Dietary lipids, gut microbiota and lipid metabolism

Short-chain fatty acids (SCFAs) such as acetate, propionate and butyrate are bacterial metabolites derived from fermentation of fibers in the colon (Fig. 2). Both butyrate and propionate have low systemic concentrations whereas acetate levels are higher [49]. SCFAs are important for host metabolism and are used as substrates for energy production, lipogenesis, gluconeogenesis and cholesterol synthesis [50, 51]. Butyrate is an energy source for colonocytes while propionate is mainly metabolized by the liver.

In addition to being metabolic substrates, SCFAs act as signaling molecules, notably through the G-protein coupled receptors GPR43/FFAR2 and GPR41/FFAR3. GPR43 protects against diet-induced-obesity in mice [52‐55]. Activation of GPR43 on L-cells increases secretion of glucagon-like peptide-1 (GLP-1) [52, 56] and acetate induces anti-lipolytic activity [57] and improves glucose and lipid metabolism [53] through GPR43 in WAT. GRP41 has also been shown to regulate metabolism through interaction with the gut microbiota. CONV-R Gpr41 knockout mice are leaner and weigh less than their wild-type littermates, while these differences are not found in GF mice. Furthermore, the microbiota increases peptide YY (PYY) production through GPR41 [58]. Butyrate and propionate have also been shown to activate PPARγ [59], and SCFA-induced activation of PPARγ modulates lipid metabolism through increased energy expenditure [60], reduced body weight and decreased liver triglyceride accumulation [61].

Overall, SCFAs have been shown to have a positive impact on metabolic health [62]. Supplementation with acetate reduces weight gain and improves glucose tolerance in obese and diabetic rats [63], butyrate protects against obesity and increases thermogenesis in mice [60] and propionate or butyrate improves glucose homeostasis in mice [64]. Some minor clinical trials have also found beneficial effects of SCFA or fiber supplementation on body weight [56, 65]. Ingestion of a propionate precursor increases postprandial plasma PYY and GLP-1 and reduces energy intake while long-term treatment results in a reduction in weight gain [56]. Plasma concentrations of PYY and GLP-1 are also increased by acetate in humans [66].

Primary bile acids are synthesized from cholesterol and conjugated to taurine or glycine in the liver. The bile acids are stored in the gallbladder and excreted into the duodenum after food ingestion to aid emulsification of dietary lipids. Most of the bile acids are reabsorbed and recirculated to the liver, but bacterially mediated deconjugation of the glycine or taurine group reduces reabsorption. Deconjugated bile acids can be further metabolized to secondary bile acids through dehydrogenation, dehydroxylation and epimerization by colonic bacteria [67]. Microbial processing results in a more hydrophobic bile acid pool and facilitates excretion in the feces. Fecal excretion of bile acids is a major sink for cholesterol and bile acids lost in the process need to be replaced by de novo synthesis from cholesterol [67].

In addition to their role in lipid digestion, bile acids can act as signaling molecules that regulate host metabolism by binding to the nuclear receptor farnesoid X receptor (FXR) and the Takeda G-protein coupled bile acid receptor TGR5. Microbial processing of bile acids increases the diversity of the bile acid pool and the different bile acids vary in their affinity to the receptors and can act as agonists or antagonists. Both of the primary bile acids cholic acid (CA) and chenodeoxycholic acid (CDCA) and the secondary bile acids lithocholic acid (LCA) and deoxycholic acid (DCA) are FXR agonists, but with different affinities [68]. In humans, CDCA can be transformed into ursodeoxycholic acid (UDCA), which is a FXR antagonist [69]. Furthermore, the taurine-conjugated murine bile acid TβMCA, but not its deconjugated counterpart βMCA, is a potent FXR antagonist [70].

FXR is involved in the regulation of lipid metabolism, especially triglyceride trafficking, synthesis and utilization [71]. Microbial processing of bile acids may therefore influence lipid metabolism through interaction with FXR. By feeding wild-type and Fxr knockout mice with or without bacteria a high-fat diet, Parséus et al showed that microbiota-induced weight gain, steatosis and inflammation were dependent on FXR signaling [72]. FXR also changed the gut microbiota composition, and transplantation of the gut microbiota into GF mice transferred the lean phenotype of Fxr knockout donor mice, demonstrating that FXR may contribute to increased adiposity by altering the microbiota composition. Comparison of whole body and tissue-specific Fxr knockout mice have revealed that activation of the liver and intestinal FXR result in distinct metabolic outcomes in obesity models [73‐77]. Several studies indicate that inhibition of intestinal FXR improves metabolic phenotypes [75, 77] but the underlying mechanisms are still unknown.

Bile acids have also been shown to influence host lipid metabolism through TGR5. TGR5 activation in skeletal muscle and brown adipose tissue promotes energy expenditure [78]. In addition, TGR5 signaling induces GLP-1 release from enteroendocrine L-cells, resulting in improved liver and pancreatic function in obese mice [79], with potential influences on lipid synthesis and storage. The microbially produced bile acids LCA and DCA act as agonists to TGR5 [78, 80] but the impact of the gut microbiota on host metabolism via TGR5 remains to be determined.

Bile acids have been implicated in the pathogenesis of fatty liver disease. Patients with NASH have been shown to have altered fecal bile acid composition [81]. In addition, an inverse relation between fibroblast growth factor 19 (FGF19), an FXR-regulated hormone produced in the ileum, and NASH has been reported [82, 83] . The importance of understanding the interplay between the gut microbiota, bile acids and lipid homeostasis is highlighted by efforts to use bile acids as treatments for NAFLD and NASH. One example is UDCA, which has been shown to have beneficial impact on steatosis and serum lipid levels after short-term treatment in severely obese patients [69] while others have reported negative results regarding improvement of NASH in response to UDCA treatment [84, 85]. Another recent example is the semi-synthetic FXR agonist obeticholic acid, which has been shown to improve NASH after 72 weeks’ treatment in a randomized, controlled clinical trial [86]. However, accompanying elevations in serum LDL levels have raised the question of the overall benefit of such treatment. Short-term treatment with the FGF19 analog NGM282 resulted in reduction of steatosis in NASH patients [87]. To date, no clinical study targeting the gut microbiota to specifically modify FXR signaling in NAFLD or NASH has been performed.

Lipopolysaccharides (LPS), also known as endotoxins, are structural compounds in the outer membrane of Gram-negative bacteria. LPS induces inflammation through activation of TLR4, which is expressed on immune cells such as macrophages as well as on many other cell types including hepatocytes and adipocytes. The intestinal epithelium works as a barrier to prevent translocation of bacterially derived factors. However, weight gain, high-fat diet [88] and increased exposure of fatty acids [89, 90] may disrupt the gut barrier function allowing translocation of LPS [91, 92]. This results in moderately increased levels of LPS in the blood which is defined as metabolic endotoxemia [91], a condition linked to metabolic perturbations such as dyslipidemia, insulin resistance, NAFLD and cardiovascular disease [93].

LPS interacts with blood lipids in various ways. First, it increases the concentration of blood triglycerides by multiple mechanisms. In rats, low-dose LPS increases hepatic synthesis of VLDL, whereas high-dose LPS decreases lipoprotein catabolism [94]. Mice lacking the TLR4 co-receptor CD14 are resistant to hyperinsulinemia, insulin resistance and steatosis induced by a high-fat diet or LPS [91]. Second, plasma lipoproteins, in particular HDL, have the ability to neutralize the toxic effects of LPS [95, 96]. The capacity of HDL to bind LPS may protect against inflammation. This is supported by the observations that infusion of HDL prior to a LPS challenge reduced release of proinflammatory cytokines in humans [97] and that LPS induces higher levels of TNFα in hypolipidemic rats compared with controls [98].

LPS has been shown to promote atherosclerosis and cardiovascular disease. LPS-treated hypercholesterolemic rabbits have increased atherosclerosis compared with controls [99] and mice lacking TLR4 have reduced atherosclerosis and plaques with decreased amounts of lipid [100]. In humans, high LPS levels during chronic infections are predictors of increased atherosclerotic risk [101] and metabolic endotoxemia increases the risk for cardiovascular disease and mortality in patients with chronic kidney disease [102]. Human TLR4 mutations have been shown to be associated with a decreased response to LPS [103], reduced risk of carotid artery atherosclerosis [104] and acute coronary events [105].

Mouse studies have shown that hepatic steatosis is induced by a high-fat diet and associated with dysbiosis and increased intestinal permeability [91]. Moreover, chemically induced colitis in rats increases the levels of circulating LPS and worsens steatohepatitis during high-fat diet[106]. Dysbiosis-induced permeability increases the levels of TLR ligands in the portal vein, thereby activating hepatic Kupffer cells and stellate cells to stimulate pro-inflammatory and pro-fibrotic pathways via inflammatory cytokines[107, 108]. In addition, mucosal TLR activation appears to contribute to hepatic steatosis via the TLR adaptor MYD88 expressed in the intestine[109]. Mice with an intestinal epithelial cell-specific deletion of MYD88 fed a high-fat diet have improved glucose homeostasis and decreased hepatic lipid content compared with wild-type mice[110]. TLR signaling in the mucosa can also induce production of inflammasomes, multiprotein oligomers responsible for the activation of inflammatory responses. Inflammasomes activate a variety of pro-inflammatory and pro-fibrotic processes involved in the progression of liver disease[111]. For example, activation of the NLRP3-inflammasome by LPS via TLR4 and TLR9 is involved in the development of fibrosis in NAFLD[112]. Gut permeability has also been linked to NAFLD in humans. NAFLD patients have been shown to have increased gut permeability compared to healthy controls and gut permeability correlated with severity of steatosis but not with steatohepatitis in patients with NAFLD[113]. Interestingly, patients with steatosis have also been shown to have a gut microbiota with increased genetic potential for biosynthesis of endotoxin[46].

The gut microbiota metabolizes methylamine-containing nutrients such as choline, lecithin and L-carnitine to generate trimethylamine (TMA), which is further processed to trimethylamine N-oxide (TMAO) by flavin monooxygenases (FMO) in the liver. TMAO levels have been correlated with risk of cardiovascular events[114] and prevalence of cardiovascular disease[115, 116]. Plasma TMAO levels in different mouse strains have been positively correlated with lesion size [117] and transfer of microbiota from high- and low-TMAO-producing mice to atherosclerosis-prone Apoe knockout mice show that increased microbial capacity for TMA production increases aortic lesions[118].

FMO3 is the primary enzyme converting TMA into TMAO. Knockdown of Fmo3 results in reduced atherosclerotic lesion areas, altered lipid and cholesterol metabolism, and decreased TMAO plasma levels [119, 120]. FMO3 expression is regulated by bile acids by a mechanism that involves FXR [121]. Gut microbiota processing of bile acids could therefore be an alternative mechanism by which the gut microbiota regulates TMAO production.

The mechanisms by which TMAO contributes to atherosclerosis appears to be complex and not fully understood. Antibiotic treatment reduces production of TMA and has been shown to suppress foam cell formation. TMAO can also contribute to atherosclerosis by inhibiting reverse cholesterol transport [115] and by inducing atherosclerosis-promoting inflammatory proteins in vascular cells[122]. In addition to atherosclerosis, increased levels of TMAO are also associated with a frequency of thrombotic events and platelet activation[123].

Although many studies have reported associations between plasma levels of choline, TMAO, and cardiovascular disease [114, 116, 124, 125] the role of TMAO in disease development is still under debate. A recent study comparing conventional and GF Apoe knockout mice fed diets with or without choline supplementation found no effect of choline enrichment on aortic root atherosclerosis in mice[126]. TMAO production was dependent on the presence of intestinal bacteria but no relationship between TMAO levels and lesions was observed. These results are in contrast to Wang et al who found a significant correlation between TMAO and lesion size[116]. Similarly, studies using diets supplemented with L-carnitine, which result in increased levels of TMAO, have shown opposite effects on aortic lesions in two different laboratories, further emphasizing context dependency on experimental outcome[115, 127].