Non-alcoholic fatty liver disease (NAFLD) and the more severe non-alcoholic steatohepatitis (NASH) represent a spectrum of diseases involving hepatic fat accumulation and histological features essentially identical to alcoholic liver disease; however, they occur in the absence of excessive alcohol intake. They typically arise in conjunction with one or more features of the metabolic syndrome. Lifestyle mediated weight loss remains the primary mode of therapy for NAFLD and NASH, but this is often ineffective and adjunctive medical and surgical treatments are presently lacking. Prebiotic fibres are a group of non-digestible carbohydrates that modulate the human microbiota in a manner that is advantageous to host health. Rodent studies suggest that dietary supplementation with prebiotic fibres positively impacts NAFLD by modifying the gut microbiota, reducing body fat, and improving glucoregulation. Future research should focus on placebo-controlled, human, clinical trials using histological endpoints to address the effects of prebiotics on NAFLD and NASH. The aim of this review is to summarize current knowledge about prebiotics as an emerging therapeutic target for NAFLD.
Prebiotic fibres refer to a group of non-digestible carbohydrates that alter the composition and activity of the gut microbiota . The health effects of prebiotics are likely conveyed to the host via two chief mechanisms that include improved glucoregulation and modified lipid metabolism, and selective modulation of the gut microbiota . There are several different selectively fermented ingredients that qualify as prebiotics including inulin-type fructans and galactans. Other potential candidates exist but lack of sufficient human trials preclude their identification as true prebiotics at present. Accordingly, the focus of this review will be on the most widely studied prebiotics in relation to the metabolic syndrome, the inulin-type fructans, and their potential role in managing non-alcoholic fatty liver disease (NAFLD).
Non-alcoholic fatty liver disease is an umbrella term for a spectrum of diseases involving hepatic fat accumulation and hallmark histopathological features that are essentially identical to alcoholic liver disease but occur in individuals consuming <10 g of alcohol per day . NAFLD encompasses the spectrum of histological abnormalities ranging from simple steatosis that traditionally has a benign clinical course, to non-alcoholic steatohepatitis (NASH) characterized by inflammation, fibrosis and occasionally cirrhosis. NASH may progress to decompensated liver disease and result in liver failure .
In developed countries, it is estimated that 20–30% of the population has NAFLD [3-6]; whereas, the prevalence of NASH in the same population is estimated at 2–3% [6, 7]. The prevalence of NAFLD increases in parallel with obesity, type 2 diabetes and dyslipidemia  and NAFLD is considered an emerging epidemic in light of the dramatic increase in obesity rates. Of considerable concern is the identification of NAFLD in children, with prevalences ranging from 2.6 to 9.6% in Japan and the United States respectively [8, 9]. Furthermore, the prevalence has been reported to be as high as 38% among obese children .
The pathogenesis of NAFLD
The defining feature of NAFLD is hepatic steatosis, which is linked to abnormalities in lipid synthesis and oxidation. The pathogenesis of NAFLD remains unclear, however, several theories have been proposed. In the ‘two hit’ theory, the first hit, steatosis, is the accumulation of fat in hepatocytes. Steatosis is believed to chiefly occur as a consequence of insulin resistance wherein transport of fatty acids is inappropriately shifted to non-adipose tissues including the liver. The second hit is oxidative stress, triggered by pro-inflammatory cytokines that further exacerbate insulin resistance and hepatocyte injury [2, 10]. The variable progression of NAFLD to NASH and cirrhosis has led to the description of the ‘multiple hit’ hypothesis where genetic or environmental susceptibilities allow inflammatory mediators, reactive oxygen species, and abnormal apoptotic mechanisms to serve as additional ‘hits’ that lead to hepatic fibrosis and ultimately cirrhosis in some patients .
Obesity, particularly excess visceral fat, is characterized by dysregulation of inflammatory cytokines. Tumour necrosis factor-α (TNFα) plays a key role in the pathogenesis of NAFLD and NASH through the induction of insulin resistance in animals and humans [12, 13], as well as through increased reactive oxygen species formation and hepatocyte apoptosis in vitro. Increases in other proinflammatory cytokines such as interleukin-6 (Il-6) and interleukin-8, and a reduction in the ‘protective’ anti-inflammatory cytokine, adiponectin, are also anticipated to play a role in NAFLD [15-18]. Furthermore, increased plasma free fatty acids associated with insulin resistance may contribute to hepatocyte apoptosis .
The role of gut microbiota in obesity and NAFLD
There is emerging evidence that the abundant and diverse communities of bacteria that live in the human intestinal tract affect energy metabolism and thus contribute to the pathogenesis of obesity [20, 21]. The interplay between host metabolism and gut microbiota is summarized and illustrated in Fig. 1. Very little is known about the composition of intestinal microbiota in patients with NAFLD or the metabolic or functional roles of gut bacteria relevant to liver steatosis, however, a phylum-wide increase in Firmicutes and a decrease in Bacteroidetes has been described in obesity , although not unanimously . Dysbiosis in the gut microbiota may trigger hepatic de novo lipogenesis by increasing the expression of lipogenic enzymes, acetyl co-A carboxylase (ACC), and fatty acid synthase (FAS) [20, 24]. Conventionalizing germ-free mice resulted in a doubling of hepatic triglyceride content, and a concomitant increase in hepatic mRNA levels of sterol-responsive element-binding protein (SREBP-1), and carbohydrate-responsive element-binding protein (ChREBP), both positive regulators of the aforementioned lipogenic enzymes [20, 25].
Gut microbial communities may be linked to NAFLD via their proinflammatory metabolic by-products. Metabolic endotoxemia, a low-grade inflammatory tone, is increased in patients with NAFLD and NASH [26-28]. A link between the gut microbiota and metabolic endotoxemia is further corroborated by the observation that antibiotic treatment reduces metabolic endotoxemia . Lipopolysaccharide (LPS), derived from Gram-negative bacterial cell walls, has potent effects on host expression of TNF-α and other pro-inflammatory cytokines, and may cross tight junctions of the intestine . Although clinical studies confirm higher levels of LPS in patients with type 2 diabetes , NAFLD and NASH [26-28], further research will be needed to understand the integrity of the intestinal epithelial barrier in NAFLD and its influence on the disease.
Small intestinal bacterial overgrowth is associated with NAFLD and is plausibly linked to its pathogenesis [28, 32-34]. A comprehensive review of the evidence has recently been published , which suggests that intestinal bacteria may be involved in NAFLD via endogenous production of ethanol, and the resultant metabolite acetaldehyde. Chronic delivery of ethanol could promote steatosis and reactive oxygen species culminating in liver injury [10, 33]. Measurable levels of ethanol are present in some patients with intestinal yeast infections or bacterial overgrowth, a condition known as auto-brewery syndrome . Furthermore, obese women have higher breath ethanol concentrations than lean women . More than just ethanol, Probert et al.  have demonstrated the presence of up to 300 different volatile organic compounds (VOC) in human faeces. The mean number of faecal VOC per individual is 154, with 44 compounds being present in at least 80% of healthy individuals and 22 being ubiquitous . The majority of VOC are derived from gut bacterial metabolism, with the remaining compounds coming from host metabolism, diet and environmental sources. Emerging evidence from our group suggests that the VOC profile is altered in obese NAFLD patients, notably with greater representation of ester VOC compared to that in healthy controls (Drs Raman and Rioux unpublished data). The continuous low-grade production of VOC by microbes in the gut, mucosal absorption and delivery to the liver via portal circulation, may have liver toxicity akin to the well described effects of ethanol. Collectively, the evidence to date provides a strong rationale for therapeutic manipulation of gut microbiota as a potential means of treating NAFLD .
Current treatment strategies for NAFLD
A paucity of effective treatments exists for the management of NAFLD and NASH. Several pharmacological and surgical strategies are being investigated [reviewed in Dowman et al. ], however, the current gold standard is weight loss achieved through dietary modification in conjunction with exercise . A recent meta-analysis summarized the outcomes from the handful of randomized clinical trials evaluating the efficacy of various treatments on NAFLD . Weight loss results in histological improvements but adherence to weight management therapies is a persistent concern, given that greater than 50% of the patients did not achieve target weight loss . Thiazolidinediones, ursodeoxycholic acid, lipid lowering drugs, weight loss medications, polyunsaturated fatty acids, metformin and the antioxidant vitamins E and C treatments were also evaluated. Thiazolidinediones, polyunsaturated fatty acids and lipid-lowering drugs may confer some benefits; however, the majority of the trials have been of short duration and did not include histological endpoints . The PIVENS trail was a large randomized controlled trial of vitamin E or pioglitazone for 96 weeks in non-diabetic NASH patients, which demonstrated significant improvements in hepatic steatosis and lobular inflammation with either treatment, but no significant effect on hepatic fibrosis scores . Concerns with vitamin E, however, persist as supplementation has been associated with increased haemorrhagic stroke . Pioglitazone, although promising, did result in modest weight gain . Overall, more large-scale randomized controlled trials, with histological endpoints, are needed to accurately assess the effectiveness of current and novel treatments for NASH.
Known actions of prebiotic fibres
Prebiotics are defined as ‘a non-digestible food ingredient that beneficially affects the host, by selectively stimulating the growth and/or activity of one or a limited number of bacteria in the colon’ . The inulin-type fructans can be differentiated by their degree of polymerization (DP) which refers to the number of repeat units in the polymer chain, specifically D-fructose in the case of fructans . Inulin is a longer chain polymer with an average DP of 12, whereas, oligofructose (OFS) and fructo-oligosaccharides (FOS) are shorter-chain inulin-type fructans with average DPs of 4 and 3–6 respectively. An additional group, the galacto-oligosaccharides (GOS, galactan) is created by enzymatic transgalactosylsation of lactose with an average DP of 2–8; however, only the inulin-type fructans have been extensively studied for improvements in the metabolic syndrome .
Prebiotics stimulate bacterial production of short-chain fatty acids (SCFA), favour the growth of indigenous bifidobacteria and/or lactobacilli, lower luminal pH and thereby impede the growth of pathogens . The ability of prebiotic-containing foods, such as Jerusalem artichokes and chicory, and more often prebiotics given in supplementary form, to modulate gut microbiota is well established in humans, with increases in Bifidobacterium and Lactobacillus spp. consistently reported [48, 49]. There is growing interest in the potential for prebiotics to serve as an effective dietary treatment for NAFLD and numerous promising animal studies support this premise (Table 1).
Table 1. Studies examining the effects of prebiotic fibres on serum lipids and liver physiology in rodents and dogsa
ALT, alanine aminotransferase; BW, body weight; DEX, dexamethasone; DG, D-galactose; DP, degree of polymerization; FOS, fructo-oligosaccharide; HF, high-fat diet; HS, high sucrose diet; n, number of animals; OFS, oligofructose; PB, phenobarbital sodium; SD, standard diet; TC, total cholesterol; TG, triglycerides.
All rodent studies used male animals. Studies in Beagles used two intact males and six neutered females.
Animal studies with prebiotic fibre and proposed mechanisms
Prebiotic fibre supplementation has consistently been shown to reduce plasma lipid and hepatic triglyceride (TG) concentrations in animal studies as summarized in Table 1. The dose of prebiotic utilized in studies is typically 10% on a weight basis although a range of doses (5–20%), and a variety of prebiotics [oligofructose (OFS), inulin, Synergy1®, lactulose] have been examined. In addition to reductions in serum TG, some studies also demonstrate a cholesterol-lowering effect of prebiotics. Several mechanisms have been proposed to explain the beneficial effects of prebiotic fibres on serum lipids and liver TG accumulation observed in animals, including reduced de novo fatty acid synthesis and SCFA production, body weight and fat loss, improved glycemic control, microbial modulation and reduced inflammation.
Reduction in de novo fatty acid synthesis
Fat deposition in hepatocytes originates from a number of sources including fatty acids produced by the liver (de novo lipogenesis), adipocyte-derived non-esterified fatty acids in circulation and dietary fatty acids . The contribution of de novo lipogenesis to hepatic TG is considered low in healthy humans, however, has been reported to account for 26% of liver TG content in NAFLD patients with hyperinsulinemia . Prebiotic-rich diets may ameliorate NAFLD by attenuating de novo fatty acid synthesis as has been shown in animal models [52-55]. Fiordaliso et al.  reported a reduced capacity for free fatty acid esterification to form cellular TG in hepatocytes isolated from animals treated with prebiotics. The decreased lipogenic capacity is thought to be caused by alterations in hepatic gene expression of enzymes that regulate lipogenesis including ACC, malic enzyme, ATP citrate lyase and FAS [52, 53, 57-59]. Conversely, some studies have not reported reductions in FAS, the rate limiting enzyme in de novo fatty acid synthesis, with prebiotic consumption [60, 61]. Comparisons between lean and genetically obese rats on a standard chow diet, without any prebiotic intervention, found elevated FAS expression in the genetically obese rats, suggesting that increased de novo TG production is, in fact, a key determinant of the obese phenotype .
Beyond the enzymatic regulators of lipogenesis, the effects of prebiotics on transcription factors involved in lipid metabolism have also been examined. SREBP 1c and ChREBP were reduced by inulin supplementation in rats fed a cafeteria diet, a finding the authors attribute to a decrease in portal plasma glucose. SREBP-1c is a positive regulator of ACC and FAS genes, and acts to stimulate hepatic fat accumulation . Given the observation that hyperinsulinemia upregulates SREBP-1c [63, 64], it is thought that SREBP-1c may provide a mechanistic link between insulin resistance and lipogenesis. Indeed, it has been postulated that abnormalities in SREBP-1c function may contribute to the development of NAFLD . On the contrary, peroxisome proliferator-activated receptor α, a regulator of several genes involved in lipid catabolism, was not altered by inulin. Similarly, the expression of peroxisome proliferator-activated receptor γ, a nuclear transcription factor that improves insulin sensitivity and promotes adipogenesis, was not affected by adding inulin to a cafeteria diet in rats .
Prebiotics may also attenuate TG accumulation in hepatocytes via alterations to the by-products of their fermentation by microbiota, namely SCFA . Acetate, propionate and butyrate are the major end products of bacterial metabolism in the large intestine. The majority of butyrate is metabolized by colonocytes; however, propionate and acetate are delivered to the liver via the portal vein [67-69]. Prebiotic fibres have been shown to increase the ratio of propionate to acetate [55, 67, 70], an action that could potentially decrease lipogenesis given that propionate inhibits lipogenesis whereas acetate promotes the process . A more recent study assessing the effects of a high-fat and high-sucrose diet on SCFA production noted that rats fed a cafeteria diet for 3 weeks had reduced plasma propionate and butyrate levels. Provision of dietary inulin to cafeteria diet-fed rats, however, restored portal plasma propionate back to control levels .
Although improvements in serum lipids and liver physiology are seen with inulin, OFS and a mixture of the two, there is currently insufficient evidence to suggest that one inulin-type fructan is superior in action to another. Although it is plausible that the DP could influence the effectiveness of select prebiotics, this remains relatively unexplored. One hypothesis is that the shorter-chain FOS and OFS are more metabolically active in the proximal colon, whereas, the longer-chain inulin would be more likely to act in the distal colon . One group has compared synthetic inulin of varying DPs (6–8, 16–17 and 23) and reported that all three forms reduce liver TG and cholesterol compared to a cafeteria diet, but only DP 16–17 reduced portal glucose . Although the majority of animal studies use a 10% dose by weight with a minimum 4 week supplementation period, Sugatani et al.  did report reductions in liver TG and cholesterol in cafeteria diet-fed rats supplemented with 5% synthetic inulin for 3 weeks. In comparison, we have recently reported improvements in serum and liver lipids with prebiotic doses as high as 20% . At present, the minimum dose and necessary duration has not been fully defined, highlighting an area for future exploration. Importantly, continued ingestion of the prebiotic is likely to be necessary to maintain changes.
Improvements in obesity and insulin resistance
Given the multitude of factors that influence hepatic lipid accumulation, consideration of the effects of prebiotics on major risk factors for NAFLD, including obesity and insulin resistance, is warranted. Prebiotic fibre supplementation is associated with reduced body weight or attenuated weight gain in lean [71, 72], high-fructose fed , high-fat, high-sucrose fed [54, 59, 73] and genetically obese [60, 67] rodent models. Other studies, however, have not been able to demonstrate weight loss with prebiotics including Fiordaliso et al. , and a more recent study in leptin-receptor deficient rats fed a combination of inulin and oligofructose . In addition, studies by Sugatani et al. [54, 59] do not report weight loss when the background diet is of normal energy density (standard diet), but do see weight loss when a high-fat, high-sucrose diet is supplemented with inulin. Traditionally, the beneficial effects of prebiotics on body weight have been attributed to increased satiety through stimulation of the anorexigenic hormones glucagon-like peptide-1 and peptide YY, and suppression of the orexigenic hormone ghrelin [71, 72, 75]. An additional benefit of prebiotic fibres with respect to weight control is their low energy density. Inulin-type fructans supply approximately 6.3 kJ/g of energy as opposed to digestible carbohydrates that provide 16.7 kJ/g . It has been suggested that rats demonstrate hyperphagia in response to the dilution of energy in the diet with cellulose ; although no difference has also been reported . Prebiotic supplementation, with a 50:50 blend of inulin and OFS, lowered the energy content of the diet without increasing food intake, in leptin-receptor deficient rats; suggesting it may not trigger hyperphagia to compensate for dilution of energy in the diet .
The gut microbial community and metabolic endotoxemia
A defining characteristic of prebiotic fibres is their ability to alter the gut microbiota in a manner that is advantageous to the host . This property of prebiotics is appealing for the management of NAFLD, particularly in light of research demonstrating that small bowel bacterial overgrowth and dysbiosis of the gut microbiota may play a role in the development of NAFLD [32-34] and obesity [22, 79]. Supporting the suggestion that manipulation of the gut microbiota modifies the course of inflammatory diseases, administration of two strains of Lactobacillus was found to reduce the proinflammatory cytokines, IL-6 and TNF α, in the colon . Both IL-6 and TNF α have been shown to be elevated in NAFLD and contribute to increased hepatic lipogenesis . The gut microbiota may also propagate obesity-associated metabolic disorders through an increase in LPS, a trigger for systemic release of proinflammatory cytokines. Indeed, the reduced numbers of Bifidobacteria that are commonly reported in obesity are associated with increased LPS . Prebiotic fibres have a bifidogenic effect and are associated with reduced LPS levels ; therefore, they are a promising target for correcting detrimental microbiota-host interactions. Prebiotic-induced changes in gut microbiota also influence production of the gut trophic hormone, glucagon-like peptide-2, which could potentially modulate lipid and LPS absorption via its effects on intestinal permeability and epithelial tight junctions .
Given that prebiotic fibres modify the gut microbial community and that VOC are a metabolic by-product of the gut bacteria, it is plausible that prebiotic fibres may impact health through alterations in VOC. This is an emerging area of research; however, studies by De Preter et al.  using an in vitro human faecal model have shown that OFS-enriched inulin increased the concentration and numbers of esters and inhibited sulphur-containing compounds. Additional research is required to fully understand the impact of prebiotic fibre supplementation on VOC and their impact on the hepatocytes.
Human studies with prebiotic fibre
Studies to assess the effects of prebiotic fibre supplementation on NAFLD in humans are lacking, however, the effects of prebiotic fibre supplementation on serum lipids have been assessed in several human trials. Of five interventional studies, two reported a decrease in serum triglycerides (16 and 27% reduction) or cholesterol (7% reduction) following inulin supplementation in healthy participants . In patients with type 2 diabetes or hyperlipidemia, improvements were more robust, with reductions in either cholesterol (6–20% reduction) or TG (14–27% reduction) or both reported in six of eight studies where participants consumed either inulin or OFS . One explanation for the discrepancy, in consistent lipid-lowering effects with inulin and OFS supplementation, between human and animal studies could be the relatively lower dose administered to humans. Humans also have a lower rate of hepatic de novo lipogenesis than that of rats , although elevated lipogenesis in patients with NAFLD may afford them greater relative benefit from prebiotics . Indeed, the enhanced lipid-lowering effects of prebiotics, in humans with obesity and hypertriglyceridemia, may be linked to an inherently greater magnitude of hepatic lipogenesis [85, 86]. Prebiotic fibres have also shown promise in treating the risk factors associated with NAFLD but effects on the disease per se have been minimally examined. Specifically, Parnell and Reimer  reported modest but significant weight loss following 3 months of oligofructose supplementation in overweight and obese adults. Self-reported energy intake was reduced alongside improvements in glycemia and modifications to plasma glucagon-like peptide-1, peptide YY and ghrelin. To our knowledge, only one pilot study specifically evaluating prebiotic fibres in NAFLD has been performed. Seven patients with NASH were supplemented with 16 g/d of OFS for 8 weeks. The authors reported reductions in aspartate aminotransferase and insulin levels but no definitive histological measures were made . Large-scale human studies of sufficient duration, with clinically relevant endpoints, such as liver biopsy, are required to better understand the potential to translate the positive effects of prebiotic fibres noted in animals to human clinical application.
Prebiotic fibre safety
Prebiotic fibres are found naturally in many common foods including asparagus, garlic, leeks, onions and chicory root . Average daily intake is estimated at 1–4 g in the United States and 3–11 g in Europe . Recently, prebiotic fibres have been added as ingredients to many common food products such as bread, cereal bars and ready-to-eat breakfast cereals. Prebiotics have generally recognized as safe (GRAS) status in the United States and are officially recognized as natural food ingredients in most European countries . There are few concerns with supplementation, however, Kaur et al.  report that doses of prebiotic fibre in excess of 30 g/d may cause adverse gastrointestinal effects primarily flatulence.
Animal studies provide promising evidence of the potential for prebiotic fibres to alter the course of NAFLD. There is currently an unmet need in terms of human clinical data available to advance the dietary management of NAFLD. Of particular interest, is the use of prebiotics in patients with NASH where hepatocellular steatosis, inflammation and degeneration can lead to liver fibrosis, dysfunction and malignancy. Future research should focus on adequately powered, randomized, placebo-controlled, human clinical trials to determine the effect of prebiotic fibres on composite histological markers of NASH. Finally, further research into potential mechanisms is required to simultaneously assess the effects of prebiotic fibres on gut microbial composition and function in patients with obesity and NAFLD, as well as host energy balance, inflammatory cytokines and regulators of metabolism.
Author contributions: J. A. P. wrote the first draft of the article and formatted the final draft. K. P. R., M. R., and R. A. R. all contributed to editing the article. R. A. R. had primary responsibility for final content. All authors read and approved the final manuscript.
Conflict of interest: J. A. Parnell and M. Raman declare no conflicts of interest. K. Rioux and R. Reimer have participated in prebiotics advisory boards for Abbott Laboratories. R. A. Reimer holds a research contract with BENEO-Orafti Inc., a manufacturer of prebiotic fibre.
Funding source: Supported in part by research grants from the Canadian Institutes of Health Research (RAR), Mount Royal University (JAP, RAR), and Canadian Institutes of Health Research – Human Microbiome Catalyst Grant (KPR, MR).