTCTGTCCAGTCCCCTGCT-3′; Nox2, 5′-CATCATCTCTTTGTGATCTTCT-3′ and 5′-CTTAGGTAGTTTCCACGCATC-3′; Nox4, 5′-GGTCCTTTTGGAAGTCCATTTGAGG-3′ and 5′-CACAGCTGATTGATTCCGCTGAG-3′; p22phox, 5′-ATGGGGCAGATCGAGTGGGCCATGT-3′ and 5′-GTAGATGCCGCTCGCAATGGCCAG-3′; p67phox, 5′-TCCCGGATTTGCTTCAACATT-3′ and 5′-TTGGCCAGCTGAGCCACTT-3′; p47phox-A, 5′-ATCCGTCACATCGCCCTGCT-3′ and 5′-CCAACCGCTCTCGCTCTTCT-3′; p47phox-B, 5′-AACAGGATCATCCCCCACCT-3′ and 5′-CAGGTACATGGACGGAAAGT-3′; p40phox, 5′-TGACATCGAGGAGAGAGGCT-3′ and 5′-GGAAGATCACATCTCCAGCTTTGA-3′; p41nox, 5′-TTTGCCTTCTCTGTGCGCTGG-3′ and 5′-TCTGGGGTGGGCAGGATCACC-3′; p51nox, 5′-CAAGCAGTGACTAAGGACACCTG −3′ and 5′-CACACAGGACATCCACCGTGTC-3′; Rac1, 5′-TGCAGGCCATCAAGTGTGTGGT-3′ and 5′-GCTGAGACATTTACAACAGCAGGCAT-3′; Rac2, 5′-TGCAGGCCATCAAGTGTGTGGT −3′ and 5′-TAGAGGAGGCTGCAGGCGCGCTT-3′; and GAPDH, 5′-TCATGACCACAGTCCATGCCATCACT-3′ and 5′-GCCTGCTTCACCACCTTCTTGATGT-3′. PCR products were sequenced with a DNA sequencer and confirmed to be the corresponding cDNA fragments.
Expression vectors and cDNA constructs
The cDNA encoding human p51nox was provided, as previously described (18), and the cDNA encoding p67phox was a gift from H. Nunoi (Miyazaki Medical School, Miyazaki, Japan). The p41nox cDNA (GenBank, AF539796) was amplified by PCR using human digestive system multiple tissue cDNA (Clontech Laboratories, Palo Alto, CA). Full-length p67phox cDNA and hemagglutinin-tagged p41nox and p51nox cDNAs were recombined into pAdTrack-CMV vector (21). T84 cells were plated at a concentration of 5 × 105 cells/well in 24-well plates and transfected with the vectors using the FuGENE transfection reagent (Roche Biomedical Laboratories, Burlington, NC).
Measurement of IL-8
T84 cells were transfected with mock (pAdTrack-CMV vector) or p51nox plus p41nox vectors for 48 h. These cells growing in 35-mm-diameter culture dishes at 70–90% confluence were incubated in 1 ml of serum-free DF medium for 2 h and then treated with rFliC (5 μg/ml) in the absence or presence of 200 U/ml SOD plus 700 U/ml catalase. After treatment for 24 h, the medium was collected and centrifuged at 1000 × g for 10 min. The concentrations of IL-8 in the supernatants were measured using a human IL-8 ELISA kit (R&D Systems, Abington, U.K.), according to the manufacturer’s protocol. The amount of IL-8 was expressed as pg/ml per mg cell protein.
Nox1-expressing and O2−-producing cells in guinea pig LIEC
Contaminated macrophages were removed as attaching cells during an initial 2-h cultivation of isolated guinea pig colonic mucosal cells. Nonadherent cells were collected and cultured in DF medium supplemented with 10% FBS. These cells began to adhere to collagen-coated plates within 6 h and became confluent at ∼24 h. After a 48-h culture, they started to undergo spontaneous apoptosis, mirroring their rapid turnover in vivo. At the 24-h cultivation, 90 ± 3% (mean ± SD, n = 8) of cultured cells possessed PAS reaction-positive granules characteristic of surface mucous cells (Fig. 1, A and B). Although vimentin-positive fibroblasts were less than 5% at 24 h (Fig. 1,C), they had grown to be an exclusive population in 2 wk (Fig. 1,D). Macrophages were not detected in the 24-h LIEC culture (Fig. 1,E) after their removal by the adherent method (Fig. 1 F), and the growing LIEC were used in the following experiments.
The cultured guinea pig LIEC spontaneously secreted O2− at 156 ± 9 nmol/mg protein/h (mean ± SD, n = 12). This rate was higher than that of cultured gastric pit cells primed with H. pylori LPS (112 ± 5 nmol O2−/mg protein/h; mean ± SD, n = 12). O2−-producing cells contained granules positive for the PAS reaction (Fig. 2,A) and expressed Nox1 protein (Fig. 2,B). These Nox1-expressing cells (Fig. 2,C) were identical with the cells covered with precipitates of blue formazan (Fig. 2,D). In the guinea pig colon, surface mucous cells possessed immunoreactive materials to anti-Nox1-C1 Ab (Fig. 2, E and F), and the synthetic Ag polypeptide completely abolished this immunoreactivity (Fig. 2 G). Thus, surface mucous cells of the guinea pig colon constitutively expressed Nox1 protein both in vitro and in vivo. We also confirmed that the prepared fibroblasts released O2− at <1 nmol/mg protein/h, but these amounts were not enough to form visible blue formazan precipitates even after incubation with 0.1 mM of NBT for 2 h (data not shown).
RT-PCR with two different primer sets amplified the Nox1 mRNA fragments in guinea pig LIEC and Caco2 (Fig. 2,H), and their nucleotide sequences were identical with the guinea pig and human Nox1 mRNAs, respectively (data not shown). We also confirmed that guinea pig LIEC did not express the Nox2 nor Nox4 mRNA (Fig. 2, I and J).
Expression of NADPH oxidase components in guinea pig LIEC
We next screened the expression of Nox components. As shown in Fig. 3,A, guinea pig LIEC expressed the p22phox, p67phox, and Rac1 transcripts, while the p47phox, p40phox, and Rac2 mRNAs were not detected. Immunoblot analysis showed that they had p67phox and p22phox proteins, but not p47phox and p40phox in line with the RT-PCR results (Fig. 3,B). Moreover, we have cloned the p41nox (GenBank, AB105906) and p51nox (GenBank, AB105907) mRNAs expressed in guinea pig LIEC (Fig. 3 C). Guinea pig PBL also expressed a small amount of p51nox mRNA, but not p41nox.
Roles of p41nox and p51nox in Nox1 activity
T84 and Caco2 cells generated only small amounts of O2− (<2 nmol/mg protein/h). The low output of O2− in these cells was mainly due to lower levels of Nox1 expression, compared with that in guinea pig LIEC or guinea pig gastric pit cells. It may be also due to the absence or insufficiency of distinct component(s) supportive for the Nox1 activity. The p67phox and p41nox mRNAs were absent in T84 and Caco2 cells, and the p51nox mRNA level was much lower than that in guinea pig LIEC (Fig. 4,A). As shown in Fig. 4,B, transduction of p67phox or overproduction of p51nox did not change the spontaneous or PMA-stimulated release of O2− from T84 cells, but p41nox-overexpressing cells significantly increased O2− generation when stimulated by PMA (Fig. 4,B). Although transfection of the p41nox-overexpressing cells with the p67phox vector failed to increase both spontaneous and PMA-stimulated O2− productions, overproduction of p51nox in the p41nox-transfected cells further increased the PMA-responsive O2− generation (Fig. 4 B). P67phox is an essential activator for Nox2, and p51nox has conserved domains that possibly interact with Nox1 in a similar way as p67phox does with cytochrome b558 in phagocytes (16, 17, 18, 22). However, our results suggest that p51nox is most likely a better partner of Nox1 for achieving a high output of O2− production.
Effects of bacterial components on Nox1 activity
To elucidate physiological functions of Nox1 in LIEC, we explored possible up-regulator(s) of the oxidase. Mirroring the rapid turnover of surface mucous cells in vivo, guinea pig LIEC in primary culture appeared to be already in a fully matured and activated status; therefore, we investigated whether T84 cells overexpressing p41nox and p51nox could be primed with bacterial components for O2− generation. LPS from H. pylori or E. coli was demonstrated to act as a potent stimulator of Nox1 in primary cultures of guinea pig gastric mucosal cells (14). Although T84 cells express the TLR4 mRNA (data not shown), they were insensitive to LPS priming: E. coli LPS up to 20 μg/ml did not change the basal O2− generation (Fig. 5,A). Neither peptidoglycan from S. aureus nor CpG DNA increased the O2− production (Fig. 5,A). In contrast, rFliC from S. enteritidis at 2 μg/ml or higher concentrations significantly enhanced O2− generation within 12 h (Fig. 5, B and C). Boiling did not affect the priming action of rFliC, but treatment with trypsin completely abolished it (Fig. 5,A), which are well-known characteristics of FliC (19). T84 cells expressed the TLR5 mRNA (Fig. 5,D), and TLR5 protein was mainly present in the membrane fraction (Fig. 5,E). TAK1 is a member of mitogen-activated protein kinase kinase kinase. TAB 1 is a specific activator for TAK1. Phosphorylation of TAK1 and TAB 1 is a crucial event for NF-κB activation through the TLR and IL-1R signaling pathways (23, 24). We confirmed that the treatment of T84 cells with rFliC promptly phosphorylated TAK1 and TAB 1 within 10 min (Fig. 5 F).
ROS-dependent IL-8 release from T84 cells
To address physiological roles of Nox1-derived ROS, we tested whether O2− production up-regulated IL-8 secretion from T84 cells. Treatment of T84 cells with rFliC for 24 h increased IL-8 release (Fig. 5,G). Overproduction of p41nox and p51nox in T84 cells enhanced O2− generation (Fig. 5, A–C), and consequently augmented the rFliC-stimulated IL-8 production, which was significantly cancelled by inclusion of SOD plus catalase (Fig. 5 G).
Induction of Nox1 protein with rFliC
Finally, we studied the mechanisms by which rFliC increased Nox1 activity in T84 cells. rFliC did not change expression of the p67phox, p41nox, and p51nox mRNAs (data not shown), but RT-PCR suggested that treatment with rFliC was likely to stimulate the Nox1 mRNA expression (Fig. 6,A). The immunoblotting clearly demonstrated the rFliC-primed induction of Nox1 protein (Fig. 6,B). The induction of Nox1 was associated with the increase in O2− generation when primed with rFliC in the presence or absence of PMA (Fig. 6,C). Recent reports have shown that TLR5-expressing cells including T84 cells activate NF-κB in response to bacterial flagellin (25, 26). As shown in Fig. 6, a potent inhibitor for NF-κB (PDTC) significantly blocked the rFliC-induced increase in O2− production at concentrations over 1 μM (Fig. 6,D), and the induction of Nox1 protein was also concomitantly hampered by PDTC in a dose-dependent manner (Fig. 6 E).
The Nox1 mRNA is dominantly expressed in the colon (6). Intracellular regions of the Nox family are composed of highly conserved domains; therefore, specific Ab for Nox1 had not been available. We have developed polyclonal Abs against human Nox1 useful for immunoblot and immunohistochemical analyses, and found that Nox1 protein was constitutively expressed in surface mucous cells of the guinea pig colon, supporting an in situ hybridization study showing that the Nox1 transcript was expressed in epithelial cells of human colonic mucosa (27). A small amount of Nox1 mRNA is detectable in primary cultures of guinea pig gastric mucosal cells even in LPS-free conditions (12, 14), while Nox1 protein was absent in normal gastric and small intestinal mucosal tissues of both guinea pigs and humans (data not shown). The expression of Nox1 in primary cultures of guinea pig gastric mucosal cells is probably due to oxidative stress during the preparation and cultivation, because the induction of Nox1 is associated with activation of a redox-sensitive transcription factor NF-κB (our unpublished observation). Quiescent guinea pig gastric pit cells (surface mucous cells) in primary culture produced small amounts of O2− (<10 nmol/mg protein/h), but once primed with H. pylori LPS, they increased O2− generation 10-fold in association with the induction of Nox1 (13) (our unpublished observation). In contrast, abundant Nox1 protein was constitutively expressed in surface mucous cells of the guinea pig colon, and these cells in primary culture spontaneously secreted O2− at a higher rate (∼160 nmol/mg protein/h), suggesting that Nox1 in the guinea pig colon is constitutively active. In fact, possible activators, such as IL-1β, TNF-α, epidermal growth factor, TGF-β, IFN-γ, or bacterial components, did not further up-regulate the O2− generation (data not shown).
In contrast to primary cultures of guinea pig LIEC, human colon cancer cell lines (T84 and Caco2 cells) produced only small amounts of O2− (<2 nmol/mg protein/h). RT-PCR analysis roughly estimated that primarily cultured LIEC appeared to express larger amounts of the Nox1 mRNA than the cell lines. Moreover, freshly isolated and cultured guinea pig LIEC constitutively expressed p67phox, its homologue p51nox, and a p47phox homologue p41nox, but they were absent or poorly expressed in T84 and Caco2 cells. Mouse and human p41nox lack the regulatory domain corresponding to the aa 286–340 of human p47phox (16, 17, 18), which may explain why Nox1 of guinea pig LIEC was in a self-activated status to generate O2− without any stimulants. When the p41nox was transfected, T84 cells augmented O2− production in response to PMA. T84 cells expressed a low level of the p51nox mRNA, and overproduction of p51nox further increased O2−-generating capability, but transfection of the p67phox was not effective (Fig. 4 B). P51nox lacks putative domains to interact with p40phox (15, 16, 17, 18), and colonic epithelial cells did not have p40phox. Based on these findings, p51nox, rather than p67phox, may be a physiological partner with Nox1 and p41nox in catalyzing an electron transfer from NADPH to molecular oxygen.
At present, physiological roles of colonic Nox1 are not fully understood. The finding that transfected Nox1 confers mitogenic properties on NIH 3T3 cells bore the scenario that Nox1 may be involved in the process of cell transformation (6). We demonstrated that the terminally differentiated surface mucous cells in the colon constitutively expressed Nox1 protein, suggesting that Nox1 may have other roles besides mitogenic properties. We examined whether Nox1-derived ROS exhibited bactericidal activities. For this purpose, 2 × 107 or 2 × 108 CFU/ml of S. enteritidis was cocultured with guinea pig LIEC for different incubation times (up to 8 h), and the bacterial growth was estimated by colony counts grown on tryptic soy agar for12 h. Guinea pig LIEC did not affect the growth rate of bacteria, and inclusion of SOD and catalase in the culture medium also did not change the growth (data not shown).
Surface mucous cells serve a primary protective role against irritants by providing mucous coat. Recently, these cells have been shown to play an important role in host defense as well, producing proinflammatory mediators after the interaction with pathogenic microbes (28). In fact, stimulation of TLR4 in guinea pig gastric mucosal cells by H. pylori LPS activated NF-κB within 30 min, followed by up-regulation of Nox1 activity within 8 h (11, 14). Enhanced production of ROS further augmented NF-κB activation, leading to prolonged expression of the TNF-α and cyclooxygenase II mRNAs (11, 12). T84 cells specifically responded to rFliC and induced Nox1, although they were insensitive to LPS, CpG DNA, and peptidoglycan. Abreu et al. (29, 30) have also demonstrated that T84 and Caco2 cells are broadly unresponsive to TLR2 and TLR4 signalings. When T84 cells transfected with both p41nox and p51nox vectors were primed with rFliC, they increased O2−-producing ability to higher than 5-fold of the vector alone control (Fig. 6,C). This enhanced O2− production by transfection of T84 cells with p41nox and p51nox cDNAs significantly enhanced the rFliC-stimulated IL-8 release from the cells (Fig. 5 G). These results suggest that the TLR5-mediated up-regulation of Nox1 activity may contribute innate immune response by enhancing inflammatory responses of LIEC, rather than by directly killing pathogenic bacteria.
It has been shown that intestinal epithelial cells, including T84 and Caco2 cells, express TLR5, and bacterial flagellin stimulates the TLR5 signaling, leading to the activation of proinflammatory signals, particularly NF-κB pathway (25, 26). T84 cells are known to show highly polarized expression of TLR5 on the basolateral surface (26, 31). Certain flagellated bacteria are capable of translocating flagellin and stimulating TLR5 (31). Gastric pit cells were sensitive to LPS (11, 13, 14), while T84 cells were not. This difference in the sensitivity may reflect physiological environments: LIEC are always exposed to Gram-negative bacteria. Thus, surface mucous cells of the stomach and colon may use different TLR members to recognize respective pathogenic microbes, activate Nox1, and finally produce defensive mediators. Rapid phosphorylation of TAK1 and TAB 1 with rFliC confirmed that it actually stimulated TLR5 signaling. Stimulation of TLR5 is suggested to activate multiple signaling pathways (25, 26, 32). Transduction of dominant-negative TAK1-adenovirus vector failed to inhibit rFliC-primed Nox1 induction and O2− increase (data not shown). PDTC, however, significantly blocked both responses, which suggest an important role of a putative NF-κB binding site at −253 bp in the human NOX1 5′-flank (GenBank, NT010552). Further studies are necessary to settle this issue.
The present study clearly demonstrated that p41nox and p51nox, novel p47phox and p67phox homologues, respectively, are essential components for Nox1 in achieving a potent oxidase activity. Additionally, the Nox1 may constitute early responses in epithelial cells against pathogens for the host defense.
This study was supported by a grant-in-aid for scientific research from Japan Society for the Promotion of Science (14370184), and Japan Society for the Promotion of Science Research Fellowships for Young Scientists (04127).
Abbreviations used in this paper: ROS, reactive oxygen species; DF, DMEM-Ham’s F-12; Duox, dual oxidase; FAD, flavin adenine dinucleotide; LIEC, large intestinal epithelial cell; NBT, nitroblue tetrazolium; Nox, NADPH oxidase; O2−, superoxide anion; PAS, periodic acid-Schiff; PDTC, pyrrolidine dithiocarbamate; PVDF, polyvinylidene difluoride filter; rFliC, recombinant structural protein of flagella filament; SOD, superoxide dismutase; TAB 1, TAK1-binding protein 1; TAK1, TGF-β-activated kinase 1; TLR, Toll-like receptor.