|Home | About | Journals | Submit | Contact Us | Français|
NADPH oxidase/dual-oxidase (Nox/Duox) family members have been implicated in nuclear factor kappa-B (NFκB)-mediated inflammation and inflammation-associated pathologies. We sought to examine, for the first time, the role of Nox/Duox and NFκB in rats treated with the cooked meat heterocyclic amine carcinogen 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP). In the PhIP-induced colon tumors obtained after 1 year, Nox1, Nox4, NFκB-p50 and NFκB-p65 were all highly overexpressed compared with their levels in adjacent normal-looking colonic mucosa. Nox1 and Nox4 mRNA and protein levels also were markedly elevated in a panel of primary human colon cancers, compared with their matched controls. In HT29 human colon cancer cells, Nox1 knockdown induced G1 cell cycle arrest, whereas in Caco-2 cells there was a strong apoptotic response, with increased levels of cleaved caspase-3, -6, -7 and poly(ADP-ribose)polymerase. Nox1 knockdown blocked lipopolysaccharide-induced phosphorylation of IκB kinase, inhibited the nuclear translocation of NFκB (p50 and p65) proteins, and attenuated NFκB DNA binding activity. There was a corresponding reduction in the expression of downstream NFκB targets, such as MYC, CCND1 and IL1β. The results provide the first evidence for a role of Nox1, Nox4 and NFκB in PhIP-induced colon carcinogenesis, including during the early stages before tumor onset. Collectively, the findings from this investigation and others suggest that further work is warranted on the role of Nox/Duox family members and NFκB in colon cancer development.
Nuclear factor kappa-B (NFκB) is a key transcription factor regulating the expression of genes involved in inflammation, immune modulation and apoptosis, as well as in various stages of cancer development.1–3 In mammals, five members of the NFκB family have been identified, p50, p65 (RelA), RelB, c-Rel and p52.4 These subunits exist as homodimers or heterodimers, under the control of inhibitors of NFκB (IκB). The IκB family comprises IκBα, IκBβ, IκBγ, IκBε, IκBζ, BCL3, p100 and p105. Activation of NFκB involves IκB kinase (IKK)-dependent phosphorylation and degradation of IκB proteins.5 IKK contains the catalytic subunits IKKα and IKKβ, and a regulatory component, IKKγ/NEMO. Nuclear trafficking of NFκB results in the activation of genes encoding cytokines, chemokines, growth factors and antiapoptotic factors.1
A plethora of physiological stimuli activate NFκB. These include lipopolysaccharide (LPS) and proinflammatory cytokines, such as interleukin-6 (IL6) and interleukin-1β (IL-1β), as well as tumor necrosis factor-α (TNF-α), acting via the TNF receptor (TNFR). Activation of NFκB is associated with the production of reactive oxygen species (ROS), and the best-studied ROS-producing enzyme is the phagocyte-derived NADPH oxidase (Nox), which plays a pivotal role during bacterial infection and inflammation.6,7 Non-phagocyte-derived Nox homologues also have been identified, designated collectively as the Nox/dual oxidase (Duox) family, which comprises Nox1, Nox2, Nox3, Nox4 and Nox5, plus Duox1 and Duox2.8 Each Nox/Duox isoform exhibits a distinct cellular and tissue distribution pattern.7 In lung, aberrant expression of Nox2, Duox1 and Duox2 contributes to chronic obstructive pulmonary disease, asthma and cystic fibrosis,9 whereas in forebrain abnormal levels of Nox2 and Nox4 have been implicated in the pathogenesis of schizophrenia.10
Accumulating evidence supports a role for Nox/Duox members in other pathologies, including cancer. Nox1 stimulates mitogenesis, cell transformation and tumorigenesis when ectopically expressed in NIH3T3 fibroblasts and DU-145 prostate epithelial cells, with a corresponding increase in angiogenesis.11,12 Overexpression of Nox1 was observed in prostate, breast and ovarian cancers,13,14 and Nox4 was detected at high levels in glioblastoma cells.15 Information on Nox/Duox involvement in colon cancer is somewhat inconsistent. Geiszt et al.16 first reported that Nox 1 was expressed mainly in differentiated colonic epithelial cells, and Szanto et al.17 detected no statistical difference for NOX1 mRNA expression between adenomas and poorly- or well-differentiated colon adenocarcinomas. Szanto et al.17 concluded that Nox1 is an enzyme that is constitutively expressed in colonic epithelium and is not associated with tumorigenesis. However, using immunohistochemical analyses, Fukuyama et al.18 observed Nox1 overexpression in human colon adenomas and well-differentiated adenocarcinomas, and Laurent et al.19 reported that Nox1 was overexpressed in human colon cancers and was correlated with activating mutations in K-ras.
To clarify the role of Nox/Duox family members during colon carcinogenesis, rats were treated with the heterocyclic amine 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), as reported before,20,21 and the expression levels of Nox/Duox and NFκB were examined before and after the appearance of frank tumors. A panel of primary human colon cancers and human colon cancer cell lines also was studied for expression of NFκB and Nox/Duox isoforms.
Colon tumors and adjacent normal-looking mucosa were from a study in which male F344 rats received intermittent exposure to PhIP alternating with a high-fat diet, as reported elsewhere.20 An interim sacrifice was included to assess early molecular changes, before the onset of tumors. Specifically, 24 h after the last dose of PhIP, rats in each group (n = 5) were euthanized, the colon was removed and opened longitu-dinally, and the mucosa was scraped and frozen in liquid nitrogen before storing at −80°C. The remaining animals in each group (n = 36) were euthanized at 52 weeks (Supporting Information Fig. 1). A complete necropsy examination was performed on each animal.20 This work received prior approval from the Institutional Animal Care and Use Committee.
Ten pairs of primary human colon cancers and their matched adjacent normal-looking tissues were kindly provided by Steven F. Moss, M.D. and Lelia Simao (Rhode Island Hospital, Providence, RI). The patients (6 female, 4 male, 53–93 years of age) had been diagnosed with adenocarcinoma of the colon.
Frozen colon tumor samples and their matched controls were thawed, and mRNA was extracted using the RNeasy kit (Qiagen, Valencia, CA). RNA (2 μg) was reverse-transcribed in 20 μl of 1× RT buffer, containing 10 U RNase inhibitor (Invitrogen, Carlsbad, CA), 0.5 mM each dNTP, 4 U Omni-script Reverse Transcriptase (Qiagen) and 50 ng random hexamers (Invitrogen). Primers were as listed in Supporting Information Table I. Forty cycles of PCR were run on an Opticon Monitor 2 system (Finnzymes, Finland), in 20 μl total reaction volume containing cDNAs, SYBR Green I dye (DyNAmo master solution, Finnzymes) and primer set. The PCR conditions were 95°C/10s, 58°C/20s and 72°C/20s, except for rat Nox1 and rat Nox4, in which the annealing temperature was 60°C. The amount of specific mRNA was quantified by determining the point at which the fluorescence accumulation entered the exponential phase (Ct), and the Ct ratio of the target gene to glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was calculated for each sample. At least two separate experiments were performed for each sample. In some experiments, malignant regions were microdissected from PhIP-induced colon tumors using a Zeiss PALM MicroBeam IV Laser Capture System (Carl Zeiss, Thorn-wood, NY), and compared with microdissected normal colon. Briefly, mRNA from captured cells was extracted using an RNeasy Micro Kit (Qiagen) and cDNA was synthesized using the Superscript III kit (Invitrogen). Forty-five cycles of qPCR (95°C/10s, 60°C/10s, 72°C/10s) were run on a Roche Light-Cyler 480 II (Roche, Indianapolis, IN), in a 10 μl reaction containing primer set, cDNA and SYBR Green I dye (Roche).
ProteoExtract Native membrane protein extraction kit (EMD Biosciences, San Diego, CA) was used for enrichment of membrane proteins, whereas NE-PER reagents (Pierce Biotechnology, Rockford, IL) were used to separate cytoplasmic and nuclear fractions. Membrane-enriched fractions were subjected to SDS-PAGE and immunoblotted using anti-Nox1 (H-75, sc-25545, 1:400 dilution, Santa Cruz Biotechnology, Santa Cruz, CA) and anti-Nox4 (rat, 1:100 dilution, abcam, Cambridge, MA, ab41886; human, 1:500 dilution, Novus, Biologicals, NB110-58851, Littleton, CA). The correct molecular weights were confirmed by reference to the marker ladder on each blot (Nox1 65-kD, Nox4 67-kD), and antibody specificity was corroborated via the use of blocking peptides (sc-5821p against sc-5821 Nox1) (Santa Cruz), 1:200 dilution; Nox4 peptide (ND110-58851 PEP Novus Biologicals), 1:250 dilution). Whole cell lysates were immunoblotted with rabbit polyclonal antibody to IKKα and IKKβ (1:1000 dilution, Cell Signaling, Nos. 2682 and 2684) or phospho-IKKα/β (1:500 dilution, Cell Signaling, no. 2681). Cytoplasmic extracts were probed with antibodies to cleaved caspase 3 (1:1000 dilution, Cell Signaling, no. 9661), cleaved caspase 6 (1:1000 dilution, Cell Signaling, no. 9761), and cleaved caspase 7 (1:1000 dilution, Cell Signaling, no. 9492). Nuclear fractions were immunoblotted with polyclonal rabbit antibody to NFκB p50 (1:600 dilution, Santa Cruz, sc-7178), NFκB p65 (1:600 dilution, Santa Cruz, sc-109), and poly-(ADP-ribose)polymerase (PARP, 1:1000 dilution, Cell Signaling, no. 9532). Amido black staining was used to ensure equal protein loading, followed by β-actin. For nuclear extracts, histone H1 also was used as a loading control (mouse monoclonal antibody sc-8030, Santa Cruz, 1:500 dilution). Proteins were visualized by Western Lightning Chemiluminescence Reagent Plus (Perkin–Elmer Life Sciences, Boston MA), with quantification via an AlphaInnotech photodocumentation system and associated software (AlphaInnotech, San Leandro, CA).
Rat colon tumors were processed to paraffin, sectioned at 4 lm, and placed on charged slides. Sections were rehydrated through xylene, 100% ethanol, 95% ethanol, 80% ethanol and water. Antigen retrieval was carried out in a microwave pressure cooker for 10 min, followed by 20 min at room temperature. Antigen retrieval solution was Dako Target Retrieval Solution pH 6.0 (Dako, Carpentaria, CA). Slides were washed in water and loaded into a Dako autoimmunostainer. Endogenous peroxides were blocked with 3% H2O2 in TBST (Dako Tris-buffered saline with Tween 20) for 10 min, and slides were then washed in TBST. Dako serum-free protein block was applied for 10 min, followed by a burst of air to blot the slides. Incubation with the primary antibody was for 30 min (Nox4, 1:250 dilution, Novus Biologicals, NB110-58851, Littleton, CA). As immunohistochemical controls, the corresponding blocking peptide (see above) was used to confirm antibody specificity, and Dako Universal Negative Rabbit control was used in place of the primary antibody. After washing in TBST, Dako Envision+ anti-rabbit HRP was applied for 30 min, followed by Nova Red (Vector Laboratories, Burlingame, CA) and hematoxylin (Dako) counter staining.
The same protocol was used to immunostain Nox4 in human tissues. Tissue microarrays (TMAs) were constructed using a Beecher Instruments MTA-1 tissue arrayer (Beecher Instruments, Sun Prairie, WI). At least duplicate tumor samples were taken from donor tissue blocks, and a retrospective analysis for outcome assessment was based on detailed clini-copathological information linked to the TMA specimens. For further details, see Ashktorab et al.22
Human colorectal cancer lines HT29 and Caco2 (American Type Culture Collection, Manassas, VA) were maintained in McCoy's 5A medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS, Hyclone Laboratories), 100 units/ml penicillin, and 100 μg/ml streptomycin at 37°C in 5% CO2. SmartPool siRNA against human NOX1 and non-specific control siRNAs were purchased from Dhar-macon and transfected according to the manufacturer's instructions. Briefly, cells (1 × 105) were seeded in 6-well plates overnight. The media was aspirated and replaced with fresh antibiotic free-transfection media containing Dharma-FECT (4 μl for Caco2 cells, 8 μl for HT29 cells), 100 nM NOX1 siRNA, or 100 nM non-specific siRNAs (non-target controls), or DharmaFECT only (mock controls). Untreated controls received antibiotic-free media only. At 48 and 72 h after transfection, cells were harvested for flow cytometry, morphological assessment of apoptosis using Acridine Orange/Ethidium Bromide staining, and protein and mRNA analyses. In additional experiments, 72 h after transfection of NOX1 siRNAs the media was aspirated and replaced with fresh media containing 1.2 μg/μl of LPS (Sigma), or media alone. Thirty minutes later the cells were washed twice with PBS and lysed in IP buffer supplemented with protease inhibitor cocktail and phosphatase inhibitor cocktail I and II (Sigma). Insoluble debris was removed by centrifugation at 10,000g for 5 min at 4°C. Cell lysates were subjected to SDS-PAGE, as described above, and immunoblotted for phospho-IKKα/β (p-IKK, Ser176/180, 1:500, Cell Signaling, no. 2697).
Cells (5 × 103) in 100 μl media were seeded in 96-well plates overnight and transfected with siRNA, as described above. At 24, 48, 72 and 96 h after transfection 3-(4,5-dimethythiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was added and incubated for 3 h, followed by 100 μl of 10% SDS in 0.01 N HCl. Formation of colored formazan dye was assessed colorimetrically at 550 nm.
Cells treated with NOX1 siRNA for 72 h (see above) were harvested in cold PBS, fixed in 70% ethanol, and stored at −20°C for 48 h. Fixed cells were washed with PBS and resuspended in propidium iodide/Triton X-100 staining solution containing RNase A. Samples were incubated in the dark for 30 min before determining the DNA content on an EPICS XL Beckman Coulter flow cytometer. Cell cycle distribution was assessed using Multicycle Software (Phoenix Flow Systems, San Diego, CA).
Cell suspensions (25 μl) were incubated with 1 μl acridine orange/ethidium bromide solution (50 μg/ml of each reagent in PBS), mixed gently, and placed onto a microscope slide. Cell morphology was examined under a fluorescence microscope for signs of chromatin condensation, fragmented nuclei and/or membrane blebbing, indicative of apoptosis. At least 500 cells were counted per treatment, and three independent experiments were performed.
Nuclear extracts were examined via enzyme-linked immunosorbent assay (NFκB-p50 transcription factor assay kit, Cat no. 10006912, Cayman Chemical, Michigan), according to the manufacturer's instructions.
Unless stated otherwise, results shown in each figure were from a single experiment, and are representative of the findings from three or more independent experiments. Data were expressed as mean ± standard error (mean ± SE), and comparisons between control and treatment groups were made using the paired t-test (SigmaPlot 8.0). In the figures, significant results were indicated as follows: *p < 0.05, **p < 0.01 and ***p < 0.001.
In the colon tumors obtained from rats after 1 year, Nox/Duox mRNA levels were highly overexpressed compared with adjacent normal looking tissue (T vs. N in Fig. 1a, solid vs. open bars). The increase was highly significant for Nox1 (133-fold, p < 0.001) and Nox4 (6.5-fold, p < 0.001), and significant for Duox2 (2.1-fold, p < 0.05), but not significant for Nox2 (2.0-fold, p > 0.05, NS). Adenocarcinoma cells also were captured by laser microdissection from PhIP-induced colon tumors, and these areas expressed significantly increased levels of Nox1 mRNA (Supporting Information Fig. 2). Because Nox1 and Nox4 were most highly overexpressed at the mRNA level, the corresponding proteins were examined by immunoblotting (Fig. 1b). After normalizing to bactin, densitometric analyses confirmed that Nox1 and Nox4 proteins were elevated 2- to 3-fold in tumors versus adjacent normal-looking tissue (p < 0.01, n = 12). A doublet was detected for Nox1, with the upper band stronger in some tumors and the lower band dominating in others (Fig. 1b, upper panel). This Nox1 doublet has been observed previously, for example in immortalized human keratinocytes,23 Ras(+) NIH3T3 cells,24 rat smooth muscle cells,25 and human T84 large intestinal cells,26 but the significance remains unclear. Subsequent studies examined nuclear NFκB levels (Fig. 1c); after normalizing to histone H1, the increase in both NFκB-p50 and NFκB-p65 protein expression was highly significant in tumors compared with adjacent normal-looking tissue (p < 0.001, n = 12). qPCR analyses revealed downstream NFκB targets to be increased in colon tumors (Fig. 1d), including IL1β (2.2-fold, p < 0.05), IL6 (2.9-fold, p < 0.01), TNFα (16.6-fold, p < 0.001) and TNFR1 (4.0-fold, p < 0.01).
We next examined the expression of Nox/Duox family members in a panel of primary human colon cancers. By conventional RT-PCR (Fig. 2a), cancer specimens nos. 1, 2, 3, 6 and 10 had high levels of both NOX1 and NOX4 mRNA, whereas NOX2 and DUOX2 mRNA levels were more inconsistent, being higher in some cancers and lower or undetectable in others. These findings were confirmed by qPCR, which revealed a 2.5- to 3.0-fold increase in NOX1 and NOX4 mRNA expression in primary colon cancers versus matched controls (p < 0.05, Fig. 2b). No significant difference was seen for NOX2 and DUOX2 in the qPCR analyses. Thus, Nox1 and Nox4 were selected for subsequent immunoblotting studies (Fig. 2c). After normalizing to β-actin, the relative expression in cancers versus matched controls was as follows (mean ± SE, n = 10): Nox1, 2.36 ± 0.40 versus 1.08 ± 0.09 (p < 0.01); Nox4, 2.87 ± 0.22 versus 1.03 ± 0.09 (p < 0.001).
In immunohistochemistry studies, focal areas of intense Nox4 staining were detected in PhIP-induced colon tumors (Fig. 3a, right), whereas normal-looking tissue adjacent to the tumor had no such focal Nox4 expression (Fig. 3a, left). Staining of serial sections in the absence (Fig. 3b) and presence of a Nox4 blocking peptide (Fig. 3c) confirmed the spec-ificity of the primary antibody. Cells containing high levels of Nox4 protein typically were interspersed with other cells having little or no such staining. In some of the rat colon tumors, however, islands of intense Nox4 staining were observed, surrounded by areas that were predominantly negative (Fig. 3d). These “islands” corresponded to the more differentiated regions within a tumor. In human colon cancer tissue microarrays, a core occasionally stained entirely negative for Nox4 (not shown), but most cores had areas that were strongly Nox4 positive (Fig. 3e). Nox4 was detected in the cancer epithelial cells, and was absent from the surrounding stroma (Fig. 3f). In some cores, islands of Nox4 staining were associated with differentiated regions within the human colon cancer (Fig. 3g), as seen in the rat (Fig. 3d). Our interpretation is that Nox4 staining patterns were similar in rat and human colon tumors, and that the PhIP model might provide useful insights into the role of Nox4 in colon cancer development.
Attempts to immunolocalize Nox1 protein in rat and human colon tissues were unsuccessful. In our hands, antibodies from commercial sources (cs-25545 from Santa Cruz, ab78016 from abcam, and LS-B1832 from Lifespan Biosciences, Seattle, WA) gave high background expression with non-specific labeling, despite exhaustive protocol modifications and antibody dilution experiments (data not shown).
HT29 and Caco2 cells were selected from among a panel of human colon cancer cell lines due to their high NOX1 mRNA content (data not shown), and NOX1 siRNA was used as a knockdown strategy. Based on the results of three replicate qPCR experiments, NOX1 mRNA levels in HT29 cells typically were diminished by 48% at 48 h and 92% at 72 h, whereas in CaCo2 cells the reduction was 60% at 48 h and 94% at 72 h. Despite the similar knockdown efficiency for NOX1 mRNA at 72 h, Nox1 protein expression was reduced more dramatically in Caco2 cells than in HT29 cells (Fig. 4a). Densitometry measurements were performed on Nox1 normalized to β-actin. Compared with the non-target siRNA control at 72 h, the relative expression of Nox1 protein was reduced by 48% in HT29 cells (1.03 ± 0.12 vs. 0.54 ± 0.07, p < 0.05) and by 82% in Caco2 cells (0.96 ± 0.07 vs. 0.18 ± 0.05, p < 0.001).
Inhibitory responses in the MTT assay following Nox1 knockdown were significant at all time-points after 24 h in Caco2 cells, and after 72–96 h in HT29 cells (Fig. 4b). To clarify whether the MTT assay data might be indicative of changes in cell cycle kinetics or apoptosis, further experiments were performed. In HT29 cells, changes in the cell cycle were detected 72 h post-transfection, as follows: 77.4 ± 2.5% versus 52.5 ± 2.9% in G1, 14.6 ± 2.2% versus 30.5 ± 1.6% in S, and 8% versus 16.6% in G2/M (NOX1 siRNA vs. nontarget siRNA, mean ± SE, n = 3 separate experiments). No significant changes were detected in the proportion of HT29 cells undergoing apoptosis (2–3% in all groups, data not shown). In contrast, Caco2 cells exhibited clear hallmarks of apoptosis (Fig. 4c, inset), and the percentage of apoptotic cells at 72 h increased from 1–2% after exposure to non-target siRNA to >30% following treatment with NOX1 siRNA (p < 0.001). In Caco2 cells treated with NOX1 siRNA, there was an increase in the cleaved (active) forms of Caspase-3, -6 and -7, as well as elevated PARP cleavage (Fig. 4d). Taken together, these findings suggested that human colon cancer cells undergo cell cycle arrest or apoptosis, depending on the cell line and extent of Nox1 knockdown.
In HT29 cells, low constitutive levels of phospho-IKK were detected in the absence of LPS treatment, whereas an increase in phospho-IKK was evident within 30 min of LPS exposure (Fig. 5a, LPS− vs. LPS+, respectively). Interestingly, LPS-induced phospho-IKK expression was blocked in cells that had been transfected with NOX1 siRNA (Fig. 5a, top right). When the nuclear extracts were examined 6 h after treatment with LPS or media alone, NFkB-p50 and -p65 proteins were reduced markedly by Nox1 knockdown (Fig. 5b). Knockdown of Nox1 reduced significantly the DNA binding activity of NFκB in nuclear extracts (p < 0.05, Fig. 5c), and lowered the mRNA expression levels of downstream NFκB targets, such as MYC, CCND1 and IL1β (p < 0.05, Fig. 5d).
Oxidative stress in the colon originally was thought to involve Nox2, and was ascribed to resident and recruited phagocytic cells with important roles in host-defense mechanisms.7,27 However, this idea has evolved with the discovery of non-phagocyte derived Nox/Duox homologues. Nox1 is expressed in gastric pits and colonic epithelial cells and is required for normal gut physiology. For example, Nox1-induced oxidative stress, involving intermediates such as and H2O2, regulates mucosal 5-hydroytryptamine levels, which affects normal secretion and motility in the colon.28 Transcriptional activation of NOX1 by IFN-γ produces in colonic epithelial cells, contributing to mucosal host defense mechanisms.26 Duox2 is expressed in normal human colorectal barrier epithelial cells,16 and its loss following duox2 silencing in flies markedly enhances mortality following infection.29 Nox family members also have been implicated in Crohn's disease and other inflammatory bowel disorders.7
Although the cancer-associated expression of Nox1 has been reported in human stomach,30 there is some debate over the precise role of Nox1 in colon cancer, as noted in the introduction. The most recent study found that Nox1 was overexpressed in human colon cancers and was correlated with activating mutations in K-ras.19 Colon tumors induced by PhIP and other heterocyclic amines lack mutations in K-ras, but harbor genetic changes in β-catenin or Apc.20,21,31 We performed mutation screening for the corresponding genes in the PhIP-induced colon tumors reported here, but observed no correlations with the expression of Nox1 or other Nox/Duox members (data not presented). Nonetheless, results from the present investigation support a role of Nox1 in human colon cancer and expand this observation to Nox4, including the first such evidence using a rat colon carcinogenesis model.
Interestingly, increased Nox1, NFκB-p50, and NFκB-p65 protein levels were detected in colonic mucosa obtained from rats immediately after completing PhIP treatment, several weeks before the appearance of frank tumors (Supporting Information Fig. 3). Thus, Nox1, NFκB-p50, and NFκB-p65 might play a role during the early stages of PhIP-induced colon carcinogenesis. Charalambous et al.32 recently postulated that NFκB-p65 and IKKα are early post-initiation events in human colorectal tissues, perhaps involved in tumor progression. Heterocyclic amine mutagens have been studied extensively,31–36 but this is the first report to show that PhIP-induced colon tumors have increased nuclear NFκB-p50 and NFκB-p65 expression, and higher levels of IL1β, IL6, TNF-α and TNFR1.
In HT29 cells, LPS-induced phosphorylation of IKK was blocked following Nox1 knockdown. This suggests that NFκB activation induced by cytokines, growth factors, and other external stimuli might be mediated, at least in part, by Nox1. Even in the absence of LPS treatment, however, Nox1 knockdown lowered nuclear NFκB-p50 and NFκB-p65 protein expression, attenuated NFκB-p50 DNA binding activity, and reduced transcription of NFκB downstream targets. We were also interested in the reverse scenario, namely that loss of NFκB might attenuate Nox1 levels. However, NFκB knockdown did not alter Nox1 mRNA or protein expression in human colon cancer cells (data not presented), suggesting that alternative mechanisms regulate Nox1.
Finally, we observed that Nox1 knockdown in HT29 cells resulted in the accumulation of cells in G1, whereas the same knockdown strategy in Caco2 cells strongly induced apoptosis. These cell-specific effects are intriguing and may suggest a causal role in tumorigenesis. However, the findings should be interpreted cautiously, since it is still not entirely clear whether Nox/Duox overexpression is a cause of tumorigenesis, or whether it results from tumor formation. The data showing increased Nox1 expression in colonic mucosa several weeks before colon tumor formation (Supporting Information Fig. 3) suggest a causal role, but additional work is needed to confirm this possibility. Moreover, immunohistochemical analyses of rat colon tumors revealed increased NFκB expression in areas that were either positive or negative for Nox4 (data not presented). This is perhaps not surprising, given that multiple pathways activate NFκB,2 and further studies appear to be warranted on other Nox/Duox family members in the context of NFκB signaling and colon cancer development. The present report is the first to show that PhIP-induced colon tumors have increased Nox/Duox expression and NFκB activation, but these events are probably not specific to heterocyclic amines, and other colon carcinogens may act in a similar fashion.
In summary, we provide here the first evidence for the involvement of Nox1, Nox4 and NFκB during PhIP-induced colon carcinogenesis, and provide further support for a role of Nox/Duox isoforms in primary human colon cancers and colon cancer cell lines. In cultured human colon cancer cells, Nox1 knockdown blocked LPS-induced phosphorylation of IKK, reduced nuclear NFκB levels and DNA binding activity, and attenuated the transcription of downstream NFκB targets. We conclude that the interplay between Nox/Duox family members and NFκB signaling during colon cancer development is worthy of further investigation.
The authors thank Dr. Steven F. Moss and Dr. Lelia Simao (Rhode Island Hospital, Providence, RI) who kindly provided the panel of primary human colon cancers and their matched controls. They thank Dr. Jeffery Green-wood for access to the Zeiss laser-capture microdissection instrument in the Cell Imaging and Analysis (CIA) Core of the Environmental Health Sciences Center. Partial support for R.H.D. was provided by the Foundation for Promotion of Cancer Research, Tokyo, Japan.
Grant sponsor: National Cancer Institute; Grant numbers: CA90890, CA65525, CA90176, CA122959; Grant sponsor: National Institute of Environmental Health Sciences; Grant number: P30 ES00210; Grant sponsor: Foundation for Promotion of Cancer Research
Additional supporting information may be found in the online version of this article.