Gastroenterology
Volume 132, Issue 2 , Pages 551-561, February 2007

Extracellular Superoxide Production by Enterococcus faecalis Promotes Chromosomal Instability in Mammalian Cells

  • Xingmin Wang
    • ,
  • Mark M. Huycke

      Affiliations

    • Corresponding Author InformationAddress reprint requests to: Mark M. Huycke, MD, Medical Service (111), 921 NE 13th Street, Oklahoma City, Oklahoma 73104. fax: (405) 297-5948.

The Muchmore Laboratories for Infectious Disease Research, Department of Veterans Affairs Medical Center, and Department of Medicine, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma

Received 28 July 2006; accepted 19 October 2006. published online 06 December 2006.

Article Outline

Background & Aims: We investigated whether Enterococcus faecalis, a Gram-positive intestinal commensal that produces extracellular superoxide, could promote chromosomal instability (CIN) in mammalian cells. Methods: We measured the ability of E faecalis to promote CIN using hybrid hamster cells (ALN) containing human chromosome 11. Results: E faecalis promoted CIN in ALN cells with average mutant fractions per 105 survivors (±SD) of 72.3 ± 6.7 at 1 × 109 cfu mL−1 compared with 22.2° ± 4.5 for the no bacteria control. γ-Irradiation at 2 Gray similarly resulted in 74.7 ± 5.7 mutant clones per 105 survivors. Deletions in chromosome 11 consistent with CIN were verified in 80% of mutant clones. E faecalis-treated ALN cells were protected from CIN by superoxide dismutase, γ-tocopherol, and cyclooxygenase-2 (COX-2) inhibitors. In a dual-chamber tissue culture model designed to mimic stromal-epithelial cell interactions, macrophages pretreated with E faecalis grown on permeable supports increased mutant fractions 2.5-fold for ALN cells. COX-2 was up-regulated by superoxide from E faecalis and mutant fractions decreased when COX-2 was silenced using short interfering RNA. Escherichia coli, a Gram-negative commensal that produces negligible extracellular superoxide, only modestly promoted CIN in this model. Conclusions: These findings indicate that macrophage COX-2 is induced by superoxide from E faecalis and promotes CIN in mammalian cells through diffusible factors. This mechanism links the oxidative physiology of E faecalis to propagation of genomic instability through a bystander effect, and offers a novel theory for the role of commensal bacteria in the etiology of sporadic colorectal cancer.

Abbreviations used in this paper: γ-CEHC, γ-carboxyethyl-hydroxychroman, CIN, chromosomal instability, COX-2, cyclooxygenase-2, MOI, multiplicity of infection, αT, α-tocopherol, γT, γ-tocopherol, TLR4, Toll-like receptor 4

 

See editorial on page 797.

Each year on a global basis nearly 1 million people are diagnosed with colorectal cancer and many die from complications.1 Most cases occur sporadically with tumors arising from normal epithelium through an accumulation of somatic mutations followed by clonal selection that result in malignant transformation.2 Genomic instability is fundamental to this process with chromosomal instability (CIN) found in >80% of sporadic colorectal cancer. CIN is typified by gene rearrangements, losses and gains of large DNA fragments, aneuploidy, and loss of heterozygosity.2 Recent mathematical modeling and experimental evidence suggest CIN precedes neoplastic transformation and may be causative.3, 4, 5 The genetic or molecular basis for CIN in human cancers, however, remains elusive, and is an area of active investigation.2, 6

The colonic microbiota have been implicated in the etiology of sporadic colorectal cancer, although no studies have convincingly defined which bacteria or what mechanisms promote this common tumor.7, 8, 9 Among intestinal commensals Enterococcus faecalis is unique in generating extracellular superoxide (·O2−).10 The discovery of extracellular radical production by this Gram-positive commensal bacterium led to our interest in its potential role in the etiology of sporadic colorectal cancer. The in vitro production of ·O2− by E faecalis (∼30 μmol/min per 109 cfu) is substantial and comparable to activated neutrophils undergoing a respiratory burst when rates are normalized to cell surface area. This phenotype is common to >95% of E faecalis strains, but rare among other bacteria.11 In one study, ·O2−-producing strains were found to colonize 40% of elderly veterans.12 Although this same study found no association between colonization with ·O2−-producing enterococcal strains and large colon adenomas or cancer by multivariate analysis, colonization patterns changed during 1 year of follow-up, suggesting that the composition of the elderly fecal microbiota is not stable.

Extracellular ·O2− from E faecalis most likely arises through the nonenzymatic reduction of oxygen by membrane-associated demethylmenaquinone.10 This results from a rudimentary respiratory chain that only contains cytochrome bd and fumarate reductase as terminal quinol oxidases. E faecalis grown in hematin or with fumarate does not generate extracellular ·O2− because these cofactors complete the respiratory chain and permit unimpeded flow of reducing equivalents to terminal acceptors. Limitations in fumarate and hematin must occur in the colon because intestinal colonization with E faecalis in murine models indicate that extracellular ·O2− is produced in vivo in quantities sufficient to cause colonic epithelial cell DNA damage.13, 14

In this article, we demonstrate that the oxidative physiology of E faecalis, a common intestinal commensal, promotes CIN in mammalian cells. In addition, we describe a role for macrophage cyclooxygenase-2 (COX-2) and identify COX-2 inhibitors and γ-tocopherol (γT) as agents that protect against genomic damage. These findings are consistent with the observations of Zhou et al,15 who reported COX-2 signaling as essential to CIN that is generated by the radiation-induced bystander effect. From these data we propose a novel hypothesis for sporadic colorectal cancer that involves triggering stromal immune responses by redox-active commensal bacteria to produce CIN in epithelial cells through a bystander effect. This mechanism is analogous to bystander effects described for mononuclear myeloid cells activated by superoxide to generate clastogens or chromosome breaking factors that damage DNA in neighboring cells.16, 17, 18

Back to Article Outline

Materials and Methods 

Cell and Bacterial Culture 

ALN human-hamster hybrid cells (a gift from Elizabeth McNiel) were maintained in Ham’s F12 Nutrition Mixture containing 4% heat-inactivated FCS and 3% heat-inactivated newborn calf serum, penicillin G (50 units mL−1), streptomycin (50 μg mL−1), G418 (400 μg mL−1), and 20 mmol/L HEPES buffer, pH 7.4, at 37°C in 5% CO2. The murine macrophage RAW264.7 cell line (American Type Culture Collection, Rockville, MD) was maintained in high glucose DMEM supplemented with 10% fetal bovine serum, penicillin G, streptomycin, and sodium bicarbonate (1.5 g L−1). E faecalis OG1RF is a Gram-positive human oral isolate that produces extracellular ·O2−.13 OG1RF was grown overnight in closed rotating tubes using brain–heart infusion (Difco, Detroit, MI) at 37°C and washed with phosphate-buffered saline. Extracellular ·O2− production was verified using ferricytochrome c as previously described.19 For purposes of comparison, Escherichia coli strain DH5α was grown overnight in Luria-Bertani broth (Difco) at 37°C with shaking (200 rpm) and washed with phosphate-buffered saline. This Gram-negative intestinal commensal generates extracellular superoxide at rates >25-fold less than E faecalis OG1RF and, unlike Gram-positive bacteria, has an outer membrane that contains lipopolysaccharide.20 Arachidonic and docosahexaenoic acids were purchased from Cayman Chemicals (Ann Arbor, MI). Manganese superoxide dismutase, catalase, and NS-398 were purchased from Sigma (St. Louis, MO). Tocopherols (Tama Chemicals, Kawasaki, Japan) and γ-carboxyethyl-hydroxychroman (γ-CEHC; Encore Pharmaceuticals, Inc, Riverside, CA) were provided by Kenneth Hensley and celecoxib by C.V. Rao.

Complement Lysis Assay 

CIN in ALN cells was determined as previously described using complement lysis.21, 22 E faecalis and E coli cytotoxicity to ALN cells were initially determined using adherent cells exposed to bacteria for 2 hours at 37°C. As a control, cells were also exposed to 137Cs γ-irradiation at room temperature. Complete F-12 medium was added to treated cells and included penicillin G and streptomycin to kill remaining enterococci or E coli when necessary. Cells were recovered at day 7, fixed with methanol, and counted after staining with Giemsa. Survival curves were normalized to the number of colony counts from untreated cells.

Following treatments to promote CIN, ALN cells were exposed to E7.1 anti-CD59 monoclonal antibody (a gift from Elizabeth McNeil) and rabbit complement.21 Complement fixation by E7.1 antibody leads to the lysis of CD59+ cells, whereas CD59− mutants survive to form colonies. After treatments ALN cells were plated to a minimum surviving fraction of at least 1 × 105 cells. Cells were subcultured for 10 days and incubated in complete UltraCULTURE (Cambrex, Walkersville, MD) at 37°C for 5 hours prior to complement lysis with 2% normal rabbit serum (Biomeda, Foster City, CA), 0.1% normal human serum, and 0.5% E7.1 antibody. Unlysed cells were counted when surviving colonies had grown to >50 cells. Plating efficiency was calculated by dividing the number of colonies that grew with complement by colonies that grew without complement. Colonies surviving complement lysis were normalized by plating efficiency and defined as the mutant fraction. Data represent a minimum of 3 independent experiments.

Chromosome 11 Deletions 

DNA from complement-resistant ALN cells was extracted and CD59 exons amplified by multiplex PCR using ExTaq DNA polymerase (Takara Shuzo, Otsu, Japan) and a program of 30 cycles at 94°C for 20 seconds, 55°C for 15 seconds, and 72°C for 30 seconds with preheating at 94°C for 1 minute. PCR primers were: exon 1 (forward) 5′-GAG CCT TGC GGG CTG GAG-3′ and (reverse) 5′-CCT TCG GGC CTT CTT ACC T-3′; exon 2 (forward) 5′-AAC CAG CAG TCA TTT GTT CG-3′ and (reverse) 5′-CCA AGG CAG CCT AGA TCC AT-3′; exon 3 (forward) 5′-GGA AGT ATA CCA CAA GTT GC-3′ and (reverse) 5′-GCC TAA TGA GGA TTA CAG TG-3′; and exon 4 (forward) 5′-TCT CCC CAC AGG GTT ACA AG-3′ and (reverse) 5′-TCC CTG CAA ACA GGA CTG-3′. These primers gave amplicons for exons 1–4 that were 101, 180, 360, and 460 bp in size, respectively. Loss of DNA from chromosome 11 was verified by single nucleotide polymorphism microarray technology according to the manufacturer’s instructions (GeneChip® Human Mapping 10K Array Xba 142 2.0, Affymetrix, Santa Clara, CA). Array data was analyzed using Affymetrix Analysis Software (v3.0) and the Web-based DNA-Chip Analyzer chromosome copy number tool (http://biosun1.harvard.edu/complab/dchip).

Dual-Chamber Model 

To investigate whether macrophages promoted CIN in ALN cells following phagocytosis of E faecalis or E coli, we developed a dual-chamber model with RAW264.7 cells separated from ALN cells by a permeable insert with 0.4 μm pores (Figure 5A, 6 Well Transwell® Permeable Support, Corning, NY). RAW264.7 cells were initially exposed to OG1RF or DH5α for 2 hours at 37°C before being plated at 2 × 105 cells on inserts. ALN cells in the lower compartment served as targets for clastogen production by RAW264.7 cells. These cells were cocultured in the dual-chamber model in F-12/DMEM medium (1:1) for 48 hours at 37°C in 5% CO2, after which time ALN cells were separately passaged for 8 days. Complement lysis and calculation of mutant fractions were then performed as described previously.

  • View full-size image.
  • Figure 5. 

    Phagocytosis of E faecalis or E coli by macrophages causes CIN in ALN cells. (A) Murine macrophages (RAW264.7 cells) exposed to OG1RF were cultured on insert in upper compartment of a dual-chamber model with ALN cells in lower compartment as the target. (B) OG1RF- or DH5α-treated RAW264.7 cells induced CIN in ALN cells at MOIs >1 (*P < .05, **P < .01, ***P < .001, NS, not significant; P values in comparison to no bacteria control); similarly, DH5α-treated RAW264.7 cells modestly increased CIN in ALN cells, but not at an MOI of 1000 (*P < .05, **P < 0.01, ***P < .001, NS, not significant; P values in comparison to no bacteria control; black bar, OG1RF; open bar, DH5α). (C) OG1RF treatment of RAW264.7 cells up-regulated COX-2; RT-PCR (upper) and Western blotting (lower). (D) OG1RF caused dose-dependent expression of COX-2 in RAW264.7 cells; RT-PCR (upper) and Western blotting (lower); (E) DH5α also caused COX-2 expression; RT-PCR (upper) and Western blotting (lower); RNA extracted at 24 hours and protein at 48 hours.

RT-PCR 

Total RNA was extracted from cells and cDNA synthesized using TaqMan Reverse Transcription Reagents (Applied Biosystems, Bedford, MA). A 181-bp fragment for Cox-2 and 282-bp fragment for β-actin were amplified from murine and hamster cDNA using ExTaq DNA polymerase. PCR primers for Cox2 were (forward) 5′-CAA CTC CCT TGG GTG TGA AAG GAA-3′ and (reverse) 5′-AAG CCA GGT CCT CGT TTC TGA TCT-3′, and for β-actin (forward) 5′-GGA ATC CTG TGG CAT CCA CGA A-3′ and (reverse) 5′-TCG TAC TCC TGC TTG CTG ATC C-3′.

Western Blotting 

Total proteins were extracted from cells and equal amounts analyzed by sodium dodecyl sulfate/PAGE. Fractionated proteins were transferred to polyvinylidene difluoride membranes (Amersham Biosciences, Piscataway, NJ) and COX-2 identified using goat polyclonal antihuman COX-2 antibody with alkaline phosphatase conjugated rabbit antigoat IgG as a secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA). β-Actin loading controls were developed using murine monoclonal antibody and alkaline phosphatase conjugated rabbit antimouse IgG (Abcam, Cambridge, MA). Bands were detected using the ECF™ Western Blotting Detection System (Amersham Biosciences).

SiRNA 

Knock-down of Cox-2 in RAW264.7 cells was performed using short interfering RNA (siRNA) from Dharmacon (Lafayette, CO) according to the manufacturer’s instructions. Following a 20-hour incubation, OG1RF was added at a multiplicity of infection (MOI) of 1000 bacteria per cell and incubated for 2 hours to induce COX-2. Treated cells were placed in the dual-chamber model as described previously. Cox-2 knock-down was verified by RT-PCR and Western blotting.

Statistical Analysis 

Data were calculated as means and standard deviations. Comparisons between treated and control groups were performed by the student’s t-test. P-values <.05 were considered significant.

Back to Article Outline

Results 

E faecalis Promotes CIN in ALN Cells 

We initially determined the ability of E faecalis OG1RF to promote CIN in ALN human–hamster hybrid cells. This engineered cell line contains 1 copy of human chromosome 11 that encodes CD59 surface antigen at 11p13.5, confers neomycin resistance at another locus, and is a validated assay for measuring CIN.22 As a control for comparing E faecalis treatments, we also exposed ALN cells to γ-irradiation. Both conditions produced cytotoxic effects with decreasing cell survival as doses of E faecalis and γ-irradiation increased (Figures 1A and B). Average mutant fractions for ALN cells increased in a dose-dependent manner for OG1RF and, as previously reported,23 following γ-irradiation. The average mutant fraction per 105 survivors (±SD) was 22.2 ± 4.5 for a no bacteria control, and 39.1 ± 9.3, 49.6° ± 6.8, and 72.3 ± 6.7 for OG1RF at 1 × 107, 1 × 108, and 1 × 109 cfu mL−1, respectively (Figure 1C). The average mutant fraction produced by OG1RF at the highest dose was equivalent to 2 Gy of γ-irradiation (Figure 1D; 72.3 ± 6.7 vs 74.7 ± 5.7 per 105 survivors, respectively).

  • View full-size image.
  • Figure 1. 

    E faecalis causes CIN in target cells. (A) Survival curve for ALN cells exposed to increasing doses of OG1RF at 37°C for 2 hours. (B) Survival curve for ALN cells exposed to increasing doses of γ-irradiation. (C) A dose–response in mutant fraction was noted for ALN cells exposed to OG1RF at 37°C for 2 hours. (D) The mutant fraction for ALN cells also increased with radiation dose (R2 = 0.99). Mutant fraction produced by OG1RF at 1 × 109 cfu mL−1 was equivalent 2 Gy γ irradiation (hashed lines). Mutant fractions were calculated as the number of surviving colonies divided by the total number of cells plated after correction for plating efficiency.

We next verified the presence of CIN in mutant fractions using 10 randomly selected complement-resistant ALN clones. By multiplex PCR 8 of 10 clones had lost all 4 exons for CD59 (Figure 2A). Two clones still contained CD59 exons and, presumably, were complement-resistant through mutations in other genes necessary for CD59 expression.23 Because genomic deletions are characteristic of CIN, we used single nucleotide polymorphism array technology to survey the entire length of chromosome 11 for additional losses and found 4 clones had large DNA deletions that ranged in size from 12 to 110 Mb (Figure 2B). Four other clones had multiple smaller deletions. All deletions, whether large or small, included the CD59 locus at 11p13.5. We similarly found loss of CD59 exons among complement-resistant ALN clones for untreated controls (7 of 11 clones). Large deletions were noted in 6 of these clones (data not shown), suggesting that OG1RF and γ-irradiation served to augment preexisting low-level CIN in these cells.

  • View full-size image.
  • Figure 2. 

    CIN in ALN cells. (A) Multiplex PCR verified deletions of CD59 exons for 8 of 10 randomly selected colonies. The 4 CD59 exons (arrows) were present in an ALN control colony and for 2 colonies that survived complement lysis (C3 and C26). No amplicons were generated for any CD59 exon in 8 of 10 surviving colonies (C6, C11, C21–C25, and C27). (B) Using human SNP mapping, clones negative by multiplex PCR for CD59 exons showed large deletions in chromosome 11 (12 to 110 Mb) in 4 clones (C6, C11, C23, and C27) and smaller deletions in 4 other clones (C21, C22, C24, and C25). CD59 is located at 11p13 (arrow) and signal shift to the left of baseline represents loss of SNPs.

E coli Does Not Promote CIN in ALN Cells 

To compare extracellular superoxide producing E faecalis OG1RF to a Gram-negative commensal strain that makes minimal extracellular superoxide, we exposed ALN cells to E coli strain DH5α. Unlike OG1RF, DH5α was more cytotoxic to ALN cells than OG1RF (Figure 3A). Average mutant fractions, however, did not increase as the dose of DH5α was increased. The average mutant fraction per 105 survivors was 27.1 ± 6.8 for a no bacteria control, and 27.1 ± 4.5, 33.5° ± 7.2, and 29.4 ± 4.1 for DH5α at 1 × 106, 1 × 107, and 1 × 108 cfu mL−1, respectively (Figure 3B; all P values not significant compared to the no bacteria control).

  • View full-size image.
  • Figure 3. 

    Commensal E coli does not promote CIN. (A) Survival curve for ALN cells exposed to increasing doses of DH5α at 37°C for 2 hours. (B) No significant difference in mutant fraction was noted for ALN cells exposed to DH5α at 37°C for 2 hours compared to the no bacteria control.

E faecalis Induces CIN Through Superoxide-Mediated COX-2 Expression 

E faecalis uniquely produces extracellular ·O2−,10 a reactive oxygen species associated with clastogenesis.18 To further assess the role of extracellular ·O2− production by E faecalis on the augmentation of CIN in eukaryotic cells, we added radical scavengers and lipid peroxidation substrates to ALN cells. Manganese superoxide dismutase (1200 units mL−1) decreased the average mutant fraction by 50% for ALN cells exposed to OG1RF (Figure 4A; P < .001). No additional decrease occurred when catalase was added (1500 units mL−1; P = .78). This suggested that superoxide, but not hydrogen peroxide (H2O2), affected the rate of CIN in these cells. Polyunsaturated fatty acids are facile substrates for ·O2−-initiated lipid peroxidation and their breakdown products are potential clastogens.18, 24 To assess the role of polyunsaturated fatty acids on CIN in OG1RF-treated ALN cells, we added exogenous arachidonic acid, an ω-6 polyunsaturated fatty acid, and noted a 28% increase in the average mutant fraction (Figure 4B, left; 90.1 ± 4.8 vs 70.3 ± 2.1; P = .003). A similar increase, however, did not occur when docosahexaenoic acid, an ω-3 polyunsaturated fatty acid, was used (78.0 ± 6.6 vs. 70.3 ± 2.1; P = .13). Arachidonic acid and docosahexaenoic acid alone had no effect on mutant fractions in ALN cells not treated with OG1RF (data not shown). These findings suggested ω-6, but not ω-3, polyunsaturated fatty acids promoted ·O2−-mediated CIN in this system.

  • View full-size image.
  • Figure 4. 

    CIN in ALN cells is due to extracellular ·O2− from E faecalis and COX-2. (A) Mutant fraction for ALN cells exposed to OG1RF (109 cfu mL−1) decreased 50% with MnSOD, but not catalase. (B) Mutant fraction increased with arachidonic acid but not docosahexaenoic acid (left) and decreased with αT, γT, and γ-CEHC (right). (C) OG1RF induced Cox-2 mRNA in ALN cells (upper); Cox-2 mRNA was decreased by MnSOD (middle); E coli DH5α did not induce Cox-2 expression (lower). (D) Mutant fractions decreased following treatment with COX-2 inhibitors (**P < .01, ***P < .001, NS, not significant; P values in comparison to OG1RF).

Because arachidonic acid promoted CIN in E faecalis-treated ALN cells, we explored the role of extracellular ·O2− by adding tocopherols as prototypic lipid radical chain terminators, or a tocopherol metabolite, γ-CEHC, as a control. Initially, we cultured OG1RF in 200 μmol/L α-tocopherol (αT), 200 μmol/L γT, or 100 μmol/L γ-CEHC, and in no instance found changes in the rate of extracellular ·O2− production (data not shown). Addition of 200 μmol/L αT or 200 μmol/L γT to ALN cells, however, decreased mutant fractions by 42% (38.3 ± 5.5 vs. 65.8 ± 4.4; P = .002) and 34% (43.3 ± 4.6 vs. 65.8 ± 4.4; P = .004), respectively (Figure 4B, right). Surprisingly, mutant fractions also decreased 47% when 100 μmol/L γ-CEHC was added (34.8 ± 4.2 vs. 65.8 ± 4.4; P < .001).

Because γ-CEHC is not a lipid radical scavenger,25 but instead selectively inhibits COX-2,25, 26 we sought to determine whether COX-2 was mechanistically linked to E faecalis-augmented CIN. We first evaluated Cox-2 expression in ALN cells by RT-PCR. Detectable mRNA was not found at baseline but increased substantially 2–3 hours following OG1RF treatment (Figure 4C, upper). When manganese superoxide dismutase was added the increase in mRNA was blocked (Figure 4C, middle), indicating that Cox-2 expression was primarily due to the production of extracellular ·O2− by E faecalis. In contrast, when we evaluated Cox-2 expression by RT-PCR in ALN cells exposed to DH5α, Cox-2 mRNA was minimally detectable at 3–5 hours following treatment (Figure 4C, lower) and remarkably less than the expression produced by OG1RF. We also tested the effect of arachidonic acid and docosahexaenoic acid and found that these fatty acids had no effect on Cox-2 expression in ALN cells when used alone (data not shown). These findings, and knowledge that COX-2 inhibitors protect against colorectal cancer,27 led us to consider the effect of these anti-inflammatory drugs on OG1RF-treated ALN cells. Indeed, two selective COX-2 inhibitors, celecoxib and NS-398, each caused dose-dependent decreases in mutant fractions for cells exposed to OG1RF (Figure 4D).

In sum, these results indicate that E faecalis, but not E coli, promotes CIN in ALN cells via mechanisms that: (1) require extracellular ·O2−, (2) are inhibited by lipid radical chain terminators, (3) involve COX-2, and (4) can be promoted by exogenous ω-6 polyunsaturated fatty acids. The latter may serve as COX-2 substrates or a target for membrane peroxidation.

Phagocytosis of E faecalis by Macrophages Causes CIN in ALN Cells 

Commensal intestinal bacteria, including E faecalis and E coli, are not usually in contact with colonic epithelial cells because of dense mucin overlying these cells. Therefore, in vitro experiments involving direct bacterial contact with epithelial cells may not accurately reflect in vivo interactions. We considered that enterococci and E coli are more likely delivered to the colonic mucosa via M cells found scattered throughout the colon.28 These specialized epithelial cells allow normal sampling of luminal antigens by host immune effector cells. In theory, macrophages residing beneath M cells would ingest commensal bacteria, become activated, and induce CIN in neighboring epithelial cells. This hypothesis is analogous to the radiation-induced bystander effect wherein irradiated cells are activated to generate diffusible factors that cause CIN in unirradiated neighboring cells.16, 17

To investigate this hypothesis, we exposed RAW264.7 cells to E faecalis OG1RF and plated them in the upper compartment of a dual-chamber tissue culture system. Macrophages on the insert were separated from the lower compartment by a support with 0.4-μm pores that prevented cell-to-cell contact and transit of bacteria (Figure 5A). This system was used to model stromal–epithelial cell interactions in the colonic mucosa with ALN cells in the lower compartment serving as epithelial targets for clastogens produced by treated macrophages. The results showed significant increases in average mutant fractions for ALN target cells exposed to E faecalis-treated macrophages (Figure 5B). Responses varied from no change in mutant fractions using an MOI of 1 cfu per RAW264.7 cell, to more than a doubling of mutant fractions at an MOI of 1000. Viable OG1RF persisted inside RAW264.7 cells for >3 days following ingestion (data not shown). This finding is consistent impaired bactericidal killing of enterococci by peritoneal macrophages as noted in previous reports.29

In contrast to E faecalis OG1RF, mutant fractions only modestly increased for ALN cells exposed to E coli DH5α-activated RAW264.7 cells at an MOI of 10 or 100 (Figure 5B). Surprisingly, and unlike E faecalis, significant increases were not noted at the highest MOI of 1000 compared with the no bacteria control. Finally, unlike OG1RF, intracellular DH5α survived <24 hours following ingestion by RAW264.7 cells (data not shown).

COX-2 Is Necessary for Macrophage-Induced CIN 

Because COX-2 inhibitors protected ALN cells from CIN when in direct contact with E faecalis, we evaluated the effect of COX-2 in the dual-chamber model. OG1RF induced COX-2 in RAW264.7 cells with maximal mRNA expression occurring at 5–24 hours post-treatment (Figure 5C). The magnitude of expression increased as mutant fractions similarly increased from an MOI of 1 to 1000 (Figure 5D). E coli similarly activated COX-2 expression in RAW264.7 cells at all doses including the MOI of only 1 (Figure 5E). In comparison, an MOI of at least 10 was necessary for COX-2 induction by OG1RF.

To determine the role of COX-2 expression in the dual-chamber model, we pretreated RAW264.7 cells with COX-2 inhibitors and then exposed them to OG1RF. These inhibitors significantly decreased ALN mutant fractions (Figure 6A). We next employed siRNA to knock-down macrophage Cox-2 and directly test the role of this enzyme. Murine Cox-2 siRNA (100 nM) silenced Cox-2 mRNA in OG1RF-treated RAW264.7 cells by >80% and COX-2 protein by >90% (Figure 6B). Cox-2-silenced macrophages exposed to OG1RF showed a marked decrease in their ability to promote ALN mutant fractions compared with nonsilenced macrophages (Figure 6A; 30.2 ± 6.3 vs 68.1 ± 5.5; P < .001). The degree of protection was similar to that provided by COX-2 inhibitors. These data strongly support the hypothesis that E faecalis induces CIN in target cells by activating macrophage COX-2.

  • View full-size image.
  • Figure 6. 

    CIN in ALN cells is caused by E faecalis-induced COX-2 expression in macrophages in dual-chamber model. (A) ALN mutant fractions decreased when OG1RF-treated RAW264.7 cells were pretreated with COX-2 inhibitors or Cox-2 silenced by siRNA (***P < .001; P values in comparison to an MOI of 1000). (B) COX-2 knock-down verified by RT-PCR (left) and Western blotting (right).

Extracellular Superoxide Induces COX-2 in Macrophages and CIN in ALN Cells 

We added manganese superoxide dismutase and used heat-inactivated E faecalis OG1RF in the dual-chamber model to test whether extracellular ·O2− contributed to the promotion of CIN in ALN cells. Each experiment reduced the average ALN mutant fraction (Figure 7A, upper), supporting the hypothesis that the oxidative physiology of E faecalis promotes macrophage-induced CIN. The average mutant fractions, however, were still greater than a no bacteria control, suggesting that phagocytosis of antigens also contributes to an increase in CIN observed in these experiments.

  • View full-size image.
  • Figure 7. 

    Extracellular ·O2− from E faecalis promotes COX-2 expression in macrophages and induces CIN in ALN cells in the dual-chamber model. (A) ALN cell mutant fraction using heat-inactivated OG1RF or MnSOD (upper; *P < .05, **P < .01, ***P < .001; P values in comparison to OG1RF). COX-2 expression by Western blotting (middle) and normalized to β-actin control (lower). (B) γT (200 μmol/L) and γ-CEHC (100 μmol/L), but not αT (200 μmol/L), protect ALN cells from CIN (upper; **P < .01, ***P < .001, NS, not significant; P values in comparison to OG1RF). COX-2 expression decreased by γT, but not αT or γ-CEHC. COX-2 protein by Western blotting (middle) and normalized to β-actin control (lower).

COX-2 expression in RAW264.7 cells was markedly less when heat-inactivated OG1RF was substituted for live bacteria and when manganese superoxide dismutase was added to live OG1RF (Figure 7A, middle and lower). Addition of OG1RF to the upper compartment, without macrophages, led to no change in baseline mutant fraction (Figure 7A, upper). This lack of response was likely due to the short half-life of ·O2− in physiologic buffer (t1/2 <1 msec).30 The distance between upper and lower compartments in the dual-chamber model is 1 mm, a gap too great for ·O2− to diffuse and act directly on target cells. The lack of change in baseline mutant fraction for ALN cells exposed to OG1RF in the upper compartment also eliminated H2O2 as a possible mediator for CIN. This observation is consistent with Emerit who found ·O2−, but not H2O2, strongly induced clastogenesis.18

Finally, because lipid radical chain terminators successfully inhibited E faecalis-induced CIN in ALN cells directly exposed to OG1RF (Figure 4B, right), we tested the effect of tocopherols in the dual-chamber model. γT and γ-CEHC protected against increases in CIN, but αT had no effect (Figure 7B, upper). γT also significantly decreased COX-2 expression (P = .003), unlike αT and γ-CEHC, where no change was noted (Figure 7B, middle and lower). These results demonstrate that extracellular ·O2− production by E faecalis can induce COX-2 in macrophages. These activated cells are then able to generate diffusible mediators that promote CIN in neighboring cells through a bystander effect.

Back to Article Outline

Discussion 

Plausible theories for sporadic colorectal cancer should successfully address the key questions concerning carcinogenesis for these tumors, ie, the origin of CIN, a role for COX-2, and causal relationships between submucosal cell reactivity and transformed epithelial cells. Our data support a model that mechanistically links these characteristics. In this scheme, extracellular ·O2− production by E faecalis triggers COX-2 expression in macrophages that drives carcinogenesis in epithelial cells through a bystander effect. The focus on the oxidative properties of E faecalis as a proof-of-concept should not be construed to exclude other commensal bacteria or environmental agents that might also induce COX-2 and trigger this effect. Indeed, E coli, which produces negligible quantities of extracellular superoxide, was able to modestly promote CIN in target cells using the dual-chamber system. Notwithstanding these possibilities, redox-active bacteria deserve particular attention. E faecalis specifically causes colon cancer in IL-10 knock-out mice that are monoassociated with this bacterium, and promotes more aggressive colitis compared with E coli.31, 32 Similarly, E faecalis can induce colitis and dysplasia in mice that have a double knockout of glutathione peroxidase isoforms 1 and 2 and are monoassociated with OG1RF (Huycke and Chu, unpublished observations, December, 2005). Our findings and these reports fit the increasingly recognized association of cancer with radicals, endogenous inflammation, and host genetic determinants.33, 34, 35

According to this hypothesis, extracellular ·O2− production by E faecalis activates macrophages to generate a bystander effect. This effect would require an interaction between E faecalis and colonic mucosal immune cells. Such an assumption is plausible because enterococci are known to readily translocate the intact intestinal mucosa.36 The average human colon contains 3–4 lymphoid follicles per square centimeter.37 A unique feature of these follicles are M cells that lack microvilli, have no overlying mucin, and are associated with a porous basement membrane.38 These specialized epithelial cells present luminal bacteria to effector cells on the basolateral side to help orchestrate innate and adaptive mucosal immune responses.28 E faecalis is able to persist inside macrophage phagolysosomes and this survival likely provides ongoing oxidative stress to these activated cells.

A bystander theory for sporadic colorectal cancer is one of long-term, endogenous mutagenesis by innate immune effector cells. This process would most readily occur in older persons who are at highest risk for sporadic colorectal cancer because of declining counterregulatory mechanisms that no longer effectively control minimal inflammatory stimuli from commensal bacteria.39 E faecalis is a redox-active commensal we used to test this hypothesis. Findings from our dual-chamber model showed that extracellular ·O2− from this bacterium strongly induced COX-2 in macrophages and, as a result, led to clastogens that promoted CIN in target cells. One or more products, or byproducts, of COX-2 probably mediate the bystander effect because COX-2 silencing protected ALN cells from increased CIN. The nature of these mediators is an area of ongoing investigation.

The permissive role of COX-2 in sporadic colorectal cancer is based on numerous clinical and population-based studies that show COX-2 inhibitors are effective chemopreventive agents.27 Cox knock-out models of colorectal cancer in mice provide further evidence.40, 41 The mechanisms by which COX-2 promotes carcinogenesis, however, remain obscure. COX-2 is an immediate early response gene not normally expressed in the colonic mucosa, but instead found in submucosal macrophages (and myofibroblasts) in human colonic adenomas.42, 43 Localization of COX-2 to these cells, but not epithelial cells, is consistent with a bystander hypothesis and spatially fits with the top-down morphogenesis described for colorectal cancer by Shih et al.44 The endoperoxide product of COX-2 is prostaglandin H2, an unstable intermediate that is enzymatically converted to structurally related bioactive lipids that have antiapoptotic, immunomodulatory, and angiogenic effects that promote tumor growth, but are otherwise not considered mutagenic and therefore less likely to be mediators for CIN.45 Several peroxidation byproducts generated by COX-2 have been partially characterized by HPLC.46 The potential importance of these as mediators for CIN merit further study.

The protection afforded by γT, but not αT, in our dual-chamber model suggests that this desmethyl tocopherol variant of vitamin E, which is the primary form found in foodstuffs,25 is more biologically active than αT in limiting CIN. Reduction in oxidative stress through the termination of lipid radical chain reactions might explain a decrease in COX-2 expression in macrophages pretreated with γT. αT, however, is an equally potent radical scavenger,25 but had no effect on COX-2 expression. Such differences between tocopherols have been reported before and, for example, include decreased synthesis of prostaglandin E2 and reduction in inflammation markers for γT but not αT.47 γT also inhibits COX-2,26 a trait not shared by αT, and this, similar to the COX-2 selective inhibitors, may contribute to protection against CIN. These biologic distinctions between γT and αT may also explain several large clinical trials that found αT in the form of oral vitamin E supplements was not effective as a chemopreventive agent for colon adenomas and cancer.48, 49, 50 Clinical trials using γT have yet to be performed.

Our focus on redox properties of E faecalis does not exclude other commensal bacteria that may also potentially induce COX-2 expression. Results with E coli, a Gram-negative commensal bacterium, suggest that other intestinal microbiota might also activate macrophages and lead to release of mediators that promote CIN. Lipopolysaccharide in the outer membrane of Gram-negative bacteria are known to activate NF-κB and COX-2 via Toll-like receptor 4 (TLR4) in RAW264.7 cells.51 Conversely, inactivation of TLR4 in mice leads to decreased proliferation, increased apoptosis, and blunted COX-2 expression in the intestinal mucosa following dextran sodium sulfate injury.52 Oshima et al reported that Helicobacter pylori, a Gram-negative stomach pathogen, stimulates gastric epithelial cells through TLR4 and activates macrophages to induce hyperplastic gastric tumors in transgenic mice.53 Similarly, E coli produces lipopolysaccharide that can activate macrophage COX-2 through TLR4.54 The lack of COX-2 expression in ALN cells following direct exposure to E coli may be due to the absence of CD14 on the cell surface of Chinese hamster ovary cells, a cofactor that is essential for lipopolysaccharide-induced COX-2 expression via TLR4.55

Although exposure of RAW264.7 cells to low doses of E coli modestly promoted CIN in ALN cells using the dual-chamber model, this effect did not occur at the highest MOI despite continued expression of COX-2 in the RAW264.7 cells. An explanation for this observation is unclear, but perhaps marked excess lipopolysaccharide at the highest MOI decreased RAW264.7 cell proliferation and/or induced apoptosis and mitigated responsiveness.

Finally, another potential mechanism for promoting CIN in epithelial cells that we did not explore in our experiments might occur with E coli strains that cause double-stranded DNA breaks in eukaryotic cells.56 This genome-damaging effect has been reported in subsets of commensal and extraintestinal pathogenic strain, is contact-dependent, and mediated by a unique polyketide-peptide hybrid cytotoxin encoded on a genomic island. The E coli DH5α strain used in this study is related to the noncytotoxin producing DH10B strain,56 and so promotion of CIN by this mechanism was not tested in our model. We note, however, that commensal bacteria are usually not in direct contact with colonic epithelial cells in vivo because of dense overlying mucin. In noninjured mucosa the physical barrier afforded by mucin would likely limit DNA damage from this cytotoxin.

One limitation to this study concerns our use of ALN cells as targets for activated macrophages. These transformed cells already express CIN. Although increased rates of CIN were generated by extracellular ·O2− from E faecalis, the complement-lysis assay could not determine whether E faecalis can induce CIN in cells that might otherwise be chromosomally stable. To address this question will require epithelial cells of colonic origin that do not express CIN. This is an area of ongoing investigation in our laboratory.

In summary, our findings demonstrate that E faecalis promotes CIN when in direct contact with mammalian cells and can cause similar effects in nearby cells through the induction of COX-2 in macrophages. An increased rate of CIN produced by diffusible clastogens from macrophages is analogous to the propagation of genomic instability to neighboring cells through a radiation-induced bystander effect.16, 17 γT, but not αT, attenuated COX-2 expression and CIN in the dual-chamber model and highlights potentially important biologic differences between these forms of vitamin E. These findings support a theory that links the oxidative physiology of commensal bacteria with CIN and elucidates the permissive role of COX-2 in sporadic colorectal cancer.

Back to Article Outline

 

The authors thank Elizabeth McNeil and Charles Waldren for providing reagents, Kenneth Hensley for helpful advice and tocopherols, and the Genotyping Laboratory at Oklahoma Medical Research Foundation for SNP array analysis.

Back to Article Outline

References 

  1. Colorectal cancer. In: Stewart BW, Kleihues P, eds. World cancer report. Lyon: IARC Press; 2003.
  2. Grady WM. Genomic instability and colon cancer. Cancer Metastasis Rev. 2004;23:11–27
  3. Nowak MA, Komarova NL, Sengupta A, Jallepalli PV, Shih IM, Vogelstein B, et al. The role of chromosomal instability in tumor initiation. Proc Natl Acad Sci U S A. 2002;99:16226–16231
  4. Stoler DL, Chen N, Basik M, Kahlenberg MS, Rodiguez-Bigas MA, Petrelli NJ, et al. The onset and extent of genomic instability in sporadic colorectal tumor progression. Proc Natl Acad Sci U S A. 1999;96:15121–15126
  5. Shih I-M, Zhou W, Goodman SN, Lengauer C, Kinzler KW, Vogelstein B. Evidence that genetic instability occurs at an early stage of colorectal tumorigenesis. Cancer Res. 2001;61:818–822
  6. Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat Med. 2004;10:789–799
  7. Moore WEC, Moore LH. Intestinal floras of populations that have a high risk of colon cancer. Appl Environ Microbiol. 1995;61:3202–3207
  8. McGarr SE, Ridlon JM, Hylemon PB. Diet, anaerobic bacterial metabolism, and colon cancer: a review of the literature. J Clin Gastroenterol. 2005;39:98–109
  9. Augenlicht LH, Mariadason JM, Wilson A, Arango D, Yang W, Heerdt BG, et al. Short chain fatty acids and colon cancer. J Nutr. 2002;132:3804S–3808S
  10. Huycke MM, Moore D, Shepard L, Joyce W, Wise P, Kotake Y, et al. Extracellular superoxide production by Enterococcus faecalis requires demethylmenaquinone and is attenuated by functional terminal quinol oxidases. Mol Microbiol. 2001;42:729–740
  11. Huycke MM. Physiology of enterococci. In:  Gilmore MS editors. Enterococci: pathogenesis, molecular biology and antibiotic resistance. Washington, DC: ASM Press; 2002;p. 133–175
  12. Winters MD, Schlinke TL, Joyce WA, Glore WR, Huycke MM. Prospective case–cohort control study of intestinal colonization with enterococci that produce extracellular superoxide and the risk for colorectal adenomas or cancer. Am J Gastroenterol. 1998;93:2491–2500
  13. Huycke MM, Moore DR. In vivo production of hydroxyl radical by Enterococcus faecalis colonizing the intestinal tract using aromatic hydroxylation. Free Radic Biol Med. 2002;33:818–826
  14. Huycke MM, Abrams V, Moore DR. Enterococcus faecalis produces extracellular superoxide and hydrogen peroxide that damages colonic epithelial cell DNA. Carcinogenesis. 2002;23:529–536
  15. Zhou H, Ivanov VN, Gillespie J, Geard CR, Amundson SA, Brenner DJ, et al. Mechanism of radiation-induced bystander effect: role of the cyclooxygenase-2 signaling pathway. Proc Natl Acad Sci U S A. 2005;102:14641–14646
  16. Little JB. Genomic instability and bystander effects: a historical perspective. Oncogene. 2003;22:6978–6987
  17. Lorimore SA, Wright EG. Radiation-induced genomic instability and bystander effects: related inflammatory-type responses to radiation-induced stress and injury? A review. Int J Radiat Biol. 2003;79:15–25
  18. Emerit I. Reactive oxygen species, chromosome mutation, and cancer: possible role of clastogenic factors in carcinogenesis. Free Radic Biol Med. 1994;16:99–109
  19. Huycke MM, Joyce W, Wack MF. Augmented production of extracellular superoxide production by blood isolates of Enterococcus faecalis. J Infect Dis. 1996;173:743–746
  20. Korshunov S, Imlay JA. Detection and quantification of superoxide formed within the periplasm of Escherichia coli. J Bacteriol. 2006;188:6326–6334
  21. McGuinness SM, Shibuya ML, Ueno AM, Vannais DB, Waldren CA. Mutant quantity and quality in mammalian cells (AL) exposed to cesium-137 gamma radiation: effect of caffeine. Radiat Res. 1995;142:247–255
  22. Waldren CA, Vannais DB, Ueno AM. A role for long-lived radicals (LLR) in radiation-induced mutation and persistent chromosomal instability: counteraction by ascorbate and RibCys but not DMSO. Mutat Res. 2004;551:255–265
  23. Kraemer SM, Vannais DB, Kronenberg A, Ueno A, Waldren CA. Gamma-ray mutagenesis studies in a new human–hamster hybrid, ALCD59+/-, which has two human chromosomes 11 but is hemizygous for the CD59 gene. Radiat Res. 2001;156:10–19
  24. Buettner GR. The pecking order of free radicals and antioxidants: lipid peroxidation, α-tocopherol, and ascorbate. Arch Biochem Biophys. 1993;300:535–543
  25. Hensley HH, Chang WC, Clapper ML. Detection and volume determination of colonic tumors in Min mice by magnetic resonance micro-imaging. Magn Reson Med. 2004;524–529
  26. Jiang Q, Elson-Schwab I, Courtemanche C, Ames BN. γ-Tocopherol and its major metabolite, in contrast to α-tocopherol, inhibit cyclooxygenase activity in macrophages and epithelial cells. Proc Natl Acad Sci U S A. 2000;97:11494–11499
  27. Marnett LJ, DuBois RN. COX-2: a target for colon cancer prevention. Annu Rev Pharmacol Toxicol. 2002;42:55–80
  28. Kraehenbuhl J-P, Neutra MR. Epithelial M cells: differentiation and function. Annu Rev Cell Dev Biol. 2000;16:301–332
  29. Gentry-Weeks CR, Karkhoff-Schweizer R, Pikis A, Estay M, Keith JM. Survival of Enterococcus faecalis in mouse peritoneal macrophages. Infect Immun. 1999;67:2160–2165
  30. Saran M, Bors W. Oxygen radicals acting as chemical messengers: a hypothesis. Free Radic Res Commun. 1989;7:213–220
  31. Balish E, Warner T. Enterococcus faecalis induces inflammatory bowel disease in interleukin-10 knockout mice. Am J Pathol. 2002;160:2253–2257
  32. Kim SC, Tonkonogy SL, Albright CA, Tsang J, Balish EJ, Braun J, et al. Variable phenotypes of enterocolitis in IL-10 deficient mice monoassociated with two different commensal bacteria. Gastroenterology. 2005;128:891–906
  33. Marnett LJ, Riggins JN, West JD. Endogenous generation of reactive oxidants and electrophiles and their reactions with DNA and protein. J Clin Invest. 2003;111:583–593
  34. Macarthur M, Hold GL, El-Omar EM. Inflammation and Cancer II (Role of chronic inflammation and cytokine gene polymorphisms in the pathogenesis of gastrointestinal malignancy). Am J Physiol Gastrointest Liver Physiol. 2004;286:G515–G520
  35. Hussain Z, Kuhn M, Lannigan R, Austin TW. Microbiological investigation of an outbreak of bacteraemia due to Streptococcus faecalis in an intensive care unit. J Hosp Infect. 1988;12:263–271
  36. Wells CL, Erlandsen SL. Localization of translocating Escherichia coli, Proteus mirabilis, and Enterococcus faecalis within cecal and colonic tissues of monoassociated mice. Infect Immun. 1991;59:4693–4697
  37. Dukes C, Bussey HJR. The number of lymphoid follicles of the human large intestine. J Pathol. 1926;29:111–116
  38. Didierlaurent A, Sirard J-C, Kraehenbuhl J-P, Neutra MR. How the gut senses its content. Cell Microbiol. 2002;4:61–72
  39. Plackett TP, Boehmer ED, Faunce DE, Kovacs EJ. Aging and innate immune cells. J Leukoc Biol. 2004;76:291–299
  40. Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E, et al. Suppression of intestinal polyposis in ApcΔ716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell. 1996;87:803–809
  41. Chulada PC, Thompson MB, Mahler JF, Doyle CM, Gaul BW, Lee C, et al. Genetic disruption of Ptgs-1, as well as of Ptgs-2, reduces intestinal tumorigenesis in Min mice. Cancer Res. 2000;60:4705–4708
  42. Bamba H, Ota S, Kato A, Adachi A, Itoyama S, Matsuaki F. High expression of cyclooxygenase-2 in macrophages of human colonic adenoma. Int J Cancer. 1999;83:470–475
  43. Chapple KS, Cartwright EJ, Hawcroft G, Tisbury A, Bonifer C, Scott N, et al. Localization of cyclooxygenase-2 in human sporadic colorectal adenomas. Am J Pathol. 2000;156:545–553
  44. Shih I-M, Wang T-L, Traverso G, Romans K, Hamilton SR, Ben-Sasson S, et al. Top-down morphogenesis of colorectal tumors. Proc Natl Acad Sci U S A. 2001;98:2640–2645
  45. Wang D, Dubois RN. Prostaglandins and cancer. Gut. 2006;55:115–122
  46. Landino LM, Crews BC, Gierse JK, Hauser SD, Marnett LJ. Mutational analysis of the role of the distal histidine and glutamine residues of prostaglandin-endoperoxide synthase-2 in peroxidase catalysis, hydroperoxide reduction, and cyclooxygenase activation. J Biol Chem. 1997;272:21565–21574
  47. Salisbury BA, Pungliya M, Choi JY, Jiang R, Sun XJ, Stephens JC. SNP and haplotype variation in the human genome. Mutat Res. 2003;526:53–61
  48. Popov SG, Villasmil R, Bernardi J, Grene E, Cardwell J, Wu A, et al. Lethal toxin of Bacillus anthracis causes apoptosis of macrophages. Biochem Biophys Res Commun. 2002;293:349–355
  49. Greenberg ER, Baron JA, Tosteson TD, Freeman DHJ, Beck GJ, Bond JH, et al. A clinical trial of antioxidant vitamins to prevent colorectal adenoma. N Engl J Med. 1994;331:141–147
  50. Jacobs EJ, Connell CJ, Patel AV, Chao A, Rodriguez C, Seymour J, et al. Vitamin C and vitamin E supplement use and colorectal cancer mortality in a large American Cancer Society cohort. Cancer Epidemiol Biomarkers Prev. 2001;10:17–23
  51. Rhee SH, Hwang D. Murine TOLL-like receptor 4 confers lipopolysaccharide responsiveness as determined by activation of NF kappa B and expression of the inducible cyclooxygenase. J Biol Chem. 2000;275:34035–34040
  52. Fukata M, Chen A, Klepper A, Krishnareddy S, Vamadevan AS, Thomas LS, et al. Cox-2 is regulated by Toll-like receptor-4 (TLR4) signaling: role in proliferation and apoptosis in the intestine. Gastroenterology. 2006;131:862–877
  53. Oshima H, Oshima M, Inaba K, Taketo MM. Hyperplastic gastric tumors induced by activated macrophages in COX-2/mPGES-1 transgenic mice. EMBO J. 2004;23:1669–1678
  54. Hwang D. Modulation of the expression of cyclooxygenase-2 by fatty acids mediated through toll-like receptor 4-derived signaling pathways. Faseb J. 2001;15:2556–2564
  55. Lien E, Means TK, Heine H, Yoshimura A, Kusumoto S, Fukase K, et al. Toll-like receptor 4 imparts ligand-specific recognition of bacterial lipopolysaccharide. J Clin Invest. 2000;105:497–504
  56. Nougayrede JP, Homburg S, Taieb F, Boury M, Brzuszkiewicz E, Gottschalk G, et al. Escherichia coli induces DNA double-strand breaks in eukaryotic cells. Science. 2006;313:848–851

 Supported by Department of Veterans Affairs Merit Review Program (to M.M.H.) and the Frances Duffy Endowment.

PII: S0016-5085(06)02521-2

doi:10.1053/j.gastro.2006.11.040

Refers to article:

  • Sporadic Colorectal Cancer: An Infectious Disease? , 03 February 2007

    Frank A. Sinicrope
    Gastroenterology February 2007 (Vol. 132, Issue 2, Pages 797-801)

Gastroenterology
Volume 132, Issue 2 , Pages 551-561, February 2007

Article Tools

* Image Usage & Resolution