Association between endotoxemia and histological features of nonalcoholic fatty liver disease

AIM To assess whether surrogate biomarkers of endotoxemia were correlated with the histological features of nonalcoholic fatty liver disease (NAFLD). METHODS One hundred twenty-six NAFLD patients who had undergone percutaneous liver biopsy were enrolled. Serum lipopolysaccharide (LPS)-binding protein (LBP) and anti-endotoxin core immunoglobulin G (EndoCab IgG) antibody concentrations at the time of liver biopsy were measured using the enzyme-linked immunosorbent assays to examine for relationships between biomarker levels and histological scores. RESULTS Serum LBP concentration was significantly increased in nonalcoholic steatohepatitis (NASH) patients as compared with nonalcoholic fatty liver (NAFL) subjects and was correlated with steatosis (r = 0.38, P < 0.0001) and ballooning scores (r = 0.23, P = 0.01), but not with the severity of lobular inflammation or fibrosis. Multivariate linear regression analysis revealed that LBP was associated with steatosis score and circulating C-reactive protein, aspartate aminotransferase, and fibrinogen levels. Serum EndoCab IgG concentration was comparable between NASH and NAFL patients. No meaningful correlations were detected between EndoCab IgG and histological findings. CONCLUSION LBP/EndoCab IgG were not correlated with lobular inflammation or fibrosis. More accurate LPS biomarkers are required to stringently assess the contribution of endotoxemia to conventional NASH.


INTRODUCTION
The prevalence of nonalcoholic fatty liver disease (NAFLD) is increasing worldwide. NAFLD includes a wide spectrum of disorders, ranging from nonalcoholic fatty liver (NAFL) to nonalcoholic steatohepatitis (NASH) and resultant liver cirrhosis and hepatocellular carcinoma [1][2][3][4] . NASH is characterized by the presence of hepatocyte ballooning, lobular inflammation and/or various degree of fibrosis in addition to macrovesicular steatosis [1][2][3] . Although several pathogenic factors, such as lipotoxicity, endoplasmic reticulum stress, iron accumulation, and inflammatory signaling, reportedly contribute to the progression from steatosis to steatohepatitis/steatofibrosis, the mechanism of NASH development has not been fully clarified. Recent murine studies have demonstrated a key role of endotoxin/lipopolysaccharide (LPS) in the onset of NASH [5,6] . For example, repeated LPS injection into ob/ob mice led to steatohepatitis [7] , while mice lacking the gene encoding Toll-like receptor (TLR) 4, a central molecule in LPS-mediated signaling, were resistant to NASH development [8] . It is generally accepted that Kupffer cells are activated when the gut mucosa becomes inflamed and fragile or when gut bacteria overgrow and LPS subsequently flows into the portal vein. However, it remains unclear whether the gut barrier is disrupted and portal LPS levels are elevated in typical obesity-, metabolic syndrome-related NAFLD patients without accompanying active inflammatory bowel disease [9,10] or a history of gastrointestinal surgery, such as pancreaticoduodenectomy and blind loop construction [11,12] . Another issue requiring attention when considering the contribution of endotoxin/LPS to human NASH is the lack of appropriate systems to determine portal LPS concentration. Since portal LPS is rapidly eliminated in the liver, systemic LPS levels often do not mirror those in the portal vein, and the halflife of circulating LPS is as short as 2 h [13] . These shortcomings obscure the evaluation of endotoxin/LPS contribution to NASH. Circulating LPS-binding protein (LBP) and antiendotoxin core immunoglobulin G (EndoCab IgG) antibody are commercially available surrogate markers of endotoxemia [10,14] . LBP is a soluble acutephase protein that binds to bacterial endotoxin. When the liver senses bacterial endotoxin, LBP is rapidly synthesized by hepatocytes to neutralize the toxin and then secreted into the circulation. The EndoCab IgG assay measures endotoxin core antibodies to reflect the immune response against persistent endotoxin exposure. Since both of these biomarkers are more stable than LPS, they represent possible indicators of endotoxemia.
In the current study, serum levels of LBP and EndoCab IgG were measured in biopsy-proven NAFLD patients and their correlation with the histological severity of NAFLD was assessed to clarify associations between endotoxemia and NASH development.

Patients
This study was approved by the Committee for Medical Ethics of Shinshu University School of Medicine (Approval number: 2802) and conducted in accordance with the 1983 revision of the Helsinki declaration of 1975. Informed written consent was obtained from all patients. One hundred twenty-six NAFLD patients who were admitted to Shinshu University Hospital between February 2009 and April 2015 for percutaneous liver biopsy were enrolled. NAFLD had been suspected based on the following criteria: (1) the presence of hepatorenal contrast and increased hepatic echogenicity on abdominal ultrasonography; (2) ethanol consumption of < 20 g/d; and (3) the absence of other causes of liver dysfunction, such as viral hepatitis, druginduced liver injury, autoimmune liver diseases, primary sclerosing cholangitis, Wilson's disease, hereditary hemochromatosis, and citrin deficiency [15][16][17][18] . The diagnosis of NAFLD/NASH was confirmed based on the histological findings of biopsied specimens. Body weight and height were measured before liver biopsy performed in a fasting state. The presence of obesity was defined as a body mass index of ≥ 25 kg/m 2 according to criteria released by the Japan Society for the Study of Obesity [19] . Medical information was also recorded, and the presence of hypertension, hyperlipidemia, and diabetes was evaluated as described previously [20,21] . Blood samples were obtained on the day of the liver biopsy in a fasting state and routine examinations, such as complete blood counts, coagulation, and blood chemistry that included serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT), were carried out using standard laboratory methods. The remaining sera samples were immediately frozen and kept at -80 ℃ until further use.

Histopathological analysis
Biopsy specimens were obtained from liver segment 5 or 8 using a 14-gauge needle as described previously and immediately fixed in 10% neutral formalin.

Measurement of serum LBP and EndoCab IgG levels
Frozen serum samples obtained at the time of liver biopsy were diluted in 1000-and 200-fold with dilution buffer, and serum LBP and EndoCab IgG concentrations were measured in duplicates using the LBP ELISA kit (HK315-01, Hycult Biotech, Uden, the Netherlands) and the ENDOCAB ELISA kit (HK504, Hycult Biotech), respectively, according to the manufacturer's instructions.

Statistical analysis
Data are expressed as number (percentage) or median (25 th , 75 th percentiles). Comparisons between the groups were carried out using the Mann-Whitney U test, χ 2 test, or one-way ANOVA with Bonferroni's correction. Spearman's test was adopted to examine for correlations among LBP, EndoCab IgG, and biochemical/ histological data. Multivariate linear regression analysis was conducted to search for independent predictors of LBP. Statistical analyses were performed using StatFlex Ver6.0 software (Artech Co., Ltd., Osaka, Japan). A P value of < 0.05 was considered to be statistically significant.

Serum LBP/EndoCab IgG levels in NAFLD patients
The clinicopathological features of the 126 NAFLD patients are summarized in Table 1. One-hundred patients (79%) were diagnosed as having NASH and 29 (23%) had advanced fibrosis of stage 3 or 4. Serum LBP concentration was significantly increased in NASH patients as compared with NAFL patients ( Figure  1A), whereas serum EndoCab IgG was comparable between the groups ( Figure 1B).

Relationships between LBP/EndoCab IgG levels and histological findings
We next examined whether serum LBP levels correlated with histological severity in NAFLD patients. Serum LBP was significantly higher in the steatosis grade 3 group as compared with the steatosis grade 1 or 2 groups ( Figure 2A) and was positively correlated with steatosis score (Table 2). Significantly increased serum LBP was seen in the ballooning grade 2 group over the ballooning grade 0 group ( Figure 2B), with a positive correlation between LBP and ballooning score (Table 2). There were no significant relationships   Table 2) or fibrosis stage ( Figure 2D and Table 2).
Similar analyses were carried out for serum EndoCab IgG, which yielded comparable findings among all subgroups ( Figure 3) and no correlations with histological scores ( Table 2). The lack of agreement between LBP and EndoCab IgG (Figure 4) may explain the discrepant results between LBP and EndoCab IgG for histological scores.

Correlations between LBP/EndoCab levels and NAS
There was a significant positive correlation for serum LBP and NAS, an indicator of histological NAFLD activity, but none for EndoCab IgG ( Figure 5).

Figure 4 Correlation between serum lipopolysaccharide-binding protein and EndoCab IgG in 126 biopsy-proven nonalcoholic fatty liver disease patients.
A correlation efficient (r) and P value were calculated using Spearman's test. LBP: Lipopolysaccharide-binding protein. Histological findings were scored according to the criteria proposed by Kleiner et al [22] Correlation coefficients (r) and P values were calculated by Spearman's test. LBP: Lipopolysaccharide-binding protein.  (Table 3). Factor analysis using several laboratory data parameters distinguished a clear separation between the LBP and EndoCab IgG groups; the former group included inflammationrelated parameters, such as CRP, FIBG, and leukocyte count, whereas the latter group contained HbA1c, age, and hyaluronic acid. Multivariate linear regression analysis revealed that steatosis score (P = 0.0033), circulating CRP (P = 0.0032), AST (P = 0.0151), and FIBG (P = 0.0181) were independent predictors of LBP.

DISCUSSION
To our knowledge, this is the first study examining the relationship between two surrogate LPS markers and histological severity in NAFLD. Unexpectedly, serum EndoCab IgG did not correlate with any histological findings, nor was serum LBP associated with lobular inflammation grade or fibrosis stage. Based on these results, we could not conclude whether LPS played a crucial role in hepatitis development or fibrosis progression in NAFLD patients, which deviated from findings obtained in murine models.
Although serum LPS is reportedly elevated in NASH patients [23] , the direct measurement of LPS has several flaws. First, LPS has a very short half-life and is rapidly eliminated in the liver, and thus serum LPS values rarely reflect actual endotoxemia. Second, the limulus amoebocyte lysate assay is widely used for LPS determination but may be influenced by exogenous LPS contamination due to its high sensitivity. The assay is also disrupted by detergents, urea, and pH [24] .
We therefore adopted the more stable biomarkers LBP and EndoCab IgG as indicators of endotoxemia.
Since LBP is rapidly induced by LPS and EndoCab IgG reflects the immune response to endotoxin core, these biomarkers measure acute/intermittent and chronic/persistent endotoxemia, respectively. Previous studies have examined either serum LBP or EndoCab IgG in NAFLD patients. LBP was similar [25] between control and NAFLD subjects and increased in NAFLD/NASH patients with severe fibrosis [26,27] .
EndoCab IgG was comparable [28] between control and NAFLD/NASH groups and higher in NASH [29] . However, the number of NAFLD patients was relatively small (less than 40) in these investigations. Moreover, there have been no reports simultaneously measuring LBP and EndoCab IgG in histologically-proven NAFLD patients or examining these factors for associations with histological findings. The absence of a correlation between LBP and EndoCab IgG has been supported by a large recent study of 920 participants [14] . Although LBP and EndoCab IgG were simultaneously measured in the previous study [14] , histological evaluation of the liver was not performed.
In the present series, the severity of lobular inflammation and fibrosis did not correlate with either serum LBP or EndoCab IgG. EndoCab IgG tended to associate with age, hyaluronic acid, and IgG, which was indicative of the possibility of nonspecific IgG elevation due to chronic liver disease and/or aging. Thus, evidence that LPS promoted fibrosis and hepatitis in the context of conventional human NAFLD/ NASH could not be demonstrated.
The degree of steatosis correlated with serum LBP level. An earlier study also demonstrated a significant correlation for serum LBP and intrahepatic triglyceride content as determined by proton-magnetic resonance spectroscopy (r = 0.366, P < 0.001) [14] , but not for EndoCab IgG. These similar findings prompted us to consider that LBP might be induced independently of LPS in NAFLD patients. Indeed, the fact that disruption of the LBP-encoding gene in mice decreased basal expression levels of fatty acid-synthesizing enzymes and suppressed steatogenesis [30] implied a direct link between increased LBP and hepatosteatosis. LBP is also up-regulated in hypertrophied adipocytes and acts as an adipokine [31] . Along with CRP and

A B
FIBG, LBP is an acute phase reactant as well. It is noteworthy that LBP can be induced by interleukin-6 (IL-6) [32] , a pro-inflammatory cytokine, in hepatocytes. IL-6 is markedly up-regulated in steatotic livers [33] , indicating a possible link among steatosis, IL-6, and LBP. Although the possibility that LPS directly induced steatosis cannot be ruled out completely, the above findings corroborate a possible association between serum LBP and steatosis independently of LPS/ endotoxemia.
This study uncovered a weak, but statistically significant, correlation between serum LBP level and the incidence of ballooned hepatocytes. In hepatocytes with ballooning degeneration, activated c-Jun N-terminal kinase (JNK) and ensuing lipoapoptosis have been documented [34] . Escherichia coli LPS induces LBP in human oral keratinocytes through the activation of JNK in addition to nuclear factor kappa B and p38 mitogen-activated protein kinase [35] . Therefore, a positive relationship between LBP expression and ballooning score might reflect activated JNK-mediated signaling in degenerated hepatocytes. Yuan et al [36] demonstrated that circulating LPS levels in pediatric NASH patients were distributed dichotomously at either high or normal levels, suggesting that endotoxemia was present in specific NASH patients only. Serum LPS levels exhibited no impact on NAS or fibrosis stage and no meaningful relationship was detected between LPS and the proportion of intestinal Gram-negative bacteria. Therefore, they concluded that gut microbiome composition did not contribute to endotoxemia in NASH, nor was endotoxemia always required in the pathogenesis of NASH, which were partially consistent with our results. The observation that antibiotics/probiotics can attenuate NAFLD may not directly support a key role of LPS in NASH development since intestinederived metabolites/toxicants other than LPS, such as deoxycholic acid and ceramides, promote NAFLD/ NASH development [37,38] . The therapeutic use of probiotics/prebiotics has not been supported by highquality clinical studies [39] . Additionally, the notion that increased hepatic TLR4 expression in NASH indicates an important role of portal endotoxemia may be inappropriate since TLRs are activated by several molecules other than LPS, including palmitic acid [40] .
A key limitation of this study was that it did not directly measure LPS concentrations. Since the present investigation used cryogenically stored samples, we were concerned about the accuracy of LPS value measurement due to the abovementioned flaws in the LPS assay system. Future studies will benefit from assessment of LPS using freshly prepared serum samples along with improvements in LPS assay systems. Another limitation was that we could not examine the association between LBP and PNPLA3 polymorphisms, which might impact the degree of hepatic steatosis. However, a recent study showed no relationship between PNPLA3 variants and circulating LBP levels in chronic hepatitis C patients [41] .
It is reasonable to consider that chronic LPS challenge to steatotic livers, such as NAFLD accompanied with severe gingivitis or inflammatory bowel disease, may be detrimental to liver condition. However, establishing a key contribution of LPS to the histological

ACKNOWLEDGMENTS
The authors thank Mr. Trevor Ralph for his editorial assistance.

Background
Recent murine studies have demonstrated a key role of endotoxin/ lipopolysaccharide (LPS) in the onset of nonalcoholic steatohepatitis (NASH). The contribution of intestinal bacterial overgrowth, increased intestinal permeability, and portal endotoxemia to the progression from nonalcoholic fatty liver (NAFL) to NASH has attracted considerable recent attention. However, evaluating the LPS contribution to human NASH is challenging since LPS is rapidly eliminated in the liver, and therefore venous LPS concentrations often do not reflect portal ones.

Research frontiers
Circulating LPS-binding protein (LBP) and the anti-endotoxin core immunoglobulin G (EndoCab IgG) antibody are commercially available surrogate markers of endotoxemia that are more stable than LPS. The research hotspot is to examine whether these endotoxemia markers correlate with histological severity in nonalcoholic fatty liver disease (NAFLD).

Innovations and breakthroughs
This is the first study simultaneously measuring two surrogate endotoxemia markers, LBP and EndoCab IgG, in biopsy-proven NAFLD patients in order to assess for relationships with the histological scores of NAFLD. Serum LBP concentration was significantly increased in NASH patients and was correlated with steatosis and ballooning scores, but not with the severity of lobular inflammation or fibrosis. Serum EndoCab IgG concentration was comparable between NASH and NAFL patients.

Applications
It is reasonable to consider that chronic LPS challenge to steatotic livers, such as NAFLD accompanied with severe gingivitis or inflammatory bowel disease, may be detrimental to liver condition. However, the contribution of LPS to the histological severity of human primary NASH could not be confirmed in the current study. The lack of appropriate endotoxin markers is a major limitation at present, as is the low sensitivity/specificity of LBP/EndoCab IgG assays. Further improvements in LPS detection systems may provide novel information on the role of LPS/endotoxemia in the pathogenesis of conventional NASH.

Terminology
NAFLD is a chronic liver disease increasing worldwide that includes a wide spectrum of disorders, ranging from NAFL to NASH and resultant liver cirrhosis and hepatocellular carcinoma. NASH is characterized by the presence of hepatocyte ballooning, lobular inflammation and/or various degrees of fibrosis in addition to macrovesicular steatosis. Increased intestinal permeability and ensuing portal endotoxemia is presumed to be one of the contributors to the progression of NASH.

Peer-review
This report, written by Kitabatake et al, is of an important retrospective study that has been carried out to disclose pathological mechanisms of NAFLD/NASH as well as its diagnosis. Especially, the close relationship between LBP and NASH is very interesting.