Activated intestinal macrophages in patients with cirrhosis release NO and IL-6 that may disrupt intestinal barrier function.

Du Plessis J, Vanheel H, Janssen CE, et al. Activated intestinal macrophages in patients with cirrhosis release NO and IL-6 that may disrupt intestinal barrier function. J Hepatol 2013;58:1125–1132.Bacterial infections commonly complicate decompensated cirrhosis and are thought to be associated with increased bacterial translocation across the intestinal barrier (Gastroenterology 2011;141:1220–1230; Hepatology 2006;44:633–639). Alterations in circulating bacterial DNA, even in the presence of negative cultures, increases the risk of hepatic decompensation, variceal hemorrhage, and hepatorenal syndrome (Gastroenterology 2007;133:818–824). In health, the gut epithelial barrier provides an effective barrier against microorganisms while simultaneously providing a semipermeable membrane for nutrient absorption (Nat Rev Immunol 2009;9:799–809). In cirrhosis, studies have confirmed that intestinal bacterial flora is altered (Hepatology 2006;45:744–757). However, it is unclear why the intestinal epithelial barrier fails, facilitating translocation of bacterial products and DNA (Hepatology 2008;48:1924–1931; Hepatology 2010;52:2044–2052).The authors of this study hypothesize that, similar to other inflammatory states, intestinal macrophage activation occurs in cirrhosis. They therefore determined the intestinal macrophage phenotype in decompensated cirrhosis and whether these macrophages are capable of modulating intestinal permeability.This study included South African patients with nonalcoholic steatohepatitis or alcohol-related cirrhosis aged 18–80 with either decompensated (n = 29) or compensated (n = 15) disease. Patients underwent peripheral blood sampling and gastroscopy for variceal screening, where duodenal biopsies were taken. Controls (n = 19) were patients undergoing gastroscopy for reflux or dyspepsia.Plasma levels of circulating endotoxin (lipopolysaccharide [LPS]) and LPS-binding protein were used as a surrogate marker of bacterial translocation. Both LPS and LPS-binding protein levels were significantly increased in patients with cirrhosis.Mucosal mononuclear cells were isolated and examined for activation status (expression of CD68, CD14, CD16, Trem-1), monocyte/macrophage lineage (CD33, CD14), costimulatory molecules (CD80, CD86), and surface expression of Toll-like receptors -2 and -4 using FACS analysis and immunohistochemistry. This demonstrated a significant increase in an activated macrophage phenotype (expressing CD68+) with increased frequency of CD33/CD14 expression and coexpression of innate immune receptors for LPS (CD14), CD33, and Trem-1 in cirrhosis compared with controls. This demonstrates that intestinal macrophages in cirrhosis have an activated phenotype and express innate immune receptors for bacterial translocation. There was no difference in the prevalence of dendritic cells, natural killer cells, or activated B cells.To analyze the transcriptional profile of the intestinal mucosa, analysis of inflammatory and housekeeping gene expression was performed. Quantitative reverse transcriptase polymerase chain reaction confirmed up-regulation of inducible nitrous oxide synthase (iNOS) and other genes associated with inflammatory cell and monocyte recruitment, including interleukin (IL)-8, chemokine ligand (CCL)2, and CCL13. Increased levels of IL-6, IL-8, and CCL2/monocyte chemotactic protein (MCP)-1 were found in the supernatant of biopsy cultures. IL-8 and CCL2/MCP-1 are secreted by activated CD14+Trem-1+ macrophages in response to microbial stimuli in inflammatory bowel disease, which may be important in the pathogenesis of Crohn’s disease (J Clin Invest 2008;118:2269–2280; J Immunol 2010;184:4069–4073). Furthermore, IL-8 production is known to be increased by Escherichia coli in Crohn’s disease (Gastroenterology 2004;127:80–93), suggesting that increased IL-8 observed at the transcriptional and protein levels in cirrhosis may reflect a response to bacterial induced stimulation.The relevance of increased iNOS mRNA at the transcription level was further assessed by immunohistochemistry. This demonstrated increased iNOS synthesis, with increased iNOS-activated macrophages in cirrhotics.Because intestinal macrophages that express innate response receptors such as CD14+ and TREM-1 have been associated with inflammation and pro-inflammatory cytokine production, colocalization studies were performed. These demonstrated that CD14+ cells were iNOS positive, thus confirming the presence of classically activated intestinal macrophages. Both increased numbers of IL-6–positive cells in decompensated cirrhosis and colocalization of IL-6 and IL-8 with CD68+ and iNOS+ macrophages were demonstrated by immunohistochemistry, indicating that macrophages are the major source of intestinal IL-6 released in cirrhosis. IL-6 was predominantly present in CD11c– cells, which are not dendritic cells, thus indicating that dendritic cells are not a major source of this pro-inflammatory cytokine. Interestingly, Trem-1+ macrophages and intestinal epithelial cells have been shown as a major source of IL-6 in Crohn’s disease. IL-8 colocalized in iNOS+ cells, but only in a subpopulation of CD68+ cells, suggesting that other inflammatory cells also produce IL-8.Histopathologic and ultrastructural analysis of duodenal wall and epithelial barrier was performed by light microscopy and transmission electron microscopy. There was no difference in morphology of the interepithelial junctions between groups. However, functional analysis demonstrated a reduced transepithelial electrical resistance and higher transepithelial passage of FITC-dx4 in cirrhosis compared with controls, indicating impaired duodenal barrier function. Structural tight junction proteins (ZO-1,11 Occludin, and claudin-1) and gap junction protein (connexion-43) were not different at the mRNA and protein levels between groups, but increased levels of claudin-2, a known pore-forming tight junction protein, were observed via Western blot, and immunohistochemistry documented a vesicular staining pattern of the protein on the apical pole of epithelial cells in decompensated cirrhosis, suggesting that it is responsible for increased intestinal permeability. Indeed, epithelial barrier dysfunction and elevated claudin-2 expression associated with bacterial translocation have been previously observed in Crohn’s disease and HIV infection (Gut 2007;56:61–72; J Acquir Immune Defic Syndr 2010;55:306–315). Importantly, it was recently demonstrated that IL-6 regulates claudin-2 expression and permeability in cultured intestinal epithelial cells, establishing a clear link between localized inflammation and intestinal permeability (J Biol Chem 2011;286:31263–31271).CommentThe results from this study demonstrate that activated intestinal macrophages are the major source of IL-6 and NO in patients with cirrhosis, most likely as a result of bacterial stimulation. These cells are capable of producing factors that influence intestinal permeability. Furthermore, the transepithelial resistance and tracer flux studies confirm that permeability of the duodenal epithelium is compromised in this patient group.The liver receives 70% of its blood supply directly from the intestine via the portal vein. It is therefore continually exposed to gut bacteria and bacterial components. Recent research has focused on the “gut–liver axis” and the possible role of the intestinal microbiota in liver disease and cirrhosis (Nat Rev Gastroenterol Hepatol 2011;9:72–74).In cirrhosis, spontaneous bacterial peritonitis, variceal hemorrhage, and encephalopathy are thought to result from impaired peristalsis, alteration in blood flow through the gut–liver axis, bacterial overgrowth, immune dysfunction, and sepsis (Curr Microbiol 2012;65:7–13). It is felt that spontaneous bacterial peritonitis, with a mortality of 20%, is a result of changes in fecal microbiota, increased intestinal permeability and bacterial translocation across the bowel wall (Hepatology 2011;54:562–572, Gut 2012;61:297–310). Studies demonstrate that there is a change in the intestinal microbiome of cirrhotics, with a dramatic increase in Enterobacteriaceae and Enterococcus species in the fecal microbiome (Curr Microbiol 2012;65:7–13). More recently, it has been shown that treatment with rifaximin improves cognitive function in hepatic encephalopathy through alteration of the gut microbiota and reduced endotoxemia (PLOS One 2013;8:e60042).Existing evidence for the role of intestinal microbiome in liver disease has so far concentrated on nonalcoholic fatty liver disease and the subsequent development of nonalcoholic steatohepatitis. LPS, a key constituent of many bacteria present in our microbiota is thought to be of importance (Gastroenterology 1998;114:842–845). Circulating levels of LPS are higher in humans on a high-fat/high-carbohydrate diet. Moreover, similar to a high-fat diet, endotoxemia induced increased liver and adipose tissue weight and insulin resistance in lean mice. Furthermore, these changes were prevented by antibiotic therapy (Hepatology 2010;52:1836–1846). Other studies have demonstrated that liver macrophages are particularly sensitive to gut-derived endotoxin (Gut Liver 2012;6:149–171).Expression and turnover of tight junctions maintaining the structural integrity and permeability of the intestinal epithelium are known to be influenced by oxidative stress and inflammation (Antioxid Redox Signal 2011;15:1255–1270; J Biol Chem 2011;286:31263–31271). The first line of defense to microorganisms breaching the epithelial barrier is from lamina propria-localized intestinal macrophages. In health, these highly anergic macrophages express a CD33+CD14–Trem-1–phenotype and are inert to microbial stimuli, LPS-induced cytokine and iNOS production (J Clin Invest 2005;115:66–75; Mucosal Immunol 2011;4:31–42). However, in inflammatory bowel disease intestinal macrophages are known to express innate response receptors such as CD14 and TREM-1, respond to microbial stimulation, produce high levels of inflammatory cytokines (interferon-γ, IL-23, IL-1β), and secrete IL-8, CCL2/MCP-1, which may be important in the pathogenesis of Crohn’s disease (J Clin Invest 2008;118:2269–2280; J Immunol 2010;184:4069–4073). Furthermore, activated CD14+ macrophages in necrotizing enterocolitis produce nitric oxide that impairs endothelial repair (Gastroenterology 2007;132:2395–2411; Gastrointest Liver Physiol 2008;294:G109–G119), and macrophage activation in HIV correlates with bacterial translocation and persistent immune activation (J Infect Dis 2010;202:723–733). In this study, colocalization studies demonstrated that CD14+ cells were iNOS positive, thus confirming the presence of classically activated intestinal macrophages.The intestinal epithelial barrier has a number of components, including the intestinal mucus layer and secreted antimicrobial peptides including defensins and RegIIIγ (J Intern Med 2012;271:421–428; Science 2011;334:255–258). Another component is the tight junctional complex located above the adherent complex between the epithelial cells. Two physiologically distinct pathways are recognized through the tight junction. The pore pathway has claudins as the major molecular component and allows passage of small ions and uncharged molecules. The current study suggests that there is an increase in the permeability of this pathway that occurs in cirrhosis as claudin 2 expression is increased. In contrast, the leak pathway has occludin as its major membrane protein and allows passage of large molecules, including proteins (Annu Rev Physiol 2011;73:283–309). Further investigation into the contribution of the leak pathway in cirrhosis is required, because LPS is thought to be too large to traverse the pore pathway.Epithelial cells are constantly being shed from villi of the small intestine or the surface of the colon into the lumen. Barrier function is maintained at sites of cell shedding by a redistribution of tight junction proteins to plug the gap left by the shedding cell (Gastroenterology 2011;140:1208–1218). LPS increases the rate of cell shedding (Gastroenterology 2013;144:S668). When shedding rates are high sometimes the redistribution of tight junction proteins fails to seal the epithelium, particularly if ≥2 adjacent cells are shed simultaneously (Gastroenterology 2007;133:1769–1778; Gut 2012;61:1146–1153). Thus, it is plausible that failure of the intestinal barrier as a result of increased cell shedding might be the mechanism of increased systemic LPS in cirrhotic patients.It is evident that the gut–liver axis is an integral pathway in the pathogenesis of liver disease and this study further highlights some of the potential mechanisms, which result in increased levels of proinflammatory cytokines, increased gut permeability, bacterial translocation, and subsequent liver disease. The authors have described a phenomenon whereby intestinal macrophages are activated and mucosal integrity compromised in this patient group. It is, however, not clear why this occurs. It may result from hemodynamic changes in intestinal blood flow as a result of changes in portal pressure, or indeed may be related to enterohepatic sharing of lymphocytes (Gastroenterology 2009;137:320–329; Lancet 2002;359:150–157). Further studies are required to explore treating cirrhotic patients by altering gut flora by antibiotic therapy, recolonization with therapeutically defined microbial populations, or even fecal transplantation from healthy donors. This study provides further evidence to support the use of ciprofloxacin and rifaximin in clinical practice for the prevention of spontaneous bacterial peritonitis and hepatic encephalopathy. It suggests a mechanism via which these antibiotics may be efficacious in prophylaxis, and thus supports current clinical practice. Du Plessis J, Vanheel H, Janssen CE, et al. Activated intestinal macrophages in patients with cirrhosis release NO and IL-6 that may disrupt intestinal barrier function. J Hepatol 2013;58:1125–1132. Bacterial infections commonly complicate decompensated cirrhosis and are thought to be associated with increased bacterial translocation across the intestinal barrier (Gastroenterology 2011;141:1220–1230; Hepatology 2006;44:633–639). Alterations in circulating bacterial DNA, even in the presence of negative cultures, increases the risk of hepatic decompensation, variceal hemorrhage, and hepatorenal syndrome (Gastroenterology 2007;133:818–824). In health, the gut epithelial barrier provides an effective barrier against microorganisms while simultaneously providing a semipermeable membrane for nutrient absorption (Nat Rev Immunol 2009;9:799–809). In cirrhosis, studies have confirmed that intestinal bacterial flora is altered (Hepatology 2006;45:744–757). However, it is unclear why the intestinal epithelial barrier fails, facilitating translocation of bacterial products and DNA (Hepatology 2008;48:1924–1931; Hepatology 2010;52:2044–2052). The authors of this study hypothesize that, similar to other inflammatory states, intestinal macrophage activation occurs in cirrhosis. They therefore determined the intestinal macrophage phenotype in decompensated cirrhosis and whether these macrophages are capable of modulating intestinal permeability. This study included South African patients with nonalcoholic steatohepatitis or alcohol-related cirrhosis aged 18–80 with either decompensated (n = 29) or compensated (n = 15) disease. Patients underwent peripheral blood sampling and gastroscopy for variceal screening, where duodenal biopsies were taken. Controls (n = 19) were patients undergoing gastroscopy for reflux or dyspepsia. Plasma levels of circulating endotoxin (lipopolysaccharide [LPS]) and LPS-binding protein were used as a surrogate marker of bacterial translocation. Both LPS and LPS-binding protein levels were significantly increased in patients with cirrhosis. Mucosal mononuclear cells were isolated and examined for activation status (expression of CD68, CD14, CD16, Trem-1), monocyte/macrophage lineage (CD33, CD14), costimulatory molecules (CD80, CD86), and surface expression of Toll-like receptors -2 and -4 using FACS analysis and immunohistochemistry. This demonstrated a significant increase in an activated macrophage phenotype (expressing CD68+) with increased frequency of CD33/CD14 expression and coexpression of innate immune receptors for LPS (CD14), CD33, and Trem-1 in cirrhosis compared with controls. This demonstrates that intestinal macrophages in cirrhosis have an activated phenotype and express innate immune receptors for bacterial translocation. There was no difference in the prevalence of dendritic cells, natural killer cells, or activated B cells. To analyze the transcriptional profile of the intestinal mucosa, analysis of inflammatory and housekeeping gene expression was performed. Quantitative reverse transcriptase polymerase chain reaction confirmed up-regulation of inducible nitrous oxide synthase (iNOS) and other genes associated with inflammatory cell and monocyte recruitment, including interleukin (IL)-8, chemokine ligand (CCL)2, and CCL13. Increased levels of IL-6, IL-8, and CCL2/monocyte chemotactic protein (MCP)-1 were found in the supernatant of biopsy cultures. IL-8 and CCL2/MCP-1 are secreted by activated CD14+Trem-1+ macrophages in response to microbial stimuli in inflammatory bowel disease, which may be important in the pathogenesis of Crohn’s disease (J Clin Invest 2008;118:2269–2280; J Immunol 2010;184:4069–4073). Furthermore, IL-8 production is known to be increased by Escherichia coli in Crohn’s disease (Gastroenterology 2004;127:80–93), suggesting that increased IL-8 observed at the transcriptional and protein levels in cirrhosis may reflect a response to bacterial induced stimulation. The relevance of increased iNOS mRNA at the transcription level was further assessed by immunohistochemistry. This demonstrated increased iNOS synthesis, with increased iNOS-activated macrophages in cirrhotics. Because intestinal macrophages that express innate response receptors such as CD14+ and TREM-1 have been associated with inflammation and pro-inflammatory cytokine production, colocalization studies were performed. These demonstrated that CD14+ cells were iNOS positive, thus confirming the presence of classically activated intestinal macrophages. Both increased numbers of IL-6–positive cells in decompensated cirrhosis and colocalization of IL-6 and IL-8 with CD68+ and iNOS+ macrophages were demonstrated by immunohistochemistry, indicating that macrophages are the major source of intestinal IL-6 released in cirrhosis. IL-6 was predominantly present in CD11c– cells, which are not dendritic cells, thus indicating that dendritic cells are not a major source of this pro-inflammatory cytokine. Interestingly, Trem-1+ macrophages and intestinal epithelial cells have been shown as a major source of IL-6 in Crohn’s disease. IL-8 colocalized in iNOS+ cells, but only in a subpopulation of CD68+ cells, suggesting that other inflammatory cells also produce IL-8. Histopathologic and ultrastructural analysis of duodenal wall and epithelial barrier was performed by light microscopy and transmission electron microscopy. There was no difference in morphology of the interepithelial junctions between groups. However, functional analysis demonstrated a reduced transepithelial electrical resistance and higher transepithelial passage of FITC-dx4 in cirrhosis compared with controls, indicating impaired duodenal barrier function. Structural tight junction proteins (ZO-1,11 Occludin, and claudin-1) and gap junction protein (connexion-43) were not different at the mRNA and protein levels between groups, but increased levels of claudin-2, a known pore-forming tight junction protein, were observed via Western blot, and immunohistochemistry documented a vesicular staining pattern of the protein on the apical pole of epithelial cells in decompensated cirrhosis, suggesting that it is responsible for increased intestinal permeability. Indeed, epithelial barrier dysfunction and elevated claudin-2 expression associated with bacterial translocation have been previously observed in Crohn’s disease and HIV infection (Gut 2007;56:61–72; J Acquir Immune Defic Syndr 2010;55:306–315). Importantly, it was recently demonstrated that IL-6 regulates claudin-2 expression and permeability in cultured intestinal epithelial cells, establishing a clear link between localized inflammation and intestinal permeability (J Biol Chem 2011;286:31263–31271). CommentThe results from this study demonstrate that activated intestinal macrophages are the major source of IL-6 and NO in patients with cirrhosis, most likely as a result of bacterial stimulation. These cells are capable of producing factors that influence intestinal permeability. Furthermore, the transepithelial resistance and tracer flux studies confirm that permeability of the duodenal epithelium is compromised in this patient group.The liver receives 70% of its blood supply directly from the intestine via the portal vein. It is therefore continually exposed to gut bacteria and bacterial components. Recent research has focused on the “gut–liver axis” and the possible role of the intestinal microbiota in liver disease and cirrhosis (Nat Rev Gastroenterol Hepatol 2011;9:72–74).In cirrhosis, spontaneous bacterial peritonitis, variceal hemorrhage, and encephalopathy are thought to result from impaired peristalsis, alteration in blood flow through the gut–liver axis, bacterial overgrowth, immune dysfunction, and sepsis (Curr Microbiol 2012;65:7–13). It is felt that spontaneous bacterial peritonitis, with a mortality of 20%, is a result of changes in fecal microbiota, increased intestinal permeability and bacterial translocation across the bowel wall (Hepatology 2011;54:562–572, Gut 2012;61:297–310). Studies demonstrate that there is a change in the intestinal microbiome of cirrhotics, with a dramatic increase in Enterobacteriaceae and Enterococcus species in the fecal microbiome (Curr Microbiol 2012;65:7–13). More recently, it has been shown that treatment with rifaximin improves cognitive function in hepatic encephalopathy through alteration of the gut microbiota and reduced endotoxemia (PLOS One 2013;8:e60042).Existing evidence for the role of intestinal microbiome in liver disease has so far concentrated on nonalcoholic fatty liver disease and the subsequent development of nonalcoholic steatohepatitis. LPS, a key constituent of many bacteria present in our microbiota is thought to be of importance (Gastroenterology 1998;114:842–845). Circulating levels of LPS are higher in humans on a high-fat/high-carbohydrate diet. Moreover, similar to a high-fat diet, endotoxemia induced increased liver and adipose tissue weight and insulin resistance in lean mice. Furthermore, these changes were prevented by antibiotic therapy (Hepatology 2010;52:1836–1846). Other studies have demonstrated that liver macrophages are particularly sensitive to gut-derived endotoxin (Gut Liver 2012;6:149–171).Expression and turnover of tight junctions maintaining the structural integrity and permeability of the intestinal epithelium are known to be influenced by oxidative stress and inflammation (Antioxid Redox Signal 2011;15:1255–1270; J Biol Chem 2011;286:31263–31271). The first line of defense to microorganisms breaching the epithelial barrier is from lamina propria-localized intestinal macrophages. In health, these highly anergic macrophages express a CD33+CD14–Trem-1–phenotype and are inert to microbial stimuli, LPS-induced cytokine and iNOS production (J Clin Invest 2005;115:66–75; Mucosal Immunol 2011;4:31–42). However, in inflammatory bowel disease intestinal macrophages are known to express innate response receptors such as CD14 and TREM-1, respond to microbial stimulation, produce high levels of inflammatory cytokines (interferon-γ, IL-23, IL-1β), and secrete IL-8, CCL2/MCP-1, which may be important in the pathogenesis of Crohn’s disease (J Clin Invest 2008;118:2269–2280; J Immunol 2010;184:4069–4073). Furthermore, activated CD14+ macrophages in necrotizing enterocolitis produce nitric oxide that impairs endothelial repair (Gastroenterology 2007;132:2395–2411; Gastrointest Liver Physiol 2008;294:G109–G119), and macrophage activation in HIV correlates with bacterial translocation and persistent immune activation (J Infect Dis 2010;202:723–733). In this study, colocalization studies demonstrated that CD14+ cells were iNOS positive, thus confirming the presence of classically activated intestinal macrophages.The intestinal epithelial barrier has a number of components, including the intestinal mucus layer and secreted antimicrobial peptides including defensins and RegIIIγ (J Intern Med 2012;271:421–428; Science 2011;334:255–258). Another component is the tight junctional complex located above the adherent complex between the epithelial cells. Two physiologically distinct pathways are recognized through the tight junction. The pore pathway has claudins as the major molecular component and allows passage of small ions and uncharged molecules. The current study suggests that there is an increase in the permeability of this pathway that occurs in cirrhosis as claudin 2 expression is increased. In contrast, the leak pathway has occludin as its major membrane protein and allows passage of large molecules, including proteins (Annu Rev Physiol 2011;73:283–309). Further investigation into the contribution of the leak pathway in cirrhosis is required, because LPS is thought to be too large to traverse the pore pathway.Epithelial cells are constantly being shed from villi of the small intestine or the surface of the colon into the lumen. Barrier function is maintained at sites of cell shedding by a redistribution of tight junction proteins to plug the gap left by the shedding cell (Gastroenterology 2011;140:1208–1218). LPS increases the rate of cell shedding (Gastroenterology 2013;144:S668). When shedding rates are high sometimes the redistribution of tight junction proteins fails to seal the epithelium, particularly if ≥2 adjacent cells are shed simultaneously (Gastroenterology 2007;133:1769–1778; Gut 2012;61:1146–1153). Thus, it is plausible that failure of the intestinal barrier as a result of increased cell shedding might be the mechanism of increased systemic LPS in cirrhotic patients.It is evident that the gut–liver axis is an integral pathway in the pathogenesis of liver disease and this study further highlights some of the potential mechanisms, which result in increased levels of proinflammatory cytokines, increased gut permeability, bacterial translocation, and subsequent liver disease. The authors have described a phenomenon whereby intestinal macrophages are activated and mucosal integrity compromised in this patient group. It is, however, not clear why this occurs. It may result from hemodynamic changes in intestinal blood flow as a result of changes in portal pressure, or indeed may be related to enterohepatic sharing of lymphocytes (Gastroenterology 2009;137:320–329; Lancet 2002;359:150–157). Further studies are required to explore treating cirrhotic patients by altering gut flora by antibiotic therapy, recolonization with therapeutically defined microbial populations, or even fecal transplantation from healthy donors. This study provides further evidence to support the use of ciprofloxacin and rifaximin in clinical practice for the prevention of spontaneous bacterial peritonitis and hepatic encephalopathy. It suggests a mechanism via which these antibiotics may be efficacious in prophylaxis, and thus supports current clinical practice. The results from this study demonstrate that activated intestinal macrophages are the major source of IL-6 and NO in patients with cirrhosis, most likely as a result of bacterial stimulation. These cells are capable of producing factors that influence intestinal permeability. Furthermore, the transepithelial resistance and tracer flux studies confirm that permeability of the duodenal epithelium is compromised in this patient group. The liver receives 70% of its blood supply directly from the intestine via the portal vein. It is therefore continually exposed to gut bacteria and bacterial components. Recent research has focused on the “gut–liver axis” and the possible role of the intestinal microbiota in liver disease and cirrhosis (Nat Rev Gastroenterol Hepatol 2011;9:72–74). In cirrhosis, spontaneous bacterial peritonitis, variceal hemorrhage, and encephalopathy are thought to result from impaired peristalsis, alteration in blood flow through the gut–liver axis, bacterial overgrowth, immune dysfunction, and sepsis (Curr Microbiol 2012;65:7–13). It is felt that spontaneous bacterial peritonitis, with a mortality of 20%, is a result of changes in fecal microbiota, increased intestinal permeability and bacterial translocation across the bowel wall (Hepatology 2011;54:562–572, Gut 2012;61:297–310). Studies demonstrate that there is a change in the intestinal microbiome of cirrhotics, with a dramatic increase in Enterobacteriaceae and Enterococcus species in the fecal microbiome (Curr Microbiol 2012;65:7–13). More recently, it has been shown that treatment with rifaximin improves cognitive function in hepatic encephalopathy through alteration of the gut microbiota and reduced endotoxemia (PLOS One 2013;8:e60042). Existing evidence for the role of intestinal microbiome in liver disease has so far concentrated on nonalcoholic fatty liver disease and the subsequent development of nonalcoholic steatohepatitis. LPS, a key constituent of many bacteria present in our microbiota is thought to be of importance (Gastroenterology 1998;114:842–845). Circulating levels of LPS are higher in humans on a high-fat/high-carbohydrate diet. Moreover, similar to a high-fat diet, endotoxemia induced increased liver and adipose tissue weight and insulin resistance in lean mice. Furthermore, these changes were prevented by antibiotic therapy (Hepatology 2010;52:1836–1846). Other studies have demonstrated that liver macrophages are particularly sensitive to gut-derived endotoxin (Gut Liver 2012;6:149–171). Expression and turnover of tight junctions maintaining the structural integrity and permeability of the intestinal epithelium are known to be influenced by oxidative stress and inflammation (Antioxid Redox Signal 2011;15:1255–1270; J Biol Chem 2011;286:31263–31271). The first line of defense to microorganisms breaching the epithelial barrier is from lamina propria-localized intestinal macrophages. In health, these highly anergic macrophages express a CD33+CD14–Trem-1–phenotype and are inert to microbial stimuli, LPS-induced cytokine and iNOS production (J Clin Invest 2005;115:66–75; Mucosal Immunol 2011;4:31–42). However, in inflammatory bowel disease intestinal macrophages are known to express innate response receptors such as CD14 and TREM-1, respond to microbial stimulation, produce high levels of inflammatory cytokines (interferon-γ, IL-23, IL-1β), and secrete IL-8, CCL2/MCP-1, which may be important in the pathogenesis of Crohn’s disease (J Clin Invest 2008;118:2269–2280; J Immunol 2010;184:4069–4073). Furthermore, activated CD14+ macrophages in necrotizing enterocolitis produce nitric oxide that impairs endothelial repair (Gastroenterology 2007;132:2395–2411; Gastrointest Liver Physiol 2008;294:G109–G119), and macrophage activation in HIV correlates with bacterial translocation and persistent immune activation (J Infect Dis 2010;202:723–733). In this study, colocalization studies demonstrated that CD14+ cells were iNOS positive, thus confirming the presence of classically activated intestinal macrophages. The intestinal epithelial barrier has a number of components, including the intestinal mucus layer and secreted antimicrobial peptides including defensins and RegIIIγ (J Intern Med 2012;271:421–428; Science 2011;334:255–258). Another component is the tight junctional complex located above the adherent complex between the epithelial cells. Two physiologically distinct pathways are recognized through the tight junction. The pore pathway has claudins as the major molecular component and allows passage of small ions and uncharged molecules. The current study suggests that there is an increase in the permeability of this pathway that occurs in cirrhosis as claudin 2 expression is increased. In contrast, the leak pathway has occludin as its major membrane protein and allows passage of large molecules, including proteins (Annu Rev Physiol 2011;73:283–309). Further investigation into the contribution of the leak pathway in cirrhosis is required, because LPS is thought to be too large to traverse the pore pathway. Epithelial cells are constantly being shed from villi of the small intestine or the surface of the colon into the lumen. Barrier function is maintained at sites of cell shedding by a redistribution of tight junction proteins to plug the gap left by the shedding cell (Gastroenterology 2011;140:1208–1218). LPS increases the rate of cell shedding (Gastroenterology 2013;144:S668). When shedding rates are high sometimes the redistribution of tight junction proteins fails to seal the epithelium, particularly if ≥2 adjacent cells are shed simultaneously (Gastroenterology 2007;133:1769–1778; Gut 2012;61:1146–1153). Thus, it is plausible that failure of the intestinal barrier as a result of increased cell shedding might be the mechanism of increased systemic LPS in cirrhotic patients. It is evident that the gut–liver axis is an integral pathway in the pathogenesis of liver disease and this study further highlights some of the potential mechanisms, which result in increased levels of proinflammatory cytokines, increased gut permeability, bacterial translocation, and subsequent liver disease. The authors have described a phenomenon whereby intestinal macrophages are activated and mucosal integrity compromised in this patient group. It is, however, not clear why this occurs. It may result from hemodynamic changes in intestinal blood flow as a result of changes in portal pressure, or indeed may be related to enterohepatic sharing of lymphocytes (Gastroenterology 2009;137:320–329; Lancet 2002;359:150–157). Further studies are required to explore treating cirrhotic patients by altering gut flora by antibiotic therapy, recolonization with therapeutically defined microbial populations, or even fecal transplantation from healthy donors. This study provides further evidence to support the use of ciprofloxacin and rifaximin in clinical practice for the prevention of spontaneous bacterial peritonitis and hepatic encephalopathy. It suggests a mechanism via which these antibiotics may be efficacious in prophylaxis, and thus supports current clinical practice.