Regulation of Gut Microbiota and Metabolic Endotoxemia with Dietary Factors

Metabolic endotoxemia is a condition in which blood lipopolysaccharide (LPS) levels are elevated, regardless of the presence of obvious infection. It has been suggested to lead to chronic inflammation-related diseases such as obesity, type 2 diabetes mellitus, non-alcoholic fatty liver disease (NAFLD), pancreatitis, amyotrophic lateral sclerosis, and Alzheimer’s disease. In addition, it has attracted attention as a target for the prevention and treatment of these chronic diseases. As metabolic endotoxemia was first reported in mice that were fed a high-fat diet, research regarding its relationship with diets has been actively conducted in humans and animals. In this review, we summarize the relationship between fat intake and induction of metabolic endotoxemia, focusing on gut dysbiosis and the influx, kinetics, and metabolism of LPS. We also summarize the recent findings about dietary factors that attenuate metabolic endotoxemia, focusing on the regulation of gut microbiota. We hope that in the future, control of metabolic endotoxemia using dietary factors will help maintain human health.


Introduction
Lipopolysaccharide (LPS) is a component of the outer membrane of gram-negative bacteria and is known to induce a variety of inflammatory reactions through Toll-like receptor 4 (TLR4). Injection of LPS into human blood elicits an inflammatory response [1,2], but it was thought that LPS is rarely detected in human blood, except under pathological conditions such as infection and colitis. However, in 2007, Cani et al. showed that mice fed with a high-fat diet had higher blood LPS levels than normal chow-fed mice, resulting in inflammation of the liver and adipose tissue, which led to the development of NAFLD and insulin resistance, and the authors defined this condition as metabolic endotoxemia [3]. Since then, studies on metabolic endotoxemia have been conducted for a variety of diseases. It has been reported that blood LPS levels are higher in humans with obesity [4], type 2 diabetes [5], NAFLD [6], pancreatitis [7], amyotrophic lateral sclerosis [8], and Alzheimer's disease [8] than those in healthy individuals. Although the causal relationship between metabolic endotoxemia and disease onset is unclear, it is expected to be an interesting target in the future from the viewpoint of disease prevention and treatment. In recent years, the association between metabolic endotoxemia and dietary factors, and the mechanism by which fat intake induces metabolic endotoxemia have been actively studied. In contrast, dietary factors that suppress metabolic endotoxemia have also been explored. Here, we review the relationship between fat intake and induction of metabolic endotoxemia, focusing on gut dysbiosis

Kinetics and Activity of LPS
The LPS concentration in the portal blood is approximately 10 times higher than the LPS concentration in the peripheral blood [34], suggesting that a part of the LPS released in the intestinal tract is flowing from the portal vein. On the other hand, LPS which is concomitantly absorbed with lipids binds to lipoproteins in chylomicrons via LPS-binding protein (LBP) [35], and is thought to pass through the lymphatic system, flow into the blood stream from the left subclavian vein, and then circulate throughout the body. It is reported that blood LPS is bound to various lipoproteins, with plasma LPS concentrations of 31%, 30%, 29%, and 10% for the very low-density lipoprotein (VLDL) fraction, low-density lipoprotein (LDL) fraction, high-density lipoprotein (HDL) fraction, and free LPS, respectively [36]. In addition, LPS bound to lipoproteins of HDL has been reported to be transferred to VLDL and LDL by LBP and phospholipid transfer protein [37], suggesting that the LPS concentration of each lipoprotein fraction changes actively. There are several reports that bioactivity of LPS bound to lipoprotein varies with the type of lipoprotein. First, Vreugdenhil et al. evaluated the effect of chylomicrons, HDL, LDL, and VLDL on the production of tumor necrosis factor-α (TNF-α) from human peripheral blood mononuclear cells on LPS stimulation and showed that chylomicrons inhibited TNF-α production the most [35]. Emansipator et al. reported that a mix of LPS with LDL or HDL decreased the spike recovery of LPS activity in the limulus amebocyte lysate test, and that incubation of LPS with apo A1 decreased the febrile response of rabbits when injected compared to those without apo A1 [38]. In a study using human mononuclear cells [39] and the mouse macrophage cell line Raw 264.7 [40], it was reported that LPS bound to HDL showed reduced interleukin-6 (IL-6) and TNF-α production. VLDL has also been reported to inhibit LPS-induced activation of nuclear factor κB (NF-κB) [41]. On the other hand, oxidized LDL has been shown to promote NFκB activation with LPS in macrophages [42], suggesting that binding to lipoproteins not only decreases LPS activity but also may promote inflammatory responses.
Increased LPS content has been reported in the livers of mice fed a high-fat diet [43], suggesting that the liver is an important site for LPS clearance. Ninety percent of the free LPS that entered the bloodstream is captured by liver resident macrophages (i.e., Kupffer cells) within 1 h [44]. LPS bound to HDL attaches primarily to sinusoidal epithelial cells of the liver [40,44], but it shows slower blood kinetics than free LPS, with 50% present in plasma even 1 h after administration and the amount accumulated in the liver accounted for only 15% of the dose [44]. LPS bound to HDL on the other hand is distributed widely to organs other than the liver, such as the kidney and adipose tissue [44]. LPS accumulated in the liver is inactivated by acyloxyacyl hydroxylase produced by Kupffer cells regardless of free or HDL-bound form [44]. Previously, in a mouse model of high-fat diet plus streptozotocin-induced non-alcoholic steatohepatitis-hepatocellular carcinoma, fecal LPS levels were continuously elevated from six weeks, while liver LPS levels were transiently elevated at eight weeks, followed by increased plasma LPS levels [45]. This report suggests that the liver acts as the first barrier against LPS entering from the intestinal tract and that liver dysfunction leads to elevated blood LPS levels. Interestingly, LPS administration in mice increased the expression of apolipoprotein AIV in the liver via TLR4, suggesting that the liver has a mechanism to increase HDL production and protect itself against LPS stimulation [46].

Dietary Factors that Decrease Blood LPS Levels
Previous reports investigating the effects of dietary factors on blood LPS levels are summarized in Table 1 (human interventional studies), Table 2 (human epidemiological studies) and Table 3 (animal studies). The findings about representative food categories are reviewed in the following sections.

Probiotics
Probiotics were defined by Fuller in 1989 as "A live microbial feed supplement which beneficially affects the host animal by improving its intestinal microbial balance" [47], and has been studied mainly for lactic acid bacteria and bifidobacteria. Since metabolic endotoxemia has been implicated in gut dysbiosis, the effects of probiotics have been investigated. However, the results in humans are unfavorable (Table 1). Lever et al. administered 195 mL of Yakult light (containing 2 × 10 10 colony-forming unit (CFU) of Lactobacillus casei Shirota) for three months to individuals with metabolic syndrome. The absence of detectable blood LPS in this study led to an assessment of the surrogate LBP level, which was significantly higher in the Yakult light-fed group than in the non-fed group [48]. Pei et al. conducted a nine-week study in which low-fat yogurt was ingested in healthy or obese individuals, however no significant decrease in blood LPS or LBP levels was observed [49]. In addition, Pei et al. studied whether low-fat yogurt could be administered before a meal to suppress the increase in blood LPS after a meal [50] and found no efficacy. On the other hand, there have been several reports of the efficacy of probiotics in animal studies (Table 3) [43,[51][52][53][54][55]. Lactobacillus rhamnosus, Lactobacillus sakei, Lactobacillus acidophilus, Lactobacillus plantarum, Bifidobacterium longum, Bifidobacterium infantis, and Bacillus cereus are used as species, and the dosage ranges from 10 7 to 10 10 CFU/day for four to twelve weeks. These animal studies used a high-fat diet, a high-fat high-sucrose diet, or a Zucker-Lep fa/fa obesity model. In addition to a significant decrease in blood LPS or LBP levels, improvement of obesity, glucose metabolism, and dyslipidemia was also observed. Since the effects of probiotics are strain-specific, it is expected that the effects of strains that have been effective in animal studies will be verified in humans.

Prebiotics
Prebiotics was defined by Gibson and Roberfroid in 1995 as "nondigestible food ingredients that beneficially affect the host by selectively stimulating the growth and/or activity of one or a limited number of bacterial species already resident in the colon, and thus attempt to improve host health" [56], and among the food components, dietary fiber and oligosaccharides are known as typical prebiotics. To date, human intervention studies have been conducted with oligofructose [57], inulin [58,59], galacto-oligosaccharides [60][61][62], resistant dextrin [63], insoluble dietary fiber [64], and whole grains (Table 1) [65]. Oligofructose is an oligosaccharide containing one molecule of glucose and several molecules of fructose and is found in many fruits and vegetables. Inulin is a type of fructose-polymerized polysaccharide that is abundant in vegetables such as burdock and onion. In intervention studies with oligofructose [57] and inulin [58,59], subjects with obesity, overweight subjects, and subjects with type 2 diabetes consumed 10-21 g of test substances for 8-12 weeks. Two of the three studies showed a significant decrease in blood LPS levels [57,58]. One study also showed a decrease in plasminogen activator inhibitor-1 (PAI-1), a risk indicator for thrombosis [57], and the other study showed an improvement in glucose metabolism [58]. Galacto-oligosaccharides are oligosaccharides in which multiple molecules of galactose are attached to one molecule of glucose. Similar to oligofructose, there have been three reports of interventional trials for galacto-oligosaccharide in obese, overweight, and type 2 diabetic patients. One study showed that galacto-oligosaccharides reduced blood LPS levels, and improved obesity by suppressing appetite [62]. In mice, chronic administration of LPS has been reported to induce hyperphagia by decreasing leptin sensitivity of afferent vagal nerves [66], and the reduced blood LPS levels and appetite suppression seen with galacto-oligosaccharide administration are of interest in supporting an association between LPS and appetite.

Polyphenols
Polyphenols are secondary metabolites found in plants and are responsible for protection against oxidative stress, UV damage, and pathogenic microorganisms [67]. Polyphenols are found in a wide range of foods, including vegetables, fruits, tea, beans, and spices, and their consumption has been reported to improve metabolic syndrome (decreased body weight, decreased blood pressure, improved glucose metabolism, and improved lipid metabolism) [68]. However, up to 27% of ingested polyphenols are detected in urine [69], suggesting that many of them are not absorbed and reach the large intestine [70]. Since polyphenols reaching the large intestine have been reported to alter the proportions of microbiota [71], it is expected that the effect of polyphenols against metabolic syndrome is mediated through the improvement of dysbiosis and of the accompanying metabolic endotoxemia. There are two human intervention studies investigating the relationship between polyphenol intake and blood LPS, both of which evaluated the inhibitory effect on postprandial elevation of blood LPS levels (Table 1) [72,73]. In the study performed by Ghanim et al., healthy individuals ingested capsules containing 100 mg of resveratrol and 75 mg of polyphenol 10 min before a 930-kcal high-fat, high-carbohydrate meal. Blood LBP levels up to 5 h after a meal were evaluated and showed increased blood LBP levels in the placebo group but not in the capsule group [72]. On the other hand, Clemente-Postigo et al. administered 272 mL of red wine to humans simultaneously with excessive fat and found no effect on either blood LPS or LBP levels [73]. The efficacy of polyphenols has been also reported in animal studies. The effects of grape seed proanthocyanin [29,33], resveratrol [74], apple-derived polymeric procyanidins [75], genistein [76], isoflavone [77], and syringarecinol [78] on blood LPS levels in animal models have been reported (Table 3). In particular, L'openz et al. reported that six-month administration of genistein to high-fat diet-fed mice reduced their blood LPS levels and improved their spatial memory ability [76]. Cho et al. administered syringalesinol to 40-week-old mice for 10 weeks and showed that the decrease in blood LBP levels was accompanied with suppression of changes in immune cells due to aging (decreased naive T cells and decreased T-cell proliferation) [78]. It has also been reported that adoption of a high-fat diet results in abnormal differentiation of bone marrow hematopoietic stem cells due to increased blood LPS levels [79], suggesting that the effect of syringalecinol on immunoaging might be also exerted in other models of metabolic endotoxemia.

Sulfated Polysaccharide
Sulfated polysaccharides are widely present in animal tissues and seaweed and are used industrially as anticoagulants, pharmaceuticals, and gelling agents for foods. The effect of sulfated polysaccharides on metabolic endotoxemia has been studied only in animals (Table 3). Intervention studies with sea cucumber-derived sulfated polysaccharides [80,81], acaudina molpadioides-derived fucosylated chondroitin sulfate [82], chicken-derived chondroitin sulfate [83] or fucoidan [84] have been performed. Of these studies, two showed that administration of sulfated polysaccharides to high-fat diet-fed mice increased the amount of short-chain fatty acids in the intestinal tract, decreased the blood LPS or LBP concentration and attenuated weight gain [80,82]. Zhu et al. also reported the same effect of sulfated polysaccharides in chow-fed lean mice [81]. Liu et al. demonstrated that exhaustive exercise with a treadmill significantly impaired kidney function, decreased fecal butyrate levels, changed intestinal morphology, and induced metabolic endotoxemia [83]. Their study is interesting in showing that exercise stress also increased blood LPS levels, and that dietary factors are also effective in the model mice.

Other Dietary Components/Extracts/Foods
In the study by Abboud et al., obese or over weight subjects ingested 30 g of glutamine per day for eight weeks (Table 1) [85]. As a result, their blood LPS levels and waist circumference decreased. In an epidemiological study conducted with healthy subjects, 25-hydroxy vitamin D was reported to negatively correlate to blood LPS levels ( Table 2) [86]. The protective effect of vitamin D is supported by animal studies in which vitamin D-deficient mice, exposed to a bacterial pathogen, exhibited lower LPS detoxification activity of the intestine and greater endotoxin translocation [87]. The effect of other dietary components, including tetrahydro iso-alpha acid [88], rhein [89], phlorizin [90], capsaicin [91], rutin [92], and lycopene [93] on blood LPS levels in animals has also been reported (Table 3). Among them, administration of tetrahydro iso-alpha acid [88], phlorizin [90], or rutin [92] to high-fat diet-fed mice or db/db mice improved metabolic impairment. Administration of rhein [89], or lycopene [93] to high-fat diet-fed mice showed a unique effect; they not only reduced blood LPS levels but also prevented high-fat diet-induced memory impairment. Kang et al. showed that administration of antibiotics to mice given capsaicin abolished the effect of capsaicin on blood LPS levels [91]. They also showed that capsaicin-induced protection against high-fat diet-induced blood LPS increase is transferrable by fecal microbiota transplantation.
It has also been reported that intervention with crude food extracts or the food itself can lower blood LPS levels in animals (Table 3). We studied the effect of broccoli sprout extract, enriched in functional glucosinolate "glucoraphanin" (details are described in Section 4) [94]. Anhê et al. examined the effects of extracts from cranberry [95] or camu camu [96]. Camu camu is an Amazonian fruit that contains an abundance of vitamin C and flavonoids such as ellagic acid, ellagitannins, and proanthocyanidins. Administration of camu camu extracts to high-fat/high-sucrose diet-fed mice reduced plasma bile acid pool size, altered gut microbiota composition, and reduced blood LPS levels. Dey et al. reported that administration of green tea extract to high-fat diet-fed mice suppressed inflammation and gut permeability especially in the ileum and colon, and reduced LPS influx from the portal vein [34]. The reduction of blood LPS levels by feeding with Tartary buckwheat protein was reported by Zhou et al. [32]. This study is valuable in that it elucidates one of the underlying mechanisms by which plant protein intake leads to improvement of metabolic abnormalities. Intervention studies with cocoa [97], nopal [98], and steamed fish meat [99] have been performed. Among these, Zhang et al. performed unique experiments [99]. They divided mice into four groups, and fed them ad libitum with normal chow, steamed fish, pork or beef at 9:00 and 18:00 daily for eight weeks. As a result, only mice group fed with steamed fish showed decreased blood LBP levels compared to the other three groups.

Chinese Medicines
The effect of the Chinese medicines; geniposide + chlorogenic acid [100], potentilla discolor bunge water extract [101], ganoderma lucidum mycelium water extract [102], semen hoveniae extract [103], and shenling baizhu powder [104] on blood LPS levels have been reported in animals ( Table 3). The combination of geniposide and chlorogenic acid is included in a traditional Chinese medicine, Qushi Huayu Decoction. Peng et al. indicated that administration of geniposide and chlorogenic acid to high-fat diet-fed mice restored colonic tight junctions by inhibiting down-regulation of RhoA/Rho-associated kinase signaling, and reduced blood LPS levels and hepatic LBP protein levels [100]. Han et al. examined the effect of potentilla discolor bunge water extract in type 2 diabetic mice induced by high-fat diet feeding and streptozotocin injection [101]. The results showed that fecal LPS levels in the type 2 diabetic model mice were significantly increased compared to the control normal mice. The administration of potentilla discolor bunge water extract to mice reduced fecal LPS levels, decreased blood LPS levels and increased the expression levels of tight junction proteins (Claudin-3, ZO-1, and Occludin) in the colon. Chang et al. studied the effect of ganoderma lucidum mycelium, a Basidiomycete fungus [102]. They showed the dose-dependent effect of ganoderma lucidum mycelium water extract on blood LPS reduction, suggesting that high molecular weight polysaccharides (>300 kDa) isolated from the extract is an effective component. Ping et al. reported that the extract of semen hoveniae, a seed of Hovenia dulcis Thunb rich in dihydromyricetin and quercetin, decreased blood LPS levels in a mouse model of alcohol-induced liver injury [103]. It has been reported that administration of shenling baizhu, a mixture of ten different traditional Chinese medicinal herbs, to high-fat diet-fed mice decreased LPS levels in the portal vein [104].

Dietary Habits
In relation to dietary habits, Kopf et al. conducted an intervention study in humans with BMI > 25 kg/m 2 and low intake of whole grains, fruits, and vegetables (Table 1) [65]. During the weekly interview, the subjects themselves selected the vegetables and fruits to be eaten the following week from apples, bananas, blueberries, clementines, grapes, pears, strawberries, broccoli, carrots, cauliflower, celery, green beans, green leaf lettuce, peas, spinach, sweet pepper, and tomatoes. The subjects ate these fruits and vegetables for 21 to 30 servings/week (at least three servings/day) for six weeks. As a result, compared to control group in that dietary habits were not or minimal changed, average daily intake of refined grains was 1/3, fruit intake was doubled, and vegetable intake was four times, leading to a significant reduction in blood LBP levels and IL-6 levels. An epidemiological study by Ahola et al. in patients with type 1 diabetes has shown a negative correlation between several dietary patterns and blood LPS levels: These dietary patterns are "Fish" (frequently eats fish dishes), "Healthy snack" (frequently eats fruits, berries, fresh vegetables, yoghurt, low-fat cheese, and does not drink many soft drinks) and "Modern" (frequently eats poultry, pasta, rice, meat dishes, fried and grilled foods, and fresh vegetables) ( Table 2) [105]. In the epidemiological study by Ahola et al., no significant correlation was found between blood LPS levels and intake of energy, carbohydrates, fats, proteins, or dietary fiber. In regard to the absence of a significant positive correlation between blood LPS levels and fat intake (the believed cause of blood LPS elevation in humans and animals), the authors consider that the previously reported amount or proportion of fat intake may be greater than the intake in the normal diet. Similarly, Amar et al. reported no significant correlation between fat intake and blood LPS levels in 201 subjects [106]. In the same study, Amar et al. reported a positive correlation between total energy intake and blood LPS levels [106]. The effect of caloric restriction on blood LPS levels have been reported in both humans and mice. Ott et al. reported that, in women with a BMI of 30 kg/m 2 or more, intake of a defined formula diet of 800 kcal/day for four weeks decreased blood LBP levels, and following intake of the normal diet (1800 kcal/day), blood LBP levels returned to the initial levels ( Table 1) [107]. Even in mice, caloric restriction of 30% [108] or 40% [109] has been reported to decrease blood LPS or LBP levels (Table 3). A common finding in these reports in mice is that blood LPS or LBP levels are reduced by calorie restriction compared to ad libitum even in normal chow-fed mice. This suggests that the influx of LPS into the bloodstream is not limited to the specific conditions of excessive fat intake but can also occur by some mechanism in the normal diet.

Association of Dietary Factor-Induced Reduction of Blood LPS and Modulation of Gut Microbiota
Although few studies have evaluated the relationship between the effect of dietary factors on blood LPS and intestinal flora in humans, several studies have evaluated intestinal flora in oligosaccharide intervention studies (Table 1). A common finding in these reports is an increase in Bifidobacterium. Bifidobacterium has been reported to enhance the intestinal tight junction by preserving claudin 4 and occludin localization at tight junctions, and inhibit permeability in mice with colitis [115]. Similarly, in human colonic epithelial cell line T84, the addition of culture supernatant of Bifidobacterium has been reported to enhance barrier function through increased expression of tight junction protein, suggesting that some humoral factors contribute to improved intestinal barrier function [116]. Increased expression of tight junction protein in Bifidobacterium-treated mice has been reported to be associated with increased short-chain fatty acids (acetic acid, butyric acid, and propionic acid) in the intestinal tract [117]. These short-chain fatty acids have been reported in the human colonic epithelial cell line caco-2 to act as an energy source for epithelial cells to protect themselves, and also act as a histone deacetylase inhibitor which inhibit Nod-like receptor P3 inflammasomes to maintain the barrier function of epithelial cells [118]. These results suggest that the increase in Bifidobacterium induced by oligosaccharide intake decreases blood LPS levels through the improvement of the barrier function of the intestinal tract. In addition, dietary factors that increase Bifidobacterium are expected to reduce blood LPS levels.
Changes in intestinal flora by the dietary factors listed in Table 4 was greatly dependent on the study. However, all of the dietary factors commonly lowered blood LPS or LBP levels in animals, as described in Table 3. In other words, by finding bacteria that have decreased or increased in many dietary factor intervention studies, we can find specific bacteria that contribute to the increase or decrease in blood LPS levels. To this end, we have organized the number of reports that show increases or decreases of each bacterial genus (Figure 1). We selected eight of these genera (Lactobacillus, Bacteroides, Akkermansia, Clostridium, Escherichia, Roseburia, Prevotella and Desulfovibrio) as bacteria included in a sufficient number (five or more) of reports, and a biased number of reports (Bifidobacterium was excluded because it was discussed above. Faecalibacterium was also excluded because there is almost no bias in the number of reports).
Lactobacillus, Bacteroides, Akkermansia, Roseburia, and Prevotella are possible bacterial genera that may contribute to the reduction of blood LPS levels. Lactobacillus is a gram-positive bacterium that produces large amounts of lactic acid during carbohydrate fermentation. The probiotic contribution of Lactobacillus to the regulation of metabolic endotoxemia is studied ( Table 3). Administration of Lactobacillus rhamnosus CNCM I-4036 to obese Zucker-Lepr fa / fa rats decreased the mRNA expression levels of endothelin receptor type B (Ednrb) in the intestinal mucosa, and reduced the blood LBP level [52]. Reduction of Ednrb decreases the density of negative charge of the colonic mucin layer, leading to an increase in the ability of the mucin layer to adsorb microparticles and bacteria, thereby inhibiting their penetration through the colonic mucosa [119]. Lactobacillus sakei OK67 and PK16 are reported to suppress high-fat diet-induced colitis, and to reduce the fecal Proteobacteria population and fecal LPS levels in mice [53]. In addition to the previous reports described in Table 3, it has been reported that oral administration of Lactobacillus reuteri ZJ617 suppresses LPS-induced apoptosis of intestinal epithelial cells and maintains the intestinal barrier function [120]. We have described in Section 2.2 that LPS is absorbed from the intestinal tract during lipid absorption. Interestingly, oral ingestion of Lactobacillus acidophilus ATCC 4356 in mice has been reported to reduce the mRNA levels of Niemann-Pick C1-like 1, which is involved in lipid absorption in the intestine, and in the suppression of cholesterol absorption [121]. Taken together, this suggests that Lactobacillus contributes to a decrease in blood LPS levels through strengthening the intestinal barrier, reducing the amount of LPS in feces, and suppressing lipid absorption. As described in Table 4, Bifidobacterium, oligofructose, galacto-oligosaccharide, syringaresinol, acaudina molpadioides-derived fucosylated chondroitin sulfate, green tea extract, Tartary buckwheat protein, nopal, semen hoveniae extract, and 30% caloric restriction are dietary factors that increase the proportion of Lactobacillus in the gut microbiota. Among them, the amylolytic Bifidobacterium strain is reported to stimulate the growth of a nonamylolytic Lactobacillus probably by producing intermediate metabolites of starch metabolism [122]. Oligosaccharides (oligofructose and galacto-oligosaccharide) were reported to support the growth of Lactobacillus as prebiotics [123]. Green tea extract [124] and buckwheat-resistant starch [125] were reported to promote the growth of Lactobacillus in a fermentation assay. On the other hand, in an in vitro fermentation assay using gut microbiota, it was reported that fucosylated chondroitin sulfate promotes the growth of Bacteroides, Bifidobacterium, and Clostridium, while the number of Lactobacillus decreases [126]. Thus, the mechanism by which Lactobacillus increased in mice fed with fucosylated chondroitin sulfate needs to be further studied. The mechanism by which the proportion of Lactobacillus in gut microbiota increases due to calorie restriction also remains unknown. As it has been reported that the bacteria adapted to the nutritional environment can grow predominantly in the gut microbiota consortium [127], Lactobacillus might be able to grow even under malnutrition. The effect of syringaresinol, nolpal, and semen hoveniae on the growth of Lactobacillus has not been revealed.
Bacteroides is a gram-negative obligate anaerobe. Hooper et al. reported that Bacteroides thetaiotaomicron, a prominent component of the normal mouse and human intestinal microflora, modulates expression of genes involved in mucosal barrier fortification [128]. The administration of Bacteroides fragilis HCK-B3 and Bacteroides ovatus ELH-B2 to mice attenuated LPS-induced intestinal inflammation, by either modulating cytokine production or restoring the Treg/Th-17 balance [129]. On the other hand, in a state in which no dietary fiber is ingested, it has been suggested that Bacteroides degrades the mucin layer of the intestinal tract, decreases the barrier function of mucus, and induces inflammation [130]. Therefore, it should be noted that depending on the diet of the host, Bacteroides can act as either probiotics or pathobionts. As described in Table 4, an increase in Bacteroides was reported in four out of five intervention studies with sulfated polysaccharides. Bacteroides is a unique bacterium among gut flora that has degrading enzymes corresponding to various sulfated polysaccharides [131] and is able to utilize sulfated polysaccharides such as heparin [131], heparan sulfate [131], and chondroitin sulfate [132] as energy sources. It is therefore thought that intake of sulfate polysaccharide preferentially nourishes Bacteroides in gut flora and suppresses metabolic endotoxemia via its anti-inflammatory and barrier function-enhancing effects.
Akkermansia is a mucin-adherent intestinal bacterium [133], which grows by degrading mucin [134], and produces propionic acid, a short-chain fatty acid [135]. In addition, Akkermansia promotes butyrate production, by supporting the growth of Anaerostipes caccae through mucin degradation [136]. As noted above, these short-chain fatty acids are known to enhance intestinal barrier function. In addition, it has been reported that Akkermansia-derived extracellular vesicles administered in mice are localized to the large intestine, and directly enhance intestinal barrier function by increasing epithelial cell expression of tight junction proteins [137]. Furthermore, oral administration of Akkermansia to mice inhibited high-fat diet-induced thinning of the mucin layer, reduced blood LPS concentration, and inhibited obesity and abnormal glucose metabolism [138]. Akkermansia has been reported to be negatively correlated with obesity (waist-to-hip ratio and subcutaneous adipocyte diameter) and diabetes mellitus (glucose intolerance states), and is attracting attention as a next-generation probiotic [139]. Among the dietary factors that increase the proportion of Akkermansia in the gut flora, polyphenols are intriguing because most of intervention studies with polyphenols (apple-derived polymeric procyanidins, genistein, and isoflavone) or polyphenol-rich food extracts (camu camu extract, cranberry extract, and green tea extract) consistently reported an increase of Akkermansia (Table 4). Anhê et al. reported that cranberry extract administration to mice increased colonic Kruppel-like factor 4 (a marker of goblet cells) and Muc2 mRNA expression, suggesting that polyphenols enhance mucin production and support the growth of Akkermansia [95]. On the other hand, direct prebiotic action of polyphenols to Akkermansia has been reported in a study using the Simulator of Human Intestinal Microbial Ecosystem (SHIME ® ) [140].        Lactobacillus, Bacteroides, Akkermansia, Roseburia, and Prevotella are possible bacterial genera that may contribute to the reduction of blood LPS levels. Lactobacillus is a gram-positive bacterium that produces large amounts of lactic acid during carbohydrate fermentation. The probiotic contribution of Lactobacillus to the regulation of metabolic endotoxemia is studied ( Table 3). Administration of Lactobacillus rhamnosus CNCM I-4036 to obese Zucker-Lepr fa / fa rats decreased the mRNA expression levels of endothelin receptor type B (Ednrb) in the intestinal mucosa, and reduced the blood LBP level Roseburia [141] is an enteric bacterium that utilizes dietary fiber and may enhance intestinal barriers by producing butyric acid. It has been reported that administration of Roseburia to mice enhanced differentiation of regulatory T cells in the intestinal lamina propria and suppressed intestinal inflammation [142]. As described in Table 4, oligofructose, apple-derived polymeric procyanidins, sea cucumber-derived sulfated polysaccharide, camu camu extract, and ganoderma lucidum mycelium water extract were reported to increase the proportion of Roseburia in the gut flora. Roseburia metabolizes oligofructose into fructose, which is used for growth, but for this process, acetic acid that is produced by Bifidobacterium is required [143]. Therefore, in order to grow Roseburia by oligofructose intake, it is necessary to pay attention to the symbiotic relationship with other intestinal bacteria and the amount of short-chain fatty acids in the intestine. Other dietary factors, procyanidins, sea cucumber-derived sulfated polysaccharide, camu camu extract, and ganoderma lucidum mycelium, have not been studied for their prebiotic function for Roseburia.
It has been suggested that LPS from Prevotella has fewer phosphate and acyl moieties contributing to endotoxin activity, resulting in a lower TLR4 stimulatory capacity than LPS from Salmonella [144]. Therefore, by increasing the population of Prevotella in the intestinal flora, endotoxin activity in the intestinal contents and damage to intestinal epithelial cells might be decreased, leading to the reduction of blood LPS levels. On the other hand, Prevotella produces succinate as a metabolite of sugar metabolism [145]. It has also been reported that succinate from intestinal bacteria is utilized by and promotes growth of Salmonella serovar Typhimurium [146] and Clostridium difficile [147], which are the pathogens of pseudomembranous colitis. Succinate has also been reported to induce colitis via succinate receptors and to promote colonic fibrosis [148]. In addition, proportion of Prevotella in the gut flora has been reported to be positively correlated with blood LPS levels in patients with type 2 diabetes [149]. Thus, an increase in the proportion of Prevotella does not necessarily have a positive effect on intestinal health. It is necessary to carefully investigate the contribution of Prevotella to blood LPS levels.
Clostridium, Escherichia, and Desulfovibrio are bacterial genera that may contribute to the increase of blood LPS levels. Many pathogenic bacteria (such as enterohemorrhagic Escherichia coli, Clostridium botulinum, Clostridium tetani, and Clostridium perfringens), which produce effector proteins or enterotoxins that disrupt epithelial tight junction belong to these genera [150]. In addition, the endotoxin activity of LPS in non-pathogenic Escherichia is also higher than in Bacteroides, and an increased proportion of these Escherichia in enteric flora aggravate colitis [151]. Clostridium species catabolize cholic acid to deoxycholic acid for their growth [152]. It is reported that, in mice, deoxycholic acid increases intestinal permeability through the reduction of goblet cell number, suppression of mucin production, induction of low-grade inflammation, and suppression of tight junction protein (ZO-1) expression [153]. In terms of dietary factors that reduce Escherichia, there are many reports of sulfated polysaccharides (Table 4). We could not find any reports that suggested the direct inhibitory effect of sulfated polysaccharide on growth of Escherichia. On the other hand, it is suggested that Bacteroides, that can be preferentially grown in sulfated polysaccharide feeding, compete with Escherichia in the co-culture assay [127]. In order to elucidate the mechanism by which sulfated polysaccharides reduce the proportion of Escherichia, it is hoped to study focusing on the interaction between gut microbes. Among the dietary factors that reduce Clostridium, procyanidin is reported to decrease the growth of Clostridium in fecal batch culture [154]. The bactericidal activity of methanol extract of nopal against Clostridium has also been reported [155].
Desulfovibrio is a gram-negative, obligate anaerobe, sulfate-reducing bacterium. Desulfovibrio utilizes electrons supplied by the oxidation of lactic acid in the electron transport system of the respiratory chain, uses sulfuric acid as the final electron acceptor, and produces hydrogen sulfide as a metabolite [156]. Desulfovibrio is ubiquitous in the intestines of humans and mice. Of the studies that showed significant changes in the proportion of Desulfovibrio, most studies reported that the proportion was increased associated to the reduction of blood LPS levels (Table 4). However, it is also reported that proportions of Desulfovibrio increased in the colons of patients with ulcerative colitis [157] and has attracted attention as a pathogen of colitis. In addition, Xie et al. reported in mice that the increase of Desulfovibrio in feces was positively correlated with the increase of LPS levels in feces, liver, and blood [45]. Qui et al. reported that ingestion of a high-fat diet in mice increased fecal Clostridium and Desulfovibrio, and oral administration of these bacteria to the normal chow-fed mice increased fecal and blood LPS levels [158]. These reports suggest that Desulfovibrio plays an important role as a source of LPS in the intestine. Desulfovibrio also competes with Anaerostipes caccae for lactic acid produced by Bifidobacterium, and reduces butyric acid production by inhibiting the growth of Anaerostipes caccae [159]. In addition, as the coexistence of Desulfovibrio and Bifidobacterium inhibits the growth of Bifidobacterium [159], this suggests that the amount of acetic acid produced by Bifidobacterium might be also reduced. On the other hand, it has also been reported that oral administration of Desulfovibrio increases the amount of hydrogen sulfide in the intestinal tract and inhibits intestinal peristalsis [160]. Desulfovibrio is thought to play an important limiting role in increasing blood LPS levels by supplying LPS, decreasing intestinal barrier function due to reduction of short-chain fatty acid content, and prolonging retention time of intestinal contents due to inhibition of peristalsis ( Figure 2). However, despite Desulfovibrio being an important target for metabolic endotoxemia, few dietary factors have been reported to reduce the proportion of Desulfovibrio (Tables 1-4).
that showed significant changes in the proportion of Desulfovibrio, most studies reported that the proportion was increased associated to the reduction of blood LPS levels (Table 4). However, it is also reported that proportions of Desulfovibrio increased in the colons of patients with ulcerative colitis [157] and has attracted attention as a pathogen of colitis. In addition, Xie et al. reported in mice that the increase of Desulfovibrio in feces was positively correlated with the increase of LPS levels in feces, liver, and blood [45]. Qui et al. reported that ingestion of a high-fat diet in mice increased fecal Clostridium and Desulfovibrio, and oral administration of these bacteria to the normal chow-fed mice increased fecal and blood LPS levels [158]. These reports suggest that Desulfovibrio plays an important role as a source of LPS in the intestine. Desulfovibrio also competes with Anaerostipes caccae for lactic acid produced by Bifidobacterium, and reduces butyric acid production by inhibiting the growth of Anaerostipes caccae [159]. In addition, as the coexistence of Desulfovibrio and Bifidobacterium inhibits the growth of Bifidobacterium [159], this suggests that the amount of acetic acid produced by Bifidobacterium might be also reduced. On the other hand, it has also been reported that oral administration of Desulfovibrio increases the amount of hydrogen sulfide in the intestinal tract and inhibits intestinal peristalsis [160]. Desulfovibrio is thought to play an important limiting role in increasing blood LPS levels by supplying LPS, decreasing intestinal barrier function due to reduction of short-chain fatty acid content, and prolonging retention time of intestinal contents due to inhibition of peristalsis ( Figure 2). However, despite Desulfovibrio being an important target for metabolic endotoxemia, few dietary factors have been reported to reduce the proportion of Desulfovibrio (Tables  1-4).  Sulforaphane (1-isothiocyanato-4-methylsulfinylbutane) is an isothiocyanate with an N=C=S functional group and is abundant in broccoli (especially the sprout) and other cruciferous vegetables as the precursor glucoraphanin. Sulforaphane is thought to play a role in plant protection through its antimicrobial action [161], induction of programmed cell death of infected tissue [162], and inhibition of insect feeding [163]. On the other hand, in humans and rodents, sulforaphane activates NF-E2-related factor 2 (NRF2), which induces expression of genes expressing antioxidant and detoxication enzymes, including phase II enzymes, and then exerts anti-cancer [164], anti-liver damage [165], and anti-depressive effects [166].
We found that dietary administration of broccoli sprout extract reduced blood LPS levels and attenuated obesity, glucose intolerance, hepatic steatosis, and inflammation in mice fed a high-fat diet [94]. We also reported that the proportion of Desulfovibrionaceae [upper taxa (family) of Desulfobivrio] was positively correlated with blood LPS levels, and that ingestion of broccoli sprout extract reduced Desulfovibrionaceae in cecal contents. Subsequently, Wu et al. also reported that broccoli powder reduced the proportion of Desulfovibrio in the large intestinal contents of mice [167]. They reported that the decrease in Desulfovibrio composition was negatively correlated with the activity of myrosinase-like activity, isothiocyanate content, and NAD(P)H:quinone dehydrogenase 1 (NQO1) in the colonic mucosa. Ingested glucoraphanin is metabolized by myrosinase-like enzymes in enteric bacteria, which then produce sulforaphane [168]. Since sulforaphane enhances NQO1 activity through activation of NRF2 [168], it is suggested that sulforaphane metabolized and formed from glucoraphanin in broccoli sprouts may have an inhibitory effect on Desulfovibrio (Figure 2). Sulforaphane has been reported to exert antibacterial activity against the Proteobacteria (Desulfovibrio belongs this phyla) [169], but its direct effect on Desulfovibrio is not well understood. It is hoped that the mechanism by which sulforaphane decreases the proportion of Desulfovibrio will be elucidated.

Conclusions
In this article, we summarized previous reports about the regulation of metabolic endotoxemia through dietary factors, focusing on gut microbiota. Although changes in the composition of Firmicutes and Bacteroides due to excessive fat intake have been reported to contribute to metabolic endotoxemia in many reports, the results differ between studies and between species, and further investigation is needed to find true pathobionts. Moreover, since human epidemiological studies have not found a correlation between fat intake and blood LPS levels, it is necessary to search for dietary factors other than fat that cause metabolic endotoxemia. Regarding dietary factors that improve metabolic endotoxemia, human intervention studies have focused on probiotics, prebiotics, polyphenols and dietary habits, and it has been reported that prebiotics, including oligosaccharides, are effective. On the other hand, few studies have evaluated the effects of dietary intervention on gut flora in humans. The development and popularization of next-generation sequencing has made it possible to comprehensively analyze the "fecal" microbiota in humans. On the other hand, as mentioned above, there are also mucin-adherent bacteria that are thought to be involved in metabolic endotoxemia (e.g., Akkermansia and Bacteroides). In a colitis mouse model, it has been reported that the bacterial flora in the mucin layer exhibits changes from 12 weeks before the onset of colitis, and that the mucin layer was thinned [170]. In this study, changes in the fecal flora occurred at the same time as the onset of colitis, indicating that the bacteria in the mucin layer play an important role in understanding the physiological state of the intestinal tract. However, although it is possible to collect mucin layer samples in animals, it is not easy to do so in humans, due to ethical and technical obstacles. In the future, if a method for collecting the mucin layer in a noninvasive manner is established in humans, the research field of metabolic endotoxemia can be further advanced. Then, it is expected that we will comprehensively understand the relationship between dietary factors, dysbiosis, and metabolic endotoxemia in humans by conducting human intervention studies and epidemiological studies with dietary surveys, gut microbiota analysis using next-generation sequencers and evaluation of blood LPS levels.