Distance burning

Abstract

Gut microbes play a major role in carcinogenesis of the gastrointestinal tract. We and others have shown in mouse models that colonic bacteria also influence the development of extraintestinal cancers including hepatocellular and mammary carcinomas. Microbes such as Helicobacter hepaticus invoke a proinflammatory microenvironment in the lower bowel that may extend to distant organs, often in the absence of histologically evident inflammation. Innate immunity plays a crucial role in the promotion of liver cancer and other systemic diseases by gut microbes. Additional mechanisms include type 1 adaptive immunity, altered metabolism and oxidative stress. Emerging links between host genetics, gut microbes, inflammatory bowel disease and colorectal cancer also may prove useful for the correlation of specific bacterial populations with extraintestinal neoplasms. Interruption of deleterious host-microbe networks through judicious use of antibiotics and targeted molecular therapies may help reduce the incidence of liver, breast and other human cancers.

By now it is well established that gut microbes influence tumor formation in the gastrointestinal (GI) tract. Helicobacter pylori is the leading cause of gastric carcinoma worldwide and lower bowel organisms play a significant though incompletely defined role in colorectal cancer (CRC).^1–3 However, until recently the influence of gut bacteria on diseases beyond the GI tract was underappreciated. The laboratory of Jeffrey Gordon at Washington University in St. Louis changed that with a landmark paper in 2006, showing that obesity could be induced in lean germfree mice simply by transferring intestinal bacteria from overweight ob/ob animals.⁴ In a vicious cycle, ob/ob mice lacking leptin ingested too much food, creating a selective advantage for energy-efficient Firmicutes relative to Bacteroidetes, resulting in even greater caloric absorption from the gut. Remarkably, bacteria transmitted from ob/ob mice to lean germfree animals retained the ability to harvest energy, and produced the same obesity phenotype in wild-type (WT) mice with intact leptin signaling. Subsequent studies by others linked gut microbes to increased risk of type II diabetes (T2D).⁵ Wen et al. showed that gut microbes even influence the development of type I diabetes (T1D), classically regarded as an autoimmune disease resulting in pancreatic β-cell depletion.⁶ When T1D-prone nonobese diabetic (NOD) mice were crossed with MyD88^-/- mutants, offspring were protected from disease. Moreover, in a reversal of the negative host-microbe cycle found in the ob/ob system, gut microbes transmitted from MyD88^-/- mice had the positive effect of abrogating diabetes in NOD recipients when compared with bacteria derived from WT donors. Together these studies showed that gut microbes, acting through metabolic and immune pathways, have wide-ranging effects on systemic health and disease. Our group and others extended this paradigm by demonstrating that gut bacteria also influence cancer risk in extraintestinal organs.^7,8 Precise mechanisms remain to be elucidated, but preliminary results point to the twin pillars of immunity and metabolism as key effector mechanisms of distant tumor promotion by intestinal bacteria. Though currently limited to animal models, these findings have significant implications for human cancer screening and intervention.

Our study set out to determine whether Helicobacter hepaticus, a murine enterohepatic bacterium that causes hepatitis and hepatocellular carcinoma (HCC) in some strains of mice, would promote liver cancer initiated by aflatoxin B1 (AFB1) and/or hepatitis C virus (HCV) transgenes.⁷ The initial hypothesis was that bacteria would translocate from the lower bowel to the liver and incite hepatic inflammation that would accelerate tumorigenesis, particularly in males. Indeed, as expected, we observed that H. hepaticus persistently colonized the lower bowel of all inoculated mice and increased the number and malignancy of liver tumors (Fig. 1). However, to our surprise we found no correlation between microbial induction of hepatitis and tumor incidence, multiplicity or grade in males. Moreover, H. hepaticus invoked preneoplastic and neoplastic changes in AFB1-initiated female mice, none of which exhibited hepatitis. To determine whether transmigration of the organism to the liver was required for tumor promotion, we performed highly sensitive nested bacterial DNA PCR with near single-copy sensitivity. In agreement with previous work, we found a direct correlation between intrahepatic bacterial colonization and histologically overt hepatitis in infected males.^9–11 However, there was a complete absence of concordance between detection of bacteria in the liver and tumor development in either sex (p = 0.99). Additionally, we found no evidence of bacterial invasion of the lymphatic system by PCR or special stains. In short, the microbes exerted protumorigenic effects on the liver without any requirement for hepatic translocation or hepatitis induction. We were left with one inescapable conclusion: H. hepaticus promoted liver cancer from its endogenous niche in the lower bowel.

Figure 1.

Enterically confined H. hepaticus promoted aflatoxin B1-induced liver cancer in mice in association with NFκB activation. (A) Liver histopathology of male (top row) and female (bottom row) mice in the 4 treatment groups; *adenoma, ^†HCC. (B) Tumor incidence comparison in male and female mice from each treatment group. (C) Adenoma versus carcinoma multiplicity and total tumor surface area comparison in AFB1-exposed males with and without H. hepaticus. (D) Pathways analysis of NFκB-anchored molecular networks altered by H. hepaticus in both the lower bowel and liver. Images reproduced from Fox et al. Gut 2010. {Fox 2010 #4672}.

To identify potential mechanisms by which the intestinally confined bacteria accomplished this surprising feat, we performed a variety of assays. By microarray, we found that H. hepaticus colonization altered transcription of thousands of genes both in the lower bowel and liver. This was true even in animals with no histologically evident inflammation in either tissue compartment. Pathways analysis revealed NFκB at the hub of the most disrupted signaling networks both in the colon and liver (Fig. 1D). Additionally, lower bowel colonization by H. hepaticus was associated with upregulated expression of tumor-necrosis factor α (TNFα) and interferon γ in the liver. In addition to immunity, H. hepaticus colonization produced altered transcription of many metabolic genes in the lower bowel and liver including cytochrome P450 isoforms, cholesterol and steroid metabolizing genes, sulfotransferases and others. The combination of H. hepaticus and AFB1 increased baseline hepatocyte proliferation and liver oxidative stress, and invoked well known oncogenes and tumor-associated signaling pathways including Fos/Jun, cyclin D1 and Wnt/β-catenin. Taken together, these results demonstrated that liver tumor promotion by H. hepaticus was associated with enterohepatic metabolic derangement and induction of proinflammatory tissue microenvironments resulting in hepatocellular oxidative stress and altered proliferation kinetics. Further work will be required to clarify specific mechanisms involved in transportal tumor promotion by gut microbes. Nevertheless, it seems clear that immunity and metabolism will anchor pathways relevant to microbial crosstalk with distant cancers.

Whereas ours was the first paper to demonstrate extraintestinal tumor promotion by gut microbes in a WT mouse model, a prior study by Rao et al. reported an increased risk of mammary cancer in Apc^Min/+Rag2^-/- mice infected with H. hepaticus.⁸ In that study, H. hepaticus increased both intestinal polyp numbers and mammary tumors in the mutant mice. Interestingly, adenomatous polyps were increased not only in the lower bowel but also the small intestine, which is not a preferred niche of the organism. Intestinal and mammary tumor promotion was mediated by innate immunity, as demonstrated by inhibition either with a neutralizing anti-TNFα antibody or adoptive transfer of regulatory T cells (T_R). Microbial induction of disease was suggested by the requirement for interleukin-10 (IL10)-competent T_R to suppress tumorigenesis. It has been shown that IL10 is required to modulate antimicrobial responses but not autoimmunity.¹² Moreover, adoptively transferred T_R primed by prior bacterial exposure were more effective at inhibiting tumors than lymphocytes from H. hepaticus-naïve donors. However, there were two limitations to the study. First, it is difficult to directly translate results from genetically engineered mice to humans. Second, the squamofollicular mammary tumors that develop spontaneously in Apc^Min/+ mice do not resemble human breast cancer. Nevertheless, this important study offered the first proof-of-principle that gut microbes can promote extraintestinal cancer, and that innate immunity is a major driver of the process.

Whereas immunity in general, and innate immunity in particular, is implicated in extraintestinal tumor promotion by gut microbes, precise molecular mechanisms remain to be elucidated. With regard to liver tumor promotion, there are at least six nonexclusive pathways that may be involved (Fig. 2). First, gut microbes may compromise mucosal integrity of the colonic epithelium, resulting in uptake of bacterial and/or luminal toxins into portal circulation. Second, lower bowel microbes may incite secretion of proinflammatory cytokines and/or growth factors from the colonic epithelium that are taken directly to the liver via portal circulation. Third, intestinal bacteria may provoke a generalized proinflammatory state of the immune system with systemic ramifications. The dependence on innate immunity for microbial promotion of T1D and mammary tumors described above highlights the importance of this pathway. Fourth, lower bowel bacteria may interrupt bile recirculation, resulting in increased chemical and/or oxidative stress and/or altered fibroblast growth factor homeostasis.¹³ Alterations in lipid signaling and cholesterol/steroid metabolism are prominent features of microbially-induced liver cancer in mouse models.^10,14 Fifth, altered energy dynamics associated with bacterial nutrient processing, as shown by Turnbaugh et al. and others,^4,5 may produce systemic perturbations in the metabolic state of the host favoring the expansion of neoplastic cells at distant sites. Finally, gut microbes may disrupt homeostasis of circulating blood protein sensors such as complement, clotting factors, proteases and others. In a prior study of H. hepaticus-induced HCC, we noted a tumor-associated increase in expression of numerous proteases and liver-derived circulating factors and their receptors including factor II/thrombin receptor (protease-activated receptor 1), serine peptidase inhibitors, serine proteases and cathepsins.¹⁰ Much work needs to be done to elucidate mechanisms by which gut microbes promote cancer of the liver and other extraintestinal organs, but it seems clear that targets in addition to the “usual suspects” of cytokines and growth factors must be considered as candidate effectors.

Figure 2.

Potential mechanisms of liver tumor promotion by gut microbes. Intestinal bacteria may promote hepatic (and other) distant neoplasms through at least six nonexclusive pathways: (1) altered mucosal permeability with toxin absorption; (2) induction of proinflammatory mucosal/epithelial microenvironments; (3) generalized immune activation; (4) activation of circulating proteases and other serum factors; (5) bacterial nutrient processing and host metabolomics; and (6) disrupted bile and cholesterol/fatty acid homeostasis. APC, antigen-presenting cells; FGFs, fibroblast growth factors; GALT, gut-associated lymphoid tissue; MLN, mesenteric lymph node.

The demonstration that H. hepaticus can promote extraintestinal tumors in mice has important ramifications for biomedical research. Enterohepatic Helicobacter spp. are virtually ubiquitous in research mouse colonies, and have only been eliminated from commercial breeding facilities in the past 15 years.¹⁵ Thus, historical discrepancies in rodent experimental cancer outcomes from different laboratories may reflect institutionspecific background flora. The NIEHS National Toxicology Program identified H. hepaticus as a potential confounder of at least twelve chronic bioassay studies, and reported distinct findings when a historical study of triethanoloamine was repeated under Helicobacter-free conditions.^16–18 With regard to our own studies, we found that HCV-transgenic mice required the presence of H. hepaticus to develop tumors.⁷ This may help explain reported discrepancies in liver cancer incidence from HCV FL-N/35 and other strains of genetically engineered mice housed in different animal facilities (S. Lemon, personal communication).¹⁹ Indeed, it was previously shown in a mouse gallstone disease model that C57/L mice fed a lithogenic diet failed to develop choleliths unless also infected with Helicobacter spp.²⁰ Because of the difficulty and expense of eliminating murine Helicobacter spp., most research facilities do not routinely test for them. “Specific pathogen-free” rarely means “Helicobacter-free”. Moreover, non-Helicobacter bacteria including Enterococcus spp., certain E. coli strains, and others may invoke intestinal and extraintestinal inflammation.^21,22 Thus, gut microbes in general, and Helicobacter spp. in particular, must be taken into account when interpreting mouse models of cancer regardless of the tissue of origin.²³

Do gut microbes promote HCC and other non-GI cancers in humans? At present we don't know. Helicobacter spp. have been associated with biliary and hepatocellular carcinomas in humans, but only when isolated from the affected tissue.^24–27 Given the association of Helicobacter spp. and other microbes with inflammatory bowel disease (IBD) in mouse models, it seems plausible that IBD-associated bacteria also may promote distant cancers in humans.²³ However, to date no single organism has been identified as a cause human IBD. This may reflect a polymicrobial basis of disease. In mouse models of IBD, H. hepaticus requires the presence of other intestinal bacteria to exert its effects; moreover, combinations of non-Helicobacter bacteria may incite chronic colonic inflammation and/or cancer where neither bacterium alone is pathogenic.^22,28,29 Sophisticated multivariate analyses correlating mixed bacterial populations with human IBD and/or CRC will be facilitated by the advent of cost-effective genomic screening and characterization of the human microbiome.³⁰ The same approach may prove useful for testing whether enteric bacteria influence the risk of human extraintestinal cancers, including liver and breast. If so, new therapeutic strategies, including the judicious use of antibiotics, may prove useful to reduce the overall incidence of cancer in humans. One thing is certain: we have only scratched the surface in understanding how gut microbes influence health and disease in the GI tract and beyond.

Abbreviations

GI

gastrointestinal

CRC

colorectal cancer

WT

wild-type

T2D

type 2 diabetes

T1D

type 1 diabetes

NOD

nonobese diabetic

HCC

hepatocellular carcinoma

AFB1

aflatoxin B1

HCV

hepatitis C virus

TNFα

tumor necrosis factor α

IFNγ

interferon γ

IL-10

interleukin 10

IBD

inflammatory bowel disease

Addendum to: Fox JG, Feng Y, Theve EJ, Fry RC, Raczynski AR, Fiala JLA, et al. Gut microbes define liver cancer risk in mice exposed to chemical and viral transgenic hepatocarcinogens. Gut. 2010;59:88–97. doi: 10.1136/gut.2009.183749.