| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Muchmore Laboratories for Infectious Disease Research, Department of Veterans Affairs Medical Center, Oklahoma City, OK 73104, USA
2 Department of Medicine, University of Oklahoma Health Sciences Center, Oklahoma City, OK 73104, USA
3 Department of Surgery, University of Oklahoma Health Sciences Center, Oklahoma City, OK 73104, USA
4 Research Service, Department of Veterans Affairs Medical Center, Oklahoma City, OK 73104, USA
Correspondence
Mark M. Huycke
mark-huycke{at}ouhsc.edu
Received 4 December 2007
Accepted 10 June 2008
B (NF-B) signalling, apoptosis and cell-cycle regulation. Colon biopsies showed no histological abnormalities by haematoxylin and eosin staining. Immunohistochemical staining, however, detected NF-B activation in tissue macrophages using antibodies to the nuclear localization sequence for p65 and the F4/80 marker for murine macrophages. Similarly, haematin-starved E. faecalis strongly activated NF-B in murine macrophages in vitro. Furthermore, primary and transformed colonic epithelial cells activated the G2/M checkpoint in vitro following exposure to haematin-starved E. faecalis. Modulation of this cell-cycle checkpoint was due to extracellular superoxide produced as a result of the respiratory block in haematin-starved E. faecalis. These results demonstrate that the uniquely dichotomous metabolism of E. faecalis can significantly modulate gene expression in the colonic mucosa for pathways associated with inflammation, apoptosis and cell-cycle regulation.
Abbreviations: CRC, colorectal cancer; DAPI, 4',6-diamidino-2-phenylindole; FBS, fetal bovine serum; FITC, fluorescein isothiocyanate; HRP, horseradish peroxidase; IL, interleukin; MnSOD, manganese superoxide dismutase; NF-B, nuclear factor-B; NLS, nuclear localization sequence; p.i. post-inoculation; qRT-PCR, quantitative real-time RT-PCR.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Perhaps the best evidence supporting a role for colonic commensals in CRC derives from genetically engineered mice that develop intestinal tumours when conventionally colonized but have fewer tumours or no pathology when housed in pathogen-free or germ-free environments (Balish & Warner, 2002; Chu et al., 2004; Dove et al., 1997; Engle et al., 2002; Kado et al., 2001; Kim et al., 2005; Maggio-Price et al., 2006). The mechanisms by which commensals trigger inflammation or initiate genomic instability (a characteristic feature of sporadic CRC) remain uncertain. Commensals can modulate the intestinal mucosa through the metabolism of faecal steroids, by producing short-chain fatty acids and by inducing host genes (Augenlicht et al., 2002; Bäckhed et al., 2005; Debruyne et al., 2001; McGarr et al., 2005). None of these effects, however, is known to initiate or promote genomic or epigenetic changes in epithelial cells as antecedents to oncogenic transformation. Commensals that cause epithelial-cell DNA damage, in contrast, are more likely to initiate chromosomal instability than bacteria that modulate epithelial-cell metabolism but are otherwise not mutagenic.
Enterococcus faecalis is a Gram-positive minority constituent of the colonic microbiota that can directly damage epithelial-cell DNA (Huycke et al., 2002), promote chromosomal instability through a macrophage-induced bystander effect (Wang & Huycke, 2007) and trigger colitis and cancer in interleukin (IL)-10 knockout mice (Balish & Warner, 2002; Kim et al., 2005). The bystander effect refers to oxidatively stressed or irradiated cells that produce chromosomal instability in neighbouring cells through the production of diffusible mutagens (Lorimore & Wright, 2003). These effects are derived in part from the redox-active phenotype of E. faecalis that occurs when this micro-organism is grown without haematin. E. faecalis cannot synthesize porphyrins and requires this nutrient to form functional cytochrome bd, establish a proton gradient and respire (Bryan-Jones & Whittenbury, 1969; Ritchey & Seeley, 1974). In contrast, growth of E. faecalis in the absence of haematin leads to a block in respiration and the production of superoxide, hydrogen peroxide and hydroxyl radical (Huycke et al., 2001, 2002). These reactive oxygen species are known to damage DNA (Marnett, 2000), although it remains to be determined whether this redox-active phenotype promotes inflammation and colon cancer in IL-10 knockout mice that are monoassociated with E. faecalis.
To explore further the effect of E. faecalis on the colonic mucosa, we developed an in vivo colonic ligation model to examine host gene expression in wild-type mice. This model preserves tissue architecture and homeostasis, and permits an analysis of gene expression in stromal and epithelial-cell compartments. We compared E. faecalis supplied with haematin with E. faecalis grown without haematin to determine the effect of altered bacterial metabolism on colonic mucosal gene expression. We believe this comparison reflects the naturally occurring potential dichotomous metabolism of E. faecalis in the intestinal environment. We found the induction and suppression of a small subset of genes involved in cell-cycle regulation, apoptosis and nuclear factor-B (NF-B) signalling by haematin-starved E. faecalis. These findings indicate that the colonic mucosa rapidly responds to alternate metabolisms of E. faecalis and suggest novel mechanisms by which this commensal may promote inflammatory or potentially transforming events.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
20 µmol min–1 per 109 c.f.u. in vitro), promotes chromosomal instability, and produces colonic inflammation and cancer when monoassociated with IL-10 knockout mice (Balish & Warner, 2002; Huycke & Moore, 2002; Kim et al., 2005; Wang & Huycke, 2007). E. faecalis was grown overnight in closed tubes using brain heart infusion broth (BD Diagnostics) with or without 10 µM haematin (Sigma), and washed in sterile PBS before use in experiments. E. faecalis cannot synthesize porphyrins and is unable to form functional cytochrome bd unless supplied with exogenous haematin (Huycke, 2002). Cytochrome bd is one of two terminal quinol oxidases expressed by E. faecalis that, when active, allows oxidative phosphorylation, promotes growth and suppresses extracellular superoxide (Huycke et al., 2001). Growth with haematin attenuates superoxide production by >10-fold, an effect that persists for >6 h in vitro. For all experiments, haematin-replete growth of E. faecalis was confirmed by measuring the attenuation of superoxide using a ferricytochrome c assay as described previously (Huycke et al., 1996). Unless otherwise specified, all chemicals were of analytical or molecular biology grade from Sigma. Murine colonic ligation model. To assess the short-term effects of haematin-starved E. faecalis on the intact colonic mucosa, we developed an intestinal ligation model. The technique is analogous to the ileal loop model used to investigate diarrhoeal toxins. Conventionally housed and fed 25–28 g adult male BALB/c mice (Jackson Laboratory) were anaesthetized using 1–2 % isoflurane in a carrier gas composed of 95 % O2 and 5 % CO2. Through a 5 mm midline abdominal incision, the proximal colon was identified at its juncture with the caecum. Two ligatures 0.5 cm apart were placed around the colon and a 2 mm incision was made into the colon. A gavage needle was inserted into the colon and the contents were completely flushed through the rectum using sterile PBS. The rectum was closed with a purse-string ligature and the colon backfilled with 1.0 ml PBS alone or PBS containing enterococci at a concentration of 1x108 c.f.u. ml–1. Immediately preceding instillation, D-glucose was added to the PBS or the bacterial inoculum to a final concentration of 5 mM. This sugar initiates extracellular superoxide production by E. faecalis, but not for bacteria grown with haematin (Huycke et al., 2001). Following inoculation, both colonic ligatures were tied to prevent backflow of enterococci and peritoneal contamination from proximal intestinal contents. Care was taken to preserve blood flow to the colon. The surgical area was washed with sterile PBS, the colon was gently returned to the peritoneal cavity, the abdominal incision was closed and mice were allowed to recover.
At 1 or 6 h post-inoculation (p.i.), mice were anaesthetized, the abdomens reopened and the colons surgically removed. The surgical manipulations were well tolerated, with only one mouse not surviving to the end of the protocol. Contents were cultured for enterococci as described previously (Huycke et al., 1992). Colon biopsies of 5 mm were obtained for histopathology and immunohistochemistry. Biopsies were examined using haematoxylin and eosin staining, and a modified Brown and Brenn stain. The remaining colon segments were opened longitudinally and the mucosal surfaces scraped with sterile razors for RNA extraction. Biopsies were fixed in formalin and scrapings were snap frozen in liquid nitrogen. Mice (n=20) were exposed to E. faecalis or PBS (n=6) with independent experiments analysed by group (1 and 6 h p.i.) for mice exposed to haematin-starved E. faecalis (n=5 per group), E. faecalis grown with haematin (n=5 per group) or PBS (n=3 per group). The animal protocol was approved by the Animal Studies Subcommittee of the Veterans Affairs Research and Development Committee.
Gene expression, network response and transcriptional regulatory element analyses.
Total RNA was isolated from colonic scrapings for mice exposed to haematin-replete or haematin-starved E. faecalis at 1 and 6 h p.i. using an Atlas pure total RNA labelling system (BD Biosciences Clontech). Probes were synthesized by reverse transcription using [
-33P]dATP. As the quantities of mucosal scrapings from individual colons were small, each sample was used in its entirety for probe synthesis. Extracted and labelled cDNA probes were hybridized overnight to 5000 cDNA murine arrays (BD Biosciences Clontech). Separate arrays (n=5 per group for a total of 20 arrays) were used for each probe prepared from colon scrapings. After high-stringency washes, membranes were quantified (Storm 820 PhosphorImager; Amersham Biosciences) and expression of individual genes was determined as absorbance readings minus background (ArrayVision software; Imaging Research). Significantly upregulated and downregulated genes were analysed using GeneSpring software version 6.2 (Silicon Genetics). After background subtraction, raw signals were normalized per spot and by array, using an intensity-dependent Lowess protocol. Signal intensities were normalized to the 50th percentile, and comparisons between array results at the 1 and 6 h time points were made using Student’s t-test with P<0.005 considered significant. Fold changes were calculated using the GeneSpring fold change filter option. The Benjamini–Hochberg method was used to correct for multiple testing and to minimize false discovery rates.
Biologically relevant response networks for significantly modulated genes were constructed using Ingenuity Pathways Analysis (Ingenuity Systems; www.ingenuity.com). The Ingenuity Pathways Knowledge Base is the largest curated database on mammalian biology in the published literature. Findings on genes in human, mouse and rat studies from peer-reviewed publications are encoded into an ontology by content and modelling experts. Manual extraction and curation identifies specific interactions that result in fewer false positives than automated methods. Networks are algorithmically generated based on their connectivity, and pathways of highly interconnected genes are identified by statistical likelihood testing (Calvano et al., 2005).
In silico analysis of differentially expressed genes was performed for transcription factor-binding sites using the web-based Promoter Analysis and Interaction Network Tool software (Vadigepalli et al., 2003). Comparisons were carried out using all genes on the murine array as the reference library.
Immunohistochemistry.
Immunohistochemical analysis of the p65 component of NF-B was performed on serial sections of paraffin-embedded murine colon tissue. Antigen retrieval of deparaffinized sections was performed using a decloaking chamber (Biocare Medical) with citrate buffer or 0.1 % pronase (Dako) and processed using the Sequenza staining method (Thermo Scientific). Endogenous peroxidase activity was quenched using peroxidase-blocking reagent (Dako) followed by a blocking step with buffer containing 1 % BSA (Jackson ImmunoResearch), 1 % normal horse serum (Jackson ImmunoResearch), coldwater fish gelatin and Tween 20. Sections were stained using nuclear localization sequence (NLS)-specific anti-p65 antibody (diluted 1 : 300; Rockland Immunochemicals) or anti-F4/80 mAb (diluted 1 : 150; AbD Serotec). The former antibody recognizes the NLS on p65 that is masked by IB. The F4/80 antigen is a surface marker expressed by mature murine macrophages (Austyn & Gordon, 1981). After primary incubation with NLS-specific anti-p65 antibody, sections were incubated in horseradish peroxidase (HRP)-labelled EnVision+ (Dako). Sections stained with anti-F4/80 antibody were developed using anti-rat secondary antibody (diluted 1 : 1000; Jackson ImmunoResearch), followed by incubation with ready-to-use streptavidin–HRP solution (Dako). Following incubation, sections were developed with 3,3-diaminobenzidine substrate or Bajoran Purple (Biocare Medical) and counterstained with haematoxylin (Biocare Medical). The distribution of positive cells per field (magnification 100x) between groups was assessed in a randomized and blind fashion, and compared using ridit analysis, with P<0.05 considered significant (Fleiss, 1981). This method assumes that discrete measures represent intervals in an underlying continuous distribution without any assumptions about the distribution. Ridits range from 0 to 1 and the ridit for the control (or comparator) distribution is 0.50. A mean ridit is >0.50 when more than half of the time randomly selected measures from the experimental distribution have a value greater than randomly selected measures from the control distribution.
Colon sections were processed for netrin-1 immunohistochemistry using UltraVision LP detection system HRP polymer and AEC chromogen (LabVision). Sections were blocked with H2O2 for 10 min to inhibit endogenous peroxidase activity, followed by washes in Tris-buffered saline with Tween 20 at pH 8.0. Following antigen retrieval, Ultra V block (Dako) was applied for 5 min followed by a 1 : 20 dilution of rabbit anti-netrin-1 (Ab-1) primary antibody or control peptide following the manufacturer’s instructions (Calbiochem). Sections were counterstained with Immuno* master haematoxylin (American Master*Tech Scientific).
Cell lines. Chromosomally stable HCT116 colonic epithelial cells (AmericanTypeCultureCollection) were grown in 5 % CO2 at 37 °C using McCoy’s 5A medium modified by L-glutamine and 25 mM HEPES (Invitrogen) and supplemented with 10 % fetal bovine serum (FBS). RAW264.7 murine macrophages (American Type Culture Collection) were grown under the same conditions using Dulbecco’s modified Eagle’s medium modified with 4.5 g glucose l–1 and L-glutamine (Invitrogen) and supplemented with 10 % FBS. For experiments involving co-incubation with E. faecalis, bacteria were diluted to 1x109 c.f.u. ml–1 in fresh medium without FBS. YAMC cells are a non-transformed intestinal epithelial cell line derived from healthy tissue and were a gift from the Ludwig Institute for Cancer Research (Whitehead et al., 1993). These cells were grown in 5 % CO2 at 33 °C using RPMI 1620 (Invitrogen) supplemented with 5 % FBS, 5 U recombinant murine gamma interferon (PeproTech) ml–1 and ITS Premix (BD) according to the manufacturer’s instructions. After treatment, cells were washed with PBS and complete medium was added containing gentamicin (10 µg ml–1) and penicillin (100 U ml–1) to kill any remaining extracellular bacteria. For H2O2-treated cells, catalase (1200 U ml–1) was also included in the complete medium to eliminate residual H2O2.
Immunofluorescent assay.
To visualize NF-B activation, E. faecalis-treated RAW264.7 cells and HCT116 cells grown on chambered slides were fixed with paraformaldehyde and incubated with polyclonal anti-p65 IgG (diluted 1 : 100; Santa Cruz Biotechnology). A fluorescein isothiocyanate (FITC)-conjugated IgG (diluted 1 : 200; Santa Cruz Biotechnology) was used as the secondary antibody and nuclei were counterstained with 4',6-diamidino-2-phenylindole (DAPI) prior to laser-scanning confocal microscopy (LSM-510 META; Zeiss).
NF-B–luciferase reporter assay.
To verify NF-B activation, RAW264.7 cells were transfected with the pNFB-Luc reporter vector (Clontech) using Lipofectin reagent (Invitrogen). Transfected cells were treated with E. faecalis for 1 h and further incubated for 24 h. LPS treatment (10 µg ml–1) served as a control. Cell lysates were prepared using reporter lysis buffer according to the manufacturer’s instructions (Promega). Luciferase activity was measured using a luciferase assay system (Promega) and a TD-20/20 luminometer (Turner Designs). Values were normalized to protein concentration.
Cell-cycle and apoptosis assays. The ability of E. faecalis to activate colonic epithelial-cell checkpoints or initiate apoptosis was assessed by flow cytometry. Following a 1 h treatment with E. faecalis, HCT116 or YAMC cells were incubated in fresh medium containing FBS, gentamicin and penicillin. Cells were fixed overnight with 70 % ethanol and stained with propidium iodide (0.02 mg ml–1) containing 0.1 % Triton X-100 and 0.2 mg RNase A ml–1. Stained cells were analysed using a FACSCalibur flow cytometer (BD Immunocytometry Systems). Apoptotic cells were stained using the Annexin V FITC apoptosis detection kit according to the manufacturer’s instructions (Calbiochem, EMD Biosciences) and quantified by flow cytometry. Data were analysed using CellQuest Pro software. Statistical analyses were performed using ModFit version 2.2 software (Verity Software House). For each sample, >10 000 events were collected and groups were compared using Student’s t-test with P<0.05 considered significant.
Quantitative real-time RT-PCR (qRT-PCR) and gene silencing.
Total mRNA was isolated from E. faecalis-treated cells using a NucleoSpin RNA II kit (BD Biosciences) and 2 µg was reverse-transcribed with an iScript cDNA synthesis kit according to the manufacturer’s instructions (Bio-Rad Laboratories). qRT-PCR was performed using an Mx3005P Q-PCR System following the manufacturer’s instructions (Stratagene). Primers used to assess expression of genes identified in the array experiment (Table 1
) were purchased from Integrated DNA Technologies. The gene for netrin-1 was silenced by RNA interference using siGENOME SMARTpool reagent specific for human NTN1 or with an siCONTROL non-targeting siRNA pool (Dharmacon). Transient transfections were performed using DharmaFECT 4 transfection reagent (Dharmacon) according to the manufacturer’s protocol and gene silencing was confirmed by qRT-PCR. Groups were compared using Student’s t-test with P<0.05 considered significant.
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
B in colonic macrophages
As haematin-starved E. faecalis induces COX-2 expression in macrophages in vitro (Wang & Huycke, 2007) and COX-2 is regulated via NF-B (Karin & Greten, 2005), we initially determined whether E. faecalis activated this redox-sensitive signalling pathway in the colon. NF-B was detected using an NLS-specific anti-p65 antibody that detects p65 only after it dissociates from IB. p65 is a member of the canonical NF-B pathway and is activated by many stimuli including redox stress and exposure to bacteria (Karin & Greten, 2005). We found significantly increased numbers of cells with p65 nuclear staining in colons exposed to E. faecalis at 6 h compared with PBS controls (ridit=0.57, P=0.008). Although an increase in NF-B activation was found for colons exposed to haematin-starved compared with haematin-replete E. faecalis, this difference was not statistically significant (ridit=0.52, P=0.25). To identify mucosal cells with NF-B activation, we stained serial colon sections with the NLS-specific anti-p65 antibody and an anti-F4/80 mAb. Nearly all cells positive for p65 also stained positive for F4/80, indicating that this redox-sensitive signalling pathway had been activated in tissue macrophages (Fig. 1a, b). In contrast, no staining for the NLS of p65 was noted in epithelial cells.
|
B in vitro, we exposed macrophage and epithelial cell lines to haematin-starved bacteria. Strong nuclear staining, along with cytoplasmic staining, was noted in macrophages by laser-scanning confocal microscopy (Fig. 1c
). NF-B activation was noted as early as 3 h p.i. and persisted for >48 h. In comparison, HCT116 cells exposed to E. faecalis did not lead to nuclear localization of p65 (data not shown). To verify NF-B activation in macrophages, we transfected RAW264.7 cells with the pNFB-Luc reporter plasmid. Compared with the control, there was a >25-fold increase in NF-B activation at 24 h following exposure to haematin-starved E. faecalis (Fig. 1d). Furthermore, manganese superoxide dismutase (MnSOD) significantly reduced NF-B activation (P=0.005), indicating that extracellular superoxide from haematin-starved bacteria contributed to this effect. The addition of catalase did not further decrease NF-B activation by haematin-starved E. faecalis, although H2O2 alone activated NF-B in these cells.
NF-B regulates genes involved in cellular proliferation, immunity and apoptosis (Karin & Greten, 2005). Activation of NF-B requires the phosphorylation of IB by IB kinase. This results in IB degradation and release of NF-B homo- and heterodimers to translocate into the nucleus. NF-B promotes tumorigenesis by inhibiting apoptosis, dysregulating tumour-specific immune responses and producing reactive oxygen species that can damage genomic DNA. Our findings indicate that haematin-starved E. faecalis activates NF-B, in part, by producing extracellular superoxide. Although reactive oxygen species (including superoxide) can activate NF-B, this effect is unpredictable and typically cell-dependent (Gloire et al., 2006). Many studies use H2O2 as an oxidative stress, although superoxide should also be considered as it may lead to dissimilar effects. For example, superoxide is required for IL-1-dependent NF-B activation in chondrocytes (Mendes et al., 2003), enhances LPS-dependent NF-B activation in macrophages (Khadaroo et al., 2003) and initiates NF-B activation in neutrophils (Mitra & Abraham, 2006). In this study, MnSOD significantly decreased E. faecalis-dependent induction of NF-B in RAW264.7 cells, confirming that this anionic radical can potentiate NF-B activation beyond that seen with H2O2 alone.
Colonic mucosal gene response to haematin-starved E. faecalis
To determine how broadly the haematin-starved physiology of E. faecalis modulated gene expression in the colonic mucosa, we compared mRNA from mice for >5000 genes following exposure to haematin-starved E. faecalis with mRNA following exposure to haematin-replete E. faecalis. At 1 h p.i., six colonic mucosal genes were differentially regulated (Sod2, Sod3, Agtrl1, Vav1, Car4 and Nme6; P<0.01 for each). At 6 h, 25 genes were significantly downregulated and 17 genes upregulated (Table 2). Of the differentially regulated genes at the 1 h time point, only Sod3 and Car4 were still differentially regulated at 6 h. For the 42 colonic mucosal genes differentially expressed by haematin-starved E. faecalis, nine (21 %) were related to inflammatory or stress responses and ten (24 %) involved pathways for cell-cycle control, signalling and apoptosis. In addition, several were expressed by immune effector cells, suggesting that acute colonic mucosal responses to haematin-starved E. faecalis involve innate and/or adaptive immunity.
|
; P<0.0001). Functions within this network included cell-cycle regulation, inositol phosphate metabolism, NF-B signalling (RelA or p65), ERK/MAPK signalling, chemokines, T-cell receptors, integrins and fibroblast growth factor. Exploration of potential regulatory responses within the network was performed using in silico transcriptional regulatory element analysis. One hundred and eleven transcriptional regulatory elements were associated with the forty-two differentially regulated genes. When we compared the frequency of these elements with elements for all genes on the murine array, there were only five significantly over-represented elements: three for NF-B (including the sequence for p65 binding), HEN1 and GATA-1. In addition, 66 (59 %) of the 111 transcriptional regulatory elements were significantly under-represented. Overall, these findings indicate that haematin-starved E. faecalis acutely activates NF-B signalling in the colonic mucosa.
|
B activity localized to tissue macrophages. The mechanism by which E. faecalis contacts tissue macrophages was not investigated, but may involve translocation of enterococci through follicle-associated M cells in the colon (Kraehenbuhl & Neutra, 2000). These specialized epithelial cells facilitate uptake of luminal bacteria and coordinate their interaction with innate and adaptive immune effector cells. Enterococci readily translocate across the intact intestinal epithelium and, in murine models, are often recovered from the liver, spleen and mesenteric lymph nodes (Wells et al., 1990). This phenomenon may derive, in part, from ineffective killing of E. faecalis by macrophages (Gentry-Weeks et al., 1999). In the colonic ligation model, the concentration of luminal bacteria at 6 h was 10-fold lower than the original inoculum. Epithelial translocation is one possible explanation for a decrease in colony counts.
Several genes within the mucosal response network were associated with NF-B signalling including C3ar1, Cyr61 and Akap8l. C3AR1 (complement component 3a receptor 1) induces NF-B activation when coupled to G16 (Yang et al., 2001). Similarly, Cyr61 (cysteine-rich protein 61) is associated with NF-B signalling, inflammation and angiogenesis (Klein et al., 2002), as well as anti-apoptotic effects when overexpressed in breast cancer (Lin et al., 2004). Although CYR61 has been implicated in the progression of breast cancer (Xie et al., 2001), it can also act as a tumour suppressor (Chien et al., 2004). Finally, AKAP8L (nuclear protein kinase A anchoring protein) can bind NF-B, although the significance of this interaction is unclear (Bouwmeester et al., 2004). Several other genes in the mucosal response network have been implicated in cancer biology. For example, MCM2 (minichromosome maintenance 2 protein) binds to the nuclear scaffold created by AKAP8L and promotes apoptosis in cancer cells (Feng et al., 2003). In contrast, TIMP2 (inhibitor of matrix metalloproteinase 2) inhibits apoptosis and allows tumour growth (Egeblad & Werb, 2002), although its overexpression is inhibitory (Gomez et al., 1997). The net long-term effect of haematin-starved E. faecalis on the colonic mucosa cannot be discerned from this study, although the complexity of the early response is apparent.
E. faecalis blocks G2/M transition in intestinal epithelial cells
As the mucosal response network indicated differential expression of genes involved in apoptosis and because reactive oxygen species from E. faecalis can damage colonic epithelial-cell DNA to initiate programmed cell death (Huycke et al., 2002), we investigated the effect of superoxide on cell-cycle checkpoints and apoptosis. Both HCT116 and YAMC cells exposed to haematin-starved E. faecalis for 1 h developed marked arrest at the G2/M transition by 48 h (Fig. 3a, b, c, d). This effect was partially reversed with MnSOD and completely abolished when MnSOD and catalase were both added to bacteria-treated cells. The effect of catalase suggested that H2O2, which spontaneously arises from the disproportionation of superoxide, also contributed to activation of the G2/M checkpoint. Treatment of HCT116 cells with H2O2 alone is known to cause this arrest (Chang et al., 2003), but our findings showed a greater proportion of arrested cells following exposure to haematin-starved E. faecalis than H2O2 alone.
|
). In contrast, HCT116 cells exposed to H2O2 showed significantly increased apoptosis at 24 h. These findings did not involve NF-B, as nuclear localization of p65 was not found by laser-scanning confocal microscopy at 6, 24, 48 or 72 h (data not shown). In contrast to transformed HCT116 cells, the primary non-transformed YAMC cells showed significantly increased apoptosis at 48 h following exposure to haematin-starved E. faecalis or H2O2 treatment (Fig. 3f). We demonstrated previously that E. faecalis can damage colonic epithelial-cell DNA (Huycke et al., 2002). In the current study, we found that extracellular superoxide from E. faecalis activates the G2/M checkpoint in HCT116 and YAMC cells. This effect was partially reversed by MnSOD and completely eliminated when both MnSOD and catalase were used, suggesting that extracellular superoxide and H2O2 each contributed to cell-cycle modulation. Arrest at the G2/M transition can be triggered by DNA double-strand breaks to activate pathways that allow mitosis to proceed after DNA repair or, alternatively, to initiate apoptosis (Nougayrede et al., 2006; Taieb et al., 2006). Differing outcomes following exposure to haematin-starved E. faecalis were apparent for HCT116 and YAMC cells, and demonstrated how cellular responses to DNA damage vary by cell type. Finally, we noted in a prior study that a minority of cells exposed to haematin-starved E. faecalis failed to repair DNA damage or to initiate apoptosis and subsequently developed chromosomal instability (Wang & Huycke, 2007).
The DNA-damaging effects of commensals on the colonic mucosa may not be limited to E. faecalis. Pathogenic and commensal strains of E. coli express hybrid peptide–polyketide and cytolethal distending toxins that produce DNA double-strand breaks, arrest at the G2/M transition and cell death (Nougayrede et al., 2006; Taieb et al., 2006). Other examples include commensal bacteria that utilize sulfate as an oxidant (in the assimilatory pathway) or terminal electron acceptor (in the dissimilatory pathway) to dispose of hydrogen-reducing equivalents (Gibson et al., 1988). The net result of this metabolism is hydrogen sulfide that, like superoxide, can be genotoxic to epithelial cells (Attene-Ramos et al., 2006). DNA double-strand breaks created by reactive oxygen species, hydrogen sulfide or other clastogens should activate DNA damage repair responses and activate the G2/M checkpoint (Su, 2006). Ongoing DNA damage could lead to the accumulation of mutations important to oncogenic transformation. Investigations are underway in our laboratory to explore these issues.
Regulation of netrin-1 by extracellular superoxide from E. faecalis
To investigate further the gene response network in vitro, we screened HCT116, YAMC and RAW264.7 cells by qRT-PCR for differential regulation of selected upregulated genes in the network known to be transcriptionally regulated by NF-B or involved in apoptosis (Table 1). Of these, only NTN1 in HCT116 cells was significantly upregulated by haematin-starved E. faecalis (Fig. 4). The inability to detect similar changes in gene expression in vitro using these cells compared with the in vivo model probably represents inherent differences between transformed cells and complex multicellular tissues in living animals (Waddell et al., 2007). Indeed, the rationale for the in vivo model was to avoid oversimplification of host–commensal interactions found in homogeneous in vitro culture systems.
|
).
Although we detected abundant expression of netrin-1 in the murine colonic mucosa by immunohistochemistry, no significant differences in immunoreactivity were found for any group of mice. The lack of differential staining may have been due to insufficient time for tissue protein concentrations to change (i.e. only 6 h). To determine whether increased netrin-1 expression in HCT116 cells explained the lack of increased apoptosis in HCT116 cells following exposure to E. faecalis (Fig. 3e), we used short interfering RNA to knock down NTN1 expression. Gene silencing led to an 83–90 % reduction in NTN1 mRNA for cells exposed to haematin-starved E. faecalis compared with untransfected controls or cells transfected with scrambled short interfering RNA. The proportion of cells undergoing apoptosis in NTN1-silenced cells, however, proved no different than for untransfected controls (data not shown). This result could have been due to a lack of functional receptors for netrin-1 on HCT116 cells or to defects in the secretion of netrin-1. When Western blots were performed, netrin-1 was not detected (data not shown), implying that other mechanisms are responsible for the inhibition of apoptosis in these cells following exposure to haematin-starved E. faecalis.
Several factors merit consideration when interpreting the results of this study. Because cDNA from mucosal scrapings were used, by necessity, in their entirety to synthesize probes, we could not check changes in gene expression for arrays by qRT-PCR. To overcome this limitation, we performed ten arrays per time point to increase the number of independent replicates. This resulted in a robust dataset for the identification of genes with differing expression. Furthermore, we defined altered gene expression using conservative rules in order to identify only those genes that were highly likely to have undergone induction or repression by haematin-starved E. faecalis. Additional genes may have been identified using less rigorous cut-offs. This, in turn, could have expanded the gene response network or identified other networks, but would not have changed the primary conclusion of this study, which is that the metabolic activity of intestinal commensals can acutely alter colonic mucosal gene expression. The short-term nature of our colonic ligation model only permitted an examination of early-response genes. Longer-term studies will require intestinal colonization or a gnotobiotic design. Finally, although superoxide from haematin-starved E. faecalis induced NTN1, knockdown of this gene had no affect on apoptosis, as secretion of this ligand appeared defective in these cells.
In summary, we found that the uniquely dichotomous metabolism of E. faecalis, a colonic commensal with a rudimentary respiratory chain that requires exogenous haematin for oxidative phosphorylation, can significantly modulate gene expression in the colonic mucosa. In vivo, in silico and in vitro analyses identified genes and signalling pathways that are associated with inflammation, apoptosis and cell-cycle regulation. Overall, these results suggest mechanisms by which E. faecalis might enhance epithelial-cell susceptibility to DNA damage through the activation of tissue macrophages and by modulating apoptosis in epithelial cells.
|
ACKNOWLEDGEMENTS |
|---|
B-Luc, Jim Henthorn at the University of Oklahoma Health Sciences Center Cytometry Laboratory and Ben Fowler at the Oklahoma Medical Research Foundation Core Imaging Facility. |
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
/NF-B signal transduction pathway. Nat Cell Biol 6, 97–105.[CrossRef][Medline]B activation by reactive oxygen species: fifteen years later. Biochem Pharmacol 72, 1493–1505.[CrossRef][Medline]B: linking inflammation and immunity to cancer development and progression. Nat Rev Immunol 5, 749–759.[CrossRef][Medline]5β1 integrin activates an NF-B-dependent program of gene expression important for angiogenesis and inflammation. Mol Cell Biol 22, 5912–5922.B-dependent XIAP up-regulation. J Biol Chem 279, 24015–24023.B activation and iNOS expression in bovine articular chondrocytes. J Cell Biochem 88, 783–793.[CrossRef][Medline]16 couples chemoattractant receptors to NF-B activation. J Immunol 166, 6885–6892.This article has been cited by other articles:
|
|
 |
|
  X. Wang, T. D. Allen, R. J. May, S. Lightfoot, C. W. Houchen, and M. M. Huycke Enterococcus faecalis Induces Aneuploidy and Tetraploidy in Colonic Epithelial Cells through a Bystander Effect Cancer Res., December 1, 2008; 68(23): 9909 - 9917. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
